<%BANNER%>

Bioanalytical Applications of Affinity-Based Nanotube Membranes for Sensing and Separations

Permanent Link: http://ufdc.ufl.edu/UFE0022016/00001

Material Information

Title: Bioanalytical Applications of Affinity-Based Nanotube Membranes for Sensing and Separations
Physical Description: 1 online resource (212 p.)
Language: english
Creator: Caicedo, Hector
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: affinity, alumina, antibodies, bioanalytical, dna, fluorescence, flux, membrane, nanometer, nanopore, nanotechnology, nanotube, partition, permeability, protein, sensing, separations, synthesis, template
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Nanotechnology has played an important role in the development of research and technology during the last two decades. The contribution of nanotechnology in different fields, along with the versatility of the constructed nanoscale materials, have made nanotechnology one of the most suitable tools to develop particular nanostructures to realize a desired function and application. A nanostructure is simply an entity at the nanometer scale with one, two or three dimensional features. Since nanotechnology covers a broad range of nanoscale materials, to simplify nanotechnology, it can be classified into two categories based on how the nanostructures are prepared: top-down and bottom-up. In the top-down methods, the nanostructures are constructed by chiseling larger bulk materials into entities of smaller size. Conversely, in the bottom-up case, small units are grown or assembled into their desired size and shape. The nanoporous materials specifically have attracted a lot of attention because they can be used for the synthesis of a variety of functional nanostructures of great usefulness in technology. These porous nanostructures usually combine many of the advantages of the top-down and bottom-up methodologies such as flexibility, size controllability, and cost. The research presented in this work utilizes nanoporous membranes to develop porous nanostructured platforms with potential applications in sensing and separations. In particular, this work is centered in fundamental studies for bioanalytical applications of affinity-based nanotube membranes for sensing and separations. A bottom-up methodology like the template synthesis was used to produce silica nanotubes inside of the pores of alumina membrane. The functionalization of the inside walls of these silica nanotube membranes allowed control of the functional behavior and properties of the nanostructured membrane during membrane-based separations and sensing. The general scheme of the work presented here, is distributed in seven chapters. The first chapter consists of a general description of the theory and fundamentals as background supporting evidence for the work presented here. The template synthesis method, the fabrication of the porous alumina membranes and the possibilities that silane chemistry offers in order to functionalize the nanostructured membranes is discussed. In addition, a brief overview of the fundamentals of biological production of proteins, antibodies, and recombinant proteins is described. After this introduction, the studies presented herein are centered in potential bioanalytical applications of the silica nanotube membranes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hector Caicedo.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Martin, Charles R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022016:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022016/00001

Material Information

Title: Bioanalytical Applications of Affinity-Based Nanotube Membranes for Sensing and Separations
Physical Description: 1 online resource (212 p.)
Language: english
Creator: Caicedo, Hector
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: affinity, alumina, antibodies, bioanalytical, dna, fluorescence, flux, membrane, nanometer, nanopore, nanotechnology, nanotube, partition, permeability, protein, sensing, separations, synthesis, template
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Nanotechnology has played an important role in the development of research and technology during the last two decades. The contribution of nanotechnology in different fields, along with the versatility of the constructed nanoscale materials, have made nanotechnology one of the most suitable tools to develop particular nanostructures to realize a desired function and application. A nanostructure is simply an entity at the nanometer scale with one, two or three dimensional features. Since nanotechnology covers a broad range of nanoscale materials, to simplify nanotechnology, it can be classified into two categories based on how the nanostructures are prepared: top-down and bottom-up. In the top-down methods, the nanostructures are constructed by chiseling larger bulk materials into entities of smaller size. Conversely, in the bottom-up case, small units are grown or assembled into their desired size and shape. The nanoporous materials specifically have attracted a lot of attention because they can be used for the synthesis of a variety of functional nanostructures of great usefulness in technology. These porous nanostructures usually combine many of the advantages of the top-down and bottom-up methodologies such as flexibility, size controllability, and cost. The research presented in this work utilizes nanoporous membranes to develop porous nanostructured platforms with potential applications in sensing and separations. In particular, this work is centered in fundamental studies for bioanalytical applications of affinity-based nanotube membranes for sensing and separations. A bottom-up methodology like the template synthesis was used to produce silica nanotubes inside of the pores of alumina membrane. The functionalization of the inside walls of these silica nanotube membranes allowed control of the functional behavior and properties of the nanostructured membrane during membrane-based separations and sensing. The general scheme of the work presented here, is distributed in seven chapters. The first chapter consists of a general description of the theory and fundamentals as background supporting evidence for the work presented here. The template synthesis method, the fabrication of the porous alumina membranes and the possibilities that silane chemistry offers in order to functionalize the nanostructured membranes is discussed. In addition, a brief overview of the fundamentals of biological production of proteins, antibodies, and recombinant proteins is described. After this introduction, the studies presented herein are centered in potential bioanalytical applications of the silica nanotube membranes.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Hector Caicedo.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Martin, Charles R.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022016:00001


This item has the following downloads:


Full Text





BIOANALYTICAL APPLICATIONS OF AFFINITY-BASED NANOTUBE MEMBRANES
FOR SENSING AND SEPARATIONS




















By

HECTOR MARIO CAICEDO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2008


































2008 Hector Mario Caicedo
































To my parents Herlinda and Luis, my wife Jessica, and my daughter Isabela









ACKNOWLEDGMENTS

I would like to thank my Ph.D. advisor, Dr. Charles R. Martin for supporting my work, and

for all his lessons on becoming a more effective scientist.

I would like to express a very special gratitude to Dr. Jim Deyrup for his tremendous help

and encouragement when I started pursuing my Ph.D.

I would also like to thank Dr. Nicolo Omenetto for being part of my committee, for his

invaluable time and discussions about my research work, and for his contagious enthusiasm for

science.

I would like to thank Dr. Amelia Dempere, Dr. David Powell, and Dr. Thomas Lyons for

serving on my graduate committee and for their time, feedback, and willingness to proofread this

dissertation.

I am very grateful to Dr. Myungchan Kang, former postdoc in the Martin group, with who

I had the opportunity to work with for two years, during which he always provided many helpful

insights about my research work.

I would like to thank my lovely wife Jessica and my beautiful daughter Isabela for their

love, patience, and continuous support during the last years of my graduate studies.

Finally, my greatest gratitude goes to my family, for their encouragement in pursuit of my

educational goals.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ............... .............................. ............................................................4

LIST OF TA BLES ....................... ....................................................... 9

LIST OF FIGURES .................................. .. .... ..... ................. 10

ABSTRAC T ................................................... ............... 17

CHAPTER

1 INTRODUCTION AND BACKGROUND ........................................ ....................... 19

In tro d u ctio n ................... ...................1...................9..........
B ack g rou n d ...................................... ................................................................. ............... 2 0
The Concept of the Template Synthesis.......................................................................20
Description of the Nanoporous Alumina Membrane Template ................................25
Tw o-step anodization m ethod ............................................................................ 27
Pore grow th m echanism ........................................................................ 30
Sol-G el C hem istry at a G lance ................................................................................... 32
Silane Chemistry a Versatile Tool for Surface Functionalization........................ ...35
Nature's Transport across Membranes: An Inspiration for Membrane-Based
Separations ............................................................................................36
Fundam entals about Proteins................................................. .............................. 41
A n tib o d ie s ................................................................4 2
R ecom binant proteins ................. ........ ............ ....... .. .............. ........ 44
Synopsis and Analytical Visualization of the Instrumental Techniques.........................46
Fluorescence spectroscopy ......................................................... ..................... 46
Fluorescence quenching ............................................................. ............... 48
FTIR -A TR spectroscopy ........................................................... ............... 49
X-ray Photoelectron Spectroscopy (XPS).....................................................51
Scanning Electron M icroscopy (SEM ) ....................................... ............... 52
Im pedance spectroscopy ................................................ .............................. 53

2 TRANSPORT AND FUNCTIONAL BEHAVIOR OF ANTIBODY-
FUNCTIONALIZED NANOTUBE MEMBRANES ......................................................55

In tro d u ctio n ................... ...................5...................5..........
E x p e rim e n ta l ..................................................................................................................... 5 7
M materials ............................................................................................. 57
Preparation of the Silica Nanotube Membranes............................................................58
A antibody Im m mobilization ....................................................................... ................... 58
T ran sport E xperim ents .......................................................................... .................... 59
C om petitive B finding A ssay s ........................................ ............................................60









Determination of Static Protein Binding Capacity and Binding Constants through
A dsorption Isotherm s ................... ... .......... .... ... .. .... ............ ................... 6 1
Effect of Equilibration Time on Flux and Selectivity via Dynamic Binding
C a p a c ity ..............................................................................6 2
R results and D discussion ......................... ..................................... ............. ..............62
Electron M icroscopy and Calibration Data................................................................ 63
Transport Properties of the anti-IgGs through PEG-Si-modified Control
M e m b ra n e s ............................................................................................................. 6 4
C om petitive B finding Studies........................... .................................................... 65
Functional Behavior of the M-IgG- and H-IgG-modified Membranes...........................67
Observed Functional Behavior is not a Transient Phenomenon ...................................68
Flux vs. Feed C concentration ........................ ........... ......................................... 70
Effect of the Pore Size on the Flux and Selectivity Coefficient.............................. 71
Effect of the Immobilized-Antibody Orientation on the Flux .......................................73
Separation Based on a Simultaneous Combination of Size and Affinity Selectivity......74
Binding Capacities and Binding Constants.................................................................. 77
Effect of the Temperature on the Flux and Binding Constant............... ... ...............78
Effect of the Low Ionic Strength on the Selectivity ..................................................83
C onclu sions.......... ............................... ................................................85

3 MODELING AND CHARACTERIZATION OF THE TRANSPORT PROPERTIES
OF PROTEINS ACROSS ANTIBODY-FUNCTIONALIZED NANOTUBE
MEMBRANES BY USING THE DUAL-MODE MODEL ...............................................87

Introduction ............................... ...... ... ......................................87
D definition of D ual-M ode M odel...................... ...... ...................................... ............... 89
R results and D iscu ssion ................ .. .. .. .... .... .... ... ......... ................................95
Simulations of Permeability vs. Feed Concentration Based on the Dual-Mode
M odel .............................. ............. .................... ............ 95
Experimental Description of the Dual-M ode M odel.................................................... 100
Effect of feed concentration on the flux........................................... ...........100
Experimental permeabilities as function of the feed concentration .......................101
Experimental evaluation of the dual-mode model parameters............................. 102
Experimental evaluation of Csites and b ...................................... ...............103
Experimental evaluation of the partition coefficients k, ..................................107
Experimental evaluation of the diffusion coefficients and solubility ....................110
Fitting Experimental Data with the Theoretical Dual-Mode Model ...........................114
Validation of the Dual-M ode M odel ....................... ........................... .................... 119
C onclu sions.......... .............................. ...............................................12 1

4 METAL ION AFFINITY NANOTUBE MEMBRANES: PREPARATION,
CHARACTERIZATION, AND CONTROLLED TRANSPORT OF SIX-HISTIDINE-
TAGGED RECOM BINAN T PROTEIN ................................................... ..................... 122

Introdu action ................... ......................................................................... 122
E x p erim mental ............................................................................... 124
M a te ria ls .........................................................................................................1 2 4









Preparation of the Silica Nanotube Membranes................. .....................................125
Preparation of Imidazole Functionalized Membranes............................125
D ivalent M etal Ions Com plexation ............................. ............ ............... .... 126
Instrumental Characterization of the Imidazole Functionalized Membranes..............127
T ran sport E xperim ents ......................................................................... ................... 12 8
R results and D discussion ..................................... ............................. .... .. .. .. ...... ... 129
Imidazole Surface Modification and Characterization.................. ...............129
Fluorescence microscopy characterization............................................ 130
Characterization by FTIR-ATR spectroscopy................................ .................. 135
Surface chemical composition by XPS ............ .............................................141
Transport Properties of Imidazole M embranes ............................................................146
C onclu sions.......... .............................. ...............................................149

5 TRANSPORT ANALYSIS OF PROTEINS ACROSS SILICA NANOTUBE
MEMBRANES THROUGH A FLUORESCENCE QUENCHING-BASED SENSING
A P P R O A C H .............................................................................................. 15 1

In tro d u ctio n ................... ...................1.............................1
E x p erim mental ............................................................................... 152
M ateria ls .................. ............................................................................................... 1 5 2
Preparation of the Silica N anotube M embranes...........................................................154
A antibody Im m mobilization .................................................... .............................. 154
Labeling of anti-H-IgG with QSY-7 ................................. 155
T ran sport E xperim ents ......................................................................... ................... 156
R results and D discussion ........... ....... ....... ....... .. ..... .. .. ... .................. ........ .........157
Experimental Principle Used for the Fluorescence Quenching Detection ....................157
Verification of the Fluorescence Quenching System via H-IgG-FITC and anti-H-
IgG -Q SY R action in Solid State ................................................... .....................159
Control Transport Studies................................................... ..... .... ...................... 160
Effect of the Feed Solution Concentration on the Rate of Quenching ........................162
Data Analysis Based on the Stern-Volmer Plots................................... ... ..................164
Data Analysis Based on log-log Plots of Fluorescence vs. Time..............................167
Transport across Antibody-Modified Silica Nanotube Membranes..............................169
C onclu sions.......... .............................. ...............................................170

6 HYBRIDIZATION STUDIES OF DNA IN A PNA-MODIFIED NANOTUBE
M EM B RAN E PLA TFORM ........................................................................... .............171

In tro d u ctio n ................... ...................1.............................1
E x p erim mental ............................................................................... 173
M ateria ls .................. ............................................................................................... 17 3
Preparation of the Silica N anotube M embranes...........................................................174
Preparation of PNA-Modified Nanotube Membranes ............................................... 174
N anotube M em brane A ssem bly .................... ...... ................... ............... .... 176
Impedance Measurement for Membrane Characterization .......................................177
Im pedance D ata A analysis ......................................................................... .. 177
R results and D discussion ..................................... ........... .......... .. ........ .... 178









Impedance Characterization of the Unmodified Membranes ............. .... ...............179
Functionalization of the Silica nanotube Membranes ........... ......... ............... 180
Impedance Sensing of DNA-PNA Hybridization .............. ...................182
Effect of the Pore Size on R ................................................................................... 183
Stability and Reproducibility of the PNA-modified Membranes .............................. 185
C onclu sions.......... .............................. ...............................................187

7 C O N C L U SIO N S ................. ......................................... .......... ........ .. ............... .. 189

L IST O F R E F E R E N C E S ..................................................................................... ..................194

B IO G R A PH IC A L SK E T C H ............................................................................. ....................212









LIST OF TABLES


Table page

2-1. Fluxes and selectivity coefficients a for anti-M-IgG and anti-H-IgG across the M-
IgG and H-IgG-modified membranes................................ ......................67

2-2. Thermodynamic parameters of AGo associated with the affinity interaction of anti-
M-IgG and M-IgG-modified membranes as a function of the temperature..................82

3-1. Dual-mode model parameters used for the permeability plot simulations ......................97

3-2. Thermodynamic parameters from adsorption isotherms of anti-M-IgG and anti-H-
IgG on M-IgG modified membranes, according to the Langmuir monolayer model......107

3-3. The kp determined by various approaches ............................................ ............... 108

3-4. Validation of the dual-mode model parameters............... ............ ...................121

4-1. FTIR-ATR characteristic bands of imidazole functionalized nanotube membranes.......138

4-2. XPS surface composition and binding energies for bare alumina, chlorosilane
modified, and imidazole functionalized membranes ...................................................... 143

4-3. Fluxes and selectivity coefficients a for 6His-UHRP across imidazole functionalized
m etal ion affinity nanotube m embranes..................................... .......................... 149

5-1. Relationship between the Stern-Volmer quenching constant (Ksv) with the feed
solution concentration of anti-H -IgG -Q SY .....................................................................167

5-2. Relationship between slope of the log-log plots for the fluorescence signal vs. time
with the feed solution concentration of anti-H-IgG-QSY ..............................................168

5-3. Data for the Stern-Volmer quenching constant (Ksv) and slope of the log-log plots
for antibody-m odified m em branes........................................................ ............... 169

6-1. Variations of Rm at the different steps of modification on a 100 nm pore diameter
alumina template .................. .. ....... ................... ........... ........... 182









LIST OF FIGURES


Figure pe

1-1. Scanning electron micrograph pointing up the honeycomb structure of nanoporous
alu m in a ................... ...................2...................5..........

1-2. Structure of nanoporous alumina. .................................................................................. 26

1-3. Electrochemical cell setup for nanoporous alumina growth............................................28

1-4. Scanning electron micrographs of a nanoporous alumina membrane anodized at 50 V
in 5% oxalic acid, (A) before and (B) after elimination of the barrier layer. ..................29

1-5. Scanning electron micrograph of a highly-ordered nanoporous alumina membrane
obtained at 50 V in 5% oxalic acid. ...... ......................................................................30

1-6. Outline of the steps involved in the sol-gel synthesis............. ........... ............... 32

1-7. Fluorescence spectra obtained from the biofunctionalized (a) alumina template and
(b) templated silica nanotube membranes with anti-Human IgG labeled with Alexa
4 8 8 .......................................................... .................................. . 3 4

1.8. Reactions involved in the alkoxysilane functionalization. Scheme obtained at
http://www.dowcoring.com/content/publishedlit/SILANE-GUIDE.pdf ......................35

1-9. Types of specialized transport in living membranes (A) carriers, and (B) channels.........37

1-10. Flux concentration dependence for passive and facilitated transport.............................38

1-11. Concentration gradients formed across the membrane during passive and facilitated
tran sport ........................................................ ..................................39

1-12. Basic structure of the immunoglobulin molecule. .......... ..................... ........... .... 42

1-13. Intermolecular attractive forces involves in the antibody-antigen binding. Scheme
obtained from www.fleshandbones.com/readingroom/pdf/291.pdf. ............................44

1-14. Jablonski diagram of the electronic states during the fluorescence process....................47

2-1. Immobilization of antibodies on the template synthesized silica nanotube
m em branes. ................................................................................56

2-2. Flow-through fluorescence system used to measure the flux of the permeating
proteins. ................................................................. ........... ............ 59

2-3. Electron micrographs of (A) branch side, (B) solution anodization side, and (C)
cross-section of the alumina template membranes, and (D) silica nanotubes after
rem oval from the m em brane ...................................................................... ..................63









2-4. Typical plots of (A) calibration data, and (B) calibration curve for Alexa-488 labeled
anti-H -IgG ...................... ..................................... .. .................. ........ 64

2-5. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) through PEG-Si-modified control
membranes. The feed concentration was 2.65 nM (lower curves) and 21.2 nM
(upper curv es). .......................................................... ................. 65

2-6. Fluorescence spectra of (A) M-IgG- and (B) H-IgG-modified membranes after
equilibration with a solution that was 21.2 nM in both (a) anti-H-IgG, and (b) anti-
M -IgG .......................................................... .................................. 66

2-7. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in (A) M-IgG-modified
membrane, and (B) H-IgG-modified membrane. The feed solution concentration was
21.2 nM in both of the permeating proteins.............. ............................................. 68

2-8. Dynamic binding capacity study using M-IgG-modified membranes. The flux and
selectivity coefficient vs. days of equilibration are shown for the transport of anti-M-
IgG and anti-H-IgG. The feed solution concentration was 21.2 nM in both of the
perm eating proteins .................. ................ ........................... .. ....... ..... 69

2-9. Flux versus feed concentration for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG
modified membrane. The feed solution concentrations were equimolar in both of the
permeating proteins............... .......... .. ....... ............. 70

2-10. Typical plot of flux versus feed concentration for anti-M-IgG or anti-H-IgG across
PEG-Si modified membrane. The data plotted is an average of the fluxes for anti-M-
IgG and anti-H-IgG for each specific feed concentration............................................... 71

2-11. Effect of the pore size on (A) the flux, and (B) selectivity coefficients for M-IgG
modified membranes. The feed solution concentrations was 2.65 nM both in anti-H-
IgG and anti-M -IgG ................... .. .................... ... .............72

2-12. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in well-oriented M-IgG-modified
membrane. The feed solution concentration was 21.2 nM in both of the permeating
p ro te in s ................... ................................ ......................... ................ 7 3

2-13. Scanning electron micrograph of (A) the surface, and (B) cross section of a highly-
ordered nanoporous alumina membrane obtained at 30 V in 5% oxalic acid. .................75

2-14. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in (A) PEG-Si-modified
membrane, and (B) well-oriented M-IgG -modified membrane. The feed solution
concentration was 21.2 nM in both of the permeating proteins.................. ...............75

2-15. Flux plots for anti-M-IgG (A) and F(ab)2-anti-H-IgG (m) in well-oriented M-IgG -
modified membrane. The feed solution concentration was 21.2 nM in both of the
perm eating proteins ................................................................. .. ....... ..... 76









2-16. Dissociation constants and binding capacities from adsorption isotherms of anti-M-
IgG and anti-H-IgG on H-IgG- and M-IgG-modified membranes, according to the
L angm uir m onolay er m odel ....................................................................... ..................78

2-17. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in well-oriented M-IgG -modified
membranes showing the effect of temperature. The feed solution concentration was
21.2 nM in both of the permeating proteins.................. .. .... .....................79

2-18. Effect of temperature on the flux and selectivity coefficients......................... ....... 79

2-19. Adsortion isotherms carried out with well-oriented M-IgG-modified membranes for
anti-M-IgG at three different temperatures: 23, 32, and 42 C......................................80

2-20. Affinity temperature dependence of anti-M-IgG in well-oriented M-IgG-modified
m em branes ........... .............................................................................. 8 1

2-21. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG -modified
membranes, (A) without equilibration, and (B) after equilibration. The feed solution
concentration was 2.65 nM in both of the permeating proteins.................. ...........84

3-1. D ual-m ode transport m odel. ................................................................... .....................88

3-2. Permeability plot simulations for the transport of (a) BM and (b) NBM across
carrier-modified membranes. The parameters are summarized in Table 3-1 ..................96

3-3. Permeability plot simulations for the transport of BM across carrier-modified
membranes illustrating the effect of the term F on the dual-mode model. (a) F = 1,
(b) F = 0.5, and (c) F = 0. The other parameters are summarized in Table 3-1 ...............97

3-4. Permeability plot simulations for the transport of BM across carrier-modified
membranes showing the effect of the binding affinity on the permeability. (a) b = 1 x
1011 cm3/moles, and (b) b = 3 x 1010 cm3/moles. The other parameters are
sum m arized in Table 3-1. .............................. ................ .. .... .... ............... 100

3-5. Flux versus feed concentration for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG
modified membrane. The feed solution concentration was equimolar in both of the
p erm eating protein s................................................................................ ...... 10 1

3-6. Experimental obtained permeability versus feed concentration for anti-M-IgG (A)
and anti-H-IgG (m) across M-IgG modified membrane. ............................................102

3-7. Curve of adsorption dynamics for anti-M-IgG (A) in a M-IgG modified membrane.
Inset shows the changes of fluorescence signal with time ............................ ........103

3-8. Adsorption isotherms for anti-M-IgG (A), and anti-H-IgG (m) binding on M-IgG
modified membranes. Also the adsorption isotherm for anti-M-IgG (*) using a
control m em brane is show n. ................................................................. ..................... 104









3-9. Adsorption isotherm for anti-M-IgG (A) binding on M-IgG-PG modified
m em branes. ...............................................................................106

3-10. Extrapolations of the adsorption isotherms at infinite dilute concentration for anti-M-
IgG (A), and anti-H-IgG (m) in M-IgG modified membranes............................ 108

3-11. Permeability plot simulations illustrating the effect of the partition coefficient on the
dual-mode model. Simulation parameters: D = 5 x 10-8 cm2/sec, b = 9 x 1010
cm3/moles, Cstes = 3.5 x 10-8 moles/cm3, and kp were respectively, 1 (V), 0.1 (A),
0 .0 1 (e ), an d 0 .0 0 1(* ).................................................................... ..... ..........10 9

3-12. Concentration-dependence of the solubility in term of the dual-mode model for anti-
M-IgG (A), and anti-H-IgG (m) across M-IgG modified membranes...........................111

3-13. Correlation plots between permeability and solubility based on the concentration
dependence for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified
m em b ran e s ............. .................................................................... 1 12

3-14. Experimental and calculated obtained permeability versus feed concentration for
anti-M-IgG (A) and anti-H-IgG (m) across M-IgG modified membrane. ....... .......... 114

3-15. Experimental obtained diffusion coefficients versus feed concentration for (A) anti-
H-IgG and (B) anti-M-IgG across M-IgG modified membrane................................ 117

3-16. Plots for experimental and calculated obtained permeability against (1 + bCfeed)-1 for
(A) anti-H-IgG, and (B) anti-M-IgG across M-IgG modified membrane .....................120

4-1. Reaction mechanism of imidazole functionalized nanotube membranes......................126

4-2. Flow-through UV-Vis detection system used to measure the flux of six-histidine-
tagged ubiquitin human recombinant protein. ...................................... ............... 129

4-3. Fluorescence spectra of (a) silica nanotube and (b) imidazole functionalized
m em branes. ...............................................................................130

4-4. Fluorescence spectra of imidazole functionalized membranes before (upper) and
after (lower) metal ion immobilization: (A) Co+2, (B) Cu+2, (C) Ni+2, and (D) Zn+2. ....131

4-5. Fluorescence quenching (F/Fo) for imidazole functionalized membranes in the
presence of nitrate solutions prepared in ethanol that were 1 mM of the respective
divalent m etal ion.............................................................................................. 132

4-6. Fluorescence spectra of imidazole functionalized membranes (a) before and (b) after
metal ion immobilization using an ethanol solution that was 1 mM in CuC12 ...............133

4-7. Copper concentration dependence on fluorescence enhancement for imidazole
functionalized m em branes. ..................................................................... ...................134









4-8. (A) FTIR-ATR spectra of bare alumina and imidazole functionalized membranes.
(B) FTIR-ATR spectra of imidazole functionalized membranes in a narrow spectral
region illustrating the effect of using two different backgrounds for the spectrum
subtraction: (a) air, and (b) 3-chloropropyl silica nanotube membrane...........................136

4-9. Transmission FTIR spectrum of pure 2-mercaptoimidazole in KBr. ............................136

4-10. FTIR-ATR spectra of imidazole functionalized membranes showing the effect of the
metal ion immobilized on the thioamide bands: (a) Zn+2, (b) Co+2, and (c) Cu+2. .........137

4-11. Transmission FTIR spectra of the alumina template film for each of the modification
steps illustrated in Figure 4-1: (a) bare alumina template, (b) silica, (c) 3-
chloropropyl silane modified and (d) imidazole functionalized nanotube membranes,
respectively ............................................................................ 139

4-12. XPS survey spectrum for the bare alumina template............................141

4-13. XPS survey spectra for (a) chlorosilane and (b) imidazole modified membranes. .........142

4-14. XPS multiplexes analysis at high resolution for the imidazole functionalized
m em b ran e ............................................... ................. ............................ 14 4

4-15. High-resolution XPS data of the Cls peak for (a) bare alumina template, (b)
chlorosilane modified, and (c) imidazole functionalized membranes ...........................144

4-16. XPS survey spectra for imidazole modified membranes after immobilization of (a)
Cu+2, and (b) Zn+2. .......................................... .......... .. .............. ................ 145

4-17. UV-Vis spectrum for 50 gM of 6His-UHRP measured in 50 mM phosphate-buffered
saline solution pH 8.0 that also contained 300 mM sodium chloride and 2 mM
im id a z o le ................................ ................... ...................... ................ 14 6

4-18. Typical plots of(A) calibration data and (B) calibration curve for 6His-UHRP.............147

4-19. Flux plots for 6His-UHRP in imidazole functionalized metal ion affinity nanotube
membranes that had immobilized different metal ions: (a) Cu+2, (b) Ni+2, (c) Co+2,
and (d) Zn+2. The feed solution concentration was 1.87 pM for all permeation
experim ents. ..............................................................................147

4-20. Effect of the divalent metal ion immobilized on the flux of 6His-UHRP across
imidazole functionalized metal ion affinity nanotube membranes.............................148

5-1. Flow-through fluorescence system used to measure the changes of fluorescence
intensity in the receiver solution. .............................................. ............................ 157

5-2. Structures of(a) QSY-7 and (b) FITC .............................................................. ... 158

5-3. Spectra for F IT C and Q SY -7 .................................................................................158









5-4. Molecular recognition reaction used to promote the fluorescence quenching. .............159

5-5. Fluorescence image and spectra respectively for a H-IgG-FITC-modified membrane,
(A-C) before exposure, and (B-D) after exposure to a sample solution of anti-H-IgG-
Q S Y ..................................................................................... ................................... 16 0

5-6. Typical plots for (A) changes of fluorescence signal, and (B) concentration of
unquenched protein in function of time for two different control transport/detection
systems: (a) anti-H-IgG/H-IgG-FITC, and (b) anti-H-IgG-QSY/R-IgG-FITC.............161

5-7. Typical plots for (A) changes of fluorescence signal and (B) concentration of
unquenched protein in function of time for three different feed solution
concentrations: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control) ........163

5-8. Stern-Volmer plots ofloI against concentration of anti-H-IgG-QSY in the receiver
solution transported across silica nanotube membranes. The effect of the feed
solution concentrations of anti-H-IgG-QSY is shown: (a) 16.6 nM, (b) 33.3 nM, (c)
66.6 nM and (d) 33.3 nM (control). ........................................... .......................... 165

5-9. Effect of the feed concentration on the Stem-Volmer quenching constant Ksv. .............166

5-10. log-log plots of the of fluorescence signal vs. time showing the effect of feed
solution concentration of anti-H-IgG-QSY: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM,
and (d) 33.3 nM (control)........... ........................................................ ...... ........ 167

5-11. Linear relationship between slopes obtained from the log-log plots with the feed
concentration of anti-H -IgG -Q SY ................................................. ........................... 168

5-12. Typical plots for (A) changes of fluorescence signal and (B) concentration of
unquenched protein in function of time illustrating the transport of anti-H-IgG-QSY
across (a) H-IgG-, and (b) M-IgG-modified membranes. The feed solution
concentration was 33.3 nM for both experiments ..................... .................169

6-1. Structures of (a) PNA, and (b) DNA. ................................. 172

6-2. Immobilization of biotinylated-PNA on templated synthesized silica nanotube
m em branes. ..................................................................175

6-3. N anotube m embrane assem bly. ............................................. .............................. 176

6-4. General equivalent circuit model used for impedance data analysis............................178

6-5. Nyquist plots (Zim vs. Zre) for impedance measurement of (a) alumina and (b) silica-
nanotube assembled membranes in highly purified water ..............................................179

6-7. Effect of the wetting time on the resistance impedance measurement for an
assembled unmodified alumina membrane 200 nm pore diameter. .............................180









6-8. Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water: (a)
alumina, (b) after synthesis of silica-nanotube, (c) after silanization, (d) after
biotinylation, (e) after streptavidin immobilization, (f) after PNA immobilization,
and (g) after DNA hybridization. The plots were recorded applying 5 mV alternative
voltage in the frequency range from 1 MHz to 0.05 Hz. ............................. ............... 181

6-9. Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water for a
PNA-modified nanotube membrane: (a) before, and (b) after 10 minutes of DNA
hybridization. The plots were recorded applying 5 mV alternative voltage in the
frequency range from 1 MHz to 0.05 Hz............................................... 183

6-10. Effect of the membrane pore size on the membrane resistance: (A) 200 nm, and (B)
100 nm. The resistance values were obtained from the impedance data recorded by
applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz.......184

6-11. Study of the hybridization process stability with time: (A) unmodified control
alumina membrane, and (B) PNA-modified membrane. The resistance values were
obtained from the impedance data recorded by applying 5 mV alternative voltage in
the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of
com plem entry D N A ...................... ........................ .. .... .................. 185

6-12. Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water for a
PNA-modified nanotube membrane: (a) before, (b) after adding 3 nmoles of DNA,
and (c) after 6 nmoles of DNA. The plots were recorded applying 5 mV alternative
voltage in the frequency range from 1 MHz to 0.05 Hz. ............................. ............... 186

6-13. Reproducibility study of the PNA-modified nanotube membranes, before and after
the hybridization process. The resistance values were obtained from the impedance
data recorded by applying 5 mV alternative voltage in the frequency range from 1
MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA. .................... 187









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

BIOANALYTICAL APPLICATIONS OF AFFINITY-BASED NANOTUBE MEMBRANES
FOR SENSING AND SEPARATIONS

By

Hector Mario Caicedo

August 2008

Chair: Charles R. Martin
Major: Chemistry

Nanotechnology has played an important role in the development of research and

technology during the last two decades. The contribution of nanotechnology in different fields,

along with the versatility of the constructed nanoscale materials, have made nanotechnology one

of the most suitable tools to develop particular nanostructures to realize a desired function and

application. A nanostructure is simply an entity at the nanometer scale with one, two or three

dimensional features. Since nanotechnology covers a broad range of nanoscale materials, to

simplify nanotechnology, it can be classified into two categories based on how the

nanostructures are prepared: top-down and bottom-up. In the top-down methods, the

nanostructures are constructed by chiseling larger bulk materials into entities of smaller size.

Conversely, in the bottom-up case, small units are grown or assembled into their desired size and

shape. The nanoporous materials specifically have attracted a lot of attention because they can be

used for the synthesis of a variety of functional nanostructures of great usefulness in technology.

These porous nanostructures usually combine many of the advantages of the top-down and

bottom-up methodologies such as flexibility, size controllability, and cost. The research

presented in this work utilizes nanoporous membranes to develop porous nanostructured

platforms with potential applications in sensing and separations. In particular, this work is









centered in fundamental studies for bioanalytical applications of affinity-based nanotube

membranes for sensing and separations. A bottom-up methodology like the template synthesis

was used to produce silica nanotubes inside of the pores of alumina membrane. The

functionalization of the inside walls of these silica nanotube membranes allowed control of the

functional behavior and properties of the nanostructured membrane during membrane-based

separations and sensing. The general scheme of the work presented here, is distributed in seven

chapters. The first chapter consists of a general description of the theory and fundamentals as

background supporting evidence for the work presented here. The template synthesis method, the

fabrication of the porous alumina membranes and the possibilities that silane chemistry offers in

order to functionalize the nanostructured membranes is discussed. In addition, a brief overview

of the fundamentals of biological production of proteins, antibodies, and recombinant proteins is

described. After this introduction, the studies presented herein are centered in potential

bioanalytical applications of the silica nanotube membranes.









CHAPTER 1
INTRODUCTION AND BACKGROUND

Introduction

The use of membranes in industrial separations, substituting the traditional methods of

separation, has become a more accepted procedure every day. Features such as compactness,

simple and efficient operation, and low energy consumption have made membrane-based

separations a fine alternative procedure.1 2 Membrane-based separations have found an extensive

range of applications in technology, such as water and wastewater treatment, electrodialysis, gas

separations, and fuel cell development.2 The most important property of membranes is their

ability to control the rate of permeation of different species. Traditionally, membranes have been

used for size- and charge-based separations with great success by using relatively low-resolution

requirements.3-7 However, some applications require the development of new membranes,

because these rudimentary forms of selectivity are insufficient to allow membranes to compete

with the traditional chromatographic methods. Current research and development efforts in

membrane-based separations are directed toward drastic improvements in selectivity keeping the

inherent characteristics of membranes.8 In the search of new advanced membrane materials, the

development of functionalized-membranes with a selective transport function executed through a

specific molecule "carrier" has become in an interesting alternative to modulate the transport

during membrane-based separations.9' 10 This concept has been inspired in analogy by the

strategies that biological membranes use to accomplish transport. Thus, the combined use of

synthetic membranes with biological molecules such as antibodies and nucleotides has become

important in the development of applications both in the biomedical and biotechnological field.1l

By using this combination, membranes have also acquired an important role in the design and

development of biosensors.12 In particular, nanotube membranes prepared by the general method









called template synthesis to prepare nanomaterials have been extensively investigated.46' 9, 10, 13-24

This method entails synthesis of the desired material within the cylindrical and monodisperse

pores of a nanopore membrane or other solid.25'26 Both silica and gold nanotube membranes

prepared by this method have demonstrated the capacity of being a novel platform for

membrane-base separations and sensors.14, 18 One of the most important features of the nanotube

membranes prepared by the template synthesis when they are used in membrane-base

separations, is that the deposited nanotubes are embedded within a mechanical strong support

membrane, making it possible that they can act as conduits during the transport of permeating

species between the solutions placed on either side of the membrane. Previous studies have

shown that, by controlling the inside diameter of the nanotubes and the chemical functionalities

present in the tube walls, highly-selective membranes for chemical and biochemical separations

and sensors can be prepared.4- 6 910, 13-24 This chapter provides a general description of the theory

and fundamentals as background supporting evidence for the work presented here. The template

synthesis method, the fabrication of the porous alumina membranes and the possibilities that

silane chemistry offers in order to functionalize the nanostructured membranes it is discussed. In

addition, a brief overview of the fundamentals of biological production of proteins, antibodies

and recombinant proteins is described. At the end of the chapter, a brief summary of some of the

instrumental analytical techniques used during the development and characterization of the

applications described in this work is provided.

Background

The Concept of the Template Synthesis

Our group has been pioneered in investigating a method called template synthesis for the

fabrication of a variety of nanomaterials.27 This template method is a general approach for

preparing nanomaterials that involves the synthesis or deposition of the desired material within









the pores or cavities of a nanopore membrane or other solid surface. Therefore, depending on the

template structure and synthetic method used, a diverse array of nanostructured materials can be

prepared.28-31 Thus for example, from porous membranes is possible to prepare hollow nanotubes

or solid nanowires, and by using solid spheres or porous spheres respectively, capsules or

nanoporous particles can also be fashioned.

Martin et al. have prepared nanotubes and nanowires composed of many types of

materials, including conductive polymers,32-34 metals,35-37 oxides,38 semiconductors,39 carbon,29'

31 and DNA.40,41 The variety of synthetic methodologies used for these nanofabrication include

electroless plating of metals; electrodeposition of insulating polymers, conductive polymers,

semiconducting oxides and metals; chemical vapor deposition; sol-gel chemistry; chemical

polymerization; carbonization; and layer-by-layer template synthesis.

Using either electrochemical33 34 or chemical42'43 oxidative polymerization of a equivalent

monomer, Martin et al. have prepared nanowires and nanotubes out of conductive polymers such

as polypyrrole, poly(3-methylthiophene) and polyaniline.

Metals having nanoscopic dimensions with attractive optical properties have also been

prepared by template synthesis.4445 For example, a macroscopic sample of pure gold has only

one color, but nanoparticles of gold can show essentially all the colors in the visible region of the

electromagnetic spectrum, depending on the size and the shape of the nanoparticles.25 Hornyak et

al. explored the optical properties of nanoscopic gold particles prepared by electrochemical

deposition of gold within the pores of nanoporous alumina membranes.4445 They found that

these particles achieve a plasmon resonance absorption limit that depended on the gold

nanoparticle diameter.









In the Martin group there is a commonplace methodology for the electroless deposition of

gold within the pores of polycarbonate (PC) 10, 37 and polyethylene terephthalate (PET) 16, 46

template membranes. Depending on the plating time, it is possible to generate gold nanowires or

nanotubes. Thus, in the nanotube situation, by controlling the plating time, the inside diameter

can be varied. The electroless deposition of gold on the pore wall of these polymeric membranes

is currently the best known method to precisely control the inside diameter of the template-

synthesized nanotubes.47

Martin et al. have shown that because the inside diameter of these gold nanotube

membranes can be small relative to molecular dimensions, the nanotube membrane can be used

to cleanly separate small molecules on the basis of molecular size.47 By varying the rate of the

plating reaction the diameter of the nanotubes was smaller in the faces of the template

membrane. Thus, the flux and selectivity were determined for the bottleneck of the nanotubes.

Together with size-based selectivity, they also demonstrated that generic chemical transport

selectivity can be introduced into this membranes.48'49

More recently, Martin et al. have investigated gold nanotube membranes where the inner

pore surfaces were systematically modified for biochemical functionalization to accomplish

selective permeation of DNA.10 They demonstrated a DNA-functionalized nanotube membrane

for the separation of DNA by single-base mismatches selectivity via facilitated transport.10 These

DNA-functionalized nanotube membranes selectively recognize and transport the DNA strand

that was complementary to the immobilized hairpin-DNA (carrier) within the pores, relative to

DNA strands that are not complementary to the immobilized DNA.

Another synthetic method used for the manufacture of nanotubes is the sol-gel chemistry to

produce nanotubes of silica and other inorganic materials. The silica nanotubes have been used









as paradigm to show the power of the template method for preparing nanotubes for

biotechnological applications.o5 Silica nanotubes are easy to make, and they can be derivatized

with an enormous variety of different chemical functional groups using simple silane chemistry

with commercially available reagents.

Mitchell et al. developed the concept of smart nanotubes that can be used for biological

extraction and biocatalysis.38 Nanotubes removed from the template and differentially

functionalized both in the inner and outer nanotube surfaces were used for this application. Such

differentially functionalized nanotube was used as smart nanophase extractor to remove specific

molecules from solution. Thus, hydrophobic functionalized nanotube on the inside, with a

hydrophilic nanotube exterior was used to extract lipophilic compounds from aqueous solutions.

In the same work, Mitchell et al. showed that nanotube functionalized with an enantioselective

antibody selectively extracts the enantiomer of a drug molecule that binds to the antibody from

the racemic mixture.38

Lee et al. demonstrated the use of silica nanotube membranes for enantioseparations.9 They

found that nanotubes embedded within the pores of the alumina template and functionalized with

a Fab fragment antibody that binds the RS isomer of a drug molecule over the SR isomer,

facilitated the transport of the RS enantiomer. They also found that the transport selectivity and

binding affinity of RS over SR could be tuned by controlling the inside diameter of the

membrane containing silica nanotube, and by addition of DMSO to the protein buffer solution,

respectively.

Gasparac et al. explored the template synthesized nano test tubes as an alternative payload-

release strategy by using the template synthesis within the pores of nanopore alumina films.51

These hollow nanotubes were closed on one end and open on the other. Therefore, if these nano









test tubes could be filled with a payload and then the open end corked with a chemically labile

cap, they might function as a universal delivery vehicle. More recently, Hillebrenner et al.

demonstrated nano test tubes corked by chemical self-assembly.52' 53

Buyukserin et al. have also prepared silica nano test tubes by using nanopore photoresist

films as templates.54 They used a plasma etch method, by using a nanopore alumina film as a

mask, to etch a replica of the alumina pore structure into the surface of a photoresist polymer

film. They found that the length of the test tubes is determined by the thickness of the porous

part of the photoresist film.

It is very important to control the inside diameter of the fashioned nanotubes, when the

alumina membranes are used as template. In order to control the nanotube inside diameter a

substitute template synthetic methodology has been considered, where the deposition of the film-

forming material is based on layer-by-layer deposition. In this case, the wall thickness, and

equally the inside diameter of the nanotubes would be determined by the number of layers of the

material deposited along the pore wall. Using this approach, Ai et al. attempted to deposited

layers of polyelectrolyte in the pores of a nanopore alumina membrane.5 Kovtyukhova et al.

have developed a method based on alternate SiC14/H20 deposition/reaction cycles to make and

control the thickness of silica nanotubes.56 Hou et al. used Mallouk's alternating a,co-

diorganophosphonate Zr(IV) chemistry to deposit layered nanotubes along the pore walls of an

alumina template membrane.40-41 For these nanotubes, the layers were held together by bonds

between the phosphonate groups and the Zr(IV) ions, and the polyvalence of this interaction

provided cross-linking that imparts the structural integrity of the layered nanotubes.40 Hou et al.

also used this synthetic strategy for the synthesis of DNA nanotubes where the bulk of the tube









was composed of DNA, and the layers were held together by hybridization of complementary

DNA strands.4041

The remarkable versatility offered by the material fabricated using template synthesis such

as: size, shape, chemistry methodology applied, and systematic functionalization, make those

nanostructured-template materials practical for particular applications.

Description of the Nanoporous Alumina Membrane Template

Nanoporous alumina films, which are produced by the electrochemical anodization of

aluminum in suitable acidic or basic electrolytic solution,44' 57- 63 have been investigated and used

as starting material for the fabrication several kinds of nanostructures, due to its very well

ordered porous structure with high aspect ratio (thickness divided by the diameter). In recent

years, simultaneously with the improvements in degree of ordering, these nanoporous alumina

membranes have become a popular template system for the synthesis of various functional

nanostructures.9,64-70

The porous structure appears as a honeycomb-like arrangement consisting of a close-

packed array of columnar alumina units called cells, each containing a central straight pore, as

shown in the scanning electron micrograph in Figure 1-1 from an alumina membrane obtained in

our laboratory.













Figure 1-1. Scanning electron micrograph pointing up the honeycomb structure of nanoporous
alumina.









By changing the anodization conditions such as electrolyte used, its concentration, the

applied voltage and the period of anodization, the pores length, radius and the cell population

density can be easily varied. The pore diameter depends on the applied voltage used for the

anodization, and the thickness is a function of the anodization time. Pore diameters can be varied

between 20 and 200 nm, with lengths between few tens of nanometers up to hundreds of

microns. The pore size can be also controlled by the pore-widening treatment of dipping the

anodized porous alumina in an appropriate acid solution after the anodization. This acidic

solution is generally used to remove the barrier layer, which is a dense insulated oxide layer that

separates the underlying aluminum surface and the porous oxide film (Figure 1-2). 58 71 However

before the barrier layer is eliminated, the porous alumina film has to be removed from the

underlying aluminum metal in order to collect a free-standing membrane. There are two methods

of removing the remaining aluminum substrate in order to obtain the free-standing membrane.

The first one is by dissolution with a saturate HgC12 solution, and the second is a voltage-

reduction method. The latter entails stepwise reduction of the potential to obtain an interfacial

oxide layer that has a network of much smaller pores.44 After formation of this highly porous

layer, the alumina film is immersed into an acidic detachment solution which results in rapid

dissolution of the interfacial oxide.44











Figure 1-2.Structure of nanoporous alumina.


Figure 1-2.Structure of nanoporous alumina.









There are two reported methods for the fabrication of highly ordered anodic nanoporous

alumina membranes. One is the two-steps anodization method, which is the commonplace

method used in our laboratory to produce the alumina membranes. Since the ordering of the pore

configuration (self-organization) depends on the anodization time in the acidic electrolyte

solution, two steps of anodization are required to obtain a highly ordered pore structure.61' 62, 72

This method will be discussed in more detail in the next section.

The second method is based on pretexturing of the aluminum substrate by using

nanoindentation with a master mold.73'74 Masuda et al. used a SiC mold with a hexagonal array

structure prepared by conventional electron beam lithography.73'74 In this pretexturing method,

the depressions formed in the aluminum serve as initiation sites for the pore creation at the initial

stage of the anodization. An ideally ordered nanoporous alumina membrane is formed after

anodization of the aluminum substrate with the transferred textured pattern.

Two-step anodization method

The degree of the ordering of the pore initiation at the surface of the anodic porous alumina

is low because the pores develop randomly at the initial stage of the anodization. To improve the

ordering of the surface side of the anodic porous alumina, two-steps anodization is effectively

adopted.61 62,72 Two separated anodization processes are required for this procedure. The first

anodization process consists of a long-period anodization to form the highly ordered concave

nano-array of nucleation sites at the alumina/aluminum interface. The second anodization is

performed after the removal of the oxide formed in the first anodization step. Thus, after the

removal of the oxide, an array of highly ordered dimples is formed on the underlying aluminum

substrate. These dimples act as initiation sites for the pore development during the second

anodization. This process generates an ordered pore structure through the entire oxide layer. In

the following paragraphs, the methodology used in our laboratory for the fabrication of alumina










membranes will be discussed, which consists of three stages: polishing of the aluminum

substrate, the two anodization steps, and the detaching step in order to have a free-standing

membrane.

Polishing of aluminum foils: In order to obtain a high quality alumina film, the starting

aluminum material, must be very smooth. Initially, the high purity aluminum foils (99.999%) are

mechanically polished with slurry of aluminum particles. If, to the naked eye, the aluminum foil

shows severe scratches a mechanical polishing with fine sand paper is applied until the scratches

disappear. The mechanical polishing is then followed by an electrochemical polishing. In this

process, a potential difference of 15 V is applied between the anode aluminum foil and a lead

plate which serve as cathode, by using a polishing electrolyte solution (95% concentrated

phosphoric acid, 5% concentrated sulfuric acid and 20 g/L chromic oxide) heated to 70 OC.

Electropolishing steps of 5 minutes are successively repeated as many times as it is necessary,

until the aluminum foil has a mirror-like surface.

Two anodization steps: Electropolished aluminum is oxidized by using an electrochemical

cell setup like the one presented in Figure 1-3.


Anode Cathode
2AI(s)+ 3H20() A203()+ 6Hq)+ 6e 6H' iq) + 6e" 3H2)

Supply
Aluminum

si S t a in le s s
Porous Steel
Alumina

A Electrolyte




Stirring
Bar _


Figure 1-3.Electrochemical cell setup for nanoporous alumina growth.









The first anodization step is performed in 5% oxalic acid as electrolyte for 12 hours, at 2

C, under a constant voltage. It should be pointed out that the pore diameter is controlled via

anodization voltage. Thus, smaller pores require lower applied voltage and more conducted

electrolytes such as sulfuric acid.7 Large pores need larger voltages, which causes a high rate of

dissolution in highly conductive electrolytes.76 For this reason, the fabrication of larger pores

requires lower conductive electrolytes, such as oxalic acid.76

After the first anodization, the formed alumina film is removed in an etching solution that

contains 0.4 M in phosphoric acid and 0.2 M in chromic oxide at 60 OC. This anodization

removal step leaves a uniform concave nano-array of nucleation sites in the underlying

aluminum metal that correspond to each pore.

The second anodization is performed at the same voltage and electrolyte conditions used in

the first anodization. This second anodization step is carried out for a length of time, depending

on the thickness of alumina membrane desired. In our laboratory, we obtained alumina

membranes of thicknesses ranging from 0.3 to 150 inm.












Figure 1-4.Scanning electron micrographs of a nanoporous alumina membrane anodized at 50 V
in 5% oxalic acid, (A) before and (B) after elimination of the barrier layer.

Detaching step: The residual unoxidized aluminum is removed first with a saturated

solution of HgC12. Then, the remaining barrier oxide layer is removed by etching the alumina

film in a 10% phosphoric acid solution. Since the barrier layer thickness also varies with









increasing voltage the etching step requires longer times for anodized membranes with bigger

pores. Figure 1-4 shows the scanning electron micrographs of an alumina membrane obtained in

our laboratory before and after elimination of the barrier layer.

Thus, the resulting free-standing alumina membrane is obtained. In our laboratory, it is

common to use a method with a two-step anodization to obtain highly ordered porous alumina

membranes, as can be seen in Figure 1-5.








obtained at 50 V in 5% oxalic acid.
anodizing potential.5861627172

























The proposed mechanism for the formation of highly ordered porous alumina membranes

has been explained as the result of two competing processes: 1) the pore initiation process due to

a geometric effect, which is called "field-assisted dissolution process"72 and 2) the self-
-000i0000 po066 ,
Thpopse ecansmfrt










Figure i-5. Scanning electron micrograph of a highly-ordered nanoporous alumina membrane



















a geometric effect, which is called "field-assisted dissolution process S,,7 and 2) the self-









organization of the pores which is considered to be triggered by mechanical stress at the

alumina/aluminum interface.76

In a first step the barrier layer is formed jointly with numerous defects in the aluminum

surface for the pore initiation. Due to the decreasing current with the increase of the thickness of

the barrier layer the formation of pores is initiated at defect positions. Then, assisted by the

electric field, the alumina film developed at the metal-oxide interface at the pore walls are

preferentially dissolved by the acidic electrolyte. For this reason, when the electric field

increases by increasing the voltage, the dissolution process is accelerated and larger pores are

formed. Thus, pores grow perpendicular to the surface via field-assisted oxide dissolution at the

oxide/electrolyte interface in dynamic equilibrium with the oxide growth at the metal/oxide

interface.77'78 It is believed that the anodization current is mainly related to the movement of

ionic species (02-, OH-, Al3+) through the oxide layer at the bottom of the pore.79 Therefore, as a

consequence of the pore development, the electric field and the ionic current become

concentrated in the barrier layer underneath the major pores. The oxide growth at the metal/oxide

interface is due to the migration of oxygen containing ions (O-2/OH-) from the electrolyte

through the oxide layer at the pore bottom, with simultaneous ejection of Al3+ ions drifted

through the oxide layer into the solution at the oxide/electrolyte interface.79 The fact that Al3+

ions are lost into the electrolyte and that the atomic density of aluminum in alumina is by a factor

of two lower than in metallic aluminum, have suggested a possible origin of forces between

neighboring pores that contributes to the mechanical stress which is associated with the

expansion during oxide formation at the metal/oxide interface.76 Since the oxidation takes place

at the entire pore bottom simultaneously, the material can only expand in the vertical direction,

so that the existing pore walls are pushed upward. This somewhat explain the self-organized










arrangement of neighboring pores in hexagonal arrays by repulsive interactions between the

pores during the steady-state film formation after the pore initiation process.

Sol-Gel Chemistry at a Glance

The sol-gel processing of inorganic ceramic and glass materials has been known for a long

time but it has been commonly used since the mid-1900s, and comprehensively studied during

the past 30 years.80 81 In recent years, the sol-gel chemistry has evolved as a powerful approach

for preparing micro- and nano-structured materials.8285 Sol-gel materials display attractive

properties such as ease of preparation, low temperature synthesis, multiple structures with large

open spaces of tunable porosity, controllable hydrophobicity, capacity for chemical modification

of entrapment of various components, mechanical stability, and optical transparency.86-89

The molecular precursors for preparation of sol-gel materials are essentially alkoxysilanes.

The basic steps involved in the sol-process using tetraethoxysilane (TEOS) as a precursor

molecule are shown in the Figure 1-6. The initiation step consists in the hydrolysis of the

alkoxide precursor which leads to the formation of silanol groups (Si-OH) and ethanol. The

silanols can then undergo polymerization reactions with other silanols or with other

alkoxysilanes and this condensation reaction leads to the formation of siloxane groups (Si-O-Si)

along with water and ethanol as additional reaction products.

1 Hv.h-rolysis
S,(OEt)4+ H20+H* -~ ... r, OHI), + nEtO

2)('oindensation
~Si(()Ol:f}OH)+'.-, ,. ), -- (OH).,(1,,t ... 4./ ,OH)_ + H,
01
",,' !'- ,_ ,(H),I +'."' I"; -- '{(H)_,(O.Er),_,,Si~-SSI(OEtn,_ (G( \_, +tlWH


l--Si-O--Si- --- i--()--SiO-Si-()-Si --S--()-, + 1.1p


rrI[-Si-O-Si--St-O-Si-O-L -*Gdl

Figure 1-6.Outline of the steps involved in the sol-gel synthesis.









These reactions lead initially to the formation of molecular clusters, which then grow upon

polycondensation to produce a sol. This sol stage is produced when the particles formed are well

dispersed in the liquid phase to form colloids. In addition, depending on the experimental

conditions; this sol can be transformed in different ways. The usual case is gelation which occurs

when the colloidal particles link together to become a three-dimensional network. Upon solvent

extraction the gel is transformed into aerogel,90'91 while solvent evaporation results in the

formation of xerogel.92 93 Those are the two most common processes of converting gels into

silica. The maximum temperature for both processes can be kept under 100 C.

The active silica surface with large specific surface area is of enormous utility to produce a

variety of modified silica via covalent bond by using silane reagents.94 By the insertion of

organic functional groups to silica surface there is a partial conversion of surface silanol to a new

organofunctional surface that acquires organophilic properties. Thus, the silane chemistry gives a

set of properties to the silica surface, which differs considerably from the original matrix.95

At the silica surface, the structure is composed of both siloxane groups (Si-O-Si) with the

oxygen atom on the surface, and one of the three forms of silanol groups (Si-OH).96 One of the

types of silanol groups found in silica surface is the isolated free silanol groups, where the

surface silicon atom has three bonds into the bulk structure and the fourth to the OH group. The

second type corresponds to the occurrence of two vicinal free silanol groups where the OH

groups are bridged by a H-bond. A third type of silanols called geminal silanol consists of two

hydroxyl groups attached to one silicon atom.97

Since the discovery of silanol groups on silica surfaces in 1936 by Kiselev,98 studies based

on theoretical calculations, physical methods and chemical methods have been explored in order

to quantify the number of silanol groups on the surface. At present, it is accepted that the










estimated number of OH groups accessible on porous silica surface is between 4 and 5 OH

groups per nm2 with an average surface area of 20.4-21.7 A2 per silanol group.94 99-101 In the

research reported here, we took advantage of this special attribute of the higher silanol groups on

silica surface for further silane functionalization, because when compared with alumina, the

number of estimated hydroxyl groups accessible on alumina surface is only 0.1-2.5 OH groups

per nm square.30'102

In order to prove the benefit of the silica nanotubes in the alumina template, some

fluorescence experiments were performed. For these studies, both a templated silica nanotube

membrane and alumina membrane were biofunctionalized with anti-Human IgG fluorescent

labeled with Alexa 488. This anti-Human IgG was covalently attached via the Schiff s base

reaction to the walls of the aldehyde functionalized nanotube membranes.9 These results are

shown in Figure 1-7, where the Alexa 488 fluorescence spectra from alumina membrane and

from silica nanotube membrane are compared. The Alexa 488 fluorescence signal from silica

nanotube templated membrane was approximately 6 times higher than the analogous alumina

membrane. Hence, these results suggest that the deposition of silica nanotubes is essential to

provide higher degree of functionalization on the membranes.

60000 b
Ssoooo -

4 30000



0 I -- -



Wavelength (nm)

Figure 1-7.Fluorescence spectra obtained from the biofunctionalized (a) alumina template and
(b) templated silica nanotube membranes with anti-Human IgG labeled with Alexa
488.
g 20000
a


400 500 500 700 Soo
Wavelength (nm)

Figure 1-7.Fluorescence spectra obtained from the biofunctionalized (a) alumina template and
(b) templated silica nanotube membranes with anti-Human IgG labeled with Alexa
488.










Silane Chemistry a Versatile Tool for Surface Functionalization

The chemical modification of silica surfaces consist in chemical bonding of molecules

using silane chemistry, which allow to modify the chemical or physical properties of the surface

in a controlled way. Silane functionalization proceeds mainly via the reaction of silanol groups

with organofunctional silanes, especially chloro- and alkoxy-silanes.103 The chemistry

modification for alkoxysilanes is analogous to that one described for sol-gel production. They are

easily prepared through the hydrolysis and condensation of organotrialkoxysilanes. The most

common alkoxy groups are methoxy and ethoxy, which give methanol and ethanol respectively,

as additional reaction products. The main difference between the trimethoxysilyl and

triethoxysilyl monomers, when they are used in silane functionalization, is that reaction with

trimethoxysilyl composite is faster.104 The reactions involved in the alkoxysilane

functionalization are illustrated schematically in Figure 1-8.


Hydrolysis of lkn\i ilunel.. Bonding to an inorganic surface.

Hydrolysis RSi(OCH,) 3 R R R
I I I
H2O CHOH HO-SI-O-SI-O-Si-OH
H20I I I
Hydrogen /0\ /O\ /0\
Condensation RSI(OH), Bonding H HH HH H
I HO Substrate
R R R A |O HO
I I I
HO-SI-O-SI-OSIH R R R
I | Bond HO-SI-O-SI-O-SI-OH
0 0 0 Formation I I I
I Fo O O 0 Substrate
H H H I h



Figure 1.8.Reactions involved in the alkoxysilane functionalization. Scheme obtained at
http://www.dowcoming.com/content/publishedlit/SILANE-GUIDE.pdf

For the chlorosilanes situation, the reaction with alcohol produces alkoxysilanes and

hydrochloric acid. The extent of this reaction can be monitored by measuring the pH.









The decision of using organosilanes with one or more than one hydrolyzable groups for

silane functionalization depends on the particular set of properties desired in the new

functionalized surface. Thus, when a hydrophobic surface is preferred, organosilanes with one

hydrolyzable group are used.103 If surfaces with higher degree of surface modification are

desired, organosilanes with two or three hydrolyzable groups are used.103

An alternative way of generating one monolayer of the desired silane monomer on silica

substrates is by conducting silane chemistry of chlorosilanes in aprotic solvents.105, 106 In

conditions where the aprotic solvent and surface are water free, no alkoxysilane is produced, and

no polymerization of the chlorosilane can take place. Therefore, the reaction must be promoted

by nucleophilic substitution; where the nucleophile hydroxyl groups in the surface attack the

primary halide carbon in the chlorosilane moiety. Surface modification in those cases takes

longer time, usually 12-24 hours.

Thereby, surface modification via silane functionalization provides a unique opportunity to

regulate the interfacial properties of silica substrates while their basic structure and mechanical

strength are preserved.

Nature's Transport across Membranes: An Inspiration for Membrane-Based Separations

Biological membranes are vital components of all living systems, forming the external

boundary of the living cells and the internal organelles into the cytoplasm.107 They are

constituted mainly of various lipids arranged in a bilayer that imparts a fluid character. Thus,

both lipids and proteins are free to move within a bilayer. The proteins embedded in the bilayer

and the carbohydrates attached to its surface, facilitate communication and transport across the

membrane. In the living membranes, there are two types of membrane proteins which are

involved in transport and signaling activities. Those are the peripheral and integral (or intrinsic)









proteins. Peripheral proteins interact through electrostatic forces and hydrogen bonds with the

surface of bilayers. While integral proteins are strongly associated with the bilayer itself.

These functional characteristics described above, enable membranes to act as essential

filters in different living processes that depend on membrane transport systems, such as the

acquisition of nutrients, the elimination of waste materials, the transport of metabolites, the

generation of concentration gradients vital for nerve impulse transmission and for the normal

function of brain, heart, kidneys and other organs.107

All transport processes in living membranes are mediated by the transport proteins which

may function either as channels or carriers. In transport mediated by carrier proteins, only the

specific binding molecule is transported. Carrier proteins cycle between conformations that

allow molecules to pass to the other side of the membrane. There is never an open conduit all the

way through the membrane with carrier proteins. Channels on the other hand, form open pores

through the membrane that permits the free diffusion of molecules of appropriated size and

charge. These two types of specialized transport systems are illustrated schematically in Figure

1-9.

A 0 B
conformational conformational conformational
change change change



closed open


Figure 1-9.Types of specialized transport in living membranes (A) carriers, and (B) channels.

In living membranes there are three classes of transport: passive, facilitated, and active. In

passive transport, the transported specie moves across the membrane in a thermodynamically

favored direction without the assistance of a specific transport system. For neutral species the

driving force across the membrane is the chemical potential, while for charged species the










driving force is the electrochemical potential. In addition, the flux of uncharged molecules across

the membrane is linearly dependent on concentration.

In facilitated transport, the transported species moves also according to a

thermodynamically favored direction, but with the collaboration of a specific transport system

"the carrier". Facilitated transport systems display saturation behavior, where the flux is still

dependent on concentration, but at high concentration the dependence is no longer linear. This

behavior is due to the saturation of the binding carrier sites as the concentration of the

transported species increases. Figure 1-10 compares the dependence on concentration of the flux

for the passive and facilitated transport systems.



Facilitated
Transport


jj Passive
u- / Transport






Concentration

Figure 1-10.Flux concentration dependence for passive and facilitated transport.

The third class of transport in living membranes corresponds to the active transport

systems. These use energy in order to transport the species against their thermodynamic

potential. The most usual energy supplier is the ATP hydrolysis, but light energy and the energy

stored in ion gradients are also used.107

Thus, inspired in the strategies that biological membranes use to accomplish transport,

scientists have explored the use of analog synthetic membranes with specific functionalization to

selectively conduce transport. Likewise, these systems function on the base of facilitated









transport concept. This facilitated transport synthetic functionalized-membranes integrate a

mediator in the membrane "the carrier".9, 10 This carrier usually interacts chemically with one of

the components in the feed to be transported across the membrane in a facilitated mode.

Therefore, in the same way that living membranes facilitated transport, the flux for this preferred

permeant is enhanced when is compared with the passive transport situation (Figure 1-10). As in

the case of facilitated transport in living membranes, the facilitated transport in synthetic

facilitated-membrane systems also displays saturation behavior. The basic mechanism for this

enhanced transport is ascribed to the sequential binding and unbinding events of the facilitated

specie with the immobilized carrier in the membrane. The concentration gradients of facilitated

process are shown schematically in Figure 1-11, in which the reaction of the facilitated specie (F)

with the immobilized carrier (C) in the membranes is given by Equation 1-1.

F+C The equilibrium constant for this reaction is characterized by Equation 1-2.

[F*C]
K =[F (1-2)
[F] x [C]
The terms in the square brackets represents the molar concentration of the chemical species

involve in the reaction.


Feed Membrane Permeate Concentration gradient
Feed Membrane Permeate for facilitated transport

C Concentration gradient
o0 for passive transport

[ F ]Feed



[ F ]Permeate
Length -----

Figure 1-11.Concentration gradients formed across the membrane during passive and facilitated
transport.









Thus, as can be seen from Figure 1-11, in the facilitated transport membranes, the

presence of immobilized carriers that interact with the permeate molecules creates larger

concentration gradients downstream the membrane length. Besides, the selectivity acquired with

facilitated transport systems is an extra important aspect compared with the nonselective

characteristic of the passive transport.

Therefore, membranes with higher fluxes can be achieved by creating membranes with

larger concentration gradients using immobilized carriers. This effect is also supported by

Fick's first law of diffusion:

ac
J =-D- (1-3)
9x
Since according to Fick's first law, the flux (J) of a permeate molecule across a membrane

is directly proportional to the concentration gradient across the membrane. The term CC/ax in

Equation 1-3 represents the concentration gradient of the permeate specie and it is the driving

force that leads to molecular movement.108 The term D is the diffusion coefficient and it is a

measure of the mobility of the individual molecules across the membrane. The negative sign

indicates that the direction of flow is down the concentration gradient.

The above description clearly demonstrates that carrier-facilitated transport membranes

promote higher concentration gradients inside the membrane that enhance the flux of the

molecule that interact selectively and reversely with the carrier. In simplest terms, the selective

binding interaction enhances the local concentration gradient, and thus the flux, of the permeate

molecule across the membrane.9' 10, 109

Systems using synthetic membranes based on the concept of facilitated transport have been

the subject of numerous investigations.9' 10, 109- 119 Thus, facilitated transport membrane systems









using different type of carriers such as chemical completing agents,111- 113 metalloproteins,114, 115

crown ethers,116' 117 antibodies,9 DNA,10 and others118' 119 have been already demonstrated.

Fundamentals about Proteins

Proteins constitute the most important material that composes all living organisms.107 They

are part without exception of each cell. A brief review of the fundamentals of biological

production of proteins and protein structure is described next.

Proteins consist of a linear sequence of aminoacids, and their order determines the primary

structure of the protein. Primary structure of a protein is encoded in the base pair sequence of

DNA encoding the gene for the protein. Thus, the four different bases of nucleic acid sequences

(adenine (A), cytosine (C), guanine (G), and thymine (T)) are translated into the 20 aminoacids

of proteins by a genetic code and grouped into triplets of bases units called codons to code each

particular aminoacid. The start codon of the gene is AUG which encodes methionine and the

three stop codons are UAA, UAG, or UGA. The genetic expression is initiated when the RNA

polymerase binds the start codon and transcripts a single-stranded copy of the DNA into the

messenger RNA molecule (mRNA), this process takes place until the stop codon is reached. This

mRNA molecule that carries now the genetic code of the DNA is then fed into the ribosome.

Likewise, chemical polymerization processes, the biological synthesis of protein has three

phases: initiation, elongation, and termination. In initiation, the mRNA binds in a particular

complex to the transfer RNA (tRNA), which serves as a carrier of aminoacids to hybridize with

the mRNA. During the elongation, the ribosome moves along the mRNA via codon-directed

association. The ribosome covalently attaches the new amino acid to the end of the growing

peptide chain and so forth the process continues. The termination of the peptide synthesis is

activated by the appearance of a stop codon (UAA, UAG, or UGA).









The ribosome produces an unfolded linear chain of denatured peptides. Then, the chain of

aminoacids folds to form structures such as alpha-helices and beta-sheets, which are the elements

of the protein secondary structure. The various secondary structures then fold into a compact

tridimensional structure, which correspond to the tertiary structure of a protein. This tertiary

structure, in many cases contains the functional form of the protein. In some cases, the

combination of two or more subunits of tertiary structure peptides may be required for the

functional form of the protein; this assembly is called the quaternary structure.

Antibodies

Antibodies, also called immunoglobulins, are proteins produced by the immune system for

the defense of the organism against foreign species.120 They exhibit very specific binding

capacity for specific molecules. Immunoglobulins are divided in five classes: IgG IgM, IgA,

IgD, and IgE.120 They vary in size, charge, aminoacid composition and carbohydrate content.

Immunoglobulins are made of two identical heavy and two identical light chains of aminoacids,

forming a characteristic Y-shape (Figure 1-12).


antigen-binding site VH domain


Domain Fab fragments
-- VL domain


hinge region
light chain Fc fragment




heavy chain


Figure 1-12.Basic structure of the immunoglobulin molecule.









The molecular weight of the heavy chains varies with the class of antibody. They are called

gamma, mu, alpha, delta and epsilon for IgG, IgM, IgA, IgD, and IgE, respectively. Based on

small polypeptide structural differences, there are two types of light chains classified as kappa

and lambda. This means that the polypeptides chains of immunoglobulins are encoded by three

different clusters of genes. One cluster encodes the heavy chains for all the classes and

subclasses of antibodies, a second one encode the kappa light chains and a third one the lambda

light chains. The four glycoprotein chains are connected to one another by disulfide (S-S) bonds

and noncovalent interactions. The variations in the number of disulfides bonds and the length of

the hinge region determine the subclasses of antibodies. This hinge region is a flexible structure

composed of aminoacids that enables the antibody to bind a couple of epitopes separated apart

on an antigen.

Immunoglobulins are bifunctional molecules. They have one region involved with the

antigen binding which is called the Fab portions of the antibody, and a second region called the

Fc portion which is involved in several biological activities. This Fc region is also used as a

binding site for manipulating the orientation of the antibody during many immunochemical

procedures, due to its capability to bind proteins G and A.121

The Fab portions have a well-defined tridimensional shape specifically designed to bind

the antigen, by a lock-and-key mechanism type. The binding between antigen-antibody entails

the formation of multiple intermolecular noncovalent forces in the pocket of the binding site as is

schematically shown in Figure 1-13.

There must be appropriate atomic groups on opposite sites of the antigen and antibody to

facilitate the binding between both proteins. Besides, the shape of the combining site must fit the

epitope in order to hold the simultaneous formation of noncovalent bonds. Thus, the formation










of hydrogen bonds results in hydrogen bridges between appropriated atoms. In the same way, the

attractions between the opposite charged groups in both sides of the proteins contribute with the

electrostatic forces, while the Van der Waals forces are created by the interaction between

electrons clouds. The hydrophobic forces on the other hand, which may contribute up to half of

the total strength of the antigen-antibody bond, rely in the association of non-polar hydrophobic

groups, minimizing the contact with water molecules. All these attractive forces are weak if they

are considered individually and compared with covalent bonds. However, if they are put together

they become a large component of the total binding energy.

antibody antigen
hydrogen H-
oondnbg from www.fles.
N -- H 1 JT_
electrostaticcmbint
i -^fl/ -+ .
Van der Waals
I ____'_ + "- l-. \ |I +."t- *
hydrophobic
watere r"
excluded


Figure t-13.Intermolecular attractive forces involves in the antibody-antigen binding. Scheme
obtained from www.fleshandbones.com/readingroom/pdf/291.pdf

Recombinant proteins

Proteins produced by in vitro transcription and translation can be expressed in organisms

such as Escherichia coli (E. coli). Since, the genome and structure of the E. coli have been long

studied and well understood by biologists and geneticists. One of the main characteristics that

make the E. coli as a very powerful and versatile system for genetic engineering and protein

expression is its ability to rapidly proliferate; E. coli cells can be doubled in number under

optimal conditions every five minutes.122,123









Recombinant proteins are produced from clone DNA (cDNA) sequences which usually

encode an enzyme or protein with a specific known function. Therefore, the process of genetic

engineering begins with the isolation and extraction of the gene of interest from the clone DNA.

This corresponds to the isolated DNA that is complementary to the mRNA that encodes the

protein of interest. The next step is inserting the gene of interest by cloning techniques into a

vector. The vector is just a small piece of DNA where foreign fragments of DNA may be

inserted. Thus, a vector functions like a "molecular carrier", which will carry fragments of the

gene of interest into a host cell. Those vectors are usually derived from plasmids, which are

small, circular, double-stranded DNA molecules occurring naturally in the cytoplasm of bacteria.

The vector thus incorporates an origin of DNA replication for the gene of interest into a host cell

like E. coli.122, 123 The production of the protein starts when these special engineered vectors are

inserted into E. coli where the recombinant protein of interest is expressed. Because this E. coli

duplicates rapidly, large amounts ofE. coli cells containing the foreign gene can be grown in a

matter of short time.

In order to purify the expressed protein from the complex cell culture, it is a common

practice to tag on the vector, a sequence of DNA that introduces a specific functionality into the

expressed peptide to be used during protein purification. This tag or purification tool is placed in

the vector to create a small extension of aminoacids added to the N-terminal of the target protein.

This aminoacid sequence is usually designed to be a point of attack for a specific protease during

purification. Thus, after the recombinant protein is expressed and extracted from E. coli, the N-

terminal extension can be used to purify the protein and then cleavage it with a specific protease

to generate an almost natural N-terminus sequence of the target peptide.122' 123









Frequently, the introduction of a metal binding site in the expressed recombinant peptide is

placed for further purification. It has been demonstrated that an aminoacid sequence consisting

of 6 or more histidine residues in a row operate as a metal binding site for recombinant

proteins.124, 125 This hexa-His sequence is called a His-Tag. By using commercially available

vectors, the His-Tag sequence can be placed on the N-terminal of a target protein. This expressed

His-Tag recombinant protein can be simply purified by using immobilized-metal affinity

chromatography (IMAC).126, 127

Synopsis and Analytical Visualization of the Instrumental Techniques

Fluorescence spectroscopy

The transport performance in membrane-based separation is normally investigated by

permeation experiments of the systems in buffered-aqueous solutions of the permeant solutes to

separate. One of the analytical methods used in this investigation to characterize the fluxes of the

transported proteins across the membranes is fluorescence spectroscopy. This was achieved

through the use of simple photophysical probes such as fluorescent-labeled antibodies. The use

of fluorescence techniques for the study of proteins in solution is widespread. Fluorescence

spectroscopy is a very sensitive technique, making it possible to perform experiments on protein

solutions at very low concentration, minimizing the amount of molecules required for the

analysis and eliminating problems associated with aggregation and low solubility.128- 130

Fluorescence is a characteristic process that occurs in certain type of molecules called

fluorophores that have the property of absorbing energy and then immediately releasing this

energy in the form of light.131 The fluorescence process basically consists of three stages:

excitation, lifetime of the excited state, and fluorescence emission.










In the excitation stage, characteristic photon energy is absorbed for the flourophore

molecules creating an excited electronic singlet state (S2) as is illustrated in the Jablonski

diagram in Figure 1-14.


Sn
excited vibrational states
(excited rotational states not shown)
S2 2 L IC
I 2 I Stokes shift
SA = photon absorption
F = fluorescence (emission)
> P = phosphorescence
Po Ts S= singlet state
S absorbed I 2 =triplet tate
S exciting e IC = internal conversion
exci in A emitted T ISC= intersystem crossing
S light fluorescence 1 intersyse
F light p phosphorescence
So
electronic ground state


Figure 1-14.Jablonski diagram of the electronic states during the fluorescence process.

This excited state lasts typically for a few nanoseconds (1 to 10 nanoseconds). During the

excitation process, part of the excess of energy gained is dissipated by internal conversion to

reach the lowest energy mode of a relaxed excited state (Si). The final stage is the fluorescence

emission, the electron returns to the ground state energy (So), where the remaining energy is

released in the form of a photon. Due to the loss of energy during the internal conversion in the

excited state, the total energy of the fluorescent emitted photons is lower than the absorbed

excited photons. Thus, there is a difference between the excitation and emission wavelength

which is called the Stokes shift. The ratio of the number of fluorescence photons emitted to the

number of photons absorbed is called the fluorescence quantum yield, and it is a measure of the

extent for the entire process.

The intensity of the fluorescent emitted light, I, is described by the relationship showed in

Equation 1-4.132

I =QI (- e-c) (1-4)









Where Q is the quantum yield, Io is the incident radiant power, e is the molar absorptive

coefficient, b is the path length of the cell, and c is the molar concentration of the fluorophore.

However, for dilute concentration (ebc < 0.05), the Equation 1-4 reduces to the form:

I = QIobc (1-5)
Therefore, at dilute concentration of the fluorophore, the relationship between fluorescent

signal and concentration should be linear, if Q, Io, and e are keeping constant.132

Fluorescence quenching

The fluorescence quenching phenomena refers to any process that decreases the

fluorescence intensity of the fluorophore molecules.131 The two general types of quenching

processes encountered are the collisional and static quenching. The collisional quenching is a

process that occurs during the excited state lifetime. In this case, the excited fluorophore

experience contact with an atom or molecule (the quencher) that facilitate non-radiant transitions

to the ground state. In the case of static quenching, the deactivation of the fluorescence intensity

is due to the reaction between fluorophore and quencher that leads to formation of a stable

complex in the ground state.

The role of fluorescence quenching can be experimentally studied by determining the

quenching rate parameters using the Stem-Volmer plots that are drawn in accordance with the

Stern-Volmer equation (Equation 1-6).131


0 =1+Ks[Q] (1-6)
I
Where [Q] is the concentration of the quencher, Io and I are respectively, the fluorescence

intensity in the absence and presence of quencher concentration, and Ksv is the Stern-Volmer

quenching constant.

In the cases where the relationship of lo/I with [Q] is linear the quenching process can be

either collisional or static. However, those processes can be distinguished experimentally by









carrying out fluorescence measurements at diverse temperatures. Thus, the quenching

mechanism is mainly ascribed to a collisional process when at higher temperatures the system

present larger values ofKsv. Conversely, for the cases of quenching via complex formation,

higher temperatures show smaller Ksv; in this case Ksv is related to the association constant of

the complex (Ka). In situations, where the fluorescence ratio as a function of the quencher

concentration reveals a positive deviation of the linear Stern-Volmer plot, it corresponds to a

combined process of both collisional and static quenching. For such a case a modified Stern-

Volmer equation is given by Equation 1-7.

2 =(1+ Ksv[Q])(1+ Ka[Q]) (1-7)
I
Therefore, the upward curvature in the Stern-Volmer plot is due to the contribution of the

square term (second power) of [Q] in the Equation 1-7.

FTIR-ATR spectroscopy

Fourier Transform Infrared Attenuated Total Reflection spectroscopy is an appropriate

technique for the characterization of the surface of materials. This technique combines the

capabilities of chemical structure identification offered by the infrared spectroscopy mutually

with the fundamental properties of the total internal reflection techniques. The origin of the

internal reflection techniques is involved with the fact that radiation propagating in an optical

medium having higher refractive index (nj) experiences total internal reflection at the interface

of a contiguous sample medium with lower refractive index (n2). This formed propagated wave,

which penetrates a small portion of the sample medium, is called evanescent.133 This

phenomenon takes place only when the angle of incidence of the radiant wave exceeds a critical

angle, 0c, which can be determined by using Snell's law (Equation 1-8).

sin 01 n
sin 0(1-8)
sin 82 n1









It is known that in the position of the critical angle, the angle of refraction, 02, is 90.

Consequently, the angle of incidence is the critical angle and the Snell's equation is reduced to

Equation 1-9:


sin 0c = n- (1-9)
ni
The critical angle, 0c, can therefore be found simply by knowing the refractive indexes of

the two mediums.

It is worth mentioning that the evanescent wave is nontransverse, which allows all its

vector components to interact with the dipoles in all directions, thus this wave is a very

informative probe of the sample material.134

During FTIR-ATR analysis, the infrared radiation is passed through an infrared

transmitting crystal characterized by having a high refractive index, allowing the radiation to

reflect within the crystal several times. Along with each reflection in the transmitting crystal, at

the top surface, the evanescent wave penetrates the sample in order to explore the surface

material. It should be pointed out that the depth of penetration and the total number of

reflections along the crystal can be controlled either by varying the angle of incidence or by

selection of crystals. According to Equation 1-10, the penetration depth of the evanescent wave

in the sample material can be controlled by the selection of transmitting crystal, since each

crystal posses its own refractive index.


d = A (1-10)



Where 2 is the wavelength of the IR radiation, n2 and nl are the refractive index of the

sample and the refractive index of the crystal respectively, and 0 is the angle of incidence.

Therefore, crystals with higher refractive index allow lower penetration depths. The depth of









penetration of the evanescent wave, d, is defined as the distance generated at the crystal-sample

interface where the intensity of the evanescent decays exponentially to 37% of its original

value.135

X-ray Photoelectron Spectroscopy (XPS)

XPS also known as ESCA (Electron Spectroscopy for Chemical Analysis) alike to FTIR-

ATR is a surface analysis technique, which defer in the depth of sample analysis and obviously

in the incident source of excitation. For example, XPS probes only 5 to 10 nm depth whereas

FTIR-ATR allows the detection of chemical species found in the first few microns of the sample

material.136 The XPS analysis is based upon the principles of the photoelectric effect, where the

surface material is irradiated with monochromatic X-rays creating photoelectrons that are

emitted from the sample surface. These ejected electrons come from the core level electrons of

the atoms dispose on the surface material.137 The emitted energy from the photoelectrons is then

analyzed by an electron analyzer spectrometer to determine the binding energy of the electrons.

From the binding energy and intensity of a photoelectron signal, it is possible to determine the

elemental identity, chemical state, and quantity of an element. It should be pointed out that

experimentally the kinetic energy (Ek) of the electrons is the physical parameter measured by the

electron analyzer spectrometer and not its binding energy (Eb). However, it is really Eb that

provides the specific properties of an electron, because Ek is not an intrinsic property of the

analyzed material since this also depends on the photon energy of the X-rays employed during

the analysis. The Equation 1-11 includes all the parameters involved in the XPS analysis and

combines the relationship between Eb and Ek.

Eb h -E -W (1-11)









Where hv is the photon energy and Wis the spectrometer work function. Thus, the binding

energy of the electron is simply measured by the other three quantifiable parameters included in

Equation 1-11.

The XPS spectrum is composed only for the peaks of those electrons that escape the

surface without energy loss. The electrons that suffer inelastic scattering and undergo energy loss

contribute to the background of the spectrum.137

As any other energy excitation process, the excess of energy in the excited state must be

released or transformed in some way of energy. Thus in XPS, the ionized atom formed after the

emission of a photoelectron must have some type of relaxation mechanism. This might be

accomplished by X-ray fluorescence emission or by ejection of an Auger electron. This

relaxation mechanism explains why during the XPS process, peaks for Auger electrons are also

observed.

Scanning Electron Microscopy (SEM)

SEM is a microscope that uses electrons rather than light to form an image. It is created by

focusing a high energy beam of electrons onto the surface. Once the focused electrons hit the

surface, secondary and backscattered electrons are ejected from the specimen.138 Those are

collected in specific detectors and converted into a signal that is sent to a TV scanner where the

image is processed and viewed as a digital image.

It is worth mentioning that secondary electrons are low energy electrons (< 50 eV) that are

originated within a few nanometers depth of the specimen. Whereas backscattered electrons are

those electrons generated after incident electrons are scattered within the specimen volume and

they are reflected or backscattered out of the specimen interaction volume keeping high energy.

Backscattered electrons are useful to distinguish contrast between areas with different chemical

composition because the image created with them depends on the mean atomic number of the









substances that constitute the specimen.138 Usually the SEM instruments are equipped with a set

of two detectors to allow separate observations of topography (TOPO) image and composition

(COMPO) image.

Impedance spectroscopy

Impedance spectroscopy is an effective method to probe interface properties of surface-

modified electrodes. Generally there are two categories in impedance measurements: non-

faradaic and faradaic impedance. Non-faradaic impedance is performed in the absence of any

redox probe, which has been a method used for measuring changes in the electrical properties of

synthetic membranes.139' 140 Instead, electrochemical impedance is a faradaic impedance

technique that is performed in the presence of a redox probe. Impedance is usually measured

applying an AC potential to an electrochemical cell or electrical circuit and measuring the

current going trough these.141,142

Impedance is expressed as a complex number. The complex impedance can be represented

as the sum of real, Zre (co), and imaginary, Zim (co), components originated mainly from the

resistance and capacitance of the system, respectively. The general electronic circuit, includes

the ohmic resistance of the electrolyte, Rs, the Warburg impedance, Zw, resulting from the

diffusion of ions from the bulk electrolyte to the electrode, the double layer capacitance, Cdl, and

the interfacial electron transfer resistance, Ret, present when a redox probe exist in the

electrolyte. In the cases of non-faradic impedance, as measuring the electrical properties of

synthetic membranes, this resistance can be associated with the changes in resistance across the

membrane, Rm. The components Rs and Rw represent bulk properties of the electrolyte solution

and diffusion features of the probe in solution. The other two components in the circuit, Cdl and

Ret, represent interfacial properties of the electrode and they are affected by immobilization of

bio-organic material onto electrode surfaces.141 Thus, analysis of Cdl and Ret can give important









information about the extent of changes of the surfaces properties resulting from coupling of

biomaterial to the studied interface.142

A typical impedance data is represented in the form of Nyquist plot. The shape of this plot

includes a semicircle region lying on the Zre-axis followed by an incline line. The semicircle

region observed at higher frequencies corresponds to the kinetic limited process. Whereas the

linear part characteristic of the lower frequency range represents diffusional limited process. In

the Nyquist plot Rs can be found by reading the real axis value at the high frequency intercept.

While the sum of Rm and Rs can be found by fitting the semicircle until the real axis value at the

lower frequency intercept.

Nyquist plot present one shortcoming, when looking at any point on the plot, it cannot tell

what frequency was used to record that point. There is another popular representation of the

impedance data called the Bode plot. For this, the impedance is plotted with log frequency on the

x-axis, and both the absolute value of impedance and the phase-shift on the y-axis.141









CHAPTER 2
TRANSPORT AND FUNCTIONAL BEHAVIOR OF ANTIBODY-FUNCTIONALIZED
NANOTUBE MEMBRANES

Introduction

Today separation and analyses of proteins present important challenges for modern

bioanalytical chemistry. It is clear that the most important frontier in the biomedical sciences

and in the disease diagnostic is proteomics. The fact that mainly proteins execute and control

cellular phenomena opens up the use of proteins as important molecular targets for drugs, and

biomarkers for diseases.143 Proteomics aims to understand complex biological systems by

analyzing protein expression, protein function, protein modifications, and protein interactions.144

Within the proteome-analysis strategies, separation techniques play an important role.145 The 2D

gel electrophoresis in combination with mass spectrometry is the most widely used

separation/identification technology in proteomics.145- 147 However, there are a number of

approaches to separate proteins based on their biochemical and biophysical features such as

molecular weight, hydrophobicity, surface charge, isoelectric point, etc. that have also been

developed and employed. 14- 151 In addition to these chromatographic methods, synthetic

membranes have been used for protein separations.3 7 152-164 Nevertheless, the membranes used

to date typically separate proteins on the basis of size and charge, and these rudimentary forms of

selectivity are insufficient to allow membranes to compete with chromatographic methods for

protein separations.

The use of affinity chromatography, however, has the advantage of offering a process that

is specific for the protein of interest, overcoming the lack of selectivity and simplicity that many

of the common methods of analysis undergo. We report here systems that exploit the advantages

of affinity chromatography in a nanometer-scale membrane-based platform. In that search for

new advanced membrane materials, our group has conducted fundamental studies in the










development of functionalized-membranes with a selective transport function executed through a

specific molecule "carrier" to modulate the transport during membrane-based separations.9'10

The "carrier" is a mediator that selectively binds to the specific chemical solute, the target

compound to be separated from a sample solution. Depending on its kinetic interactions with the

target compound, the carrier allows enhanced or retarded transport across the membrane.

Here we describe the use of antibody-functionalized nanotube membranes for the

separation of proteins. Since antibodies that selectively bind proteins are well known,165, 166 it

occurred to us that this antibody-based approach might provide a general route for preparing

highly selective membranes for protein separations. To test this hypothesis, membranes

containing silica nanotubes were prepared using the template synthesis method.9' 25, 85 In the

inside walls of these nanotubes, purified mouse immunoglobulin IgG (M-IgG) or human

immunoglobulin IgG (H-IgG) were covalently attached, using an aldehyde-containing silane

reagent (Figure 2-1).9






NH,
0 N-H N
II Ij -H20 II
H-C -- H-C-OH H-C
Covalent Schlff
coupling base by
Aldehyde-Silane dhydraton
Nanotube Membranes
Figure 2-1.Immobilization of antibodies on the template synthesized silica nanotube membranes.

To study the functional behavior of the IgG-functionalized membranes, competitive

transport experiments were done, using antibodies of anti-mouse IgG (anti-M-IgG) and anti-

human IgG (anti-H-IgG) fluorescent labeled with Alexa 488 or Alexa 594. The studies of

transport properties for these fluorescent labeled antibodies demonstrated that the nature of

immobilized antibody changes the functional behavior of the functionalized membranes and the









transport properties of the permeant proteins across them. Additional characterization of these

antibody-based membranes showed a Langmuirian shape plot characteristic of a typical

facilitated transport membrane when the effect of feed solution concentration on the fluxes of

these fluorescent-labeled antibodies was investigated.

Experimental

Materials

Commercially available nanopore alumina membranes (60 [m thick, nominal pore

diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS)

(Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes.

The purified mouse immunoglobulin (M-IgG) immobilized within the silica nanotube

membranes was obtained from Sigma Aldrich. Membranes modified with purified human

immunoglobulin (H-IgG) obtained from Sigma Aldrich were also prepared. A recombinant

protein G obtained from Sigma Aldrich was used for well oriented antibody-immobilization

membranes. A trimethoxysilyl poly(ethylene glycol) (PEG-Si, MW 460-590 Da) obtained from

Gelest (Morrisville, PA) was used to modify the silica nanotubes control membranes. A

trimethoxysilyl propyl aldehyde (United Chemical Technologies) was used to attach the M-IgG

or H-IgG to the nanotubes. Two whole antibodies raised in goat against M-IgG were used to

investigate the transport properties of the M-IgG-modified membranes. These anti-M-IgGs were

obtained from Molecular Probes (Eugene, OR) conjugated with either the dye Alexa 488 or the

dye Alexa 594.

The transport properties of two control IgGs, raised in goat against human IgG (H-IgG),

were also investigated. These anti-H-IgGs were also obtained from Molecular Probes

conjugated with either Alexa 488 or Alexa 594. It is worthy to mention that these four

conjugated antibodies were also used to investigate the transport properties of the H-IgG









modified membranes. In this case the two control IgGs were the anti-M-IgG. A F(ab)2 fragment

antibody raised in goat against human IgG (F(ab)2-anti-H-IgG) and labeled with FITC

fluorophore was also used for transport experiments. This F(ab)2-anti-H-IgG antibody was

obtained from Molecular Probes. SuperBlock blocking buffer solution in phosphate-buffered

saline was obtained from Pierce. SuperBlock solution contains a proprietary protein for

blocking binding sites for non-specific protein adsorption. Tween 20 was obtained from Sigma

Aldrich. All other chemicals were of reagent grade and used as received. Purified water,

obtained by passing house-distilled water through a Barnstead, E-pure water-purification system,

was used to prepare all solutions.

Preparation of the Silica Nanotube Membranes

Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-

based template-synthesis method.9'25, 85 Briefly, a sol-gel precursor solution was prepared by

mixing absolute ethanol, TEOS and 1 M HC1 (50:5:1 by volume), and this solution was allowed

to hydrolyze for 30 minutes. The alumina membrane was then immersed into this solution for

two minutes under sonication. The membrane was then removed and the surfaces swabbed with

ethanol to remove precursor from the alumina surface. The membrane was then air dry for ten

minutes at room temperature and cured in an oven for 12 hours at 150 oC. This yields silica

nanotubes, with wall thickness of 3 nm, embedded within the pores of the alumina membrane.9

Antibody Immobilization

The trimethoxy aldehyde silane was first attached to the inside walls of the silica

nanotubes to introduce aldehyde terminal groups onto the pores of the membrane.9 The

membrane was then immersed into a solution that was 1.0 mg per mL in either M-IgG or H-IgG

dissolved in 10 mM phosphate-buffered saline with pH = 7.4 that also contained 150 mM

sodium chloride and 5 mM sodium azide. The membrane was incubated overnight at 4 oC,









which allowed free amino sites on the antibody to attach via the Schiff's base reaction to the

nanotube-immobilized aldehyde groups9. After immobilization, the membranes were rinsed with

buffer and then incubated for 5 hours in the SuperBlock solution that also contained 0.05%

Tween 20.

Transport Experiments

The M-IgG- or H-IgG-modified membranes were sandwiched between two pieces of

Scotch tape that had 0.31 cm2 area holes punched through them. These holes defined the area of

membrane exposed to the contacting solution phases. This membrane assembly was mounted

between the two halves of a U-tube permeation cell, with half-cell volumes of -10 mL.4 6, 10,22,23






i'e
I -






Figure 2-2.Flow-through fluorescence system used to measure the flux of the permeating
proteins.

The transport properties of three types of membranes were investigated membranes

containing immobilized M-IgG, membranes containing immobilized H-IgG, and control

membranes with attached PEG-Si. The permeating proteins, whose transport properties were

investigated, were the fluorescently-labeled anti-M-IgG and anti-H-IgG. The feed solution for

all experiments was equi-molar in these two proteins, dissolved in 10 mM phosphate-buffered

saline pH 7.4 that also contained 150 mM sodium chloride and 5 mM sodium azide (buffer used

for all studies). The receiver solution on the other side of the membrane was identical except it

initially contained no protein.









The fluorescence intensity of the receiver solution was periodically measured using two

flow-through fluorescence detectors connected in series (Figure 2-2).

This system consisted of an L-7485 fluorescence detector (Hitachi) and an RF-2000

fluorescence detector (Dionex). The flow stream was delivered using an L-7100 HPLC pump

(Hitachi). We discovered that if the receiver solution was sent continuously through the

fluorescence detectors during the entire course of a transport experiment, significant photo

bleaching of the Alexa dyes occurred. To prevent this, we turned the pump on to deliver receiver

solution to the detectors for only 3 minutes per hour of transport experiment. That is, we used a

duty cycle of 57 minutes with pump off followed by 3 minutes with pump on (flow rate = 3 mL

per min) to collect fluorescence data. The whole system was controlled using Chromeleon 6.50

chromatography management software (Dionex) through a UCI-100 universal chromatography

interface (Dionex).

The Hitachi detector was used exclusively to detect the Alexa-594-labeled anti-M-IgG and

the Dionex detector was used for the Alexa-488-labeled anti-H-IgG. Alexa 488 was excited at

495 nm, and the emission was detected at 515 nm. Alexa 594 was excited at 590 nm and

detected at 617 nm. Calibration curves were obtained with the same flow through system on

solutions containing known concentrations of both of the flourophore-labeled antibodies. These

calibration curves were used to obtain the concentrations of the transport proteins in the receiver

solution.

Competitive Binding Assays

Our experimental design assumes that the permeating protein anti-M-IgG binds strongly to

the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier.

Likewise we assume that anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-IgG

binds weakly to this carrier. We proved these assumptions to be true by conducting competitive









binding assays. This was accomplished by equilibrating H-IgG- and M-IgG-modified

membranes to solutions that were 21.2 nM in both anti-M-IgG and anti-H-IgG, and then

measuring the fluorescence intensities from the membranes.

Calibration data were used to convert these fluorescence intensities to pmoles of anti-M-

IgG and anti-H-IgG bound to the membrane. The standards used to obtain the calibration curves

were 0.31 cm2 membrane samples to which were applied known quantities of the desired protein.

The fluorescence intensities from these standards were measured using a fluorescence

microscope that was equipped with both a fluorescence detector and a CCD camera. This system

combines an Axioplan 2 imaging microscope (Zeiss) with a J&M-PMT photometry system

detector (SpectrAlliance), for measuring the fluorescence intensity using a BP450-490 filter for

excitation and a LP515 filter for green emission of light, and a BP546/12 filter for excitation and

a LP590 filter for red emission of light.

Determination of Static Protein Binding Capacity and Binding Constants through
Adsorption Isotherms

The adsorption isotherms were carried out with M-IgG modified, H-IgG modified, and

unmodified control membranes at 23 C in stirred U-tube permeation cells using the same buffer

conditions used for the permeation studies, via solutions containing single concentrations of anti-

M-IgG or anti-H-IgG fluorescent labeled with Alexa 488. In the case of M-IgG-modified but

immobilized via protein G, the adsorption isotherms were carried out at three different

temperatures: 23, 32, and 42 C. The fluorescence intensity of all the membranes was measured

under the fluorescence microscope. Calibration curves were used to obtain the concentration of

the Alexa 488-labeled proteins adsorbed to the membrane sample from the measured

fluorescence intensity. The data were analyzed according to the monolayer Langmuir model.167









Effect of Equilibration Time on Flux and Selectivity via Dynamic Binding Capacity

At the start of a transport experiment, all of the M-IgG or H-IgG binding sites are empty

and these sites fill with anti-H-IgG or anti-M-IgG during the course of the experiment. We were

concerned that binding equilibrium might not be achieved during the course of a transport

experiment and that our results reflected this lack of equilibration. To explore this issue we

investigated the flux and selectivity as a function of equilibration time. This was accomplished

using seven membranes modified with M-IgG, which were exposed to a solution that was 21.2

nM in both of the permeating proteins, anti-H-IgG and anti-M-IgG.

After one day of exposure, the first membrane was rinsed with buffer and a transport experiment

was run (for 24 hours) using a feed solution that was 21.2 nM in both of the permeating proteins.

After the second day of exposure, the same procedure was repeated for the second membrane,

and so on, until the seventh membrane was investigated after seven days of exposure. Thus, after

the binding sites available in the membrane are filled, the dynamic binding capacity is reached to

conduct the transport experiments.168

Results and Discussion

The nanotube membranes investigated here contain a bound carrier, either M-IgG or H-IgG.

The transport selectivity coefficient (ca) across of the membranes was evaluated by comparing

the flux of a permeating species that weakly binds to these carriers to the flux of a permeating

species that binds stronger. In the case of the M-IgG-modified membrane, the strongly and

weakly binding species are anti-M-IgG and anti-H-IgG, respectively. In the case of the H-IgG-

modified membrane, the strongly and weakly binding species are anti-H-IgG and anti-M-IgG,

respectively. These studies begin, however, with electron microscopy characterization of the

alumina template and investigations of the fluxes of the two permeating proteins, anti-H-IgG and

anti-M-IgG, across PEG-Si-modified control nanotube membranes.









Electron Microscopy and Calibration Data

Figure 2-3 shows electron micrographs of the surface and cross section of the alumina

template membrane and of silica nanotubes deposited within the pores of such membranes. The

alumina membrane has two different faces the face that was in contact with the underlying

aluminum during anodization (branching side), and the face that was in contact with the

electrolyte during anodization. The shape of the pores has a polygonal aspect on the branching

side, while in the solution anodization side is rounded. Note that in the transport experiments the

nanotubes were left embedded within the pores of the membrane, but they were removed here so

that they could be imaged. The tubes have an outside diameter of 18522 nm, as would be

expected given the pore diameter, 18815 nm in the solution anodization side and 11325 nm

for the branch side. The diameters for the alumina template were determined by electron

microscopy based on the analysis of twenty pores in four different membranes. It is well known

that the pores in these alumina templates branch at one surface due to the voltage-reduction

method used to remove the membrane from the underlying Al surface.44









Figure 2-3.Electron micrographs of (A) branch side, (B) solution anodization side, and (C) cross-
section of the alumina template membranes, and (D) silica nanotubes after removal
from the membrane.

Evidence for this branching can be seen at the ends of some of the silica nanotubes (Figure

2-3D). The length of the pores with larger size cover most of the 60 jtm thickness of the alumina

template, while the smaller pores cover less than 200 nm of the membrane.










Figure 2-4 shows typical calibration data and a typical calibration curve. The calibration

data (Figure 2-4A) were obtained by pumping the indicated concentrations of Alexa 488-labeled

anti-H-IgG through the Dionex detector (Figure 2-2). The calibration curve obtained from these

data (Figure 2-4B) showed excellent linearity (correlation coefficient = 0.996) over the range

from 20 pM to 1280 pM. Analogous data were obtained for Alexa 594-labeled anti-M-IgG using

the Hitachi detector.

180
A 128pM 160 B
-E 150 = 140 -
S120 -
120
S. 1 0 0
90 640 p 80
'0 p0
320 pM 60 /
S160 pM
80M 40 -
S8300PM
0 40pM p 20
L P20 pM
S OpM -0
0 0 -
0 10 20 30 40 0 200 400 600 800 1000 1200 1400
Time (nin) Anti-H-IgG Concentration (pM)
Figure2-4.Typical plots of (A) calibration data, and (B) calibration curve for Alexa-488 labeled
anti-H-IgG.

Transport Properties of the anti-IgGs through PEG-Si-modified Control Membranes

The transport data were processed via "flux plots"9' 10 moles of permeating protein

transported across the membrane vs. time. Flux plots for feed solutions containing equi-molar

concentrations of anti-M-IgG and anti-H-IgG across PEG-Si-modified control membranes are

shown in Figure 2-5; data for two different feed concentrations are shown. These plots were

linear (correlation coefficients >0.998) indicating that Fick's first law is obeyed and that over the

time window investigated, the feed solution was not appreciably depleted of the transporting

proteins.

Figure 2-5 shows clearly that the fluxes of anti-M-IgG and anti-H-IgG are the same in the

PEG-Si-modified control membranes. This means that the diffusion coefficients for these two










proteins are the same, which in turn means that the proteins are identical in size, as would be

expected for two IgG molecules. Therefore, the transport properties across these control

membranes are the same for these two antibodies. These points are further reinforced by the

concentration dependence of the flux. The concentration used for the lower two curves was

eight-times lower than the concentration used for the upper two curves. Correspondingly, the

flux at the high concentration (1.45 pmoles h-1 cm-2) is eight times higher than the flux at the low

concentration (0.18 pmoles h-1 cm-2). After the permeation experiments, the membranes were

observed under the fluorescence microscope, but no fluorescence was detected. This indicates

PEG-Si-modification suppresses non-specific protein adsorption to the silica nanotubes.

10
9 -
8d


7
0 6 }
5 -
0
i8
4 4
0 4
o 3


A0 A

l;lelrlsin--is
0 200 400 600 800 1000 1200 1400
Time (min)

Figure 2-5. Flux plots for anti-M-IgG (A) and anti-H-IgG (m) through PEG-Si-modified control
membranes. The feed concentration was 2.65 nM (lower curves) and 21.2 nM (upper
curves).

Competitive Binding Studies

As noted above, our experimental design assumes that the permeating protein anti-M-IgG

binds strongly to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to

this carrier. Likewise we assume that anti-H-IgG binds strongly to the carrier H-IgG and that










anti-M-IgG binds weakly to this carrier. We proved these assumptions to be true by conducting

competitive binding assays.

Figure 2-6A shows the fluorescence spectra of the permeating proteins bound to a M-IgG-

modified membrane after equilibration with a solution that was 21.2 nM in both anti-M-IgG and

anti-H-IgG. The membrane shows strong fluorescence from anti-M-IgG and very weak

fluorescence from anti-H-IgG. Calibration data allowed us to convert these fluorescence

intensities to moles of each protein bound to the surface. In agreement with our design

paradigm, 10-times more anti-M-IgG was bound to the M-IgG-modified membrane than anti-H-

IgG. The small amount of anti-H-IgG bound is due to antigen cross reactivity, as a result of the

polyclonal characteristic of the antibodies.169- 172 Similarly, 14-times more anti-H-IgG was

bound to the H-IgG-modified membrane than anti-M-IgG (Figure 2-6B).


35000 45000
A B
S3000040000
.9 30000 b P a
35000
U 25000 -
C 30000
( 20000 25000
15000 20000
1L 15000
10000 -
10000 -
5000 5000
0 I 0
400 500 600 700 800 400 500 600 700 800
Wavelength (nm) Wavelength (nm)
Figure 2-6.Fluorescence spectra of (A) M-IgG- and (B) H-IgG-modified membranes after
equilibration with a solution that was 21.2 nM in both (a) anti-H-IgG, and (b) anti-M-
IgG.

These experiments also confirmed the success of the chemistry immobilization and the

capability of the antibodies immobilized in the nanotube membranes to preserve their bioactive

properties on the porous surface.









Functional Behavior of the M-IgG- and H-IgG-modified Membranes

In these studies, we have been exploring the transport properties, and separation of

different anti-IgGs through antibody-modified nanotube membranes, and how the type of

immobilized antibody alters the functional behavior of these functionalized membranes.

Although, porous membranes have been used for size-based protein separation with great

success, separation of proteins with similar size is more difficult.7 This process is even more

challenging, when the separation is between proteins that have practically the same molecular

structure and size as in the case of different anti-IgGs.

Flux plots for the transport of anti-M-IgG and anti-H-IgG across the M-IgG-modified

nanotube membrane (Figure 2-7A) are linear, as are the flux plots for these permeating proteins

for a H-IgG-modified membrane (Figure 2-7B). The fluxes obtained from the slopes of these

plots are shown in Table 2-1. As mentioned above, the selectivity coefficient (a) across of the

membranes was evaluated by comparing the flux of a permeating species that binds weakly to

the carrier to the flux of a permeating species that binds more strongly.

Table 2-1.Fluxes and selectivity coefficients a for anti-M-IgG and anti-H-IgG across the M-IgG
and H-IgG-modified membranes
IgG-attached to the Flux of anti-M-IgG Flux of anti-H-IgG
membrane (pmoles h-1 cm-2) (pmoles h-1 cm-2)

M-IgG 0.60 1.31 2.2

H-IgG 1.33 0.48 2.8



It was observed that for both membranes the flux of the species that binds strongly to the

carrier (e.g., anti M-IgG for the M-IgG-modified membrane) is lower than the flux of species

that binds weakly to the carrier. This is a clear indication that the transport properties for these










two permeating proteins are different, which in this situation, depends on the functional behavior

of the IgG-functionalized nanotube membrane.

It is worth mentioning that under steady state conditions observed in all the transport

experiments, only an averaged transport rate can be determined because proteins being

transported do not enter membrane pores at the same time so that the measured signal is a

spatially and temporally averaged one.

10
8- A .* B
8C A
A
t:6 -* t
6. n 6 -
T C( i )
A 4- A

A A A A
0 .A A
o ^ I
E2 A A A E A
0 A 0 2 A
A. ono A, I

0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200
Time (min) Time (min)
Figure 2-7.Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in (A) M-IgG-modified membrane,
and (B) H-IgG-modified membrane. The feed solution concentration was 21.2 nM in
both of the permeating proteins.

Observed Functional Behavior is not a Transient Phenomenon

One way to interpret the results in Figure 2-7 and Table 2-1 is that they might reflect a

transient phenomenon. That is, one could hypothesize that during the permeation experiments

summarized in Figure 2-7, the flux of the species that bound strongly to the carrier were

suppressed because all the sites for binding were initially empty. According to this hypothesis

the flux for the strongly binding species would remain low until all sites were filled and

breakthrough occurred.

It is difficult to reconcile this hypothesis with the constant flux observed for the strongly

binding species over these nearly one-daylong transport experiments (Figure 2-7). One would










anticipate instead, very low rates of transport at short times and higher rates at longer times as

the sites fill. Nevertheless, we designed a set of experiments to test this hypothesis (Figure 2-8).

We found that the flux of the weakly-binding species (in this case anti-H-IgG) remained higher

than that for the strongly binding species (anti-M-IgG) even after 1 week of equilibration with a

solution equi-molar in both proteins. Interestingly, the selectivity coefficient (a) improved

slightly over the time course of this experiment. The other important point to glean from Figure

2-8 is that the data indicate that the membrane-bound M-IgG remains bioactive for the weak-

long period summarized in the figure.

1.6
anti-H-IgG
1.4 anti-M-IgG





0.6
0.8


L. 0.4

0.2

0.0
0 1 2 3 4 5 6 7
Equilibration Day
Figure 2-8.Dynamic binding capacity study using M-IgG-modified membranes. The flux and
selectivity coefficient vs. days of equilibration are shown for the transport of anti-M-
IgG and anti-H-IgG. The feed solution concentration was 21.2 nM in both of the
permeating proteins.

It was also observed that some of the protein transmission was lost across membranes with

more than 4 days of saturation. This effect was ascribed to the fact that of protein bound to the

M-IgG-modified membranes decreased the effective pore size for the remaining diffusing protein

molecules. Besides, such diminution in flux is not uncommon with protein-transporting

membranes and results from gradual fouling of the membrane.162, 173- 176










It was also found that after equilibration the weakly binding protein (anti-H-IgG) diffuse

more rapidly than the strongly-binding protein (anti-M-IgG). Therefore, there is a trade-off

between selectivity and flux after the membranes are equilibrated. The equilibration reduces the

flux for the strongly binding protein with respect to the weakly binding one, but leads at the same

time to an increase in selectivity. Thus, it can be concluded that the M-IgG carrier without

equilibration in some extend facilitate the transport of anti-M-IgG when is compared with

transport across membranes that have been previously equilibrated.

Flux vs. Feed Concentration

A number of competitive permeation experiments were carried out at different feed

concentration, ranging from 2.65 nM to 0.5 [M using M-IgG modified membranes (Figure 2-9).



3.0 -
3.0









S1.5 0 100 200 3 40 500
1.0
L. 0.5
0.0
S 100 400 500
Feed Concentration (nM)
Figure 2-9.Flux versus feed concentration for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG
modified membrane. The feed solution concentrations were equimolar in both of the
permeating proteins.

Figure 2-9 shows flux-vs.-feed plots for transport of anti-H-IgG and anti-M-IgG across M-

IgG-modified membranes. The flux for anti-H-IgG was at all times higher than the flux for anti-

M-IgG. These plots show a characteristic "Langmuirian" shapes,9' 10, 109 demonstrating that the

M-IgG attached in the membrane indeed interacts with the penetrant-antibody, acting as a carrier

during the permeation process. A straight line would be expected, if there is not interaction










between carrier and permeating antibodies at the very low feed concentration used during these

studies.

A key way to prove that transport across a membrane is facilitated is to plot the flux vs. the

feed concentration.9' 10' 109 For facilitated-transport membranes, such plots are linear at low feed

concentration but flatten (Langmuirian shape) 9, 10, 109 at high concentrations. Data for both of the

permeating proteins show the characteristic Langmuirian shape of facilitated transport.

Membranes that contained attached PEG-Si, instead of attached M-IgG showed linear plots as is

shown in Figure 2-10. Therefore, the antibody-functionalized membranes showed here exhibit a

type of behavior for facilitated-transport membranes during protein separations.


8 -

E
Q 6-

4-
04
E
0

0
L2



0 20 40 60 80 100
Feed Concentration ( nM )
Figure 2-10.Typical plot of flux versus feed concentration for anti-M-IgG or anti-H-IgG across
PEG-Si modified membrane. The data plotted is an average of the fluxes for anti-M-
IgG and anti-H-IgG for each specific feed concentration.

Effect of the Pore Size on the Flux and Selectivity Coefficient

The influence of the pore size on the flux and selectivity coefficient was explored by using

three different sizes of commercially available (Whatman) nanopore alumina membranes having

60 rm thickness, and nominal pore diameters respectively of 200, 100, and 20 nm. To study this

effect, nanotube membranes of the specified pore size were modified with M-IgG. For the










transport experiments, the feed solution concentration was in all cases 2.65 nM in both of the

permeating proteins anti-H-IgG and anti-M-IgG (Figure 2-11A).

0.16
A 4.5 -
0.14 A 4
o.r / 34.0 -
E 0.12 3.5 -

0.10 3.0
E 0
0.08 -2.5
X '^ 2.0 --
S0.06 .
S1.5- *
0.0 1.0
201200 200120 100 200 20/200 200/20 100 200
Pore Size (nm) Pore Size (nm)
Figure 2-11.Effect of the pore size on (A) the flux, and (B) selectivity coefficients for M-IgG
modified membranes. The feed solution concentrations was 2.65 nM both in anti-H-
IgG and anti-M-IgG.

At this specific feed concentration (2.65 nM) not appreciable differences for the flux of

anti-M-IgG across these M-IgG modified membranes were observed. However, lower fluxes

were obtained for anti-H-IgG as the pore size decreases (Figure 2-11A). This reduction on flux is

due mainly to the resistance that the transported antibodies experience in the pores that arises

from the hydrodynamic interaction of the protein molecules with the wall of the pore, which

include far field and near field interactions.177 178 Interestingly, the flux for the 200 nm

membrane decreases slightly compared with the 100 nm membrane. This reduction in flux was

ascribed to the fact that 200 nm pore size membranes are much easier of functionalize than 100

nm one. Therefore, the data suggests that the number of binding carrier sites immobilized in this

membrane is possibly higher than the 100 nm, which in turn decreases the effective pore size of

the membrane. Additional evidence for this conclusion was obtained from the comparison of

fluorescence spectra of the membranes after transport experiments, which showed higher

fluorescence intensities for the 200 nm M-IgG modified membrane.











It was observed that the selectivity coefficient (a) increases as the pore diameter increases

(Figure 2-11B). The increase in selectivity with bigger pores was attributed to the availability of

much more room to diffusive for the protein that weakly binds the M-IgG carrier, which in this

specific case corresponds to anti-H-IgG. In addition, this effect was observed in the case of the

20/200 nm membranes when the transport was carried out placing the feed solution on the 200

nm side of the membrane (Figure 2-11 A).

Effect of the Immobilized-Antibody Orientation on the Flux

This study was accomplished by taking advantage of the specific affinity that protein G has for

the Fc region of the IgG molecule.121' 170, 179, 180 Thus, it was possible to manipulate the

orientation of the immobilized antibody on the walls of the nanotubes.


4 .


3 n
0
o U

'2


E A
0 -
"C AL Ak
A

0 200 400 600 800 1000 1200
Time (min)
Figure 2-12.Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in well-oriented M-IgG-modified
membrane. The feed solution concentration was 21.2 nM in both of the permeating
proteins.

It is known that due to steric hindrance and random orientation of the antibody molecules

on the solid-phase surface, the binding activity of antibody immobilized directly on the solid-

phase surface is usually less than that of soluble antibody.179' 180 In order to overcome this

problem, protein G was used to modify silica nanotube membrane for antibody immobilization.

Thus, the well oriented immobilization of M-IgG was carried out on nanotube membranes









modified with protein G that was previously immobilized via aldehyde silane chemistry on the

nanotube walls. The results for the transport experiments of anti-M-IgG and anti-H-IgG across

this well-oriented M-IgG modified-membrane are shown in Figure 2-12. The feed solution

concentration was 21.2 nM in both of the permeating proteins.

The fluxes of anti-H-IgG and anti-M-IgG across this well-oriented M-IgG modified

membrane were 0.64 and 0.20 pmoles h-1 cm2, respectively. As the M-IgG-modified membranes

immobilized by direct immobilization on the nanotube surface, this membranes also

preferentially transport the protein that binds weakly to the immobilized carrier. The fluxes for

anti-H-IgG and anti-M-IgG were respectively two times and almost three times lower compared

with those fluxes for membranes modified with M-IgG but where the immobilization was

directly attached to the silane aldehyde surfaces in the membrane. The reason for these lower

fluxes was attributed to the decrease in the effective pore size in the membrane with the addition

of protein-G and the possibility of having all the M-IgG molecules in a well-oriented

immobilization which also contributes to the decrease in the effective pore size. Besides, the

well-oriented immobilization contributes with a stronger binding affinity, because this helps to

preserve the binding activity of the antibody, which at the same time lowers the diffusion

coefficients of the permeating proteins.

Separation Based on a Simultaneous Combination of Size and Affinity Selectivity

In order to prove this concept of a simultaneous combination of size- and affinity-based

separation of proteins, a nanotube membranes templated in a monodisperse pore size alumina of

-40 nm crosswise the whole 60 |tm thickness was used. Scanning electron micrographs of the

surface and cross section of this membrane are shown in Figure 2-13.











It should be pointed out that the 40 nm pore size of this alumina template is about 3 times

the hydrodynamic diameter for the IgG whole molecule (14 nm).3 Therefore, after M-IgG

immobilization the effective pore size of the nanotube membranes is reduced to a magnitude that

allows a closer interaction of the permeating proteins both with the M-IgG immobilized carrier

and the wall of the pores.














Figure 2-13.Scanning electron micrograph of (A) the surface, and (B) cross section of a highly-
ordered nanoporous alumina membrane obtained at 30 V in 5% oxalic acid.

The next set of experiments demonstrated the capacity of these antibody-modified

membranes to be used with a combination of both size and affinity selectivity.

First, the transport of anti-H-IgG and anti-M-IgG across a control PEG-Si modified

membrane and a well-oriented M-IgG modified membrane are compared (Figure 2-14).

0.6
0.5
A A B
0.5 -

0.4-
o 03 0 .
0 A A
00 A


S0 A
.- 0.1 A* .0 0.1 -, A
' '* o'o
0.0 A.

0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Time (min) Time (min)
Figure 2-14.Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in (A) PEG-Si-modified
membrane, and (B) well-oriented M-IgG -modified membrane. The feed solution
concentration was 21.2 nM in both of the permeating proteins.










The permeation process through membranes can be divided into its transient and steady

state components.181,182 The transient component can be characterized by the time lag. This time

lag observed phenomenon correspond to the time at which the permeating specie enters the

membrane to the time that the flow rate of diffusing species into the close volume reaches a

steady state of permeation.181, 182

Figure 2-14A shows the transport of anti-M-IgG and anti-H-IgG across a PEG-Si modified

control membrane. It is clearly observed a time lag event for both of the permeating proteins.

This effect was even more noticeable for the antibody-modified membrane as can be seen in

Figure 2-14B. By just comparing the time in which anti-M-IgG reaches a steady state of

permeation across PEG-Si control and M-IgG-modified membranes, ten and twelve hour

respectively, we can see that antibody immobilization has a significant effect in the functional

behavior of the membrane.

Therefore, by taking advantage of this time lag effect, transport of antibodies with different

sizes were carried out in order to prove the concept of a simultaneous combination of size- and

affinity-based separation of proteins (Figure 2-15).

0.6

0.5 -

o 0.4- -

0.3

0 0.2 -
0.
o A A

"" 0.1 A

0.0 fLIAAAAAAAA
0 200 400 600 800 1000 1200 1400
Time (min)
Figure 2-15. Flux plots for anti-M-IgG (A) and F(ab)2-anti-H-IgG (m) in well-oriented M-IgG -
modified membrane. The feed solution concentration was 21.2 nM in both of the
permeating proteins.









The difference between the experiment in Figure 2-14B and Figure 2-15 is that for the

experiments in Figure 2-15 a F(ab)2-anti-H-IgG labeled with FITC was transported instead of the

anti-H-IgG whole antibody labeled with Alexa 488. It can be seen from Figure 2-15 that the time

lag is less evident for transport of the Fab fragment antibody than for the whole antibody IgG as

a result of the smaller size of this protein. The other important point to glean from Figure 2-15 is

that during the initial three hours of the experiment the flux of the fragment antibody seems

lower that the whole antibody in Figure 2-14B. The reason for this lower flux was ascribed to the

fact that the sensitivity reach with Alexa 488 is much higher than with the FITC fluorophore.

Therefore, these experiments demonstrated the capacity of these antibody-modified

nanotube membranes to be used with a combination of both size and affinity selectivity for

protein separation.

Binding Capacities and Binding Constants

Adsorption isotherms were carried out to evaluate the protein binding capacity and the

binding constants for the permeating proteins to the antibody carriers. The data were analyzed

using a monolayer Langmuir model and nonlinear regression methods for the fitting. A summary

of the results obtained from the adsorption isotherms carried out with M-IgG modified, and H-

IgG modified membranes, using the 100 nm pore size commercially alumina templates is shown

in Figure 2-16.

By observing the binding capacities in Figure 2-16 obtained from each isotherm, it is

obvious, that these antibody-modified membranes preferentially bound the protein that binds

stronger to the carrier on the membrane as was demonstrated above. It was also observed that a

small amount of the weakly bounding proteins were bound to the carriers, which was due to

antigen cross reactivity. Likewise, the results for the dissociation constant showed the same trend

observed for the binding capacities. Lower dissociation constants (higher binding constant) were










obtained by the proteins that strongly bind to the carrier. The binding constant errors found in

Figure 2-16 for the proteins that weakly bind the carriers may result not only from the possible

lack of equilibrium due to the weak affinity with the carrier, but also from the random orientation

effect of the immobilized antibody. It was also demonstrated that by controlling the orientation

of the immobilized antibody using protein G the potency affinity of the antibodies is effectively

preserved (bars in the right side of Figure 2-16).

90 1600
M K = Dissociation Cta.
80 C = Binding Capacity 1400

70 anti-M-IgGi antiIgG 1200

60 T I-ntiH-lgG N
1000
50 -
anti-H-gG 800
40 -antiM
30 600 ,
30 U
400
20 -

10 200


H-IgG H-IgG M-IgG M-IgG Protein G
Membrane Modification M-IgG
Figure 2-16.Dissociation constants and binding capacities from adsorption isotherms of anti-M-
IgG and anti-H-IgG on H-IgG- and M-IgG-modified membranes, according to the
Langmuir monolayer model.

Effect of the Temperature on the Flux and Binding Constant

For these studies, nanotube membranes templated on the 100 nm pore size commercially

alumina were modified with M-IgG well-oriented immobilized via the protein G immobilization

technique. Transport experiments (Figure 2-17), and adsorption isotherms (Figure 2-19) were

carried out at three different temperatures: 23, 32, and 42 C.

Figure 2-17 shows the flux plots for the transport experiments carried out at the three

studied temperatures. It was observed that as the temperature increases the flux for anti-M-IgG












also increases, which in turn decreases the selectivity coefficient between for these two


permeating proteins. It should be pointed out that the flux of anti-H-IgG also increased at higher


temperatures, but in a less extent compared with the fluxes for anti-M-IgG. A much clearer


illustration of this flux and selectivity behavior with the changes in temperature is shown in


Figure 2-18.

10 10 10

23 C 32C 42 C
8 8 8

t *
S6 6 6





21 2 2
o o


0 0 0
I I I I I I I I
0 400 800 1200 0 400 800 1200 0 400 800 1200
Time ( min )
Figure 2-17.Flux plots for anti-M-IgG (A) and anti-H-IgG (m) in well-oriented M-IgG -modified
membranes showing the effect of temperature. The feed solution concentration was
21.2 nM in both of the permeating proteins.


1.6
anti-H-IgG
1.4 anti-M-IgG
E 1.2
0

1.0

S0.8

0.6

0.4

0.2
0.0
23 32 42
Temperature ('C)
Figure 2-18.Effect of temperature on the flux and selectivity coefficients.


The effect of the temperature in the affinity was study by carrying out adsorption isotherms


for anti-M-IgG with well-oriented immobilized M-IgG at 23, 32, and 42 C (Figure 2-19). As










should be expected adsorption isotherms perform to higher temperatures displayed a more linear

character than the typical Langmuirian shape obtained at room temperature, indicating a lower

affinity for anti-M-IgG.


300 A- 23 C

250 A 32 C

-200 A
S42'C
150 -
0
100 r
0
0. 50 A


0 5 10 15 20 25
Free concentration ( nM)
Figure 2-19.Adsortion isotherms carried out with well-oriented M-IgG-modified membranes for
anti-M-IgG at three different temperatures: 23, 32, and 42 C.

The results of the data analyzed according to the Langmuir monolayer model are

summarized in the Figure 2-20. Higher dissociation constants (Lower binding constant) were

obtained as the temperature is raised. Interestingly, it is the fact that the binding capacity of the

antibody-modified nanotube membranes remains almost constant with the change in

temperature, which is a clearly indication that the functional behavior of these antibody-modified

membranes is governed for the immobilized carrier in the membrane.

It is worth mentioning that previous studies reported in literature have demonstrated that

alumina membranes have low thermal expansion and good stability at high temperatures.183 In

addition, temperature studies ranging from 20 to 60 OC have demonstrated that the permeability

of solutes across these porous membranes increases slightly with temperature compared with the

diffusivity in the bulk solution.184 Therefore, it can be concluded that the enhancement in the

fluxes observed in Figure 2-17, as the temperature was raised, is mainly due to the lower affinity

between the permeating protein and the antibody carrier immobilized on the membrane. An











additional evidence for this conclusion is that the fluxes at higher temperatures for anti-M-IgG

were increased in a more extent compared with the fluxes for anti-H-IgG.

55
-900
50 K = Dissociation Cta.
45 C. Binding Capacity 800
40 700
35 600
3 500
25 -
400
20 --300 UM
15 -
10 200
5 100
0 0
23 32 42
Temperature (C)
Figure 2-20.Affinity temperature dependence of anti-M-IgG in well-oriented M-IgG-modified
membranes

The data obtained in Figure 2-19 and plotted in Figure 2-20, can be used to determine


thermodynamic parameters such as AG, AH, and AS of the system in study. Thermodynamic


costs for a system indicate whether processes are likely to occur and they are conveniently


quantified in terms of free energy AG.185 Therefore, the feasibility of the protein and antibody to


interact (at a constant pressure P, and temperature T) is determined by the change in free energy

of the system, which for a reversible process can be written as (Equation 2-1):186


AGO = AH TASO (2-1)
Where AHo and ASo are the enthalpic and entropic changes experienced during the


biomolecular reaction. The enthalpic part is measure of the average potential energy of

interaction between molecules, and the entropic part is a measure of the order or intermolecular

correlations.185 By assuming that the biomolecular reaction is a reversible process, the

relationship between binding affinity constant (Ka), and free energy change can be expressed by


Equation 2-2.186


AGo =-RT nK, (2-2)









Since the values Ka, T and R the gas constant (8.314 Joules moles- K-1) are known, by

using Equation 2-2 was possible to calculate AGo. The results of these energetic calculations are

summarized in Table 2-2.

Table 2-2.Thermodynamic parameters of AGo associated with the affinity interaction of anti-M-
IgG and M-IgG-modified membranes as a function of the temperature
Temperature Temperature Ka AGO
(C) ( K) (M-1 ) (KJ / moles)
23 296 3.76 x 108 -48.59

32 305 7.81 x 107 -46.08

42 315 3.19 x107 -45.24



As can be seen from Table 2-2, the resulting AGo values were negative, which indicate that

the process is exothermic and proceed spontaneously toward the formation of the complex

protein-antibody. In addition, we can see how small changes in free energy generate significant

changes in the binding affinity constant. Therefore, these changes in energy can be associated

with the variations in the fluxes observed in Figure 2-17 when the temperature was changed.

Now that that the changes of free energy as a function of the temperature were established,

from a plot of AGo vs. Tthe values of AHo and ASo can be extracted from Equation 2-1, by

assuming that AHo is independent of temperature. From the slope and intercept of the obtained

plot with a linear correlation of 0.918 the estimated values for ASo and AHo were 0.17 0.06

KJ/mol and -99.95 17.12 KJ/mol, respectively. Those estimated values confirm the spontaneity

of this biomolecular reaction toward the formation of the protein-antibody complex. Interesting

is that ASo is still positive even though the loss of rotational-translational degrees of freedom

after complex formation. This effect has been extensively studied and had been associated with









partial desolvation and amount of water excluded during the formation of multiple

intermolecular noncovalent forces in the pocket of the binding site.187-190

Thus, these studies confirm that the functional behavior of these antibody-modified

membranes is governed for the changes in free energy before and after interaction of the

permeating proteins with the carrier antibody, which can be attributed to the changes in enthalpy

and entropy of the system as was demonstrated here.

Effect of the Low Ionic Strength on the Selectivity

One method to improve selectivity in membrane-based separations is to shut down the

transport of one of the permeating species. It was found that by using transport solution with low

ionic strength it is possible to contemplate this option with antibody-modified membranes. To

test this option, we studied the rate and selectivity of transport of anti-H-IgG and anti-M-IgG in

only water solution across M-IgG modified nanotube membranes templated in commercial

alumina with 100-nm-diameter pores (Figure 2-21). It was observed that after two hours of

transport in low ionic strength conditions (only water), the fluxes of both the anti-H-IgG and

anti-M-IgG were depleted (Figure 2-21A). The selectivity for these two permeating proteins

through M-IgG membranes without equilibration was only 2.7. Alternatively, it was found that

transport across M-IgG-modified membranes that were previously equilibrated for six hours in a

water solution that was 2.65 nM in both of the permeating proteins (anti-H-IgG and anti-M-IgG),

the selectivity coefficient was increased to 17.3. This effect is shown in Figure 2-21B.

The effect of low ionic strength in the selectivity was attributed to the fact that IgG has no

unique isoelectric point, since each specific antibody molecule has its own variable region

containing different charged residues.191 Thus, for transport experiments carried out at low ionic

strength the thickness of the electric double-layer has a considerable influence in the mobility of

the permeating proteins across the antibody-modified membranes.186, 192 Consequently, for










transport experiments carried out at low ionic strength in antibody-modified membranes that has

been previously equilibrate with a protein that strongly binds the antibody carrier, both the

modified membrane and this permeating protein bearing similar electric double-layer that creates

double-layer repulsions when they approach each other. Besides, at low ionic strength the

overlapping between the double-layers both of the membrane pore wall, and that one of the

permeating protein affects the mobility across the membrane.193, 194 Therefore, the repulsion

effects together with the double-layer overlapping are the responsible factors of the enhanced

selectivity at low ionic strength observed in Figure 2-21B. Then, the high selectivity obtained

under these conditions might be attributed to the electrostatic exclusion of the charged anti-M-

IgG molecules from the membrane pores that were equilibrated with the same protein (anti-M-

IgG) having identical isoelectric point.162, 193


A B..
A B *



I I rI
T) ( n"
0 0
1.5. i of o *te .
2 1.0 -

0 0 0.5 -
o 0
E E
0 0.5 -

A AA.AAA"AA.A
A A 0.0 A A A A a
0.0 ..0.0
0 200 400 600 800 1,000 1,200 0 200 400 600 800 1,000 1,200
Time (min) Time (min)
Figure 2-21.Flux plots for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG -modified
membranes, (A) without equilibration, and (B) after equilibration. The feed solution
concentration was 2.65 nM in both of the permeating proteins.

These data suggest that this antibody-modified nanotube-membranes at low ionic strength

can be used for situations were the transport of an unwanted protein needs to be shut down, in

order to improve selectivity. One potential application that this approach might be useful is in the

sample complexity reduction in proteomic analysis. Since the analysis and characterization of









complex mixtures of proteins are the central aims of proteomics.195 Thus, it would make possible

to deplete the flux of high abundance proteins across the antibody-modified membranes while at

the same time it would allow improving the separation and detection of low abundance proteins

by subsequent analytical techniques.

Conclusions

A detailed experimental study on antibody-functionalized nanotube membranes has been

carried out, including the effect of surface modification, pore size, feed concentration,

temperature, and ionic strength, both on the transport as on the functional behavior of the

membranes. It was demonstrated that the type of immobilized antibody changes the functional

behavior of the antibody-modified membranes and the transport properties of the permeating

proteins across them. These membranes showed a Langmuirian shape plot characteristic of a

typical behavior of facilitated transport membranes, when the effect of feed solution

concentration on the fluxes of the permeating proteins was investigated. Therefore, the antibody-

functionalized membranes showed here exhibit a type of behavior for facilitated-transport

membranes during protein separations. Important parameters such as the binding capacities of

the antibody-modified membranes and the binding constants for the permeating proteins to the

antibody carriers confirmed that the permeating protein anti-M-IgG binds strongly to the carrier

M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier, likewise, the

permeating protein anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-IgG binds

weakly to this carrier. It was also demonstrated that by controlling the orientation of the

immobilized antibody using protein G the potency affinity of the antibodies is effectively

preserved. Through these studies was also demonstrated that by selecting a template membrane

with an appropriated pore size diameters these antibody-modified nanotube membranes have the

ability to be used with a combination of both size and affinity selectivity for protein separation.









Additional studies on the effect of temperature on the fluxes and the binding constant

demonstrated that the functional behavior of these antibody-modified membranes is governed for

the changes in free energy before and after interaction of the permeating proteins with the carrier

antibody, which was attributed to the changes in enthalpy and entropy of the system. Other

important finding during this investigation was that by controlling the ionic strength of the

transport solution, these antibody-modified nanotube membranes can be used for situations were

the transport of an unwanted protein needs to be shut down, in order to improve selectivity. This

latter approach might be useful in the sample complexity reduction in proteomic analysis. Since

the analysis and characterization of complex mixtures of proteins are the central aims of

proteomics. It is also noteworthy to mention the low pM detection limits achieved during these

studies. In general, the ability to alter the surface biochemistries and thus tailor the membrane

for specific analytes, promises the use of these membranes for the analysis of a wide range of

proteins.









CHAPTER 3
MODELING AND CHARACTERIZATION OF THE TRANSPORT PROPERTIES OF
PROTEINS ACROSS ANTIBODY-FUNCTIONALIZED NANOTUBE MEMBRANES BY
USING THE DUAL-MODE MODEL

Introduction

In the design and development of more effective membranes for membrane-based

separations, the value of the transport mechanism must be considered in order to understand the

different interactions that govern transport through the pores of the membranes. In addition, the

transport mechanism, supported by mathematical models and experimental evidences provides a

grater understanding of the underlying transport phenomena.

The separation properties of molecules through membranes are based on the difference of

permeation. Therefore, it is important to understand permeation behavior, not only to understand

the permeation mechanism but also to elucidate the separation properties.

In order to elucidate the transport properties across selective membranes a wide variety of

models have been developed.196- 202 However, a model that explains the sorption and permeation

mechanism is of fundamental importance to fully understand the transport properties in

membrane separation.

The well-known dual-mode model allows evaluation of both sorption and permeation

mechanism, for those cases where the model applies. The dual-mode model was initially

developed by Vieth and Sladek203 and later modified by Paul and Koros204 to describe the

sorption and permeation of gas molecules in glassy polymeric films. In literature there are

various modified dual-mode models that have also been explored.182, 205 208

The model assumes that the membrane contains two distinct regions in which sorption can

occur.204 One region exhibits solubility of the molecule based on Henry's law and the second










region has a Langmuir sorption isotherm. For permeation, it is assumed that each region is

equilibrated and permeation is governed by a solution-diffusion mechanism.

feed na.otue membran permta
colnpament D conpanent
exa Phasial
Pernmeation
SPathway

DD+ DH

Dual-Mode
Model
Permeation
Patlhway


Figure 3-1.Dual-mode transport model.

The dual-mode model has also been useful to explain membranes with higher transport

selectivity obtained via a methodology called facilitated transport.114 115, 209 These models

typically draw an analogy between the facilitated transport process and the dual-mode model for

gas transport in glassy polymers. During facilitated transport, the enhanced transport of a specific

permeate is mediated by a "carrier" immobilized within the membrane, plus the normal physical

permeation pathway, since this carrier in the membrane reversibly binds the specific permeate

molecule. In Figure 3-1 is shown a schematic illustration of the dual-mode model presented here.

This chapter presents the results on the characterization of the transport properties of anti-

M-IgG and anti-H-IgG across M-IgG modified membranes by using the dual-mode model. We

have studied the effect of systematic chemical modification on the transport properties of silica

nanotube membranes by functionalization with covalently attached-carrier-antibodies as a means

of selectively improving transport properties of proteins. It is discussed how the presence of this

carrier in the nanotube-functionalized membranes enhance the permeability of the transported

anti-IgG. This enhancement in permeability at low concentration is explained with the dual-

mode model and it is caused by a contribution of the diffusion due to the Langmuir mode









species. Based on these results, the concept of modulated facilitated transport using antibody-

modified nanotube membranes is proposed. Thus, depending on the affinity parameters, the

transport in the carrier situation can be modulated in comparison to the situation without carrier

modification where the rate of transport is only controlled by physical diffusion. The present

study test the dual-mode model under circumstance where is possible to have modulated

facilitated transport, and simulations and experimental results are discuses to validate this

concept.

Definition of Dual-Mode Model

The dual-mode model is widely used to describe the sorption of gas molecules on glassy

polymers210,211. Depending on their internal structure, the polymers are characterized as either

glassy or rubbery. Sorption of gases to the rubbery state occurs by dissolution, while sorption to

glassy state occurs by concurrent dissolution and hole-filling mechanisms.212' 213 The sorption

mechanism of gas molecules in glassy polymers has been understood in term of the dual-mode

model.203 208 The model is based on two types of sorption sites. One obeys the Henry's law

dissolution and the second a Langmuir-type sorption. According to the dual-mode model, total

sorption to surface membrane is the sum of sorption in the dissolution domain and sorption in the

Langmuir-type sorption domain (Equation 3-1): 214


C = C +H = kDp + Cstes (3-1)
S+bp
By analogy but making a couple of changes, we have used the dual-mode model used for

gas-transport to interpret transport of proteins in aqueous solution through antibody-modified

nanotube membranes. Instead of reporting the concentration of penetrant in the membrane for a

given gas pressure, the concentration is reported as a function of the feed protein concentration.

The Henry's law coefficient, kD, was also adjusted and it was given in terms of partition









coefficient kp. Since, the sorption coefficient given in the Equation 3-1 is a gas-phase coefficient;

in solution therefore the sorption coefficient should be considered as a liquid-phase coefficient.

The adapted dual-mode model is shown in Equation 3-2.

C bC
C = CD C = kpCfeed + tes feed (3-2)
l+bC feed
Where C is the total protein concentration in the membrane, CD is the protein concentration

based on Henry's law sorption, CH is the protein concentration based on Langmuir sorption.

Whereas, kp is defined as the partition coefficient, b and Cstes are the Langmuir affinity constant

and maximum capacity parameters, respectively.

The Langmuir affinity constant, b, characterizes the sorption affinity for a particular

protein-antibody membrane system, and Cstes is used to measure the amount of antibody

immobilized in the membrane.

The three dual-mode sorption parameters, kp, Cstes, and b, can be determined through the

experimental evaluation of protein-adsorption isotherms. The estimation for Cstes, and b,

sorption parameters can be accomplished by analyzing the data according to the monolayer

Langmuir model (Equation 3-3):167

CBd -C Stes Free (3-3)
1 +bCFree
Where CBound is the specific concentration of protein bound to the carriers in the

membrane, which is related to each specific concentration of free protein (Cfree) in the

equilibration solution.

The other of the three sorption parameters contained in the dual-mode model is the

partition coefficient. Through the analysis of the boundary conditions of the Langmuir-mode

model, an optimum parameter value for the partition coefficient could be estimated.









Thus, at the high concentration range, the Langmuir model approaches to a constant value

that is equal to Cst,,e, so no more sorption will take place when all the sites are filled. On the

other hand, at very low concentrations, the Langmuir model obeys the correct thermodynamic

boundary condition of Henry's law over infinitely dilute concentration range.215 216

In the limit of zero concentration the Langmuir model leads to the following equation

which satisfies the condition of a linear isotherm (Equation 3-4):

lim- C = CsutebC,ee kDC Fed (3-4)
C-O
Hence, a good estimation of Henry's law coefficient, kD, may be written as the product of

CsItes and b. Consequently, this kD parameter can be used for the calculation of the partition

coefficient (kp), which is defined as the product of kD times the ratio of volumes between volume

inside the membrane (VMembrane) and volume in the free equilibration solution (VFree Solution)

(Equation 3-5).


k = kD VMembrane (3-5)
SFree-Soluhon )
It should be pointed out that for linear isotherms, the Henry's law coefficient, kD, can be

determined from the slope, while for dual-mode adsorption isotherms, it can be determined from

the initial slope at infinite dilute concentration.

The partition coefficients can also be calculated using Equation 3-6, which expresses the

ratio of the intramembrane concentration to external concentration at equilibrium.


k V MembraneCMembrane (3-6)
VFree-Soluhon Free-Soluhon )
All this dual-mode sorption model theory explained until here can be validated using

linearization methods for the Equation 3-2. Several linear plots of the dual-mode sorption

model have been studied,217 and for instance, the equation can be rearranged and transformed

into the Equation 3-7.









CF 1 1
Feed Feed (3-7)
C kp CFeed CSitesb C Sites
A plot of CFeed / (C kp CFeed) against CFeed is expected to be linear, where the Langmuir

sorption parameters, Cstes and b, can be evaluated from the slope and intercept of the straight

line, respectively.217

Now, if the solute-adsorption isotherms are represented by the dual-mode model, as

shown above, the permeation mechanism can also be described in terms of the dual-mode model.

The permeation mechanism using the dual-mode model assumes partial immobilization of

the Langmuir-mode species.204 Therefore, independent dual diffusion coefficients are given for

each sorption mode. One diffusion coefficient (DD) is used for the fraction of the solute dissolved

in the membrane according to Henry's law expression and a second (DH) for the fraction of the

solute contained in the Langmuir-type sorption mode. The expression for the permeation flux

through the membrane is derived by Fick's law, and it can be written in the form given in

Equation 3-8.

dC dC FK C
J = -DD D -DH H = -D D (3-8)
D9x x D (I +aCD )2 CD x
F corresponds to the mobile fraction of the Langmuir species (Equation 3-9).


F DH (3-9)
DD
Where a and K are constants that contain the three sorption parameters (Equation 3-10 and

Equation 3-11).

C~b b
K Cste b (3-10); a (3-11)
k k
P P
The mathematical solution that Koros and Paul found for the Equation 3-8 is known as the

partial immobilization model.204 The model assumes that there is always local equilibrium

between the two sorption modes, thus, CH can be written in terms of CD. In addition, the









diffusion coefficient is assumed constant so it does not depend on concentration. Besides, the

model splits the total concentration into two parts, one a mobile part with a diffusion coefficient

DD and concentration Cm, and the second remainder part (C Cm), corresponds to the part that is

totally immobilized. In other words, the concentration associated with CD as well as a fraction F

of that associated with CH has finite mobility while the remaining fraction, 1-F, of CH has zero

mobility. Therefore, the concentration for the mobile part can be written as (Equation 3-12):

Cm = CD +FC (3-12)
The flux for this mobile part may be written as is shown in Equation 3-13.

QC
J = -DD (3-13)
8x
Whereas the equation of continuity (Fick's second law), represented by the dual-mode

model for a solute transported through the membrane is given by the Equation 3-14.

OC 8(C + CH) aJ
(3-14)
at at ax
The equation of continuity is nothing more that the mass balance for the system, which

relates the accumulation of solute in the porous membrane to the changes in flux at any time and

any direction of the flux. In other words it is the rate of solute balance. Thus, the difference

between the rate of solute that comes in to the membrane, and the rate of solute out, is equal to

the rate of solute accumulation. The insertion of the Equation 3-13 into the Equation 3-14, results

in Equation 3-15.

+(CD H ) 8(-DD C )(3-15)
at x
By assuming that the diffusion coefficient does not depend on concentration, and inserting

the Equation 3-12 into the mass balance equation given by the Equation 3-15, the equation

becomes (Equation 3-16)

a(C, + C) D 2 (CD +FCH)
= D (3-16)
Qt D2









Finally, by setting all the assumptions and writing the concentration of CH in terms of CD,

(KCD / l+aCD), a numerical solution of the partial immobilization model yields to an expression

for the permeability coefficient given by Equation 3-17.204


P =kDD 1+ lFK (3-17)
1 1 + -bCFeed
The use of the model requires the determination of both DD and F. This can be

accomplished through the experimental evaluation of the permeability coefficients; so, optimum

parameters for the diffusion coefficients can be determined.

Throughout every experimental permeation, measurements of the flux per unit area, J,

(moles cm-2 sec-1), is obtained. Therefore, if a number of permeation experiments are carried out

at different feed concentration using membranes with homogeneous thickness, L, the steady-state

permeability coefficient for a specific feed concentration can be calculated using the following

relationship showed in Equation 3-18.218 219

JL
P =- (3-18)
CFeed
This procedure would also allow comparing the experimental permeability plots with those

calculated obtained ones.

In order to explain permeation, the dual-mode model assumes that the permeation is

governed via a solution-diffusion mechanism. If the permeation of the protein in the membrane

is controlled by a solution-diffusion mechanism, the permeability can be expressed as the

product of diffusivity, D, and solubility, S as is shown in Equation 3-19.

P = DS (3-19)
Although the solubility of a permeating protein in the membrane can be also expressed in

term of the dual-mode sorption parameters by (Equation 3-20):210, 211

C Cs b
S = SD + SH = kp + (3-20)
CFeed + bCFeed









Where S is the solubility of permeating protein, and the terms SD and SH corresponds to the

solubility values based on Henry's law and Langmuir-type sorption, respectively.

If the diffusion coefficient does not depend on concentration, a plot of permeability against

solubility is expected to be linear, with a slope equals to the diffusion coefficient (Equation 3-

19). Through this correlation between P and S experimentally the diffusion coefficients can be

estimated.

With the aim of determining the F parameter, the experimental permeability data can be

plotted against (1 + bCFeed)-1. A value for F can be estimated from the slope of this line.211

Additionally, the diffusion coefficient DD might be evaluated from the intercept as well, but only

if the plot is straight line with a very high linear correlation coefficient (Equation 3-21).

Cstes bFDD
P = kDD StesFDD (3-21)
1 + bCFeed
In fact, the dual-mode model makes feasible the evaluation of the sorption and permeation

mechanism. This is of fundamental importance to fully understand the transport properties in

membrane separation. Hence, using the dual-mode model, a comprehensible explanation of the

transport mechanism might be accomplished through a detailed elucidation of the transport

properties such as sorption, diffusion and permeability.

Results and Discussion

Simulations of Permeability vs. Feed Concentration Based on the Dual-Mode Model

The permeability is the product between diffusivity and solubility as is described in

Equation 3-19. The permeability dependence in terms of the dual-mode model is not only related

with the concentration of solute molecules in the feed solution, but also with the different

properties and interactions between solute-membrane. Thus, the permeation of solute molecules










through carrier-modified membranes, which is governed by both the Henry mode and the

Langmuir mode, can be modulated depending on the affinity parameters between solute-carrier.

In order to understand how the dual-mode model works, simulations of the model were

initially carried out (Figure 3-2). The simulation analysis was based on the assumption of

competitive permeation studies for two different solute species transported across carrier-

modified nanoporous membranes. With the aim of simplifying the analysis an idealized case

was compared. Thus, one of the species "the spectator" or nonbinding molecule (NBM) is

assumed to have no interaction with the carriers in membrane and therefore permeates the

membrane by only physical diffusion. On the other hand, the second specie, the binding

molecule (BM), permeates the membrane through a dual-mode model mechanism. The different

dual-mode parameters used for the simulations are listed in Table 3-1.

1.2x106
a
1.0x10 -

0 8.0x10 -
E
6.0x107

4.0x10 -
E
2.0X107

o.o F b

0.0 1.0x10"0 2.0x1010 3.0x1010 4.0x1010 5.0x1010
Feed Concentration (moles/cm3)
Figure 3-2.Permeability plot simulations for the transport of (a) BM and (b) NBM across carrier-
modified membranes. The parameters are summarized in Table 3-1.

By simple visual analysis of the simulations showed in Figure 3-2, we can notice that for a

permeation mechanism that follows the dual-mode model, the permeability is highly dependent

on the concentration, while for a mechanism based on physical diffusion only, the permeability is

totally independent of the concentration. The permeability in terms of the dual-mode model










decreases with increasing feed concentration, which is concave to the concentration axis as can

be seen in the plot (a) in Figure 3-2.

Table 3-1.Dual-mode model parameters used for the permeability plot simulations

D Csites b
Permeant specie kp D 3Cie 3b
Permeant specie (cm2/sec) (moles/cm3) (cm3/moles)

Nonbinding Molecule 613 910
(NBM)
Binding Molecule 3 10
Ml 6 x 10-3 9 x 10- 1.5 x 10 1 x 011
(BM)


Figure 3-3 shows simulations for the permeability versus feed concentration, illustrating

the effect of the F parameter in dual-mode model mechanism. Figure 3-3 compares three

simulation plots for the case of the BM situation, which correspond to three different values ofF.

The term F is associated with the mobile fraction of the Langmuir species as was stated before,

and its magnitude is equal to unity or zero, when the model assumes no immobilization or total

immobilization of the Langmuir species, respectively.

1.2x10-
a
1.0x10 -
I




6.0x10'
b"

4.0x107
1*
E
S2.0x107 -
0. \

0.0 C
I I I I I
0.0 1.0x1010 2.0x1010 3.0x1010 4.0x1010 5.0x10"1

Feed Concentration (moles/cm3)
Figure 3-3.Permeability plot simulations for the transport of BM across carrier-modified
membranes illustrating the effect of the term F on the dual-mode model. (a) F= 1, (b)
F = 0.5, and (c) F = 0. The other parameters are summarized in Table 3-1.









Hence, the permeability for a solute molecule that permeates a membrane through a dual-

mode mechanism depends also on the magnitude ofF. For the BM situation, the permeability is

highly dependent on the feed concentration as F gets closer to unity (plot a, in Figure 3-3).

Consequently, as the value ofF decreases or the degree of immobilization increases, the

permeability becomes less dependent on the feed concentration (plot b, in Figure 3-3). When

there is total immobilization, when F is equal to zero, the permeability is independent of the

concentration, and so the BM dissolves and diffuses through the membrane via Henry's mode

dissolution only (plot c, in Figure 3-3). It should be pointed out that regardless of the value ofF,

the permeability decreases gradually with the increase of feed concentration until a steady level

that reach the value of kpDD.

The above simulations in Figures 3-2 and 3-3 suggested that when no immobilization of

the Langmuir species is achieved, the permeability across carrier-modified membranes is always

enhanced compared with the physical diffusion case. Therefore, the transport through carrier-

modified membranes that fulfill the dual-mode model mechanism is always facilitated when the

Langmuir species no experience immobilization.

The concentration dependence of the permeability for the dual-mode model can also be

analyzed through the evaluation of the limit boundaries conditions in concentration for the

equation of permeability (Equation 3-17). At the high limit concentration range the equation for

permeability becomes equal to the product between kp times D, this is due to the behavior of the

factor [1 + FK / (1 + b CFeed)], which becomes closer to unity as the concentration increases. At

the low limit concentration, the highest permeability value is reached when the denominator

term, (1 + b CFeed), becomes unity. As the latter gets closer to one, there is a range in

concentrations where the variations in permeability are less noticeable, and the permeability









becomes less dependent on the concentration. This region resembles to the time lag region for

gases in glassy polymers.181' 182

In general, as stated by Koros and Paul the limits of the dual-mode model are confined to

two situations.204 One is the limit where total immobilization occurs, and only the Henry's mode

species contributes to the permeability. The other limit is a situation with a nonlinear isotherm

with no distinctions for sorption mechanism or transport behavior between any of the adsorbed

solutes. However, the concentration-dependence of the permeability is a consequence of the

shape of the adsorption isotherm, and therefore a result of the binding affinity strength between

Langmuir permeating species and Langmuir binding sites.

In order to demonstrate the concentration dependence on the permeability as consequence

of the shape of the adsorption isotherm, permeability plot simulations were carried out

comparing two different sorption mechanisms via two different binding affinities (Figure 3-4).

The plot (a) in Figure 3-4 clearly indicates that a carrier-modified membrane with stronger

affinity for the solute molecules displays higher concentration-dependence as a result of the

sorption isotherm. So, the concentration dependence of permeability is determined by the

strength of the binding affinity for the solute-carrier complex in the membrane. However, the

total permeability is highly dependent on the diffusion coefficient.

It is worth mentioning that by using simulations was verified that the order of influence of

the dual-mode model parameters in the permeability was D > b > C,t,, >>> kp, respectively.

According to these results, the dual-mode model permeation analysis is a useful model to

explain transport across carrier-modified membranes. The analysis showed until here

demonstrated how the fluctuations in permeability are a result of the strength of the sorption

mechanism. Therefore, the differences in permeability and the differences in permeability










concentration-dependence can be used to elucidate the separation properties for the solute

molecules that permeate a carrier-modified membrane. For the reason that differences in

permeability are a consequence of the sorption mechanism, the separation properties of one

solute might differ from that of another, depending on the affinity solute-carrier interaction.

a




6.0xtO -
1 4.0x10' b
2.OxlOX


0.0
0.0 1.0X10'" 2.0X10 10X11" 4.0X0lIO' 5.OX0
Feed Concentration (moles/cm')
Figure 3-4.Permeability plot simulations for the transport of BM across carrier-modified
membranes showing the effect of the binding affinity on the permeability. (a) b = 1 x
1011 cm3/moles, and (b) b = 3 x 1010 cm3/moles. The other parameters are summarized
in Table 3-1.

Experimental Description of the Dual-Mode Model

Effect of feed concentration on the flux

A number of competitive permeation experiments were carried out at different feed

concentration, ranging from 2.65 x 10-12 to 5.0 x 10-10 moles cm-3 using M-IgG modified

membranes (Figure 3-5). Throughout every experimental permeation measurements, the flux per

-22
unit area (moles cm-2 sec-1) was obtained, using 0.31 cm2 as the area of membrane exposed to the

contacting solution phases.

Figure 3-5 shows the studies of the effect of feed concentration on the flux during

competitive permeation experiments of anti-H-IgG and anti-M-IgG across M-IgG-modified

membranes. The flux for anti-H-IgG was at all times higher than the flux for anti-M-IgG. The

linear dynamic range for anti-M-IgG was 3 times lower than for anti-H-IgG, as a consequence of










a stronger binding affinity with the M-IgG carrier attached in the membrane. These plots show a

characteristic "Langmuirian" shapes, demonstrating that the immobilized M-IgG indeed interacts

with the permeating proteins, acting as a carrier during the permeation process.9' 10, 109


1.0x10'"

S 8.0x10'6

D 6.0X1016

o 4.0x10- '
E

2.0x10
LL
0.0
0.0 1.0x1010 2.0x1O0 3.0x101 4.0x10" 5.0x1010
Feed Concentration (moles / cm3)
Figure 3-5.Flux versus feed concentration for anti-M-IgG (A) and anti-H-IgG (m) across M-IgG
modified membrane. The feed solution concentration was equimolar in both of the
permeating proteins.

Experimental permeabilities as function of the feed concentration

In order to conduct fundamental studies of any membrane process, a well defined

membrane system is required. One of the worthy characteristics of the alumina template

membranes used in these studies is their well defined and characterized structure. Thus, it is

possible compare theoretical models with the experimental results as is demonstrated in this

dissertation. Because the thickness of the membranes and feed concentration for each permeation

experiment were known, the Equation 3-18 for permeability was used to calculate the

experimental permeability coefficient of anti-M-IgG and anti-H-IgG across M-IgG modified

membranes (Figure 3-6).

Figure 3-6 shows the plots for the experimental permeability versus feed concentration. It was observed

that the permeability decreases as the concentration increases both for anti-H-IgG and for anti-M-IgG

across M-IgG modified membranes. This diminution in permeability may be understood in terms of the










dual-mode model,204 206, 211, 220 since the permeability in terms of the dual-mode model decreases with

increasing feed concentration.221


1.2x10' -

1.0x10'

7 8.0x108

6.0x10-



2.0x108

0.0 ,i- i I t -
0.0 1.0x10" 2.0x10" 3.0x1O0 4.0x10" 5.0x10"
Feed Concentration (moles / cm')
Figure 3-6. Experimental obtained permeability versus feed concentration for anti-M-IgG (A)
and anti-H-IgG (m) across M-IgG modified membrane.

At first sight from the data of concentration dependence of the experimental permeability

in Figure 3-6, we might anticipate that the transport mechanism for these two proteins across M-

IgG modified membranes approaches to the dual mode model mechanism. In order to prove this

transport mechanism for these two proteins across the M-IgG modified membranes, an

evaluation of the sorption and permeation parameters were carried out, and they are discussed

next.

Experimental evaluation of the dual-mode model parameters

The mathematical solution that Koros and Paul found for the permeability coefficients

based on the dual mode model is known as the partial immobilization model (Equation 3-17).204

The use of the model requires the determination of both DD and F, which were evaluated

experimentally by using the permeability coefficients in Figure 3-6. Additionally, to use the

model is a prerequisite to have the information of the three dual-mode sorption parameters, kp,

Csites, and b, which were also determined experimentally via protein-adsorption isotherms.










Experimental evaluation of Csites and b

These Langmuir sorption parameters were determined experimentally using adsorption

isotherms. The relationship that exists between concentrations of protein in the porous membrane

isothermally in equilibrium with those in the feed solution is called adsorption isotherm. In order

to assess the equilibration time for the sorption experiments, a sorption dynamic study was

initially carried out (Figure 3-7). For this, a solution that was 21.2 nM of anti-M-IgG labeled

with Alexa 488 was presented to M-IgG modified membranes. The amount of anti-M-IgG

adsorbed was investigated observing the changes in fluorescence intensity as a function of time.


600

500

C 400

C 300
0-
200-
0
S100



0 2 4 6 8
Time (Hours)
Figure 3-7.Curve of adsorption dynamics for anti-M-IgG (A) in a M-IgG modified membrane.
Inset shows the changes of fluorescence signal with time.

As can be seen from Figure 3-7, the amount of anti-M-IgG adsorbed in the M-IgG

modified membrane increases rapidly during the first 3 hours. After that, the rate of adsorption

decreases and the system reaches a plateau of apparently equilibrium.

All the adsorption isotherms were carried out with M-IgG modified membranes at 23 C in

stirred U-tube permeation cells using the same buffer conditions used for the permeation studies,

via solutions containing single concentrations of anti-M-IgG or anti-H-IgG fluorescent labeled










with Alexa 488. Data were analyzed according to the monolayer Langmuir model (Equation 3-

3).167

By using the experimental values of Cfee and Cbound in moles per cm3 in Equation 3-3 the

parameter Csites was calculated in moles per cm3 while b in cm3 per moles. Figure 3-8 shows and

compares adsorption isotherms for anti-M-IgG and anti-H-IgG carried out with M-IgG modified

membranes; also the adsorption isotherm for anti-M-IgG using unmodified nanotube membranes

is showed. Table 3-2 summarizes the corresponding binding constant and protein binding

capacities, as calculated by fitting the data using nonlinear regression methods.


1.2x109 Control Membrane
12x -- M-IgG Modified Membrane
o -A- M-IgG Modified Membrane
1.0x109 [--

E 8.0x10 -

Santi-M-IgG
6.0x10 10
0
S4.0x10 1 anti-H-IgG
A
2 2.0x10" anti-M-IgG
S A / /AAA
0.0 -*-*.-0 ,---- =

0.0 5.0x1012 1.0x11 1.5x101 2.0x1011 2.5x101

Free Concentration (moles I cm3)
Figure 3-8. Adsorption isotherms for anti-M-IgG (A), and anti-H-IgG (m) binding on M-IgG
modified membranes. Also the adsorption isotherm for anti-M-IgG (*) using a
control membrane is shown.

Figure 3-8 shows the adsorption isotherms obtained from the fluorescence intensity of the

wet membranes exposed to different protein concentrations. For the unmodified silica nanotube

membranes the adsorption of anti-M-IgG labeled with Alexa 488 was negligible in comparison

with the antibody-modified membranes. It was also observed, that the adsorption for the

unmodified membranes increases proportionally with the concentration of fluorescent-labeled









antibody, in the entire range of concentration used for the saturation binding experiments (from

1.0 x 10-12 to 2.5 x 10-11 moles cm-3), obeying clearly a Henry's linear isotherm. The linear shape

of the unmodified isotherm also indicates a very low affinity for the anti-M-IgG. Thus, these

results also suggested that the silica nanotube membranes exhibit low protein binding (non-

specific antibody adsorption) to the surface membrane, and minimal autofluorescence.

On the other hand, the M-IgG modified membranes exhibited clearly nonlinear isotherms

and they bound more protein than the unmodified ones. When anti-H-IgG labeled with Alexa

488 was presented to M-IgG modified membranes the adsorption isotherm was nonlinear. This

effect is due mainly to the existence of cross-reactivity between the permeating anti-H-IgG and

the attached M-IgG. That kind of cross-reactivity of IgG has been also seen in the literature.169-

172

By fitting the experimental adsorption isotherm for anti-H-IgG, the existence of one

binding site for the antibody was demonstrated. A binding affinity constant equal to (2.90 + 1.35)

x 1010 cm3 moles-1 was obtained. The amount of protein bound to the membranes exposed to

anti-H-IgG was almost twice compared to the unmodified silica nanotubes ones (Table 3-2).

In contrast to the anti-H-IgG sorption isotherm case, when anti-M-IgG labeled with Alexa

488 was presented to the M-IgG modified membranes the protein bound to the membrane was

approximately 20 times higher than the anti-H-IgG situation. These data confirmed the

successful modification of the membranes, and the remarkable selective attribute of the

antibody-modified membranes.

The fitting process for the M-IgG modified sorption isotherms exposed to anti-M-IgG

indicated the existence of two binding sites in the membrane for the permeating protein (Figure

3-8). This is probably due to the spatial orientation of the M-IgG molecules immobilized on the










pore surface of the membrane. Therefore, the binding site with the stronger binding constant,

(4.46 0.91) x 1011 cm3 moles-1 was attributed to the concentration of well-oriented immobilized

IgG molecules. The other site with an affinity constant of (5.27 1.58) x 1010 cm3 moles-1 was

ascribed to the random-oriented concentration. Thus, anti-M-IgG first binds to the site of well-

oriented immobilized M-IgG. After a fixed amount of anti-M-IgG is adsorbed the random

immobilized M-IgG become available as cooperative binding sites.


3.0x10' 0

S2.5x10'Io

5 2.0x10 10

"g 1.5x10 -
1.0x10'0


CL 5.0x10 -

0.0
0.0 5.0x10 1.0x1011 1.5x10-" 2.0x10- 2.5x10."
Free Concentration (moles / cms)
Figure 3-9.Adsorption isotherm for anti-M-IgG (A) binding on M-IgG-PG modified
membranes.

Additional confirmation for this conclusion about two binding sites with different spatial

orientation was obtained through adsorption isotherms of anti-M-IgG using M-IgG well-oriented

antibody-modified membranes via protein-G (PG) immobilization (Figure 3-9). The well-

oriented immobilization of M-IgG was achieved through silica nanotube membranes already

modified with PG. This PG was covalently attached via the Schiff's base reaction to the walls of

aldehyde modified nanotube membranes. The adsorption isotherm with M-IgG-PG modified

membranes revealed the existence of only one binding site by using this well-oriented antibody-

immobilization (Figure 3-9). For this specific case, the value of the binding affinity constant was

(3.77 0.53) x 1011 cm3 moles-1 (Table 3-2).









It should be pointed out that the antibodies used in these studies were all polyclonal

antibodies. Therefore, the found binding affinity constants are indeed an average of the pool of

the different clonally related subpopulations of antibodies existing in the sample solution.222'223


Table 3-2.Thermodynamic parameters from adsorption isotherms of anti-M-IgG and anti-H-IgG
on M-IgG modified membranes, according to the Langmuir monolayer model
Adsorption Csltesl b Csltes2 b2 R
Isotherms (moles/cm3) (cm3/moles) (moles/cm3) (cm3/moles)

anti-H-IgG (7.24 + 1.33) x 10-11 (2.90 + 1.35) x 1010 0.983
/ M-IgG

anti-M-IgG (1.16 + 0.19) x 10-10 (4.46 + 0.91) x 101 (1.30 + 0.39) x 10-9 (5.27 1.58) x 1010 0.928
/ M-IgG

anti-M-IgG (3.49 + 0.14) x 10-1 (3.77 0.53) x 1011 0.992
/M-IgG-PG


Experimental evaluation of the partition coefficients kp

For definition, partition coefficients provide a thermodynamic measurement of the

solubility of the permeating proteins in the membrane. Therefore, considering this

thermodynamic fact, proteins with higher diffusion coefficient should have lower partition

coefficients. Contrarily, an increase of the partition coefficient increases the resident time of the

permeating proteins in the membrane.

The partition coefficient is one the other three sorption parameters contained in the dual-

mode model. The experimental evaluation of the partition coefficients across M-IgG modified

membranes for the proteins studied here were investigated using three different approaches.

Table 3-3 compiles the results for these studies.

The first approach for the analysis of kp was based on the examination of the boundary

conditions of the Langmuir-mode in the limit of zero concentration, that determine that the

Henry's law coefficient (kD), may be written as the product of Csite times b (Equation 3-4).215, 216










Since the values of Csites and b were already estimated for both of the permeating proteins

studied here as was shown above. The partition coefficients calculated using the Equation 3-5

were 7.24 x 10-3 and 2.94 x 10-4 for anti-M-IgG and for anti-H-IgG, respectively.

Table 3-3.The kp determined by various approaches

Method anti-M-IgG Anti-H-IgG

Equations 3-4 and 3-5 7.24 x 10-3 2.94 x 10-4

Slopes Figure 3-10 6.14 x 10-3 2.77 x 10-4

Equation 3-6 (5.84 1.06) x 10-3 (2.79 0.69) x 10-4



It is important to mention that for linear isotherms, the Henry's law coefficient, kD, can be

determined from the slope of the isotherm plot, while for dual-mode sorption isotherms, it can be

determined from the initial slope at infinite dilute concentration (Figure 3-10). To prove this

assertion the fitting plots of the adsorption isotherms for anti-M-IgG and anti-H-IgG were

extrapolated to a very dilute concentration range (from 0 to 6 x 10-13 moles cm-3), and the

respective slopes from the linear plots were estimated, these results are shown in Figure 3-10.


2.5x10" -
E
2.0x10 -
o0
E 1.5x10" -

1.0x10"
0

5.0x10.2
0
0.0

0.0 1.0x1013 2.0x10" 3.0x10 "4.0x10"3 5.0x101 6.0x10"
Free Concentration (moles / cm3)
Figure 3-10. Extrapolations of the adsorption isotherms at infinite dilute concentration for anti-
M-IgG (A), and anti-H-IgG (m) in M-IgG modified membranes.










The kD obtained for anti-M-IgG and anti-H-IgG from the slopes in Figure 3-10 were 43.85

and 1.98, respectively. Likewise, using Equation 3-5 a partition coefficient of 6.14 x 10-3 for

anti-M-IgG and 2.77 x 10-4 for anti-H-IgG were found.

The partition coefficients were also calculated using Equation 3-6, which expresses the

ratio of the intramembrane concentration to external concentration at equilibrium. The partition

coefficients estimated with this equation were comparable to those calculated by using the direct

method via kD. As is shown in Table 3-3, the partition coefficients in this case were (5.84 1.06)

x 10-3 and (2.79 0.69) x 10-4 for anti-M-IgG and anti-H-IgG, respectively.

In order to validate these partition coefficients, dual-mode model permeability plot

simulations were carried out (Figure 3-11). For these simulations, four different values for the

partition coefficients were assigned by keeping the rest of the dual-mode model parameters

constant. The results shown in Figure 3-11 demonstrated that variations in the partition

coefficients have a negligible effect in the permeability coefficients when the binding affinity is

fairly strong as the studies showed here.


1.6x10" -
v k =1
k =0.1

E 1.2x10 -* 1

1.0x104

a 8.0x10 -
E
5 6.0x10G
a.
4.0x105
0.0 5.0x10" 1.0x10"- 1.5x10" 2.0x10" 2.5x10"
Feed Concentration ( moles / cm )
Figure 3-11.Permeability plot simulations illustrating the effect of the partition coefficient on the
dual-mode model. Simulation parameters: D = 5 x 10-8 cm2/sec, b = 9 x 1010
cm3/moles, C,,te, = 3.5 x 108 moles/cm3, and kp were respectively, 1 (V), 0.1 ( ),
0.01 (e), and 0.001(*).









Experimental evaluation of the diffusion coefficients and solubility

The most direct method of experimental determination of S and D requires steady-state

permeation data, complemented with equilibrium sorption data.199 Therefore, effective values of

S, D, and hence P characterized for a membrane-penetrant system that has been restricted to

equilibrium sorption and steady-state permeation conditions represent appropriate averages of

the transport parameter of the system.224 To use the dual-mode model requires knowing D and S.

This was accomplished experimentally through an evaluation of the permeability and solubility

in terms of the dual-mode model as will be shown next.

As was state before, the solubility for a penetrant protein in the membrane can be

expressed in term of the dual-mode sorption parameters by the Equation 3-20.210 For the reason

that the three sorption parameters required for this calculation were already established, an

experimental study of the concentration dependence of the solubility was conducted (Figure 3-

12).

Figure 3-12 shows the concentration-dependence of the solubility in term of the dual-mode

model for anti-M-IgG and anti-H-IgG in M-IgG modified membranes. The concentration-

dependence of the solubility in term of the dual model is a consequence of the shape of the

adsorption isotherm. Hence, from Figure 3-12 M-IgG modified membranes possess higher

affinity for anti-M-IgG than for anti-H-IgG, as was demonstrated before (Table 3-2). These data

also indicate where a protein molecule prefers to dissolve, whether in the Langmuir binding sites

or in the membrane matrix based on Henry's law. According to Figure 3-12, anti-M-IgG

molecules would rather dissolve in the Langmuir sorption mode (upper plot), while anti-H-IgG

molecules in the Henry's mode (lower plot). It was also observed that regardless of the protein at

low concentrations the Langmuir mode contributes the most to the total solubility, which is

probably due to the strong affinity exerted for that fraction of well-oriented immobilized carriers










in the membrane. Therefore, at low concentration the solubility of the proteins was highly

increased by the presence of the M-IgG carriers, so those can be used as facilitated partition

domains of the transport. In addition, the data revealed that as the concentration increases, the

solubility decreases and reaches a plateau where the solubility in the membrane is only

controlled by the solubility in the Henry's mode.

100

80 -

60 -

Z 40
.2
20


0II
,. *.----. ------- -

0.0 1.0x101~ 2.0x10" 3.0x10 4.0x10 5.0x1010
Feed Concentration (moles/cm3)
Figure 3-12. Concentration-dependence of the solubility in term of the dual-mode model for anti-
M-IgG (A), and anti-H-IgG (m) across M-IgG modified membranes.

As was described before in Equation 3-19, the permeability is the product between

diffusion and solubility. Since both the concentration dependence of the permeability and

solubility were experimentally determined, they were used to estimate the effective values for the

diffusion coefficients of anti-M-IgG and anti-H-IgG across M-IgG modified membranes. By

using this correlation method a plot of permeability against solubility is expected to be linear,

with a slope equals to the diffusion coefficient (Equation 3-19).

Figure 3-13 shows the correlation plots of permeability and solubility based on the

concentration dependence for the anti-H-IgG (Figure 3-13A) and the anti-M-IgG (Figure 3-13B)

across M-IgG modified membranes. These plots in Figure 3-13 were carried out by using the

experimental concentration-dependence data from Figure 3-6 and Figure 3-12 both for the











permeability and solubility of anti-H-IgG and anti-M-IgG in M-IgG modified membranes. The


plots in Figure 3-13 demonstrated a good correlation between permeability and solubility. The


slopes of these correlation plots will provide the diffusion coefficients of the protein transported


across the M-IgG modified membranes, if the diffusion coefficients are concentration-


independent as the dual-mode model assumes.

Feed Concentration (moles/cm3) Feed Concentration (moles/cm')
5.0x10' 0.0 5.0x10.o 0.0
1.2x10 A B
S6.I0x10 B
i 1.0x10io -
mA
8.0x10 -E A
4.0x108
Z' 6.0x108 A

a 4.x 2.00 A
E O 2.0x10
t A
2.0x10 A
AA
0.0 I I II 0.0
0.0 0.5 1.0 1.5 2.0 0 20 40 60 80 100
Solubility Solubility
Figure 3-13.Correlation plots between permeability and solubility based on the concentration
dependence for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified
membranes.

However, the lines in Figure 3-13 were not linear in the total range of concentration, but


instead they showed two straight-line slope segments. The segment with negative slope


corresponds to the region at very low feed concentration. As mentioned above, in this region


(low feed concentration) the Langmuir sorption sites contributed the most to the solubility of


these proteins as was demonstrated in Figure 3-12.


Therefore, the lower diffusion coefficients at low concentration are attributed to the


presence of slow-moving adsorbed proteins in this Langmuir sorption mode. A more detailed


analysis of this conclusion will be demonstrated below in the discussion of fitting experimental


data with the theoretical dual-mode model. On the other hand, the segment with positive slope


was used to estimate the diffusion coefficients. By fitting of the positive linear segments in









Figures 3-13A and 3-13B, diffusion coefficients of (7.61 0.27) x 10-8 cm2 sec-1 and (1.02 +

0.06) x 10-9 cm2 sec-1 were found for anti-H-IgG and anti-M-IgG, respectively.

These obtained values compare favorably with other IgG diffusion coefficients found in

the literature using different porous materials.225- 229 A lower diffusion coefficient of 4.09 x 10-10

cm2 sec-1 was reported for a bare nanoporous alumina with a pore size of 75 nm and membrane

thickness of 70 nm.228 A diffusion coefficient of 2.71 x 10-9 cm2 sec-1 across nanoporous silicon

with a pore size of 49 nm and membrane thickness of 6 im was also reported.229

It should be mentioned that the most general equation used to obtain a rough estimate of

the diffusion coefficients in liquids is the Stokes-Einstein equation.230 233 This equation is

derived for a sphere with hydrodynamic radius RH, moving in a fluid (Equation 3-22).

kT
D kT (3-22)
6;rqRH
Where, kB is the Boltzmann constant, Tthe temperature, q the solvent viscosity, and RH the

hydrodynamic radius of the molecule, respectively. This Stokes-Einstein relationship described

that the diffusion process is basically controlled by the size of the molecules when they are in

free solution and in absence of any interactions with other molecules. However, it is known that

not only the size but also the intermolecular interaction and the molecular shape affect the

magnitude of Do.234 235 Therefore, we can concluded that the differences in the diffusion

coefficients for the anti-H-IgG and anti-M-IgG across M-IgG modified membranes are mainly

due to the changes in the molecular energy of the proteins as the molecular recognition

interaction takes place. It should be pointed out that the estimated diffusion coefficients were

significantly lower than the reported free diffusion coefficient for IgG in water (3.85 x 10-7 cm2

sec-1).228,229 The reasons for the decreased diffusion coefficients found in these studies compared









with those in free solution are probably due to the intrinsic membrane resistance itself, together

with the changes in molecular energy as was explained above.

The above results show clearly that the correlation analysis between permeability and

solubility is a direct method to estimate diffusion coefficients when equilibrium sorption and

steady-state permeation data are experimentally available. In addition, the two straight-line slope

segments observed through this correlation made also possible to distinguish the heterogeneity in

the spatial orientation of the immobilized carrier in the membrane which demonstrated to have

minor effects in the functional behavior of the membrane (Figure 3-4).

Fitting Experimental Data with the Theoretical Dual-Mode Model

Figure 3-14 compares the experimental and calculated concentration dependence of

permeability for anti-M-IgG and anti-H-IgG across M-IgG modified membranes.


1.5x107 ...........Calculated anti-H-IgG
-. Calculated anti-M-IgG
7 Experimental anti-H-IgG
O 1.2x10
AA Experimental anti-M-IgG

E U.
9.0x10 -


6.0x10' anti-H-IgG

) 3 8 anti-M-IgG
3.Ox10 -
S............... .......... .......
0 0. . . . . . *.. .. .. .. .. .. .. ... .
0.0 -. -.

0.0 1.0x0101 2.0x1010 3.0x1010 4.0x1010 5.0x1010

Feed Concentration (moles/cm3)
Figure 3-14. Experimental and calculated obtained permeability versus feed concentration for
anti-M-IgG (A) and anti-H-IgG (m) across M-IgG modified membrane.









At first glance observing the plots in Figure 3-14, one can conclude that according to the

dual-mode model there is not total immobilization on the Langmuir-mode sites for these two

permeating proteins across M-IgG modified membranes. If total immobilization occurs in the

Langmuir-mode sites (F=0) as the dual-mode model explain, the permeability become

independent of the concentration. These studies demonstrated a high concentration dependence

on permeability both for anti-H-IgG and anti-M-IgG. These also demonstrated that the M-IgG

carriers in the membrane indeed enhance the permeability at low concentration for both of the

permeating proteins as shown in Figure 3-14.

It was found that the permeability enhancement for these two proteins was only up to a

certain threshold of the binding affinity and appeared to depend strongly on the concentration.

This enhancement in permeability at low concentration is caused by a contribution of the

diffusion due to the Langmuir mode species.204' 211

As a consequence of its lower binding affinity, the permeabilities for anti-H-IgG were

higher than those for anti-M-IgG. At the beginning we set a hypothesis that the differences in

permeation for anti-H-IgG and anti-M-IgG observed could be due to a sort of antifacilitated

transport. This was initially considered based on our previous works,9' 10 which showed that

higher fluxes were always observed for the molecule that binds stronger to the carrier in the

membrane. The mechanism for the facile kinetic conquered, in our previous works, was ascribed

by the sequential binding and unbinding events of the transported permeating species with the

carrier in the membrane. This behavior might be mathematically rationalized in a more detailed

manner by the model described here.

As indicated in Figure 3-14, the experimental and calculated data were in good agreement

at concentrations higher than 2.0 x 10-1 moles per cm3. Nevertheless, at very low concentration









that agreement was missing. The explanation for this behavior was that the dual-mode model

assumes a constant diffusion coefficient, which in practice does not happen. Therefore, the

model is not able to discriminate a different regimen for the concentration dependence on the

diffusion coefficient as the case at low concentration observed here. As was said above, at low

concentration the adsorption of these two proteins in M-IgG modified membranes is dominated

basically for the binding sites with stronger binding affinity (Figure 3-12). Then, these

Langmuir-mode sites impart higher sorption rates for the permeating proteins, slowing down the

diffusion coefficients. Consequently this is also reflected in the permeability. Thus, at low

concentration the experimental permeabilities were lower than the calculated permeabilities. This

reduction in permeability is mostly due to a combined effect of both the lower diffusion

coefficients, and concentration-dependence of diffusion coefficients at this lower region in

concentration (Figure 3-14).

Additional evidence for this conclusion was also demonstrated by observing the

concentration dependence on the experimentally obtained diffusion coefficients in Figure 3-15.

As can be seen from Figure 3-15, lower diffusion coefficients were obtained at very low

concentration. Thus, these results demonstrate highly concentration-dependent diffusion

coefficients at low concentration as a result of the two types of binding sites present in the M-

IgG modified membranes (well-oriented and random).

Lower permeabilities were also observed for anti-H-IgG at the low concentration region

(Figure 3-14). This effect was explained taking into consideration the differences in diffusion

coefficients between anti-H-IgG and anti-M-IgG. Due to anti-H-IgG diffuses faster and thereby

reaches the binding sites inside the pores faster; the concentration of anti-H-IgG is initially

greater in the membrane than that of anti-M-IgG. Thus, anti-H-IgG initially at low concentration











binds more sites but those are displaced when the anti-M-IgG molecules catch up to them.

Additional confirmation for this conclusion is obtained by comparing the binding capacities for

these two proteins (Table 3-2).

1.6x109
? 8.0x108 A
0 1.4x10 -
E E
I 1.2x10-9 A
7.0x10
MA
.- 1.0x109
o .o
o 0
S6.0x10 8.0x10" A
o

0 5.Ox1 O A
4.0x10" A
0.0 1.0x101 2.0x10" 3.0x10 4.0x10" 5.0x10' 0.0 1.0x10.' 2.0x10.' 3.0x10.' 4.0x10(O 5.0x10~
Feed Concentration (moles/cm3) Feed Concentration (moles/cm3)
Figure 3-15.Experimental obtained diffusion coefficients versus feed concentration for (A) anti-
H-IgG and (B) anti-M-IgG across M-IgG modified membrane.

At the region of low limit concentration it was also observed that as the concentration

increases the permeability increases as well (Figure 3-14). It was deduced that between the two

set of binding sites (well-oriented and random) present in M-IgG membranes, there is a transition

stage in concentration before the random binding sites with a higher binding capacity become in

a positive cooperative binding sites. This transition involves a mutual cooperation of these two

binding sites in the membrane. This effect was observed at concentrations lower than 2.0 x 10-11

moles per cm3 (Figure 3-14). These results are in favorable agreement with the positive

cooperative contribution observed in the adsorption isotherms in Figure 3-8. Besides as

mentioned previously, the Langmuir binding sites were the dominant factor in determining the

total solubility of anti-H-IgG and ant-M-IgG in the M-IgG modified membranes. And regardless

of the protein these Langmuir sorption sites contributed the most to the solubility at low

concentration, as was demonstrated in Figure 3-12.









The above results demonstrated an enhancement of permeability both for anti-H-IgG and

anti-M-IgG across M-IgG modified membranes. It was also confirmed that the extent of this

enhancement and the linear dynamic range for each system can be controlled by the strength of

the binding affinity between permeant-protein and the antibody-carrier immobilized in the

membrane. Thus, we have demonstrated that antibody-modified nanotube membranes can be

used for the transport of proteins at a rate determined by the affinity between carrier antibody

and transported protein. Hence, it demonstrated that it is possible to develop systems of

antibody-functionalized membranes with different dynamic range and controlled rate of

transport.

The result described herein clearly demonstrated a suitable principle for the concept of

controlled facilitated transport systems for proteins using antibody-modified nanotube

membranes. Of course, each system will require its own development of an appropriate affinity

for the carrier/protein complex, in order to reach the desired rate of diffusion. Since the useful

dynamic range of separation with these antibody-functionalized membranes depends on the

binding affinity, the gamma of affinity constants found with antibodies makes those as proper

elements for this principle.

Besides, with the current scientific development both in genetically engineered and cloned

binding proteins, the system proposed here may also be extended to other system by using

proteins different than antibodies as carriers. Additional research into redesigned proteins may

eventually permit the production of binding proteins against any desired analyte, just as

antibodies can be produced against any antigen.236 The concept of binding proteins genetically

fused as affinity tags to other proteins without the loss of their structural and functional

characteristics has been already demonstrated.237 It has been also demonstrated that a common









way to tune the affinity of a binding protein is to perform rational mutations at the binding site238

240. Thus, depending on the mutations performed in the binding pocket of binding proteins,

mutants with different binding affinities can be obtained;240 this approach would allow the

development of membrane-based separation systems with diverse dynamic ranges.

Moreover, if this tuning affinity can be integrated with a system of membranes where the

number of carriers can be augmented in a way that the permeability of the desired permeating

protein can be selectively facilitated with respect to other present proteins in the separation

sample, a selective and amplified separation system of membranes can be developed. This can be

accomplished by taking advantage of the enhancement in permeability at low concentration

caused by a contribution of the diffusion due to the Langmuir mode species as was explained

above.204,211 Thus, this system would permit to the large number of accessible binding sites to

promote higher concentration gradients inside the membrane that would enhance the flux of the

selected target proteins, which imply a notable key aspect in designing more effective

membranes for membrane-based separation.

Validation of the Dual-Mode Model

The validity of the above results can be assessed by the plot of experimental obtained

permeabilities against (1 + bCfeed)-1. From the slope of the straight line in this plot, the Langmuir-

mode diffusion coefficient (DH) can be evaluated and consequently the mobile fraction of the

Langmuir species F.211 The results for these data are shown and summarized in Figure 3-16 and

Table 3-4, respectively.

Figure 3-16 shows and compares the experimentally and calculated obtained plots of anti-

H-IgG and anti-M-IgG permeabilities vs. (1 + bCfeed)-1. Taking into consideration the cooperation

of the two binding sites at low concentration, the examination of the linear relationship of these

plots demonstrates that the concentration dependence of the permeability for these two proteins










agrees fairly well with the dual-mode model (Equation 3-17). It worth mentioning, that this

validation method requires the use of only one binding affinity. Therefore, in the case of anti-M-

IgG that has two binding sites in the membrane; an average affinity was defined as being the sum

of the free energy changes that occur in the first and second binding site. Thus, in order to keep

the consistency with the conventional units the average binding affinity constant was given by

the square root of the product between the stronger and weaker binding constants (b = 1.53 x

1011 cm3 moles-'). It should be pointed out that comparable results were also obtained for the

experimental concentration dependence on permeability for anti-M-IgG when they were fitted

using this nominal average binding affinity constant.

1.6x10-
1.4 07 1.2x10' B
,3- 1.2x107 1.2x10 -
S1.0x10 x
2 *0 9.0x10 -
8.0x108 -
S6.0x10 6.0x10.8

S42.0x108 A

0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.1 0.2 0.3 OA 0.5 0.6 0.7 0.8
1 / (1 + bCeed) 1 / (1 + bCeed)
Figure 3-16. Plots for experimental and calculated obtained permeability against (1 + bCfeed)-1 for
(A) anti-H-IgG, and (B) anti-M-IgG across M-IgG modified membrane.

The above results demonstrated that the permeation mechanism for anti-H-IgG and anti-M-

IgG across M-IgG modified membranes obeys the dual-mode model proposed in these studies.

The differences in dual-mode model parameters for these two proteins suggested the existence of

a distribution of binding sites inside of these antibody-modified nanotube membranes that are

accessed for each permeating protein in a particular manner. In fact the permeability plots in

Figure 3-14 and Figure 3-16 demonstrated the differentiation of specific transport mechanisms

for each one of the permeating proteins across the M-IgG modified membranes. The higher









permeability for anti-H-IgG, and the higher concentration-dependence in permeability for the

anti-M-IgG are clearly evidences that the permeation mechanisms across M-IgG modified

membranes for each one of these proteins are different and highly depending on the affinity

interactions with the immobilized carrier in the membrane.

The closeness of the two obtained diffusion coefficients DD and DH, also confirmed the

partial immobilization of the Langmuir species (F=1) on the Langmuir sorption mode (Table 3-

4). This also confirms the contribution of the Langmuir species were the responsible factor of the

enhancement of permeability observed in Figure 3-14.

Table 3-4.Validation of the dual-mode model parameters
Permeating DD DH
Protein (cm2/sec) (cm2/sec) F
anti-H-IgG 7.61 x 10- 8.11 x 10- 1.06

anti-M-IgG 1.02 x 10-9 1.02 x 10-9 1.00


Conclusions

With this work, we have demonstrated the dual-mode model proposes here is suitable

model for the characterization of the transport properties of anti-H-IgG and anti-M-IgG in M-

IgG modified membranes. The model demonstrated a good agreement with the experimental

system studied here. Besides, the findings showed in these studies allow addressing the concept

of modulated facilitated transport systems using antibody-modified nanotube membranes, and

how the extent of the facilitation is determined and controlled by the binding affinity between

carrier and permeating protein molecules. These studies can provide a new direction in the

development of more effective membranes for membrane-based separation.










CHAPTER 4
METAL ION AFFINITY NANOTUBE MEMBRANES: PREPARATION,
CHARACTERIZATION, AND CONTROLLED TRANSPORT OF SIX-HISTIDINE-TAGGED
RECOMBINANT PROTEIN

Introduction

Since the introduction of Immobilized Metal Ion Affinity (IMAC) by Porath and

coworkers126 in 1975, IMAC has become in one of the most widely used methods for protein

purification.241 250 IMAC exploits the fact that many transition metal ions, such as, copper,

nickel, cobalt and zinc, can be coordinated to amino acid sites exposed on the surface of the

proteins. It has also been proposed that metal ions in IMAC can be classified into three groups:

hard, intermediate and soft.251, 252 This classification has been based on the preferred reactivity of

the metal ions toward their nucleophiles. Thus, metal ions such as Fe+3, Ca+2, and Al+3, which

show preference for oxygen belongs to the hard metal ion group. Those ions that prefer sulfur

(Cu+, Hg+2, Ag+, etc.) correspond to soft metal ions. Then, the Cu+2, Ni+2, Co+2, and Zn+2 ions,

which coordinate nitrogen, oxygen and sulfur, compose the intermediate ions. Given that the

number of cysteine residues on the surface of proteins is limited, makes the histidine residues the

main targets for intermediate metal ions. For example, a binding constant of 4.5 x 103 M-1 has

been determined for an average protein with a single histidyl residue.253

Thus, multipoint histidyl interactions can be used to control the strength of binding of a

protein in IMAC applications. The advanced understanding of which aminoacid site chains on

the surface of proteins are responsible of the interaction with the metal ions, has been one of the

most important contributions for the development of tagged peptides containing multiple

histidine residues.254 255 In particular, the hexahistidine tags attached to the 5' or 3' end of the

target gene that encodes the protein are usually used as purification tags in recombinant

proteins.256,257









Today, recombinant protein purification represents one of the principal protein purification

markets, and many systems have been developed in recent years for the purification of these

proteins.241 250,257 The system proposed here, exploits the advantages of IMAC separations, but

in a nanopore-membrane- platform. The use of imidazole-modified silica nanotube membranes

for the controlled transport of histidine-tagged recombinant proteins is described.

Imidazole is a functional aromatic nitrogen heterocycle important in ligand coordination

chemistry. The structure of imidazole is a five-membered ring with two different nitrogen atoms.

One of the nitrogen is bound to hydrogen, having two lone pair of electrons in an unhybridized

p-orbital donated to the aromatic ring, and it is structurally related with the pyrrole nitrogen. The

second nitrogen structurally related with the pyridine nitrogen, donates its single electron in an

unhybridized p-orbital to the ring, containing also a free lone pair of electrons in a hybrid

orbital.258 In aqueous solution imidazole behaves like a weak organic base with pKa = 6.99.259

Therefore, the lone pair of electrons in the pyridine-like nitrogen atom, are only involved in

coordination when the pH of aqueous solutions is close or higher than 6.99. It is worth

mentioning, that the mode of coordination of imidazole can be also influenced by the nature of

the metal.

In this investigation, the silica nanotube membrane was prepared by sol-gel template

synthesis of silica nanotubes within the pores of a nanopore alumina template. Imidazole was

attached to the inside walls of chloropropyl-modified silica nanotubes by nucleophilic

substitution reaction with 2-mercaptoimidazole in anhydrous tetrahydrofuran. The modified

membrane was then exposed to solutions of the desired divalent metal ion (copper, nickel, cobalt

or zinc) to attach the metal via coordination with the imidazole. The histidine-tag on the









recombinant protein reversibly coordinates with the metal ion sites, and via this chemistry the

flux of the protein across the membrane could be modulated.

In these studies the concept of IMAC separations was exploited in a nanopore-membrane-

platform. The transport of histidine-tagged recombinant proteins in metal ion affinity nanotube

membranes is described. It was demonstrated that transport of histidine-tagged recombinant

proteins across imidazole-modified silica nanotube membranes can be controlled by the nature of

the immobilized metal ions on the imidazole surface. The imidazole functionalized membrane

was extensively characterized using three different analytical techniques in order to identify how

the coordination chemistry with different divalent metal ions affects the functional behavior and

transport properties during the permeation of six-histidine recombinant proteins. Furthermore,

additional results shown that by the appropriate choice of the transition metal ion for

coordination with the imidazole ligand, the binding strength and optical properties within the

functional membrane can also be modified. These studies demonstrated an alternative way of

immobilized metal ion affinity by using nanotube membranes.

Experimental

Materials

Commercially available nanopore alumina membranes (60 gm thick, nominal pore

diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS)

(Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes.

A 3-chloropropyltrimethoxysilane (United Chemical Technologies) was used for silane

modification of the silica nanotubes. 2-mercaptoimidazole from Sigma Aldrich was used as

attacking nucleophile to substitute the chlorine atoms in the 3-chloropropyl silica nanotube

functionalized membrane. As an aprotic solvent anhydrous tetrahydrofuran >99.9%, inhibitor

free from Sigma Aldrich was used. Divalent salts of copper, nickel, cobalt and zinc nitrate









obtained from Fulka, were used to attach the metal via coordination with the imidazole. A six-

histidine-tagged ubiquitin human recombinant protein (6His-UHRP) obtained from Sigma

Aldrich was used to investigate the transport properties of the imidazole-modified membranes.

All other chemicals were of reagent grade and used as received. Purified water, obtained by

passing house-distilled water through a Barnstead, E-pure water-purification system, was used to

prepare all solutions.

Preparation of the Silica Nanotube Membranes

Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-

based template-synthesis method described previously. Briefly, a sol-gel precursor solution was

prepared by mixing absolute ethanol, TEOS and 1 M HC1 (50:5:1 by volume), and this solution

was allowed to hydrolyze for 30 minutes. The alumina membranes were then immersed into this

solution for two minutes under sonication. The membranes were then removed and the surfaces

swabbed with ethanol to remove precursor from the alumina surface. The membranes were then

air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. This

yields silica nanotubes, with wall thickness of 3 nm, embedded within the pores of the alumina

membrane.9

Preparation of Imidazole Functionalized Membranes

The 3-chloropropyltrimethoxysilane was first attached to the inside walls of the silica

nanotubes. A solution that was 2% (v/v) in 3-chloropropyltrimethoxysilane was prepared in

ethanol; to this solution was added acetate buffer (50 mM, pH=5.1) to make the solution 5%

(v/v) in this buffer. This ethanol-based solution was stirred for 20 minutes and the alumina

membranes were then immersed. The soak membranes were sonicated for 2 minutes, and after

20 minutes the membranes were removed from the solution and rinsed with ethanol. To make

sure that the loose silane was removed from the surface, the membranes were also sonicated for










2 minute in ethanol. The membranes were then dry with nitrogen and placed under vacuum for

12 hours. The resulting 3-chloropropyl silica nanotube functionalized membranes were then

immersed into different vials containing a solution that was 0.5 M in 2-mercaptoimidazole

dissolved in anhydrous THF at 45 C. This reaction was kept at this temperature for a period of 1

hour, with periodically additions of THF to continually replace the evaporated THF. Then, the

lids of the vials were tightly closed and they were wrapped with paraffin film, and the

membranes were incubated in the dark for 24 hours at 25 C to allow the completion of the

reaction. After imidazole functionalization, the membranes were rinsed copiously with ethanol

and sonicated for 2 minutes to remove the unreacted mercaptoimidazole. A schematic illustration

of the reaction mechanism for the preparation of imidazole functionalized membranes is shown

in Figure 4-1.





M o H UC:.o.C u O + CH30-SI-(CH2),-CI
-OH /
OCHOHIHO "A E 0//


C2HOHICHCOONa
CHOH




M 01-(C SI -SI-(CH,2)-C, N
UB 0N THF 0U 0 o
MR N\ .. R 0 N '
N N SI 0 I-(CH2)3 S N N SI-0-I-8i(CH2)3--CI H
SH MCI / /
I

Figure 4-1.Reaction mechanism of imidazole functionalized nanotube membranes.

Divalent Metal Ions Complexation

The imidazole modified membranes were then exposed to solutions of the desired divalent

metal ion (copper, nickel, cobalt or zinc) to attach the metal ion via coordination with the









imidazole. Briefly, the imidazole functionalized membranes were immersed in different vials

containing solutions prepared in ethanol that were 1 mM in Cu+2, Ni+2, Co+2, and Zn+2

respectively. The vials were placed in an orbital shaker platform and the membranes were

equilibrated for 12 hours at 25 C to affirm the complexation of the metal ions on the membrane.

After equilibration the membranes were removed and rinsed copiously with ethanol.

Instrumental Characterization of the Imidazole Functionalized Membranes

To characterize the imidazole functionalized nanotube membranes fluorescence

microscopy, Fourier transformed infrared attenuated total reflectance spectroscopy (FTIR-

ATR), and X-ray photoelectron spectroscopy (XPS) measurement were taken.

We discovered that the imidazole membranes are fluorescent after functionalization.

Therefore, the observation of these imidazole functionalized membranes under the fluorescence

microscope was used as the first control to confirm the success of the chemistry modification.

The fluorescence intensity was measured using a fluorescence microscope which was equipped

with both a fluorescence detector and CCD camera. This system combines an Axioplan 2

imaging microscope (Zeiss) with a J&M-PMT photometry system detector (SpectrAlliance), for

measuring the fluorescence intensity using a BP450-490 filter for excitation and a LP515 filter

for emission of light.

The FTIR-ATR was employed to examine the presence of characteristic bands for 2-

mercaptoimidazole in the functionalized membranes. The FTIR-ATR spectra were recorded on

a Thermo Nicolet Nexus 670 FTIR spectrometer equipped with a GATRTM as the ATR crystal.

All spectra were run at 4000cm-1 to 400cm-1, 512 scans, 4 cm-1 resolution, and gain equal to 1.

The spectrometer aperture was set to 100 for all runs. A background spectrum (air) was taken

prior to each sample spectrum, which was used for background subtraction using the OMNIC

ATR software for the ATR correction.









The surface composition of imidazole functionalized membranes and the presence of the

immobilized transition metal ions were also determined through XPS analysis. The analysis was

performed using a XPS/ESCA Perkin-Elmer PHI 5100 ESCA system. For the XPS measurement

a Mg Ka (1253.6 eV) or Al Ka (1486.6 eV) monochromatized X-ray source was used to

stimulate photoemission. Survey and high resolution spectra were obtained with a hemispherical

energy analyzer operated in the constant analyzer transmission mode to measure the binding

energies of emitted photoelectrons.

Transport Experiments

The imidazole-modified membranes were sandwiched between two pieces of Scotch tape

that had 0.13 cm2 area holes punched through them. These holes defined the area of membrane

exposed to the contacting solution phases. This membrane assembly was mounted between the

two halves of a U-tube permeation cell, with half-cell volumes of 2.5 mL.

The feed solution for the recombinant protein experiments was 1.87 [M dissolved in 50

mM phosphate-buffered saline pH 8.0 that also contained 300 mM sodium chloride and 2 mM

imidazole. The receiver solution on the other side of the membrane was identical except it

initially contained no protein. The concentrations of the transporting protein in the feed solution

were in all cases the same. It is important to mention that prior to starting each transport

experiment the metal-immobilized imidazole membranes were equilibrated in the PBS buffer

solution for two hours, and that the metal ion immobilization for each divalent cation was

completed right before this equilibration.

The rate of protein transport across the membrane was determined by periodically

checking the absorbance at 276 nm in the receiver solution using a flow-through UV-Vis

detector (Figure 4-2).























Figure 4-2.Flow-through UV-Vis detection system used to measure the flux of six-histidine-
tagged ubiquitin human recombinant protein.

This system consisted of a L-7455 UV-Vis detector (Hitachi), where the flow stream was

delivered using an L-7100 HPLC pump (Hitachi) using a flow rate of 3 mL per min. The whole

system was controlled using Chromeleon 6.50 chromatography management software (Dionex)

through a UCI-100 universal chromatography interface (Dionex).

Calibration curve was obtained with the same flow through system on solutions containing

known concentrations of the transport recombinant protein. This calibration curve was used to

obtain the concentrations of the transport proteins in the receiver solution.

Results and Discussion

Imidazole Surface Modification and Characterization

The imidazole functionalized nanotube membranes were prepared by covalent

immobilization of imidazole inside of the silica nanotubes. This coupling between silica

nanotube surface and imidazole was accomplished via simple silane chemistry using 3-

chloropropyltrimethoxysilane. Silica surfaces offer the positive attribute that they can be

derivatized with an enormous variety of chemical functional groups using simple silane

chemistry95. Here, the covalent immobilization of imidazole was achieved by nucleophilic

substitution of chlorine atoms in the resulting silane modified 3-chloropropyl silica nanotube











membranes. It has been demonstrated that mercapto compounds are appropriated modification

reagents for this strategy.106 260 263 Furthermore, it has been confirmed that the use of aromatic

thiols leads to degrees of modification which are much higher than those achieved with aliphatic

ones. In addition, when aromatic thiols are used the possibilities of elimination as a side reaction

are eradicated.261

In order to promote the substitution nucleophilic bimolecular mechanism (SN2) the

imidazole modification was carried out in an aprotic solvent (THF). Since aprotic solvents do not

solvate the nucleophiles, the reactivity of the 2-mercaptoimidazole can be enhanced making it

more accessible to approach to the primary halide carbon as the conditions in 3-chloropropyl

silica nanotube modified membranes. A high concentration of 2-mercaptoimidazole was also

used to promote the SN2. It is worth mentioning that some indications that the reaction is taking

place were the sulfur smell during the reaction and the lower pH of the solution at the end of it.

Fluorescence microscopy characterization

We found that after imidazole immobilization the nanotube membranes were fluorescent,

and this observation become in our first control to monitor the success of the modification

chemistry.


25000
o) :
i 20000 -

S15000
09 20000
15000 -
0 I
10000 -,

5000 '



400 450 500 550 600 650 700
Wavelength (nm)
Figure 4-3.Fluorescence spectra of (a) silica nanotube and (b) imidazole functionalized
membranes.











Figure 4-3 shows the fluorescence spectra of the nanotube membranes before and after

imidazole functionalization. It is clearly seen from these data that the fluorescence intensity for

the imidazole functionalized membrane was nine-fold higher compared to the bare silica

nanotube membranes. This fluorescence signal was attributed to the imidazole moiety in the

functional membrane. The origin of this fluorescence signal was also confirmed by conducting

fluorescence experiments of pure 2-mercaptoimidazole in methanol.

Additionally, we found that by the choice of the appropriate transition metal ion for the

complexation with the imidazole, the optical properties within the functional membrane can be

modified (Figures 4-4 and 4-5).

00000
sooo0 o .- .Imdazole hemb Imidazole Mbmb
A ;: Co Deposit ... D B --- co oepositon
S 0020000
2 0 2000 '
o lo0 o

5000











Wavelength (nm)
Figure 4-4.Fluorescence spectra of imidazole functionalized membranes before (upper) and after
(lower) metal ion immobilization: (A) Co+2, (B) Cu+2, (C) Ni+2, and (D) Zn+2

Figure 4-4 shows the fluorescence spectra for a set of four imidazole functionalized


membranes that were exposed respectively to four different nitrate salts solutions of Cu+2, Ni+2

Co+2, and Zn2 videe infra). It was observed that after immobilization of these transition metal

ions the fluorescence intensity of the functional membranes was quenched.


Figure 4-5 reports the fluorescence quenching ratio F/Fo, where Fo is the fluorescence of

imidazole functionalized membrane and F is the corresponding fluorescence in the presence of










the metal ions. It is evident that fluorescence quenching was much higher for Cu+2 and Co+2 ions,

while it was weaker in the case ofNi+2 and very low for Zn+2


0.7

0.6 -

0.5

o 0.4
LL.
"L 0.3

0.2

0.1

0.0 ] L
Cu (ll) Ni (II) Co (II) Zn (II)
Metal Ion Immobilized
Figure 4-5.Fluorescence quenching (F/Fo) for imidazole functionalized membranes in the
presence of nitrate solutions prepared in ethanol that were 1 mM of the respective
divalent metal ion.

Fluorescence quenching by transition metal ions is a common phenomenon.264 The

fluorescence quenching by transition-metal ions and their coordination complexes is a process

which has been known for a long time and is still the subject of intense investigations.265 267 It is

now well recognized that electron transfer and electronic energy transfer are considered as the

two main deactivation pathways responsible for efficient fluorescence inhibition.264- 267

These metallosystems based on fluorescence detection have found applications in many

areas such as metal ion sensing,268' 269 artificial photosynthesis,270 and optical sensing of DNA.271

-273


On the other hand, when the imidazole functionalized membrane was exposed to solutions

of CuC12 instead of Cu(NO3)2, the optical behavior for this metal ion immobilized membrane

was different than in the case of nitrate salts. Under this condition, a fluorescence enhancement

instead of quenching was observed in the presence of the metal ion (Figure 4-6).










Figure 4-6 compares the fluorescence spectra for imidazole functionalized membranes

before and after coordination with CuC12. Simultaneously with the fluorescence enhancement, a

small bathochromic shift of the maximum emission wavelength was observed in the fluorescence

spectra of the copper immobilized membrane from 443 nm to 550 nm.

45000
S40000
35000
1 30000 -
25000 a
0
0 20000
S15000 -
1 0000 -
5000 -


400 450 500 550 600 650 700
Wavelength (nm)
Figure 4-6. Fluorescence spectra of imidazole functionalized membranes (a) before and (b) after
metal ion immobilization using an ethanol solution that was 1 mM in CuC12.

Besides, a dependence on the copper concentration for the fluorescence enhancement was

found (Figure 4-7). As can be seen from Figure 4-7 the addition of the chloride salt of copper

leads to an increase in fluorescence intensity of the functional membrane and at a certain

concentration the fluorescence intensity remains constant. Such an increase in fluorescence

enhancement can result from conformational restriction induced upon binding.274

It is beyond the scope of the present work to discuss in detail the mechanism of interaction

of the functional imidazole membrane with these metal ions. Nevertheless, the observed

fluorescence intensity changes can be reasonably explained on the basis of the complex structure

formed after coordination takes place. It was ascribed that for the situation of the imidazole

membranes coordinated with the nitrate salts, the nitrate anions in [M(Im)n](N03)2 complexes

are strictly counterions, hence they are no part of the sphere of coordination. Therefore, the










fluorescence quenching through an electron transfer mechanism can be accredited.264' 265 The

great 7t-acceptor properties of the imidazole permit it to accept electronic charge from d orbitals

on the metal ion.258 These charge transfer transitions correspond roughly to the transfer of an

electron from the metal ion to a ligand. More precisely, an electron is transferred from a bonding

orbital of mostly metal ion character to an empty antibonding orbital of mostly ligand

character.275'276




12 -


8 -
0
u- /
U-
4 /


0
0 20 40 60 80 100
CuCI, Concentration (mM)
Figure 4-7. Copper concentration dependence on fluorescence enhancement for imidazole
functionalized membranes.

Conversely, in the situation using chloride salts, the chloride anions can be part of the

sphere of coordination. Therefore, in the [M(Im)n(C1)2] complex both the imidazole and chloride

ligands can be coordinated to the metal ion. Furthermore, for that reason that chloride is known

as a 7t-trans-directing ligand, the enhancement in fluorescence can be ascribed to a Rt-trans-

influence, so that a trans-Tt-acceptor ligand as chloride, can form Rt bonds with the metal and

withdraw electron density away from the other imidazole ligand.275'276 Then, the 7t-trans bonding

have a cooperative effect allowing more delocalization of the n electrons.277 The geometrical

restraint required for having this interaction, support that fact that the increase in fluorescence

enhancement can result from conformational restriction induced upon binding.269 274 McFarland

and Finney269 274 demonstrated that metal ion binding could restrict the rotational excited-state of









a fluorophore suppressing some of the fluorescence deactivation pathways, producing

fluorescence enhancement.

Thus, those changes in fluorescence intensity showed how the optical properties of the

functional imidazole membranes changes depending on the transition metal ion immobilized.

Besides, these results confirm the accomplishment in the preparation of imidazole functionalized

membranes.

Characterization by FTIR-ATR spectroscopy

The successful modification of the silica nanotube membrane with the imidazole moiety

was also demonstrated by FTIR-ATR spectroscopy. Figure 4-8 compares the FTIR-ATR spectra

for bare alumina and imidazole functionalized membranes. It should be pointed out that all

spectra were corrected for spectral distortion using the OMNIC ATR software. However, by

using this ATR correction it is possible to lose spectral information of the fundamental

vibrational modes of a molecule, overall when the entire spectral region is examined at once. To

avoid this, it is a common method to analyze the data in limited spectral regions in order to give

a more detailed assignment of the characteristic bands. The Figure 4-8B illustrates this issue, and

a narrow spectral region, from 2000 to 600 cm-1, for the imidazole functionalized membrane is

used.

The differences observed in the spectra of bare alumina and imidazole-modified

membranes clearly demonstrated the modification of the nanotube membranes by using 2-

mercaptoimidazole. For the FTIR-ATR spectrum of bare alumina two broad bands appear in the

3900-3200 cm-1, and 3000-1600 cm-1 regions, which are attributed mainly to the O-H stretching

modes and O-H deformation modes, respectively.278'279 Although, for these two broad bands the

contribution of water (at about 3600 cm-1 and about 1600 cm-1) also has to be considered, since it

is known that hydrogen bonds with water molecules contribute to broaden peak areas.280'281


















Imidazole functionalized --- II
a


Bare alumina --. i

A B v
4000 3500 3000 2500 2000 1500 1000 500 2000 1800 1600 1400 1200 1000 800 600
Wavenumber (cm') Wavenumber (cm1)
Figure 4-8. (A) FTIR-ATR spectra of bare alumina and imidazole functionalized membranes. (B)
FTIR-ATR spectra of imidazole functionalized membranes in a narrow spectral
region illustrating the effect of using two different backgrounds for the spectrum
subtraction: (a) air, and (b) 3-chloropropyl silica nanotube membrane.

For the case of imidazole membranes, the observed bands were assigned by comparison of

their frequencies with the transmission FTIR spectrum of pure 2-mercaptoimidazole in KBr

(Figure 4-9).263, 282-285


100

-80 -

S60

E40

I-
20

0 -


4000 3500 3000 2500 2000 1500 1000 500
Wavenumber (cm')
Figure 4-9.Transmission FTIR spectrum of pure 2-mercaptoimidazole in KBr.

The FTIR-ATR spectrum of the imidazole functionalized membrane showed clearly three

of the characteristic vibrations for 2-mercaptoimidazole. One was the ring vibration modes for

C=C and C=N bonds in the region -1684 1630 cm-1. A second band, observed at -1486 1400

cm-1 region that was assigned to the thioamide bands I and II, which are coupled vibrations










involving stretching motion of C-S + N-H and bending motion C-N + N-H, respectively.286 The

third band was the very strong band at 2890 cm-1 assigned to N-H stretching mode. This

imidazole N-H stretching appeared as a broad band in a wide range from -3400 to 2200 cm1.

When the characteristic band of N-H was compared with the band of pure 2-mercaptoimidazole a

shift from 3145 cm- to 2890 cm- was observed. This shifting might be due to the solid state

characteristic of the sample, but also to the formation of hydrogen bonds between imidazole

moieties. All the assigned bands were also observed in metal-immobilized imidazole

functionalized membranes, and some of them were enhanced by the presence of the immobilized

metal ion as is shown in Figure 4-10.


a
b 1420

1368




S1320



2000 1800 1600 1400 1200 1000 800 600
Wavenumber (cm1)
Figure 4-10.FTIR-ATR spectra of imidazole functionalized membranes showing the effect of the
metal ion immobilized on the thioamide bands: (a) Zn+2, (b) Co+2, and (c) Cu+2

Figure 4-10 compares the results for three different imidazole functionalized membranes

which were exposed to solutions of divalent metal ions of copper, cobalt and zinc, respectively.

It was found that the enhancement of the band intensity is related to the strength of the binding

constant between the metal ions with the functional membrane.258 In addition, it was observed

that those bands shifted toward lower frequencies as is demonstrated in Figure 4-10. These

changes of the band intensity may be attributed to changes of the electric dipole moment after

coordination with the metal ions.287 It may be possible that IR energy is preferential absorbs for









some favored orientations of the imidazole moiety in the functional membrane after

coordination.

The observed band at 1110 cm 1 was also assigned, and it is clearly illustrated in Figure 4-

8B. This band was labeled to the asymmetric C-S-C stretching vibrations.284, 288 For the correct

assignment of this band, the contribution of Si-O stretching vibration at this region was also

considered. In order to confer a proper assignment, the FTIR-ATR spectrum of imidazole

functionalized membrane was also acquired by using 3-chloropropyl silica nanotube membrane

as background for the spectrum subtraction instead of air (Figure 4-8B).

Figure 4-8B compares the FTIR-ATR spectra of an imidazole functionalized membrane,

which were collected using both air and 3-chloropropyl silica nanotube membrane for

background subtraction, respectively. It was clearly observed that when the chlorosilane-

modified membrane was used for the background subtraction, the fundamental vibrational mode

from the imidazole moiety at 1118 cm-1 was more clearly discriminated. Therefore, at 1118 cm-1

the contribution of both Si-O and asymmetric C-S-C stretching vibrations should be

considered.284, 288, 289

Table 4-1.FTIR-ATR characteristic bands of imidazole functionalized nanotube membranes

Assignment Frequency (cm-)


N-H stretching -2900

N-H bending in the plane -1560

C=C and C=N ring vibration modes -1684-1630
Thioamide bands I and II, which are respectively coupled vibrations
involving stretching motion of C-S + N-H, and bending motion C-N + N-H. -1486 and 1412

C-S-C anti-symmetric and Si-O stretching vibrations. 1118
11 lo










Table 4-1 shows the assignment of the characteristic bands for the imidazole functionalized

membranes. In some cases the overlapping between bands made this assignment difficult. This

was particularly more apparent in the region for N-H bending vibration and the coupled

vibrations involving C-S and N-H of the thioamide band I.263

Transmission FTIR spectra for the alumina template, silica nanotube, and nanotube

modified membrane films were also studied. Figure 4-11 shows the transmission FTIR spectra of

the alumina template film for each of the modification steps illustrated in Figure 4-1.


100 -

80 \




S40 -

20 -

oa
4500 4000 3500 3000 2500 2000 1500 1000 500 0
Wavenumber (cm"1)
Figure 4-11. Transmission FTIR spectra of the alumina template film for each of the
modification steps illustrated in Figure 4-1: (a) bare alumina template, (b) silica, (c)
3-chloropropyl silane modified and (d) imidazole functionalized nanotube
membranes, respectively.

By observing the transmission FTIR spectra in Figure 4-11, there were no obvious

variations of the characteristic bands related with each chemistry modification. However, it was

possible to distinguish through the different modification steps by monitoring the changes in

transmittance in each transmission FTIR spectrum.

It was clearly observed that the absorption of stretching vibrations band for the O-H groups

in the region between 3000 and 3600 cm-' changes throughout the different modification steps.

Likewise, differences in the region 1500 and 1600 cm-1 for the bending vibration of O-H were









also observed. In the latter region, the differences between the bending vibrations of the O-H for

the alumina membrane and the silica nanotube membrane were not significant (spectra a, and b

in Figure 4-11). This was reasonable because part of the OH groups for the alumina are replaced

for the silanols OH groups in silica; besides silica surface introduce higher silanol groups on

surface for further silane functionalization.30 94, 99-102

In the case of chlorosilane modification (spectrum c in figure 4-11), lower absorption for

the silane-modified membrane was observed in the bending O-H region compared with alumina

and silica. This difference in transmittance clearly confirmed the success of the silanization

reaction where the O-H free groups in the alumina matrix and the silanol groups in the silica

nanotubes were replaced by the condensation reaction with 3-chloropropyltrimethoxysilane.

The differences in transmittance were even more evident for the imidazole functionalized

membrane (spectrum d in Figure 4-11). Additionally this exhibited two observable absorption

bands at 1118 cm-1 and 474 cm-1 assigned respectively to the anti-symmetric C-S-C, and to the

C-S stretching vibrations, which confirmed the imidazole functionalization.

These differences in the intensity of the bands might be attributed once more to changes of

the electric dipole moment after each chemistry modification.287 Besides, each chemistry

modification introduces a variety of vibrational motions, and degrees of freedom characteristic of

its component atoms that should produce a net change in the dipole moment. It should be also

considered the fact that may be after the modifications the membranes become less hydrophilic

than the bare alumina template, thereby the contribution of water on surface must be less

significant, even though all the membranes were dried at the same conditions.

These transmission FTIR data in Figure 4-11 may perhaps be interpreted on the basis of

the infrared absorption spectral region. Thus, the shorter wavelengths can penetrate less the










alumina matrix and transmit more IR energy through the porous structure of the membrane to

induce vibrational excitation of the covalently bonded atoms on the nanotube surface. On the

other hand, longer wavelengths should have deeper penetration on the alumina matrix, allowing

less transmission of the IR energy. Therefore, the latter region at the same time should be more

sensitive to changes associated with electric dipole moment, and more information may be

obtained in this region.

Surface chemical composition by XPS

XPS analysis was performed not only to assure the presence of chlorine and imidazole

groups in the modified membranes, but also to confirm the immobilization of the transition metal

ions on the functionalized imidazole membranes. This analysis begins, however, with the XPS

survey studies for the bare alumina membrane (Figure 4-12).

80000
0 Auger
70000 -

60000 -





P 2p3 Al 2s
SCis P2s AI2p

20000 02s

10000 1
1000 800 600 400 200 0
Binding Energy (eV)
Figure 4-12.XPS survey spectrum for the bare alumina template.

The survey spectrum for the alumina template is shown in Figure 4-12. The performed

analysis showed the presence of elemental chemical composition peaks for Al, O, P, and C. The

unexpected peak for Cls at 289.3 eV was attributed mainly to impurities on the surface; this

carbon impurity is explained below in more detail.










Survey scans were taken to determine the elemental surface composition of various

elements present in the chlorosilane and imidazole modified membranes. Figure 4-13 compares

the XPS survey scan of the chlorosilane and imidazole modified membranes.

120000

100000

80000

", 60000
c Cls

40000 CI 2p

20000
a
o b

1200 1000 800 600 400 200 0
Binding Energy (eV)
Figure 4-13.XPS survey spectra for (a) chlorosilane and (b) imidazole modified membranes.

The chlorosilane modified surface shown the presence of XPS line for chlorine C12p,

which was absent in the bare alumina, which confirmed the accomplishment of the chlorosilane

modification of the silica nanotubes (spectrum a Figure 4-13). In addition, this survey scan

shown distinct peaks for A12p as well as for C Is compared to the imidazole functionalized and

bare membranes.

It was noticed that after functionalization with 2-mercaptoimidazole, an increase in C1s

peak and a diminution in the A12p peak was observed when it is compared with the chlorosilane

membrane (spectrum b Figure 4-13). Considering the surface sensitivity of XPS, the decrease of

the A12p peak in the XPS measurements for the imidazole functionalized membranes suggested

the presence of imidazole on the nanotube membrane surface.136 Additionally, the

functionalization of the membranes with the imidazole moiety was confirmed for the presence of

Nls and Sis lines in the XPS spectra.









The chemical composition and binding energies data from such spectra are summarized in

Table 4-2. It is worth mentioning that owing to the insulator characteristics of the alumina, the

identification of the specific peaks was demanding. Moreover, due to the surface charging a

positive shift toward higher binding energy was observed for some peaks.

Table 4-2.XPS surface composition and binding energies for bare alumina, chlorosilane
modified, and imidazole functionalized membranes
Bare Chlorosilane Imidazole
Alumina Modified Functionalized

XPS Signal % BE(eV) % BE(eV) % BE(eV)

Na ls 4.4 1075.6 1.0 1074.5
O ls 60.3 536.2 52.5 534.5 49.8 534.0
Nls 3.9 402.0
C Is 7.4 289.3 11.9 288.0 20.4 287.0

C1 2p 1.4 203.0
S 2p 1.5 171.0
P 2p 2.8 138.4 2.3 137.0 2.1 136.0
Si 2p 1.8 104.5 1.4 104.0
Al 2p 29.5 78.4 25.7 77.5 19.9 76.5


Furthermore, the XPS multiplexes analysis at high resolution for the imidazole membranes

(Figure 4-14) confirmed the imidazole functionalization. An analysis that was based only on the

relative areas of the XPS peaks at high resolution demonstrated an elemental compositional ratio

of 6:2:1 for C: N: S. However, after the individual peak areas were divided by their respective

atomic sensitivity factor (ASF) the elemental compositional ratio for C: N: S respectively was

14:2:1. This increase in C signal was attributed to impurities on the surface of the alumina

membranes. These carbon impurities were corroborated by observing the high-resolution scan of

the Cis peak (Figure 4-15).



























Figure 4-14.XPS multiplexes analysis at high resolution for the imidazole functionalized
membrane.


Figure 4-15 compares the high-resolution C s peaks for the bare alumina membrane, the


chlorosilane modified membrane and the imidazole functionalized membrane, respectively.


4000



3000



2000



1000


294 292 290 288 286 284 282
Binding Energy (eV)

Figure 4-15.High-resolution XPS data of the C s peak for (a) bare alumina template, (b)
chlorosilane modified, and (c) imidazole functionalized membranes.


It was clearly seen that the C s peak for Imidazole modified membranes consisted both of


a well defined peak at 287 eV which was ascribed to the C-C peak and a shoulder at higher


energy binding where the C-N, C-S peaks are overlapped. On the other hand, the C s peak for


chlorosilane modified membranes revealed a better indication of a second peak at 291 eV which


may be assigned to the C-C1 environment. The high-resolution scan of C s peak for the bare


-a i
-


-' ~`""
rar rrnw
~II
NI)
\II)










alumina membrane consisted of only one peak the C-C peak at 287 eV, which confirmed the

presence of carbon impurities in the membrane.

The immobilization of the transition metal ions in the imidazole functionalized membranes

was also verified by XPS analysis. Figure 4-16 compares the XPS survey scan of two imidazole

functionalized membranes after immobilization of Cu and Zn, respectively. For the XPS survey

scan of Cu and Zn metal-immobilized nanotube membranes, the Mg anode and the Al anode

were used respectively to stimulate photoemission.

80000
Cu 2pl
70000 a
Cu 2p3
60000 Zn 2p1

50000 Zn 2p3

S40000 -
b
S30000 Cu Auger
Zn Auger
20000
10000

0
1200 1000 800 600 400 200 0
Binding Energy (eV)
Figure 4-16. XPS survey spectra for imidazole modified membranes after immobilization of (a)
Cu+2, and (b) Zn+2

The presence of XPS spectral lines for Zn and Cu were observed in the respective

spectrum of the imidazole functionalized membranes. The XPS survey scan for Zn shown the

presence ofZn2pl at 1050 eV and Zn2p3 at 1026.5 eV peaks. Whereas the survey for Cu shown

the presence of Cu2pl and Cu2p3 peaks assigned at 954 eV and 934 eV, respectively. It should

be mentioned that analogous experiments were also carried out for the immobilization of Ni and

Co, and Auger lines for all these four metal ions were also observed. Thus, these data confirmed

the coordination and immobilization of the transition metal ions used in these studies with the

imidazole functionalized membranes.










Transport Properties of Imidazole Membranes

The transport properties of the imidazole functionalized membranes were explored by

monitoring the flux of six-histidine tagged ubiquitin human recombinant protein (6His-UHRP) in

function of time. For the transport experiments, the rate of 6His-UHRP across the membrane was

determined by periodically checking the absorbance at 276 nm in the receiver solution using the

flow-through UV-Vis detector illustrated in Figure 4-2.

Figure 4-17 shows the absorption spectrum of 50 pM solution of 6His-UHRP that was

dissolved in 50 mM phosphate-buffered saline pH 8.0 that also contained 300 mM sodium

chloride and 2 mM imidazole. It is clearly observed that the absorption maximum occurs in 276

nm.

0.5

276
0.4 258


S0.3

-9 0.2


0.1 -


0.0
250 300 350 400 450 500 550 600
Wavelength (nm)
Figure 4-17.UV-Vis spectrum for 50 pM of 6His-UHRP measured in 50 mM phosphate-buffered
saline solution pH 8.0 that also contained 300 mM sodium chloride and 2 mM
imidazole.

Calibration curve at 276 nm was obtained with the flow through system showed in Figure

4-2 on solutions containing known concentrations of 6His-UHRP. This calibration curve was

used to obtain the concentration of the transport protein in the receiver solution. Figure 4-18

shows the typical calibration data (Figure 4-18A) and typical calibration curve (Figure 4-18B)












for 6His-UHRP. The calibration curve shown an excellent linear correlation (r=0.999) within a


range from 0.374 to 1.496 tM in protein concentration.


> 7 B

6 6

3 5

S4

- 3

S2

1
o I
0 -


0.0 2.5 5.0 7.5 10.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Time (min) Protein Concentration (pM)
Figure 4-18.Typical plots of (A) calibration data and (B) calibration curve for 6His-UHRP.


Figure 4-19 shows the nanomoles of the recombinant protein transported across the metal


ion affinity nanotube membranes versus permeation time. In addition, the results of the


permeation experiments for four different metal-immobilized imidazole membranes are


compared.


o 3.0 *

0.
12.5 -

*




( .0 -
G -

E o.5 -
A AA


0.0 *

0 200 400 600 800 1000
Time (min)
Figure 4-19.Flux plots for 6His-UHRP in imidazole functionalized metal ion affinity nanotube
membranes that had immobilized different metal ions: (a) Cu+2, (b) Ni+2, (c) Co+2,
and (d) Zn+2. The feed solution concentration was 1.87 pM for all permeation
experiments.


1.496 iM


0.935 jIM


0.374 uM










We found that the flux of 6His-UHRP through modified membranes can be controlled by

the appropriate divalent metal ion (copper, nickel, cobalt or zinc) immobilized in the membrane

via coordination with the imidazole, as is shown in Figure 4-19. The data clearly demonstrated

that the functional transport of the imidazole functionalized membrane depends on the

appropriate immobilized divalent metal to coordinate with the imidazole ligands in the

membrane. When the Zn ion was immobilized in the membrane the flux was considerably higher

than the flux for the other three divalent metal ions (Co, Ni, and Cu). The flux for the Zn

situation was not linear after 3 hours of permeation due mainly to the rapid equilibration reached

between the feed and receiver solutions. Thus, the coordinated Zn in the imidazole functional

membrane acts as binding sites to facilitate the transport of the 6His-UHRP.


4.0 -

3.5
o 3.0
"C
g 2.5

E 2.0
0
1.5

1.0
x
FL 0.5
0.0
Cu (II) Ni (II) Co (II) Zn (II)
Metal Ion Immobilized
Figure 4-20.Effect of the divalent metal ion immobilized on the flux of 6His-UHRP across
imidazole functionalized metal ion affinity nanotube membranes.

Analogous permeation experiments were obtained for 6His-UHRP with the other three

divalent metal ions. The transport plots of 6His-UHRP with these three metal ions were not quite

linear and their fluxes were always lower than the Zn immobilized situation. The slope of the

flux plots for the four different immobilized metal ions was determined by using transport data









only during the initial three hours of permeation, period where fluxes were apparently linear for

all four metal-immobilized membranes. These results are summarized in Figure 4-20.

In order to quantify the effect of the different metal ions immobilized in the transport of

6His-UHRP across the metal ion affinity nanotube membranes, an enhancement selective

coefficient, a, was defined as the calculated fluxes during three hours of transport for each

divalent metal ions divided by the flux of the Cu situation. Thus, it was observed that during the

three initial hours of permeation, the fluxes across the metal ion affinity nanotube membranes

with Ni, Co and Zn immobilized ions are enhanced respectively in -2, -3 and -9 times,

compared with the Cu immobilized membrane. These results are shown in table 4-3.

Table 4-3.Fluxes and selectivity coefficients a for 6His-UHRP across imidazole functionalized
metal ion affinity nanotube membranes

Metal Ion Flux of 6His-UHRP Flux M +2
Immobilized (nanomoles h- cm-2) Flux. Cu+2

Cu+2 0.43 1.0
Ni+2 0.77 1.8
Co+2 1.47 3.4
Zn+2 4.05 9.4


Thus, these data demonstrated that by the selection of the appropriate transition metal ion

to coordinate the imidazole, the transport properties within the functional membrane can be

controlled.

Conclusions

The development of metal ion affinity nanotube membranes has been demonstrated by

exploiting the concept of IMAC separations in a nanopore-membrane-platform. Silica nanotube

membranes were functionalized with imidazole through a nucleophilic substitution reaction

between 2-mercaptoimidazol and 3-chloropropyl modified silica nanotubes carried out in an









aprotic medium. It was demonstrated that transport of histidine-tagged recombinant proteins

across these imidazole-modified silica nanotube membranes is controlled by the nature of

immobilized metal ion on the imidazole surface. Besides, the extensive characterization of the

imidazole functionalized membrane allowed the understanding of how coordination chemistry

with different divalent metal ions affects the functional behavior of the membrane and transport

properties during the permeation of six-histidine recombinant proteins. Furthermore, additional

results shown that by the appropriate choice of the transition metal ion for coordination with the

imidazole ligand, the binding strength and optical properties within the functional membrane are

also modified. Hence, these studies revealed how the presence of the immobilized metal ion

affects the functionality of the imidazole membrane and its transport properties. The metal ion

affinity nanotube membrane described here has the potential to be a versatile analytical platform,

where different metal ions can be immobilized on the same prevailing conditions, additionally

offering the possibilities of regeneration without apparent loss of the metal coordination

properties of the functional membrane. These studies demonstrate an alternative way of

immobilized metal ion affinity by using nanotube membranes.









CHAPTER 5
TRANSPORT ANALYSIS OF PROTEINS ACROSS SILICA NANOTUBE MEMBRANES
THROUGH A FLUORESCENCE QUENCHING-BASED SENSING APPROACH

Introduction

Molecular recognition is a very important and commonly used tool in both clinical

diagnostics and pharmacology. The practical methods based on recognition depend

fundamentally on specific biomolecular recognition reactions. In some cases, these biomolecular

recognition reactions confer such extremely high specificity that it has allowed the development

of analytical assays for the determination of many chemicals and biochemicals at very low

concentrations.290 293 This is important because in this era of proteomics in particular, the

quantitative determination and identification of proteins at low concentrations are essential

requirements. Many very sensitive methods have been developed for the detection of proteins

and other biological molecules on the basis and capacity of fluorescent compounds to transfer

energy absorbed from photon radiation to close molecules by taking advantage of the molecular
294-297
recognition.294 297

This called fluorescence resonant energy transfer (FRET) phenomenon takes place when

the energy in the excited state of a donor fluorophore is transferred to a close acceptor molecule

via an appropriate orientation of the dipole-dipole interaction between these donor-acceptor

molecules.298, 299 One of the basic conditions for the FRET process to occur is that the emission

spectrum of the donor must overlap the absorption spectrum of the acceptor. Therefore, FRET

efficiency is a distance-dependent interaction process.131' 298 This distance dependence of FRET

has been extensively used to study both structure and dynamics of proteins and nucleic acids, as

well for the detection and visualization of intermolecular associations, and for the development

of biomolecular devices based on intermolecular binding interaction.298 300'301









If the donor and acceptor molecules are brought close together within the range of 20 to

100 A, the fluorescence intensity of the donor fluorophore is reduced (quenched) as a result of

the FRET process. However, if the donor and acceptor molecules are brought within an intimate

distance, most of the absorbed energy is dissipated as heat and only a small amount of the energy

is emitted as fluorescence; this phenomenon is commonly referred to as static or contact

quenching.131 The rate of fluorescence quenching can be studied experimentally by calculating

the Stem-Volmer quench (Ksv) constant.131

In the present study, a new approach is described to study the transport of proteins across

silica nanotube membranes by using a fluorescence quenching-based analytical method. This

analysis was accomplished by simply observing the fluorescence quenching of a donor-

fluorescent labeled protein through a ligand-receptor binding mechanism. Thus, the flux of anti-

Human IgG conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by using a

known concentration of Human IgG FITC-labeled (H-IgG-FITC) as fluorescence emitter and

donor in the permeation compartment. In order to preserve the chemical potential as the actual

driving force for the flux, a much higher concentration of protein in the feed compartment (anti-

H-IgG-QSY) was intentionally kept with respect to the concentration of emitter-labeled protein

in the permeation compartment. It is described that the rate of fluorescence quenching depends

on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of fluorescence

quenching can be used to determine the concentration of the analyzed protein on the feed

solution. The results of this investigation are reported here.

Experimental

Materials

Commercially available nanopore alumina membranes (60 tm thick, nominal pore

diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS)









(Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes.

The purified mouse immunoglobulin (M-IgG) immobilized within the silica nanotube

membranes was obtained from Sigma Aldrich. Membranes modified with purified human

immunoglobulin (H-IgG) obtained from Sigma Aldrich were also prepared. A trimethoxysilyl

propyl aldehyde (United Chemical Technologies) was used to attach the M-IgG or H-IgG to the

nanotubes. A purified human immunoglobulin conjugated with the FITC dye (H-IgG-FITC)

obtained from Sigma Aldrich was used as emitter during the fluorescence quenching detection. A

purified rabbit immunoglobulin conjugated with the FITC dye (R-IgG-FITC) obtained from

Sigma Aldrich was used as control emitter during the fluorescence quenching detection. To

investigate the transport properties across silica nanotube membranes the whole antibody raised

in goat against H-IgG was used was used as permeating protein. This anti-human IgG (anti-H-

IgG) with specific affinity for the Fab fragments of the H-IgG was obtained from Sigma Aldrich.

In order to observe the fluorescence quenching of the donor-fluorescent labeled protein through a

molecular recognition reaction, the anti-H-IgG permeating protein was conjugated with the

QSY-7 quencher. This protein labeling was accomplished via acylation reaction of an N-

Hydroxysuccinimidyl (NHS) ester of QSY-7 to the primary amine residues of the protein to form

a stable amide bond. This NHS ester of QSY-7 was purchased from Molecular Probes (Eugene,

OR). The BioGel P-30 resin provided in the amine-reactive protein labeling kit obtained from

Molecular Probes was used for purification of the anti-H-IgG conjugated with QSY-7.

SuperBlock blocking buffer solution in phosphate-buffered saline was obtained from Pierce.

SuperBlock solution contains a proprietary protein for blocking binding sites for non-specific

protein adsorption. Tween 20 was obtained from Sigma Aldrich. All other chemicals were of









reagent grade and used as received. Purified water, obtained by passing house-distilled water

through a Barnstead, E-pure water-purification system, was used to prepare all solutions.

Preparation of the Silica Nanotube Membranes

Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-

based template-synthesis method described previously. Briefly, a sol-gel precursor solution was

prepared by mixing absolute ethanol, TEOS and 1 M HC1 (50:5:1 by volume), and this solution

was allowed to hydrolyze for 30 minutes. The alumina membranes were then immersed into this

solution for two minutes under sonication. The membranes were then removed and the surfaces

swabbed with ethanol to remove precursor from the alumina surface. The membranes were then

air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. This

yields silica nanotubes, with wall thickness of 3 nm, embedded within the pores of the alumina

membrane.9

Antibody Immobilization

The immobilization chemistry has also been described previously. The aldehyde silane

was first attached to the inside walls of the silica nanotubes. The membrane was then immersed

into a solution that was 1.0 mg per mL either in M-IgG or in H-IgG dissolved in 10 mM

phosphate-buffered saline with pH = 7.4 that also contained 150 mM sodium chloride and 5 mM

sodium azide. The membrane was incubated overnight at 4 oC, which allowed free amino sites

on the antibody to attach via the Schiff s base reaction to the nanotube-immobilized aldehyde

groups. After immobilization, the membranes were rinsed with buffer and then incubated for 5

hours in the SuperBlock solution that also contained 0.05% Tween 20. It is worth mentioning

that the unmodified silica nanotube membranes were also incubated for 5 hours in the

SuperBlock solution containing 0.05% Tween 20 before the transport experiments.









Labeling of anti-H-IgG with QSY-7

Forty microliters of 6 mM QSY-7 solution prepared in absolute DMSO was added to a 1

mL of 13.3 [LM of anti-H-IgG solution prepared in 10 mM phosphate-buffered saline with pH =

7.4 that also contained 150 mM sodium chloride. This solution was mixed thoroughly by gentle

vortexing. The reaction was incubated at room temperature for 2 hours on an orbital shaker

platform. Unconjugated QSY-7 was separated from the protein-quencher conjugated by size

exclusion chromatographic using BioGel P-30 resin equilibrated with 10 mM phosphate-buffered

saline with pH = 7.4 that also contained 150 mM sodium chloride. The same buffer was also

used as eluent. It should be pointed out that in order to have a separation of the elute protein with

minimal sample dilution, the dimensions of the column separation, which was provided in the

amine-reactive protein labeling kit, were approximately 20 cm in length and 0.5 cm in diameter.

The final concentration and degree of labeling of the purified anti-H-IgG-QSY was

determined spectrophotometrically by measuring the absorbance of the protein at 280 nm and the

absorbance of the QSY-7 at 560 nm using a sample dilution factor of 10 times of the original

purified fraction. The protein concentration was calculated by the Beer-Lambert law using a

correction factor of 0.22 for the absorbance of the QSY-7 at 280 nm,302 and a molar extinction

coefficient (CigG) for the IgG of 210,000 M-1 cm-1, in the Equations 5-1 and 5.2.302

Aconjugated protein = A280 A560 (0.22) (5-1)
Aconjugated-protein = IgG b CgG (5-2)
The degree of labeling in the protein was also determined by using the Equation 5-3.

[QSY 7] A56 XgG (M- cm 1)
(5-3)
[IgG] Aconjugated.protein 'QS- 7(M 1. cm 1)
It was found a degree of labeling of 4.6 by using 90,000 M-1 cm-1 as the molar extinction

coefficient (cSY-7) for the QSY-7 quencher molecule.302









Transport Experiments

The bare silica nanotube membranes, M-IgG- or H-IgG-modified membranes were

sandwiched between two pieces of Scotch tape that had 0.31 cm2 area holes punched through

them. These holes defined the area of membrane exposed to the contacting solution phases.

This membrane assembly was mounted between the two halves of a U-tube permeation cell, with

half-cell volumes of 10 mL.

The feed solution contained the quencher conjugated antibody (anti-H-IgG-QSY)

dissolved in 10 mM phosphate-buffered saline pH 7.4 that also contained 150 mM sodium

chloride and 5 mM sodium azide (buffer used for all studies). The receiver solution on the other

side of the membrane contained the buffered solution of the fluorescence emitter labeled protein

the H-IgG-FITC. The concentration of the transporting protein in the feed solution was in all

cases higher than the emitter-donor-protein in the receiver side. Thus, the chemical potential was

preferentially preserved from the feed to receiver solutions as the actual driving force for the

flux. A protocol was established to carry out all the transport experiments.

First, the membrane mounted in the U-tube permeation cell was equilibrate for one hour in

the buffer solution, and the receiver solution were frequently measured using a flow-through

fluorescence detector until a stable base line for the fluorescence signal was reached. After that, a

known concentration of anti-H-IgG-QSY was placed in the feed solution, and it was permitted to

equilibrate for another fifteen minutes. Then, a known concentration of H-IgG-FITC was put in

the receiver solution, after five minutes when the fluorescence signal was stable the experimental

data for the transport experiment started being recorded.

The changes of fluorescence intensity in the receiver solution were periodically measured

using a flow-through fluorescence detector (Figure 5-1).














", ------------ piB




Figure 5-1.Flow-through fluorescence system used to measure the changes of fluorescence
intensity in the receiver solution.

This system consisted of an L-7485 fluorescence detector (Hitachi). The flow stream was

delivered using an L-7100 HPLC pump (Hitachi). We discovered that if the receiver solution

was sent continuously through the fluorescence detectors during the entire course of a transport

experiment, significant photobleaching of the FITC dye occurred. To prevent this, we turned the

pump on to deliver receiver solution to the detectors for only 3 minutes per hour of transport

experiment. That is, we used a duty cycle of 57 minutes with pump off followed by 3 minutes

with pump on (flow rate = 3 mL per min) to collect fluorescence data. The whole system was

controlled using Chromeleon 6.50 chromatography management software (Dionex) through a

UCI-100 universal chromatography interface (Dionex).

FITC was excited at 495 nm, and the emission was detected at 515 nm. Calibration curve

was obtained with the same flow through system on solutions containing known concentrations

of the flourophore-labeled antibody. The calibration curve was used to obtain the correct

concentration of the unquenched fluorescent labeled protein in solution.

Results and Discussion

Experimental Principle Used for the Fluorescence Quenching Detection

For these studies, the transport analysis design of the anti-H-IgG-QSY across silica

nanotubes was mainly based on the FRET process. Therefore, the analysis required the use of a










fluorescence donor (FITC) and fluorescence acceptor (QSY-7) molecules, which their structures

are shown in Figure 5-2.


CH CH3 HO 0O 0
Q"C3 .+ o
N O N C1

0 C -OH
SSO2_-N -O0-N | 0

0 N=C=S

a) QSY-7 b) FITC

Figure 5-2.Structures of (a) QSY-7 and (b) FITC

One of the basic conditions for the FRET process to occur is that the emission spectrum of

the donor must overlap the absorption spectrum of the acceptor. Figure 5-3 illustrates both the

excitation and emission spectra for FITC and the absorption spectrum for QSY-7.



S FITC / \QSY-7
Excitation Spectrum /' / \ Absorption Spectrum

/ / \
,, -I.-. \
(- .'s 'c- Spe"-';.;T 0 "
/i
s- --- "
400 440 480 520 560 600 640
Wavelength (nm)
Figure 5-3.Spectra for FITC and QSY-7

It is clearly seen how the emission spectrum of the FITC fluorophore overlaps significantly

the absorption spectrum of the QSY-7 dye. It demonstrated that these selected pair of donor and

acceptor molecules fulfills the overlapping condition for the FRET process required for the

studies described here. The QSY-7 quencher is a nonfluorescent acceptor fashioned to quench

specifically the fluorescence within the range of 500 to 600 nm.303 The FRET efficiency is a

distance-dependent interaction process, and the FRET efficiency is related to the Forster radius









(Ro), and this radius corresponds to the distance at which fluorescence resonance energy transfer

from the donor dye to the acceptor dye is 50% efficient (Equation 5-4).131,298

E = + 6 -> r = distance between fluorophore and quencher (5-4)

Ro
The magnitude of Ro is dependent on the spectral properties of the donor and acceptor dyes

and the typical value reported for the FITC and QSY-7 pair is 61 A.303

The suitable proximity for the fluorescence quenching process between these fluorescence

donor and acceptor molecules was accomplished through a molecular recognition reaction using

two antibodies with specific affinity for each other. This antigen-antibody immunoassay is

illustrated in Figure 5-4.






10'(D
0 P + i4


tWH40g G4Y H4gGTC Compl.x An n.nAnttbody
Figure 5-4.Molecular recognition reaction used to promote the fluorescence quenching.

Thus, the flux of anti-H-IgG-QSY was monitored by using a known concentration of H-

IgG-FITC as fluorescence emitter in the permeation compartment.

Verification of the Fluorescence Quenching System via H-IgG-FITC and anti-H-IgG-QSY
Reaction in Solid State

In order to verify that the molecular recognition reaction between H-IgG-FITC and anti-H-

IgG-QSY leads to an inhibition of the fluorescence process, a silica nanotube membrane was

modified with H-IgG-FITC. The fluorescence signal for this modified membrane was observed

under microscope before and after six hours of exposure to a solution of anti-H-IgG-QSY that











was 100 nM in protein. The results of the fluorescence images and fluorescence spectra of the H-

IgG-FITC-modified membrane are shown in Figure 5-5.
















50C D
4000 20
8000
2 10
S2000
0
1000
0ooo 10


400 500 600 700 800 900 1000 400 500 600 700 800 900 1000
Wavelength (nm) Wavelength [nm)
Figure 5-5.Fluorescence image and spectra respectively for a H-IgG-FITC-modified membrane,
(A-C) before exposure, and (B-D) after exposure to a sample solution of anti-H-IgG-
QSY.

It was observed that the fluorescence intensity of the H-IgG-FITC-modified membrane is


completed inhibited through the immunoassay with anti-H-IgG-QSY. These results confirmed


that the fluorescence signal of H-IgG-FITC can be quenched by anti-H-IgG-QSY by taking


advantage of the antibody-antigen molecular recognition reaction, which situates the


fluorescence donor and acceptor molecules within the intimate distance required for the


quenching process.


Control Transport Studies

In order to demonstrate that the transport of proteins across silica nanotube membranes can


be measured using fluorescence quenching detection through a biomolecular recognition reaction











between a fluorescence donor fluorescent- labeled antibody and a fluorescence acceptor

quencher- conjugated antibody, two different types of controls were carried out.

In the first control experiment (control 1), the fluorescence of a 3.35 nM solution of R-

IgG-FITC was measured in the buffered receiver solution, whereas a 33.3 nM solution of the

quencher-conjugated protein (anti-H-IgG-QSY) was placed in the buffered feed solution.

For the second control (control 2), the fluorescence of a 3.35 nM solution of H-IgG-FITC

was measured in the buffered receiver solution, but in this case, the protein place in the buffered

feed solution was the anti-H-IgG without QSY-7 conjugation. The same concentration of protein,

a 33.3 nM of anti-H-IgG was also used in the buffered feed solution during the control 2. The

results for these control experiments are shown in the Figure 5-6.

4.0
180 A B
A -r3.5
S160 2 A

140 3.0 -
M A- A
AA 0
Tm A T A
0 120 A A 2.5 A:,
U. A r- A
t A o r A
100 -)
100 sA tr 2.0 AAA a o b
S 80 .. '. M. b
1.5 a
60 I
0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800
Time (min) Time (min)
Figure 5-6.Typical plots for (A) changes of fluorescence signal, and (B) concentration of
unquenched protein in function of time for two different control transport/detection
systems: (a) anti-H-IgG/H-IgG-FITC, and (b) anti-H-IgG-QSY/R-IgG-FITC.

Figure 5-6A shows the changes of fluorescence signal in the receiver solution during the

course of the transport experiments both for control 1 and control 2. It was observed that for both

controls the fluorescence signal is reduced as the time of the transport experiment increases.

Since any of these two controls comply with the conditions required for FRET, this inhibition in

fluorescence signal was attributed mostly to two factors different of FRET process. One factor









was endorsed to photobleaching of the fluorophore, and the second to a partial diffusion of the

FITC-labeled proteins toward the direction of the feed solution.

By assuming that during the time of the transport experiment the degree of labeling (DL)

remains constant for the fluorescently protein, calibration data allowed us to convert these

fluorescence intensities to molar concentration of unquenched protein in the receiver solution

(Figure 5-6B). The observed variations in scale for the plots in Figure 6-6A and Figure 6-6B are

due to the difference in the DL between H-IgG-FITC and R-IgG-FITC proteins (sensitivity). The

DL for H-IgG-FITC was 4.2, whereas it was 4.7 for R-IgG-FITC.

The plots in Figure 6-6B give the impression that the deactivation of the fluorescence

intensity for these two controls was carried out via two different mechanisms. In the case of the

anti-H-IgG-QSY/R-IgG-FITC pair, it was observed that during the initial hour of the experiment

the rate of fluorescence quenching for this system was higher than for the anti-H-IgG/H-IgG-

FITC pair. Although, there is not a specific affinity for these two proteins (control 1), this effect

might be assigned to the possibility that the donor (FITC) and acceptor (QSY-7) fluorescence

label moieties in the proteins collided with each other to deactivate the fluorescence intensity

faster than the other control system, for which the presence of QSY-7 quencher is absence. The

anti-H-IgG-QSY/R-IgG-FITC pair (control 1) was chosen as the control for the transport

experiments performed through the specific molecular recognition reaction proposes in this

approach.

Effect of the Feed Solution Concentration on the Rate of Quenching

The experimental system demonstrated here, relies on the strength of the binding affinity

and interaction between the labeled antibodies to bring together the fluorophore and quencher to

an appropriated distance in order that the fluorescence quenching occurs. To study the flux

across silica nanotube membranes using this fluorescence quenching-mediated detection, we










carried out experiments to determine the effect of the feed solution concentration on the rate of

fluorescence quenching.

For these experiments, the concentration of H-IgG-FITC in the receiver solution was 3.35

nM in all cases, whereas three different concentrations for anti-H-IgG-QSY were used in the

feed solution. The three concentrations of anti-H-IgG-QSY were respectively, 16.6 nM, 33.3

nM, and 66.6 nM. These results are shown in Figure 5-7.

3.5
180- A B

S160 3.0 -
a) 0 V ~ A '7
S140 A
a vv I 2.5 7 v,
g 12 0 A v v7 A e 7 77
z I AA
1 0 0 7 7* A v 2 .0 7 A 7
L 0 .:*, 77 v d C .. "*A .. 7 ,aV V d


< 60 b 0
S1.0
40 I
0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800
Time (min) Time (min)
Figure 5-7.Typical plots for (A) changes of fluorescence signal and (B) concentration of
unquenched protein in function of time for three different feed solution
concentrations: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control).

Figure 5-7 shows the influence of feed solution concentration on the rate of fluorescence

quenching in the receiver solution. As can be observed from Figure 6-7, transport experiments

with higher concentration of anti-H-IgG-QSY yields to a faster rate of the fluorescence

quenching. These results confirmed that the molecular recognition reaction between the labeled

antibodies generated the appropriate distance between the donor (FITC) and acceptor (QSY-7)

molecules necessary for this fluorescence quenching-mediated detection. This was also verified

by comparing this effect with the control experiment (plot d in Figure 5-7). The significant

differences in fluorescence intensity for the experiments shown in Figure 5-7were observed only

during the initial four hours of the transport experiments. After this period of time, the changes in









fluorescence signal were less evident. This behavior can be rationalized in that only a small

fraction of the labeled proteins fulfilled an appropriate orientation of the dipole-dipole interaction

between the donor-acceptor molecules for FRET. However, the possibility of contact quenching

should be considered as well.

Data Analysis Based on the Stern-Volmer Plots

The role of fluorescence quenching can be experimentally studied by determining the

quenching rate parameters using the Stem-Volmer plots that are drawn in accordance with the

Stern-Volmer equation (Equation 5-5)131


S=I+Ks[Q] (5-5)
I
From the experimental data of this investigation, the terms Io and I in Equation 5-5 are

known, and they are respectively the fluorescence intensity in the absence and presence of the

quencher-conjugated protein. However, the concentration of the quencher-conjugated protein

(anti-H-IgG-QSY) is not absolutely defined. Therefore, some assumptions were made in order to

quantitatively describe the fluorescence quenching process in terms of the Stern-Volmer

equation. By assuming that during the time of the transport experiment the degree of labeling

(DL) remains constant for the fluorescently protein, the fluorescence intensities were converted

to molar concentration of unquenched protein via calibration curve. It was also assumed that the

reaction between anti-H-IgG-QSY and H-IgG-FITC during the quenching phenomenon was 1:1,

since the molar concentration of unquenched protein was already established. Besides, it was

assumed that the differences in molar concentration of unquenched protein with time represent

the concentration of the quencher-conjugated protein (anti-H-IgG-QSY) that was able to have the

right conditions to interact with the donor-labeled protein (H-IgG-FITC) to yield the quenching










process. Thus, the Io/I ratio was associated with the assumed concentration of anti-H-IgG-QSY

in the Stern-Volmer plot (Figure 5-8).

1.25 C
b
a
1.20 -- d


1.15 .

1.10 --^--


1.05


1.00 ant,-H-lgG- SY (nMI
0.0 0.1 0.2 0.3 0.4 0.5 0.6
anti-H-IgG-QSY (nM)
Figure 5-8.Stern-Volmer plots of Io/I against concentration of anti-H-IgG-QSY in the receiver
solution transported across silica nanotube membranes. The effect of the feed solution
concentrations of anti-H-IgG-QSY is shown: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM,
and (d) 33.3 nM (control).

Figure 5-8 shows the linear segment of the Stem-Volmer plots from the data compiled in

Figures 5-7A and 5-7B. The linear part of the data analysis was used with the purpose of

comparing the slopes of these Stem-Volmer plots (Ksv) as a function of the feed solution

concentration of anti-H-IgG-QSY (Table 5-1). The insert in Figure 5-8 displays the complete

range of the data analysis. It should be mentioned that those plots were not linear, and a positive

deviation in all of them was observed, which is clearly noticed in the insert in Figure 5-8. The no

linearity of the plots suggested that the observed fluorescence quenching is not merely

collisional; thereby the contribution of contact quenching should be also considered.

It was found that the rate of fluorescence quenching depends on the flux of anti-H-IgG-

QSY across the nanotube membrane, and that the rate of fluorescence quenching can be used to

determine the concentration of the analyzed protein on the feed solution. Figure 5-9 and Table 5-

1 summarize these results.










The slopes (Ksv) calculated from the linear segments of the Stern-Volmer plots were

compared on function of the feed solution concentration of anti-H-IgG-QSY (Table 5-1). The

values obtained from the slopes (Ksv) confirmed that higher fluxes lead to faster rates of

quenching.



0.60 /

0.55

0.50

> 0.45

0.40 -

0.35
0 1 I i I I I ,
0 10 20 30 40 50 60 70
Feed Concentration of anti-H-lgG-QSY (nM)
Figure 5-9.Effect of the feed concentration on the Stern-Volmer quenching constant Ksv.

A linear relationship was found when the Stern-Volmer quenching constant (Ksv) was

plotted versus the concentration of anti-H-IgG-QSY in the feed solution; this plot is shown in

Figure 5-9. These data demonstrated that the linear relationship between Ksv and feed solution

concentration can be used to determine the concentration of the analyzed protein on the feed

solution.

Observing the values of Ksv, an additional conclusion that might be inferred, is that they

can be associated with the binding affinity (nM-1) for the permeating protein to the fluorescent

labeled antibody in the receiver solution. It should be pointed out that Ksv is associated with the

rate constant of the dynamic reaction between quencher and fluorophore. Therefore, this

analytical approach proposed in these studies might also be used to estimate binding affinities of

antibodies in solution.










Table 5-1.Relationship between the Stern-Volmer quenching constant (Ksv) with the feed
solution concentration of anti-H-IgG-QSY
Feed Concentration of anti-H-IgG-QSY (nM) Ksv (nM'1)
0 (control) 0.369
16.6 0.431
33.3 0.491
66.6 0.621



Data Analysis Based on log-log Plots of Fluorescence vs. Time

A second and much simpler analysis for the obtained data was based on the logarithmic

plots of the changes of fluorescence in function of time. The results for this analysis are shown in

Figure 5-10.


2.2 -

2.1 A -v-V


19 2.0 .

v, 1.9 d
*^ **.ee 4 a

1.8 b
1.7 E C


1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
log (time)
Figure 5-10.log-log plots of the of fluorescence signal vs. time showing the effect of feed
solution concentration of anti-H-IgG-QSY: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM,
and (d) 33.3 nM (control).

The respective log-log plots of the Figure 5-7A is shown in Figure 5-10. After the

experimental data was represented in a logarithmic scale the relation between fluorescence and

time was almost linear, which confirmed that the relationship between these two physical

variables is described by a power law formula. It was observed that as the concentration of anti-

H-IgG-QSY in feed solution increases the slope of the log-log plot tends toward more negative

values as a result of faster rate of quenching. The trend observed with this simple analysis (log-










log plots) confirmed also that transport experiments with higher concentration of anti-H-IgG-

QSY yields to a faster rate of the fluorescence quenching. These results are summarized in Table

5-2.

Table 5-2.Relationship between slope of the log-log plots for the fluorescence signal vs. time
with the feed solution concentration of anti-H-IgG-QSY.
Feed Concentration of
danti-H-IgGC SY (nM) Slope of log-log plots Intercept of log-log plots
anti-H-IgG-QSY (nM)
0 (control) -0.1198 2.4360
16.6 -0.1374 2.3406
33.3 -0.1693 2.3488
66.6 -0.2095 2.3344


The relation between the slopes obtained from the log-log plots with the feed concentration

is shown in Figure 5-11.


-0.12


"- -0.14 -


-0.16 -

-0.18 -


m -0.20

0 10 20 30 40 50 60 70
Feed Cocentration of anti-H-IgG-QSY (nM)
Figure 5-11.Linear relationship between slopes obtained from the log-log plots with the feed
concentration of anti-H-IgG-QSY.

A linear relationship was also obtained between the slopes of the log-log plots with the

feed concentration of anti-H-IgG-QSY. It showed excellent linearity (correlation coefficient =

0.996) over the range from 0 nM to 66.6 nM. Therefore, by this simple data analysis using log-

log plots, it is also possible to determine the concentration of the analyzed protein on the feed

solution by simply knowing the slope of the log-log plot.










Transport across Antibody-Modified Silica Nanotube Membranes

Transport experiments were also carried out across antibody-functionalized nanotube

membranes. For these experiments silica nanotube membranes were modified both with H-IgG

or M-IgG. The results for these experiments are compared in Figure 5-12.

160 3.5
o A o B
0 0 0 A



P 0 2.5
140 A 0
S120 0 a

o (M A
S120 A 0 A
A 0 AAA



00ncentration ws3.nf t 2
< 80 O00ooooooooob ooooob
I I 1 1.5 '
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Time (min) Time (min)
Figure 5-12.Typical plots for (A) changes of fluorescence signal and (B) concentration of
unquenched protein in function of time illustrating the transport of anti-H-IgG-QSY
across (a) H-IgG-, and (b) M-IgG-modified membranes. The feed solution
concentration was 33.3 nM for both experiments.

The data analysis for the transport across antibody-modified membranes based on Stern-

Volmer plots and log-log plots are summarized in Table 5-3.

Table 5-3.Data for the Stem-Volmer quenching constant (Ksv) and slope of the log-log plots for
antibody-modified membranes
Antibody-Modified Membrane* Ksv (nM-') Slope of log-log plots
H-IgG 0.382 -0.0995
M-IgG 0.415 -0.1218
*The feed solution concentration of anti-H-IgG-QSY was 33.3 for both experiments.

The results shown that the fluorescence quenching in the receiver solution was higher for

the transport experiments carried out across M-IgG modified than H-IgG modified membranes.

This is due to the faster rate of transport of anti-H-IgG-QSY across M-IgG as a result of the

lower binding constant between these two proteins. Therefore, it demonstrated that antibody-

modified membranes can be use to modulated the rate of fluorescence quenching for any specific









pair of quencher-fluorophore labeled antibodies. These results confirmed that the observed

fluorescence quenching phenomenon is a result of the biomolecular recognition reaction between

anti-H-IgG-QSY and H-IgG-FITC in the receiver solution.

Conclusions

It was described a new method to study transport of proteins across silica nanotube

membranes by using a fluorescence quenching-based sensing approach. This analysis was

accomplished by simply observing the fluorescence quenching of a donor-fluorescent labeled

protein (H-IgG-FITC) in the receiver solution after immunoassay reaction with the transported

acceptor-fluorescent conjugated protein (anti-H-IgG-QSY). The flux of anti-Human IgG

conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by observing the

changes in fluorescence of a known concentration of Human IgG FITC-labeled (H-IgG-FITC) in

the receiver solution. The rate of fluorescence quenching was studied both in the absence and the

presence of quencher labeled protein transported across of bare silica nanotube membranes and

antibody-modified membranes. It was demonstrated that the rate of fluorescence quenching

depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of

fluorescence quenching can be used to determine the concentration of the analyzed protein on the

feed solution.









CHAPTER 6
HYBRIDIZATION STUDIES OF DNA IN A PNA-MODIFIED NANOTUBE MEMBRANE
PLATFORM

Introduction

The development of DNA-sensors devices have attracted substantial research efforts

directed to gene analysis, detection of genetic disorders, analysis of single nucleotide

polymorphism in genes, tissue matching and forensic applications among others.304- 307 Various

techniques, including quartz crystal microbalance,308 electrochemical,309 fluorescence,310

electrophoretic,311 electrochemiluminescence,312 enzymatic,313 surface plasmon resonance,314 and

atomic force microscopy,315 have been developed to detect DNA hybridization. However, the

label-free detection methods have attracted more interest than other methods because they can be

simple, rapid, highly sensitive, and inexpensive.307

In 1953, Watson and Crick proposed a model for the structure of DNA.316 This model

predicted that DNA would exist as a helix of two complementary anti-parallel strands.

Soon after the first description of the double helix by Watson and Crick, it was shown, that

the two strands could be separated by heat or treatment with alkali. The reverse process, which

underlies all the methods based on DNA renaturation or molecular hybridization, was first

described by Marmur and Doty.317 It was quickly established that the two sequences involved in

duplex formation must have some degree of sequence complementarity and that the stability of

the duplex formed depends on the extent of complementarity. These remarkable properties

suggested ways to analyze relationships between nucleic acid sequences, and analytical methods

based on molecular hybridization. Since then, these methods have been rapidly developed and

applied to a variety of biological malfunctions.

Base-pairing (A-T and G-C) or hybridization is the underlying principle of DNA-sensors,

namely the strong interaction between two complementary strands. Accordingly, such devices









rely on the immobilization of a single-stranded DNA molecule (the "probe") for hybridizing with

the complementary ("target") strand in a given sample.

Then, the specificity of DNA hybridization devices depends primarily on the selection of

the probe (sequence and length) and secondary upon the hybridization conditions (mainly the

temperature, and ionic strength as well).

Peptide nucleotide acid (PNA) was introduced in 1991 as a structural mimic of DNA with

the ability to invade double helical DNA and bind its complementary target sequence.318 Due to

its higher specificity in the recognition of complementary DNA; PNA has already established

itself as an attractive recognition layer for DNA sensing.319





o G





o C-o
0
(a) PNA "U (b) DNA 0o-6
Figure 6-1.Structures of (a) PNA, and (b) DNA.

PNA is a DNA analogue with a neutral N-(2-aminoethyl) glycine backbone replacing the

negatively charged phosphate backbone of DNA (Figure 6-1).318 Since the spacing between the

nucleotides is the same as in DNA, the conventional Watson-Crick base pairing rules apply

between mixed PNA/DNA sequences. PNA highly discriminates mismatch DNA and has a

stronger binding affinity for complementary DNA than its DNA counterpart. Then, the favorable

hybridization properties of PNA have also inspired the incorporation of PNA into array and

nucleic acid biosensor formats.32- 323









This chapter describes the use of PNA-modified nanotube membranes to study a label-free

detection for DNA hybridization. This was achieved by exploiting the unique structural and

hybridization features of PNA. This approach was characterized by observing the changes in the

membrane resistance using impedance spectroscopy. This method allowed carrying out an in situ

characterization of the interfacial properties of the modified nanotube membranes during the

hybridization process in the duplex formation PNA-DNA. The novel idea was based on the fact

that PNA is a neutral molecule; thereby during hybridization events achieved in just highly

purified water without any electrolyte in the solution, the only source of new charges in the

system during the hybridization process emerge from the DNA molecules. This event exhibited a

change of the membrane resistant that was detected by impedance spectroscopy which is an

effective technique to probe the interfacial properties of modified electrodes.

Experimental

Materials

Commercially available nanopore alumina membranes (60 rm thick, nominal pore

diameter 100 nm and 200 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999%

(TEOS) (Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these

membranes. A trimethoxysilyl propyl diethylenetriamine (DETA) was used to introduce amino

groups to the nanotubes was obtained from United Chemical Technologies. Sulfosuccinimidyl-

6-(biotinamido) hexanoate (NHS-LC-Biotin) obtained from Pierce was used as biotin-

streptavidin linker. A purified streptavidin was obtained from Sigma Aldrich. A biotinylated-

PNA (PNA) with twelve nucleotides bases (Biotin- GCG CTA GCC TAC) obtained from

Applied Biosystems was used as the probe to detect DNA. To investigate the hybridization

properties of the PNA-modified nanotube membranes a complementary twelve bases DNA (3'-

CGC GAT CGG ATG -5') with a MW=3607 g/moles was used. This complementary target









DNA was obtained from Alpha-DNA. All other chemicals were of reagent grade and used as

received. Purified water, obtained by passing house-distilled water through a Barnstead, E-pure

water-purification system, was used to prepare all solutions.

Preparation of the Silica Nanotube Membranes

Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-

based template-synthesis method. Briefly, a sol-gel precursor solution was prepared by mixing

absolute ethanol, TEOS and water (40:5:10 by volume), and this solution was allowed to

hydrolyze for 1 hour. The alumina membranes were then immersed into this solution for two

minutes under sonication, and they were kept in this solution for ten additional minutes. The

membranes were then removed and the surfaces swabbed with ethanol to remove precursor from

the alumina surface. The membranes were then air dry for ten minutes at room temperature and

cured in an oven for 12 hours at 150 oC.

Preparation of PNA-Modified Nanotube Membranes

The biotinylated-PNA probe was immobilized on nanotube membrane surfaces using the

streptavidin-biotin affinity system.297 324 Streptavidin is a 60-kDa protein purified from the

bacterium Streptomyces avidini. The interest in streptavidin is the remarkable and unique

interaction with its low-molecular weight ligand biotin (Kd_10-15 M). Because of its extremely

strong affinity, the streptavidin-biotin affinity system has been developed as a general tool to use

in various experimental systems.325

The immobilization process used in this approach consisted of three basic steps. The first

and second steps were respectively, a silanization which introduced amino groups to the surface,

and conjugation between the amino groups and the NHS-LC-Biotin. The latter acylation reaction

of an N-Hydroxysuccinimidyl (NHS) ester of biotin to the primary amino groups of the silane-

modified surface yields to a stable amide bond. The next step was the coupling of streptavidin to










the biotin-modified surface. It should be pointed out that the NHS-LC-Biotin had a spacer arm of

22.4 A that can reduce steric hindrance associated with the binding many biotinylated molecules

on one streptavidin, which in turns may enhance the sensitivity of the system.


SIu C LC
S NHS LC-Biotin Streptavidin




o o/ DN
ALUSAN MEMBRANE ALM 9'"






P -


DETA: a T
Sulfosuccinimidyl-6-(biotinamido) hexanoate BRAN








membranes.
1 (v/v) solution of DETA in 1 mM acetic acid for 30 minutes at room temperature. The DETA-
OH OH H
0 LiC I




DETA: -oThe a- mio--si
Trimethoxysilylpropydiethylenetriamin e d u r non ad fd at 10 -ov o
NHS-LC-Biotin: .0 a_
Sulfosuccinimidyl-6-(biotinamido) hexanoate L. LLM iioUIE,
Figure 6-2.Immobilization of biotinylated-PNA on templated synthesized silica nanotube
membranes.

During silanization the silica nanotube membranes were immersed into freshly prepared

1% (v/v) solution of DETA in 1 mM acetic acid for 30 minutes at room temperature. The DETA-

modified membranes were rinsed with water to remove the unreacted DETA. The amino-silane

modified nanotube membranes were dried under nitrogen and fixed at 120 C in the oven for 15

minutes. For the biotinylation process the amino-silane modified membranes were exposed for 6

hours to a 0.5 mg/mL solution of NHS-LC-Biotin prepared in 0.1 M bicarbonate buffer pH 8.5.

The biotinylated nanotube membranes were then incubated overnight at 4 OC in a 1 mg/mL









solution of streptavidin prepared in 10 mM phosphate buffer pH 7.4. The PNA immobilization

was carried out by incubating the streptavidin-modified membranes for 8 hours in a solution that

contained 1 iM ofbiotinylated PNA dissolved in water at 4 OC. A scheme showing the

immobilization steps for the PNA-modified nanotube membranes and the hybridization reaction

with the complementary DNA is shown in Figure 6-2.

Nanotube Membrane Assembly

After the silanization with DETA, the silica nanotube membranes were assembled in a

suitable format for their characterization with impedance spectroscopy.

In order to make the nanotube membrane conductive, they were sputter-coated for 90 sec

on both sides with Gold-Palladium at 75 mTorr of system pressure, 45 mA of current, conditions

that gave a coating of approximately 10 nm thicknesses. The metal sputtered on the membranes

enhanced the conduction properties of the membrane, and provided the facilities for the

characterization of the membranes.


-q in (i' Coefrtli ka--






-//

flumina 9dtrrufrane
sputtfr-coatldwitk )Au- d

Figure 6-3.Nanotube membrane assembly.

The sputtered membranes together with two thin platinum leads of 60 microns thickness

(one in each opposite side of the membrane) were placed between two pieces of plastic tape

punched with a hole of 0.31 cm2 area. The two platinum leads were sandwiched between the two

pieces of tape in such a way that the holes were aligned and the platinum leads were covered









with the tape to avoid their contact with the solution during the impedance measurements. A

schematic illustration for this assembly is shown in Figure 6-3.

These holes in the assembled nanotube membranes defined the area of the nanotube

membrane exposed for the subsequent modification steps and the area exposed during the

impedance measurements.

It is worth noticing that the external electrical connections involves dry metal-to-metal

contacts, so that, the membrane itself performs the function of the electrodes (working and

reference).

Impedance Measurement for Membrane Characterization

For the impedance measurements the assembled membrane was mounted between the two

halves of a U-tube permeation cell. This sputter-coated nanotube membrane performed and

replaced the function of the typically used remote electrodes located on either side of the

membrane. For all the impedance experiments both half-cells were filled with 10 mL of highly

purified water. During the hybridization experiments one of the half-cell was chosen to spike the

concentration of DNA target, this side also belonged to the side where the reference electrode

connections were placed.

The impedance data was obtained using a Solartron 1287 electrochemical interface module

connected to a Solartron 1255B frequency response analyzer and a personal computer. The

impedance data was recorded applying 5 mV alternative voltage in the frequency range from 1

MHz to 0.05 Hz. Z-plot and Z-view software packages (Scribner Asc. Inc., NC) were used to

control the impedance experiments and analyzed the impedance data, respectively.

Impedance Data Analysis

The main interest of in these studies was to monitor the changes in the membrane

resistance during the hybridization process in the duplex formation PNA-DNA. Even though,









that the experimental impedance data revealed the contribution of Warburg impedance, in order

to simplify the data analysis this contribution was not considered. In addition, since this represent

a bulk property of the electrolyte solution (diffusion of ions from the bulk electrolyte); it is not

affected for the chemistry modification occurring at the surface of the pores in the membranes.

Therefore, the experimental data of impedance were analyzed by using the model equivalent

circuit shown in Figure 6-4.

CPE



Rm
Figure 6-4.General equivalent circuit model used for impedance data analysis.

A scan be seen from Figure 6-4, the data were fitted using a constant phase element (CPE).

The reason of this is because in impedance experiments the capacitors often do not behave

ideally.326 This CPE is defined by the Equation 6-1.

Z(CPE) = A(jo) (6-1)
Where is an imaginary number, co is the angular frequency, and A is a constant. When

this equation describes a capacitor, the constant A is equal to the inverse of the capacitance, and

the exponent n is equal to 1. For a CPE the exponent n is less than 1. Thus, by fitting the

experimental data with the equivalent circuit shown in Figure 6-4, the changes in membrane

resistance, Rm, were extracted.

Results and Discussion

In order to evaluate the effects of the various immobilization steps on the surface of the

sputter-coated nanotube membranes, impedance measurements were carried out during the all

steps of the immobilization process. All the assays were conducted in highly purified water. A

biotinylated PNA probe with a sequence Biotin- GCG CTA GCC TAC, which is fully










complementary to the DNA target with a sequence 3'- CGC GAT CGG ATG -5', was used

as receptor in the nanotube membrane sensing-platform. These studies begin, however, with the

impedance measurements for the alumina template, and silica nanotube membranes and

investigations of the equilibration times required for the impedance measurements.

Impedance Characterization of the Unmodified Membranes

The Nyquist plots for bare alumina template (200 nm nominal pore size) and silica

nanotube membranes are shown in Figure 6-5. The x-axis and y-axis in this plot corresponds to

real impedance, Zre, and imaginary impedance, Zim, respectively. Zre and Zim are mainly

originated from the resistance and capacitance of the cell system, respectively.

-40000


-30000



N -20000 b -
Sa

-10000


0
0 10000 20000 30000 40000
Z'
Figure 6-5.Nyquist plots (Zim vs. Zre) for impedance measurement of (a) alumina and (b) silica-
nanotube assembled membranes in highly purified water.

As shown in Figure 6-5, changes after the deposition of the silica nanotubes in the alumina

template were observed in semicircle part at high frequency region (usually associated with

kinetic controlled process), whereas the linear tail at lower frequencies (usually associated with

mass transfer process) maintained almost the same slope. The variations in the high frequency

region were probably due to a change in the isoelectric point (pI) of the nanotube membrane.

Since it is known that the pi of the alumina membranes range from 6.9 to 10.5, while the pi for










silica has been ranged from 1.9 to 4.5.327 Because of the different electrical characteristics of

these materials, a change in the surface characteristics of the membranes should be expected

when silica nanotubes are synthesized within the alumina membrane, as was observed in the

Nyquist plots in Figure 6-5.

In order to assess the stability of the impedance measurements a study of the wettability of

the bare alumina membrane was carried out by monitoring the changes in resistance with the

time. These results are summarized in Figure 6-7.



8000

1 6000 ,

r 4000 -

2000


0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Wetting Time (hours)
Figure 6-7.Effect of the wetting time on the resistance impedance measurement for an assembled
unmodified alumina membrane 200 nm pore diameter.

As can be seen from Figure 6-7, the changes in resistance from impedance measurements

of alumina membrane were pretty stable with the time. These results confirmed the stability of

the sputter-coated membranes and the stability itself of the impedance measurements.

Functionalization of the Silica nanotube Membranes

The differences in membrane resistance before and after each modification step during the

immobilization process were analyzed in order to characterize the accomplishment of the various

modifications (Figure 6-8). The obtained values of Rm from the Nyquist plots in Figure 6-8 are

compiled in Table 6-1.









The results revealed that the different immobilization steps introduced significant changes

in the membrane resistance. The most significant change was found in the membrane resistance

associated with the hybridization process of DNA with the PNA-modified nanotube membrane.

This decrease in resistance was associated with the increase of negative surface charge density

due to the duplex formation PNA-DNA on the surface.

-125000






-75000



-- (a)
---- (b)
/ -- (C)
-25000 / (d)
--o- (e)

-- (g)

0 25000 50000 75000
Z'
Figure 6-8. Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water: (a)
alumina, (b) after synthesis of silica-nanotube, (c) after silanization, (d) after
biotinylation, (e) after streptavidin immobilization, (f) after PNA immobilization, and
(g) after DNA hybridization. The plots were recorded applying 5 mV alternative
voltage in the frequency range from 1 MHz to 0.05 Hz.

The increase of Rm after the deposition of silica nanotubes was consistent with the fact that

after deposition smaller effective pore size in the membrane might be reached. Conversely, the

decrease in membrane resistance after silanization was explained through the formation of

protonated amino groups on the surface at the pH of the experiment (highly purified water).

However, part of this amino protonated groups were neutralized by the conjugation reaction with

the biotin linker, which was confirmed for the increase in resistance after the biotinylation. The









coupling of streptavidin also produced a significant change in resistance, since the size of this

protein reduced the effective pore size of the membrane. For the immobilization of the

biotinylated PNA a small decrease in resistance was observed compared with the streptavidin

coupling. This effect was difficult of explain, but it might be due to the fact that PNA oligomers

are purified by reverse phase HPLC in the presence of trifluoroacetic acid. The aqueous solutions

of PNA posses a pH lower than 3. Therefore, at this low pH some residual positive charge in the

structure of PNA might be introduced in the membrane that causes the reduction of the

resistance.

Table 6-1.Variations ofRm at the different steps of modification on a 100 nm pore diameter
alumina template
Modification step Rm (2)
Alumina membrane 8100
Silica nanotube membrane 12500
Amino-modified membrane 1030
Biotinylated-modified membrane 4515
Streptavidin-modified membrane 18000
PNA-modified membrane 14717
DNA hybridization 327



Impedance Sensing of DNA-PNA Hybridization

Figure 6-9 shows the Nyquist plots of the PNA-modified nanotube membrane before and

after 10 minutes of the hybridization process. As can be seen after the addition of 3 nmoles of the

complementary DNA to the half-cell containing highly purified water (side where the reference

electrode connections were also sited) a significant decrease of the semicircle sector of the plot

was observed. This result suggested that the resistance of the PNA-modified nanotube membrane

became lower due to the increase in new negative charge in the system brought to the membrane

surface for the negatively charged phosphate backbone of DNA molecules during the










hybridization process. There was almost no change in the solution resistance (Rs) and Warburg

impedance (Zw), which demonstrated that they were not affected during the hybridization

process on the membrane surface.

-125000

(a)
X (b)





-75000








-25000

-- (a
i n ( b )



0 25000 50000 75000
Z'
Figure 6-9. Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water for a
PNA-modified nanotube membrane: (a) before, and (b) after 10 minutes of DNA
hybridization. The plots were recorded applying 5 mV alternative voltage in the
frequency range from 1 MHz to 0.05 Hz.

Effect of the Pore Size on Rm

The influence of the pore size on the membrane resistance was explored by using two

different sizes of commercially available (Whatman) nanopore alumina membranes having 60

gm thickness, and nominal pore diameters respectively of 200, and 100 nm. To study this effect,

nanotube membranes of the specified pore size were modified with PNA. For the hybridization

impedance experiments, 3 nmoles of the complementary DNA were added to the highly purified

water in both cases (Figure 6-10).


























ALUMINA MEMBRANE 100 nm B

100000
80000
60000
S40000
20000


STREPTAMDIN PNA DNAhybridization
TYPE OF MODIFICATION STEP

Figure 6-10.Effect of the membrane pore size on the membrane resistance: (A) 200 nm, and (B)
100 nm. The resistance values were obtained from the impedance data recorded by
applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz.

Figure 6-10 compares the streptavidin immobilization, PNA immobilization, and DNA

hybridization processes, respectively, by using membranes with two different pore diameters

(100 and 200 nm). The changes in membrane resistance during the immobilization steps for the

100 nm membrane showed the same trend observed for the 200 nm pore diameter membrane

explained above. However, the resistance was for all modification steps higher in the 100 nm

pore diameter membranes, as a result of the smaller pore size. In the case of the PNA

immobilization, the resistance was increased in approximately 280% by using the 100 nm

membrane.

The effect of the pore size during the hybridization process revealed a reduction of 98% in

the resistance of the membrane for the 200 nm-situation (Figure 6-10A), while for the 100 nm-


ALU INA MEMBRANE 200 nm A


16000 -

S10000 -

6 5000

0 E l
STREPTAMDIN PNA DNAhybridization
TYPE OF MODIFICATION STBE











situation the reduction was of 53% (Figure 6-10B). These differences in the resistance during the

hybridization process might be attributed to the fact that membranes with higher resistance

present longer dynamic range. However, the possibility of a higher immobilization of PNA

might be also considered.

Stability and Reproducibility of the PNA-modified Membranes

The stability of the PNA-modified membrane was evaluated by comparison the effect of

DNA hybridization in time with an unmodified control membrane. These results are compared in

Figure 6-11.

16000 45000
A o B
A 40000 B
14000
35000
12000_ _
S0 30000
100000
3 25000
8000
E 20000
6000
6000 15000
4001500 --- -- -
4000 10000

2000 5000
0 L I F I I I I I IF 1 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time after adding DNA (hours) Time after adding DNA (hours)
Figure 6-11. Study of the hybridization process stability with time: (A) unmodified control
alumina membrane, and (B) PNA-modified membrane. The resistance values were
obtained from the impedance data recorded by applying 5 mV alternative voltage in
the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of
complementary DNA.

Figure 6-11 shows the stability of the impedance measurement for a control (Figure 6-

11A), and PNA-modified membrane (Figure 6-11B) after adding 3 nmoles of complementary

DNA to the experimental cell. It is clearly seen that the changes in membrane resistance are

mainly due to the hybridization process (Figure 6-11B). However, some reduction on the

membrane resistance was observed for the control membrane after the addition of 3 nmoles of

DNA. This effect might be attributed to nonspecific interactions on the membrane surface. The










changes in membranes resistance after 15 hours of adding DNA were 27098 ohms for the PNA-

modified membrane, while this change was only 4271 ohms for the control unmodified

membrane. However, after 20 minutes these changes were 25152 and 1806 ohms for the PNA-

modified and unmodified membranes, respectively. These results clearly demonstrated that the

assembled PNA-modified membrane configuration study herein might be used as a good

platform for DNA sensors. An additional evidence of this conclusion can be observed in Figure

6-12 were the effect of DNA concentration on the impedance measurements is illustrated.

-75000

a

-50000
-sooo -




-25000 C




0 25000 50000 75000
Z'
Figure 6-12.Nyquist plots (Zim vs. Zre) of impedance measurement in highly purified water for a
PNA-modified nanotube membrane: (a) before, (b) after adding 3 nmoles of DNA,
and (c) after 6 nmoles of DNA. The plots were recorded applying 5 mV alternative
voltage in the frequency range from 1 MHz to 0.05 Hz.

The reproducibility of the PNA-modified nanotube membranes were also evaluated (Figure

6-13). For these hybridization studies three different membranes modified with PNA were

compared. Figure 6-13 shows the results of membrane resistance for these three PNA-modified

membranes, before and after hybridization with DNA (20 minutes and 3 days). It was observed

that after 20 minutes of adding 3 nmoles of the complementary DNA solution to each of the

PNA-modified membranes, the reduction of the resistance membrane were almost in the same

order of magnitude. The reduction in resistance for membrane 1, 2, and 3 in Figure 6-13 were









54%, 42%, and 53%, respectively. Amazingly was to observe the stability of these membranes

after three days of the duplex formation during the hybridization process.


THREE PNA-MODIFIED NANOTUBE MEMBRANES

60000 54618
50000 41831
40000 35097 25459
E 30000 19440 20435 22126
1 20000 11913 15892
10000


.6 Z6 i- ,




Figure 6-13.Reproducibility study of the PNA-modified nanotube membranes, before and after
the hybridization process. The resistance values were obtained from the impedance
data recorded by applying 5 mV alternative voltage in the frequency range from 1
MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA.

The results presented in these studies demonstrated the good stability and reproducibility

of the PNA-modified nanotube membrane describe here.

Conclusions

It was described the use of PNA-modified nanotube membranes to study a label-free

detection for DNA hybridization. This was achieved by exploiting the unique structural and

hybridization features of PNA. This approach was characterized by observing the changes in the

membrane resistance using impedance spectroscopy. This method allowed carrying out an in situ

characterization of the interfacial properties of the modified nanotube membranes during the

hybridization process in the duplex formation PNA-DNA. The novel idea was based on the fact

that PNA is a neutral molecule; thereby during hybridization events achieved in just highly

purified water without any electrolyte in the solution, the only source of new charges in the

system during the hybridization process emerge from the DNA molecules. This event exhibited a









change of the membrane resistant that was detected by impedance spectroscopy which is an

effective technique to probe the interfacial properties of modified electrodes. These results

showed the feasibility of using PNA-modified nanotube membranes as a sensing-platform for

DNA hybridization methods. This assembled nanotube membrane platform present the special

characteristic that the membrane itself performed the function of the electrodes (working and

reference respectively). It was demonstrated that PNA immobilized on the nanotube membranes

function as a good probe to detect the DNA targets in condition where no electrolyte is present

during the hybridization process. It was also demonstrated the good reproducibility obtained

during the fabrication and detection with this novel sensing-platform.









CHAPTER 7
CONCLUSIONS

The main goal of the research presented in this dissertation has been to develop selective

nanotube membranes based on affinity molecular recognition with interesting applications in

biological analysis. The unique property of biomolecules and metal ions to recognized specific

molecules was exploited to elaborate functional nanotube membranes with potential application

in sensing and separations of proteins and DNA. The template synthesis method was use to

obtain the silica nanotube within the pores of alumina membranes. The walls of these nanotubes

were functionalized via silane chemistry to couple molecular recognition agents onto the

modified surfaces. Chapter 1 introduce the background information for this dissertation including

the template synthesis method, the description of the alumina template membranes, sol-gel and

silane chemistry, the inspiration of living transport in the development of membrane-based

separations, fundamental about proteins, and a synopsis of the instrumental techniques used

during the development and characterization of the applications describe in this dissertation.

In chapter 2, a detailed experimental study on antibody-functionalized nanotube

membranes was carried out, including the effect of surface modification, pore size, feed

concentration, temperature, and ionic strength, both on the transport as on the functional

behavior of the membranes. It was demonstrated that the type of immobilized antibody changes

the functional behavior of the antibody-modified membranes and the transport properties of the

permeating proteins across them. These membranes showed a Langmuirian shape plot

characteristic of a typical behavior of facilitated transport membranes, when the effect of feed

solution concentration on the fluxes of the permeating proteins was investigated. Therefore, the

antibody-functionalized membranes showed here exhibit a type of behavior for facilitated-

transport membranes during protein separations. Important parameters such as the binding









capacities of the antibody-modified membranes and the binding constants for the permeating

proteins to the antibody carriers confirmed that the permeating protein anti-M-IgG binds strongly

to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier,

likewise, the permeating protein anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-

IgG binds weakly to this carrier. It was also demonstrated that by controlling the orientation of

the immobilized antibody using protein G the potency affinity of the antibodies is effectively

preserved. Through these studies was also demonstrated that by selecting a template membrane

with an appropriated pore size diameters these antibody-modified nanotube membranes have the

ability to be used with a combination of both size and affinity selectivity for protein separation.

Additional studies on the effect of temperature on the fluxes and the binding constant

demonstrated that the functional behavior of these antibody-modified membranes is governed for

the changes in free energy before and after interaction of the permeating proteins with the carrier

antibody, which was attributed to the changes in enthalpy and entropy of the system. Other

important finding during this investigation was that by controlling the ionic strength of the

transport solution, these antibody-modified nanotube membranes can be used for situations were

the transport of an unwanted protein needs to be shut down, in order to improve selectivity. This

latter approach might be useful in the sample complexity reduction in proteomic analysis. Since

the analysis and characterization of complex mixtures of proteins are the central aims of

proteomics. It is also noteworthy to mention the low pM detection limits achieved during these

studies. In general, the ability to alter the surface biochemistries and thus tailor the membrane

for specific analytes, promises the use of these membranes for the analysis of a wide range of

proteins.









Chapter 3 describes the modeling and characterization of the transport properties of

proteins across antibody-functionalized nanotube membranes by using the dual-mode model.

This work demonstrated that the dual-mode model proposes here was a suitable model for the

characterization of the transport properties of anti-H-IgG and anti-M-IgG in M-IgG modified

membranes. The model demonstrated a good agreement with the experimental system studied

here. Besides, the findings showed in these studies allow addressing the concept of modulated

facilitated transport systems using antibody-modified nanotube membranes, and how the extent

of the facilitation is determined and controlled by the binding affinity between carrier and

permeating protein molecules. These studies can provide a new direction in the development of

more effective membranes for membrane-based separation.

In chapter 4, the development of metal ion affinity nanotube membranes was demonstrated

by exploiting the concept of IMAC separations in a nanopore-membrane-platform. Silica

nanotube membranes were functionalized with imidazole through a nucleophilic substitution

reaction between 2-mercaptoimidazol and 3-chloropropyl modified silica nanotubes carried out

in an aprotic medium. It was demonstrated that transport of histidine-tagged recombinant

proteins across these imidazole-modified silica nanotube membranes is controlled by the nature

of immobilized metal ion on the imidazole surface. Besides, the extensive characterization of the

imidazole functionalized membrane allowed the understanding of how coordination chemistry

with different divalent metal ions affects the functional behavior of the membrane and transport

properties during the permeation of six-histidine recombinant proteins. Furthermore, additional

results shown that by the appropriate choice of the transition metal ion for coordination with the

imidazole ligand, the binding strength and optical properties within the functional membrane are

also modified. Hence, these studies revealed how the presence of the immobilized metal ion









affects the functionality of the imidazole membrane and its transport properties. The metal ion

affinity nanotube membrane described here has the potential to be a versatile analytical platform,

where different metal ions can be immobilized on the same prevailing conditions, additionally

offering the possibilities of regeneration without apparent loss of the metal coordination

properties of the functional membrane. These studies demonstrate an alternative way of

immobilized metal ion affinity by using nanotube membranes.

Chapter 5 described a new method to study transport of proteins across silica nanotube

membranes by using a fluorescence quenching-based sensing approach. This analysis was

accomplished by simply observing the fluorescence quenching of a donor-fluorescent labeled

protein (H-IgG-FITC) in the receiver solution after immunoassay reaction with the transported

acceptor-fluorescent conjugated protein (anti-H-IgG-QSY). The flux of anti-Human IgG

conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by observing the

changes in fluorescence of a known concentration of Human IgG FITC-labeled (H-IgG-FITC) in

the receiver solution. The rate of fluorescence quenching was studied both in the absence and the

presence of quencher labeled protein transported across of bare silica nanotube membranes and

antibody-modified membranes. It was demonstrated that the rate of fluorescence quenching

depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of

fluorescence quenching can be used to determine the concentration of the analyzed protein on the

feed solution.

In chapter 6, the use of PNA-modified nanotube membranes to study a label-free detection

for DNA hybridization was demonstrated. This was achieved by exploiting the unique structural

and hybridization features of PNA. This approach was characterized by observing the changes

in the membrane resistance using impedance spectroscopy. This method allowed carrying out an









in situ characterization of the interfacial properties of the modified nanotube membranes during

the hybridization process in the duplex formation PNA-DNA. The novel idea was based on the

fact that PNA is a neutral molecule; thereby during hybridization events achieved in just highly

purified water without any electrolyte in the solution, the only source of new charges in the

system during the hybridization process emerge from the DNA molecules. This event exhibited a

change of the membrane resistant that was detected by impedance spectroscopy which is an

effective technique to probe the interfacial properties of modified electrodes. These results

showed the feasibility of using PNA-modified nanotube membranes as a sensing-platform for

DNA hybridization methods. This assembled nanotube membrane platform present the special

characteristic that the membrane itself performed the function of the electrodes (working and

reference respectively). It was demonstrated that PNA immobilized on the nanotube membranes

function as a good probe to detect the DNA targets in condition where no electrolyte is present

during the hybridization process. It was also demonstrated the good reproducibility obtained

during the fabrication and detection with this novel sensing-platform.









LIST OF REFERENCES

[1] Hagg, M.-B. Separation and Purification Methods 1998, 27, 51-168.

[2] Baker, R. W. Membrane Technology and Applications, 2nd ed.; John Wiley & Sons, Ltd.,
2004.

[3] Striemer, C. C.; Gaborski, T. R.; McGrath, J. L.; Fauchet, P. M. Nature 2007, 445, 749-
753.

[4] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Science 1997, 278, 655-658.

[5] Yu, S.; Lee, S. B.; Martin, C. R. Analytical Chemistry 2003, 75, 1239-1244.

[6] Yu, S.; Lee, S. B.; Kang, M.; Martin, C. R. Nano Letters 2001, 1, 495-498.

[7] Chun, K.-Y.; Stroeve, P. Langmuir 2002, 18, 4653-4658.

[8] Van Reis, R.; Zydney, A. Current Opinion in Biotechnology 2001, 12, 208-211.

[9] Lee, S. B.; Mitchell, D. T.; Trofin, L.; Nevanen, T. K.; Soederlund, H.; Martin, C. R.
Science 2002, 296, 2198-2200.

[10] Kohli, P.; Harrell, C. C.; Cao, Z.; Gasparac, R.; Tan, W.; Martin, C. R. Science 2004,
305, 984-986.

[11] Giorno, L.; De Bartolo, L.; Drioli, E. Membrane Science and Technology Series 2003, 8,
187-217.

[12] Gavalas, V. G.; Wang, J.; Bachas, L. G. Membrane Science and Technology Series 2003,
8, 379-392.

[13] Sexton, L. T.; Home, L. P.; Sherrill, S. A.; Bishop, G. W.; Baker, L. A.; Martin, C. R.
Journal of the American Chemical Society 2007, 129, 13144-13152.

[14] Baker, L. A.; Choi, Y.; Martin, C. R. Current Nanoscience 2006, 2, 243-255.

[15] Choi, Y.; Baker, L. A.; Hillebrenner, H.; Martin, C. R. Physical Chemistry Chemical
Physics 2006, 8, 4976-4988.

[16] Siwy, Z.; Trofin, L.; Kohli, P.; Baker, L. A.; Trautmann, C.; Martin, C. R. Journal of the
American Chemical Society 2005, 127, 5000-5001.

[17] Harrell, C. C.; Choi, Y.; Home, L. P.; Baker, L. A.; Siwy, Z. S.; Martin, C. R. Langmuir
2006, 22, 10837-10843.

[18] Kohli, P.; Wirtz, M.; Martin, C. R. Electroanalysis 2004, 16, 9-18.









[19] Harrell, C. C.; Kohli, P.; Siwy, Z.; Martin, C. R. Journal of the American Chemical
Society 2004, 126, 15646-15647.

[20] Steinle, E. D.; Mitchell, D. T.; Wirtz, M.; Lee, S. B.; Young, V. Y.; Martin, C. R.
Analytical Chemistry 2002, 74, 2416-2422.

[21] Lee, S. B.; Martin, C. R. Journal of the American Chemical Society 2002, 124, 11850-
11851.

[22] Wirtz, M.; Yu, S.; Martin, C. R. Analyst 2002, 127, 871-879.

[23] Martin, C. R.; Nishizawa, M.; Jirage, K.; Kang, M. Journal of Physical Chemistry B
2001, 105, 1925-1934.

[24] Lee, S. B.; Martin, C. R. Chemistry ofMaterials 2001, 13, 3236-3244.

[25] Martin, C. R. Science 1994, 266, 1961-1966.

[26] Martin, C. R.; Mitchell, D. T. Analytical Chemistry 1998, 70, 322A-327A.

[27] Hulteen, J. C.; Martin, C. R. Journal ofMaterials Chemistry 1997, 7, 1075-1087.

[28] Cepak, V. M.; Hulteen, J. C.; Che, G.; Jirage, K. B.; Lakshmi, B. B.; Fisher, E. R.;
Martin, C. R. Journal ofMaterials Research 1998, 13, 3070-3080.

[29] Miller, S. A.; Young, V. Y.; Martin, C. R. Journal of the American Chemical Society
2001, 123, 12335-12342.

[30] Kang, M.; Trofin, L.; Mota, M. O.; Martin, C. R. Analytical Chemistry 2005, 77, 6243-
6249.

[31] Xu, F.; Wharton, J. E.; Martin, C. R. Small 2007, 3, 1718-1722.

[32] Liang, W.; Martin, C. R. Journal of the American Chemical Society 1990, 112, 9666-
9668.

[33] Van Dyke, L. S.; Martin, C. R. Langmuir 1990, 6, 1118-1123.

[34] Yamada, K.; Gasparac, R.; Martin, C. R. Journal of the Electrochemical Society 2004,
151, E14-E19.

[35] Nishizawa, M.; Menon, V. P.; Martin, C. R. Science 1995, 268, 700-702.

[36] Brumlik, C. J.; Menon, V. P.; Martin, C. R. Journal ofMaterials Research 1994, 9, 1174-
1183.









[37] Harrell, C. C.; Lee, S. B.; Martin, C. R. Analytical Chemistry 2003, 75, 6861-6867.

[38] Mitchell, D. T.; Lee, S. B.; Trofin, L.; Li, N.; Nevanen, T. K.; Soederlund, H.; Martin, C.
R. Journal of the American Chemical Society 2002, 124, 11864-11865.

[39] Klein, J. D.; Herrick, R. D., II; Palmer, D.; Sailor, M. J.; Brumlik, C. J.; Martin, C. R.
Chemistry ofMaterials 1993, 5, 902-904.

[40] Hou, S.; Harrell, C. C.; Trofin, L.; Kohli, P.; Martin, C. R. Journal of the American
Chemical Society 2004, 126, 5674-5675.

[41] Hou, S.; Wang, J.; Martin, C. R. Journal of the American Chemical Society 2005, 127,
8586-8587.

[42] Cai, Z.; Lei, J.; Liang, W.; Menon, V.; Martin, C. R. Chemistry ofMaterials 1991, 3,
960-967.

[43] Parthasarathy, R.; Martin, C. R. Nature 1994, 369, 298-301.

[44] Hornyak, G. L.; Patrissi, C. J.; Martin, C. R. Journal of Physical Chemistry B 1997, 101,
1548-1555.

[45] Foss, C. A., Jr.; Hornyak, G. L.; Stockert, J. A.; Martin, C. R. Journal of Physical
Chemistry 1994, 98, 2963-2971.

[46] Scopece, P.; Baker, L. A.; Ugo, P.; Martin, C. R. Nanotechnology 2006, 17, 3951-3956.

[47] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Science 1997, 278, 655-658.

[48] Hulteen, J. C.; Jirage, K. B.; Martin, C. R. Journal of the American Chemical Society
1998, 120, 6603-6604.

[49] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Analytical Chemistry 1999, 71, 4913-4918.

[50] Martin, C. R.; Kohli, P. Nature Reviews Drug Discovery 2003, 2, 29-37.

[51] Gasparac, R.; Kohli, P.; Paulino, M. O. M.; Trofin, L.; Martin, C. R. Nano Letters 2004,
4, 513-516.

[52] Hillebrenner, H.; Buyukserin, F.; Kang, M.; Mota, M. O.; Stewart, J. D.; Martin, C. R.
Journal of the American Chemical Society 2006, 128, 4236-4237.

[53] Hillebrenner, H.; Buyukserin, F.; Stewart, J. D.; Martin, C. R. Journal ofNanoscience
and Nanotechnology 2007, 7, 2211-2221.

[54] Buyukserin, F.; Kang, M.; Martin, C. R. Small 2007, 3, 106-110.









[55] Ai, S.; Lu, G.; He, Q.; Li, J. Journal of the American Chemical Society 2003, 125, 11140-
11141.

[56] Kovtyukhova, N. I.; Mallouk, T. E.; Mayer, T. S. Advanced Materials 2003, 15, 780-785.

[57] Li, F.; Zhang, L.; Metzger, R. M. Chemistry ofMaterials 1998, 10, 2470-2480.

[58] Masuda, H.; Fukuda, K. Science 1995, 268, 1466-1468.

[59] Wood, G. C.; O'Sullivan, J. P. Electrochimica Acta 1970, 15, 1865-1876.

[60] Shingubara, S.; Morimoto, K.; Sakaue, H.; Takahagi, T. Electrochemical and Solid-State
Letters 2003, 7, E15-E17.

[61] Masuda, H.; Hasegwa, F.; Ono, S. Journal of the Electrochemical Society 1997, 144,
L127-L130.

[62] Masuda, H.; Yada, K.; Osaka, A. Japanese Journal ofApplied Physics, Part 2: Letters
1998, 37, L 1340-L 1342.

[63] Li, A. P.; Muller, F.; Bimer, A.; Nielsch, K.; Gosele, U. Journal ofAppliedPhysics 1998,
84, 6023-6026.

[64] Che, G.; Lakshmi, B. B.; Fisher, E. R.; Martin, C. R. Nature 1998, 393, 346-349.

[65] Hurst Sarah, J.; Payne Emma, K.; Qin, L.; Mirkin Chad, A. Angewandte Chemie,
International Edition 2006, 45, 2672-2692.

[66] Huber, C. A.; Huber, T. E.; Sadoqi, M.; Lubin, J. A.; Manalis, S.; Prater, C. B. Science
1994, 263, 800-802.

[67] Kovtyukhova, N. I.; Mallouk, T. E. Advanced Materials 2005, 17, 187-192.

[68] Park, S.; Lim, J.-H.; Chung, S.-W.; Mirkin, C. A. Science 2004, 303, 348-351.

[69] Lee, W.; Scholz, R.; Nielsch, K.; Goesele, U. Angewandte Chemie, International Edition
2005, 44, 6050-6054.

[70] Zhi, L.; Wu, J.; Li, J.; Kolb, U.; Muellen, K. Angewandte Chemie, International Edition
2005, 44, 2120-2123.

[71] Wood, G. C.; O'Sullivan, J. P.; Vaszko, B. Journal of the Electrochemical Society 1968,
115, 618-620.

[72] Masuda, H.; Satoh, M. Japanese Journal ofApplied Physics, Part 2: Letters 1996, 35,
L126-L129.









[73] Masuda, H.; Yamada, H.; Satoh, M.; Asoh, H.; Nakao, M.; Tamamura, T. Applied
Physics Letters 1997, 71, 2770-2772.

[74] Asoh, H.; Nishio, K.; Nakao, M.; Tamamura, T.; Masuda, H. Journal of the
Electrochemical Society 2001, 148, B152-B156.

[75] Diggle, J. W.; Downie, T. C.; Goulding, C. W. ChemicalReviews 1969, 69, 365-405.

[76] Jessensky, O.; Muller, F.; Gosele, U. Applied Physics Letters 1998, 72, 1173-1175.

[77] O'Sullivan, J. P.; Wood, G. C. Proceedings of the Royal Society ofLondon, Series A:
Mathematical, Physical and Engineering Sciences 1970, 317, 511-543.

[78] Parkhutik, V. P.; Shershul'skii, V. I. Journal of Physics D: Applied Physics 1992, 25,
1258-1263.

[79] Siejka, J.; Ortega, C. Journal of the Electrochemical Society 1977, 124, 883-891.

[80] Hench, L. L.; West, J. K. ChemicalReviews 1990, 90, 33-72.

[81] Livage, J.; Henry, M.; Sanchez, C. Progress in Solid State Chemistry 1988, 18, 259-341.

[82] Patrissi, C. J.; Martin, C. R. Journal of the Electrochemical Society 1999, 146, 3176-
3180.

[83] Sides, C. R.; Martin, C. R. Modern Aspects ofElectrochemistry 2007, 40, 75-126.

[84] Sides, C. R.; Li, N.; Patrissi, C. J.; Scrosati, B.; Martin, C. R. MRS Bulletin 2002, 27,
604-607.

[85] Lakshmi, B. B.; Patrissi, C. J.; Martin, C. R. Chemistry ofMaterials 1997, 9, 2544-2550.

[86] Lin, J.; Brown, C. W. TrAC, Trends in Analytical Chemistry 1997, 16, 200-211.

[87] Collinson, M. M. Mikrochimica Acta 1998, 129, 149-165.

[88] Avnir, D. Accounts of Chemical Research 1995, 28, 328-334.

[89] Walcarius, A. Electroanalysis 1998, 10, 1217-1235.

[90] Brinker, C. J. Current Opinion in Colloid & Interface Science 1998, 3, 166-173.

[91] Walcarius, A. Electroanalysis 2001, 13, 701-718.

[92] Stoeber, W.; Fink, A.; Bohn, E. Journal of Colloid and interface Science 1968, 26, 62-
69.









[93] Hench, L. L. Sol-Gel Silica-Properties, Processing and Technology Transfer; Noyes
Publications: New Jersey, USA, 1998.

[94] Iler, R. K. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface
Properties and Biochemistry, John Wiley & Sons, 1979.

[95] Pyell, U.; Stork, G. Fresenius'Journal ofAnalytical Chemistry 1992, 343, 576-581.

[96] Jal, P. K.; Patel, S.; Mishra, B. K. Talanta 2004, 62, 1005-1028.

[97] Kiselev, A. V.; Lygin, V. I. Infrared Spectra of Surface Compounds; Wiley: New York,
1975.

[98] Kiselev, A. V. Kolloidnyi Zhurnal 1936, 2, 17-26.

[99] Chuang, I. S.; Maciel, G. E. Journal of Physical Chemistry B 1997, 101, 3052-3064.

[100]Grushka, E.; Kikta, E. J., Jr. Analytical Chemistry 1977, 49, 1004A-1008A, 1010A,
1012A, 1014A.

[101] Van der Voort, P.; Gillis-D'Hamers, I.; Vansant, E. F. Journal of the Chemical Society,
Faraday Transactions 1990, 86, 3751-3755.

[102] Tundo, P.; Venturello, P.; Angeletti, E. Journal of the American Chemical Society 1982,
104, 6551-6555.

[103]Plueddemann, E. P. Silane CouplingAgents, 2nd ed.; Plenum Publishing Corporation:
New York, 1991.

[104]Loy, D. A.; Baugher, B. M.; Baugher, C. R.; Schneider, D. A.; Rahimian, K. Chemistry
ofMaterials 2000, 12, 3624-3632.

[105]Arkles, B. Chemtech 1977, 7, 766-778.

[106]Dias Filho, N. L.; Gushikem, Y. Separation Science and Technology 1997, 32, 2535-
2545.

[107]Lehninger, A. L. Principles ofBiochemistry, 2nd ed.; Worth Inc.: New York, 1993.

[108]Kemp, N. J.; Noble, R. D. Separation Science and Technology 1983, 18, 1147-1165.

[109]Lakshmi, B. B.; Martin, C. R. Nature 1997, 388, 758-760.

[ 110]Noble, R. D.; Koval, C. A. Materials Science of Membranes for Gas and Vapor
Separation 2006, 411-435.









[111]Wirgau, J. I.; Crumbliss, A. L. Dalton Transactions 2003, 3680-3685.


[112]Zaghbani, A.; Tayeb, R.; Bonnamour, I.; Felix, C.; Vocanson, F.; Lamartine, R.; Dhahbi,
M.; Seta, P. Journal of Membrane Science 2005, 258, 5-7.

[113]Villamo, O.; Barboiu, C.; Barboiu, M.; Yau-Chun-Wan, W.; Hovnanian, N. Journal of
Membrane Science 2002, 204, 97-110.

[114] Shentu, B.; Nishide, H. Industrial & Engineering Chemistry Research 2003, 42, 5954-
5958.

[115]Tsuchida, E.; Nishide, H.; Ohyanagi, M. Journal ofMacromolecular Science, Chemistry
1988, A25, 1327-1338.

[116]Barboiu, M.; Guizard, C.; Hovnanian, N.; Palmeri, J.; Reibel, C.; Cot, L.; Luca, C.
Journal of Membrane Science 2000, 172, 91-103.

[117] Gherrou, A.; Kerdjoudj, H.; Molinari, R.; Drioli, E. Separation Science and Technology
2002, 37, 2317-2336.

[118]Choi, H. W.; Kim, D. B.; Choi, D. K.; Ahn, B. S.; Kim, H. G.; Kim, H. S.; Lee, C. H.;
Sung, J. Y. Journal of Membrane Science 2006, 279, 403-409.

[119]Kaucher, M. S.; Peterca, M.; Dulcey, A. E.; Kim, A. J.; Vinogradov, S. A.; Hammer, D.
A.; Heiney, P. A.; Percec, V. Journal of the American Chemical Society 2007, 129,
11698-11699.

[120] Janeway, C. A., Travers, P.,Walport, M., Capra, J. D. Immunobiology: The Immune
System in Health and Disease, 5th ed.; Garland Publishing Inc.: New York, 2001.

[121]Burton, D. R. Trends in Biochemical Sciences 1990, 15, 64-69.

[122] Gellissen, G. Production ofRecombinant Proteins : Novel Microbial and Eukaryotic
Expression Systems; Wiley-VCH: Weinheim, Germany, 2004.

[123]Merten, O.-W., Mattanovich, D., Lang, C., Larsson, G., Neubauer, P., Porro, D., Postma,
P., Teixeira de Mattos, J., Cole, J.A. Recombinant Protein Production With Prokaryotic
and Eukaryotic Cells : A Comparative View on Host Physiology; Kluwer Academic Pub.:
Semmering, Austria, 2001.

[124]Hochuli, E.; Doebeli, H.; Schacher, A. Journal of Chromatography 1987, 411, 177-184.

[125] Smith, M. C.; Furman, T. C.; Ingolia, T. D.; Pidgeon, C. Journal of Biological Chemistry
1988, 263, 7211-7215.

[126]Porath, J.; Carlsson, J.; Olsson, I.; Belfrage, G. Nature 1975, 258, 598-599.









[127]Porath, J. Trends in Analytical Chemistry 1988, 7, 254-259.

[128]Eftink, M. R. Biophysical Journal 1994, 66, 482-501.

[129]Eftink, M. R. Methods in Enzymology 2000, 323, 459-473.

[ 130]Methods in Enzymology: Fluorescence Spectroscopy, 1997, 278, 1-628.

[131]Lakowicz, J. R. Principles ofFluorescence Spectroscopy, 2nd ed.; Plenum Publishers,
Inc.: New York, 1999.

[132] Guilbault, G. G. Practical Fluorescence. Modern Monographs in Analytical Chemistry,
2nd ed.; Marcel Dekker, Inc.: New York, 1990.

[133]Mirabella, F. M. Principles, Theory, and Practice ofInternal Reflection Spectroscopy, in
"Internal Reflection Spectroscopy Theory and Appliluiinii\ ", Marcel Dekker, Inc.: New
York, 1993.

[134] Skourlis, T. P.; McCullough, R. L. Journal ofAppliedPolymer Science 1994, 52, 1241-
1248.

[135]Harrick, N. J.; duPre, F. K. Applied Optics 1966, 5, 1739-1742.

[136] Tan, S., Belanger, D. Journal of Physical Chemistry B 2005, 109, 23480-23490.

[137]Watts, J. F., Wolstenholme, J. An Introduction to Surface Analysis by XPS andAES; John
Wiley & Sons Inc., 2003.

[138]Goldstein, J., Newbury, D. E., Joy, D. C., Lyman, C. E., Echlin, P., Lifshin, E., Sawyer,
L., Michael, J. Scanning Electron Microscopy andX-ray Microanalysis, 3rd ed.; Springer
Science + Business Media, Inc.: New York, 2003.

[139]Chilcott, T. C.; Chan, M.; Gaedt, L.; Nantawisarakul, T.; Fane, A. G.; Coster, H. G. L.
Journal of Membrane Science 2002, 195, 153-167.

[140]Gaedt, L.; Chilcott, T. C.; Chan, M.; Nantawisarakul, T.; Fane, A. G.; Coster, H. G. L.
Journal of Membrane Science 2002, 195, 169-180.

[141]Park, S.-M.; Yoo, J.-S. Analytical Chemistry 2003, 75, 455A-461A.

[142]Kharitonov, A. B.; Alfonta, L.; Katz, E.; Willner, I. Journal ofElectroanalytical
Chemistry 2000, 487, 133-141.

[143]Bilitewski, U. Analytica Chimica Acta 2006, 568, 232-247.









[144]Ye, M.; Jiang, X.; Feng, S.; Tian, R.; Zou, H. TrAC, Trends in Analytical Chemistry
2007, 26, 80-84.

[145]Domon, B.; Aebersold, R. Science 2006, 312, 212-217.

[146]Domon, B.; Broder, S. Journal ofProteome Research 2004, 3, 253-260.

[147]Marengo, E.; Leardi, R.; Robotti, E.; Righetti, P. G.; Antonucci, F.; Cecconi, D. Journal
ofProteome Research 2003, 2, 351-360.

[148]Barroso, B.; Lubda, D.; Bischoff, R. Journal ofProteome Research 2003, 2, 633-642.

[149]Masselon, C.; Pasa-Tolic, L.; Tolic, N.; Anderson, G. A.; Bogdanov, B.; Vilkov, A. N.;
Shen, Y.; Zhao, R.; Qian, W.-J.; Lipton, M. S.; Camp, D. G., II; Smith, R. D. Analytical
Chemistry 2005, 77, 400-406.

[150]Nedelkov, D.; Tubbs, K. A.; Niederkofler, E. E.; Kiernan, U. A.; Nelson, R. W.
Analytical Chemistry 2004, 76, 1733-1737.

[151]Zilberstein, G.; Korol, L.; Antonioli, P.; Righetti, P. G.; Bukshpan, S. Analytical
Chemistry 2007, 79, 821-827.

[152]Iritani, E.; Mukai, Y.; Murase, T. Filtration & Separation 1997, 34, 967-973.

[153]Ghosh, R.; Cui, Z. F. Journal of Membrane Science 2000, 167, 47-53.

[154]Harkensee, D.; Kokpinar, O.; Walter, J.; Kasper, C.; Beutel, S.; Reif, O. W.; Scheper, T.;
Ulber, R. Engineering in Life Sciences 2007, 7, 388-394.

[155]Guo, W.; Ruckenstein, E. Journal of Membrane Science 2003, 211, 101-111.

[156]Ramamoorthy, M.; Raju, M. D. Industrial & Engineering Chemistry Research 2001, 40,
4815-4820.

[157] Suen, S.-Y.; Liu, Y.-C.; Chang, C.-S. Journal of Chromatography, B 2003, 797, 305-319.

[158] O'Connor, A. J.; Pratt, H. R. C.; Stevens, G. W. Chemical Engineering Science 1996, 51,
3459-3477.

[159]Ho, A. K.; Perera, J. M.; Dunstan, D. E.; Stevens, G. W.; Nystrom, M. AIChE Journal
1999, 45, 1434-1450.

[160]Li, Q. Y.; Pepper, D. S. Journal of Membrane Science 1997, 136, 181-190.

[161]Nystrom, M.; Aimar, P.; Luque, S.; Kulovaara, M.; Metsamuuronen, S. Colloids and
Surfaces, A: Physicochemical and Engineering Aspects 1998, 138, 185-205.









[162]Nakatsuka, S.; Michaels, A. S. Journal of Membrane Science 1992, 69, 189-211.

[163] Saxena, A.; Shahi, V. K. Journal of Membrane Science 2007, 299, 211-221.

[164]Aravind, U. K.; Mathew, J.; Aravindakumar, C. T. Journal of Membrane Science 2007,
299, 146-155.

[165]Poetz, O.; Schwenk, J. M.; Kramer, S.; Stoll, D.; Templin, M. F.; Joos, T. O. Mechanisms
ofAgeing and Development 2005, 126, 161-170.

[166]Poetz, O.; Ostendorp, R.; Brocks, B.; Schwenk, J. M.; Stoll, D.; Joos, T. O.; Templin, M.
F. Proteomics 2005, 5, 2402-2411.

[167]Langmuir, I. Journal of the American Chemical Society 1918, 40, 1361-1402.

[168]Yang, L.; Chen, P. Journal of Membrane Science 2002, 205, 141-153.

[169]Weiss, M.; Welsch, W.; Schickfus, M. v.; Hunklinger, S. Analytical Chemistry 1998, 70,
2881-2887.

[170] Saha, K.; Bender, F.; Gizeli, E. Analytical Chemistry 2003, 75, 835-842.

[171]Igarashi, T.; Ishigatsubo, Y.; Ohno, S.; Shirai, A.; Aoki, I.; Okubo, T.; Klinman, D. M.
Clinical Immunology and Immunopathology 1994, 72, 307-311.

[172]Klinman, D. M.; Banks, S.; Hartman, A.; Steinberg, A. D. Journal of Clinical
Investigations 1988, 82, 652-657.

[173]Gueell, C.; Davis, R. H. Journal of Membrane Science 1996, 119, 269-284.

[174]Mueller, J.; Davis, R. H. Journal of Membrane Science 1996, 116, 47-60.

[175]Ho, C.-C.; Zydney, A. L. Journal of Membrane Science 1999, 155, 261-275.

[176]Jones, K. L.; O'Melia, C. R. Journal of Membrane Science 2000, 165, 31-46.

[177]Dechadilok, P.; Deen, W. M. Industrial & Engineering Chemistry Research 2006, 45,
6953-6959.

[178]Bhattacharjee, S.; Sharma, A.; Bhattacharya, P. K. Industrial & Engineering Chemistry
Research 1996, 35, 3108-3121.

[179]Johnson, C. P.; Jensen, I. E.; Prakasam, A.; Vijayendran, R.; Leckband, D. Bioconjugate
Chemistry 2003, 14, 974-978.









[180]Dancette, O. P.; Taboureau, J.-L.; Toumier, E.; Charcosset, C.; Blond, P. Journal of
Chromatography, B: Biomedical Sciences and Applications 1999, 723, 61-68.

[181]Rutherford, S. W.; Do, D. D. Adsorption 1997, 3, 283-312.

[182]Paul, D. R. Journal of Polymer Science, Polymer Physics Edition 1969, 7, 1811-1818.

[183]Tsuru, T.; Izumi, S.; Yoshioka, T.; Asaeda, M. AIChE Journal 2000, 46, 565-574.

[184]Nomura, T.; Nakao, S.; Kimura, S. Kagaku Kogaku Ronbunshu 1987, 13, 811-817.

[185] Chandler, D. Nature 2005, 437, 640-647.

[186]Lin, F.-Y.; Chen, W.-Y.; Hearn, M. T. W. Journal of Molecular Recognition 2002, 15,
55-93.

[187]Palma, E.; Klapper, D. G.; Cho, M. J. Pharmaceutical Research 2005, 22, 122-127.

[188]Norde, W.; Lyklema, J. Journal of Colloid and Interface Science 1979, 71, 350-366.

[189]Norde, W. Advances in colloid and interface science 1986, 25, 267-340.

[190]Norde, W.; MacRitchie, F.; Nowicka, G.; Lyklema, J. Journal of Colloid and Interface
Science 1986, 112, 447-456.

[191]Hamilton, R. G. Production and epitope location of monoclonal antibodies to the human
IgG subclasses. In Molecular Aspects of Immunoglobulin Subclasses; Pergamon Press:
Oxford, 1990, pp 79-91.

[192] Schmuhl, R.; Keizer, K.; van den Berg, A.; ten Elshof, J. E.; Blank, D. H. A. Journal of
Colloid and Interface Science 2004, 273, 331-338.

[193] Saksena, S.; Zydney, A. L. Biotechnology and Bioengineering 1994, 43, 960-968.

[194]Zydney, A. L.; Pujar, N. S. Colloids and Surfaces, A: Physicochemical and Engineering
Aspects 1998, 138, 133-143.

[195]Lopez, M. F.; Melov, S. Circulation Research 2002, 90, 380-389.

[196] Sun, L.; Crooks, R. M. Langmuir 1999, 15, 738-741.

[197]Bath, B. D.; Lee, R. D.; White, H. S.; Scott, E. R. Analytical Chemistry 1998, 70, 1047-
1058.

[198]Baltus, R. E. Journal of Membrane Science 1997, 130, 157-172.









[199]Galiatsatou, P.; Kanellopoulos, N. K.; Petropoulos, J. H. Journal of Membrane Science
2006, 280, 634-642.

[200]Funke, H. H.; Noble, R. D.; Koval, C. A. Journal of Membrane Science 1993, 82, 229-
236.

[201] Smith, K. A.; Meldon, J. H.; Colton, C. K. AIChE Journal 1973, 19, 102-109.

[202] Teramoto, M. Industrial & Engineering Chemistry Research 1994, 33, 2161-2171.

[203]Vieth, W. R.; Sladek, K. J. Journal of Colloid Science 1965, 20, 1014-1033.

[204]Paul, D. R.; Koros, W. J. Journal of Polymer Science, Polymer Physics Edition 1976, 14,
675-685.

[205]Vieth, W. R.; Howell, J. M.; Hsieh, J. H. Journal of Membrane Science 1976, 1, 177-220.

[206] Petropoulos, J. H. Journal of Polymer Science, Polymer Physics Edition 1970, 8, 1797-
1801.

[207]Mauze, G. R.; Stern, S. A. Journal of Membrane Science 1984, 18, 99-109.

[208]Riedl, C. M.; Spencer, H. G. Journal ofAppliedPolymer Science 1990, 41, 1685-1693.

[209]Noble, R. D. Journal of Membrane Science 1990, 50, 207-214.

[210]Kanehashi, S.; Nagai, K. Journal of Membrane Science 2005, 253, 117-138.

[211] Tsujita, Y. Progress in Polymer Science 2003, 28, 1377-1401.

[212]Barrer, R. M.; Barrie, J. A.; Slater, J. Journal of Polymer Science 1958, 27, 177-197.

[213] Stern, S. A.; Zhou, S.; Araux-Lara, J. L.; Ware, B. R. Journal of Polymer Science, Part
C: Polymer Letters 1989, 27, 427-431.

[214]Koros, W. J.; Paul, D. R. Journal of Polymer Science, Polymer Physics Edition 1978, 16,
1947-1963.

[215]Huh, J.-K.; Song, D.-I.; Jeon, Y.-W. Separation Science and Technology 1999, 34, 571-
586.

[216]Huh, J.-K.; Song, D.-I.; Jeon, Y.-W. Separation Science and Technology 2000, 35, 243-
259.

[217]Wang, J.-S.; Kamiya, Y. Journal of Membrane Science 1999, 154, 25-32.









[218]Dixon-Garrett, S. V.; Nagai, K.; Freeman, B. D. Journal of Polymer Science, Part B:
Polymer Physics 2000, 38, 1078-1089.

[219]Dixon-Garrett, S. V.; Nagai, K.; Freeman, B. D. Journal of Polymer Science, Part B:
Polymer Physics 2000, 38, 1461-1473.

[220] Meares, P. Journal of the American Chemical Society 1954, 76, 3415-3422.

[221]Budd, P. M.; McKeown, N. B.; Fritsch, D. Journal ofMaterials Chemistry 2005, 15,
1977-1986.

[222]Huang, S. H.; Margolies, M. N.; Rothstein, A. M.; Siekevitz, M.; Gefter, M. L.
Pharmacology Reviews 1982, 34, 43-49.

[223] Sem, D. S.; McNeeley, P. A.; Linnik, M. D. Archives of Biochemistry and Biophysics
1999, 372, 62-68.

[224]Petropoulos, J. H. Advances in Polymer Science 1985, 64, 93-142.

[225]Kempe, H.; Persson, P.; Axelsson, A.; Nilsson,B.; Zacchi, G. Biotechnology Bioengineer
2006, 93, 656-664.

[226]Bosma, J. C.; Wesselingh, J. A. Journal of Chromatography B 2000, 743, 169-180.

[227]Burczak, K.; Fujisato, T.; Hatada, M.; Ikada, Y. Biomaterials 1994, 15, 231-238.

[228]La Flamme, K. E.; Mor, G.; Gong, D.; La Tempa, T.; Fusaro, V. A.; Grimes, C. A.;
Desai, T. A. Diabetes Technology & Therapeutics 2005, 7, 684-694.

[229]Leoni, L.; Boiarski, A.; Desai, T. A. Biomedice Microdevices 2002, 4, 131-139.

[230]Bird, R. B.; Steward, W. E.; Lightfoot, E. N. Transport Phenomena, John Wiley, New
York, NY, 1960.

[231]Cussler, E. L. Diffusion. Mass Transfer in Fluid systems. Cambridge University Press,
Cambridge, UK, 1997.

[232]Edward, J. T. Journal of Chemical Education 1970, 47, 261-270.

[233]Pluen, A.; Netti, P. A.; Jain, R. K.; Berk, D. A. BiophysicalJournal 1999, 77, 542-552.

[234]Terazima, M. Physical Chemistry Chemical Physics 2006, 8, 545-557.

[235]Nishida, S.; Nada, T.; Terazima, M. Biophysical Journal 2004, 87, 2663-2675.

[236]Moschou, E. A.; Bachas, L. G.; Deo, S. K. Analytical Chemistry 2006, 78, 6692-6700.









[237]Hentz, N. G.; Daunert, S. Analytical chemistry 1996, 68, 3939-3944.

[238]Lager, I.; Looger, L. L.; Hilpert, M.; Lalonde, S.; Frommer, W. B. Journal of Biological
Chemistry 2006, 281, 30875-30883.

[239]Dwyer, M. A.; Hellinga, H. W. Current Opinion in Structural Biology 2004, 14, 495-504.

[240]Lager, I.; Fehr, M.; Frommer, W. B.; Lalonde, S. FEBSLetters 2003, 553, 85-89.

[241]Nilsson, B.; Forsberg, G.; Moks, T.; Hartmanis, M.; Uhlen, M. Current Opinion in
Structural Biology 1992, 2, 569-575.

[242] Seidler, A. Protein Engineering 1994, 7, 1277-1280.

[243]Zhang, Z.; Tong, K. T.; Belew, M.; Pettersson, T.; Janson, J. C. Journal of
chromatography 1992, 604, 143-155.

[244] Carlsson, J.; Porath, J.; Lonnerdal, B. FEBS letters 1977, 75, 89-92.

[245] Chadha, K. C.; Grob, P. M.; Mikulski, A. J.; Davis, L. R.; Sulkowski, E. Journal of
General Virology 1979, 43, 701-706.

[246] Chaga, G.; Andersson, L.; Ersson, B.; Porath, J. Biotechnology and applied biochemistry
1989, 11, 424-431.

[247]McCluskey, A. J.; Poon, G. M. K.; Gariepy, J. Protein Science 2007, 16, 2726-2732.

[248]Lee, K.-M.; Ma, K.-W.; Shaw, P.-C.; Wong, K.-B. Analytical Biochemistry 2006, 358,
152-154.

[249] Cantin, G. T.; Yi, W.; Lu, B.; Park, S. K.; Xu, T.; Lee, J.-D.; Yates, J. R. Journal of
Proteome Research 2008, 7, 1346-1351.

[250]Kermani, B. G.; Kozlov, I.; Melnyk, P.; Zhao, C.; Hachmann, J.; Barker, D.; Lebl, M.
Sensors and Actuators, B: Chemical 2007, B125, 149-157.

[251]Ueda, E. K. M.; Gout, P. W.; Morganti, L. Journal of Chromatography, A 2003, 988, 1-
23.

[252]Pearson, R. G. Journal of Chemical Education 1968, 45, 581-587.

[253]Arnold, F. H. Biotechnology 1991, 9, 151-156.

[254] Sulkowski, E. BioEssays 1989, 10, 170-175.

[255]Min, C.; Verdine, G. L. Nucleic Acids Research 1996, 24, 3806-3810.









[256]Hochuli, E.; Doebeli, H.; Schacher, A. Journal of Chromatography 1987, 411, 177-184.

[257]Gaberc-Porekar, V.; Menart, V. ChemicalEngineering & Technology 2005, 28, 1306-
1314.

[258] Sundberg, R. J.; Martin, R. B. Chemical Reviews 1974, 74, 471-517.

[259] Gasparac, R.; Martin, C. R.; Stupnisek-Lisac, E.; Mandic, Z. Journal of the
Electrochemical Society 2000, 147, 991-998.

[260]Meier, M. A. R.; Schubert, U. S. Journal of Polymer Science, Part A: Polymer Chemistry
2003, 41, 2964-2973.

[261]Herrero, M.; Tiemblo, P.; Reyes-Labarta, J.; Mijangos, C.; Reinecke, H. Polymer 2002,
43, 2631-2636.

[262] Tiemblo, P.; Guzman, J.; Riande, E.; Mijangos, C.; Reinecke, H. Macromolecules 2002,
35, 420-424.

[263]Dias Filho, N. L. Mikrochimica Acta 1999, 130, 233-240.

[264]Kemlo, J. A.; Shepherd, T. M. ChemicalPhysics Letters 1977, 47, 158-162.

[265] Varnes, A. W.; Dodson, R. B.; Wehry, E. L. Journal of the American Chemical Society
1972, 94, 946-950.

[266]de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.;
Rademacher, J. T.; Rice, T. E. ChemicalReviews 1997, 97, 1515-1566.

[267]Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Sacchi, D.; Taglietti, A. Analyst 1996, 121,
1763-1768.

[268]Huang, F.; Tian, Y.; Chen, C.-Y.; Cheng, Y.-J.; Young, A. C.; Jen, A. K. Y. Journal of
Physical Chemistry C 2007, 111, 10673-10681.

[269]McFarland, S. A.; Finney, N. S. Journal of the American Chemical Society 2001, 123,
1260-1261.

[270] You, C.-C.; Hippius, C.; Gruene, M.; Wuerthner, F. Chemistry--A European Journal
2006, 12, 7510-7519.

[271] Shen, Y.; Mackey, G.; Rupcich, N.; Gloster, D.; Chiuman, W.; Li, Y.; Brennan, J. D.
Analytical Chemistry 2007, 79, 3494-3503.

[272]Hartwig, J. F.; Pil, P. M.; Lippard, S. J. Journal of the American Chemical Society 1992,
114, 8292-8293.









[273] Thomton, N. B.; Schanze, K. S. New Journal of Chemistry 1996, 20, 791-800.

[274]McFarland, S. A.; Finney, N. S. Journal of the American Chemical Society 2002, 124,
1178-1179.

[275] Cotton, F. A., Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John Wiley & Sons:
NY, 1987.

[276]Housecroft, C. E., Sharpe, A. G. Inorganic Chemistry; Prentice Hall: NY, 2001.

[277] Silva, H. A. S.; Carlos, R. M.; Camargo, A. J.; Picchi, C. M. C.; Santos, R. H. d. A.;
McGarvey, B. R.; Franco, D. W. Inorganica ChimicaActa 2004, 357, 3147-3154.

[278] O'Sullivan, J. P.; Hockey, J. A.; Wood, G. C. Transactions of the Faraday Society 1969,
65, 535-541.

[279]Kimura, Y.; Shiraki, H.; Ishibashi, K.-i.; Ishii, H.; Itaya, K.; Niwano, M. Journal of the
Electrochemical Society 2006, 153, C296-C300.

[280]Abbott, T. P.; Nabetani, H.; Sessa, D. J.; Wolf, W. J.; Liebman, M. N.; Dukor, R. K.
Journal ofAgricultural and Food Chemistry 1996, 44, 2220-2224.

[281]Malhotra, V. M.; Jasty, S.; Mu, R. Applied Spectroscopy 1989, 43, 638-645.

[282] Spinner, E. Journal of the Chemical Society 1960, 1237-1242.

[283] Spinner, E. Spectrochimica Acta 1959, 95-109.

[284] Spinner, E. Journal of the Chemical Society 1962, 3127-3128.

[285]Katritzky, A. R.; Witek, R. M.; Rodriguez-Garcia, V.; Mohapatra, P. P.; Rogers, J. W.;
Cusido, J.; Abdel-Fattah, A. A. A.; Steel, P. J. Journal of Organic Chemistry 2005, 70,
7866-7881.

[286]Hayashi, T.; Mukamel, S. Journal ofPhysical Chemistry B 2007, 111, 11032-11046.

[287]Waldron, R. D. Physical Review 1955, 99, 1727-1735.

[288]Feenan, K.; Fowles, G. W. A. Journal of the Chemical Society 1965, 2449-2451.

[289]Kyotani, T.; Sato, K.; Mizuno, T.; Kakui, S.; Aizawa, M.; Saito, J.; Ikeda, S.; Ichikawa,
S.; Nakane, T. Analytical Sciences 2005, 21, 321-325.

[290]Do, J.; Ahn, C. H. Lab on a Chip 2008, 8, 542-549.









[291]VanWiggeren, G. D.; Bynum, M. A.; Ertel, J. P.; Jefferson, S.; Robotti, K. M.; Thrush, E.
P.; Baney, D. M.; Killeen, K. P. Sensors andActuators, B: Chemical 2007, B127, 341-
349.

[292]Medyantseva, E. P.; Varlamova, R. M.; Gimaletdinova, D. A.; Fattakhova, A. N.;
Budnikov, G. K. Journal of Analytical Chemistry 2008, 63, 275-279.

[293]Borisov, S. M.; Wolfbeis, O. S. ChemicalReviews 2008, 108, 423-461.

[294] Goel, G.; Kumar, A.; Puniya, A. K.; Chen, W.; Singh, K. Journal ofApplied
Microbiology 2005, 99, 435-442.

[295]Zeineldin, R.; Piyasena, M. E.; Sklar, L. A.; Whitten, D.; Lopez, G. P. Langmuir 2008,
24, 4125-4131.

[296]Ko, S.; Grant, S. A. Biosensors & Bioelectronics 2006, 21, 1283-1290.

[297]Liu, X.; Tan, W. Analytical Chemistry 1999, 71, 5054-5059.

[298] Wu, P.; Brand, L. Analytical Biochemistry 1994, 218, 1-13.

[299]Kagan, C. R.; Murray, C. B.; Nirmal, M.; Bawendi, M. G. Physical Review Letters 1996,
76, 3043.

[300] Cheung, H. C.; Dong, W.-J. Reviews in Fluorescence 2006, 3, 445-462.

[301] Schuler, B. Chemphyschem 2005, 6, 1206-1220.

[302] Invitrogen. Amine-Reactive Probes, MP00143, Molecular Probes: Eugene, OR, 2007.
http://probes.invitrogen.com/media/pis/mp00143.pdf

[303] http://probes.invitrogen.com/handbook/boxes/0422.html

[304]Piunno, P. A. E.; Krull, U. J.; Hudson, R. H. E.; Damha, M. J.; Cohen, H. Analytical
Chemistry 1995, 67, 2635-2643.

[305]Wang, J.; Xu, D.; Kawde, A. N.; Polsky, R. Analytical chemistry 2001, 73, 5576-5581.

[306]Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757-1760.

[307]Drummond, T. G.; Hill, M. G.; Barton, J. K. Nature Biotechnology 2003, 21, 1192-1199.

[308] Su, X.; Robelek, R.; Wu, Y.; Wang, G.; Knoll, W. Analytical Chemistry 2004, 76, 489-
494.









[309]Maruyama, K.; Mishima, Y.; Minagawa, K.; Motonaka, J. Analytical Chemistry 2002,
74, 3698-3703.

[310]Zhao, X.; Tapec-Dytioco, R.; Tan, W. Journal of the American Chemical Society 2003,
125, 11474-11475.

[311]Landers, J. P. Analytical Chemistry 2003, 75, 2919-2927.

[312]Miao, W.; Bard, A. J. Analytical Chemistry 2004, 76, 5379-5386.

[313]Castro, A.; Dalvit, D. A. R.; Paz-Matos, L. Analytical Chemistry 2004, 76, 4169-4174.

[314]Wang, R.; Minunni, M.; Tombelli, S.; Mascini, M. Biosensors & Bioelectronics 2004, 20,
598-605.

[315]Bui, M.-P. N.; Baek, T. J.; Seong, G. H. Analytical and Bioanalytical Chemistry 2007,
388, 1185-1190.

[316]Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 737-738.

[317]Marmur, J.; Doty, P. Journal of molecular biology 1961, 3, 585-594.

[318]Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science 1991, 254, 1497-1500.

[319]Nielsen, P. E. Current opinion in biotechnology 1999, 10, 71-75.

[320]Wang, J. Biosensors & Bioelectronics 1998, 13, 757-762.

[321]Arlinghaus, H. F.; Kwoka, M. N.; Jacobson, K. B. Analytical Chemistry 1997, 69, 3747-
3753.

[322]Wang, J.; Palecek, E.; Nielsen, P. E.; Rivas, G.; Cai, X.; Shiraishi, H.; Dontha, N.; Luo,
D.; Farias, P. A. M. Journal of the American Chemical Society 1996, 118, 7667-7670.

[323]Aoki, H.; Umezawa, Y. Analyst 2003, 128, 681-685.

[324] Abel, A. P.; Weller, M. G.; Duveneck, G. L.; Ehrat, M.; Widmer, H. M. Analytical
Chemistry 1996, 68, 2905-2912.

[325]Methods in Enzymology: Aviding-Biotin Technology, 1990, 184, 1-617.

[326] Macdonald, J. R. Impedance Spectroscopy; Emphasizing Solid Materials and Systems;
Wiley Interscience publications: NY, 1987.

[327]Winkler, B. H.; Baltus, R. E. Journal of Membrane Science 2003, 226, 75-84.









BIOGRAPHICAL SKETCH

Hector Mario Caicedo was born in Cali, Colombia. He graduated with a B.S. in chemistry

from the Universidad del Valle, in Cali. For his bachelor's thesis he conducted research in

analytical chemistry focusing on environmental pollutants under the direction of Dr. Nelly de

Palacios (his mentor in chemistry), developing an analytical methodology for the evaluation of

chromium in urine and blood by using electrothermal atomic absorption spectroscopy (ETAAS).

After that, he worked as a chemist in the industrial field, for a combined time of almost four

years, for both a biotechnology and pharmaceutical company. Then, he decided to go back to

academia to conduct his Master of Science in analytical chemistry at the Universidad del Valle.

In the summer of 2000, he visited the University of Florida as an international exchange student,

conducting research in electrochemistry, as a part of his master's research work, using carbon

fiber ultramicroelectrodes under the direction of Dr. Anna Brajter-Toth. After returning to the

Universidad del Valle, he conducted research in electroanalysis of xanthine by using carbon fiber

ultramicroelectrodes with surfaces electrochemically activated and polymer-modified, in the

group of Dr. Alonso Jaramillo. Once he graduated with a M.S. in chemistry from the University

del Valle, he continued his graduate studies in the Department of Chemistry at the University of

Florida to pursue his Ph.D. He worked under the supervision of Dr. Charles R. Martin, where he

conducted research in bioanalytical applications of affinity-based nanotube membranes for

sensing and separations. He completed his final work in April 2008, when he received a Doctor

of Philosophy degree.


212





PAGE 1

BIOANALYTICAL APPLICATIONS OF AFFINITY-BASED NANOTUBE MEMBRANES FOR SENSING AND SEPARATIONS By HECTOR MARIO CAICEDO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Hctor Mario Caicedo 2

PAGE 3

To my parents Herlinda and Luis, my wife Jessica, and my daughter Isabela 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank my Ph.D. advisor, Dr. Charles R. Martin for supporting my work, and for all his lessons on becoming a more effective scientist. I would like to express a very special gratitude to Dr. Jim Deyrup for his tremendous help and encouragement when I started pursuing my Ph.D. I would also like to thank Dr. Nicolo Omenetto for being part of my committee, for his invaluable time and discussions about my research work, and for his contagious enthusiasm for science. I would like to thank Dr. Amelia Dempere, Dr. David Powell, and Dr. Thomas Lyons for serving on my graduate committee and for their time, feedback, and willingness to proofread this dissertation. I am very grateful to Dr. Myungchan Kang, former postdoc in the Martin group, with who I had the opportunity to work with for two years, during which he always provided many helpful insights about my research work. I would like to thank my lovely wife Jessica and my beautiful daughter Isabela for their love, patience, and continuous support during the last years of my graduate studies. Finally, my greatest gratitude goes to my family, for their encouragement in pursuit of my educational goals. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................17 CHAPTER 1 INTRODUCTION AND BACKGROUND...........................................................................19 Introduction.............................................................................................................................19 Background.............................................................................................................................20 The Concept of the Template Synthesis..........................................................................20 Description of the Nanoporous Alumina Membrane Template......................................25 Two-step anodization method..................................................................................27 Pore growth mechanism...........................................................................................30 Sol-Gel Chemistry at a Glance........................................................................................32 Silane Chemistry a Versatile Tool for Surface Functionalization...................................35 Natures Transport across Membranes: An Inspiration for Membrane-Based Separations...................................................................................................................36 Fundamentals about Proteins...........................................................................................41 Antibodies................................................................................................................42 Recombinant proteins...............................................................................................44 Synopsis and Analytical Visualization of the Instrumental Techniques.........................46 Fluorescence spectroscopy.......................................................................................46 Fluorescence quenching...........................................................................................48 FTIR-ATR spectroscopy..........................................................................................49 X-ray Photoelectron Spectroscopy (XPS)................................................................51 Scanning Electron Microscopy (SEM)....................................................................52 Impedance spectroscopy..........................................................................................53 2 TRANSPORT AND FUNCTIONAL BEHAVIOR OF ANTIBODY-FUNCTIONALIZED NANOTUBE MEMBRANES............................................................55 Introduction.............................................................................................................................55 Experimental...........................................................................................................................57 Materials..........................................................................................................................57 Preparation of the Silica Nanotube Membranes..............................................................58 Antibody Immobilization................................................................................................58 Transport Experiments....................................................................................................59 Competitive Binding Assays...........................................................................................60 5

PAGE 6

Determination of Static Protein Binding Capacity and Binding Constants through Adsorption Isotherms...................................................................................................61 Effect of Equilibration Time on Flux and Selectivity via Dynamic Binding Capacity.......................................................................................................................62 Results and Discussion...........................................................................................................62 Electron Microscopy and Calibration Data.....................................................................63 Transport Properties of the anti-IgGs through PEG-Si-modified Control Membranes...................................................................................................................64 Competitive Binding Studies...........................................................................................65 Functional Behavior of the M-IgGand H-IgG-modified Membranes...........................67 Observed Functional Behavior is not a Transient Phenomenon.....................................68 Flux vs. Feed Concentration............................................................................................70 Effect of the Pore Size on the Flux and Selectivity Coefficient......................................71 Effect of the Immobilized-Antibody Orientation on the Flux.........................................73 Separation Based on a Simultaneous Combination of Size and Affinity Selectivity......74 Binding Capacities and Binding Constants.....................................................................77 Effect of the Temperature on the Flux and Binding Constant.........................................78 Effect of the Low Ionic Strength on the Selectivity........................................................83 Conclusions.............................................................................................................................85 3 MODELING AND CHARACTERIZATION OF THE TRANSPORT PROPERTIES OF PROTEINS ACROSS ANTIBODY-FUNCTIONALIZED NANOTUBE MEMBRANES BY USING THE DUAL-MODE MODEL..................................................87 Introduction.............................................................................................................................87 Definition of Dual-Mode Model.............................................................................................89 Results and Discussion...........................................................................................................95 Simulations of Permeability vs. Feed Concentration Based on the Dual-Mode Model...........................................................................................................................95 Experimental Description of the Dual-Mode Model.....................................................100 Effect of feed concentration on the flux.................................................................100 Experimental permeabilities as function of the feed concentration.......................101 Experimental evaluation of the dual-mode model parameters...............................102 Experimental evaluation of C sites and b..................................................................103 Experimental evaluation of the partition coefficients k p ........................................107 Experimental evaluation of the diffusion coefficients and solubility....................110 Fitting Experimental Data with the Theoretical Dual-Mode Model.............................114 Validation of the Dual-Mode Model.............................................................................119 Conclusions...........................................................................................................................121 4 METAL ION AFFINITY NANOTUBE MEMBRANES: PREPARATION, CHARACTERIZATION, AND CONTROLLED TRANSPORT OF SIX-HISTIDINE-TAGGED RECOMBINANT PROTEIN..............................................................................122 Introduction...........................................................................................................................122 Experimental.........................................................................................................................124 Materials........................................................................................................................124 6

PAGE 7

Preparation of the Silica Nanotube Membranes............................................................125 Preparation of Imidazole Functionalized Membranes...................................................125 Divalent Metal Ions Complexation...............................................................................126 Instrumental Characterization of the Imidazole Functionalized Membranes................127 Transport Experiments..................................................................................................128 Results and Discussion.........................................................................................................129 Imidazole Surface Modification and Characterization..................................................129 Fluorescence microscopy characterization.............................................................130 Characterization by FTIR-ATR spectroscopy........................................................135 Surface chemical composition by XPS..................................................................141 Transport Properties of Imidazole Membranes.............................................................146 Conclusions...........................................................................................................................149 5 TRANSPORT ANALYSIS OF PROTEINS ACROSS SILICA NANOTUBE MEMBRANES THROUGH A FLUORESCENCE QUENCHING-BASED SENSING APPROACH.........................................................................................................................151 Introduction...........................................................................................................................151 Experimental.........................................................................................................................152 Materials........................................................................................................................152 Preparation of the Silica Nanotube Membranes............................................................154 Antibody Immobilization..............................................................................................154 Labeling of anti-H-IgG with QSY-7.............................................................................155 Transport Experiments..................................................................................................156 Results and Discussion.........................................................................................................157 Experimental Principle Used for the Fluorescence Quenching Detection....................157 Verification of the Fluorescence Quenching System via H-IgG-FITC and anti-H-IgG-QSY Reaction in Solid State..............................................................................159 Control Transport Studies..............................................................................................160 Effect of the Feed Solution Concentration on the Rate of Quenching..........................162 Data Analysis Based on the Stern-Volmer Plots...........................................................164 Data Analysis Based on log-log Plots of Fluorescence vs. Time..................................167 Transport across Antibody-Modified Silica Nanotube Membranes..............................169 Conclusions...........................................................................................................................170 6 HYBRIDIZATION STUDIES OF DNA IN A PNA-MODIFIED NANOTUBE MEMBRANE PLATFORM.................................................................................................171 Introduction...........................................................................................................................171 Experimental.........................................................................................................................173 Materials........................................................................................................................173 Preparation of the Silica Nanotube Membranes............................................................174 Preparation of PNA-Modified Nanotube Membranes...................................................174 Nanotube Membrane Assembly....................................................................................176 Impedance Measurement for Membrane Characterization...........................................177 Impedance Data Analysis..............................................................................................177 Results and Discussion.........................................................................................................178 7

PAGE 8

Impedance Characterization of the Unmodified Membranes........................................179 Functionalization of the Silica nanotube Membranes...................................................180 Impedance Sensing of DNA-PNA Hybridization.........................................................182 Effect of the Pore Size on R m .........................................................................................183 Stability and Reproducibility of the PNA-modified Membranes..................................185 Conclusions...........................................................................................................................187 7 CONCLUSIONS..................................................................................................................189 LIST OF REFERENCES.............................................................................................................194 BIOGRAPHICAL SKETCH.......................................................................................................212 8

PAGE 9

LIST OF TABLES Table page 2-1. Fluxes and selectivity coefficients for anti-M-IgG and anti-H-IgG across the M-IgG and H-IgG-modified membranes................................................................................67 2-2. Thermodynamic parameters of G 0 associated with the affinity interaction of anti-M-IgG and M-IgG-modified membranes as a function of the temperature.......................82 3-1. Dual-mode model parameters used for the permeability plot simulations........................97 3-2. Thermodynamic parameters from adsorption isotherms of anti-M-IgG and anti-H-IgG on M-IgG modified membranes, according to the Langmuir monolayer model......107 3-3. The k p determined by various approaches.......................................................................108 3-4. Validation of the dual-mode model parameters...............................................................121 4-1. FTIR-ATR characteristic bands of imidazole functionalized nanotube membranes.......138 4-2. XPS surface composition and binding energies for bare alumina, chlorosilane modified, and imidazole functionalized membranes.......................................................143 4-3. Fluxes and selectivity coefficients for 6His-UHRP across imidazole functionalized metal ion affinity nanotube membranes...........................................................................149 5-1. Relationship between the Stern-Volmer quenching constant (K SV ) with the feed solution concentration of anti-H-IgG-QSY.....................................................................167 5-2. Relationship between slope of the log-log plots for the fluorescence signal vs. time with the feed solution concentration of anti-H-IgG-QSY................................................168 5-3. Data for the Stern-Volmer quenching constant (K SV ) and slope of the log-log plots for antibody-modified membranes...................................................................................169 6-1. Variations of R m at the different steps of modification on a 100 nm pore diameter alumina template..............................................................................................................182 9

PAGE 10

LIST OF FIGURES Figure page 1-1. Scanning electron micrograph pointing up the honeycomb structure of nanoporous alumina...............................................................................................................................25 1-2. Structure of nanoporous alumina.......................................................................................26 1-3. Electrochemical cell setup for nanoporous alumina growth..............................................28 1-4. Scanning electron micrographs of a nanoporous alumina membrane anodized at 50 V in 5% oxalic acid, (A) before and (B) after elimination of the barrier layer.....................29 1-5. Scanning electron micrograph of a highly-ordered nanoporous alumina membrane obtained at 50 V in 5% oxalic acid....................................................................................30 1-6. Outline of the steps involved in the sol-gel synthesis........................................................32 1-7. Fluorescence spectra obtained from the biofunctionalized (a) alumina template and (b) templated silica nanotube membranes with anti-Human IgG labeled with Alexa 488......................................................................................................................................34 1.8. Reactions involved in the alkoxysilane functionalization. Scheme obtained at http://www.dowcorning.com/content/publishedlit/SILANE-GUIDE.pdf.........................35 1-9. Types of specialized transport in living membranes (A) carriers, and (B) channels.........37 1-10. Flux concentration dependence for passive and facilitated transport................................38 1-11. Concentration gradients formed across the membrane during passive and facilitated transport.............................................................................................................................39 1-12. Basic structure of the immunoglobulin molecule..............................................................42 1-13. Intermolecular attractive forces involves in the antibody-antigen binding. Scheme obtained from www.fleshandbones.com/readingroom/pdf/291.pdf..................................44 1-14. Jablonski diagram of the electronic states during the fluorescence process......................47 2-1. Immobilization of antibodies on the template synthesized silica nanotube membranes.........................................................................................................................56 2-2. Flow-through fluorescence system used to measure the flux of the permeating proteins...............................................................................................................................59 2-3. Electron micrographs of (A) branch side, (B) solution anodization side, and (C) cross-section of the alumina template membranes, and (D) silica nanotubes after removal from the membrane..............................................................................................63 10

PAGE 11

2-4. Typical plots of (A) calibration data, and (B) calibration curve for Alexa-488 labeled anti-H-IgG..........................................................................................................................64 2-5. Flux plots for anti-M-IgG () and anti-H-IgG () through PEG-Si-modified control membranes. The feed concentration was 2.65 nM (lower curves) and 21.2 nM (upper curves)....................................................................................................................65 2-6. Fluorescence spectra of (A) M-IgGand (B) H-IgG-modified membranes after equilibration with a solution that was 21.2 nM in both (a) anti-H-IgG, and (b) anti-M-IgG................................................................................................................................66 2-7. Flux plots for anti-M-IgG () and anti-H-IgG () in (A) M-IgG-modified membrane, and (B) H-IgG-modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins.......................................................................68 2-8. Dynamic binding capacity study using M-IgG-modified membranes. The flux and selectivity coefficient vs. days of equilibration are shown for the transport of anti-M-IgG and anti-H-IgG. The feed solution concentration was 21.2 nM in both of the permeating proteins............................................................................................................69 2-9. Flux versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. The feed solution concentrations were equimolar in both of the permeating proteins............................................................................................................70 2-10. Typical plot of flux versus feed concentration for anti-M-IgG or anti-H-IgG across PEG-Si modified membrane. The data plotted is an average of the fluxes for anti-M-IgG and anti-H-IgG for each specific feed concentration..................................................71 2-11. Effect of the pore size on (A) the flux, and (B) selectivity coefficients for M-IgG modified membranes. The feed solution concentrations was 2.65 nM both in anti-H-IgG and anti-M-IgG...........................................................................................................72 2-12. Flux plots for anti-M-IgG () and anti-H-IgG () in well-oriented M-IgG-modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins...............................................................................................................................73 2-13. Scanning electron micrograph of (A) the surface, and (B) cross section of a highly-ordered nanoporous alumina membrane obtained at 30 V in 5% oxalic acid...................75 2-14. Flux plots for anti-M-IgG () and anti-H-IgG () in (A) PEG-Si-modified membrane, and (B) well-oriented M-IgG -modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins.........................................75 2-15. Flux plots for anti-M-IgG () and F(ab) 2 -anti-H-IgG () in well-oriented M-IgG -modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins............................................................................................................76 11

PAGE 12

2-16. Dissociation constants and binding capacities from adsorption isotherms of anti-M-IgG and anti-H-IgG on H-IgGand M-IgG-modified membranes, according to the Langmuir monolayer model...............................................................................................78 2-17. Flux plots for anti-M-IgG () and anti-H-IgG () in well-oriented M-IgG -modified membranes showing the effect of temperature. The feed solution concentration was 21.2 nM in both of the permeating proteins.......................................................................79 2-18. Effect of temperature on the flux and selectivity coefficients...........................................79 2-19. Adsortion isotherms carried out with well-oriented M-IgG-modified membranes for anti-M-IgG at three different temperatures: 23, 32, and 42 C..........................................80 2-20. Affinity temperature dependence of anti-M-IgG in well-oriented M-IgG-modified membranes.........................................................................................................................81 2-21. Flux plots for anti-M-IgG () and anti-H-IgG () across M-IgG -modified membranes, (A) without equilibration, and (B) after equilibration. The feed solution concentration was 2.65 nM in both of the permeating proteins.........................................84 3-1. Dual-mode transport model...............................................................................................88 3-2. Permeability plot simulations for the transport of (a) BM and (b) NBM across carrier-modified membranes. The parameters are summarized in Table 3-1....................96 3-3. Permeability plot simulations for the transport of BM across carrier-modified membranes illustrating the effect of the term F on the dual-mode model. (a) F = 1, (b) F = 0.5, and (c) F = 0. The other parameters are summarized in Table 3-1................97 3-4. Permeability plot simulations for the transport of BM across carrier-modified membranes showing the effect of the binding affinity on the permeability. (a) b = 1 x 10 11 cm 3 /moles, and (b) b = 3 x 10 10 cm 3 /moles. The other parameters are summarized in Table 3-1.................................................................................................100 3-5. Flux versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. The feed solution concentration was equimolar in both of the permeating proteins..........................................................................................................101 3-6. Experimental obtained permeability versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane..................................................102 3-7. Curve of adsorption dynamics for anti-M-IgG () in a M-IgG modified membrane. Inset shows the changes of fluorescence signal with time...............................................103 3-8. Adsorption isotherms for anti-M-IgG (), and anti-H-IgG () binding on M-IgG modified membranes. Also the adsorption isotherm for anti-M-IgG () using a control membrane is shown.............................................................................................104 12

PAGE 13

3-9. Adsorption isotherm for anti-M-IgG () binding on M-IgG-PG modified membranes.......................................................................................................................106 3-10. Extrapolations of the adsorption isotherms at infinite dilute concentration for anti-M-IgG (), and anti-H-IgG () in M-IgG modified membranes........................................108 3-11. Permeability plot simulations illustrating the effect of the partition coefficient on the dual-mode model. Simulation parameters: D = 5 x 10 -8 cm 2 /sec, b = 9 x 10 10 cm 3 /moles, C sites = 3.5 x 10 -8 moles/cm 3 and k p were respectively, 1 (), 0.1 (), 0.01 (), and 0.001()....................................................................................................109 3-12. Concentration-dependence of the solubility in term of the dual-mode model for anti-M-IgG (), and anti-H-IgG () across M-IgG modified membranes............................111 3-13. Correlation plots between permeability and solubility based on the concentration dependence for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified membranes.......................................................................................................................112 3-14. Experimental and calculated obtained permeability versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane.......................114 3-15. Experimental obtained diffusion coefficients versus feed concentration for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified membrane......................................117 3-16. Plots for experimental and calculated obtained permeability against (1 + bC feed ) -1 for (A) anti-H-IgG, and (B) anti-M-IgG across M-IgG modified membrane.......................120 4-1. Reaction mechanism of imidazole functionalized nanotube membranes........................126 4-2. Flow-through UV-Vis detection system used to measure the flux of six-histidine-tagged ubiquitin human recombinant protein..................................................................129 4-3. Fluorescence spectra of (a) silica nanotube and (b) imidazole functionalized membranes.......................................................................................................................130 4-4. Fluorescence spectra of imidazole functionalized membranes before (upper) and after (lower) metal ion immobilization: (A) Co +2 (B) Cu +2 (C) Ni +2 and (D) Zn +2 ......131 4-5. Fluorescence quenching (F/F 0 ) for imidazole functionalized membranes in the presence of nitrate solutions prepared in ethanol that were 1 mM of the respective divalent metal ion.............................................................................................................132 4-6. Fluorescence spectra of imidazole functionalized membranes (a) before and (b) after metal ion immobilization using an ethanol solution that was 1 mM in CuCl 2 ................133 4-7. Copper concentration dependence on fluorescence enhancement for imidazole functionalized membranes...............................................................................................134 13

PAGE 14

4-8. (A) FTIR-ATR spectra of bare alumina and imidazole functionalized membranes. (B) FTIR-ATR spectra of imidazole functionalized membranes in a narrow spectral region illustrating the effect of using two different backgrounds for the spectrum subtraction: (a) air, and (b) 3-chloropropyl silica nanotube membrane...........................136 4-9. Transmission FTIR spectrum of pure 2-mercaptoimidazole in KBr...............................136 4-10. FTIR-ATR spectra of imidazole functionalized membranes showing the effect of the metal ion immobilized on the thioamide bands: (a) Zn +2 (b) Co +2 and (c) Cu +2 ...........137 4-11. Transmission FTIR spectra of the alumina template film for each of the modification steps illustrated in Figure 4-1: (a) bare alumina template, (b) silica, (c) 3-chloropropyl silane modified and (d) imidazole functionalized nanotube membranes, respectively......................................................................................................................139 4-12. XPS survey spectrum for the bare alumina template.......................................................141 4-13. XPS survey spectra for (a) chlorosilane and (b) imidazole modified membranes..........142 4-14. XPS multiplexes analysis at high resolution for the imidazole functionalized membrane.........................................................................................................................144 4-15. High-resolution XPS data of the C1s peak for (a) bare alumina template, (b) chlorosilane modified, and (c) imidazole functionalized membranes.............................144 4-16. XPS survey spectra for imidazole modified membranes after immobilization of (a) Cu +2 and (b) Zn +2 ............................................................................................................145 4-17. UV-Vis spectrum for 50 M of 6His-UHRP measured in 50 mM phosphate-buffered saline solution pH 8.0 that also contained 300 mM sodium chloride and 2 mM imidazole..........................................................................................................................146 4-18. Typical plots of (A) calibration data and (B) calibration curve for 6His-UHRP.............147 4-19. Flux plots for 6His-UHRP in imidazole functionalized metal ion affinity nanotube membranes that had immobilized different metal ions: (a) Cu+2, (b) Ni+2, (c) Co+2, and (d) Zn+2. The feed solution concentration was 1.87 M for all permeation experiments......................................................................................................................147 4-20. Effect of the divalent metal ion immobilized on the flux of 6His-UHRP across imidazole functionalized metal ion affinity nanotube membranes..................................148 5-1. Flow-through fluorescence system used to measure the changes of fluorescence intensity in the receiver solution......................................................................................157 5-2. Structures of (a) QSY-7 and (b) FITC.............................................................................158 5-3. Spectra for FITC and QSY-7...........................................................................................158 14

PAGE 15

5-4. Molecular recognition reaction used to promote the fluorescence quenching................159 5-5. Fluorescence image and spectra respectively for a H-IgG-FITC-modified membrane, (A-C) before exposure, and (B-D) after exposure to a sample solution of anti-H-IgG-QSY..................................................................................................................................160 5-6. Typical plots for (A) changes of fluorescence signal, and (B) concentration of unquenched protein in function of time for two different control transport/detection systems: (a) anti-H-IgG/H-IgG-FITC, and (b) anti-H-IgG-QSY/R-IgG-FITC...............161 5-7. Typical plots for (A) changes of fluorescence signal and (B) concentration of unquenched protein in function of time for three different feed solution concentrations: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control)........163 5-8. Stern-Volmer plots of I 0 /I against concentration of anti-H-IgG-QSY in the receiver solution transported across silica nanotube membranes. The effect of the feed solution concentrations of anti-H-IgG-QSY is shown: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control).................................................................................165 5-9. Effect of the feed concentration on the Stern-Volmer quenching constant K SV ..............166 5-10. log-log plots of the of fluorescence signal vs. time showing the effect of feed solution concentration of anti-H-IgG-QSY: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control)................................................................................................167 5-11. Linear relationship between slopes obtained from the log-log plots with the feed concentration of anti-H-IgG-QSY...................................................................................168 5-12. Typical plots for (A) changes of fluorescence signal and (B) concentration of unquenched protein in function of time illustrating the transport of anti-H-IgG-QSY across (a) H-IgG-, and (b) M-IgG-modified membranes. The feed solution concentration was 33.3 nM for both experiments............................................................169 6-1. Structures of (a) PNA, and (b) DNA...............................................................................172 6-2. Immobilization of biotinylated-PNA on templated synthesized silica nanotube membranes.......................................................................................................................175 6-3. Nanotube membrane assembly........................................................................................176 6-4. General equivalent circuit model used for impedance data analysis...............................178 6-5. Nyquist plots (Z im vs. Z re ) for impedance measurement of (a) alumina and (b) silica-nanotube assembled membranes in highly purified water...............................................179 6-7. Effect of the wetting time on the resistance impedance measurement for an assembled unmodified alumina membrane 200 nm pore diameter.................................180 15

PAGE 16

6-8. Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water: (a) alumina, (b) after synthesis of silica-nanotube, (c) after silanization, (d) after biotinylation, (e) after streptavidin immobilization, (f) after PNA immobilization, and (g) after DNA hybridization. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz..................................................181 6-9. Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water for a PNA-modified nanotube membrane: (a) before, and (b) after 10 minutes of DNA hybridization. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz.........................................................................183 6-10. Effect of the membrane pore size on the membrane resistance: (A) 200 nm, and (B) 100 nm. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz.......184 6-11. Study of the hybridization process stability with time: (A) unmodified control alumina membrane, and (B) PNA-modified membrane. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA.......................................................................................................185 6-12. Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water for a PNA-modified nanotube membrane: (a) before, (b) after adding 3 nmoles of DNA, and (c) after 6 nmoles of DNA. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz..................................................186 6-13. Reproducibility study of the PNA-modified nanotube membranes, before and after the hybridization process. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA.......................187 16

PAGE 17

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BIOANALYTICAL APPLICATIONS OF AFFINITY-BASED NANOTUBE MEMBRANES FOR SENSING AND SEPARATIONS By Hctor Mario Caicedo August 2008 Chair: Charles R. Martin Major: Chemistry Nanotechnology has played an important role in the development of research and technology during the last two decades. The contribution of nanotechnology in different fields, along with the versatility of the constructed nanoscale materials, have made nanotechnology one of the most suitable tools to develop particular nanostructures to realize a desired function and application. A nanostructure is simply an entity at the nanometer scale with one, two or three dimensional features. Since nanotechnology covers a broad range of nanoscale materials, to simplify nanotechnology, it can be classified into two categories based on how the nanostructures are prepared: top-down and bottom-up. In the top-down methods, the nanostructures are constructed by chiseling larger bulk materials into entities of smaller size. Conversely, in the bottom-up case, small units are grown or assembled into their desired size and shape. The nanoporous materials specifically have attracted a lot of attention because they can be used for the synthesis of a variety of functional nanostructures of great usefulness in technology. These porous nanostructures usually combine many of the advantages of the top-down and bottom-up methodologies such as flexibility, size controllability, and cost. The research presented in this work utilizes nanoporous membranes to develop porous nanostructured platforms with potential applications in sensing and separations. In particular, this work is 17

PAGE 18

centered in fundamental studies for bioanalytical applications of affinity-based nanotube membranes for sensing and separations. A bottom-up methodology like the template synthesis was used to produce silica nanotubes inside of the pores of alumina membrane. The functionalization of the inside walls of these silica nanotube membranes allowed control of the functional behavior and properties of the nanostructured membrane during membrane-based separations and sensing. The general scheme of the work presented here, is distributed in seven chapters. The first chapter consists of a general description of the theory and fundamentals as background supporting evidence for the work presented here. The template synthesis method, the fabrication of the porous alumina membranes and the possibilities that silane chemistry offers in order to functionalize the nanostructured membranes is discussed. In addition, a brief overview of the fundamentals of biological production of proteins, antibodies, and recombinant proteins is described. After this introduction, the studies presented herein are centered in potential bioanalytical applications of the silica nanotube membranes. 18

PAGE 19

CHAPTER 1 INTRODUCTION AND BACKGROUND Introduction The use of membranes in industrial separations, substituting the traditional methods of separation, has become a more accepted procedure every day. Features such as compactness, simple and efficient operation, and low energy consumption have made membrane-based separations a fine alternative procedure. 1, 2 Membrane-based separations have found an extensive range of applications in technology, such as water and wastewater treatment, electrodialysis, gas separations, and fuel cell development. 2 The most important property of membranes is their ability to control the rate of permeation of different species. Traditionally, membranes have been used for sizeand charge-based separations with great success by using relatively low-resolution requirements. 3-7 However, some applications require the development of new membranes, because these rudimentary forms of selectivity are insufficient to allow membranes to compete with the traditional chromatographic methods. Current research and development efforts in membrane-based separations are directed toward drastic improvements in selectivity keeping the inherent characteristics of membranes. 8 In the search of new advanced membrane materials, the development of functionalized-membranes with a selective transport function executed through a specific molecule carrier has become in an interesting alternative to modulate the transport during membrane-based separations. 9, 10 This concept has been inspired in analogy by the strategies that biological membranes use to accomplish transport. Thus, the combined use of synthetic membranes with biological molecules such as antibodies and nucleotides has become important in the development of applications both in the biomedical and biotechnological field. 11 By using this combination, membranes have also acquired an important role in the design and development of biosensors. 12 In particular, nanotube membranes prepared by the general method 19

PAGE 20

called template synthesis to prepare nanomaterials have been extensively investigated. 4-6, 9, 10, 13-24 This method entails synthesis of the desired material within the cylindrical and monodisperse pores of a nanopore membrane or other solid. 25, 26 Both silica and gold nanotube membranes prepared by this method have demonstrated the capacity of being a novel platform for membrane-base separations and sensors. 14, 18 One of the most important features of the nanotube membranes prepared by the template synthesis when they are used in membrane-base separations, is that the deposited nanotubes are embedded within a mechanical strong support membrane, making it possible that they can act as conduits during the transport of permeating species between the solutions placed on either side of the membrane. Previous studies have shown that, by controlling the inside diameter of the nanotubes and the chemical functionalities present in the tube walls, highly-selective membranes for chemical and biochemical separations and sensors can be prepared. 46, 9,10, 13-24 This chapter provides a general description of the theory and fundamentals as background supporting evidence for the work presented here. The template synthesis method, the fabrication of the porous alumina membranes and the possibilities that silane chemistry offers in order to functionalize the nanostructured membranes it is discussed. In addition, a brief overview of the fundamentals of biological production of proteins, antibodies and recombinant proteins is described. At the end of the chapter, a brief summary of some of the instrumental analytical techniques used during the development and characterization of the applications described in this work is provided. Background The Concept of the Template Synthesis Our group has been pioneered in investigating a method called template synthesis for the fabrication of a variety of nanomaterials. 27 This template method is a general approach for preparing nanomaterials that involves the synthesis or deposition of the desired material within 20

PAGE 21

the pores or cavities of a nanopore membrane or other solid surface. Therefore, depending on the template structure and synthetic method used, a diverse array of nanostructured materials can be prepared. 28-31 Thus for example, from porous membranes is possible to prepare hollow nanotubes or solid nanowires, and by using solid spheres or porous spheres respectively, capsules or nanoporous particles can also be fashioned. Martin et al. have prepared nanotubes and nanowires composed of many types of materials, including conductive polymers, 32-34 metals, 35-37 oxides, 38 semiconductors, 39 carbon, 29, 31 and DNA. 40, 41 The variety of synthetic methodologies used for these nanofabrication include electroless plating of metals; electrodeposition of insulating polymers, conductive polymers, semiconducting oxides and metals; chemical vapor deposition; sol-gel chemistry; chemical polymerization; carbonization; and layer-by-layer template synthesis. Using either electrochemical 33, 34 or chemical 42, 43 oxidative polymerization of a equivalent monomer, Martin et al. have prepared nanowires and nanotubes out of conductive polymers such as polypyrrole, poly(3-methylthiophene) and polyaniline. Metals having nanoscopic dimensions with attractive optical properties have also been prepared by template synthesis. 44-45 For example, a macroscopic sample of pure gold has only one color, but nanoparticles of gold can show essentially all the colors in the visible region of the electromagnetic spectrum, depending on the size and the shape of the nanoparticles. 25 Hornyak et al. explored the optical properties of nanoscopic gold particles prepared by electrochemical deposition of gold within the pores of nanoporous alumina membranes. 44-45 They found that these particles achieve a plasmon resonance absorption limit that depended on the gold nanoparticle diameter. 21

PAGE 22

In the Martin group there is a commonplace methodology for the electroless deposition of gold within the pores of polycarbonate (PC) 10, 37 and polyethylene terephthalate (PET) 16, 46 template membranes. Depending on the plating time, it is possible to generate gold nanowires or nanotubes. Thus, in the nanotube situation, by controlling the plating time, the inside diameter can be varied. The electroless deposition of gold on the pore wall of these polymeric membranes is currently the best known method to precisely control the inside diameter of the template-synthesized nanotubes. 47 Martin et al. have shown that because the inside diameter of these gold nanotube membranes can be small relative to molecular dimensions, the nanotube membrane can be used to cleanly separate small molecules on the basis of molecular size. 47 By varying the rate of the plating reaction the diameter of the nanotubes was smaller in the faces of the template membrane. Thus, the flux and selectivity were determined for the bottleneck of the nanotubes. Together with size-based selectivity, they also demonstrated that generic chemical transport selectivity can be introduced into this membranes. 48, 49 More recently, Martin et al. have investigated gold nanotube membranes where the inner pore surfaces were systematically modified for biochemical functionalization to accomplish selective permeation of DNA. 10 They demonstrated a DNA-functionalized nanotube membrane for the separation of DNA by single-base mismatches selectivity via facilitated transport. 10 These DNA-functionalized nanotube membranes selectively recognize and transport the DNA strand that was complementary to the immobilized hairpin-DNA (carrier) within the pores, relative to DNA strands that are not complementary to the immobilized DNA. Another synthetic method used for the manufacture of nanotubes is the sol-gel chemistry to produce nanotubes of silica and other inorganic materials. The silica nanotubes have been used 22

PAGE 23

as paradigm to show the power of the template method for preparing nanotubes for biotechnological applications. 50 Silica nanotubes are easy to make, and they can be derivatized with an enormous variety of different chemical functional groups using simple silane chemistry with commercially available reagents. Mitchell et al. developed the concept of smart nanotubes that can be used for biological extraction and biocatalysis. 38 Nanotubes removed from the template and differentially functionalized both in the inner and outer nanotube surfaces were used for this application. Such differentially functionalized nanotube was used as smart nanophase extractor to remove specific molecules from solution. Thus, hydrophobic functionalized nanotube on the inside, with a hydrophilic nanotube exterior was used to extract lipophilic compounds from aqueous solutions. In the same work, Mitchell et al. showed that nanotube functionalized with an enantioselective antibody selectively extracts the enantiomer of a drug molecule that binds to the antibody from the racemic mixture. 38 Lee et al. demonstrated the use of silica nanotube membranes for enantioseparations. 9 They found that nanotubes embedded within the pores of the alumina template and functionalized with a Fab fragment antibody that binds the RS isomer of a drug molecule over the SR isomer, facilitated the transport of the RS enantiomer. They also found that the transport selectivity and binding affinity of RS over SR could be tuned by controlling the inside diameter of the membrane containing silica nanotube, and by addition of DMSO to the protein buffer solution, respectively. Gasparac et al. explored the template synthesized nano test tubes as an alternative payload-release strategy by using the template synthesis within the pores of nanopore alumina films. 51 These hollow nanotubes were closed on one end and open on the other. Therefore, if these nano 23

PAGE 24

test tubes could be filled with a payload and then the open end corked with a chemically labile cap, they might function as a universal delivery vehicle. More recently, Hillebrenner et al. demonstrated nano test tubes corked by chemical self-assembly. 52, 53 Buyukserin et al. have also prepared silica nano test tubes by using nanopore photoresist films as templates. 54 They used a plasma etch method, by using a nanopore alumina film as a mask, to etch a replica of the alumina pore structure into the surface of a photoresist polymer film. They found that the length of the test tubes is determined by the thickness of the porous part of the photoresist film. It is very important to control the inside diameter of the fashioned nanotubes, when the alumina membranes are used as template. In order to control the nanotube inside diameter a substitute template synthetic methodology has been considered, where the deposition of the film-forming material is based on layer-by-layer deposition. In this case, the wall thickness, and equally the inside diameter of the nanotubes would be determined by the number of layers of the material deposited along the pore wall. Using this approach, Ai et al. attempted to deposited layers of polyelectrolyte in the pores of a nanopore alumina membrane. 55 Kovtyukhova et al. have developed a method based on alternate SiCl 4 /H 2 O deposition/reaction cycles to make and control the thickness of silica nanotubes. 56 Hou et al. used Mallouks alternating ,-diorganophosphonate Zr(IV) chemistry to deposit layered nanotubes along the pore walls of an alumina template membrane. 40-41 For these nanotubes, the layers were held together by bonds between the phosphonate groups and the Zr(IV) ions, and the polyvalence of this interaction provided cross-linking that imparts the structural integrity of the layered nanotubes. 40 Hou et al. also used this synthetic strategy for the synthesis of DNA nanotubes where the bulk of the tube 24

PAGE 25

was composed of DNA, and the layers were held together by hybridization of complementary DNA strands. 40-41 The remarkable versatility offered by the material fabricated using template synthesis such as: size, shape, chemistry methodology applied, and systematic functionalization, make those nanostructured-template materials practical for particular applications. Description of the Nanoporous Alumina Membrane Template Nanoporous alumina films, which are produced by the electrochemical anodization of aluminum in suitable acidic or basic electrolytic solution, 44, 57 63 have been investigated and used as starting material for the fabrication several kinds of nanostructures, due to its very well ordered porous structure with high aspect ratio (thickness divided by the diameter). In recent years, simultaneously with the improvements in degree of ordering, these nanoporous alumina membranes have become a popular template system for the synthesis of various functional nanostructures. 9, 64 70 The porous structure appears as a honeycomb-like arrangement consisting of a close-packed array of columnar alumina units called cells, each containing a central straight pore, as shown in the scanning electron micrograph in Figure 1-1 from an alumina membrane obtained in our laboratory. Figure 1-1.Scanning electron micrograph pointing up the honeycomb structure of nanoporous alumina. 25

PAGE 26

By changing the anodization conditions such as electrolyte used, its concentration, the applied voltage and the period of anodization, the pores length, radius and the cell population density can be easily varied. The pore diameter depends on the applied voltage used for the anodization, and the thickness is a function of the anodization time. Pore diameters can be varied between 20 and 200 nm, with lengths between few tens of nanometers up to hundreds of microns. The pore size can be also controlled by the pore-widening treatment of dipping the anodized porous alumina in an appropriate acid solution after the anodization. This acidic solution is generally used to remove the barrier layer, which is a dense insulated oxide layer that separates the underlying aluminum surface and the porous oxide film (Figure 1-2). 58, 71 However before the barrier layer is eliminated, the porous alumina film has to be removed from the underlying aluminum metal in order to collect a free-standing membrane. There are two methods of removing the remaining aluminum substrate in order to obtain the free-standing membrane. The first one is by dissolution with a saturate HgCl 2 solution, and the second is a voltage-reduction method. The latter entails stepwise reduction of the potential to obtain an interfacial oxide layer that has a network of much smaller pores. 44 After formation of this highly porous layer, the alumina film is immersed into an acidic detachment solution which results in rapid dissolution of the interfacial oxide. 44 Figure 1-2.Structure of nanoporous alumina. 26

PAGE 27

There are two reported methods for the fabrication of highly ordered anodic nanoporous alumina membranes. One is the two-steps anodization method, which is the commonplace method used in our laboratory to produce the alumina membranes. Since the ordering of the pore configuration (self-organization) depends on the anodization time in the acidic electrolyte solution, two steps of anodization are required to obtain a highly ordered pore structure. 61, 62, 72 This method will be discussed in more detail in the next section. The second method is based on pretexturing of the aluminum substrate by using nanoindentation with a master mold. 73, 74 Masuda et al. used a SiC mold with a hexagonal array structure prepared by conventional electron beam lithography. 73, 74 In this pretexturing method, the depressions formed in the aluminum serve as initiation sites for the pore creation at the initial stage of the anodization. An ideally ordered nanoporous alumina membrane is formed after anodization of the aluminum substrate with the transferred textured pattern. Two-step anodization method The degree of the ordering of the pore initiation at the surface of the anodic porous alumina is low because the pores develop randomly at the initial stage of the anodization. To improve the ordering of the surface side of the anodic porous alumina, two-steps anodization is effectively adopted. 61, 62, 72 Two separated anodization processes are required for this procedure. The first anodization process consists of a long-period anodization to form the highly ordered concave nano-array of nucleation sites at the alumina/aluminum interface. The second anodization is performed after the removal of the oxide formed in the first anodization step. Thus, after the removal of the oxide, an array of highly ordered dimples is formed on the underlying aluminum substrate. These dimples act as initiation sites for the pore development during the second anodization. This process generates an ordered pore structure through the entire oxide layer. In the following paragraphs, the methodology used in our laboratory for the fabrication of alumina 27

PAGE 28

membranes will be discussed, which consists of three stages: polishing of the aluminum substrate, the two anodization steps, and the detaching step in order to have a free-standing membrane. Polishing of aluminum foils: In order to obtain a high quality alumina film, the starting aluminum material, must be very smooth. Initially, the high purity aluminum foils (99.999%) are mechanically polished with slurry of aluminum particles. If, to the naked eye, the aluminum foil shows severe scratches a mechanical polishing with fine sand paper is applied until the scratches disappear. The mechanical polishing is then followed by an electrochemical polishing. In this process, a potential difference of 15 V is applied between the anode aluminum foil and a lead plate which serve as cathode, by using a polishing electrolyte solution (95% concentrated phosphoric acid, 5% concentrated sulfuric acid and 20 g/L chromic oxide) heated to 70 C. Electropolishing steps of 5 minutes are successively repeated as many times as it is necessary, until the aluminum foil has a mirror-like surface. Two anodization steps: Electropolished aluminum is oxidized by using an electrochemical cell setup like the one presented in Figure 1-3. Figure 1-3.Electrochemical cell setup for nanoporous alumina growth. 28

PAGE 29

The first anodization step is performed in 5% oxalic acid as electrolyte for 12 hours, at 2 C, under a constant voltage. It should be pointed out that the pore diameter is controlled via anodization voltage. Thus, smaller pores require lower applied voltage and more conducted electrolytes such as sulfuric acid. 75 Large pores need larger voltages, which causes a high rate of dissolution in highly conductive electrolytes. 76 For this reason, the fabrication of larger pores requires lower conductive electrolytes, such as oxalic acid. 76 After the first anodization, the formed alumina film is removed in an etching solution that contains 0.4 M in phosphoric acid and 0.2 M in chromic oxide at 60 C. This anodization removal step leaves a uniform concave nano-array of nucleation sites in the underlying aluminum metal that correspond to each pore. The second anodization is performed at the same voltage and electrolyte conditions used in the first anodization. This second anodization step is carried out for a length of time, depending on the thickness of alumina membrane desired. In our laboratory, we obtained alumina membranes of thicknesses ranging from 0.3 to 150 m. Figure 1-4.Scanning electron micrographs of a nanoporous alumina membrane anodized at 50 V in 5% oxalic acid, (A) before and (B) after elimination of the barrier layer. Detaching step: The residual unoxidized aluminum is removed first with a saturated solution of HgCl 2 Then, the remaining barrier oxide layer is removed by etching the alumina film in a 10% phosphoric acid solution. Since the barrier layer thickness also varies with 29

PAGE 30

increasing voltage the etching step requires longer times for anodized membranes with bigger pores. Figure 1-4 shows the scanning electron micrographs of an alumina membrane obtained in our laboratory before and after elimination of the barrier layer. Thus, the resulting free-standing alumina membrane is obtained. In our laboratory, it is common to use a method with a two-step anodization to obtain highly ordered porous alumina membranes, as can be seen in Figure 1-5. Figure 1-5.Scanning electron micrograph of a highly-ordered nanoporous alumina membrane obtained at 50 V in 5% oxalic acid. Pore growth mechanism At the initial stage of anodization of aluminum, the sites of pore initiation are spread randomly over the aluminum surface. The ordering of the cell arrangement only proceeds with the anodizing time after steady state in the growth of the porous structure is established. 61, 62, 72 In addition, it is also known that the cell size and barrier-layer thickness are directly related to the anodizing potential. 58, 61, 62, 71, 72 The proposed mechanism for the formation of highly ordered porous alumina membranes has been explained as the result of two competing processes: 1) the pore initiation process due to a geometric effect, which is called field-assisted dissolution process 72 and 2) the self30

PAGE 31

organization of the pores which is considered to be triggered by mechanical stress at the alumina/aluminum interface. 76 In a first step the barrier layer is formed jointly with numerous defects in the aluminum surface for the pore initiation. Due to the decreasing current with the increase of the thickness of the barrier layer the formation of pores is initiated at defect positions. Then, assisted by the electric field, the alumina film developed at the metal-oxide interface at the pore walls are preferentially dissolved by the acidic electrolyte. For this reason, when the electric field increases by increasing the voltage, the dissolution process is accelerated and larger pores are formed. Thus, pores grow perpendicular to the surface via field-assisted oxide dissolution at the oxide/electrolyte interface in dynamic equilibrium with the oxide growth at the metal/oxide interface. 77, 78 It is believed that the anodization current is mainly related to the movement of ionic species (O 2, OH Al 3+ ) through the oxide layer at the bottom of the pore. 79 Therefore, as a consequence of the pore development, the electric field and the ionic current become concentrated in the barrier layer underneath the major pores. The oxide growth at the metal/oxide interface is due to the migration of oxygen containing ions (O -2 /OH ) from the electrolyte through the oxide layer at the pore bottom, with simultaneous ejection of Al 3+ ions drifted through the oxide layer into the solution at the oxide/electrolyte interface. 79 The fact that Al 3+ ions are lost into the electrolyte and that the atomic density of aluminum in alumina is by a factor of two lower than in metallic aluminum, have suggested a possible origin of forces between neighboring pores that contributes to the mechanical stress which is associated with the expansion during oxide formation at the metal/oxide interface. 76 Since the oxidation takes place at the entire pore bottom simultaneously, the material can only expand in the vertical direction, so that the existing pore walls are pushed upward. This somewhat explain the self-organized 31

PAGE 32

arrangement of neighboring pores in hexagonal arrays by repulsive interactions between the pores during the steady-state film formation after the pore initiation process. Sol-Gel Chemistry at a Glance The sol-gel processing of inorganic ceramic and glass materials has been known for a long time but it has been commonly used since the mid-1900s, and comprehensively studied during the past 30 years. 80, 81 In recent years, the sol-gel chemistry has evolved as a powerful approach for preparing microand nano-structured materials. 82-85 Sol-gel materials display attractive properties such as ease of preparation, low temperature synthesis, multiple structures with large open spaces of tunable porosity, controllable hydrophobicity, capacity for chemical modification of entrapment of various components, mechanical stability, and optical transparency. 86-89 The molecular precursors for preparation of sol-gel materials are essentially alkoxysilanes. The basic steps involved in the sol-process using tetraethoxysilane (TEOS) as a precursor molecule are shown in the Figure 1-6. The initiation step consists in the hydrolysis of the alkoxide precursor which leads to the formation of silanol groups (Si-OH) and ethanol. The silanols can then undergo polymerization reactions with other silanols or with other alkoxysilanes and this condensation reaction leads to the formation of siloxane groups (Si-O-Si) along with water and ethanol as additional reaction products. Figure 1-6.Outline of the steps involved in the sol-gel synthesis. 32

PAGE 33

These reactions lead initially to the formation of molecular clusters, which then grow upon polycondensation to produce a sol. This sol stage is produced when the particles formed are well dispersed in the liquid phase to form colloids. In addition, depending on the experimental conditions; this sol can be transformed in different ways. The usual case is gelation which occurs when the colloidal particles link together to become a three-dimensional network. Upon solvent extraction the gel is transformed into aerogel, 90, 91 while solvent evaporation results in the formation of xerogel. 92, 93 Those are the two most common processes of converting gels into silica. The maximum temperature for both processes can be kept under 100 C. The active silica surface with large specific surface area is of enormous utility to produce a variety of modified silica via covalent bond by using silane reagents. 94 By the insertion of organic functional groups to silica surface there is a partial conversion of surface silanol to a new organofunctional surface that acquires organophilic properties. Thus, the silane chemistry gives a set of properties to the silica surface, which differs considerably from the original matrix. 95 At the silica surface, the structure is composed of both siloxane groups (Si-O-Si) with the oxygen atom on the surface, and one of the three forms of silanol groups (Si-OH). 96 One of the types of silanol groups found in silica surface is the isolated free silanol groups, where the surface silicon atom has three bonds into the bulk structure and the fourth to the OH group. The second type corresponds to the occurrence of two vicinal free silanol groups where the OH groups are bridged by a H-bond. A third type of silanols called geminal silanol consists of two hydroxyl groups attached to one silicon atom. 97 Since the discovery of silanol groups on silica surfaces in 1936 by Kiselev, 98 studies based on theoretical calculations, physical methods and chemical methods have been explored in order to quantify the number of silanol groups on the surface. At present, it is accepted that the 33

PAGE 34

estimated number of OH groups accessible on porous silica surface is between 4 and 5 OH groups per nm 2 with an average surface area of 20.4-21.7 2 per silanol group. 94, 99 -101 In the research reported here, we took advantage of this special attribute of the higher silanol groups on silica surface for further silane functionalization, because when compared with alumina, the number of estimated hydroxyl groups accessible on alumina surface is only 0.1-2.5 OH groups per nm square. 30, 102 In order to prove the benefit of the silica nanotubes in the alumina template, some fluorescence experiments were performed. For these studies, both a templated silica nanotube membrane and alumina membrane were biofunctionalized with anti-Human IgG fluorescent labeled with Alexa 488. This anti-Human IgG was covalently attached via the Schiffs base reaction to the walls of the aldehyde functionalized nanotube membranes. 9 These results are shown in Figure 1-7, where the Alexa 488 fluorescence spectra from alumina membrane and from silica nanotube membrane are compared. The Alexa 488 fluorescence signal from silica nanotube templated membrane was approximately 6 times higher than the analogous alumina membrane. Hence, these results suggest that the deposition of silica nanotubes is essential to provide higher degree of functionalization on the membranes. Figure 1-7.Fluorescence spectra obtained from the biofunctionalized (a) alumina template and (b) templated silica nanotube membranes with anti-Human IgG labeled with Alexa 488. 34

PAGE 35

Silane Chemistry a Versatile Tool for Surface Functionalization The chemical modification of silica surfaces consist in chemical bonding of molecules using silane chemistry, which allow to modify the chemical or physical properties of the surface in a controlled way. Silane functionalization proceeds mainly via the reaction of silanol groups with organofunctional silanes, especially chloroand alkoxy-silanes. 103 The chemistry modification for alkoxysilanes is analogous to that one described for sol-gel production. They are easily prepared through the hydrolysis and condensation of organotrialkoxysilanes. The most common alkoxy groups are methoxy and ethoxy, which give methanol and ethanol respectively, as additional reaction products. The main difference between the trimethoxysilyl and triethoxysilyl monomers, when they are used in silane functionalization, is that reaction with trimethoxysilyl composite is faster. 104 The reactions involved in the alkoxysilane functionalization are illustrated schematically in Figure 1-8. Figure 1.8.Reactions involved in the alkoxysilane functionalization. Scheme obtained at http://www.dowcorning.com/content/publishedlit/SILANE-GUIDE.pdf For the chlorosilanes situation, the reaction with alcohol produces alkoxysilanes and hydrochloric acid. The extent of this reaction can be monitored by measuring the pH. 35

PAGE 36

The decision of using organosilanes with one or more than one hydrolyzable groups for silane functionalization depends on the particular set of properties desired in the new functionalized surface. Thus, when a hydrophobic surface is preferred, organosilanes with one hydrolyzable group are used. 103 If surfaces with higher degree of surface modification are desired, organosilanes with two or three hydrolyzable groups are used. 103 An alternative way of generating one monolayer of the desired silane monomer on silica substrates is by conducting silane chemistry of chlorosilanes in aprotic solvents. 105, 106 In conditions where the aprotic solvent and surface are water free, no alkoxysilane is produced, and no polymerization of the chlorosilane can take place. Therefore, the reaction must be promoted by nucleophilic substitution; where the nucleophile hydroxyl groups in the surface attack the primary halide carbon in the chlorosilane moiety. Surface modification in those cases takes longer time, usually 12-24 hours. Thereby, surface modification via silane functionalization provides a unique opportunity to regulate the interfacial properties of silica substrates while their basic structure and mechanical strength are preserved. Natures Transport across Membranes: An Inspiration for Membrane-Based Separations Biological membranes are vital components of all living systems, forming the external boundary of the living cells and the internal organelles into the cytoplasm. 107 They are constituted mainly of various lipids arranged in a bilayer that imparts a fluid character. Thus, both lipids and proteins are free to move within a bilayer. The proteins embedded in the bilayer and the carbohydrates attached to its surface, facilitate communication and transport across the membrane. In the living membranes, there are two types of membrane proteins which are involved in transport and signaling activities. Those are the peripheral and integral (or intrinsic) 36

PAGE 37

proteins. Peripheral proteins interact through electrostatic forces and hydrogen bonds with the surface of bilayers. While integral proteins are strongly associated with the bilayer itself. These functional characteristics described above, enable membranes to act as essential filters in different living processes that depend on membrane transport systems, such as the acquisition of nutrients, the elimination of waste materials, the transport of metabolites, the generation of concentration gradients vital for nerve impulse transmission and for the normal function of brain, heart, kidneys and other organs. 107 All transport processes in living membranes are mediated by the transport proteins which may function either as channels or carriers. In transport mediated by carrier proteins, only the specific binding molecule is transported. Carrier proteins cycle between conformations that allow molecules to pass to the other side of the membrane. There is never an open conduit all the way through the membrane with carrier proteins. Channels on the other hand, form open pores through the membrane that permits the free diffusion of molecules of appropriated size and charge. These two types of specialized transport systems are illustrated schematically in Figure 1-9. Figure 1-9.Types of specialized transport in living membranes (A) carriers, and (B) channels. In living membranes there are three classes of transport: passive, facilitated, and active. In passive transport, the transported specie moves across the membrane in a thermodynamically favored direction without the assistance of a specific transport system. For neutral species the driving force across the membrane is the chemical potential, while for charged species the 37

PAGE 38

driving force is the electrochemical potential. In addition, the flux of uncharged molecules across the membrane is linearly dependent on concentration. In facilitated transport, the transported species moves also according to a thermodynamically favored direction, but with the collaboration of a specific transport system the carrier. Facilitated transport systems display saturation behavior, where the flux is still dependent on concentration, but at high concentration the dependence is no longer linear. This behavior is due to the saturation of the binding carrier sites as the concentration of the transported species increases. Figure 1-10 compares the dependence on concentration of the flux for the passive and facilitated transport systems. Figure 1-10.Flux concentration dependence for passive and facilitated transport. The third class of transport in living membranes corresponds to the active transport systems. These use energy in order to transport the species against their thermodynamic potential. The most usual energy supplier is the ATP hydrolysis, but light energy and the energy stored in ion gradients are also used. 107 Thus, inspired in the strategies that biological membranes use to accomplish transport, scientists have explored the use of analog synthetic membranes with specific functionalization to selectively conduce transport. Likewise, these systems function on the base of facilitated 38

PAGE 39

transport concept. This facilitated transport synthetic functionalized-membranes integrate a mediator in the membrane the carrier. 9, 10 This carrier usually interacts chemically with one of the components in the feed to be transported across the membrane in a facilitated mode. Therefore, in the same way that living membranes facilitated transport, the flux for this preferred permeant is enhanced when is compared with the passive transport situation (Figure 1-10). As in the case of facilitated transport in living membranes, the facilitated transport in synthetic facilitated-membrane systems also displays saturation behavior. The basic mechanism for this enhanced transport is ascribed to the sequential binding and unbinding events of the facilitated specie with the immobilized carrier in the membrane. The concentration gradients of facilitated process are shown schematically in Figure 1-11, in which the reaction of the facilitated specie (F) with the immobilized carrier (C) in the membranes is given by Equation 1-1. CFCF (1-1) The equilibrium constant for this reaction is characterized by Equation 1-2. ][][][CFCFK (1-2) The terms in the square brackets represents the molar concentration of the chemical species involve in the reaction. Figure 1-11.Concentration gradients formed across the membrane during passive and facilitated transport. 39

PAGE 40

Thus, as can be seen from Figure 1-11, in the facilitated transport membranes, the presence of immobilized carriers that interact with the permeate molecules creates larger concentration gradients downstream the membrane length. Besides, the selectivity acquired with facilitated transport systems is an extra important aspect compared with the nonselective characteristic of the passive transport. Therefore, membranes with higher fluxes can be achieved by creating membranes with larger concentration gradients using immobilized carriers. This effect is also supported by Ficks first law of diffusion: xCDJ (1-3) Since according to Ficks first law, the flux (J) of a permeate molecule across a membrane is directly proportional to the concentration gradient across the membrane. The term C/x in Equation 1-3 represents the concentration gradient of the permeate specie and it is the driving force that leads to molecular movement. 108 The term D is the diffusion coefficient and it is a measure of the mobility of the individual molecules across the membrane. The negative sign indicates that the direction of flow is down the concentration gradient. The above description clearly demonstrates that carrier-facilitated transport membranes promote higher concentration gradients inside the membrane that enhance the flux of the molecule that interact selectively and reversely with the carrier. In simplest terms, the selective binding interaction enhances the local concentration gradient, and thus the flux, of the permeate molecule across the membrane. 9, 10, 109 Systems using synthetic membranes based on the concept of facilitated transport have been the subject of numerous investigations. 9, 10, 109 119 Thus, facilitated transport membrane systems 40

PAGE 41

using different type of carriers such as chemical complexing agents, 111 113 metalloproteins, 114, 115 crown ethers, 116, 117 antibodies, 9 DNA, 10 and others 118, 119 have been already demonstrated. Fundamentals about Proteins Proteins constitute the most important material that composes all living organisms. 107 They are part without exception of each cell. A brief review of the fundamentals of biological production of proteins and protein structure is described next. Proteins consist of a linear sequence of aminoacids, and their order determines the primary structure of the protein. Primary structure of a protein is encoded in the base pair sequence of DNA encoding the gene for the protein. Thus, the four different bases of nucleic acid sequences (adenine (A), cytosine (C), guanine (G), and thymine (T)) are translated into the 20 aminoacids of proteins by a genetic code and grouped into triplets of bases units called codons to code each particular aminoacid. The start codon of the gene is AUG which encodes methionine and the three stop codons are UAA, UAG, or UGA. The genetic expression is initiated when the RNA polymerase binds the start codon and transcripts a single-stranded copy of the DNA into the messenger RNA molecule (mRNA), this process takes place until the stop codon is reached. This mRNA molecule that carries now the genetic code of the DNA is then fed into the ribosome. Likewise, chemical polymerization processes, the biological synthesis of protein has three phases: initiation, elongation, and termination. In initiation, the mRNA binds in a particular complex to the transfer RNA (tRNA), which serves as a carrier of aminoacids to hybridize with the mRNA. During the elongation, the ribosome moves along the mRNA via codon-directed association. The ribosome covalently attaches the new amino acid to the end of the growing peptide chain and so forth the process continues. The termination of the peptide synthesis is activated by the appearance of a stop codon (UAA, UAG, or UGA). 41

PAGE 42

The ribosome produces an unfolded linear chain of denatured peptides. Then, the chain of aminoacids folds to form structures such as alpha-helices and beta-sheets, which are the elements of the protein secondary structure. The various secondary structures then fold into a compact tridimensional structure, which correspond to the tertiary structure of a protein. This tertiary structure, in many cases contains the functional form of the protein. In some cases, the combination of two or more subunits of tertiary structure peptides may be required for the functional form of the protein; this assembly is called the quaternary structure. Antibodies Antibodies, also called immunoglobulins, are proteins produced by the immune system for the defense of the organism against foreign species. 120 They exhibit very specific binding capacity for specific molecules. Immunoglobulins are divided in five classes: IgG IgM, IgA, IgD, and IgE. 120 They vary in size, charge, aminoacid composition and carbohydrate content. Immunoglobulins are made of two identical heavy and two identical light chains of aminoacids, forming a characteristic Y-shape (Figure 1-12). Figure 1-12.Basic structure of the immunoglobulin molecule. 42

PAGE 43

The molecular weight of the heavy chains varies with the class of antibody. They are called gamma, mu, alpha, delta and epsilon for IgG, IgM, IgA, IgD, and IgE, respectively. Based on small polypeptide structural differences, there are two types of light chains classified as kappa and lambda. This means that the polypeptides chains of immunoglobulins are encoded by three different clusters of genes. One cluster encodes the heavy chains for all the classes and subclasses of antibodies, a second one encode the kappa light chains and a third one the lambda light chains. The four glycoprotein chains are connected to one another by disulfide (S-S) bonds and noncovalent interactions. The variations in the number of disulfides bonds and the length of the hinge region determine the subclasses of antibodies. This hinge region is a flexible structure composed of aminoacids that enables the antibody to bind a couple of epitopes separated apart on an antigen. Immunoglobulins are bifunctional molecules. They have one region involved with the antigen binding which is called the Fab portions of the antibody, and a second region called the Fc portion which is involved in several biological activities. This Fc region is also used as a binding site for manipulating the orientation of the antibody during many immunochemical procedures, due to its capability to bind proteins G and A. 121 The Fab portions have a well-defined tridimensional shape specifically designed to bind the antigen, by a lock-and-key mechanism type. The binding between antigen-antibody entails the formation of multiple intermolecular noncovalent forces in the pocket of the binding site as is schematically shown in Figure 1-13. There must be appropriate atomic groups on opposite sites of the antigen and antibody to facilitate the binding between both proteins. Besides, the shape of the combining site must fit the epitope in order to hold the simultaneous formation of noncovalent bonds. Thus, the formation 43

PAGE 44

of hydrogen bonds results in hydrogen bridges between appropriated atoms. In the same way, the attractions between the opposite charged groups in both sides of the proteins contribute with the electrostatic forces, while the Van der Waals forces are created by the interaction between electrons clouds. The hydrophobic forces on the other hand, which may contribute up to half of the total strength of the antigen-antibody bond, rely in the association of non-polar hydrophobic groups, minimizing the contact with water molecules. All these attractive forces are weak if they are considered individually and compared with covalent bonds. However, if they are put together they become a large component of the total binding energy. Figure 1-13.Intermolecular attractive forces involves in the antibody-antigen binding. Scheme obtained from www.fleshandbones.com/readingroom/pdf/291.pdf Recombinant proteins Proteins produced by in vitro transcription and translation can be expressed in organisms such as Escherichia coli (E. coli). Since, the genome and structure of the E. coli have been long studied and well understood by biologists and geneticists. One of the main characteristics that make the E. coli as a very powerful and versatile system for genetic engineering and protein expression is its ability to rapidly proliferate; E. coli cells can be doubled in number under optimal conditions every five minutes. 122, 123 44

PAGE 45

Recombinant proteins are produced from clone DNA (cDNA) sequences which usually encode an enzyme or protein with a specific known function. Therefore, the process of genetic engineering begins with the isolation and extraction of the gene of interest from the clone DNA. This corresponds to the isolated DNA that is complementary to the mRNA that encodes the protein of interest. The next step is inserting the gene of interest by cloning techniques into a vector. The vector is just a small piece of DNA where foreign fragments of DNA may be inserted. Thus, a vector functions like a "molecular carrier", which will carry fragments of the gene of interest into a host cell. Those vectors are usually derived from plasmids, which are small, circular, double-stranded DNA molecules occurring naturally in the cytoplasm of bacteria. The vector thus incorporates an origin of DNA replication for the gene of interest into a host cell like E. coli. 122, 123 The production of the protein starts when these special engineered vectors are inserted into E. coli where the recombinant protein of interest is expressed. Because this E. coli duplicates rapidly, large amounts of E. coli cells containing the foreign gene can be grown in a matter of short time. In order to purify the expressed protein from the complex cell culture, it is a common practice to tag on the vector, a sequence of DNA that introduces a specific functionality into the expressed peptide to be used during protein purification. This tag or purification tool is placed in the vector to create a small extension of aminoacids added to the N-terminal of the target protein. This aminoacid sequence is usually designed to be a point of attack for a specific protease during purification. Thus, after the recombinant protein is expressed and extracted from E. coli, the N-terminal extension can be used to purify the protein and then cleavage it with a specific protease to generate an almost natural N-terminus sequence of the target peptide. 122, 123 45

PAGE 46

Frequently, the introduction of a metal binding site in the expressed recombinant peptide is placed for further purification. It has been demonstrated that an aminoacid sequence consisting of 6 or more histidine residues in a row operate as a metal binding site for recombinant proteins. 124, 125 This hexa-His sequence is called a His-Tag. By using commercially available vectors, the His-Tag sequence can be placed on the N-terminal of a target protein. This expressed His-Tag recombinant protein can be simply purified by using immobilized-metal affinity chromatography (IMAC). 126, 127 Synopsis and Analytical Visualization of the Instrumental Techniques Fluorescence spectroscopy The transport performance in membrane-based separation is normally investigated by permeation experiments of the systems in buffered-aqueous solutions of the permeant solutes to separate. One of the analytical methods used in this investigation to characterize the fluxes of the transported proteins across the membranes is fluorescence spectroscopy. This was achieved through the use of simple photophysical probes such as fluorescent-labeled antibodies. The use of fluorescence techniques for the study of proteins in solution is widespread. Fluorescence spectroscopy is a very sensitive technique, making it possible to perform experiments on protein solutions at very low concentration, minimizing the amount of molecules required for the analysis and eliminating problems associated with aggregation and low solubility. 128 130 Fluorescence is a characteristic process that occurs in certain type of molecules called fluorophores that have the property of absorbing energy and then immediately releasing this energy in the form of light. 131 The fluorescence process basically consists of three stages: excitation, lifetime of the excited state, and fluorescence emission. 46

PAGE 47

In the excitation stage, characteristic photon energy is absorbed for the flourophore molecules creating an excited electronic singlet state (S 2 ) as is illustrated in the Jablonski diagram in Figure 1-14. Figure 1-14.Jablonski diagram of the electronic states during the fluorescence process. This excited state lasts typically for a few nanoseconds (1 to 10 nanoseconds). During the excitation process, part of the excess of energy gained is dissipated by internal conversion to reach the lowest energy mode of a relaxed excited state (S 1 ). The final stage is the fluorescence emission, the electron returns to the ground state energy (S 0 ), where the remaining energy is released in the form of a photon. Due to the loss of energy during the internal conversion in the excited state, the total energy of the fluorescent emitted photons is lower than the absorbed excited photons. Thus, there is a difference between the excitation and emission wavelength which is called the Stokes shift. The ratio of the number of fluorescence photons emitted to the number of photons absorbed is called the fluorescence quantum yield, and it is a measure of the extent for the entire process. The intensity of the fluorescent emitted light, I, is described by the relationship showed in Equation 1-4. 132 bceQII10 (1-4) 47

PAGE 48

Where Q is the quantum yield, I 0 is the incident radiant power, is the molar absorptive coefficient, b is the path length of the cell, and c is the molar concentration of the fluorophore. However, for dilute concentration (bc < 0.05), the Equation 1-4 reduces to the form: bcQII 0 (1-5) Therefore, at dilute concentration of the fluorophore, the relationship between fluorescent signal and concentration should be linear, if Q, I 0 and are keeping constant. 132 Fluorescence quenching The fluorescence quenching phenomena refers to any process that decreases the fluorescence intensity of the fluorophore molecules. 131 The two general types of quenching processes encountered are the collisional and static quenching. The collisional quenching is a process that occurs during the excited state lifetime. In this case, the excited fluorophore experience contact with an atom or molecule (the quencher) that facilitate non-radiant transitions to the ground state. In the case of static quenching, the deactivation of the fluorescence intensity is due to the reaction between fluorophore and quencher that leads to formation of a stable complex in the ground state. The role of fluorescence quenching can be experimentally studied by determining the quenching rate parameters using the Stern-Volmer plots that are drawn in accordance with the Stern-Volmer equation (Equation 1-6). 131 QKIISV10 (1-6) Where [Q] is the concentration of the quencher, I 0 and I are respectively, the fluorescence intensity in the absence and presence of quencher concentration, and K SV is the Stern-Volmer quenching constant. In the cases where the relationship of Io/I with [Q] is linear the quenching process can be either collisional or static. However, those processes can be distinguished experimentally by 48

PAGE 49

carrying out fluorescence measurements at diverse temperatures. Thus, the quenching mechanism is mainly ascribed to a collisional process when at higher temperatures the system present larger values of K SV Conversely, for the cases of quenching via complex formation, higher temperatures show smaller K SV ; in this case K SV is related to the association constant of the complex (K a ). In situations, where the fluorescence ratio as a function of the quencher concentration reveals a positive deviation of the linear Stern-Volmer plot, it corresponds to a combined process of both collisional and static quenching. For such a case a modified Stern-Volmer equation is given by Equation 1-7. ][1][10QKQKIIaSV (1-7) Therefore, the upward curvature in the Stern-Volmer plot is due to the contribution of the square term (second power) of [Q] in the Equation 1-7. FTIR-ATR spectroscopy Fourier Transform Infrared Attenuated Total Reflection spectroscopy is an appropriate technique for the characterization of the surface of materials. This technique combines the capabilities of chemical structure identification offered by the infrared spectroscopy mutually with the fundamental properties of the total internal reflection techniques. The origin of the internal reflection techniques is involved with the fact that radiation propagating in an optical medium having higher refractive index (n 1 ) experiences total internal reflection at the interface of a contiguous sample medium with lower refractive index (n 2 ). This formed propagated wave, which penetrates a small portion of the sample medium, is called evanescent. 133 This phenomenon takes place only when the angle of incidence of the radiant wave exceeds a critical angle, c which can be determined by using Snells law (Equation 1-8). 1221 sin sinnn (1-8) 49

PAGE 50

It is known that in the position of the critical angle, the angle of refraction, 2 is 90. Consequently, the angle of incidence is the critical angle and the Snells equation is reduced to Equation 1-9: 12sinnnc (1-9) The critical angle, c can therefore be found simply by knowing the refractive indexes of the two mediums. It is worth mentioning that the evanescent wave is nontransverse, which allows all its vector components to interact with the dipoles in all directions, thus this wave is a very informative probe of the sample material. 134 During FTIR-ATR analysis, the infrared radiation is passed through an infrared transmitting crystal characterized by having a high refractive index, allowing the radiation to reflect within the crystal several times. Along with each reflection in the transmitting crystal, at the top surface, the evanescent wave penetrates the sample in order to explore the surface material. It should be pointed out that the depth of penetration and the total number of reflections along the crystal can be controlled either by varying the angle of incidence or by selection of crystals. According to Equation 1-10, the penetration depth of the evanescent wave in the sample material can be controlled by the selection of transmitting crystal, since each crystal posses its own refractive index. 2/121221sin2nnnd (1-10) Where is the wavelength of the IR radiation, n 2 and n 1 are the refractive index of the sample and the refractive index of the crystal respectively, and is the angle of incidence. Therefore, crystals with higher refractive index allow lower penetration depths. The depth of 50

PAGE 51

penetration of the evanescent wave, d, is defined as the distance generated at the crystal-sample interface where the intensity of the evanescent decays exponentially to 37% of its original value. 135 X-ray Photoelectron Spectroscopy (XPS) XPS also known as ESCA (Electron Spectroscopy for Chemical Analysis) alike to FTIR-ATR is a surface analysis technique, which defer in the depth of sample analysis and obviously in the incident source of excitation. For example, XPS probes only 5 to 10 nm depth whereas FTIR-ATR allows the detection of chemical species found in the first few microns of the sample material. 136 The XPS analysis is based upon the principles of the photoelectric effect, where the surface material is irradiated with monochromatic X-rays creating photoelectrons that are emitted from the sample surface. These ejected electrons come from the core level electrons of the atoms dispose on the surface material. 137 The emitted energy from the photoelectrons is then analyzed by an electron analyzer spectrometer to determine the binding energy of the electrons. From the binding energy and intensity of a photoelectron signal, it is possible to determine the elemental identity, chemical state, and quantity of an element. It should be pointed out that experimentally the kinetic energy (E k ) of the electrons is the physical parameter measured by the electron analyzer spectrometer and not its binding energy (E b ). However, it is really E b that provides the specific properties of an electron, because E k is not an intrinsic property of the analyzed material since this also depends on the photon energy of the X-rays employed during the analysis. The Equation 1-11 includes all the parameters involved in the XPS analysis and combines the relationship between E b and E k WEhEkb (1-11) 51

PAGE 52

Where h is the photon energy and W is the spectrometer work function. Thus, the binding energy of the electron is simply measured by the other three quantifiable parameters included in Equation 1-11. The XPS spectrum is composed only for the peaks of those electrons that escape the surface without energy loss. The electrons that suffer inelastic scattering and undergo energy loss contribute to the background of the spectrum. 137 As any other energy excitation process, the excess of energy in the excited state must be released or transformed in some way of energy. Thus in XPS, the ionized atom formed after the emission of a photoelectron must have some type of relaxation mechanism. This might be accomplished by X-ray fluorescence emission or by ejection of an Auger electron. This relaxation mechanism explains why during the XPS process, peaks for Auger electrons are also observed. Scanning Electron Microscopy (SEM) SEM is a microscope that uses electrons rather than light to form an image. It is created by focusing a high energy beam of electrons onto the surface. Once the focused electrons hit the surface, secondary and backscattered electrons are ejected from the specimen. 138 Those are collected in specific detectors and converted into a signal that is sent to a TV scanner where the image is processed and viewed as a digital image. It is worth mentioning that secondary electrons are low energy electrons (< 50 eV) that are originated within a few nanometers depth of the specimen. Whereas backscattered electrons are those electrons generated after incident electrons are scattered within the specimen volume and they are reflected or backscattered out of the specimen interaction volume keeping high energy. Backscattered electrons are useful to distinguish contrast between areas with different chemical composition because the image created with them depends on the mean atomic number of the 52

PAGE 53

substances that constitute the specimen. 138 Usually the SEM instruments are equipped with a set of two detectors to allow separate observations of topography (TOPO) image and composition (COMPO) image. Impedance spectroscopy Impedance spectroscopy is an effective method to probe interface properties of surface-modified electrodes. Generally there are two categories in impedance measurements: non-faradaic and faradaic impedance. Non-faradaic impedance is performed in the absence of any redox probe, which has been a method used for measuring changes in the electrical properties of synthetic membranes. 139, 140 Instead, electrochemical impedance is a faradaic impedance technique that is performed in the presence of a redox probe. Impedance is usually measured applying an AC potential to an electrochemical cell or electrical circuit and measuring the current going trough these. 141, 142 Impedance is expressed as a complex number. The complex impedance can be represented as the sum of real, Z re (), and imaginary, Z im (), components originated mainly from the resistance and capacitance of the system, respectively. The general electronic circuit, includes the ohmic resistance of the electrolyte, R s the Warburg impedance, Z W resulting from the diffusion of ions from the bulk electrolyte to the electrode, the double layer capacitance, C dl and the interfacial electron transfer resistance, R et present when a redox probe exist in the electrolyte. In the cases of non-faradic impedance, as measuring the electrical properties of synthetic membranes, this resistance can be associated with the changes in resistance across the membrane, R m The components R s and R W represent bulk properties of the electrolyte solution and diffusion features of the probe in solution. The other two components in the circuit, C dl and R et represent interfacial properties of the electrode and they are affected by immobilization of bio-organic material onto electrode surfaces. 141 Thus, analysis of C dl and R et can give important 53

PAGE 54

information about the extent of changes of the surfaces properties resulting from coupling of biomaterial to the studied interface. 142 A typical impedance data is represented in the form of Nyquist plot. The shape of this plot includes a semicircle region lying on the Z re -axis followed by an incline line. The semicircle region observed at higher frequencies corresponds to the kinetic limited process. Whereas the linear part characteristic of the lower frequency range represents diffusional limited process. In the Nyquist plot R s can be found by reading the real axis value at the high frequency intercept. While the sum of R m and R s can be found by fitting the semicircle until the real axis value at the lower frequency intercept. Nyquist plot present one shortcoming, when looking at any point on the plot, it cannot tell what frequency was used to record that point. There is another popular representation of the impedance data called the Bode plot. For this, the impedance is plotted with log frequency on the x-axis, and both the absolute value of impedance and the phase-shift on the y-axis. 141 54

PAGE 55

CHAPTER 2 TRANSPORT AND FUNCTIONAL BEHAVIOR OF ANTIBODY-FUNCTIONALIZED NANOTUBE MEMBRANES Introduction Today separation and analyses of proteins present important challenges for modern bioanalytical chemistry. It is clear that the most important frontier in the biomedical sciences and in the disease diagnostic is proteomics. The fact that mainly proteins execute and control cellular phenomena opens up the use of proteins as important molecular targets for drugs, and biomarkers for diseases. 143 Proteomics aims to understand complex biological systems by analyzing protein expression, protein function, protein modifications, and protein interactions. 144 Within the proteome-analysis strategies, separation techniques play an important role. 145 The 2D gel electrophoresis in combination with mass spectrometry is the most widely used separation/identification technology in proteomics. 145 147 However, there are a number of approaches to separate proteins based on their biochemical and biophysical features such as molecular weight, hydrophobicity, surface charge, isoelectric point, etc. that have also been developed and employed. 148 151 In addition to these chromatographic methods, synthetic membranes have been used for protein separations. 3 7, 152 164 Nevertheless, the membranes used to date typically separate proteins on the basis of size and charge, and these rudimentary forms of selectivity are insufficient to allow membranes to compete with chromatographic methods for protein separations. The use of affinity chromatography, however, has the advantage of offering a process that is specific for the protein of interest, overcoming the lack of selectivity and simplicity that many of the common methods of analysis undergo. We report here systems that exploit the advantages of affinity chromatography in a nanometer-scale membrane-based platform. In that search for new advanced membrane materials, our group has conducted fundamental studies in the 55

PAGE 56

development of functionalized-membranes with a selective transport function executed through a specific molecule carrier to modulate the transport during membrane-based separations. 9, 10 The carrier is a mediator that selectively binds to the specific chemical solute, the target compound to be separated from a sample solution. Depending on its kinetic interactions with the target compound, the carrier allows enhanced or retarded transport across the membrane. Here we describe the use of antibody-functionalized nanotube membranes for the separation of proteins. Since antibodies that selectively bind proteins are well known, 165, 166 it occurred to us that this antibody-based approach might provide a general route for preparing highly selective membranes for protein separations. To test this hypothesis, membranes containing silica nanotubes were prepared using the template synthesis method. 9, 25, 85 In the inside walls of these nanotubes, purified mouse immunoglobulin IgG (M-IgG) or human immunoglobulin IgG (H-IgG) were covalently attached, using an aldehyde-containing silane reagent (Figure 2-1). 9 Figure 2-1.Immobilization of antibodies on the template synthesized silica nanotube membranes. To study the functional behavior of the IgG-functionalized membranes, competitive transport experiments were done, using antibodies of anti-mouse IgG (anti-M-IgG) and anti-human IgG (anti-H-IgG) fluorescent labeled with Alexa 488 or Alexa 594. The studies of transport properties for these fluorescent labeled antibodies demonstrated that the nature of immobilized antibody changes the functional behavior of the functionalized membranes and the 56

PAGE 57

transport properties of the permeant proteins across them. Additional characterization of these antibody-based membranes showed a Langmuirian shape plot characteristic of a typical facilitated transport membrane when the effect of feed solution concentration on the fluxes of these fluorescent-labeled antibodies was investigated. Experimental Materials Commercially available nanopore alumina membranes (60 m thick, nominal pore diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS) (Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes. The purified mouse immunoglobulin (M-IgG) immobilized within the silica nanotube membranes was obtained from Sigma Aldrich. Membranes modified with purified human immunoglobulin (H-IgG) obtained from Sigma Aldrich were also prepared. A recombinant protein G obtained from Sigma Aldrich was used for well oriented antibody-immobilization membranes. A trimethoxysilyl poly(ethylene glycol) (PEG-Si, MW 460-590 Da) obtained from Gelest (Morrisville, PA) was used to modify the silica nanotubes control membranes. A trimethoxysilyl propyl aldehyde (United Chemical Technologies) was used to attach the M-IgG or H-IgG to the nanotubes. Two whole antibodies raised in goat against M-IgG were used to investigate the transport properties of the M-IgG-modified membranes. These anti-M-IgGs were obtained from Molecular Probes (Eugene, OR) conjugated with either the dye Alexa 488 or the dye Alexa 594. The transport properties of two control IgGs, raised in goat against human IgG (H-IgG), were also investigated. These anti-H-IgGs were also obtained from Molecular Probes conjugated with either Alexa 488 or Alexa 594. It is worthy to mention that these four conjugated antibodies were also used to investigate the transport properties of the H-IgG 57

PAGE 58

modified membranes. In this case the two control IgGs were the anti-M-IgG. A F(ab) 2 fragment antibody raised in goat against human IgG (F(ab) 2 -anti-H-IgG) and labeled with FITC fluorophore was also used for transport experiments. This F(ab) 2 -anti-H-IgG antibody was obtained from Molecular Probes. SuperBlock blocking buffer solution in phosphate-buffered saline was obtained from Pierce. SuperBlock solution contains a proprietary protein for blocking binding sites for non-specific protein adsorption. Tween 20 was obtained from Sigma Aldrich. All other chemicals were of reagent grade and used as received. Purified water, obtained by passing house-distilled water through a Barnstead, E-pure water-purification system, was used to prepare all solutions. Preparation of the Silica Nanotube Membranes Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-based template-synthesis method. 9, 25, 85 Briefly, a sol-gel precursor solution was prepared by mixing absolute ethanol, TEOS and 1 M HCl (50:5:1 by volume), and this solution was allowed to hydrolyze for 30 minutes. The alumina membrane was then immersed into this solution for two minutes under sonication. The membrane was then removed and the surfaces swabbed with ethanol to remove precursor from the alumina surface. The membrane was then air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. This yields silica nanotubes, with wall thickness of ~3 nm, embedded within the pores of the alumina membrane. 9 Antibody Immobilization The trimethoxy aldehyde silane was first attached to the inside walls of the silica nanotubes to introduce aldehyde terminal groups onto the pores of the membrane. 9 The membrane was then immersed into a solution that was 1.0 mg per mL in either M-IgG or H-IgG dissolved in 10 mM phosphate-buffered saline with pH = 7.4 that also contained 150 mM sodium chloride and 5 mM sodium azide. The membrane was incubated overnight at 4 C, 58

PAGE 59

which allowed free amino sites on the antibody to attach via the Schiffs base reaction to the nanotube-immobilized aldehyde groups 9 After immobilization, the membranes were rinsed with buffer and then incubated for 5 hours in the SuperBlock solution that also contained 0.05% Tween 20. Transport Experiments The M-IgGor H-IgG-modified membranes were sandwiched between two pieces of Scotch tape that had 0.31 cm 2 area holes punched through them. These holes defined the area of membrane exposed to the contacting solution phases. This membrane assembly was mounted between the two halves of a U-tube permeation cell, with half-cell volumes of ~10 mL. 4, 6, 10, 22, 23 Figure 2-2.Flow-through fluorescence system used to measure the flux of the permeating proteins. The transport properties of three types of membranes were investigated membranes containing immobilized M-IgG, membranes containing immobilized H-IgG, and control membranes with attached PEG-Si. The permeating proteins, whose transport properties were investigated, were the fluorescently-labeled anti-M-IgG and anti-H-IgG. The feed solution for all experiments was equi-molar in these two proteins, dissolved in 10 mM phosphate-buffered saline pH 7.4 that also contained 150 mM sodium chloride and 5 mM sodium azide (buffer used for all studies). The receiver solution on the other side of the membrane was identical except it initially contained no protein. 59

PAGE 60

The fluorescence intensity of the receiver solution was periodically measured using two flow-through fluorescence detectors connected in series (Figure 2-2). This system consisted of an L-7485 fluorescence detector (Hitachi) and an RF-2000 fluorescence detector (Dionex). The flow stream was delivered using an L-7100 HPLC pump (Hitachi). We discovered that if the receiver solution was sent continuously through the fluorescence detectors during the entire course of a transport experiment, significant photo bleaching of the Alexa dyes occurred. To prevent this, we turned the pump on to deliver receiver solution to the detectors for only 3 minutes per hour of transport experiment. That is, we used a duty cycle of 57 minutes with pump off followed by 3 minutes with pump on (flow rate = 3 mL per min) to collect fluorescence data. The whole system was controlled using Chromeleon 6.50 chromatography management software (Dionex) through a UCI-100 universal chromatography interface (Dionex). The Hitachi detector was used exclusively to detect the Alexa-594-labeled anti-M-IgG and the Dionex detector was used for the Alexa-488-labeled anti-H-IgG. Alexa 488 was excited at 495 nm, and the emission was detected at 515 nm. Alexa 594 was excited at 590 nm and detected at 617 nm. Calibration curves were obtained with the same flow through system on solutions containing known concentrations of both of the flourophore-labeled antibodies. These calibration curves were used to obtain the concentrations of the transport proteins in the receiver solution. Competitive Binding Assays Our experimental design assumes that the permeating protein anti-M-IgG binds strongly to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier. Likewise we assume that anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-IgG binds weakly to this carrier. We proved these assumptions to be true by conducting competitive 60

PAGE 61

binding assays. This was accomplished by equilibrating H-IgGand M-IgG-modified membranes to solutions that were 21.2 nM in both anti-M-IgG and anti-H-IgG, and then measuring the fluorescence intensities from the membranes. Calibration data were used to convert these fluorescence intensities to pmoles of anti-M-IgG and anti-H-IgG bound to the membrane. The standards used to obtain the calibration curves were 0.31 cm 2 membrane samples to which were applied known quantities of the desired protein. The fluorescence intensities from these standards were measured using a fluorescence microscope that was equipped with both a fluorescence detector and a CCD camera. This system combines an Axioplan 2 imaging microscope (Zeiss) with a J&M-PMT photometry system detector (SpectrAlliance), for measuring the fluorescence intensity using a BP450-490 filter for excitation and a LP515 filter for green emission of light, and a BP546/12 filter for excitation and a LP590 filter for red emission of light. Determination of Static Protein Binding Capacity and Binding Constants through Adsorption Isotherms The adsorption isotherms were carried out with M-IgG modified, H-IgG modified, and unmodified control membranes at 23 C in stirred U-tube permeation cells using the same buffer conditions used for the permeation studies, via solutions containing single concentrations of anti-M-IgG or anti-H-IgG fluorescent labeled with Alexa 488. In the case of M-IgG-modified but immobilized via protein G, the adsorption isotherms were carried out at three different temperatures: 23, 32, and 42 C. The fluorescence intensity of all the membranes was measured under the fluorescence microscope. Calibration curves were used to obtain the concentration of the Alexa 488-labeled proteins adsorbed to the membrane sample from the measured fluorescence intensity. The data were analyzed according to the monolayer Langmuir model. 167 61

PAGE 62

Effect of Equilibration Time on Flux and Selectivity via Dynamic Binding Capacity At the start of a transport experiment, all of the M-IgG or H-IgG binding sites are empty and these sites fill with anti-H-IgG or anti-M-IgG during the course of the experiment. We were concerned that binding equilibrium might not be achieved during the course of a transport experiment and that our results reflected this lack of equilibration. To explore this issue we investigated the flux and selectivity as a function of equilibration time. This was accomplished using seven membranes modified with M-IgG, which were exposed to a solution that was 21.2 nM in both of the permeating proteins, anti-H-IgG and anti-M-IgG. After one day of exposure, the first membrane was rinsed with buffer and a transport experiment was run (for 24 hours) using a feed solution that was 21.2 nM in both of the permeating proteins. After the second day of exposure, the same procedure was repeated for the second membrane, and so on, until the seventh membrane was investigated after seven days of exposure. Thus, after the binding sites available in the membrane are filled, the dynamic binding capacity is reached to conduct the transport experiments. 168 Results and Discussion The nanotube membranes investigated here contain a bound carrier, either M-IgG or H-IgG. The transport selectivity coefficient () across of the membranes was evaluated by comparing the flux of a permeating species that weakly binds to these carriers to the flux of a permeating species that binds stronger. In the case of the M-IgG-modified membrane, the strongly and weakly binding species are anti-M-IgG and anti-H-IgG, respectively. In the case of the H-IgG-modified membrane, the strongly and weakly binding species are anti-H-IgG and anti-M-IgG, respectively. These studies begin, however, with electron microscopy characterization of the alumina template and investigations of the fluxes of the two permeating proteins, anti-H-IgG and anti-M-IgG, across PEG-Si-modified control nanotube membranes. 62

PAGE 63

Electron Microscopy and Calibration Data Figure 2-3 shows electron micrographs of the surface and cross section of the alumina template membrane and of silica nanotubes deposited within the pores of such membranes. The alumina membrane has two different faces the face that was in contact with the underlying aluminum during anodization (branching side), and the face that was in contact with the electrolyte during anodization. The shape of the pores has a polygonal aspect on the branching side, while in the solution anodization side is rounded. Note that in the transport experiments the nanotubes were left embedded within the pores of the membrane, but they were removed here so that they could be imaged. The tubes have an outside diameter of 185 nm, as would be expected given the pore diameter, 188 nm in the solution anodization side and 113 nm for the branch side. The diameters for the alumina template were determined by electron microscopy based on the analysis of twenty pores in four different membranes. It is well known that the pores in these alumina templates branch at one surface due to the voltage-reduction method used to remove the membrane from the underlying Al surface. 44 Figure 2-3.Electron micrographs of (A) branch side, (B) solution anodization side, and (C) cross-section of the alumina template membranes, and (D) silica nanotubes after removal from the membrane. Evidence for this branching can be seen at the ends of some of the silica nanotubes (Figure 2-3D). The length of the pores with larger size cover most of the 60 m thickness of the alumina template, while the smaller pores cover less than 200 nm of the membrane. 63

PAGE 64

Figure 2-4 shows typical calibration data and a typical calibration curve. The calibration data (Figure 2-4A) were obtained by pumping the indicated concentrations of Alexa 488-labeled anti-H-IgG through the Dionex detector (Figure 2-2). The calibration curve obtained from these data (Figure 2-4B) showed excellent linearity (correlation coefficient = 0.996) over the range from 20 pM to 1280 pM. Analogous data were obtained for Alexa 594-labeled anti-M-IgG using the Hitachi detector. Figure2-4.Typical plots of (A) calibration data, and (B) calibration curve for Alexa-488 labeled anti-H-IgG. Transport Properties of the anti-IgGs through PEG-Si-modified Control Membranes The transport data were processed via flux plots 9, 10 moles of permeating protein transported across the membrane vs. time. Flux plots for feed solutions containing equi-molar concentrations of anti-M-IgG and anti-H-IgG across PEG-Si-modified control membranes are shown in Figure 2-5; data for two different feed concentrations are shown. These plots were linear (correlation coefficients >0.998) indicating that Ficks first law is obeyed and that over the time window investigated, the feed solution was not appreciably depleted of the transporting proteins. Figure 2-5 shows clearly that the fluxes of anti-M-IgG and anti-H-IgG are the same in the PEG-Si-modified control membranes. This means that the diffusion coefficients for these two 64

PAGE 65

proteins are the same, which in turn means that the proteins are identical in size, as would be expected for two IgG molecules. Therefore, the transport properties across these control membranes are the same for these two antibodies. These points are further reinforced by the concentration dependence of the flux. The concentration used for the lower two curves was eight-times lower than the concentration used for the upper two curves. Correspondingly, the flux at the high concentration (1.45 pmoles h -1 cm -2 ) is eight times higher than the flux at the low concentration (0.18 pmoles h -1 cm -2 ). After the permeation experiments, the membranes were observed under the fluorescence microscope, but no fluorescence was detected. This indicates PEG-Si-modification suppresses non-specific protein adsorption to the silica nanotubes. Figure 2-5. Flux plots for anti-M-IgG () and anti-H-IgG () through PEG-Si-modified control membranes. The feed concentration was 2.65 nM (lower curves) and 21.2 nM (upper curves). Competitive Binding Studies As noted above, our experimental design assumes that the permeating protein anti-M-IgG binds strongly to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier. Likewise we assume that anti-H-IgG binds strongly to the carrier H-IgG and that 65

PAGE 66

anti-M-IgG binds weakly to this carrier. We proved these assumptions to be true by conducting competitive binding assays. Figure 2-6A shows the fluorescence spectra of the permeating proteins bound to a M-IgG-modified membrane after equilibration with a solution that was 21.2 nM in both anti-M-IgG and anti-H-IgG. The membrane shows strong fluorescence from anti-M-IgG and very weak fluorescence from anti-H-IgG. Calibration data allowed us to convert these fluorescence intensities to moles of each protein bound to the surface. In agreement with our design paradigm, 10-times more anti-M-IgG was bound to the M-IgG-modified membrane than anti-H-IgG. The small amount of anti-H-IgG bound is due to antigen cross reactivity, as a result of the polyclonal characteristic of the antibodies. 169 172 Similarly, 14-times more anti-H-IgG was bound to the H-IgG-modified membrane than anti-M-IgG (Figure 2-6B). b a b a Figure 2-6.Fluorescence spectra of (A) M-IgGand (B) H-IgG-modified membranes after equilibration with a solution that was 21.2 nM in both (a) anti-H-IgG, and (b) anti-M-IgG. These experiments also confirmed the success of the chemistry immobilization and the capability of the antibodies immobilized in the nanotube membranes to preserve their bioactive properties on the porous surface. 66

PAGE 67

Functional Behavior of the M-IgGand H-IgG-modified Membranes In these studies, we have been exploring the transport properties, and separation of different anti-IgGs through antibody-modified nanotube membranes, and how the type of immobilized antibody alters the functional behavior of these functionalized membranes. Although, porous membranes have been used for size-based protein separation with great success, separation of proteins with similar size is more difficult. 7 This process is even more challenging, when the separation is between proteins that have practically the same molecular structure and size as in the case of different anti-IgGs. Flux plots for the transport of anti-M-IgG and anti-H-IgG across the M-IgG-modified nanotube membrane (Figure 2-7A) are linear, as are the flux plots for these permeating proteins for a H-IgG-modified membrane (Figure 2-7B). The fluxes obtained from the slopes of these plots are shown in Table 2-1. As mentioned above, the selectivity coefficient () across of the membranes was evaluated by comparing the flux of a permeating species that binds weakly to the carrier to the flux of a permeating species that binds more strongly. Table 2-1.Fluxes and selectivity coefficients for anti-M-IgG and anti-H-IgG across the M-IgG and H-IgG-modified membranes IgG-attached to the membrane Flux of anti-M-IgG (pmoles h -1 cm -2 ) Flux of anti-H-IgG (pmoles h -1 cm -2 ) M-IgG 0.60 1.31 2.2 H-IgG 1.33 0.48 2.8 It was observed that for both membranes the flux of the species that binds strongly to the carrier (e.g., anti M-IgG for the M-IgG-modified membrane) is lower than the flux of species that binds weakly to the carrier. This is a clear indication that the transport properties for these 67

PAGE 68

two permeating proteins are different, which in this situation, depends on the functional behavior of the IgG-functionalized nanotube membrane. It is worth mentioning that under steady state conditions observed in all the transport experiments, only an averaged transport rate can be determined because proteins being transported do not enter membrane pores at the same time so that the measured signal is a spatially and temporally averaged one. Figure 2-7.Flux plots for anti-M-IgG () and anti-H-IgG () in (A) M-IgG-modified membrane, and (B) H-IgG-modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins. Observed Functional Behavior is not a Transient Phenomenon One way to interpret the results in Figure 2-7 and Table 2-1 is that they might reflect a transient phenomenon. That is, one could hypothesize that during the permeation experiments summarized in Figure 2-7, the flux of the species that bound strongly to the carrier were suppressed because all the sites for binding were initially empty. According to this hypothesis the flux for the strongly binding species would remain low until all sites were filled and breakthrough occurred. It is difficult to reconcile this hypothesis with the constant flux observed for the strongly binding species over these nearly one-daylong transport experiments (Figure 2-7). One would 68

PAGE 69

anticipate instead, very low rates of transport at short times and higher rates at longer times as the sites fill. Nevertheless, we designed a set of experiments to test this hypothesis (Figure 2-8). We found that the flux of the weakly-binding species (in this case anti-H-IgG) remained higher than that for the strongly binding species (anti-M-IgG) even after 1 week of equilibration with a solution equi-molar in both proteins. Interestingly, the selectivity coefficient () improved slightly over the time course of this experiment. The other important point to glean from Figure 2-8 is that the data indicate that the membrane-bound M-IgG remains bioactive for the weak-long period summarized in the figure. Figure 2-8.Dynamic binding capacity study using M-IgG-modified membranes. The flux and selectivity coefficient vs. days of equilibration are shown for the transport of anti-M-IgG and anti-H-IgG. The feed solution concentration was 21.2 nM in both of the permeating proteins. It was also observed that some of the protein transmission was lost across membranes with more than 4 days of saturation. This effect was ascribed to the fact that of protein bound to the M-IgG-modified membranes decreased the effective pore size for the remaining diffusing protein molecules. Besides, such diminution in flux is not uncommon with protein-transporting membranes and results from gradual fouling of the membrane. 162, 173 176 69

PAGE 70

It was also found that after equilibration the weakly binding protein (anti-H-IgG) diffuse more rapidly than the strongly-binding protein (anti-M-IgG). Therefore, there is a trade-off between selectivity and flux after the membranes are equilibrated. The equilibration reduces the flux for the strongly binding protein with respect to the weakly binding one, but leads at the same time to an increase in selectivity. Thus, it can be concluded that the M-IgG carrier without equilibration in some extend facilitate the transport of anti-M-IgG when is compared with transport across membranes that have been previously equilibrated. Flux vs. Feed Concentration A number of competitive permeation experiments were carried out at different feed concentration, ranging from 2.65 nM to 0.5 M using M-IgG modified membranes (Figure 2-9). Figure 2-9.Flux versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. The feed solution concentrations were equimolar in both of the permeating proteins. Figure 2-9 shows flux-vs.-feed plots for transport of anti-H-IgG and anti-M-IgG across M-IgG-modified membranes. The flux for anti-H-IgG was at all times higher than the flux for anti-M-IgG. These plots show a characteristic Langmuirian shapes, 9, 10, 109 demonstrating that the M-IgG attached in the membrane indeed interacts with the penetrant-antibody, acting as a carrier during the permeation process. A straight line would be expected, if there is not interaction 70

PAGE 71

between carrier and permeating antibodies at the very low feed concentration used during these studies. A key way to prove that transport across a membrane is facilitated is to plot the flux vs. the feed concentration. 9, 10, 109 For facilitated-transport membranes, such plots are linear at low feed concentration but flatten (Langmuirian shape) 9, 10, 109 at high concentrations. Data for both of the permeating proteins show the characteristic Langmuirian shape of facilitated transport. Membranes that contained attached PEG-Si, instead of attached M-IgG showed linear plots as is shown in Figure 2-10. Therefore, the antibody-functionalized membranes showed here exhibit a type of behavior for facilitated-transport membranes during protein separations. Figure 2-10.Typical plot of flux versus feed concentration for anti-M-IgG or anti-H-IgG across PEG-Si modified membrane. The data plotted is an average of the fluxes for anti-M-IgG and anti-H-IgG for each specific feed concentration. Effect of the Pore Size on the Flux and Selectivity Coefficient The influence of the pore size on the flux and selectivity coefficient was explored by using three different sizes of commercially available (Whatman) nanopore alumina membranes having 60 m thickness, and nominal pore diameters respectively of 200, 100, and 20 nm. To study this effect, nanotube membranes of the specified pore size were modified with M-IgG. For the 71

PAGE 72

transport experiments, the feed solution concentration was in all cases 2.65 nM in both of the permeating proteins anti-H-IgG and anti-M-IgG (Figure 2-11A). Figure 2-11.Effect of the pore size on (A) the flux, and (B) selectivity coefficients for M-IgG modified membranes. The feed solution concentrations was 2.65 nM both in anti-H-IgG and anti-M-IgG. At this specific feed concentration (2.65 nM) not appreciable differences for the flux of anti-M-IgG across these M-IgG modified membranes were observed. However, lower fluxes were obtained for anti-H-IgG as the pore size decreases (Figure 2-11A). This reduction on flux is due mainly to the resistance that the transported antibodies experience in the pores that arises from the hydrodynamic interaction of the protein molecules with the wall of the pore, which include far field and near field interactions. 177, 178 Interestingly, the flux for the 200 nm membrane decreases slightly compared with the 100 nm membrane. This reduction in flux was ascribed to the fact that 200 nm pore size membranes are much easier of functionalize than 100 nm one. Therefore, the data suggests that the number of binding carrier sites immobilized in this membrane is possibly higher than the 100 nm, which in turn decreases the effective pore size of the membrane. Additional evidence for this conclusion was obtained from the comparison of fluorescence spectra of the membranes after transport experiments, which showed higher fluorescence intensities for the 200 nm M-IgG modified membrane. 72

PAGE 73

It was observed that the selectivity coefficient () increases as the pore diameter increases (Figure 2-11B). The increase in selectivity with bigger pores was attributed to the availability of much more room to diffusive for the protein that weakly binds the M-IgG carrier, which in this specific case corresponds to anti-H-IgG. In addition, this effect was observed in the case of the 20/200 nm membranes when the transport was carried out placing the feed solution on the 200 nm side of the membrane (Figure 2-11A). Effect of the Immobilized-Antibody Orientation on the Flux This study was accomplished by taking advantage of the specific affinity that protein G has for the Fc region of the IgG molecule. 121, 170, 179, 180 Thus, it was possible to manipulate the orientation of the immobilized antibody on the walls of the nanotubes. Figure 2-12.Flux plots for anti-M-IgG () and anti-H-IgG () in well-oriented M-IgG-modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins. It is known that due to steric hindrance and random orientation of the antibody molecules on the solid-phase surface, the binding activity of antibody immobilized directly on the solid-phase surface is usually less than that of soluble antibody. 179, 180 In order to overcome this problem, protein G was used to modify silica nanotube membrane for antibody immobilization. Thus, the well oriented immobilization of M-IgG was carried out on nanotube membranes 73

PAGE 74

modified with protein G that was previously immobilized via aldehyde silane chemistry on the nanotube walls. The results for the transport experiments of anti-M-IgG and anti-H-IgG across this well-oriented M-IgG modified-membrane are shown in Figure 2-12. The feed solution concentration was 21.2 nM in both of the permeating proteins. The fluxes of anti-H-IgG and anti-M-IgG across this well-oriented M-IgG modified membrane were 0.64 and 0.20 pmoles h -1 cm -2 respectively. As the M-IgG-modified membranes immobilized by direct immobilization on the nanotube surface, this membranes also preferentially transport the protein that binds weakly to the immobilized carrier. The fluxes for anti-H-IgG and anti-M-IgG were respectively two times and almost three times lower compared with those fluxes for membranes modified with M-IgG but where the immobilization was directly attached to the silane aldehyde surfaces in the membrane. The reason for these lower fluxes was attributed to the decrease in the effective pore size in the membrane with the addition of protein-G and the possibility of having all the M-IgG molecules in a well-oriented immobilization which also contributes to the decrease in the effective pore size. Besides, the well-oriented immobilization contributes with a stronger binding affinity, because this helps to preserve the binding activity of the antibody, which at the same time lowers the diffusion coefficients of the permeating proteins. Separation Based on a Simultaneous Combination of Size and Affinity Selectivity In order to prove this concept of a simultaneous combination of sizeand affinity-based separation of proteins, a nanotube membranes templated in a monodisperse pore size alumina of 40 nm crosswise the whole 60 m thickness was used. Scanning electron micrographs of the surface and cross section of this membrane are shown in Figure 2-13. 74

PAGE 75

It should be pointed out that the 40 nm pore size of this alumina template is about 3 times the hydrodynamic diameter for the IgG whole molecule (14 nm). 3 Therefore, after M-IgG immobilization the effective pore size of the nanotube membranes is reduced to a magnitude that allows a closer interaction of the permeating proteins both with the M-IgG immobilized carrier and the wall of the pores. Figure 2-13.Scanning electron micrograph of (A) the surface, and (B) cross section of a highly-ordered nanoporous alumina membrane obtained at 30 V in 5% oxalic acid. The next set of experiments demonstrated the capacity of these antibody-modified membranes to be used with a combination of both size and affinity selectivity. First, the transport of anti-H-IgG and anti-M-IgG across a control PEG-Si modified membrane and a well-oriented M-IgG modified membrane are compared (Figure 2-14). A B Figure 2-14.Flux plots for anti-M-IgG () and anti-H-IgG () in (A) PEG-Si-modified membrane, and (B) well-oriented M-IgG -modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins. 75

PAGE 76

The permeation process through membranes can be divided into its transient and steady state components. 181, 182 The transient component can be characterized by the time lag. This time lag observed phenomenon correspond to the time at which the permeating specie enters the membrane to the time that the flow rate of diffusing species into the close volume reaches a steady state of permeation. 181, 182 Figure 2-14A shows the transport of anti-M-IgG and anti-H-IgG across a PEG-Si modified control membrane. It is clearly observed a time lag event for both of the permeating proteins. This effect was even more noticeable for the antibody-modified membrane as can be seen in Figure 2-14B. By just comparing the time in which anti-M-IgG reaches a steady state of permeation across PEG-Si control and M-IgG-modified membranes, ten and twelve hour respectively, we can see that antibody immobilization has a significant effect in the functional behavior of the membrane. Therefore, by taking advantage of this time lag effect, transport of antibodies with different sizes were carried out in order to prove the concept of a simultaneous combination of sizeand affinity-based separation of proteins (Figure 2-15). Figure 2-15. Flux plots for anti-M-IgG () and F(ab) 2 -anti-H-IgG () in well-oriented M-IgG -modified membrane. The feed solution concentration was 21.2 nM in both of the permeating proteins. 76

PAGE 77

The difference between the experiment in Figure 2-14B and Figure 2-15 is that for the experiments in Figure 2-15 a F(ab) 2 -anti-H-IgG labeled with FITC was transported instead of the anti-H-IgG whole antibody labeled with Alexa 488. It can be seen from Figure 2-15 that the time lag is less evident for transport of the Fab fragment antibody than for the whole antibody IgG as a result of the smaller size of this protein. The other important point to glean from Figure 2-15 is that during the initial three hours of the experiment the flux of the fragment antibody seems lower that the whole antibody in Figure 2-14B. The reason for this lower flux was ascribed to the fact that the sensitivity reach with Alexa 488 is much higher than with the FITC fluorophore. Therefore, these experiments demonstrated the capacity of these antibody-modified nanotube membranes to be used with a combination of both size and affinity selectivity for protein separation. Binding Capacities and Binding Constants Adsorption isotherms were carried out to evaluate the protein binding capacity and the binding constants for the permeating proteins to the antibody carriers. The data were analyzed using a monolayer Langmuir model and nonlinear regression methods for the fitting. A summary of the results obtained from the adsorption isotherms carried out with M-IgG modified, and H-IgG modified membranes, using the 100 nm pore size commercially alumina templates is shown in Figure 2-16. By observing the binding capacities in Figure 2-16 obtained from each isotherm, it is obvious, that these antibody-modified membranes preferentially bound the protein that binds stronger to the carrier on the membrane as was demonstrated above. It was also observed that a small amount of the weakly bounding proteins were bound to the carriers, which was due to antigen cross reactivity. Likewise, the results for the dissociation constant showed the same trend observed for the binding capacities. Lower dissociation constants (higher binding constant) were 77

PAGE 78

obtained by the proteins that strongly bind to the carrier. The binding constant errors found in Figure 2-16 for the proteins that weakly bind the carriers may result not only from the possible lack of equilibrium due to the weak affinity with the carrier, but also from the random orientation effect of the immobilized antibody. It was also demonstrated that by controlling the orientation of the immobilized antibody using protein G the potency affinity of the antibodies is effectively preserved (bars in the right side of Figure 2-16). Figure 2-16.Dissociation constants and binding capacities from adsorption isotherms of anti-M-IgG and anti-H-IgG on H-IgGand M-IgG-modified membranes, according to the Langmuir monolayer model. Effect of the Temperature on the Flux and Binding Constant For these studies, nanotube membranes templated on the 100 nm pore size commercially alumina were modified with M-IgG well-oriented immobilized via the protein G immobilization technique. Transport experiments (Figure 2-17), and adsorption isotherms (Figure 2-19) were carried out at three different temperatures: 23, 32, and 42 C. Figure 2-17 shows the flux plots for the transport experiments carried out at the three studied temperatures. It was observed that as the temperature increases the flux for anti-M-IgG 78

PAGE 79

also increases, which in turn decreases the selectivity coefficient between for these two permeating proteins. It should be pointed out that the flux of anti-H-IgG also increased at higher temperatures, but in a less extent compared with the fluxes for anti-M-IgG. A much clearer illustration of this flux and selectivity behavior with the changes in temperature is shown in Figure 2-18. Figure 2-17.Flux plots for anti-M-IgG () and anti-H-IgG () in well-oriented M-IgG -modified membranes showing the effect of temperature. The feed solution concentration was 21.2 nM in both of the permeating proteins. Figure 2-18.Effect of temperature on the flux and selectivity coefficients. The effect of the temperature in the affinity was study by carrying out adsorption isotherms for anti-M-IgG with well-oriented immobilized M-IgG at 23, 32, and 42 C (Figure 2-19). As 79

PAGE 80

should be expected adsorption isotherms perform to higher temperatures displayed a more linear character than the typical Langmuirian shape obtained at room temperature, indicating a lower affinity for anti-M-IgG. Figure 2-19.Adsortion isotherms carried out with well-oriented M-IgG-modified membranes for anti-M-IgG at three different temperatures: 23, 32, and 42 C. The results of the data analyzed according to the Langmuir monolayer model are summarized in the Figure 2-20. Higher dissociation constants (Lower binding constant) were obtained as the temperature is raised. Interestingly, it is the fact that the binding capacity of the antibody-modified nanotube membranes remains almost constant with the change in temperature, which is a clearly indication that the functional behavior of these antibody-modified membranes is governed for the immobilized carrier in the membrane. It is worth mentioning that previous studies reported in literature have demonstrated that alumina membranes have low thermal expansion and good stability at high temperatures. 183 In addition, temperature studies ranging from 20 to 60 C have demonstrated that the permeability of solutes across these porous membranes increases slightly with temperature compared with the diffusivity in the bulk solution. 184 Therefore, it can be concluded that the enhancement in the fluxes observed in Figure 2-17, as the temperature was raised, is mainly due to the lower affinity between the permeating protein and the antibody carrier immobilized on the membrane. An 80

PAGE 81

additional evidence for this conclusion is that the fluxes at higher temperatures for anti-M-IgG were increased in a more extent compared with the fluxes for anti-H-IgG. Figure 2-20.Affinity temperature dependence of anti-M-IgG in well-oriented M-IgG-modified membranes The data obtained in Figure 2-19 and plotted in Figure 2-20, can be used to determine thermodynamic parameters such as G, H, and S of the system in study. Thermodynamic costs for a system indicate whether processes are likely to occur and they are conveniently quantified in terms of free energy G. 185 Therefore, the feasibility of the protein and antibody to interact (at a constant pressure P, and temperature T) is determined by the change in free energy of the system, which for a reversible process can be written as (Equation 2-1): 186 000STHG (2-1) Where H 0 and S 0 are the enthalpic and entropic changes experienced during the biomolecular reaction. The enthalpic part is measure of the average potential energy of interaction between molecules, and the entropic part is a measure of the order or intermolecular correlations. 185 By assuming that the biomolecular reaction is a reversible process, the relationship between binding affinity constant (K a ), and free energy change can be expressed by Equation 2-2. 186 (2-2) aKRTGln0 81

PAGE 82

Since the values K a T and R the gas constant (8.314 Joules moles -1 K -1 ) are known, by using Equation 2-2 was possible to calculate G 0 The results of these energetic calculations are summarized in Table 2-2. Table 2-2.Thermodynamic parameters of G 0 associated with the affinity interaction of anti-M-IgG and M-IgG-modified membranes as a function of the temperature Temperature ( o C) Temperature ( K ) K a ( M -1 ) G o ( KJ / moles ) 23 296 3.76 x 10 8 48.59 32 305 7.81 x 10 7 46.08 42 315 3.19 x 10 7 45.24 As can be seen from Table 2-2, the resulting G 0 values were negative, which indicate that the process is exothermic and proceed spontaneously toward the formation of the complex protein-antibody. In addition, we can see how small changes in free energy generate significant changes in the binding affinity constant. Therefore, these changes in energy can be associated with the variations in the fluxes observed in Figure 2-17 when the temperature was changed. Now that that the changes of free energy as a function of the temperature were established, from a plot of G 0 vs. T the values of H 0 and S 0 can be extracted from Equation 2-1, by assuming that H 0 is independent of temperature. From the slope and intercept of the obtained plot with a linear correlation of 0.918 the estimated values for S 0 and H 0 were 0.17 0.06 KJ/mol and -99.95 17.12 KJ/mol, respectively. Those estimated values confirm the spontaneity of this biomolecular reaction toward the formation of the protein-antibody complex. Interesting is that S 0 is still positive even though the loss of rotational-translational degrees of freedom after complex formation. This effect has been extensively studied and had been associated with 82

PAGE 83

partial desolvation and amount of water excluded during the formation of multiple intermolecular noncovalent forces in the pocket of the binding site. 187 -190 Thus, these studies confirm that the functional behavior of these antibody-modified membranes is governed for the changes in free energy before and after interaction of the permeating proteins with the carrier antibody, which can be attributed to the changes in enthalpy and entropy of the system as was demonstrated here. Effect of the Low Ionic Strength on the Selectivity One method to improve selectivity in membrane-based separations is to shut down the transport of one of the permeating species. It was found that by using transport solution with low ionic strength it is possible to contemplate this option with antibody-modified membranes. To test this option, we studied the rate and selectivity of transport of anti-H-IgG and anti-M-IgG in only water solution across M-IgG modified nanotube membranes templated in commercial alumina with 100-nm-diameter pores (Figure 2-21). It was observed that after two hours of transport in low ionic strength conditions (only water), the fluxes of both the anti-H-IgG and anti-M-IgG were depleted (Figure 2-21A). The selectivity for these two permeating proteins through M-IgG membranes without equilibration was only 2.7. Alternatively, it was found that transport across M-IgG-modified membranes that were previously equilibrated for six hours in a water solution that was 2.65 nM in both of the permeating proteins (anti-H-IgG and anti-M-IgG), the selectivity coefficient was increased to 17.3. This effect is shown in Figure 2-21B. The effect of low ionic strength in the selectivity was attributed to the fact that IgG has no unique isoelectric point, since each specific antibody molecule has its own variable region containing different charged residues. 191 Thus, for transport experiments carried out at low ionic strength the thickness of the electric double-layer has a considerable influence in the mobility of the permeating proteins across the antibody-modified membranes. 186, 192 Consequently, for 83

PAGE 84

transport experiments carried out at low ionic strength in antibody-modified membranes that has been previously equilibrate with a protein that strongly binds the antibody carrier, both the modified membrane and this permeating protein bearing similar electric double-layer that creates double-layer repulsions when they approach each other. Besides, at low ionic strength the overlapping between the double-layers both of the membrane pore wall, and that one of the permeating protein affects the mobility across the membrane. 193, 194 Therefore, the repulsion effects together with the double-layer overlapping are the responsible factors of the enhanced selectivity at low ionic strength observed in Figure 2-21B. Then, the high selectivity obtained under these conditions might be attributed to the electrostatic exclusion of the charged anti-M-IgG molecules from the membrane pores that were equilibrated with the same protein (anti-M-IgG) having identical isoelectric point. 162, 193 Figure 2-21.Flux plots for anti-M-IgG () and anti-H-IgG () across M-IgG -modified membranes, (A) without equilibration, and (B) after equilibration. The feed solution concentration was 2.65 nM in both of the permeating proteins. These data suggest that this antibody-modified nanotube-membranes at low ionic strength can be used for situations were the transport of an unwanted protein needs to be shut down, in order to improve selectivity. One potential application that this approach might be useful is in the sample complexity reduction in proteomic analysis. Since the analysis and characterization of 84

PAGE 85

complex mixtures of proteins are the central aims of proteomics. 195 Thus, it would make possible to deplete the flux of high abundance proteins across the antibody-modified membranes while at the same time it would allow improving the separation and detection of low abundance proteins by subsequent analytical techniques. Conclusions A detailed experimental study on antibody-functionalized nanotube membranes has been carried out, including the effect of surface modification, pore size, feed concentration, temperature, and ionic strength, both on the transport as on the functional behavior of the membranes. It was demonstrated that the type of immobilized antibody changes the functional behavior of the antibody-modified membranes and the transport properties of the permeating proteins across them. These membranes showed a Langmuirian shape plot characteristic of a typical behavior of facilitated transport membranes, when the effect of feed solution concentration on the fluxes of the permeating proteins was investigated. Therefore, the antibody-functionalized membranes showed here exhibit a type of behavior for facilitated-transport membranes during protein separations. Important parameters such as the binding capacities of the antibody-modified membranes and the binding constants for the permeating proteins to the antibody carriers confirmed that the permeating protein anti-M-IgG binds strongly to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier, likewise, the permeating protein anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-IgG binds weakly to this carrier. It was also demonstrated that by controlling the orientation of the immobilized antibody using protein G the potency affinity of the antibodies is effectively preserved. Through these studies was also demonstrated that by selecting a template membrane with an appropriated pore size diameters these antibody-modified nanotube membranes have the ability to be used with a combination of both size and affinity selectivity for protein separation. 85

PAGE 86

Additional studies on the effect of temperature on the fluxes and the binding constant demonstrated that the functional behavior of these antibody-modified membranes is governed for the changes in free energy before and after interaction of the permeating proteins with the carrier antibody, which was attributed to the changes in enthalpy and entropy of the system. Other important finding during this investigation was that by controlling the ionic strength of the transport solution, these antibody-modified nanotube membranes can be used for situations were the transport of an unwanted protein needs to be shut down, in order to improve selectivity. This latter approach might be useful in the sample complexity reduction in proteomic analysis. Since the analysis and characterization of complex mixtures of proteins are the central aims of proteomics. It is also noteworthy to mention the low pM detection limits achieved during these studies. In general, the ability to alter the surface biochemistries and thus tailor the membrane for specific analytes, promises the use of these membranes for the analysis of a wide range of proteins. 86

PAGE 87

CHAPTER 3 MODELING AND CHARACTERIZATION OF THE TRANSPORT PROPERTIES OF PROTEINS ACROSS ANTIBODY-FUNCTIONALIZED NANOTUBE MEMBRANES BY USING THE DUAL-MODE MODEL Introduction In the design and development of more effective membranes for membrane-based separations, the value of the transport mechanism must be considered in order to understand the different interactions that govern transport through the pores of the membranes. In addition, the transport mechanism, supported by mathematical models and experimental evidences provides a grater understanding of the underlying transport phenomena. The separation properties of molecules through membranes are based on the difference of permeation. Therefore, it is important to understand permeation behavior, not only to understand the permeation mechanism but also to elucidate the separation properties. In order to elucidate the transport properties across selective membranes a wide variety of models have been developed. 196 202 However, a model that explains the sorption and permeation mechanism is of fundamental importance to fully understand the transport properties in membrane separation. The well-known dual-mode model allows evaluation of both sorption and permeation mechanism, for those cases where the model applies. The dual-mode model was initially developed by Vieth and Sladek 203 and later modified by Paul and Koros 204 to describe the sorption and permeation of gas molecules in glassy polymeric films. In literature there are various modified dual-mode models that have also been explored. 182, 205 208 The model assumes that the membrane contains two distinct regions in which sorption can occur. 204 One region exhibits solubility of the molecule based on Henrys law and the second 87

PAGE 88

region has a Langmuir sorption isotherm. For permeation, it is assumed that each region is equilibrated and permeation is governed by a solution-diffusion mechanism. Figure 3-1.Dual-mode transport model. The dual-mode model has also been useful to explain membranes with higher transport selectivity obtained via a methodology called facilitated transport. 114, 115, 209 These models typically draw an analogy between the facilitated transport process and the dual-mode model for gas transport in glassy polymers. During facilitated transport, the enhanced transport of a specific permeate is mediated by a carrier immobilized within the membrane, plus the normal physical permeation pathway, since this carrier in the membrane reversibly binds the specific permeate molecule. In Figure 3-1 is shown a schematic illustration of the dual-mode model presented here. This chapter presents the results on the characterization of the transport properties of anti-M-IgG and anti-H-IgG across M-IgG modified membranes by using the dual-mode model. We have studied the effect of systematic chemical modification on the transport properties of silica nanotube membranes by functionalization with covalently attached-carrier-antibodies as a means of selectively improving transport properties of proteins. It is discussed how the presence of this carrier in the nanotube-functionalized membranes enhance the permeability of the transported anti-IgG. This enhancement in permeability at low concentration is explained with the dual-mode model and it is caused by a contribution of the diffusion due to the Langmuir mode 88

PAGE 89

species. Based on these results, the concept of modulated facilitated transport using antibody-modified nanotube membranes is proposed. Thus, depending on the affinity parameters, the transport in the carrier situation can be modulated in comparison to the situation without carrier modification where the rate of transport is only controlled by physical diffusion. The present study test the dual-mode model under circumstance where is possible to have modulated facilitated transport, and simulations and experimental results are discuses to validate this concept. Definition of Dual-Mode Model The dual-mode model is widely used to describe the sorption of gas molecules on glassy polymers 210, 211 Depending on their internal structure, the polymers are characterized as either glassy or rubbery. Sorption of gases to the rubbery state occurs by dissolution, while sorption to glassy state occurs by concurrent dissolution and hole-filling mechanisms. 212, 213 The sorption mechanism of gas molecules in glassy polymers has been understood in term of the dual-mode model. 203 208 The model is based on two types of sorption sites. One obeys the Henrys law dissolution and the second a Langmuir-type sorption. According to the dual-mode model, total sorption to surface membrane is the sum of sorption in the dissolution domain and sorption in the Langmuir-type sorption domain (Equation 3-1): 214 bpbpCpkCCCsitesDHD1 (3-1) By analogy but making a couple of changes, we have used the dual-mode model used for gas-transport to interpret transport of proteins in aqueous solution through antibody-modified nanotube membranes. Instead of reporting the concentration of penetrant in the membrane for a given gas pressure, the concentration is reported as a function of the feed protein concentration. The Henrys law coefficient, k D, was also adjusted and it was given in terms of partition 89

PAGE 90

coefficient k p Since, the sorption coefficient given in the Equation 3-1 is a gas-phase coefficient; in solution therefore the sorption coefficient should be considered as a liquid-phase coefficient. The adapted dual-mode model is shown in Equation 3-2. feedfeedsitesfeedpHDbCbCCCkCCC1 (3-2) Where C is the total protein concentration in the membrane, C D is the protein concentration based on Henrys law sorption, C H is the protein concentration based on Langmuir sorption. Whereas, k p is defined as the partition coefficient, b and C sites are the Langmuir affinity constant and maximum capacity parameters, respectively. The Langmuir affinity constant, b, characterizes the sorption affinity for a particular protein-antibody membrane system, and C sites is used to measure the amount of antibody immobilized in the membrane. The three dual-mode sorption parameters, k p C Sites and b, can be determined through the experimental evaluation of protein-adsorption isotherms. The estimation for C Sites and b, sorption parameters can be accomplished by analyzing the data according to the monolayer Langmuir model (Equation 3-3): 167 FreeFreeSitesBoundbCbCCC1 (3-3) Where C Bound is the specific concentration of protein bound to the carriers in the membrane, which is related to each specific concentration of free protein (C free ) in the equilibration solution. The other of the three sorption parameters contained in the dual-mode model is the partition coefficient. Through the analysis of the boundary conditions of the Langmuir-mode model, an optimum parameter value for the partition coefficient could be estimated. 90

PAGE 91

Thus, at the high concentration range, the Langmuir model approaches to a constant value that is equal to C Sites so no more sorption will take place when all the sites are filled. On the other hand, at very low concentrations, the Langmuir model obeys the correct thermodynamic boundary condition of Henrys law over infinitely dilute concentration range. 215, 216 In the limit of zero concentration the Langmuir model leads to the following equation which satisfies the condition of a linear isotherm (Equation 3-4): FeedDFeedSitesCCkbCCC 0lim (3-4) Hence, a good estimation of Henrys law coefficient, k D, may be written as the product of C Sites and b. Consequently, this k D parameter can be used for the calculation of the partition coefficient (k p ), which is defined as the product of k D times the ratio of volumes between volume inside the membrane (V Membrane ) and volume in the free equilibration solution (V Free Solution ) (Equation 3-5). SolutionFreeMembraneDpVVkk (3-5) It should be pointed out that for linear isotherms, the Henrys law coefficient, k D, can be determined from the slope, while for dual-mode adsorption isotherms, it can be determined from the initial slope at infinite dilute concentration. The partition coefficients can also be calculated using Equation 3-6, which expresses the ratio of the intramembrane concentration to external concentration at equilibrium. SolutionFreeSolutionFreeMembraneMembranepCVCVk (3-6) All this dual-mode sorption model theory explained until here can be validated using linearization methods for the Equation 3-2. Several linear plots of the dual-mode sorption model have been studied, 217 and for instance, the equation can be rearranged and transformed into the Equation 3-7. 91

PAGE 92

FeedSitesSitesFeedpFeedCCbCCkCC11 (3-7) A plot of C Feed / (C k p C Feed ) against C Feed is expected to be linear, where the Langmuir sorption parameters, C sites and b, can be evaluated from the slope and intercept of the straight line, respectively. 217 Now, if the solute-adsorption isotherms are represented by the dual-mode model, as shown above, the permeation mechanism can also be described in terms of the dual-mode model. The permeation mechanism using the dual-mode model assumes partial immobilization of the Langmuir-mode species. 204 Therefore, independent dual diffusion coefficients are given for each sorption mode. One diffusion coefficient (D D ) is used for the fraction of the solute dissolved in the membrane according to Henrys law expression and a second (D H ) for the fraction of the solute contained in the Langmuir-type sorption mode. The expression for the permeation flux through the membrane is derived by Ficks law, and it can be written in the form given in Equation 3-8. xCCFKDxCDxCDJDDDHHDD211 (3-8) F corresponds to the mobile fraction of the Langmuir species (Equation 3-9). DHDDF (3-9) Where and K are constants that contain the three sorption parameters (Equation 3-10 and Equation 3-11). pSiteskbCK (3-10); pkb (3-11) The mathematical solution that Koros and Paul found for the Equation 3-8 is known as the partial immobilization model. 204 The model assumes that there is always local equilibrium between the two sorption modes, thus, C H can be written in terms of C D In addition, the 92

PAGE 93

diffusion coefficient is assumed constant so it does not depend on concentration. Besides, the model splits the total concentration into two parts, one a mobile part with a diffusion coefficient D D and concentration C m and the second remainder part (C C m ), corresponds to the part that is totally immobilized. In other words, the concentration associated with C D as well as a fraction F of that associated with C H has finite mobility while the remaining fraction, 1-F, of C H has zero mobility. Therefore, the concentration for the mobile part can be written as (Equation 3-12): HDmFCCC (3-12) The flux for this mobile part may be written as is shown in Equation 3-13. xCDJmD (3-13) Whereas the equation of continuity (Ficks second law), represented by the dual-mode model for a solute transported through the membrane is given by the Equation 3-14. xJtCCtCHD)( (3-14) The equation of continuity is nothing more that the mass balance for the system, which relates the accumulation of solute in the porous membrane to the changes in flux at any time and any direction of the flux. In other words it is the rate of solute balance. Thus, the difference between the rate of solute that comes in to the membrane, and the rate of solute out, is equal to the rate of solute accumulation. The insertion of the Equation 3-13 into the Equation 3-14, results in Equation 3-15. xxCDtCCmDHD)()( (3-15) By assuming that the diffusion coefficient does not depend on concentration, and inserting the Equation 3-12 into the mass balance equation given by the Equation 3-15, the equation becomes (Equation 3-16) 22)()( x FCCDtCCHDDHD (3-16) 93

PAGE 94

Finally, by setting all the assumptions and writing the concentration of C H in terms of C D (KC D / 1+C D ), a numerical solution of the partial immobilization model yields to an expression for the permeability coefficient given by Equation 3-17. 204 FeedDpbCFKDkP11 (3-17) The use of the model requires the determination of both D D and F. This can be accomplished through the experimental evaluation of the permeability coefficients; so, optimum parameters for the diffusion coefficients can be determined. Throughout every experimental permeation, measurements of the flux per unit area, J, (moles cm -2 sec -1 ), is obtained. Therefore, if a number of permeation experiments are carried out at different feed concentration using membranes with homogeneous thickness, L, the steady-state permeability coefficient for a specific feed concentration can be calculated using the following relationship showed in Equation 3-18. 218, 219 FeedCJLP (3-18) This procedure would also allow comparing the experimental permeability plots with those calculated obtained ones. In order to explain permeation, the dual-mode model assumes that the permeation is governed via a solution-diffusion mechanism. If the permeation of the protein in the membrane is controlled by a solution-diffusion mechanism, the permeability can be expressed as the product of diffusivity, D, and solubility, S as is shown in Equation 3-19. DSP (3-19) Although the solubility of a permeating protein in the membrane can be also expressed in term of the dual-mode sorption parameters by (Equation 3-20): 210, 211 FeedSitespHDFeedbCbCkSSCCS1 (3-20) 94

PAGE 95

Where S is the solubility of permeating protein, and the terms S D and S H corresponds to the solubility values based on Henrys law and Langmuir-type sorption, respectively. If the diffusion coefficient does not depend on concentration, a plot of permeability against solubility is expected to be linear, with a slope equals to the diffusion coefficient (Equation 3-19). Through this correlation between P and S experimentally the diffusion coefficients can be estimated. With the aim of determining the F parameter, the experimental permeability data can be plotted against (1 + bC Feed ) -1 A value for F can be estimated from the slope of this line. 211 Additionally, the diffusion coefficient D D might be evaluated from the intercept as well, but only if the plot is straight line with a very high linear correlation coefficient (Equation 3-21). FeedDSitesDpbCbFDCDkP1 (3-21) In fact, the dual-mode model makes feasible the evaluation of the sorption and permeation mechanism. This is of fundamental importance to fully understand the transport properties in membrane separation. Hence, using the dual-mode model, a comprehensible explanation of the transport mechanism might be accomplished through a detailed elucidation of the transport properties such as sorption, diffusion and permeability. Results and Discussion Simulations of Permeability vs. Feed Concentration Based on the Dual-Mode Model The permeability is the product between diffusivity and solubility as is described in Equation 3-19. The permeability dependence in terms of the dual-mode model is not only related with the concentration of solute molecules in the feed solution, but also with the different properties and interactions between solute-membrane. Thus, the permeation of solute molecules 95

PAGE 96

through carrier-modified membranes, which is governed by both the Henry mode and the Langmuir mode, can be modulated depending on the affinity parameters between solute-carrier. In order to understand how the dual-mode model works, simulations of the model were initially carried out (Figure 3-2). The simulation analysis was based on the assumption of competitive permeation studies for two different solute species transported across carrier-modified nanoporous membranes. With the aim of simplifying the analysis an idealized case was compared. Thus, one of the species the spectator or nonbinding molecule (NBM) is assumed to have no interaction with the carriers in membrane and therefore permeates the membrane by only physical diffusion. On the other hand, the second specie, the binding molecule (BM), permeates the membrane through a dual-mode model mechanism. The different dual-mode parameters used for the simulations are listed in Table 3-1. Figure 3-2.Permeability plot simulations for the transport of (a) BM and (b) NBM across carrier-modified membranes. The parameters are summarized in Table 3-1. By simple visual analysis of the simulations showed in Figure 3-2, we can notice that for a permeation mechanism that follows the dual-mode model, the permeability is highly dependent on the concentration, while for a mechanism based on physical diffusion only, the permeability is totally independent of the concentration. The permeability in terms of the dual-mode model 96

PAGE 97

decreases with increasing feed concentration, which is concave to the concentration axis as can be seen in the plot (a) in Figure 3-2. Table 3-1.Dual-mode model parameters used for the permeability plot simulations Permeant specie k p D (cm 2 /sec) C sites (moles/cm 3 ) b (cm 3 /moles) Nonbinding Molecule (NBM) 6 x 10 -3 9 x 10 -8 Binding Molecule (BM) 6 x 10 -3 9 x 10 -8 1.5 x 10 -10 1 x 10 11 Figure 3-3 shows simulations for the permeability versus feed concentration, illustrating the effect of the F parameter in dual-mode model mechanism. Figure 3-3 compares three simulation plots for the case of the BM situation, which correspond to three different values of F. The term F is associated with the mobile fraction of the Langmuir species as was stated before, and its magnitude is equal to unity or zero, when the model assumes no immobilization or total immobilization of the Langmuir species, respectively. Figure 3-3.Permeability plot simulations for the transport of BM across carrier-modified membranes illustrating the effect of the term F on the dual-mode model. (a) F = 1, (b) F = 0.5, and (c) F = 0. The other parameters are summarized in Table 3-1. 97

PAGE 98

Hence, the permeability for a solute molecule that permeates a membrane through a dual-mode mechanism depends also on the magnitude of F. For the BM situation, the permeability is highly dependent on the feed concentration as F gets closer to unity (plot a, in Figure 3-3). Consequently, as the value of F decreases or the degree of immobilization increases, the permeability becomes less dependent on the feed concentration (plot b, in Figure 3-3). When there is total immobilization, when F is equal to zero, the permeability is independent of the concentration, and so the BM dissolves and diffuses through the membrane via Henrys mode dissolution only (plot c, in Figure 3-3). It should be pointed out that regardless of the value of F, the permeability decreases gradually with the increase of feed concentration until a steady level that reach the value of k p D D The above simulations in Figures 3-2 and 3-3 suggested that when no immobilization of the Langmuir species is achieved, the permeability across carrier-modified membranes is always enhanced compared with the physical diffusion case. Therefore, the transport through carrier-modified membranes that fulfill the dual-mode model mechanism is always facilitated when the Langmuir species no experience immobilization. The concentration dependence of the permeability for the dual-mode model can also be analyzed through the evaluation of the limit boundaries conditions in concentration for the equation of permeability (Equation 3-17). At the high limit concentration range the equation for permeability becomes equal to the product between k p times D, this is due to the behavior of the factor [1 + FK / (1 + b C Feed )], which becomes closer to unity as the concentration increases. At the low limit concentration, the highest permeability value is reached when the denominator term, (1 + b C Feed ), becomes unity. As the latter gets closer to one, there is a range in concentrations where the variations in permeability are less noticeable, and the permeability 98

PAGE 99

becomes less dependent on the concentration. This region resembles to the time lag region for gases in glassy polymers. 181, 182 In general, as stated by Koros and Paul the limits of the dual-mode model are confined to two situations. 204 One is the limit where total immobilization occurs, and only the Henrys mode species contributes to the permeability. The other limit is a situation with a nonlinear isotherm with no distinctions for sorption mechanism or transport behavior between any of the adsorbed solutes. However, the concentration-dependence of the permeability is a consequence of the shape of the adsorption isotherm, and therefore a result of the binding affinity strength between Langmuir permeating species and Langmuir binding sites. In order to demonstrate the concentration dependence on the permeability as consequence of the shape of the adsorption isotherm, permeability plot simulations were carried out comparing two different sorption mechanisms via two different binding affinities (Figure 3-4). The plot (a) in Figure 3-4 clearly indicates that a carrier-modified membrane with stronger affinity for the solute molecules displays higher concentration-dependence as a result of the sorption isotherm. So, the concentration dependence of permeability is determined by the strength of the binding affinity for the solute-carrier complex in the membrane. However, the total permeability is highly dependent on the diffusion coefficient. It is worth mentioning that by using simulations was verified that the order of influence of the dual-mode model parameters in the permeability was D > b > C sites >>> k p respectively. According to these results, the dual-mode model permeation analysis is a useful model to explain transport across carrier-modified membranes. The analysis showed until here demonstrated how the fluctuations in permeability are a result of the strength of the sorption mechanism. Therefore, the differences in permeability and the differences in permeability 99

PAGE 100

concentration-dependence can be used to elucidate the separation properties for the solute molecules that permeate a carrier-modified membrane. For the reason that differences in permeability are a consequence of the sorption mechanism, the separation properties of one solute might differ from that of another, depending on the affinity solute-carrier interaction. Figure 3-4.Permeability plot simulations for the transport of BM across carrier-modified membranes showing the effect of the binding affinity on the permeability. (a) b = 1 x 10 11 cm 3 /moles, and (b) b = 3 x 10 10 cm 3 /moles. The other parameters are summarized in Table 3-1. Experimental Description of the Dual-Mode Model Effect of feed concentration on the flux A number of competitive permeation experiments were carried out at different feed concentration, ranging from 2.65 x 10 -12 to 5.0 x 10 -10 moles cm -3 using M-IgG modified membranes (Figure 3-5). Throughout every experimental permeation measurements, the flux per unit area (moles cm -2 sec -1 ) was obtained, using 0.31 cm 2 as the area of membrane exposed to the contacting solution phases. Figure 3-5 shows the studies of the effect of feed concentration on the flux during competitive permeation experiments of anti-H-IgG and anti-M-IgG across M-IgG-modified membranes. The flux for anti-H-IgG was at all times higher than the flux for anti-M-IgG. The linear dynamic range for anti-M-IgG was 3 times lower than for anti-H-IgG, as a consequence of 100

PAGE 101

a stronger binding affinity with the M-IgG carrier attached in the membrane. These plots show a characteristic Langmuirian shapes, demonstrating that the immobilized M-IgG indeed interacts with the permeating proteins, acting as a carrier during the permeation process. 9, 10, 109 Figure 3-5.Flux versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. The feed solution concentration was equimolar in both of the permeating proteins. Experimental permeabilities as function of the feed concentration In order to conduct fundamental studies of any membrane process, a well defined membrane system is required. One of the worthy characteristics of the alumina template membranes used in these studies is their well defined and characterized structure. Thus, it is possible compare theoretical models with the experimental results as is demonstrated in this dissertation. Because the thickness of the membranes and feed concentration for each permeation experiment were known, the Equation 3-18 for permeability was used to calculate the experimental permeability coefficient of anti-M-IgG and anti-H-IgG across M-IgG modified membranes (Figure 3-6). Figure 3-6 shows the plots for the experimental permeability versus feed concentration. It was observed that the permeability decreases as the concentration increases both for anti-H-IgG and for anti-M-IgG across M-IgG modified membranes. This diminution in permeability may be understood in terms of the 101

PAGE 102

dual-mode model, 204 206, 211, 220 since the permeability in terms of the dual-mode model decreases with increasing feed concentration. 221 Figure 3-6. Experimental obtained permeability versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. At first sight from the data of concentration dependence of the experimental permeability in Figure 3-6, we might anticipate that the transport mechanism for these two proteins across M-IgG modified membranes approaches to the dual mode model mechanism. In order to prove this transport mechanism for these two proteins across the M-IgG modified membranes, an evaluation of the sorption and permeation parameters were carried out, and they are discussed next. Experimental evaluation of the dual-mode model parameters The mathematical solution that Koros and Paul found for the permeability coefficients based on the dual mode model is known as the partial immobilization model (Equation 3-17). 204 The use of the model requires the determination of both D D and F, which were evaluated experimentally by using the permeability coefficients in Figure 3-6. Additionally, to use the model is a prerequisite to have the information of the three dual-mode sorption parameters, k p C Sites and b, which were also determined experimentally via protein-adsorption isotherms. 102

PAGE 103

Experimental evaluation of C sites and b These Langmuir sorption parameters were determined experimentally using adsorption isotherms. The relationship that exists between concentrations of protein in the porous membrane isothermally in equilibrium with those in the feed solution is called adsorption isotherm. In order to assess the equilibration time for the sorption experiments, a sorption dynamic study was initially carried out (Figure 3-7). For this, a solution that was 21.2 nM of anti-M-IgG labeled with Alexa 488 was presented to M-IgG modified membranes. The amount of anti-M-IgG adsorbed was investigated observing the changes in fluorescence intensity as a function of time. Figure 3-7.Curve of adsorption dynamics for anti-M-IgG () in a M-IgG modified membrane. Inset shows the changes of fluorescence signal with time. As can be seen from Figure 3-7, the amount of anti-M-IgG adsorbed in the M-IgG modified membrane increases rapidly during the first 3 hours. After that, the rate of adsorption decreases and the system reaches a plateau of apparently equilibrium. All the adsorption isotherms were carried out with M-IgG modified membranes at 23 C in stirred U-tube permeation cells using the same buffer conditions used for the permeation studies, via solutions containing single concentrations of anti-M-IgG or anti-H-IgG fluorescent labeled 103

PAGE 104

with Alexa 488. Data were analyzed according to the monolayer Langmuir model (Equation 3-3). 167 By using the experimental values of C free and C bound in moles per cm 3 in Equation 3-3 the parameter C sites was calculated in moles per cm 3 while b in cm 3 per moles. Figure 3-8 shows and compares adsorption isotherms for anti-M-IgG and anti-H-IgG carried out with M-IgG modified membranes; also the adsorption isotherm for anti-M-IgG using unmodified nanotube membranes is showed. Table 3-2 summarizes the corresponding binding constant and protein binding capacities, as calculated by fitting the data using nonlinear regression methods. Figure 3-8. Adsorption isotherms for anti-M-IgG (), and anti-H-IgG () binding on M-IgG modified membranes. Also the adsorption isotherm for anti-M-IgG () using a control membrane is shown. Figure 3-8 shows the adsorption isotherms obtained from the fluorescence intensity of the wet membranes exposed to different protein concentrations. For the unmodified silica nanotube membranes the adsorption of anti-M-IgG labeled with Alexa 488 was negligible in comparison with the antibody-modified membranes. It was also observed, that the adsorption for the unmodified membranes increases proportionally with the concentration of fluorescent-labeled 104

PAGE 105

antibody, in the entire range of concentration used for the saturation binding experiments (from 1.0 x 10 -12 to 2.5 x 10 -11 moles cm -3 ), obeying clearly a Henrys linear isotherm. The linear shape of the unmodified isotherm also indicates a very low affinity for the anti-M-IgG. Thus, these results also suggested that the silica nanotube membranes exhibit low protein binding (non-specific antibody adsorption) to the surface membrane, and minimal autofluorescence. On the other hand, the M-IgG modified membranes exhibited clearly nonlinear isotherms and they bound more protein than the unmodified ones. When anti-H-IgG labeled with Alexa 488 was presented to M-IgG modified membranes the adsorption isotherm was nonlinear. This effect is due mainly to the existence of cross-reactivity between the permeating anti-H-IgG and the attached M-IgG. That kind of cross-reactivity of IgG has been also seen in the literature. 169 172 By fitting the experimental adsorption isotherm for anti-H-IgG, the existence of one binding site for the antibody was demonstrated. A binding affinity constant equal to (2.90 1.35) x 10 10 cm 3 moles -1 was obtained. The amount of protein bound to the membranes exposed to anti-H-IgG was almost twice compared to the unmodified silica nanotubes ones (Table 3-2). In contrast to the anti-H-IgG sorption isotherm case, when anti-M-IgG labeled with Alexa 488 was presented to the M-IgG modified membranes the protein bound to the membrane was approximately 20 times higher than the anti-H-IgG situation. These data confirmed the successful modification of the membranes, and the remarkable selective attribute of the antibody-modified membranes. The fitting process for the M-IgG modified sorption isotherms exposed to anti-M-IgG indicated the existence of two binding sites in the membrane for the permeating protein (Figure 3-8). This is probably due to the spatial orientation of the M-IgG molecules immobilized on the 105

PAGE 106

pore surface of the membrane. Therefore, the binding site with the stronger binding constant, (4.46 0.91) x 10 11 cm 3 moles -1 was attributed to the concentration of well-oriented immobilized IgG molecules. The other site with an affinity constant of (5.27 1.58) x 10 10 cm 3 moles -1 was ascribed to the random-oriented concentration. Thus, anti-M-IgG first binds to the site of well-oriented immobilized M-IgG. After a fixed amount of anti-M-IgG is adsorbed the random immobilized M-IgG become available as cooperative binding sites. Figure 3-9.Adsorption isotherm for anti-M-IgG () binding on M-IgG-PG modified membranes. Additional confirmation for this conclusion about two binding sites with different spatial orientation was obtained through adsorption isotherms of anti-M-IgG using M-IgG well-oriented antibody-modified membranes via protein-G (PG) immobilization (Figure 3-9). The well-oriented immobilization of M-IgG was achieved through silica nanotube membranes already modified with PG. This PG was covalently attached via the Schiffs base reaction to the walls of aldehyde modified nanotube membranes. The adsorption isotherm with M-IgG-PG modified membranes revealed the existence of only one binding site by using this well-oriented antibody-immobilization (Figure 3-9). For this specific case, the value of the binding affinity constant was (3.77 0.53) x 10 11 cm 3 moles -1 (Table 3-2). 106

PAGE 107

It should be pointed out that the antibodies used in these studies were all polyclonal antibodies. Therefore, the found binding affinity constants are indeed an average of the pool of the different clonally related subpopulations of antibodies existing in the sample solution. 222, 223 Table 3-2.Thermodynamic parameters from adsorption isotherms of anti-M-IgG and anti-H-IgG on M-IgG modified membranes, according to the Langmuir monolayer model Adsorption Isotherms C sites1 (moles/cm 3 ) b 1 (cm 3 /moles) C sites2 (moles/cm 3 ) b 2 (cm 3 /moles) R anti-H-IgG / M-IgG (7.24 1.33) x 10 -11 (2.90 1.35) x 10 10 0.983 anti-M-IgG / M-IgG (1.16 0.19) x 10 -10 (4.46 0.91) x 10 11 (1.30 0.39) x 10 -9 (5.27 1.58) x 10 10 0.928 anti-M-IgG /M-IgG-PG (3.49 0.14) x 10 -10 (3.77 0.53) x 10 11 0.992 Experimental evaluation of the partition coefficients k p For definition, partition coefficients provide a thermodynamic measurement of the solubility of the permeating proteins in the membrane. Therefore, considering this thermodynamic fact, proteins with higher diffusion coefficient should have lower partition coefficients. Contrarily, an increase of the partition coefficient increases the resident time of the permeating proteins in the membrane. The partition coefficient is one the other three sorption parameters contained in the dual-mode model. The experimental evaluation of the partition coefficients across M-IgG modified membranes for the proteins studied here were investigated using three different approaches. Table 3-3 compiles the results for these studies. The first approach for the analysis of k p was based on the examination of the boundary conditions of the Langmuir-mode in the limit of zero concentration, that determine that the Henrys law coefficient (k D ) may be written as the product of C Sites times b (Equation 3-4). 215, 216 107

PAGE 108

Since the values of C Sites and b were already estimated for both of the permeating proteins studied here as was shown above. The partition coefficients calculated using the Equation 3-5 were 7.24 x 10 -3 and 2.94 x 10 -4 for anti-M-IgG and for anti-H-IgG, respectively. Table 3-3.The k p determined by various approaches Method anti-M-IgG Anti-H-IgG Equations 3-4 and 3-5 7.24 x 10 -3 2.94 x 10 -4 Slopes Figure 3-10 6.14 x 10 -3 2.77 x 10 -4 Equation 3-6 (5.84 1.06) x 10 -3 (2.79 0.69) x 10 -4 It is important to mention that for linear isotherms, the Henrys law coefficient, k D, can be determined from the slope of the isotherm plot, while for dual-mode sorption isotherms, it can be determined from the initial slope at infinite dilute concentration (Figure 3-10). To prove this assertion the fitting plots of the adsorption isotherms for anti-M-IgG and anti-H-IgG were extrapolated to a very dilute concentration range (from 0 to 6 x 10 -13 moles cm -3 ), and the respective slopes from the linear plots were estimated, these results are shown in Figure 3-10. Figure 3-10. Extrapolations of the adsorption isotherms at infinite dilute concentration for anti-M-IgG (), and anti-H-IgG () in M-IgG modified membranes. 108

PAGE 109

The k D obtained for anti-M-IgG and anti-H-IgG from the slopes in Figure 3-10 were 43.85 and 1.98, respectively. Likewise, using Equation 3-5 a partition coefficient of 6.14 x 10 -3 for anti-M-IgG and 2.77 x 10 -4 for anti-H-IgG were found. The partition coefficients were also calculated using Equation 3-6, which expresses the ratio of the intramembrane concentration to external concentration at equilibrium. The partition coefficients estimated with this equation were comparable to those calculated by using the direct method via k D As is shown in Table 3-3, the partition coefficients in this case were (5.84 1.06) x 10 -3 and (2.79 0.69) x 10 -4 for anti-M-IgG and anti-H-IgG, respectively. In order to validate these partition coefficients, dual-mode model permeability plot simulations were carried out (Figure 3-11). For these simulations, four different values for the partition coefficients were assigned by keeping the rest of the dual-mode model parameters constant. The results shown in Figure 3-11 demonstrated that variations in the partition coefficients have a negligible effect in the permeability coefficients when the binding affinity is fairly strong as the studies showed here. Figure 3-11.Permeability plot simulations illustrating the effect of the partition coefficient on the dual-mode model. Simulation parameters: D = 5 x 10 -8 cm 2 /sec, b = 9 x 10 10 cm 3 /moles, C sites = 3.5 x 10 -8 moles/cm 3 and k p were respectively, 1 (), 0.1 (), 0.01 (), and 0.001(). 109

PAGE 110

Experimental evaluation of the diffusion coefficients and solubility The most direct method of experimental determination of S and D requires steady-state permeation data, complemented with equilibrium sorption data. 199 Therefore, effective values of S, D, and hence P characterized for a membrane-penetrant system that has been restricted to equilibrium sorption and steady-state permeation conditions represent appropriate averages of the transport parameter of the system. 224 To use the dual-mode model requires knowing D and S. This was accomplished experimentally through an evaluation of the permeability and solubility in terms of the dual-mode model as will be shown next. As was state before, the solubility for a penetrant protein in the membrane can be expressed in term of the dual-mode sorption parameters by the Equation 3-20. 210 For the reason that the three sorption parameters required for this calculation were already established, an experimental study of the concentration dependence of the solubility was conducted (Figure 3-12). Figure 3-12 shows the concentration-dependence of the solubility in term of the dual-mode model for anti-M-IgG and anti-H-IgG in M-IgG modified membranes. The concentration-dependence of the solubility in term of the dual model is a consequence of the shape of the adsorption isotherm. Hence, from Figure 3-12 M-IgG modified membranes possess higher affinity for anti-M-IgG than for anti-H-IgG, as was demonstrated before (Table 3-2). These data also indicate where a protein molecule prefers to dissolve, whether in the Langmuir binding sites or in the membrane matrix based on Henrys law. According to Figure 3-12, anti-M-IgG molecules would rather dissolve in the Langmuir sorption mode (upper plot), while anti-H-IgG molecules in the Henrys mode (lower plot). It was also observed that regardless of the protein at low concentrations the Langmuir mode contributes the most to the total solubility, which is probably due to the strong affinity exerted for that fraction of well-oriented immobilized carriers 110

PAGE 111

in the membrane. Therefore, at low concentration the solubility of the proteins was highly increased by the presence of the M-IgG carriers, so those can be used as facilitated partition domains of the transport. In addition, the data revealed that as the concentration increases, the solubility decreases and reaches a plateau where the solubility in the membrane is only controlled by the solubility in the Henrys mode. Figure 3-12. Concentration-dependence of the solubility in term of the dual-mode model for anti-M-IgG (), and anti-H-IgG () across M-IgG modified membranes. As was described before in Equation 3-19, the permeability is the product between diffusion and solubility. Since both the concentration dependence of the permeability and solubility were experimentally determined, they were used to estimate the effective values for the diffusion coefficients of anti-M-IgG and anti-H-IgG across M-IgG modified membranes. By using this correlation method a plot of permeability against solubility is expected to be linear, with a slope equals to the diffusion coefficient (Equation 3-19). Figure 3-13 shows the correlation plots of permeability and solubility based on the concentration dependence for the anti-H-IgG (Figure 3-13A) and the anti-M-IgG (Figure 3-13B) across M-IgG modified membranes. These plots in Figure 3-13 were carried out by using the experimental concentration-dependence data from Figure 3-6 and Figure 3-12 both for the 111

PAGE 112

permeability and solubility of anti-H-IgG and anti-M-IgG in M-IgG modified membranes. The plots in Figure 3-13 demonstrated a good correlation between permeability and solubility. The slopes of these correlation plots will provide the diffusion coefficients of the protein transported across the M-IgG modified membranes, if the diffusion coefficients are concentration-independent as the dual-mode model assumes. Figure 3-13.Correlation plots between permeability and solubility based on the concentration dependence for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified membranes. However, the lines in Figure 3-13 were not linear in the total range of concentration, but instead they showed two straight-line slope segments. The segment with negative slope corresponds to the region at very low feed concentration. As mentioned above, in this region (low feed concentration) the Langmuir sorption sites contributed the most to the solubility of these proteins as was demonstrated in Figure 3-12. Therefore, the lower diffusion coefficients at low concentration are attributed to the presence of slow-moving adsorbed proteins in this Langmuir sorption mode. A more detailed analysis of this conclusion will be demonstrated below in the discussion of fitting experimental data with the theoretical dual-mode model. On the other hand, the segment with positive slope was used to estimate the diffusion coefficients. By fitting of the positive linear segments in 112

PAGE 113

Figures 3-13A and 3-13B, diffusion coefficients of (7.61 0.27) x 10 -8 cm 2 sec -1 and (1.02 0.06) x 10 -9 cm 2 sec -1 were found for anti-H-IgG and anti-M-IgG, respectively. These obtained values compare favorably with other IgG diffusion coefficients found in the literature using different porous materials. 225 229 A lower diffusion coefficient of 4.09 x 10 -10 cm 2 sec -1 was reported for a bare nanoporous alumina with a pore size of 75 nm and membrane thickness of 70 nm. 228 A diffusion coefficient of 2.71 x 10 -9 cm 2 sec -1 across nanoporous silicon with a pore size of 49 nm and membrane thickness of 6 m was also reported. 229 It should be mentioned that the most general equation used to obtain a rough estimate of the diffusion coefficients in liquids is the Stokes-Einstein equation. 230 233 This equation is derived for a sphere with hydrodynamic radius R H moving in a fluid (Equation 3-22). HBRTkD60 (3-22) Where, k B is the Boltzmann constant, T the temperature, the solvent viscosity, and R H the hydrodynamic radius of the molecule, respectively. This Stokes-Einstein relationship described that the diffusion process is basically controlled by the size of the molecules when they are in free solution and in absence of any interactions with other molecules. However, it is known that not only the size but also the intermolecular interaction and the molecular shape affect the magnitude of D 0 234, 235 Therefore, we can concluded that the differences in the diffusion coefficients for the anti-H-IgG and anti-M-IgG across M-IgG modified membranes are mainly due to the changes in the molecular energy of the proteins as the molecular recognition interaction takes place. It should be pointed out that the estimated diffusion coefficients were significantly lower than the reported free diffusion coefficient for IgG in water (3.85 x 10 -7 cm 2 sec -1 ). 228, 229 The reasons for the decreased diffusion coefficients found in these studies compared 113

PAGE 114

with those in free solution are probably due to the intrinsic membrane resistance itself, together with the changes in molecular energy as was explained above. The above results show clearly that the correlation analysis between permeability and solubility is a direct method to estimate diffusion coefficients when equilibrium sorption and steady-state permeation data are experimentally available. In addition, the two straight-line slope segments observed through this correlation made also possible to distinguish the heterogeneity in the spatial orientation of the immobilized carrier in the membrane which demonstrated to have minor effects in the functional behavior of the membrane (Figure 3-4). Fitting Experimental Data with the Theoretical Dual-Mode Model Figure 3-14 compares the experimental and calculated concentration dependence of permeability for anti-M-IgG and anti-H-IgG across M-IgG modified membranes. Figure 3-14. Experimental and calculated obtained permeability versus feed concentration for anti-M-IgG () and anti-H-IgG () across M-IgG modified membrane. 114

PAGE 115

At first glance observing the plots in Figure 3-14, one can conclude that according to the dual-mode model there is not total immobilization on the Langmuir-mode sites for these two permeating proteins across M-IgG modified membranes. If total immobilization occurs in the Langmuir-mode sites (F=0) as the dual-mode model explain, the permeability become independent of the concentration. These studies demonstrated a high concentration dependence on permeability both for anti-H-IgG and anti-M-IgG. These also demonstrated that the M-IgG carriers in the membrane indeed enhance the permeability at low concentration for both of the permeating proteins as shown in Figure 3-14. It was found that the permeability enhancement for these two proteins was only up to a certain threshold of the binding affinity and appeared to depend strongly on the concentration. This enhancement in permeability at low concentration is caused by a contribution of the diffusion due to the Langmuir mode species. 204, 211 As a consequence of its lower binding affinity, the permeabilities for anti-H-IgG were higher than those for anti-M-IgG. At the beginning we set a hypothesis that the differences in permeation for anti-H-IgG and anti-M-IgG observed could be due to a sort of antifacilitated transport. This was initially considered based on our previous works, 9, 10 which showed that higher fluxes were always observed for the molecule that binds stronger to the carrier in the membrane. The mechanism for the facile kinetic conquered, in our previous works, was ascribed by the sequential binding and unbinding events of the transported permeating species with the carrier in the membrane. This behavior might be mathematically rationalized in a more detailed manner by the model described here. As indicated in Figure 3-14, the experimental and calculated data were in good agreement at concentrations higher than 2.0 x 10 -11 moles per cm 3 Nevertheless, at very low concentration 115

PAGE 116

that agreement was missing. The explanation for this behavior was that the dual-mode model assumes a constant diffusion coefficient, which in practice does not happen. Therefore, the model is not able to discriminate a different regimen for the concentration dependence on the diffusion coefficient as the case at low concentration observed here. As was said above, at low concentration the adsorption of these two proteins in M-IgG modified membranes is dominated basically for the binding sites with stronger binding affinity (Figure 3-12). Then, these Langmuir-mode sites impart higher sorption rates for the permeating proteins, slowing down the diffusion coefficients. Consequently this is also reflected in the permeability. Thus, at low concentration the experimental permeabilities were lower than the calculated permeabilities. This reduction in permeability is mostly due to a combined effect of both the lower diffusion coefficients, and concentration-dependence of diffusion coefficients at this lower region in concentration (Figure 3-14). Additional evidence for this conclusion was also demonstrated by observing the concentration dependence on the experimentally obtained diffusion coefficients in Figure 3-15. As can be seen from Figure 3-15, lower diffusion coefficients were obtained at very low concentration. Thus, these results demonstrate highly concentration-dependent diffusion coefficients at low concentration as a result of the two types of binding sites present in the M-IgG modified membranes (well-oriented and random). Lower permeabilities were also observed for anti-H-IgG at the low concentration region (Figure 3-14). This effect was explained taking into consideration the differences in diffusion coefficients between anti-H-IgG and anti-M-IgG. Due to anti-H-IgG diffuses faster and thereby reaches the binding sites inside the pores faster; the concentration of anti-H-IgG is initially greater in the membrane than that of anti-M-IgG. Thus, anti-H-IgG initially at low concentration 116

PAGE 117

binds more sites but those are displaced when the anti-M-IgG molecules catch up to them. Additional confirmation for this conclusion is obtained by comparing the binding capacities for these two proteins (Table 3-2). Figure 3-15.Experimental obtained diffusion coefficients versus feed concentration for (A) anti-H-IgG and (B) anti-M-IgG across M-IgG modified membrane. At the region of low limit concentration it was also observed that as the concentration increases the permeability increases as well (Figure 3-14). It was deduced that between the two set of binding sites (well-oriented and random) present in M-IgG membranes, there is a transition stage in concentration before the random binding sites with a higher binding capacity become in a positive cooperative binding sites. This transition involves a mutual cooperation of these two binding sites in the membrane. This effect was observed at concentrations lower than 2.0 x 10 -11 moles per cm 3 (Figure 3-14). These results are in favorable agreement with the positive cooperative contribution observed in the adsorption isotherms in Figure 3-8. Besides as mentioned previously, the Langmuir binding sites were the dominant factor in determining the total solubility of anti-H-IgG and ant-M-IgG in the M-IgG modified membranes. And regardless of the protein these Langmuir sorption sites contributed the most to the solubility at low concentration, as was demonstrated in Figure 3-12. 117

PAGE 118

The above results demonstrated an enhancement of permeability both for anti-H-IgG and anti-M-IgG across M-IgG modified membranes. It was also confirmed that the extent of this enhancement and the linear dynamic range for each system can be controlled by the strength of the binding affinity between permeant-protein and the antibody-carrier immobilized in the membrane. Thus, we have demonstrated that antibody-modified nanotube membranes can be used for the transport of proteins at a rate determined by the affinity between carrier antibody and transported protein. Hence, it demonstrated that it is possible to develop systems of antibody-functionalized membranes with different dynamic range and controlled rate of transport. The result described herein clearly demonstrated a suitable principle for the concept of controlled facilitated transport systems for proteins using antibody-modified nanotube membranes. Of course, each system will require its own development of an appropriate affinity for the carrier/protein complex, in order to reach the desired rate of diffusion. Since the useful dynamic range of separation with these antibody-functionalized membranes depends on the binding affinity, the gamma of affinity constants found with antibodies makes those as proper elements for this principle. Besides, with the current scientific development both in genetically engineered and cloned binding proteins, the system proposed here may also be extended to other system by using proteins different than antibodies as carriers. Additional research into redesigned proteins may eventually permit the production of binding proteins against any desired analyte, just as antibodies can be produced against any antigen. 236 The concept of binding proteins genetically fused as affinity tags to other proteins without the loss of their structural and functional characteristics has been already demonstrated. 237 It has been also demonstrated that a common 118

PAGE 119

way to tune the affinity of a binding protein is to perform rational mutations at the binding site 238 240 Thus, depending on the mutations performed in the binding pocket of binding proteins, mutants with different binding affinities can be obtained; 240 this approach would allow the development of membrane-based separation systems with diverse dynamic ranges. Moreover, if this tuning affinity can be integrated with a system of membranes where the number of carriers can be augmented in a way that the permeability of the desired permeating protein can be selectively facilitated with respect to other present proteins in the separation sample, a selective and amplified separation system of membranes can be developed. This can be accomplished by taking advantage of the enhancement in permeability at low concentration caused by a contribution of the diffusion due to the Langmuir mode species as was explained above. 204, 211 Thus, this system would permit to the large number of accessible binding sites to promote higher concentration gradients inside the membrane that would enhance the flux of the selected target proteins, which imply a notable key aspect in designing more effective membranes for membrane-based separation. Validation of the Dual-Mode Model The validity of the above results can be assessed by the plot of experimental obtained permeabilities against (1 + bC feed ) -1 From the slope of the straight line in this plot, the Langmuir-mode diffusion coefficient (D H ) can be evaluated and consequently the mobile fraction of the Langmuir species F. 211 The results for these data are shown and summarized in Figure 3-16 and Table 3-4, respectively. Figure 3-16 shows and compares the experimentally and calculated obtained plots of anti-H-IgG and anti-M-IgG permeabilities vs. (1 + bC feed ) -1 Taking into consideration the cooperation of the two binding sites at low concentration, the examination of the linear relationship of these plots demonstrates that the concentration dependence of the permeability for these two proteins 119

PAGE 120

agrees fairly well with the dual-mode model (Equation 3-17). It worth mentioning, that this validation method requires the use of only one binding affinity. Therefore, in the case of anti-M-IgG that has two binding sites in the membrane; an average affinity was defined as being the sum of the free energy changes that occur in the first and second binding site. Thus, in order to keep the consistency with the conventional units the average binding affinity constant was given by the square root of the product between the stronger and weaker binding constants (b = 1.53 x 10 11 cm 3 moles -1 ). It should be pointed out that comparable results were also obtained for the experimental concentration dependence on permeability for anti-M-IgG when they were fitted using this nominal average binding affinity constant. Figure 3-16. Plots for experimental and calculated obtained permeability against (1 + bC feed ) -1 for (A) anti-H-IgG, and (B) anti-M-IgG across M-IgG modified membrane. The above results demonstrated that the permeation mechanism for anti-H-IgG and anti-M-IgG across M-IgG modified membranes obeys the dual-mode model proposed in these studies. The differences in dual-mode model parameters for these two proteins suggested the existence of a distribution of binding sites inside of these antibody-modified nanotube membranes that are accessed for each permeating protein in a particular manner. In fact the permeability plots in Figure 3-14 and Figure 3-16 demonstrated the differentiation of specific transport mechanisms for each one of the permeating proteins across the M-IgG modified membranes. The higher 120

PAGE 121

permeability for anti-H-IgG, and the higher concentration-dependence in permeability for the anti-M-IgG are clearly evidences that the permeation mechanisms across M-IgG modified membranes for each one of these proteins are different and highly depending on the affinity interactions with the immobilized carrier in the membrane. The closeness of the two obtained diffusion coefficients D D and D H also confirmed the partial immobilization of the Langmuir species (F=1) on the Langmuir sorption mode (Table 3-4). This also confirms the contribution of the Langmuir species were the responsible factor of the enhancement of permeability observed in Figure 3-14. Table 3-4.Validation of the dual-mode model parameters Permeating Protein D D (cm 2 /sec) D H (cm 2 /sec) F anti-H-IgG 7.61 x 10 -8 8.11 x 10 -8 1.06 anti-M-IgG 1.02 x 10 -9 1.02 x 10 -9 1.00 Conclusions With this work, we have demonstrated the dual-mode model proposes here is suitable model for the characterization of the transport properties of anti-H-IgG and anti-M-IgG in M-IgG modified membranes. The model demonstrated a good agreement with the experimental system studied here. Besides, the findings showed in these studies allow addressing the concept of modulated facilitated transport systems using antibody-modified nanotube membranes, and how the extent of the facilitation is determined and controlled by the binding affinity between carrier and permeating protein molecules. These studies can provide a new direction in the development of more effective membranes for membrane-based separation. 121

PAGE 122

CHAPTER 4 METAL ION AFFINITY NANOTUBE MEMBRANES: PREPARATION, CHARACTERIZATION, AND CONTROLLED TRANSPORT OF SIX-HISTIDINE-TAGGED RECOMBINANT PROTEIN Introduction Since the introduction of Immobilized Metal Ion Affinity (IMAC) by Porath and coworkers 126 in 1975, IMAC has become in one of the most widely used methods for protein purification. 241 250 IMAC exploits the fact that many transition metal ions, such as, copper, nickel, cobalt and zinc, can be coordinated to amino acid sites exposed on the surface of the proteins. It has also been proposed that metal ions in IMAC can be classified into three groups: hard, intermediate and soft. 251, 252 This classification has been based on the preferred reactivity of the metal ions toward their nucleophiles. Thus, metal ions such as Fe +3 Ca +2 and Al +3 which show preference for oxygen belongs to the hard metal ion group. Those ions that prefer sulfur (Cu + Hg +2 Ag + etc.) correspond to soft metal ions. Then, the Cu +2 Ni +2 Co +2 and Zn +2 ions, which coordinate nitrogen, oxygen and sulfur, compose the intermediate ions. Given that the number of cysteine residues on the surface of proteins is limited, makes the histidine residues the main targets for intermediate metal ions. For example, a binding constant of 4.5 x 10 3 M -1 has been determined for an average protein with a single histidyl residue. 253 Thus, multipoint histidyl interactions can be used to control the strength of binding of a protein in IMAC applications. The advanced understanding of which aminoacid site chains on the surface of proteins are responsible of the interaction with the metal ions, has been one of the most important contributions for the development of tagged peptides containing multiple histidine residues. 254, 255 In particular, the hexahistidine tags attached to the 5 or 3 end of the target gene that encodes the protein are usually used as purification tags in recombinant proteins. 256, 257 122

PAGE 123

Today, recombinant protein purification represents one of the principal protein purification markets, and many systems have been developed in recent years for the purification of these proteins. 241 250, 257 The system proposed here, exploits the advantages of IMAC separations, but in a nanopore-membraneplatform. The use of imidazole-modified silica nanotube membranes for the controlled transport of histidine-tagged recombinant proteins is described. Imidazole is a functional aromatic nitrogen heterocycle important in ligand coordination chemistry. The structure of imidazole is a five-membered ring with two different nitrogen atoms. One of the nitrogen is bound to hydrogen, having two lone pair of electrons in an unhybridized p-orbital donated to the aromatic ring, and it is structurally related with the pyrrole nitrogen. The second nitrogen structurally related with the pyridine nitrogen, donates its single electron in an unhybridized p-orbital to the ring, containing also a free lone pair of electrons in a hybrid orbital. 258 In aqueous solution imidazole behaves like a weak organic base with pK a = 6.99. 259 Therefore, the lone pair of electrons in the pyridine-like nitrogen atom, are only involved in coordination when the pH of aqueous solutions is close or higher than 6.99. It is worth mentioning, that the mode of coordination of imidazole can be also influenced by the nature of the metal. In this investigation, the silica nanotube membrane was prepared by sol-gel template synthesis of silica nanotubes within the pores of a nanopore alumina template. Imidazole was attached to the inside walls of chloropropyl-modified silica nanotubes by nucleophilic substitution reaction with 2-mercaptoimidazole in anhydrous tetrahydrofuran. The modified membrane was then exposed to solutions of the desired divalent metal ion (copper, nickel, cobalt or zinc) to attach the metal via coordination with the imidazole. The histidine-tag on the 123

PAGE 124

recombinant protein reversibly coordinates with the metal ion sites, and via this chemistry the flux of the protein across the membrane could be modulated. In these studies the concept of IMAC separations was exploited in a nanopore-membrane-platform. The transport of histidine-tagged recombinant proteins in metal ion affinity nanotube membranes is described. It was demonstrated that transport of histidine-tagged recombinant proteins across imidazole-modified silica nanotube membranes can be controlled by the nature of the immobilized metal ions on the imidazole surface. The imidazole functionalized membrane was extensively characterized using three different analytical techniques in order to identify how the coordination chemistry with different divalent metal ions affects the functional behavior and transport properties during the permeation of six-histidine recombinant proteins. Furthermore, additional results shown that by the appropriate choice of the transition metal ion for coordination with the imidazole ligand, the binding strength and optical properties within the functional membrane can also be modified. These studies demonstrated an alternative way of immobilized metal ion affinity by using nanotube membranes. Experimental Materials Commercially available nanopore alumina membranes (60 m thick, nominal pore diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS) (Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes. A 3-chloropropyltrimethoxysilane (United Chemical Technologies) was used for silane modification of the silica nanotubes. 2-mercaptoimidazole from Sigma Aldrich was used as attacking nucleophile to substitute the chlorine atoms in the 3-chloropropyl silica nanotube functionalized membrane. As an aprotic solvent anhydrous tetrahydrofuran >99.9%, inhibitor free from Sigma Aldrich was used. Divalent salts of copper, nickel, cobalt and zinc nitrate 124

PAGE 125

obtained from Fulka, were used to attach the metal via coordination with the imidazole. A six-histidine-tagged ubiquitin human recombinant protein (6His-UHRP) obtained from Sigma Aldrich was used to investigate the transport properties of the imidazole-modified membranes. All other chemicals were of reagent grade and used as received. Purified water, obtained by passing house-distilled water through a Barnstead, E-pure water-purification system, was used to prepare all solutions. Preparation of the Silica Nanotube Membranes Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-based template-synthesis method described previously. Briefly, a sol-gel precursor solution was prepared by mixing absolute ethanol, TEOS and 1 M HCl (50:5:1 by volume), and this solution was allowed to hydrolyze for 30 minutes. The alumina membranes were then immersed into this solution for two minutes under sonication. The membranes were then removed and the surfaces swabbed with ethanol to remove precursor from the alumina surface. The membranes were then air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. This yields silica nanotubes, with wall thickness of ~3 nm, embedded within the pores of the alumina membrane. 9 Preparation of Imidazole Functionalized Membranes The 3-chloropropyltrimethoxysilane was first attached to the inside walls of the silica nanotubes. A solution that was 2% (v/v) in 3-chloropropyltrimethoxysilane was prepared in ethanol; to this solution was added acetate buffer (50 mM, pH=5.1) to make the solution 5% (v/v) in this buffer. This ethanol-based solution was stirred for 20 minutes and the alumina membranes were then immersed. The soak membranes were sonicated for 2 minutes, and after 20 minutes the membranes were removed from the solution and rinsed with ethanol. To make sure that the loose silane was removed from the surface, the membranes were also sonicated for 125

PAGE 126

2 minute in ethanol. The membranes were then dry with nitrogen and placed under vacuum for 12 hours. The resulting 3-chloropropyl silica nanotube functionalized membranes were then immersed into different vials containing a solution that was 0.5 M in 2-mercaptoimidazole dissolved in anhydrous THF at 45 C. This reaction was kept at this temperature for a period of 1 hour, with periodically additions of THF to continually replace the evaporated THF. Then, the lids of the vials were tightly closed and they were wrapped with paraffin film, and the membranes were incubated in the dark for 24 hours at 25 C to allow the completion of the reaction. After imidazole functionalization, the membranes were rinsed copiously with ethanol and sonicated for 2 minutes to remove the unreacted mercaptoimidazole. A schematic illustration of the reaction mechanism for the preparation of imidazole functionalized membranes is shown in Figure 4-1. Figure 4-1.Reaction mechanism of imidazole functionalized nanotube membranes. Divalent Metal Ions Complexation The imidazole modified membranes were then exposed to solutions of the desired divalent metal ion (copper, nickel, cobalt or zinc) to attach the metal ion via coordination with the 126

PAGE 127

imidazole. Briefly, the imidazole functionalized membranes were immersed in different vials containing solutions prepared in ethanol that were 1 mM in Cu +2 Ni +2 Co +2 and Zn +2 respectively. The vials were placed in an orbital shaker platform and the membranes were equilibrated for 12 hours at 25 C to affirm the complexation of the metal ions on the membrane. After equilibration the membranes were removed and rinsed copiously with ethanol. Instrumental Characterization of the Imidazole Functionalized Membranes To characterize the imidazole functionalized nanotube membranes fluorescence microscopy, Fourier transformed infrared attenuated total reflectance spectroscopy (FTIR-ATR), and X-ray photoelectron spectroscopy (XPS) measurement were taken. We discovered that the imidazole membranes are fluorescent after functionalization. Therefore, the observation of these imidazole functionalized membranes under the fluorescence microscope was used as the first control to confirm the success of the chemistry modification. The fluorescence intensity was measured using a fluorescence microscope which was equipped with both a fluorescence detector and CCD camera. This system combines an Axioplan 2 imaging microscope (Zeiss) with a J&M-PMT photometry system detector (SpectrAlliance), for measuring the fluorescence intensity using a BP450-490 filter for excitation and a LP515 filter for emission of light. The FTIR-ATR was employed to examine the presence of characteristic bands for 2-mercaptoimidazole in the functionalized membranes. The FTIR-ATR spectra were recorded on a Thermo Nicolet Nexus 670 FTIR spectrometer equipped with a GATRTM as the ATR crystal. All spectra were run at 4000cm-1 to 400cm-1, 512 scans, 4 cm-1 resolution, and gain equal to 1. The spectrometer aperture was set to 100 for all runs. A background spectrum (air) was taken prior to each sample spectrum, which was used for background subtraction using the OMNIC ATR software for the ATR correction. 127

PAGE 128

The surface composition of imidazole functionalized membranes and the presence of the immobilized transition metal ions were also determined through XPS analysis. The analysis was performed using a XPS/ESCA Perkin-Elmer PHI 5100 ESCA system. For the XPS measurement a Mg K (1253.6 eV) or Al K (1486.6 eV) monochromatized X-ray source was used to stimulate photoemission. Survey and high resolution spectra were obtained with a hemispherical energy analyzer operated in the constant analyzer transmission mode to measure the binding energies of emitted photoelectrons. Transport Experiments The imidazole-modified membranes were sandwiched between two pieces of Scotch tape that had 0.13 cm 2 area holes punched through them. These holes defined the area of membrane exposed to the contacting solution phases. This membrane assembly was mounted between the two halves of a U-tube permeation cell, with half-cell volumes of 2.5 mL. The feed solution for the recombinant protein experiments was 1.87 M dissolved in 50 mM phosphate-buffered saline pH 8.0 that also contained 300 mM sodium chloride and 2 mM imidazole. The receiver solution on the other side of the membrane was identical except it initially contained no protein. The concentrations of the transporting protein in the feed solution were in all cases the same. It is important to mention that prior to starting each transport experiment the metal-immobilized imidazole membranes were equilibrated in the PBS buffer solution for two hours, and that the metal ion immobilization for each divalent cation was completed right before this equilibration. The rate of protein transport across the membrane was determined by periodically checking the absorbance at 276 nm in the receiver solution using a flow-through UV-Vis detector (Figure 4-2). 128

PAGE 129

Figure 4-2.Flow-through UV-Vis detection system used to measure the flux of six-histidine-tagged ubiquitin human recombinant protein. This system consisted of a L-7455 UV-Vis detector (Hitachi), where the flow stream was delivered using an L-7100 HPLC pump (Hitachi) using a flow rate of 3 mL per min. The whole system was controlled using Chromeleon 6.50 chromatography management software (Dionex) through a UCI-100 universal chromatography interface (Dionex). Calibration curve was obtained with the same flow through system on solutions containing known concentrations of the transport recombinant protein. This calibration curve was used to obtain the concentrations of the transport proteins in the receiver solution. Results and Discussion Imidazole Surface Modification and Characterization The imidazole functionalized nanotube membranes were prepared by covalent immobilization of imidazole inside of the silica nanotubes. This coupling between silica nanotube surface and imidazole was accomplished via simple silane chemistry using 3-chloropropyltrimethoxysilane. Silica surfaces offer the positive attribute that they can be derivatized with an enormous variety of chemical functional groups using simple silane chemistry 95 Here, the covalent immobilization of imidazole was achieved by nucleophilic substitution of chlorine atoms in the resulting silane modified 3-chloropropyl silica nanotube 129

PAGE 130

membranes. It has been demonstrated that mercapto compounds are appropriated modification reagents for this strategy. 106, 260 263 Furthermore, it has been confirmed that the use of aromatic thiols leads to degrees of modification which are much higher than those achieved with aliphatic ones. In addition, when aromatic thiols are used the possibilities of elimination as a side reaction are eradicated. 261 In order to promote the substitution nucleophilic bimolecular mechanism (S N 2) the imidazole modification was carried out in an aprotic solvent (THF). Since aprotic solvents do not solvate the nucleophiles, the reactivity of the 2-mercaptoimidazole can be enhanced making it more accessible to approach to the primary halide carbon as the conditions in 3-chloropropyl silica nanotube modified membranes. A high concentration of 2-mercaptoimidazole was also used to promote the S N 2. It is worth mentioning that some indications that the reaction is taking place were the sulfur smell during the reaction and the lower pH of the solution at the end of it. Fluorescence microscopy characterization We found that after imidazole immobilization the nanotube membranes were fluorescent, and this observation become in our first control to monitor the success of the modification chemistry. Figure 4-3.Fluorescence spectra of (a) silica nanotube and (b) imidazole functionalized membranes. 130

PAGE 131

Figure 4-3 shows the fluorescence spectra of the nanotube membranes before and after imidazole functionalization. It is clearly seen from these data that the fluorescence intensity for the imidazole functionalized membrane was nine-fold higher compared to the bare silica nanotube membranes. This fluorescence signal was attributed to the imidazole moiety in the functional membrane. The origin of this fluorescence signal was also confirmed by conducting fluorescence experiments of pure 2-mercaptoimidazole in methanol. Additionally, we found that by the choice of the appropriate transition metal ion for the complexation with the imidazole, the optical properties within the functional membrane can be modified (Figures 4-4 and 4-5). Figure 4-4.Fluorescence spectra of imidazole functionalized membranes before (upper) and after (lower) metal ion immobilization: (A) Co +2 (B) Cu +2 (C) Ni +2 and (D) Zn +2 Figure 4-4 shows the fluorescence spectra for a set of four imidazole functionalized membranes that were exposed respectively to four different nitrate salts solutions of Cu +2 Ni +2 Co +2 and Zn +2 (vide infra). It was observed that after immobilization of these transition metal ions the fluorescence intensity of the functional membranes was quenched. Figure 4-5 reports the fluorescence quenching ratio F/F 0 where F 0 is the fluorescence of imidazole functionalized membrane and F is the corresponding fluorescence in the presence of 131

PAGE 132

the metal ions. It is evident that fluorescence quenching was much higher for Cu +2 and Co +2 ions, while it was weaker in the case of Ni +2 and very low for Zn +2 Figure 4-5.Fluorescence quenching (F/F 0 ) for imidazole functionalized membranes in the presence of nitrate solutions prepared in ethanol that were 1 mM of the respective divalent metal ion. Fluorescence quenching by transition metal ions is a common phenomenon. 264 The fluorescence quenching by transition-metal ions and their coordination complexes is a process which has been known for a long time and is still the subject of intense investigations. 265 267 It is now well recognized that electron transfer and electronic energy transfer are considered as the two main deactivation pathways responsible for efficient fluorescence inhibition. 264 267 These metallosystems based on fluorescence detection have found applications in many areas such as metal ion sensing, 268, 269 artificial photosynthesis, 270 and optical sensing of DNA. 271 273 On the other hand, when the imidazole functionalized membrane was exposed to solutions of CuCl 2 instead of Cu(NO 3 ) 2 the optical behavior for this metal ion immobilized membrane was different than in the case of nitrate salts. Under this condition, a fluorescence enhancement instead of quenching was observed in the presence of the metal ion (Figure 4-6). 132

PAGE 133

Figure 4-6 compares the fluorescence spectra for imidazole functionalized membranes before and after coordination with CuCl 2 Simultaneously with the fluorescence enhancement, a small bathochromic shift of the maximum emission wavelength was observed in the fluorescence spectra of the copper immobilized membrane from 443 nm to 550 nm. Figure 4-6. Fluorescence spectra of imidazole functionalized membranes (a) before and (b) after metal ion immobilization using an ethanol solution that was 1 mM in CuCl 2 Besides, a dependence on the copper concentration for the fluorescence enhancement was found (Figure 4-7). As can be seen from Figure 4-7 the addition of the chloride salt of copper leads to an increase in fluorescence intensity of the functional membrane and at a certain concentration the fluorescence intensity remains constant. Such an increase in fluorescence enhancement can result from conformational restriction induced upon binding. 274 It is beyond the scope of the present work to discuss in detail the mechanism of interaction of the functional imidazole membrane with these metal ions. Nevertheless, the observed fluorescence intensity changes can be reasonably explained on the basis of the complex structure formed after coordination takes place. It was ascribed that for the situation of the imidazole membranes coordinated with the nitrate salts, the nitrate anions in [M(Im) n ](NO 3 ) 2 complexes are strictly counterions, hence they are no part of the sphere of coordination. Therefore, the 133

PAGE 134

fluorescence quenching through an electron transfer mechanism can be accredited. 264, 265 The great -acceptor properties of the imidazole permit it to accept electronic charge from d orbitals on the metal ion. 258 These charge transfer transitions correspond roughly to the transfer of an electron from the metal ion to a ligand. More precisely, an electron is transferred from a bonding orbital of mostly metal ion character to an empty antibonding orbital of mostly ligand character. 275, 276 Figure 4-7. Copper concentration dependence on fluorescence enhancement for imidazole functionalized membranes. Conversely, in the situation using chloride salts, the chloride anions can be part of the sphere of coordination. Therefore, in the [M(Im) n (Cl) 2 ] complex both the imidazole and chloride ligands can be coordinated to the metal ion. Furthermore, for that reason that chloride is known as a -trans-directing ligand, the enhancement in fluorescence can be ascribed to a -trans-influence, so that a trans--acceptor ligand as chloride, can form bonds with the metal and withdraw electron density away from the other imidazole ligand. 275, 276 Then, the -trans bonding have a cooperative effect allowing more delocalization of the electrons. 277 The geometrical restraint required for having this interaction, support that fact that the increase in fluorescence enhancement can result from conformational restriction induced upon binding. 269, 274 McFarland and Finney269, 274 demonstrated that metal ion binding could restrict the rotational excited-state of 134

PAGE 135

a fluorophore suppressing some of the fluorescence deactivation pathways, producing fluorescence enhancement. Thus, those changes in fluorescence intensity showed how the optical properties of the functied tion by FTIR-ATR spectroscopy ube membrane with the imidazole moiety was aa nce. To he differences observed in the spectra of bare alumina and imidazole-modified memb 2-n the is known that hydrogen bonds with water molecules contribute to broaden peak areas.280, 281 onal imidazole membranes changes depending on the transition metal ion immobilized. Besides, these results confirm the accomplishment in the preparation of imidazole functionalizmembranes. Characteriza The successful modification of the silica nanot lso demonstrated by FTIR-ATR spectroscopy. Figure 4-8 compares the FTIR-ATR spectrfor bare alumina and imidazole functionalized membranes. It should be pointed out that all spectra were corrected for spectral distortion using the OMNIC ATR software. However, byusing this ATR correction it is possible to lose spectral information of the fundamental vibrational modes of a molecule, overall when the entire spectral region is examined at oavoid this, it is a common method to analyze the data in limited spectral regions in order to give a more detailed assignment of the characteristic bands. The Figure 4-8B illustrates this issue, anda narrow spectral region, from 2000 to 600 cm -1 for the imidazole functionalized membrane is used. T ranes clearly demonstrated the modification of the nanotube membranes by usingmercaptoimidazole. For the FTIR-ATR spectrum of bare alumina two broad bands appear i3900-3200 cm -1 and 3000-1600 cm -1 regions, which are attributed mainly to the O-H stretching modes and O-H deformation modes, respectively. 278, 279 Although, for these two broad bands the contribution of water (at about 3600 cm -1 and about 1600 cm -1 ) also has to be considered, since it 135

PAGE 136

Figure 4-8. (A) FTIR-ATR spectra of bare alumina and imidazole functionalized membranes. (B) FTIR-ATR spectra of imidazole functionalized membranes in a narrow spectral region illustrating the effect of using two different backgrounds for the spectrum subtraction: (a) air, and (b) 3-chloropropyl silica nanotube membrane. For tn of their freque in KBr (FiguFigure 4-9.TransmThe FTIR-ATR spectrurane showed clearly three tion modes for C=C he case of imidazole membranes, the observed bands were assigned by comparisoencies with the transmission FTIR spectrum of pure 2-mercaptoimidazol re 4-9). 263, 282 285 ission FTIR spectrum of pure 2-mercaptoimidazole in KBr. m of the imidazole functionalized memb of the characteristic vibrations for 2-mercaptoimidazole. One was the ring vibra and C=N bonds in the region ~1684 1630 cm -1 A second band, observed at ~1486 1400 cm -1 region that was assigned to the thioamide bands I and II, which are coupled vibrations 136

PAGE 137

involving stretching motion of C-S + N-H and bending motion C-N + N-H, respectively. 286 The third band was the very strong band at 2890 cm -1 assigned to N-H stretching mode. This imidazole N-H stretching appeared as a broad band in a wide range from ~3400 to 2200 cm 1 When the characteristic band of N-H was compared with the band of pure 2-mercaptoimishift from 3145 cm dazole a zed -1 to 2890 cm -1 was observed. This shifting might be due to the solid state characteristic of the sample, but also to the formation of hydrogen bonds between imidazole moieties. All the assigned bands were also observed in metal-immobilized imidazole functionalized membranes, and some of them were enhanced by the presence of the immobilimetal ion as is shown in Figure 4-10. Figure 4-10.FTIR-ATR spectra of imidazole functionalized membranes showing the effect of the metal ion immobilized on the thioamide bands: (a) Zn+2, (b) Co+2, and (c) Cu+2. Figure 4-10 compares the results for three different imidazole functionalized membranes which werely. It was for e exposed to solutions of divalent metal ions of copper, cobalt and zinc, respectiv found that the enhancement of the band intensity is related to the strength of the binding constant between the metal ions with the functional membrane. 258 In addition, it was observed that those bands shifted toward lower frequencies as is demonstrated in Figure 4-10. These changes of the band intensity may be attributed to changes of the electric dipole moment after coordination with the metal ions. 287 It may be possible that IR energy is preferential absorbs 137

PAGE 138

some favored orientations of the imidazole moiety in the functional membrane after coordination. The observed band at 1110 cm was also assigned, and it is clearly illustrated i8B. This band n Figure 4-was labeled to the asymmetric C-S-C stretching vibrations.284, 288 For the correct assignne embrane for backgnal mode 8 cm-1 branes Frequency (cm-1) ment of this band, the contribution of Si-O stretching vibration at this region was also considered. In order to confer a proper assignment, the FTIR-ATR spectrum of imidazole functionalized membrane was also acquired by using 3-chloropropyl silica nanotube membraas background for the spectrum subtraction instead of air (Figure 4-8B). Figure 4-8B compares the FTIR-ATR spectra of an imidazole functionalized membrane, which were collected using both air and 3-chloropropyl silica nanotube m round subtraction, respectively. It was clearly observed that when the chlorosilane-modified membrane was used for the background subtraction, the fundamental vibratiofrom the imidazole moiety at 1118 cm -1 was more clearly discriminated. Therefore, at 111the contribution of both Si-O and asymmetric C-S-C stretching vibrations should be considered. 284, 288, 289 Table 4-1.FTIR-ATR characteristic bands of imidazole functionalized nanotube memAssignment N-H stretching ~2900 N-H bending in the plane ing vibration modes ~1684-1630 which are respectively coupled vibrations n of C-S + N-H, and bending motion C-N + N-H. ~1486 and 1412 hing vibrations. 1118 ~1560 C=C and C=N r Thioamide bands I and II, involving stretching motio C-S-C anti-symmetric and Si-O stretc 138

PAGE 139

139 Table 4-1 shows the assignment of the characteristic bands for the imidazole functionalized embranes. In some cases the overlapping between bands made this assignment difficult. This was particularly more apparent in the region for N-H bending vibration and the coupled vibrations involving C-S and N-H of the thioamide band I. Transmission FTIR spectra for the alumina template, silica nanotube, and nanotube modified membrane films were also studied. Figure 4-11 shows the transmission FTIR spectra of the alumina template film for each of the modification steps illustrated in Figure 4-1. Figure 4-11. Transm for each of the modificationplate, (b) silica, (c) By ovariations ofith each chemistry modification. However, it was possible to distinguish through the different modification steps by monitoring the changes in transmittance in each transmission FTIR spectrum. It was clearly observed that the absorption of stretching vibrations band for the O-H groups in the region between 3000 and 3600 cm changes throughout the different modification steps. Likewise, differences in the region 1500 and 1600 cm-1 for the bending vibration of O-H were m 263 ission FTIR spectra of the alumina template film steps illustrated in Figure 4-1: (a) bare alumina tem3-chloropropyl silane modified and (d) imidazole functionalized nanotube membranes, respectively. bserving the transmission FTIR spectra in Figure 4-11, there were no obvious the characteristic bands related w -1 139

PAGE 140

also od rmed the success of the silanization reactied e C-S sf om characteristic of its cohilic the infrared absorption spectral region. Thus, the shorter wavelengths can penetrate less the bserved. In the latter region, the differences between the bending vibrations of the O-H for the alumina membrane and the silica nanotube membrane were not significant (spectra a, and b in Figure 4-11). This was reasonable because part of the OH groups for the alumina are replacefor the silanols OH groups in silica; besides silica surface introduce higher silanol groups on surface for further silane functionalization. 30, 94, 99 102 In the case of chlorosilane modification (spectrum c in figure 4-11), lower absorption for the silane-modified membrane was observed in the bending O-H region compared with alumina and silica. This difference in transmittance clearly confi on where the O-H free groups in the alumina matrix and the silanol groups in the silica nanotubes were replaced by the condensation reaction with 3-chloropropyltrimethoxysilane. The differences in transmittance were even more evident for the imidazole functionalizmembrane (spectrum d in Figure 4-11). Additionally this exhibited two observable absorptionbands at 1118 cm -1 and 474 cm -1 assigned respectively to the anti-symmetric C-S-C, and to th tretching vibrations, which confirmed the imidazole functionalization. These differences in the intensity of the bands might be attributed once more to changes othe electric dipole moment after each chemistry modification. 287 Besides, each chemistry modification introduces a variety of vibrational motions, and degrees of freed mponent atoms that should produce a net change in the dipole moment. It should be also considered the fact that may be after the modifications the membranes become less hydropthan the bare alumina template, thereby the contribution of water on surface must be less significant, even though all the membranes were dried at the same conditions. These transmission FTIR data in Figure 4-11 may perhaps be interpreted on the basis of 140

PAGE 141

alumina matrix and transmit more IR energy through the porous structure of the membraninduce vibrational excitation of the covalently bonded atoms on the nanotube su e to rface. On the other re metal azole membranes. This analysis begins, however, with the XPS rane (Figure 4-12). Figure 4-12.The survey sped analysis showed the presposition peaks for Al, O, P, and C. The the surface; this hand, longer wavelengths should have deeper penetration on the alumina matrix, allowingless transmission of the IR energy. Therefore, the latter region at the same time should be mosensitive to changes associated with electric dipole moment, and more information may be obtained in this region. Surface chemical composition by XPS XPS analysis was performed not only to assure the presence of chlorine and imidazole groups in the modified membranes, but also to confirm the immobilization of the transition ions on the functionalized imid survey studies for the bare alumina memb XPS survey spectrum for the bare alumina template. ectrum for the alumina template is shown in Figure 4-12. The performence of elemental chemical comunexpected peak for C1s at 289.3 eV was attributed mainly to impurities on lained below in more detail. carbon impurity is exp 141

PAGE 142

Survey scans were taken to determine the elemental surface composition of various elements present in the chlorosilane and imidazole modified membranes. Figure 4-13 compares the XPS survey scan of the chlorosilane and imidazole modified membranes. XPS survey spectra for (a) chlorosilane and (b) imidazole modified modified surface shown the presence of XPS line for china, which confirmed the accompmodification of the silica nanotubes (spectrum a Figure 4-13). In addition, Figure 4-13.embranes. The chlorosilane mlorine Cl2p, which was absent in the bare alumlishment of the chlorosilane this survey scan d bare mlane e XPS measurements for the imidazole functionalized membranes suggested the pr shown distinct peaks for Al2p as well as for C1s compared to the imidazole functionalized an embranes. It was noticed that after functionalization with 2-mercaptoimidazole, an increase in C1s peak and a diminution in the Al2p peak was observed when it is compared with the chlorosimembrane (spectrum b Figure 4-13). Considering the surface sensitivity of XPS, the decrease ofthe Al2p peak in th esence of imidazole on the nanotube membrane surface. 136 Additionally, the functionalization of the membranes with the imidazole moiety was confirmed for the presence of N1s and S1s lines in the XPS spectra. 142

PAGE 143

The chemical composition and binding energies data from such spectra are summarized in Table 4-2. It is worth mentioning that owing to the insulator characteristics of the aidentification of the specific peaks was demanding. Moreover, due to the surface charging a positive shift toward higher binding en lumina, the ergy was observed for some peaks. Tablemodified, and imidazole functionalized membranes 4-2.XPS surface composition and binding energies for bare alumina, chlorosilane Bare Alumina Chlorosilane Modified Imidazole Functionalized XPS Signal % BE(eV) % BE(eV) % BE(eV) Na 1s 4.4 1075.6 1 .0 1074.5 O 1s 60.3 536.2 52.5 534.5 49.8 534.0 N 1s 3.9 9.3 88.0 201.8 104.5 402.0 C 1s 7.4 28 11.9 2 .4 287.0 Cl 2p 1.4 203.0 S 2p 1.5 171.0 P 2p 2.8 138.4 2.3 137.0 2.1 136.0 Si 2p 1.4 104.0 Al 2p 29.5 78.4 25.7 77.5 19.9 76.5 Furthermore, the XPS multiplexes analysis aigh resoluor the izole mes (Figure 4-14) confirmed the imidazctionaln. An that was based only on the relative areas of the XPS peaks at hition demonstrated an elemental compositional ratio f 6:2:1 for C: N: S. However, after the individual peak areas were divided by their respective atomi t h tion f mida embran ole fun izatio analysis gh resolu o c sensitivity factor (ASF) the elemental compositional ratio for C: N: S respectively was 14:2:1. This increase in C signal was attributed to impurities on the surface of the alumina membranes. These carbon impurities were corroborated by observing the high-resolution scan of the C1s peak (Figure 4-15). 143

PAGE 144

Figure 4-14.XPS multiplexes analysis at high resolution for the imidazole functionalized membrane. Figure 4-15 compares the high-resolution C1s peaks for the bare alumina membrane, the chlorosilane modified membrane and the imidazole functionalized membrane, respectively. plate, (b) chlorosmbranes consisted both of energy binding where thnd, the C1s peak for which may b Figure 4-15.High-resolution XPS data of the C1s peak for (a) bare alumina temilane modified, and (c) imidazole functionalized membranes. It was clearly seen that the C1s peak for Imidazole modified mea well defined peak at 287 eV which was ascribed to the C-C peak and a shoulder at higher e C-N, C-S peaks are overlapped. On the other ha chlorosilane modified membranes revealed a better indication of a second peak at 291 eV e assigned to the C-Cl environment. The high-resolution scan of C1s peak for the bare 144

PAGE 145

alumina membrane consisted of only one peak the C-C peak at 287 eV, which confirmed the presence of carbon impurities in the membrane. The immobilization of the transition metal ions in the imidazole functionalized membranwas also verified by XPS analysis. Figure 4-16 compares the XPS survey scan of two imidazofunctionalized membranes after immobilization o es le f Cu and Zn, respectively. For the XPS survey scan oFigure 4-16. Xmobilization of (a) Cu+2spectrum of the impresence of Zn2p1 at 1050 eV and Zn2p3 at 1026.5 eV peaks. Whereas the survey for Cu shown the prhould f Cu and Zn metal-immobilized nanotube membranes, the Mg anode and the Al anode were used respectively to stimulate photoemission. PS survey spectra for imidazole modified membranes after im, and (b) Zn+2. The presence of XPS spectral lines for Zn and Cu were observed in the respective idazole functionalized membranes. The XPS survey scan for Zn shown the esence of Cu2p1 and Cu2p3 peaks assigned at 954 eV and 934 eV, respectively. It sbe mentioned that analogous experiments were also carried out for the immobilization of Ni andCo, and Auger lines for all these four metal ions were also observed. Thus, these data confirmed the coordination and immobilization of the transition metal ions used in these studies with the imidazole functionalized membranes. 145

PAGE 146

Transport Properties of Imidazole Membranes The transport properties of the imidazole functionalized membranes were explored by monitoring the flux of six-histidine tagged ubiquitin human recombinant protein (6His-UHRP) in function of time. For the transport experiments, the rate of 6His-UHRP across the membrane was determthe he absorption maximum occurs in 276 nm. Figure 4-17.UV-Vis spectrummM phosphate-buffered loride and 2 mM imidazole. Calibration curve at 276 nm showed in Figure used to obtain the concentration of the transport protein in the receiver solution. Figure 4-18 show ined by periodically checking the absorbance at 276 nm in the receiver solution using flow-through UV-Vis detector illustrated in Figure 4-2. Figure 4-17 shows the absorption spectrum of 50 M solution of 6His-UHRP that was dissolved in 50 mM phosphate-buffered saline pH 8.0 that also contained 300 mM sodium chloride and 2 mM imidazole. It is clearly observed that t for 50 M of 6His-UHRP measured in 50 saline solution pH 8.0 that also contained 300 mM sodium ch was obtained with the flow through system 4-2 on solutions containing known concentrations of 6His-UHRP. This calibration curve was s the typical calibration data (Figure 4-18A) and typical calibration curve (Figure 4-18B) 146

PAGE 147

for 6His-UHRP. The calibration curve shown an excellent linear correlation (r=0.999) withinrange from 0.374 to 1.496 M in protein concentration. a and (B) calibration curve for 6His-UHRP. binant protein transported across the metal Figure 4-18.Typical plots of (A) calibration dataFigure 4-19 shows the nanomoles of the recomperme ion affinity nanotube membranes versus permeation time. In addition, the results of the ation experiments for four different metal-immobilized imidazole membranes are compared. Figure 4-19.Flux plots for 6His-UHRP in imidazole functionalized metal ion affinity nanotube membranes that had immobilized different metal ions: (a) Cu+2, (b) Ni+2, (c) Co+2, and (d) Zn+2. The feed solution concentration was 1.87 M for all permeation experiments. 147

PAGE 148

We found that the flux of 6His-UHRP through modified membranes can be controlled by riate divalent metal ion (copper, nickel, cobalt or zinc) immobilized in the memb the approprane via coordination with the imidazole, as is shown in Figure 4-19. The data clearly demonstrated that the functional transport of the imidazole functionalized membrane depends on the appropriate immobilized divalent metal to coordinate with the imidazole ligands in the membrane. When the Zn ion was immobilized in the membrane the flux was considerably higher than the flux for the other three divalent metal ions (Co, Ni, and Cu). The flux for the Zn situation was not linear after 3 hours of permeation due mainly to the rapid equilibration reached between the feed and receiver solutions. Thus, the coordinated Zn in the imidazole functional membrane acts as binding sites to facilitate the transport of the 6His-UHRP. Figure 4-20.Effect of the divalent metal ion immobilized on the flux of 6His-UHRP across imidazole functionalized metal ion affinity nanotube membranes. Analogous permeation experiments were obtained for 6His-UHRP with the other three quite linear and their fluxes were always lower than the Zn immobilized situation. The slope of the flux plots for the four different immobilized metal ions was determined by using transport data divalent metal ions. The transport plots of 6His-UHRP with these three metal ions were not 148

PAGE 149

only during the initial three hours of permeation, period where fluxes were apparently linear for all four metal-immobilized membranes. These results are summarized in Figure 4-20. In order to quantify the effect of the different metal ions immobilized in the trans port of 6His-g the ionalized metal ion affinity nanotube membranes moles h-1 cm-2) UHRP across the metal ion affinity nanotube membranes, an enhancement selective coefficient, was defined as the calculated fluxes during three hours of transport for each divalent metal ions divided by the flux of the Cu situation. Thus, it was observed that durinthree initial hours of permeation, the fluxes across the metal ion affinity nanotube membranes with Ni, Co and Zn immobilized ions are enhanced respectively in ~2, ~3 and ~9 times, compared with the Cu immobilized membrane. These results are shown in table 4-3. Table 4-3.Fluxes and selectivity coefficients for 6His-UHRP across imidazole funct 2 Metal Ion Immobilized Flux of 6His-UHRP (nano 2 Cu F lux MFlux Cu +2 0.43 1.0 Ni +2 0.77 1.8 Co +2 1.47 3.4 Zn +2 4.05 9.4 Thus, these data demonstrated that by the selection of the appropriate transition metal ion to cooConclusions The development of metal ion affinexploe rdinate the imidazole, the transport properties within the functional membrane can be controlled. ity nanotube membranes has been demonstrated by iting the concept of IMAC separations in a nanopore-membrane-platform. Silica nanotubmembranes were functionalized with imidazole through a nucleophilic substitution reaction between 2-mercaptoimidazol and 3-chloropropyl modified silica nanotubes carried out in an 149

PAGE 150

aprotic medium. It was demonstrated that transport of histidine-tagged recombinant proteins across these imidazole-modified silica nanotube membranes is controlled by the nature of immobilized metal ion on the imidazole surface. Besides, the extensive characterization of imidazole functionalized membrane allowed the understanding of how coordination chemistry with different divalent metal ions affects the functional behavior of the membrane and transportproperties during the permeation of six-histidine recombinant proteins. Furthermore, additional results shown that by the appropriate choice of the transition metal ion for coordination with the imidazole ligand, the binding strength and optical properties within the functional membrane are also modified. Hence, these studies revealed how the presence of the immobilized metal ion affects the functionality of the imidazole membrane and its transport properties. The metal ioaffinity nanotube membrane described here has the potential to be a versatile analytical platformwhere different metal ions can be immobilized on the same prevailing conditions, additionally offering the possibilities of regeneration without apparent loss of the metal coordination properties of the functional membrane. These studies demonstrate an alternative way of immobilized metal ion affinity by using nanotube membranes. the n 150

PAGE 151

CHAPTER 5 TRANSPORT ANALYSIS OF PROTEINS ACROSS SILICA NANOTUBE MEMBRANES THROUGH A FLUORESCENCE QUENCHING-BASED SENSING APPROACH Introduction Molecular recognition is a very important and commonly used tool in both clinical diagnostics and pharmacology. The practical methods based on recognition depend fundamentally on specific biomolecular recognition reactions. In some cases, these biomolecular recognition reactions confer such extremely high specificity that it has allowed the development of analytical assays for the determination of many chemicals and biochemicals at very low concentrations. 290 293 This is important because in this era of proteomics in particular, the quantitative determination and identification of proteins at low concentrations are essential requirements. Many very sensitive methods have been developed for the detection of proteins and other biological molecules on the basis and capacity of fluorescent compounds to transfer energy absorbed from photon radiation to close molecules by taking advantage of the molecular recognition. 294 297 This called fluorescence resonant energy transfer (FRET) phenomenon takes place when the energy in the excited state of a donor fluorophore is transferred to a close acceptor molecule via an appropriate orientation of the dipole-dipole interaction between these donor-acceptor molecules. 298, 299 One of the basic conditions for the FRET process to occur is that the emission spectrum of the donor must overlap the absorption spectrum of the acceptor. Therefore, FRET efficiency is a distance-dependent interaction process. 131, 298 This distance dependence of FRET has been extensively used to study both structure and dynamics of proteins and nucleic acids, as well for the detection and visualization of intermolecular associations, and for the development of biomolecular devices based on intermolecular binding interaction. 298, 300, 301 151

PAGE 152

If the donor and acceptor molecules are brought close together within the range of 20 to 100 the fluorescence intensity of the donor fluorophore is reduced (quenched) as a result of the FRET process. However, if the donor and acceptor molecules are brought within an intimate distance, most of the absorbed energy is dissipated as heat and only a small amount of the energy is emitted as fluorescence; this phenomenon is commonly referred to as static or contact quenching. 131 The rate of fluorescence quenching can be studied experimentally by calculating the Stern-Volmer quench (K SV ) constant. 131 In the present study, a new approach is described to study the transport of proteins across silica nanotube membranes by using a fluorescence quenching-based analytical method. This analysis was accomplished by simply observing the fluorescence quenching of a donor-fluorescent labeled protein through a ligand-receptor binding mechanism. Thus, the flux of anti-Human IgG conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by using a known concentration of Human IgG FITC-labeled (H-IgG-FITC) as fluorescence emitter and donor in the permeation compartment. In order to preserve the chemical potential as the actual driving force for the flux, a much higher concentration of protein in the feed compartment (anti-H-IgG-QSY) was intentionally kept with respect to the concentration of emitter-labeled protein in the permeation compartment. It is described that the rate of fluorescence quenching depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of fluorescence quenching can be used to determine the concentration of the analyzed protein on the feed solution. The results of this investigation are reported here. Experimental Materials Commercially available nanopore alumina membranes (60 m thick, nominal pore diameter 100 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS) 152

PAGE 153

(Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes. The purified mouse immunoglobulin (M-IgG) immobilized within the silica nanotube membranes was obtained from Sigma Aldrich. Membranes modified with purified human immunoglobulin (H-IgG) obtained from Sigma Aldrich were also prepared. A trimethoxysilyl propyl aldehyde (United Chemical Technologies) was used to attach the M-IgG or H-IgG to the nanotubes. A purified human immunoglobulin conjugated with the FITC dye (H-IgG-FITC) obtained from Sigma Aldrich was used as emitter during the fluorescence quenching detection. A purified rabbit immunoglobulin conjugated with the FITC dye (R-IgG-FITC) obtained from Sigma Aldrich was used as control emitter during the fluorescence quenching detection. To investigate the transport properties across silica nanotube membranes the whole antibody raised in goat against H-IgG was used was used as permeating protein. This anti-human IgG (anti-H-IgG) with specific affinity for the F ab fragments of the H-IgG was obtained from Sigma Aldrich. In order to observe the fluorescence quenching of the donor-fluorescent labeled protein through a molecular recognition reaction, the anti-H-IgG permeating protein was conjugated with the QSY-7 quencher. This protein labeling was accomplished via acylation reaction of an N-Hydroxysuccinimidyl (NHS) ester of QSY-7 to the primary amine residues of the protein to form a stable amide bond. This NHS ester of QSY-7 was purchased from Molecular Probes (Eugene, OR). The BioGel P-30 resin provided in the amine-reactive protein labeling kit obtained from Molecular Probes was used for purification of the anti-H-IgG conjugated with QSY-7. SuperBlock blocking buffer solution in phosphate-buffered saline was obtained from Pierce. SuperBlock solution contains a proprietary protein for blocking binding sites for non-specific protein adsorption. Tween 20 was obtained from Sigma Aldrich. All other chemicals were of 153

PAGE 154

reagent grade and used as received. Purified water, obtained by passing house-distilled water through a Barnstead, E-pure water-purification system, was used to prepare all solutions. Preparation of the Silica Nanotube Membranes Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-based template-synthesis method described previously. Briefly, a sol-gel precursor solution was prepared by mixing absolute ethanol, TEOS and 1 M HCl (50:5:1 by volume), and this solution was allowed to hydrolyze for 30 minutes. The alumina membranes were then immersed into this solution for two minutes under sonication. The membranes were then removed and the surfaces swabbed with ethanol to remove precursor from the alumina surface. The membranes were then air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. This yields silica nanotubes, with wall thickness of ~3 nm, embedded within the pores of the alumina membrane. 9 Antibody Immobilization The immobilization chemistry has also been described previously. The aldehyde silane was first attached to the inside walls of the silica nanotubes. The membrane was then immersed into a solution that was 1.0 mg per mL either in M-IgG or in H-IgG dissolved in 10 mM phosphate-buffered saline with pH = 7.4 that also contained 150 mM sodium chloride and 5 mM sodium azide. The membrane was incubated overnight at 4 C, which allowed free amino sites on the antibody to attach via the Schiffs base reaction to the nanotube-immobilized aldehyde groups. After immobilization, the membranes were rinsed with buffer and then incubated for 5 hours in the SuperBlock solution that also contained 0.05% Tween 20. It is worth mentioning that the unmodified silica nanotube membranes were also incubated for 5 hours in the SuperBlock solution containing 0.05% Tween 20 before the transport experiments. 154

PAGE 155

Labeling of anti-H-IgG with QSY-7 Forty microliters of 6 mM QSY-7 solution prepared in absolute DMSO was added to a 1 mL of 13.3 M of anti-H-IgG solution prepared in 10 mM phosphate-buffered saline with pH = 7.4 that also contained 150 mM sodium chloride. This solution was mixed thoroughly by gentle vortexing. The reaction was incubated at room temperature for 2 hours on an orbital shaker platform. Unconjugated QSY-7 was separated from the protein-quencher conjugated by size exclusion chromatographic using BioGel P-30 resin equilibrated with 10 mM phosphate-buffered saline with pH = 7.4 that also contained 150 mM sodium chloride. The same buffer was also used as eluent. It should be pointed out that in order to have a separation of the elute protein with minimal sample dilution, the dimensions of the column separation, which was provided in the amine-reactive protein labeling kit, were approximately 20 cm in length and 0.5 cm in diameter. The final concentration and degree of labeling of the purified anti-H-IgG-QSY was determined spectrophotometrically by measuring the absorbance of the protein at 280 nm and the absorbance of the QSY-7 at 560 nm using a sample dilution factor of 10 times of the original purified fraction. The protein concentration was calculated by the Beer-Lambert law using a correction factor of 0.22 for the absorbance of the QSY-7 at 280 nm, 302 and a molar extinction coefficient ( IgG ) for the IgG of 210,000 M -1 cm -1 in the Equations 5-1 and 5.2. 302 )22.0(560280AAAproteinconjugated (5-1) IgGIgGproteinconjugatedCbA (5-2) The degree of labeling in the protein was also determined by using the Equation 5-3. )()(][]7[11711560cmMAcmMAIgGQSYQSYproteinconjugatedIgG (5-3) It was found a degree of labeling of 4.6 by using 90,000 M -1 cm -1 as the molar extinction coefficient ( QSY-7 ) for the QSY-7 quencher molecule. 302 155

PAGE 156

Transport Experiments The bare silica nanotube membranes, M-IgGor H-IgG-modified membranes were sandwiched between two pieces of Scotch tape that had 0.31 cm 2 area holes punched through them. These holes defined the area of membrane exposed to the contacting solution phases. This membrane assembly was mounted between the two halves of a U-tube permeation cell, with half-cell volumes of 10 mL. The feed solution contained the quencher conjugated antibody (anti-H-IgG-QSY) dissolved in 10 mM phosphate-buffered saline pH 7.4 that also contained 150 mM sodium chloride and 5 mM sodium azide (buffer used for all studies). The receiver solution on the other side of the membrane contained the buffered solution of the fluorescence emitter labeled protein the H-IgG-FITC. The concentration of the transporting protein in the feed solution was in all cases higher than the emitter-donor-protein in the receiver side. Thus, the chemical potential was preferentially preserved from the feed to receiver solutions as the actual driving force for the flux. A protocol was established to carry out all the transport experiments. First, the membrane mounted in the U-tube permeation cell was equilibrate for one hour in the buffer solution, and the receiver solution were frequently measured using a flow-through fluorescence detector until a stable base line for the fluorescence signal was reached. After that, a known concentration of anti-H-IgG-QSY was placed in the feed solution, and it was permitted to equilibrate for another fifteen minutes. Then, a known concentration of H-IgG-FITC was put in the receiver solution, after five minutes when the fluorescence signal was stable the experimental data for the transport experiment started being recorded. The changes of fluorescence intensity in the receiver solution were periodically measured using a flow-through fluorescence detector (Figure 5-1). 156

PAGE 157

Figure 5-1.Flow-through fluorescence system used to measure the changes of fluorescence intensity in the receiver solution. This system consisted of an L-7485 fluorescence detector (Hitachi). The flow stream was delivered using an L-7100 HPLC pump (Hitachi). We discovered that if the receiver solution was sent continuously through the fluorescence detectors during the entire course of a transport experiment, significant photobleaching of the FITC dye occurred. To prevent this, we turned the pump on to deliver receiver solution to the detectors for only 3 minutes per hour of transport experiment. That is, we used a duty cycle of 57 minutes with pump off followed by 3 minutes with pump on (flow rate = 3 mL per min) to collect fluorescence data. The whole system was controlled using Chromeleon 6.50 chromatography management software (Dionex) through a UCI-100 universal chromatography interface (Dionex). FITC was excited at 495 nm, and the emission was detected at 515 nm. Calibration curve was obtained with the same flow through system on solutions containing known concentrations of the flourophore-labeled antibody. The calibration curve was used to obtain the correct concentration of the unquenched fluorescent labeled protein in solution. Results and Discussion Experimental Principle Used for the Fluorescence Quenching Detection For these studies, the transport analysis design of the anti-H-IgG-QSY across silica nanotubes was mainly based on the FRET process. Therefore, the analysis required the use of a 157

PAGE 158

fluorescence donor (FITC) and fluorescence acceptor (QSY-7) molecules, which their structures are shown in Figure 5-2. Figure 5-2.Structures of (a) QSY-7 and (b) FITC One of the basic conditions for the FRET process to occur is that the emission spectrum of the donor must overlap the absorption spectrum of the acceptor. Figure 5-3 illustrates both the excitation and emission spectra for FITC and the absorption spectrum for QSY-7. Figure 5-3.Spectra for FITC and QSY-7 It is clearly seen how the emission spectrum of the FITC fluorophore overlaps significantly the absorption spectrum of the QSY-7 dye. It demonstrated that these selected pair of donor and acceptor molecules fulfills the overlapping condition for the FRET process required for the studies described here. The QSY-7 quencher is a nonfluorescent acceptor fashioned to quench specifically the fluorescence within the range of 500 to 600 nm. 303 The FRET efficiency is a distance-dependent interaction process, and the FRET efficiency is related to the Frster radius 158

PAGE 159

(R o ), and this radius corresponds to the distance at which fluorescence resonance energy transfer from the donor dye to the acceptor dye is 50% efficient (Equation 5-4). 131, 298 quencher and efluorophorbetween distance r 1160RrE (5-4) The magnitude of R o is dependent on the spectral properties of the donor and acceptor dyes and the typical value reported for the FITC and QSY-7 pair is 61 303 The suitable proximity for the fluorescence quenching process between these fluorescence donor and acceptor molecules was accomplished through a molecular recognition reaction using two antibodies with specific affinity for each other. This antigen-antibody immunoassay is illustrated in Figure 5-4. Figure 5-4.Molecular recognition reaction used to promote the fluorescence quenching. Thus, the flux of anti-H-IgG-QSY was monitored by using a known concentration of H-IgG-FITC as fluorescence emitter in the permeation compartment. Verification of the Fluorescence Quenching System via H-IgG-FITC and anti-H-IgG-QSY Reaction in Solid State In order to verify that the molecular recognition reaction between H-IgG-FITC and anti-H-IgG-QSY leads to an inhibition of the fluorescence process, a silica nanotube membrane was modified with H-IgG-FITC. The fluorescence signal for this modified membrane was observed under microscope before and after six hours of exposure to a solution of anti-H-IgG-QSY that 159

PAGE 160

was 100 nM in protein. The results of the fluorescence images and fluorescence spectra of the H-IgG-FITC-modified membrane are shown in Figure 5-5. Figure 5-5.Fluorescence image and spectra respectively for a H-IgG-FITC-modified membrane, (A-C) before exposure, and (B-D) after exposure to a sample solution of anti-H-IgG-QSY. It was observed that the fluorescence intensity of the H-IgG-FITC-modified membrane is completed inhibited through the immunoassay with anti-H-IgG-QSY. These results confirmed that the fluorescence signal of H-IgG-FITC can be quenched by anti-H-IgG-QSY by taking advantage of the antibody-antigen molecular recognition reaction, which situates the fluorescence donor and acceptor molecules within the intimate distance required for the quenching process. Control Transport Studies In order to demonstrate that the transport of proteins across silica nanotube membranes can be measured using fluorescence quenching detection through a biomolecular recognition reaction 160

PAGE 161

between a fluorescence donor fluorescentlabeled antibody and a fluorescence acceptor quencherconjugated antibody, two different types of controls were carried out. In the first control experiment (control 1), the fluorescence of a 3.35 nM solution of R-IgG-FITC was measured in the buffered receiver solution, whereas a 33.3 nM solution of the quencher-conjugated protein (anti-H-IgG-QSY) was placed in the buffered feed solution. For the second control (control 2), the fluorescence of a 3.35 nM solution of H-IgG-FITC was measured in the buffered receiver solution, but in this case, the protein place in the buffered feed solution was the anti-H-IgG without QSY-7 conjugation. The same concentration of protein, a 33.3 nM of anti-H-IgG was also used in the buffered feed solution during the control 2. The results for these control experiments are shown in the Figure 5-6. Figure 5-6.Typical plots for (A) changes of fluorescence signal, and (B) concentration of unquenched protein in function of time for two different control transport/detection systems: (a) anti-H-IgG/H-IgG-FITC, and (b) anti-H-IgG-QSY/R-IgG-FITC. Figure 5-6A shows the changes of fluorescence signal in the receiver solution during the course of the transport experiments both for control 1 and control 2. It was observed that for both controls the fluorescence signal is reduced as the time of the transport experiment increases. Since any of these two controls comply with the conditions required for FRET, this inhibition in fluorescence signal was attributed mostly to two factors different of FRET process. One factor 161

PAGE 162

was endorsed to photobleaching of the fluorophore, and the second to a partial diffusion of the FITC-labeled proteins toward the direction of the feed solution. By assuming that during the time of the transport experiment the degree of labeling (DL) remains constant for the fluorescently protein, calibration data allowed us to convert these fluorescence intensities to molar concentration of unquenched protein in the receiver solution (Figure 5-6B). The observed variations in scale for the plots in Figure 6-6A and Figure 6-6B are due to the difference in the DL between H-IgG-FITC and R-IgG-FITC proteins (sensitivity). The DL for H-IgG-FITC was 4.2, whereas it was 4.7 for R-IgG-FITC. The plots in Figure 6-6B give the impression that the deactivation of the fluorescence intensity for these two controls was carried out via two different mechanisms. In the case of the anti-H-IgG-QSY/R-IgG-FITC pair, it was observed that during the initial hour of the experiment the rate of fluorescence quenching for this system was higher than for the anti-H-IgG/H-IgG-FITC pair. Although, there is not a specific affinity for these two proteins (control 1), this effect might be assigned to the possibility that the donor (FITC) and acceptor (QSY-7) fluorescence label moieties in the proteins collided with each other to deactivate the fluorescence intensity faster than the other control system, for which the presence of QSY-7 quencher is absence. The anti-H-IgG-QSY/R-IgG-FITC pair (control 1) was chosen as the control for the transport experiments performed through the specific molecular recognition reaction proposes in this approach. Effect of the Feed Solution Concentration on the Rate of Quenching The experimental system demonstrated here, relies on the strength of the binding affinity and interaction between the labeled antibodies to bring together the fluorophore and quencher to an appropriated distance in order that the fluorescence quenching occurs. To study the flux across silica nanotube membranes using this fluorescence quenchingmediated detection, we 162

PAGE 163

carried out experiments to determine the effect of the feed solution concentration on the rate of fluorescence quenching. For these experiments, the concentration of H-IgG-FITC in the receiver solution was 3.35 nM in all cases, whereas three different concentrations for anti-H-IgG-QSY were used in the feed solution. The three concentrations of anti-H-IgG-QSY were respectively, 16.6 nM, 33.3 nM, and 66.6 nM. These results are shown in Figure 5-7. Figure 5-7.Typical plots for (A) changes of fluorescence signal and (B) concentration of unquenched protein in function of time for three different feed solution concentrations: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control). Figure 5-7 shows the influence of feed solution concentration on the rate of fluorescence quenching in the receiver solution. As can be observed from Figure 6-7, transport experiments with higher concentration of anti-H-IgG-QSY yields to a faster rate of the fluorescence quenching. These results confirmed that the molecular recognition reaction between the labeled antibodies generated the appropriate distance between the donor (FITC) and acceptor (QSY-7) molecules necessary for this fluorescence quenching-mediated detection. This was also verified by comparing this effect with the control experiment (plot d in Figure 5-7). The significant differences in fluorescence intensity for the experiments shown in Figure 5-7were observed only during the initial four hours of the transport experiments. After this period of time, the changes in 163

PAGE 164

fluorescence signal were less evident. This behavior can be rationalized in that only a small fraction of the labeled proteins fulfilled an appropriate orientation of the dipole-dipole interaction between the donor-acceptor molecules for FRET. However, the possibility of contact quenching should be considered as well. Data Analysis Based on the Stern-Volmer Plots The role of fluorescence quenching can be experimentally studied by determining the quenching rate parameters using the Stern-Volmer plots that are drawn in accordance with the Stern-Volmer equation (Equation 5-5) 131 QKIISV10 (5-5) From the experimental data of this investigation, the terms I 0 and I in Equation 5-5 are known, and they are respectively the fluorescence intensity in the absence and presence of the quencher-conjugated protein. However, the concentration of the quencher-conjugated protein (anti-H-IgG-QSY) is not absolutely defined. Therefore, some assumptions were made in order to quantitatively describe the fluorescence quenching process in terms of the Stern-Volmer equation. By assuming that during the time of the transport experiment the degree of labeling (DL) remains constant for the fluorescently protein, the fluorescence intensities were converted to molar concentration of unquenched protein via calibration curve. It was also assumed that the reaction between anti-H-IgG-QSY and H-IgG-FITC during the quenching phenomenon was 1:1, since the molar concentration of unquenched protein was already established. Besides, it was assumed that the differences in molar concentration of unquenched protein with time represent the concentration of the quencher-conjugated protein (anti-H-IgG-QSY) that was able to have the right conditions to interact with the donor-labeled protein (H-IgG-FITC) to yield the quenching 164

PAGE 165

process. Thus, the I 0 /I ratio was associated with the assumed concentration of anti-H-IgG-QSY in the Stern-Volmer plot (Figure 5-8). Figure 5-8.Stern-Volmer plots of I 0 /I against concentration of anti-H-IgG-QSY in the receiver solution transported across silica nanotube membranes. The effect of the feed solution concentrations of anti-H-IgG-QSY is shown: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control). Figure 5-8 shows the linear segment of the Stern-Volmer plots from the data compiled in Figures 5-7A and 5-7B. The linear part of the data analysis was used with the purpose of comparing the slopes of these Stern-Volmer plots (K SV ) as a function of the feed solution concentration of anti-H-IgG-QSY (Table 5-1). The insert in Figure 5-8 displays the complete range of the data analysis. It should be mentioned that those plots were not linear, and a positive deviation in all of them was observed, which is clearly noticed in the insert in Figure 5-8. The no linearity of the plots suggested that the observed fluorescence quenching is not merely collisional; thereby the contribution of contact quenching should be also considered. It was found that the rate of fluorescence quenching depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of fluorescence quenching can be used to determine the concentration of the analyzed protein on the feed solution. Figure 5-9 and Table 5-1 summarize these results. 165

PAGE 166

The slopes (K SV ) calculated from the linear segments of the Stern-Volmer plots were compared on function of the feed solution concentration of anti-H-IgG-QSY (Table 5-1). The values obtained from the slopes (K SV ) confirmed that higher fluxes lead to faster rates of quenching. Figure 5-9.Effect of the feed concentration on the Stern-Volmer quenching constant K SV A linear relationship was found when the Stern-Volmer quenching constant (K SV ) was plotted versus the concentration of anti-H-IgG-QSY in the feed solution; this plot is shown in Figure 5-9. These data demonstrated that the linear relationship between K SV and feed solution concentration can be used to determine the concentration of the analyzed protein on the feed solution. Observing the values of K SV an additional conclusion that might be inferred, is that they can be associated with the binding affinity (nM -1 ) for the permeating protein to the fluorescent labeled antibody in the receiver solution. It should be pointed out that K SV is associated with the rate constant of the dynamic reaction between quencher and fluorophore. Therefore, this analytical approach proposed in these studies might also be used to estimate binding affinities of antibodies in solution. 166

PAGE 167

Table 5-1.Relationship between the Stern-Volmer quenching constant (K SV ) with the feed solution concentration of anti-H-IgG-QSY Feed Concentration of anti-H-IgG-QSY (nM) K SV (nM -1 ) 0 (control) 0.369 16.6 0.431 33.3 0.491 66.6 0.621 Data Analysis Based on log-log Plots of Fluorescence vs. Time A second and much simpler analysis for the obtained data was based on the logarithmic plots of the changes of fluorescence in function of time. The results for this analysis are shown in Figure 5-10. Figure 5-10.log-log plots of the of fluorescence signal vs. time showing the effect of feed solution concentration of anti-H-IgG-QSY: (a) 16.6 nM, (b) 33.3 nM, (c) 66.6 nM, and (d) 33.3 nM (control). The respective log-log plots of the Figure 5-7A is shown in Figure 5-10. After the experimental data was represented in a logarithmic scale the relation between fluorescence and time was almost linear, which confirmed that the relationship between these two physical variables is described by a power law formula. It was observed that as the concentration of anti-H-IgG-QSY in feed solution increases the slope of the log-log plot tends toward more negative values as a result of faster rate of quenching. The trend observed with this simple analysis (log167

PAGE 168

log plots) confirmed also that transport experiments with higher concentration of anti-H-IgG-QSY yields to a faster rate of the fluorescence quenching. These results are summarized in Table 5-2. Table 5-2.Relationship between slope of the log-log plots for the fluorescence signal vs. time with the feed solution concentration of anti-H-IgG-QSY. Feed Concentration of anti-H-IgG-QSY (nM) Slope of log-log plots Intercept of log-log plots 0 (control) -0.1198 2.4360 16.6 -0.1374 2.3406 33.3 -0.1693 2.3488 66.6 -0.2095 2.3344 The relation between the slopes obtained from the log-log plots with the feed concentration is shown in Figure 5-11. Figure 5-11.Linear relationship between slopes obtained from the log-log plots with the feed concentration of anti-H-IgG-QSY. A linear relationship was also obtained between the slopes of the log-log plots with the feed concentration of anti-H-IgG-QSY. It showed excellent linearity (correlation coefficient = 0.996) over the range from 0 nM to 66.6 nM. Therefore, by this simple data analysis using log-log plots, it is also possible to determine the concentration of the analyzed protein on the feed solution by simply knowing the slope of the log-log plot. 168

PAGE 169

Transport across Antibody-Modified Silica Nanotube Membranes Transport experiments were also carried out across antibody-functionalized nanotube membranes. For these experiments silica nanotube membranes were modified both with H-IgG or M-IgG. The results for these experiments are compared in Figure 5-12. Figure 5-12.Typical plots for (A) changes of fluorescence signal and (B) concentration of unquenched protein in function of time illustrating the transport of anti-H-IgG-QSY across (a) H-IgG-, and (b) M-IgG-modified membranes. The feed solution concentration was 33.3 nM for both experiments. The data analysis for the transport across antibody-modified membranes based on Stern-Volmer plots and log-log plots are summarized in Table 5-3. Table 5-3.Data for the Stern-Volmer quenching constant (K SV ) and slope of the log-log plots for antibody-modified membranes Antibody-Modified Membrane* K SV (nM -1 ) Slope of log-log plots H-IgG 0.382 -0.0995 M-IgG 0.415 -0.1218 *The feed solution concentration of anti-H-IgG-QSY was 33.3 for both experiments. The results shown that the fluorescence quenching in the receiver solution was higher for the transport experiments carried out across M-IgG modified than H-IgG modified membranes. This is due to the faster rate of transport of anti-H-IgG-QSY across M-IgG as a result of the lower binding constant between these two proteins. Therefore, it demonstrated that antibody-modified membranes can be use to modulated the rate of fluorescence quenching for any specific 169

PAGE 170

pair of quencher-fluorophore labeled antibodies. These results confirmed that the observed fluorescence quenching phenomenon is a result of the biomolecular recognition reaction between anti-H-IgG-QSY and H-IgG-FITC in the receiver solution. Conclusions It was described a new method to study transport of proteins across silica nanotube membranes by using a fluorescence quenching-based sensing approach. This analysis was accomplished by simply observing the fluorescence quenching of a donor-fluorescent labeled protein (H-IgG-FITC) in the receiver solution after immunoassay reaction with the transported acceptor-fluorescent conjugated protein (anti-H-IgG-QSY). The flux of anti-Human IgG conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by observing the changes in fluorescence of a known concentration of Human IgG FITC-labeled (H-IgG-FITC) in the receiver solution. The rate of fluorescence quenching was studied both in the absence and the presence of quencher labeled protein transported across of bare silica nanotube membranes and antibody-modified membranes. It was demonstrated that the rate of fluorescence quenching depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of fluorescence quenching can be used to determine the concentration of the analyzed protein on the feed solution. 170

PAGE 171

CHAPTER 6 HYBRIDIZATION STUDIES OF DNA IN A PNA-MODIFIED NANOTUBE MEMBRANE PLATFORM Introduction The development of DNA-sensors devices have attracted substantial research efforts directed to gene analysis, detection of genetic disorders, analysis of single nucleotide polymorphism in genes, tissue matching and forensic applications among others. 304 307 Various techniques, including quartz crystal microbalance, 308 electrochemical, 309 fluorescence, 310 electrophoretic, 311 electrochemiluminescence, 312 enzymatic, 313 surface plasmon resonance, 314 and atomic force microscopy, 315 have been developed to detect DNA hybridization. However, the label-free detection methods have attracted more interest than other methods because they can be simple, rapid, highly sensitive, and inexpensive. 307 In 1953, Watson and Crick proposed a model for the structure of DNA. 316 This model predicted that DNA would exist as a helix of two complementary anti-parallel strands. Soon after the first description of the double helix by Watson and Crick, it was shown, that the two strands could be separated by heat or treatment with alkali. The reverse process, which underlies all the methods based on DNA renaturation or molecular hybridization, was first described by Marmur and Doty. 317 It was quickly established that the two sequences involved in duplex formation must have some degree of sequence complementarity and that the stability of the duplex formed depends on the extent of complementarity. These remarkable properties suggested ways to analyze relationships between nucleic acid sequences, and analytical methods based on molecular hybridization. Since then, these methods have been rapidly developed and applied to a variety of biological malfunctions. Base-pairing (A-T and G-C) or hybridization is the underlying principle of DNA-sensors, namely the strong interaction between two complementary strands. Accordingly, such devices 171

PAGE 172

rely on the immobilization of a single-stranded DNA molecule (the probe) for hybridizing with the complementary (target) strand in a given sample. Then, the specificity of DNA hybridization devices depends primarily on the selection of the probe (sequence and length) and secondary upon the hybridization conditions (mainly the temperature, and ionic strength as well). Peptide nucleotide acid (PNA) was introduced in 1991 as a structural mimic of DNA with the ability to invade double helical DNA and bind its complementary target sequence. 318 Due to its higher specificity in the recognition of complementary DNA; PNA has already established itself as an attractive recognition layer for DNA sensing. 319 Figure 6-1.Structures of (a) PNA, and (b) DNA. PNA is a DNA analogue with a neutral N-(2-aminoethyl) glycine backbone replacing the negatively charged phosphate backbone of DNA (Figure 6-1). 318 Since the spacing between the nucleotides is the same as in DNA, the conventional Watson-Crick base pairing rules apply between mixed PNA/DNA sequences. PNA highly discriminates mismatch DNA and has a stronger binding affinity for complementary DNA than its DNA counterpart. Then, the favorable hybridization properties of PNA have also inspired the incorporation of PNA into array and nucleic acid biosensor formats. 320 323 172

PAGE 173

This chapter describes the use of PNA-modified nanotube membranes to study a label-free detection for DNA hybridization. This was achieved by exploiting the unique structural and hybridization features of PNA. This approach was characterized by observing the changes in the membrane resistance using impedance spectroscopy. This method allowed carrying out an in situ characterization of the interfacial properties of the modified nanotube membranes during the hybridization process in the duplex formation PNA-DNA. The novel idea was based on the fact that PNA is a neutral molecule; thereby during hybridization events achieved in just highly purified water without any electrolyte in the solution, the only source of new charges in the system during the hybridization process emerge from the DNA molecules. This event exhibited a change of the membrane resistant that was detected by impedance spectroscopy which is an effective technique to probe the interfacial properties of modified electrodes. Experimental Materials Commercially available nanopore alumina membranes (60 m thick, nominal pore diameter 100 nm and 200 nm) were obtained from Whatman. Tetraethyl orthosilicate 99.999% (TEOS) (Sigma Aldrich) was used to deposit the silica nanotubes within the pores of these membranes. A trimethoxysilyl propyl diethylenetriamine (DETA) was used to introduce amino groups to the nanotubes was obtained from United Chemical Technologies. Sulfosuccinimidyl-6-(biotinamido) hexanoate (NHS-LC-Biotin) obtained from Pierce was used as biotin-streptavidin linker. A purified streptavidin was obtained from Sigma Aldrich. A biotinylated-PNA (PNA) with twelve nucleotides bases (BiotinGCG CTA GCC TAC) obtained from Applied Biosystems was used as the probe to detect DNA. To investigate the hybridization properties of the PNA-modified nanotube membranes a complementary twelve bases DNA (3CGC GAT CGG ATG -5) with a MW=3607 g/moles was used. This complementary target 173

PAGE 174

DNA was obtained from Alpha-DNA. All other chemicals were of reagent grade and used as received. Purified water, obtained by passing house-distilled water through a Barnstead, E-pure water-purification system, was used to prepare all solutions. Preparation of the Silica Nanotube Membranes Silica nanotubes were deposited in the pores of the alumina membrane using the sol-gel-based template-synthesis method. Briefly, a sol-gel precursor solution was prepared by mixing absolute ethanol, TEOS and water (40:5:10 by volume), and this solution was allowed to hydrolyze for 1 hour. The alumina membranes were then immersed into this solution for two minutes under sonication, and they were kept in this solution for ten additional minutes. The membranes were then removed and the surfaces swabbed with ethanol to remove precursor from the alumina surface. The membranes were then air dry for ten minutes at room temperature and cured in an oven for 12 hours at 150 C. Preparation of PNA-Modified Nanotube Membranes The biotinylated-PNA probe was immobilized on nanotube membrane surfaces using the streptavidin-biotin affinity system. 297, 324 Streptavidin is a 60-kDa protein purified from the bacterium Streptomyces avidini. The interest in streptavidin is the remarkable and unique interaction with its low-molecular weight ligand biotin (K d 10 -15 M). Because of its extremely strong affinity, the streptavidin-biotin affinity system has been developed as a general tool to use in various experimental systems. 325 The immobilization process used in this approach consisted of three basic steps. The first and second steps were respectively, a silanization which introduced amino groups to the surface, and conjugation between the amino groups and the NHS-LC-Biotin. The latter acylation reaction of an N-Hydroxysuccinimidyl (NHS) ester of biotin to the primary amino groups of the silane-modified surface yields to a stable amide bond. The next step was the coupling of streptavidin to 174

PAGE 175

the biotin-modified surface. It should be pointed out that the NHS-LC-Biotin had a spacer arm of 22.4 that can reduce steric hindrance associated with the binding many biotinylated molecules on one streptavidin, which in turns may enhance the sensitivity of the system. Figure 6-2.Immobilization of biotinylated-PNA on templated synthesized silica nanotube membranes. During silanization the silica nanotube membranes were immersed into freshly prepared 1% (v/v) solution of DETA in 1 mM acetic acid for 30 minutes at room temperature. The DETA-modified membranes were rinsed with water to remove the unreacted DETA. The amino-silane modified nanotube membranes were dried under nitrogen and fixed at 120 C in the oven for 15 minutes. For the biotinylation process the amino-silane modified membranes were exposed for 6 hours to a 0.5 mg/mL solution of NHS-LC-Biotin prepared in 0.1 M bicarbonate buffer pH 8.5. The biotinylated nanotube membranes were then incubated overnight at 4 C in a 1 mg/mL 175

PAGE 176

solution of streptavidin prepared in 10 mM phosphate buffer pH 7.4. The PNA immobilization was carried out by incubating the streptavidin-modified membranes for 8 hours in a solution that contained 1 M of biotinylated PNA dissolved in water at 4 C. A scheme showing the immobilization steps for the PNA-modified nanotube membranes and the hybridization reaction with the complementary DNA is shown in Figure 6-2. Nanotube Membrane Assembly After the silanization with DETA, the silica nanotube membranes were assembled in a suitable format for their characterization with impedance spectroscopy. In order to make the nanotube membrane conductive, they were sputter-coated for 90 sec on both sides with Gold-Palladium at 75 mTorr of system pressure, 45 mA of current, conditions that gave a coating of approximately 10 nm thicknesses. The metal sputtered on the membranes enhanced the conduction properties of the membrane, and provided the facilities for the characterization of the membranes. Figure 6-3.Nanotube membrane assembly. The sputtered membranes together with two thin platinum leads of 60 microns thickness (one in each opposite side of the membrane) were placed between two pieces of plastic tape punched with a hole of 0.31 cm 2 area. The two platinum leads were sandwiched between the two pieces of tape in such a way that the holes were aligned and the platinum leads were covered 176

PAGE 177

with the tape to avoid their contact with the solution during the impedance measurements. A schematic illustration for this assembly is shown in Figure 6-3. These holes in the assembled nanotube membranes defined the area of the nanotube membrane exposed for the subsequent modification steps and the area exposed during the impedance measurements. It is worth noticing that the external electrical connections involves dry metal-to-metal contacts, so that, the membrane itself performs the function of the electrodes (working and reference). Impedance Measurement for Membrane Characterization For the impedance measurements the assembled membrane was mounted between the two halves of a U-tube permeation cell. This sputter-coated nanotube membrane performed and replaced the function of the typically used remote electrodes located on either side of the membrane. For all the impedance experiments both half-cells were filled with 10 mL of highly purified water. During the hybridization experiments one of the half-cell was chosen to spike the concentration of DNA target, this side also belonged to the side where the reference electrode connections were placed. The impedance data was obtained using a Solartron 1287 electrochemical interface module connected to a Solartron 1255B frequency response analyzer and a personal computer. The impedance data was recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz. Z-plot and Z-view software packages (Scribner Asc. Inc., NC) were used to control the impedance experiments and analyzed the impedance data, respectively. Impedance Data Analysis The main interest of in these studies was to monitor the changes in the membrane resistance during the hybridization process in the duplex formation PNA-DNA. Even though, 177

PAGE 178

that the experimental impedance data revealed the contribution of Warburg impedance, in order to simplify the data analysis this contribution was not considered. In addition, since this represent a bulk property of the electrolyte solution (diffusion of ions from the bulk electrolyte); it is not affected for the chemistry modification occurring at the surface of the pores in the membranes. Therefore, the experimental data of impedance were analyzed by using the model equivalent circuit shown in Figure 6-4. Figure 6-4.General equivalent circuit model used for impedance data analysis. A scan be seen from Figure 6-4, the data were fitted using a constant phase element (CPE). The reason of this is because in impedance experiments the capacitors often do not behave ideally. 326 This CPE is defined by the Equation 6-1. njACPEZ)()( (6-1) Where j is an imaginary number, is the angular frequency, and A is a constant. When this equation describes a capacitor, the constant A is equal to the inverse of the capacitance, and the exponent n is equal to 1. For a CPE the exponent n is less than 1. Thus, by fitting the experimental data with the equivalent circuit shown in Figure 6-4, the changes in membrane resistance, R m were extracted. Results and Discussion In order to evaluate the effects of the various immobilization steps on the surface of the sputter-coated nanotube membranes, impedance measurements were carried out during the all steps of the immobilization process. All the assays were conducted in highly purified water. A biotinylated PNA probe with a sequence BiotinGCG CTA GCC TAC, which is fully 178

PAGE 179

complementary to the DNA target with a sequence 3CGC GAT CGG ATG -5, was used as receptor in the nanotube membrane sensing-platform. These studies begin, however, with the impedance measurements for the alumina template, and silica nanotube membranes and investigations of the equilibration times required for the impedance measurements. Impedance Characterization of the Unmodified Membranes The Nyquist plots for bare alumina template (200 nm nominal pore size) and silica nanotube membranes are shown in Figure 6-5. The x-axis and y-axis in this plot corresponds to real impedance, Z re and imaginary impedance, Z im respectively. Z re and Z im are mainly originated from the resistance and capacitance of the cell system, respectively. Figure 6-5.Nyquist plots (Z im vs. Z re ) for impedance measurement of (a) alumina and (b) silica-nanotube assembled membranes in highly purified water. As shown in Figure 6-5, changes after the deposition of the silica nanotubes in the alumina template were observed in semicircle part at high frequency region (usually associated with kinetic controlled process), whereas the linear tail at lower frequencies (usually associated with mass transfer process) maintained almost the same slope. The variations in the high frequency region were probably due to a change in the isoelectric point (pI) of the nanotube membrane. Since it is known that the pI of the alumina membranes range from 6.9 to 10.5, while the pI for 179

PAGE 180

silica has been ranged from 1.9 to 4.5. 327 Because of the different electrical characteristics of these materials, a change in the surface characteristics of the membranes should be expected when silica nanotubes are synthesized within the alumina membrane, as was observed in the Nyquist plots in Figure 6-5. In order to assess the stability of the impedance measurements a study of the wettability of the bare alumina membrane was carried out by monitoring the changes in resistance with the time. These results are summarized in Figure 6-7. Figure 6-7.Effect of the wetting time on the resistance impedance measurement for an assembled unmodified alumina membrane 200 nm pore diameter. As can be seen from Figure 6-7, the changes in resistance from impedance measurements of alumina membrane were pretty stable with the time. These results confirmed the stability of the sputter-coated membranes and the stability itself of the impedance measurements. Functionalization of the Silica nanotube Membranes The differences in membrane resistance before and after each modification step during the immobilization process were analyzed in order to characterize the accomplishment of the various modifications (Figure 6-8). The obtained values of R m from the Nyquist plots in Figure 6-8 are compiled in Table 6-1. 180

PAGE 181

The results revealed that the different immobilization steps introduced significant changes in the membrane resistance. The most significant change was found in the membrane resistance associated with the hybridization process of DNA with the PNA-modified nanotube membrane. This decrease in resistance was associated with the increase of negative surface charge density due to the duplex formation PNA-DNA on the surface. Figure 6-8. Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water: (a) alumina, (b) after synthesis of silica-nanotube, (c) after silanization, (d) after biotinylation, (e) after streptavidin immobilization, (f) after PNA immobilization, and (g) after DNA hybridization. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz. The increase of R m after the deposition of silica nanotubes was consistent with the fact that after deposition smaller effective pore size in the membrane might be reached. Conversely, the decrease in membrane resistance after silanization was explained through the formation of protonated amino groups on the surface at the pH of the experiment (highly purified water). However, part of this amino protonated groups were neutralized by the conjugation reaction with the biotin linker, which was confirmed for the increase in resistance after the biotinylation. The 181

PAGE 182

coupling of streptavidin also produced a significant change in resistance, since the size of this protein reduced the effective pore size of the membrane. For the immobilization of the biotinylated PNA a small decrease in resistance was observed compared with the streptavidin coupling. This effect was difficult of explain, but it might be due to the fact that PNA oligomers are purified by reverse phase HPLC in the presence of trifluoroacetic acid. The aqueous solutions of PNA posses a pH lower than 3. Therefore, at this low pH some residual positive charge in the structure of PNA might be introduced in the membrane that causes the reduction of the resistance. Table 6-1.Variations of R m at the different steps of modification on a 100 nm pore diameter alumina template Modification step R m () Alumina membrane 8100 Silica nanotube membrane 12500 Amino-modified membrane 1030 Biotinylated-modified membrane 4515 Streptavidin-modified membrane 18000 PNA-modified membrane 14717 DNA hybridization 327 Impedance Sensing of DNA-PNA Hybridization Figure 6-9 shows the Nyquist plots of the PNA-modified nanotube membrane before and after 10 minutes of the hybridization process. As can be seen after the addition of 3 nmoles of the complementary DNA to the half-cell containing highly purified water (side where the reference electrode connections were also sited) a significant decrease of the semicircle sector of the plot was observed. This result suggested that the resistance of the PNA-modified nanotube membrane became lower due to the increase in new negative charge in the system brought to the membrane surface for the negatively charged phosphate backbone of DNA molecules during the 182

PAGE 183

hybridization process. There was almost no change in the solution resistance (R s ) and Warburg impedance (Z W ), which demonstrated that they were not affected during the hybridization process on the membrane surface. Figure 6-9. Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water for a PNA-modified nanotube membrane: (a) before, and (b) after 10 minutes of DNA hybridization. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz. Effect of the Pore Size on R m The influence of the pore size on the membrane resistance was explored by using two different sizes of commercially available (Whatman) nanopore alumina membranes having 60 m thickness, and nominal pore diameters respectively of 200, and 100 nm. To study this effect, nanotube membranes of the specified pore size were modified with PNA. For the hybridization impedance experiments, 3 nmoles of the complementary DNA were added to the highly purified water in both cases (Figure 6-10). 183

PAGE 184

Figure 6-10.Effect of the membrane pore size on the membrane resistance: (A) 200 nm, and (B) 100 nm. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz. Figure 6-10 compares the streptavidin immobilization, PNA immobilization, and DNA hybridization processes, respectively, by using membranes with two different pore diameters (100 and 200 nm). The changes in membrane resistance during the immobilization steps for the 100 nm membrane showed the same trend observed for the 200 nm pore diameter membrane explained above. However, the resistance was for all modification steps higher in the 100 nm pore diameter membranes, as a result of the smaller pore size. In the case of the PNA immobilization, the resistance was increased in approximately 280% by using the 100 nm membrane. The effect of the pore size during the hybridization process revealed a reduction of 98% in the resistance of the membrane for the 200 nm-situation (Figure 6-10A), while for the 100 nm184

PAGE 185

situation the reduction was of 53% (Figure 6-10B). These differences in the resistance during the hybridization process might be attributed to the fact that membranes with higher resistance present longer dynamic range. However, the possibility of a higher immobilization of PNA might be also considered. Stability and Reproducibility of the PNA-modified Membranes The stability of the PNA-modified membrane was evaluated by comparison the effect of DNA hybridization in time with an unmodified control membrane. These results are compared in Figure 6-11. Figure 6-11.Study of the hybridization process stability with time: (A) unmodified control alumina membrane, and (B) PNA-modified membrane. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA. Figure 6-11 shows the stability of the impedance measurement for a control (Figure 6-11A), and PNA-modified membrane (Figure 6-11B) after adding 3 nmoles of complementary DNA to the experimental cell. It is clearly seen that the changes in membrane resistance are mainly due to the hybridization process (Figure 6-11B). However, some reduction on the membrane resistance was observed for the control membrane after the addition of 3 nmoles of DNA. This effect might be attributed to nonspecific interactions on the membrane surface. The 185

PAGE 186

changes in membranes resistance after 15 hours of adding DNA were 27098 ohms for the PNA-modified membrane, while this change was only 4271 ohms for the control unmodified membrane. However, after 20 minutes these changes were 25152 and 1806 ohms for the PNA-modified and unmodified membranes, respectively. These results clearly demonstrated that the assembled PNA-modified membrane configuration study herein might be used as a good platform for DNA sensors. An additional evidence of this conclusion can be observed in Figure 6-12 were the effect of DNA concentration on the impedance measurements is illustrated. Figure 6-12.Nyquist plots (Z im vs. Z re ) of impedance measurement in highly purified water for a PNA-modified nanotube membrane: (a) before, (b) after adding 3 nmoles of DNA, and (c) after 6 nmoles of DNA. The plots were recorded applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz. The reproducibility of the PNA-modified nanotube membranes were also evaluated (Figure 6-13). For these hybridization studies three different membranes modified with PNA were compared. Figure 6-13 shows the results of membrane resistance for these three PNA-modified membranes, before and after hybridization with DNA (20 minutes and 3 days). It was observed that after 20 minutes of adding 3 nmoles of the complementary DNA solution to each of the PNA-modified membranes, the reduction of the resistance membrane were almost in the same order of magnitude. The reduction in resistance for membrane 1, 2, and 3 in Figure 6-13 were 186

PAGE 187

54%, 42%, and 53%, respectively. Amazingly was to observe the stability of these membranes after three days of the duplex formation during the hybridization process. Figure 6-13.Reproducibility study of the PNA-modified nanotube membranes, before and after the hybridization process. The resistance values were obtained from the impedance data recorded by applying 5 mV alternative voltage in the frequency range from 1 MHz to 0.05 Hz after the addition of 3 nmoles of complementary DNA. The results presented in these studies demonstrated the good stability and reproducibility of the PNA-modified nanotube membrane describe here. Conclusions It was described the use of PNA-modified nanotube membranes to study a label-free detection for DNA hybridization. This was achieved by exploiting the unique structural and hybridization features of PNA. This approach was characterized by observing the changes in the membrane resistance using impedance spectroscopy. This method allowed carrying out an in situ characterization of the interfacial properties of the modified nanotube membranes during the hybridization process in the duplex formation PNA-DNA. The novel idea was based on the fact that PNA is a neutral molecule; thereby during hybridization events achieved in just highly purified water without any electrolyte in the solution, the only source of new charges in the system during the hybridization process emerge from the DNA molecules. This event exhibited a 187

PAGE 188

change of the membrane resistant that was detected by impedance spectroscopy which is an effective technique to probe the interfacial properties of modified electrodes. These results showed the feasibility of using PNA-modified nanotube membranes as a sensing-platform for DNA hybridization methods. This assembled nanotube membrane platform present the special characteristic that the membrane itself performed the function of the electrodes (working and reference respectively). It was demonstrated that PNA immobilized on the nanotube membranes function as a good probe to detect the DNA targets in condition where no electrolyte is present during the hybridization process. It was also demonstrated the good reproducibility obtained during the fabrication and detection with this novel sensing-platform. 188

PAGE 189

CHAPTER 7 CONCLUSIONS The main goal of the research presented in this dissertation has been to develop selective nanotube membranes based on affinity molecular recognition with interesting applications in biological analysis. The unique property of biomolecules and metal ions to recognized specific molecules was exploited to elaborate functional nanotube membranes with potential application in sensing and separations of proteins and DNA. The template synthesis method was use to obtain the silica nanotube within the pores of alumina membranes. The walls of these nanotubes were functionalized via silane chemistry to couple molecular recognition agents onto the modified surfaces. Chapter 1 introduce the background information for this dissertation including the template synthesis method, the description of the alumina template membranes, sol-gel and silane chemistry, the inspiration of living transport in the development of membrane-based separations, fundamental about proteins, and a synopsis of the instrumental techniques used during the development and characterization of the applications describe in this dissertation. In chapter 2, a detailed experimental study on antibody-functionalized nanotube membranes was carried out, including the effect of surface modification, pore size, feed concentration, temperature, and ionic strength, both on the transport as on the functional behavior of the membranes. It was demonstrated that the type of immobilized antibody changes the functional behavior of the antibody-modified membranes and the transport properties of the permeating proteins across them. These membranes showed a Langmuirian shape plot characteristic of a typical behavior of facilitated transport membranes, when the effect of feed solution concentration on the fluxes of the permeating proteins was investigated. Therefore, the antibody-functionalized membranes showed here exhibit a type of behavior for facilitated-transport membranes during protein separations. Important parameters such as the binding 189

PAGE 190

capacities of the antibody-modified membranes and the binding constants for the permeating proteins to the antibody carriers confirmed that the permeating protein anti-M-IgG binds strongly to the carrier M-IgG and that the permeating protein anti-H-IgG binds weakly to this carrier, likewise, the permeating protein anti-H-IgG binds strongly to the carrier H-IgG and that anti-M-IgG binds weakly to this carrier. It was also demonstrated that by controlling the orientation of the immobilized antibody using protein G the potency affinity of the antibodies is effectively preserved. Through these studies was also demonstrated that by selecting a template membrane with an appropriated pore size diameters these antibody-modified nanotube membranes have the ability to be used with a combination of both size and affinity selectivity for protein separation. Additional studies on the effect of temperature on the fluxes and the binding constant demonstrated that the functional behavior of these antibody-modified membranes is governed for the changes in free energy before and after interaction of the permeating proteins with the carrier antibody, which was attributed to the changes in enthalpy and entropy of the system. Other important finding during this investigation was that by controlling the ionic strength of the transport solution, these antibody-modified nanotube membranes can be used for situations were the transport of an unwanted protein needs to be shut down, in order to improve selectivity. This latter approach might be useful in the sample complexity reduction in proteomic analysis. Since the analysis and characterization of complex mixtures of proteins are the central aims of proteomics. It is also noteworthy to mention the low pM detection limits achieved during these studies. In general, the ability to alter the surface biochemistries and thus tailor the membrane for specific analytes, promises the use of these membranes for the analysis of a wide range of proteins. 190

PAGE 191

Chapter 3 describes the modeling and characterization of the transport properties of proteins across antibody-functionalized nanotube membranes by using the dual-mode model. This work demonstrated that the dual-mode model proposes here was a suitable model for the characterization of the transport properties of anti-H-IgG and anti-M-IgG in M-IgG modified membranes. The model demonstrated a good agreement with the experimental system studied here. Besides, the findings showed in these studies allow addressing the concept of modulated facilitated transport systems using antibody-modified nanotube membranes, and how the extent of the facilitation is determined and controlled by the binding affinity between carrier and permeating protein molecules. These studies can provide a new direction in the development of more effective membranes for membrane-based separation. In chapter 4, the development of metal ion affinity nanotube membranes was demonstrated by exploiting the concept of IMAC separations in a nanopore-membrane-platform. Silica nanotube membranes were functionalized with imidazole through a nucleophilic substitution reaction between 2-mercaptoimidazol and 3-chloropropyl modified silica nanotubes carried out in an aprotic medium. It was demonstrated that transport of histidine-tagged recombinant proteins across these imidazole-modified silica nanotube membranes is controlled by the nature of immobilized metal ion on the imidazole surface. Besides, the extensive characterization of the imidazole functionalized membrane allowed the understanding of how coordination chemistry with different divalent metal ions affects the functional behavior of the membrane and transport properties during the permeation of six-histidine recombinant proteins. Furthermore, additional results shown that by the appropriate choice of the transition metal ion for coordination with the imidazole ligand, the binding strength and optical properties within the functional membrane are also modified. Hence, these studies revealed how the presence of the immobilized metal ion 191

PAGE 192

affects the functionality of the imidazole membrane and its transport properties. The metal ion affinity nanotube membrane described here has the potential to be a versatile analytical platform, where different metal ions can be immobilized on the same prevailing conditions, additionally offering the possibilities of regeneration without apparent loss of the metal coordination properties of the functional membrane. These studies demonstrate an alternative way of immobilized metal ion affinity by using nanotube membranes. Chapter 5 described a new method to study transport of proteins across silica nanotube membranes by using a fluorescence quenching-based sensing approach. This analysis was accomplished by simply observing the fluorescence quenching of a donor-fluorescent labeled protein (H-IgG-FITC) in the receiver solution after immunoassay reaction with the transported acceptor-fluorescent conjugated protein (anti-H-IgG-QSY). The flux of anti-Human IgG conjugated with the quencher QSY-7 (anti-H-IgG-QSY) was monitored by observing the changes in fluorescence of a known concentration of Human IgG FITC-labeled (H-IgG-FITC) in the receiver solution. The rate of fluorescence quenching was studied both in the absence and the presence of quencher labeled protein transported across of bare silica nanotube membranes and antibody-modified membranes. It was demonstrated that the rate of fluorescence quenching depends on the flux of anti-H-IgG-QSY across the nanotube membrane, and that the rate of fluorescence quenching can be used to determine the concentration of the analyzed protein on the feed solution. In chapter 6, the use of PNA-modified nanotube membranes to study a label-free detection for DNA hybridization was demonstrated. This was achieved by exploiting the unique structural and hybridization features of PNA. This approach was characterized by observing the changes in the membrane resistance using impedance spectroscopy. This method allowed carrying out an 192

PAGE 193

in situ characterization of the interfacial properties of the modified nanotube membranes during the hybridization process in the duplex formation PNA-DNA. The novel idea was based on the fact that PNA is a neutral molecule; thereby during hybridization events achieved in just highly purified water without any electrolyte in the solution, the only source of new charges in the system during the hybridization process emerge from the DNA molecules. This event exhibited a change of the membrane resistant that was detected by impedance spectroscopy which is an effective technique to probe the interfacial properties of modified electrodes. These results showed the feasibility of using PNA-modified nanotube membranes as a sensing-platform for DNA hybridization methods. This assembled nanotube membrane platform present the special characteristic that the membrane itself performed the function of the electrodes (working and reference respectively). It was demonstrated that PNA immobilized on the nanotube membranes function as a good probe to detect the DNA targets in condition where no electrolyte is present during the hybridization process. It was also demonstrated the good reproducibility obtained during the fabrication and detection with this novel sensing-platform. 193

PAGE 194

LIST OF REFERENCES [1] Hagg, M.-B. Separation and Purification Methods 1998, 27, 51-168. [2] Baker, R. W. Membrane Technology and Applications, 2nd ed.; John Wiley & Sons, Ltd., 2004. [3] Striemer, C. C.; Gaborski, T. R.; McGrath, J. L.; Fauchet, P. M. Nature 2007, 445, 749-753. [4] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Science 1997, 278, 655-658. [5] Yu, S.; Lee, S. B.; Martin, C. R. Analytical Chemistry 2003, 75, 1239-1244. [6] Yu, S.; Lee, S. B.; Kang, M.; Martin, C. R. Nano Letters 2001, 1, 495-498. [7] Chun, K.-Y.; Stroeve, P. Langmuir 2002, 18, 4653-4658. [8] Van Reis, R.; Zydney, A. Current Opinion in Biotechnology 2001, 12, 208-211. [9] Lee, S. B.; Mitchell, D. T.; Trofin, L.; Nevanen, T. K.; Soederlund, H.; Martin, C. R. Science 2002, 296, 2198-2200. [10] Kohli, P.; Harrell, C. C.; Cao, Z.; Gasparac, R.; Tan, W.; Martin, C. R. Science 2004, 305, 984-986. [11] Giorno, L.; De Bartolo, L.; Drioli, E. Membrane Science and Technology Series 2003, 8, 187-217. [12] Gavalas, V. G.; Wang, J.; Bachas, L. G. Membrane Science and Technology Series 2003, 8, 379-392. [13] Sexton, L. T.; Horne, L. P.; Sherrill, S. A.; Bishop, G. W.; Baker, L. A.; Martin, C. R. Journal of the American Chemical Society 2007, 129, 13144-13152. [14] Baker, L. A.; Choi, Y.; Martin, C. R. Current Nanoscience 2006, 2, 243-255. [15] Choi, Y.; Baker, L. A.; Hillebrenner, H.; Martin, C. R. Physical Chemistry Chemical Physics 2006, 8, 4976-4988. [16] Siwy, Z.; Trofin, L.; Kohli, P.; Baker, L. A.; Trautmann, C.; Martin, C. R. Journal of the American Chemical Society 2005, 127, 5000-5001. [17] Harrell, C. C.; Choi, Y.; Horne, L. P.; Baker, L. A.; Siwy, Z. S.; Martin, C. R. Langmuir 2006, 22, 10837-10843. [18] Kohli, P.; Wirtz, M.; Martin, C. R. Electroanalysis 2004, 16, 9-18. 194

PAGE 195

[19] Harrell, C. C.; Kohli, P.; Siwy, Z.; Martin, C. R. Journal of the American Chemical Society 2004, 126, 15646-15647. [20] Steinle, E. D.; Mitchell, D. T.; Wirtz, M.; Lee, S. B.; Young, V. Y.; Martin, C. R. Analytical Chemistry 2002, 74, 2416-2422. [21] Lee, S. B.; Martin, C. R. Journal of the American Chemical Society 2002, 124, 11850-11851. [22] Wirtz, M.; Yu, S.; Martin, C. R. Analyst 2002, 127, 871-879. [23] Martin, C. R.; Nishizawa, M.; Jirage, K.; Kang, M. Journal of Physical Chemistry B 2001, 105, 1925-1934. [24] Lee, S. B.; Martin, C. R. Chemistry of Materials 2001, 13, 3236-3244. [25] Martin, C. R. Science 1994, 266, 1961-1966. [26] Martin, C. R.; Mitchell, D. T. Analytical Chemistry 1998, 70, 322A-327A. [27] Hulteen, J. C.; Martin, C. R. Journal of Materials Chemistry 1997, 7, 1075-1087. [28] Cepak, V. M.; Hulteen, J. C.; Che, G.; Jirage, K. B.; Lakshmi, B. B.; Fisher, E. R.; Martin, C. R. Journal of Materials Research 1998, 13, 3070-3080. [29] Miller, S. A.; Young, V. Y.; Martin, C. R. Journal of the American Chemical Society 2001, 123, 12335-12342. [30] Kang, M.; Trofin, L.; Mota, M. O.; Martin, C. R. Analytical Chemistry 2005, 77, 6243-6249. [31] Xu, F.; Wharton, J. E.; Martin, C. R. Small 2007, 3, 1718-1722. [32] Liang, W.; Martin, C. R. Journal of the American Chemical Society 1990, 112, 9666-9668. [33] Van Dyke, L. S.; Martin, C. R. Langmuir 1990, 6, 1118-1123. [34] Yamada, K.; Gasparac, R.; Martin, C. R. Journal of the Electrochemical Society 2004, 151, E14-E19. [35] Nishizawa, M.; Menon, V. P.; Martin, C. R. Science 1995, 268, 700-702. [36] Brumlik, C. J.; Menon, V. P.; Martin, C. R. Journal of Materials Research 1994, 9, 1174-1183. 195

PAGE 196

[37] Harrell, C. C.; Lee, S. B.; Martin, C. R. Analytical Chemistry 2003, 75, 6861-6867. [38] Mitchell, D. T.; Lee, S. B.; Trofin, L.; Li, N.; Nevanen, T. K.; Soederlund, H.; Martin, C. R. Journal of the American Chemical Society 2002, 124, 11864-11865. [39] Klein, J. D.; Herrick, R. D., II; Palmer, D.; Sailor, M. J.; Brumlik, C. J.; Martin, C. R. Chemistry of Materials 1993, 5, 902-904. [40] Hou, S.; Harrell, C. C.; Trofin, L.; Kohli, P.; Martin, C. R. Journal of the American Chemical Society 2004, 126, 5674-5675. [41] Hou, S.; Wang, J.; Martin, C. R. Journal of the American Chemical Society 2005, 127, 8586-8587. [42] Cai, Z.; Lei, J.; Liang, W.; Menon, V.; Martin, C. R. Chemistry of Materials 1991, 3, 960-967. [43] Parthasarathy, R.; Martin, C. R. Nature 1994, 369, 298-301. [44] Hornyak, G. L.; Patrissi, C. J.; Martin, C. R. Journal of Physical Chemistry B 1997, 101, 1548-1555. [45] Foss, C. A., Jr.; Hornyak, G. L.; Stockert, J. A.; Martin, C. R. Journal of Physical Chemistry 1994, 98, 2963-2971. [46] Scopece, P.; Baker, L. A.; Ugo, P.; Martin, C. R. Nanotechnology 2006, 17, 3951-3956. [47] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Science 1997, 278, 655-658. [48] Hulteen, J. C.; Jirage, K. B.; Martin, C. R. Journal of the American Chemical Society 1998, 120, 6603-6604. [49] Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Analytical Chemistry 1999, 71, 4913-4918. [50] Martin, C. R.; Kohli, P. Nature Reviews Drug Discovery 2003, 2, 29-37. [51] Gasparac, R.; Kohli, P.; Paulino, M. O. M.; Trofin, L.; Martin, C. R. Nano Letters 2004, 4, 513-516. [52] Hillebrenner, H.; Buyukserin, F.; Kang, M.; Mota, M. O.; Stewart, J. D.; Martin, C. R. Journal of the American Chemical Society 2006, 128, 4236-4237. [53] Hillebrenner, H.; Buyukserin, F.; Stewart, J. D.; Martin, C. R. Journal of Nanoscience and Nanotechnology 2007, 7, 2211-2221. [54] Buyukserin, F.; Kang, M.; Martin, C. R. Small 2007, 3, 106-110. 196

PAGE 197

[55] Ai, S.; Lu, G.; He, Q.; Li, J. Journal of the American Chemical Society 2003, 125, 11140-11141. [56] Kovtyukhova, N. I.; Mallouk, T. E.; Mayer, T. S. Advanced Materials 2003, 15, 780-785. [57] Li, F.; Zhang, L.; Metzger, R. M. Chemistry of Materials 1998, 10, 2470-2480. [58] Masuda, H.; Fukuda, K. Science 1995, 268, 1466-1468. [59] Wood, G. C.; O'Sullivan, J. P. Electrochimica Acta 1970, 15, 1865-1876. [60] Shingubara, S.; Morimoto, K.; Sakaue, H.; Takahagi, T. Electrochemical and Solid-State Letters 2003, 7, E15-E17. [61] Masuda, H.; Hasegwa, F.; Ono, S. Journal of the Electrochemical Society 1997, 144, L127-L130. [62] Masuda, H.; Yada, K.; Osaka, A. Japanese Journal of Applied Physics, Part 2: Letters 1998, 37, L 1340-L 1342. [63] Li, A. P.; Muller, F.; Birner, A.; Nielsch, K.; Gosele, U. Journal of Applied Physics 1998, 84, 6023-6026. [64] Che, G.; Lakshmi, B. B.; Fisher, E. R.; Martin, C. R. Nature 1998, 393, 346-349. [65] Hurst Sarah, J.; Payne Emma, K.; Qin, L.; Mirkin Chad, A. Angewandte Chemie, International Edition 2006, 45, 2672-2692. [66] Huber, C. A.; Huber, T. E.; Sadoqi, M.; Lubin, J. A.; Manalis, S.; Prater, C. B. Science 1994, 263, 800-802. [67] Kovtyukhova, N. I.; Mallouk, T. E. Advanced Materials 2005, 17, 187-192. [68] Park, S.; Lim, J.-H.; Chung, S.-W.; Mirkin, C. A. Science 2004, 303, 348-351. [69] Lee, W.; Scholz, R.; Nielsch, K.; Goesele, U. Angewandte Chemie, International Edition 2005, 44, 6050-6054. [70] Zhi, L.; Wu, J.; Li, J.; Kolb, U.; Muellen, K. Angewandte Chemie, International Edition 2005, 44, 2120-2123. [71] Wood, G. C.; O'Sullivan, J. P.; Vaszko, B. Journal of the Electrochemical Society 1968, 115, 618-620. [72] Masuda, H.; Satoh, M. Japanese Journal of Applied Physics, Part 2: Letters 1996, 35, L126-L129. 197

PAGE 198

[73] Masuda, H.; Yamada, H.; Satoh, M.; Asoh, H.; Nakao, M.; Tamamura, T. Applied Physics Letters 1997, 71, 2770-2772. [74] Asoh, H.; Nishio, K.; Nakao, M.; Tamamura, T.; Masuda, H. Journal of the Electrochemical Society 2001, 148, B152-B156. [75] Diggle, J. W.; Downie, T. C.; Goulding, C. W. Chemical Reviews 1969, 69, 365-405. [76] Jessensky, O.; Muller, F.; Gosele, U. Applied Physics Letters 1998, 72, 1173-1175. [77] O'Sullivan, J. P.; Wood, G. C. Proceedings of the Royal Society of London, Series A: Mathematical, Physical and Engineering Sciences 1970, 317, 511-543. [78] Parkhutik, V. P.; Shershul'skii, V. I. Journal of Physics D: Applied Physics 1992, 25, 1258-1263. [79] Siejka, J.; Ortega, C. Journal of the Electrochemical Society 1977, 124, 883-891. [80] Hench, L. L.; West, J. K. Chemical Reviews 1990, 90, 33-72. [81] Livage, J.; Henry, M.; Sanchez, C. Progress in Solid State Chemistry 1988, 18, 259-341. [82] Patrissi, C. J.; Martin, C. R. Journal of the Electrochemical Society 1999, 146, 3176-3180. [83] Sides, C. R.; Martin, C. R. Modern Aspects of Electrochemistry 2007, 40, 75-126. [84] Sides, C. R.; Li, N.; Patrissi, C. J.; Scrosati, B.; Martin, C. R. MRS Bulletin 2002, 27, 604-607. [85] Lakshmi, B. B.; Patrissi, C. J.; Martin, C. R. Chemistry of Materials 1997, 9, 2544-2550. [86] Lin, J.; Brown, C. W. TrAC, Trends in Analytical Chemistry 1997, 16, 200-211. [87] Collinson, M. M. Mikrochimica Acta 1998, 129, 149-165. [88] Avnir, D. Accounts of Chemical Research 1995, 28, 328-334. [89] Walcarius, A. Electroanalysis 1998, 10, 1217-1235. [90] Brinker, C. J. Current Opinion in Colloid & Interface Science 1998, 3, 166-173. [91] Walcarius, A. Electroanalysis 2001, 13, 701-718. [92] Stoeber, W.; Fink, A.; Bohn, E. Journal of Colloid and Interface Science 1968, 26, 62-69. 198

PAGE 199

[93] Hench, L. L. Sol-Gel Silica-Properties, Processing and Technology Transfer; Noyes Publications: New Jersey, USA, 1998. [94] Iler, R. K. The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface Properties and Biochemistry, John Wiley & Sons, 1979. [95] Pyell, U.; Stork, G. Fresenius' Journal of Analytical Chemistry 1992, 343, 576-581. [96] Jal, P. K.; Patel, S.; Mishra, B. K. Talanta 2004, 62, 1005-1028. [97] Kiselev, A. V.; Lygin, V. I. Infrared Spectra of Surface Compounds; Wiley: New York, 1975. [98] Kiselev, A. V. Kolloidnyi Zhurnal 1936, 2, 17-26. [99] Chuang, I. S.; Maciel, G. E. Journal of Physical Chemistry B 1997, 101, 3052-3064. [100] Grushka, E.; Kikta, E. J., Jr. Analytical Chemistry 1977, 49, 1004A-1008A, 1010A, 1012A, 1014A. [101] Van der Voort, P.; Gillis-D'Hamers, I.; Vansant, E. F. Journal of the Chemical Society, Faraday Transactions 1990, 86, 3751-3755. [102] Tundo, P.; Venturello, P.; Angeletti, E. Journal of the American Chemical Society 1982, 104, 6551-6555. [103] Plueddemann, E. P. Silane Coupling Agents, 2nd ed.; Plenum Publishing Corporation: New York, 1991. [104] Loy, D. A.; Baugher, B. M.; Baugher, C. R.; Schneider, D. A.; Rahimian, K. Chemistry of Materials 2000, 12, 3624-3632. [105] Arkles, B. Chemtech 1977, 7, 766-778. [106] Dias Filho, N. L.; Gushikem, Y. Separation Science and Technology 1997, 32, 2535-2545. [107] Lehninger, A. L. Principles of Biochemistry, 2nd ed.; Worth Inc.: New York, 1993. [108] Kemp, N. J.; Noble, R. D. Separation Science and Technology 1983, 18, 1147-1165. [109] Lakshmi, B. B.; Martin, C. R. Nature 1997, 388, 758-760. [110] Noble, R. D.; Koval, C. A. Materials Science of Membranes for Gas and Vapor Separation 2006, 411-435. 199

PAGE 200

[111] Wirgau, J. I.; Crumbliss, A. L. Dalton Transactions 2003, 3680-3685. [112] Zaghbani, A.; Tayeb, R.; Bonnamour, I.; Felix, C.; Vocanson, F.; Lamartine, R.; Dhahbi, M.; Seta, P. Journal of Membrane Science 2005, 258, 5-7. [113] Villamo, O.; Barboiu, C.; Barboiu, M.; Yau-Chun-Wan, W.; Hovnanian, N. Journal of Membrane Science 2002, 204, 97-110. [114] Shentu, B.; Nishide, H. Industrial & Engineering Chemistry Research 2003, 42, 5954-5958. [115] Tsuchida, E.; Nishide, H.; Ohyanagi, M. Journal of Macromolecular Science, Chemistry 1988, A25, 1327-1338. [116] Barboiu, M.; Guizard, C.; Hovnanian, N.; Palmeri, J.; Reibel, C.; Cot, L.; Luca, C. Journal of Membrane Science 2000, 172, 91-103. [117] Gherrou, A.; Kerdjoudj, H.; Molinari, R.; Drioli, E. Separation Science and Technology 2002, 37, 2317-2336. [118] Choi, H. W.; Kim, D. B.; Choi, D. K.; Ahn, B. S.; Kim, H. G.; Kim, H. S.; Lee, C. H.; Sung, J. Y. Journal of Membrane Science 2006, 279, 403-409. [119] Kaucher, M. S.; Peterca, M.; Dulcey, A. E.; Kim, A. J.; Vinogradov, S. A.; Hammer, D. A.; Heiney, P. A.; Percec, V. Journal of the American Chemical Society 2007, 129, 11698-11699. [120] Janeway, C. A., Travers, P.,Walport, M., Capra, J. D. Immunobiology: The Immune System in Health and Disease, 5th ed.; Garland Publishing Inc.: New York, 2001. [121] Burton, D. R. Trends in Biochemical Sciences 1990, 15, 64-69. [122] Gellissen, G. Production of Recombinant Proteins : Novel Microbial and Eukaryotic Expression Systems; Wiley-VCH: Weinheim, Germany, 2004. [123] Merten, O.-W., Mattanovich, D., Lang, C., Larsson, G., Neubauer, P., Porro, D., Postma, P., Teixeira de Mattos, J., Cole, J.A. Recombinant Protein Production With Prokaryotic and Eukaryotic Cells : A Comparative View on Host Physiology; Kluwer Academic Pub.: Semmering, Austria, 2001. [124] Hochuli, E.; Doebeli, H.; Schacher, A. Journal of Chromatography 1987, 411, 177-184. [125] Smith, M. C.; Furman, T. C.; Ingolia, T. D.; Pidgeon, C. Journal of Biological Chemistry 1988, 263, 7211-7215. [126] Porath, J.; Carlsson, J.; Olsson, I.; Belfrage, G. Nature 1975, 258, 598-599. 200

PAGE 201

[127] Porath, J. Trends in Analytical Chemistry 1988, 7, 254-259. [128] Eftink, M. R. Biophysical Journal 1994, 66, 482-501. [129] Eftink, M. R. Methods in Enzymology 2000, 323, 459-473. [130] Methods in Enzymology: Fluorescence Spectroscopy, 1997, 278, 1-628. [131] Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Plenum Publishers, Inc.: New York, 1999. [132] Guilbault, G. G. Practical Fluorescence. Modern Monographs in Analytical Chemistry, 2nd ed.; Marcel Dekker, Inc.: New York, 1990. [133] Mirabella, F. M. Principles, Theory, and Practice of Internal Reflection Spectroscopy, in "Internal Reflection Spectroscopy Theory and Applications"; Marcel Dekker, Inc.: New York, 1993. [134] Skourlis, T. P.; McCullough, R. L. Journal of Applied Polymer Science 1994, 52, 1241-1248. [135] Harrick, N. J.; duPre, F. K. Applied Optics 1966, 5, 1739-1742. [136] Tan, S., Belanger, D. Journal of Physical Chemistry B 2005, 109, 23480-23490. [137] Watts, J. F., Wolstenholme, J. An Introduction to Surface Analysis by XPS and AES; John Wiley & Sons Inc., 2003. [138] Goldstein, J., Newbury, D. E., Joy, D. C., Lyman, C. E., Echlin, P., Lifshin, E., Sawyer, L., Michael, J. Scanning Electron Microscopy and X-ray Microanalysis, 3rd ed.; Springer Science + Business Media, Inc.: New York, 2003. [139] Chilcott, T. C.; Chan, M.; Gaedt, L.; Nantawisarakul, T.; Fane, A. G.; Coster, H. G. L. Journal of Membrane Science 2002, 195, 153-167. [140] Gaedt, L.; Chilcott, T. C.; Chan, M.; Nantawisarakul, T.; Fane, A. G.; Coster, H. G. L. Journal of Membrane Science 2002, 195, 169-180. [141] Park, S.-M.; Yoo, J.-S. Analytical Chemistry 2003, 75, 455A-461A. [142] Kharitonov, A. B.; Alfonta, L.; Katz, E.; Willner, I. Journal of Electroanalytical Chemistry 2000, 487, 133-141. [143] Bilitewski, U. Analytica Chimica Acta 2006, 568, 232-247. 201

PAGE 202

[144] Ye, M.; Jiang, X.; Feng, S.; Tian, R.; Zou, H. TrAC, Trends in Analytical Chemistry 2007, 26, 80-84. [145] Domon, B.; Aebersold, R. Science 2006, 312, 212-217. [146] Domon, B.; Broder, S. Journal of Proteome Research 2004, 3, 253-260. [147] Marengo, E.; Leardi, R.; Robotti, E.; Righetti, P. G.; Antonucci, F.; Cecconi, D. Journal of Proteome Research 2003, 2, 351-360. [148] Barroso, B.; Lubda, D.; Bischoff, R. Journal of Proteome Research 2003, 2, 633-642. [149] Masselon, C.; Pasa-Tolic, L.; Tolic, N.; Anderson, G. A.; Bogdanov, B.; Vilkov, A. N.; Shen, Y.; Zhao, R.; Qian, W.-J.; Lipton, M. S.; Camp, D. G., II; Smith, R. D. Analytical Chemistry 2005, 77, 400-406. [150] Nedelkov, D.; Tubbs, K. A.; Niederkofler, E. E.; Kiernan, U. A.; Nelson, R. W. Analytical Chemistry 2004, 76, 1733-1737. [151] Zilberstein, G.; Korol, L.; Antonioli, P.; Righetti, P. G.; Bukshpan, S. Analytical Chemistry 2007, 79, 821-827. [152] Iritani, E.; Mukai, Y.; Murase, T. Filtration & Separation 1997, 34, 967-973. [153] Ghosh, R.; Cui, Z. F. Journal of Membrane Science 2000, 167, 47-53. [154] Harkensee, D.; Kokpinar, O.; Walter, J.; Kasper, C.; Beutel, S.; Reif, O. W.; Scheper, T.; Ulber, R. Engineering in Life Sciences 2007, 7, 388-394. [155] Guo, W.; Ruckenstein, E. Journal of Membrane Science 2003, 211, 101-111. [156] Ramamoorthy, M.; Raju, M. D. Industrial & Engineering Chemistry Research 2001, 40, 4815-4820. [157] Suen, S.-Y.; Liu, Y.-C.; Chang, C.-S. Journal of Chromatography, B 2003, 797, 305-319. [158] O'Connor, A. J.; Pratt, H. R. C.; Stevens, G. W. Chemical Engineering Science 1996, 51, 3459-3477. [159] Ho, A. K.; Perera, J. M.; Dunstan, D. E.; Stevens, G. W.; Nystrom, M. AIChE Journal 1999, 45, 1434-1450. [160] Li, Q. Y.; Pepper, D. S. Journal of Membrane Science 1997, 136, 181-190. [161] Nystrom, M.; Aimar, P.; Luque, S.; Kulovaara, M.; Metsamuuronen, S. Colloids and Surfaces, A: Physicochemical and Engineering Aspects 1998, 138, 185-205. 202

PAGE 203

[162] Nakatsuka, S.; Michaels, A. S. Journal of Membrane Science 1992, 69, 189-211. [163] Saxena, A.; Shahi, V. K. Journal of Membrane Science 2007, 299, 211-221. [164] Aravind, U. K.; Mathew, J.; Aravindakumar, C. T. Journal of Membrane Science 2007, 299, 146-155. [165] Poetz, O.; Schwenk, J. M.; Kramer, S.; Stoll, D.; Templin, M. F.; Joos, T. O. Mechanisms of Ageing and Development 2005, 126, 161-170. [166] Poetz, O.; Ostendorp, R.; Brocks, B.; Schwenk, J. M.; Stoll, D.; Joos, T. O.; Templin, M. F. Proteomics 2005, 5, 2402-2411. [167] Langmuir, I. Journal of the American Chemical Society 1918, 40, 1361-1402. [168] Yang, L.; Chen, P. Journal of Membrane Science 2002, 205, 141-153. [169] Weiss, M.; Welsch, W.; Schickfus, M. v.; Hunklinger, S. Analytical Chemistry 1998, 70, 2881-2887. [170] Saha, K.; Bender, F.; Gizeli, E. Analytical Chemistry 2003, 75, 835-842. [171] Igarashi, T.; Ishigatsubo, Y.; Ohno, S.; Shirai, A.; Aoki, I.; Okubo, T.; Klinman, D. M. Clinical Immunology and Immunopathology 1994, 72, 307-311. [172] Klinman, D. M. ; Banks, S.; Hartman, A.; Steinberg, A. D. Journal of Clinical Investigations 1988, 82, 652-657. [173] Gueell, C.; Davis, R. H. Journal of Membrane Science 1996, 119, 269-284. [174] Mueller, J.; Davis, R. H. Journal of Membrane Science 1996, 116, 47-60. [175] Ho, C.-C.; Zydney, A. L. Journal of Membrane Science 1999, 155, 261-275. [176] Jones, K. L.; O'Melia, C. R. Journal of Membrane Science 2000, 165, 31-46. [177] Dechadilok, P.; Deen, W. M. Industrial & Engineering Chemistry Research 2006, 45, 6953-6959. [178] Bhattacharjee, S.; Sharma, A.; Bhattacharya, P. K. Industrial & Engineering Chemistry Research 1996, 35, 3108-3121. [179] Johnson, C. P.; Jensen, I. E.; Prakasam, A.; Vijayendran, R.; Leckband, D. Bioconjugate Chemistry 2003, 14, 974-978. 203

PAGE 204

[180] Dancette, O. P.; Taboureau, J.-L.; Tournier, E.; Charcosset, C.; Blond, P. Journal of Chromatography, B: Biomedical Sciences and Applications 1999, 723, 61-68. [181] Rutherford, S. W.; Do, D. D. Adsorption 1997, 3, 283-312. [182] Paul, D. R. Journal of Polymer Science, Polymer Physics Edition 1969, 7, 1811-1818. [183] Tsuru, T.; Izumi, S.; Yoshioka, T.; Asaeda, M. AIChE Journal 2000, 46, 565-574. [184] Nomura, T.; Nakao, S.; Kimura, S. Kagaku Kogaku Ronbunshu 1987, 13, 811-817. [185] Chandler, D. Nature 2005, 437, 640-647. [186] Lin, F.-Y.; Chen, W.-Y.; Hearn, M. T. W. Journal of Molecular Recognition 2002, 15, 55-93. [187] Palma, E.; Klapper, D. G.; Cho, M. J. Pharmaceutical Research 2005, 22, 122-127. [188] Norde, W.; Lyklema, J. Journal of Colloid and Interface Science 1979, 71, 350-366. [189] Norde, W. Advances in colloid and interface science 1986, 25, 267-340. [190] Norde, W.; MacRitchie, F.; Nowicka, G.; Lyklema, J. Journal of Colloid and Interface Science 1986, 112, 447-456. [191] Hamilton, R. G. Production and epitope location of monoclonal antibodies to the human IgG subclasses. In Molecular Aspects of Immunoglobulin Subclasses; Pergamon Press: Oxford, 1990, pp 79-91. [192] Schmuhl, R.; Keizer, K.; van den Berg, A.; ten Elshof, J. E.; Blank, D. H. A. Journal of Colloid and Interface Science 2004, 273, 331-338. [193] Saksena, S.; Zydney, A. L. Biotechnology and Bioengineering 1994, 43, 960-968. [194] Zydney, A. L.; Pujar, N. S. Colloids and Surfaces, A: Physicochemical and Engineering Aspects 1998, 138, 133-143. [195] Lopez, M. F.; Melov, S. Circulation Research 2002, 90, 380-389. [196] Sun, L.; Crooks, R. M. Langmuir 1999, 15, 738-741. [197] Bath, B. D.; Lee, R. D.; White, H. S.; Scott, E. R. Analytical Chemistry 1998, 70, 1047-1058. [198] Baltus, R. E. Journal of Membrane Science 1997, 130, 157-172. 204

PAGE 205

[199] Galiatsatou, P.; Kanellopoulos, N. K.; Petropoulos, J. H. Journal of Membrane Science 2006, 280, 634-642. [200] Funke, H. H.; Noble, R. D.; Koval, C. A. Journal of Membrane Science 1993, 82, 229-236. [201] Smith, K. A.; Meldon, J. H.; Colton, C. K. AIChE Journal 1973, 19, 102-109. [202] Teramoto, M. Industrial & Engineering Chemistry Research 1994, 33, 2161-2171. [203] Vieth, W. R.; Sladek, K. J. Journal of Colloid Science 1965, 20, 1014-1033. [204] Paul, D. R.; Koros, W. J. Journal of Polymer Science, Polymer Physics Edition 1976, 14, 675-685. [205] Vieth, W. R.; Howell, J. M.; Hsieh, J. H. Journal of Membrane Science 1976, 1, 177-220. [206] Petropoulos, J. H. Journal of Polymer Science, Polymer Physics Edition 1970, 8, 1797-1801. [207] Mauze, G. R.; Stern, S. A. Journal of Membrane Science 1984, 18, 99-109. [208] Riedl, C. M.; Spencer, H. G. Journal of Applied Polymer Science 1990, 41, 1685-1693. [209] Noble, R. D. Journal of Membrane Science 1990, 50, 207-214. [210] Kanehashi, S.; Nagai, K. Journal of Membrane Science 2005, 253, 117-138. [211] Tsujita, Y. Progress in Polymer Science 2003, 28, 1377-1401. [212] Barrer, R. M.; Barrie, J. A.; Slater, J. Journal of Polymer Science 1958, 27, 177-197. [213] Stern, S. A.; Zhou, S.; Araux-Lara, J. L.; Ware, B. R. Journal of Polymer Science, Part C: Polymer Letters 1989, 27, 427-431. [214] Koros, W. J.; Paul, D. R. Journal of Polymer Science, Polymer Physics Edition 1978, 16, 1947-1963. [215] Huh, J.-K.; Song, D.-I.; Jeon, Y.-W. Separation Science and Technology 1999, 34, 571-586. [216] Huh, J.-K.; Song, D.-I.; Jeon, Y.-W. Separation Science and Technology 2000, 35, 243-259. [217] Wang, J.-S.; Kamiya, Y. Journal of Membrane Science 1999, 154, 25-32. 205

PAGE 206

[218] Dixon-Garrett, S. V.; Nagai, K.; Freeman, B. D. Journal of Polymer Science, Part B: Polymer Physics 2000, 38, 1078-1089. [219] Dixon-Garrett, S. V.; Nagai, K.; Freeman, B. D. Journal of Polymer Science, Part B: Polymer Physics 2000, 38, 1461-1473. [220] Meares, P. Journal of the American Chemical Society 1954, 76, 3415-3422. [221] Budd, P. M.; McKeown, N. B.; Fritsch, D. Journal of Materials Chemistry 2005, 15, 1977-1986. [222] Huang, S. H.; Margolies, M. N.; Rothstein, A. M.; Siekevitz, M.; Gefter, M. L. Pharmacology Reviews 1982, 34, 43. [223] Sem, D. S.; McNeeley, P. A.; Linnik, M. D. Archives of Biochemistry and Biophysics 1999, 372, 62-68. [224] Petropoulos, J. H. Advances in Polymer Science 1985, 64, 93-142. [225] Kempe, H.; Persson, P.; Axelsson, A.; Nilsson,B.; Zacchi, G. Biotechnology Bioengineer 2006, 93, 656-664. [226] Bosma, J. C.; Wesselingh, J. A. Journal of Chromatography B 2000, 743, 169. [227] Burczak, K.; Fujisato, T.; Hatada, M.; Ikada, Y. Biomaterials 1994, 15, 231-238. [228] La Flamme, K. E.; Mor, G.; Gong, D.; La Tempa, T.; Fusaro, V. A.; Grimes, C. A.; Desai, T. A. Diabetes Technology & Therapeutics 2005, 7, 684-694. [229] Leoni, L.; Boiarski, A.; Desai, T. A. Biomedice Microdevices 2002, 4, 131-139. [230] Bird, R. B.; Steward, W. E.; Lightfoot, E. N. Transport Phenomena, John Wiley, New York, NY, 1960. [231] Cussler, E. L. Diffusion. Mass Transfer in Fluid systems. Cambridge University Press, Cambridge, UK, 1997. [232] Edward, J. T. Journal of Chemical Education 1970, 47, 261-270. [233] Pluen, A.; Netti, P. A.; Jain, R. K.; Berk, D. A. Biophysical Journal 1999, 77, 542-552. [234] Terazima, M. Physical Chemistry Chemical Physics 2006, 8, 545-557. [235] Nishida, S.; Nada, T.; Terazima, M. Biophysical Journal 2004, 87, 2663-2675. [236] Moschou, E. A.; Bachas, L. G.; Deo, S. K. Analytical Chemistry 2006, 78, 6692-6700. 206

PAGE 207

[237] Hentz, N. G.; Daunert, S. Analytical chemistry 1996, 68, 3939-3944. [238] Lager, I.; Looger, L. L.; Hilpert, M.; Lalonde, S.; Frommer, W. B. Journal of Biological Chemistry 2006, 281, 30875-30883. [239] Dwyer, M. A.; Hellinga, H. W. Current Opinion in Structural Biology 2004, 14, 495-504. [240] Lager, I.; Fehr, M.; Frommer, W. B.; Lalonde, S. FEBS Letters 2003, 553, 85-89. [241] Nilsson, B.; Forsberg, G.; Moks, T.; Hartmanis, M.; Uhlen, M. Current Opinion in Structural Biology 1992, 2, 569-575. [242] Seidler, A. Protein Engineering 1994, 7, 1277-1280. [243] Zhang, Z.; Tong, K. T.; Belew, M.; Pettersson, T.; Janson, J. C. Journal of chromatography 1992, 604, 143-155. [244] Carlsson, J.; Porath, J.; Lonnerdal, B. FEBS letters 1977, 75, 89-92. [245] Chadha, K. C.; Grob, P. M.; Mikulski, A. J.; Davis, L. R.; Sulkowski, E. Journal of General Virology 1979, 43, 701-706. [246] Chaga, G.; Andersson, L.; Ersson, B.; Porath, J. Biotechnology and applied biochemistry 1989, 11, 424-431. [247] McCluskey, A. J.; Poon, G. M. K.; Gariepy, J. Protein Science 2007, 16, 2726-2732. [248] Lee, K.-M.; Ma, K.-W.; Shaw, P.-C.; Wong, K.-B. Analytical Biochemistry 2006, 358, 152-154. [249] Cantin, G. T.; Yi, W.; Lu, B.; Park, S. K.; Xu, T.; Lee, J.-D.; Yates, J. R. Journal of Proteome Research 2008, 7, 1346-1351. [250] Kermani, B. G.; Kozlov, I.; Melnyk, P.; Zhao, C.; Hachmann, J.; Barker, D.; Lebl, M. Sensors and Actuators, B: Chemical 2007, B125, 149-157. [251] Ueda, E. K. M.; Gout, P. W.; Morganti, L. Journal of Chromatography, A 2003, 988, 1-23. [252] Pearson, R. G. Journal of Chemical Education 1968, 45, 581-587. [253] Arnold, F. H. Biotechnology 1991, 9, 151-156. [254] Sulkowski, E. BioEssays 1989, 10, 170-175. [255] Min, C.; Verdine, G. L. Nucleic Acids Research 1996, 24, 3806-3810. 207

PAGE 208

[256] Hochuli, E.; Doebeli, H.; Schacher, A. Journal of Chromatography 1987, 411, 177-184. [257] Gaberc-Porekar, V.; Menart, V. Chemical Engineering & Technology 2005, 28, 1306-1314. [258] Sundberg, R. J.; Martin, R. B. Chemical Reviews 1974, 74, 471-517. [259] Gasparac, R.; Martin, C. R.; Stupnisek-Lisac, E.; Mandic, Z. Journal of the Electrochemical Society 2000, 147, 991-998. [260] Meier, M. A. R.; Schubert, U. S. Journal of Polymer Science, Part A: Polymer Chemistry 2003, 41, 2964-2973. [261] Herrero, M.; Tiemblo, P.; Reyes-Labarta, J.; Mijangos, C.; Reinecke, H. Polymer 2002, 43, 2631-2636. [262] Tiemblo, P.; Guzman, J.; Riande, E.; Mijangos, C.; Reinecke, H. Macromolecules 2002, 35, 420-424. [263] Dias Filho, N. L. Mikrochimica Acta 1999, 130, 233-240. [264] Kemlo, J. A.; Shepherd, T. M. Chemical Physics Letters 1977, 47, 158-162. [265] Varnes, A. W.; Dodson, R. B.; Wehry, E. L. Journal of the American Chemical Society 1972, 94, 946-950. [266] de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chemical Reviews 1997, 97, 1515-1566. [267] Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Sacchi, D.; Taglietti, A. Analyst 1996, 121, 1763-1768. [268] Huang, F.; Tian, Y.; Chen, C.-Y.; Cheng, Y.-J.; Young, A. C.; Jen, A. K. Y. Journal of Physical Chemistry C 2007, 111, 10673-10681. [269] McFarland, S. A.; Finney, N. S. Journal of the American Chemical Society 2001, 123, 1260-1261. [270] You, C.-C.; Hippius, C.; Gruene, M.; Wuerthner, F. Chemistry--A European Journal 2006, 12, 7510-7519. [271] Shen, Y.; Mackey, G.; Rupcich, N.; Gloster, D.; Chiuman, W.; Li, Y.; Brennan, J. D. Analytical Chemistry 2007, 79, 3494-3503. [272] Hartwig, J. F.; Pil, P. M.; Lippard, S. J. Journal of the American Chemical Society 1992, 114, 8292-8293. 208

PAGE 209

[273] Thornton, N. B.; Schanze, K. S. New Journal of Chemistry 1996, 20, 791-800. [274] McFarland, S. A.; Finney, N. S. Journal of the American Chemical Society 2002, 124, 1178-1179. [275] Cotton, F. A., Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; John Wiley & Sons: NY, 1987. [276] Housecroft, C. E., Sharpe, A. G. Inorganic Chemistry; Prentice Hall: NY, 2001. [277] Silva, H. A. S.; Carlos, R. M.; Camargo, A. J.; Picchi, C. M. C.; Santos, R. H. d. A.; McGarvey, B. R.; Franco, D. W. Inorganica Chimica Acta 2004, 357, 3147-3154. [278] O'Sullivan, J. P.; Hockey, J. A.; Wood, G. C. Transactions of the Faraday Society 1969, 65, 535-541. [279] Kimura, Y.; Shiraki, H.; Ishibashi, K.-i.; Ishii, H.; Itaya, K.; Niwano, M. Journal of the Electrochemical Society 2006, 153, C296-C300. [280] Abbott, T. P.; Nabetani, H.; Sessa, D. J.; Wolf, W. J.; Liebman, M. N.; Dukor, R. K. Journal of Agricultural and Food Chemistry 1996, 44, 2220-2224. [281] Malhotra, V. M.; Jasty, S.; Mu, R. Applied Spectroscopy 1989, 43, 638-645. [282] Spinner, E. Journal of the Chemical Society 1960, 1237-1242. [283] Spinner, E. Spectrochimica Acta 1959, 95-109. [284] Spinner, E. Journal of the Chemical Society 1962, 3127-3128. [285] Katritzky, A. R.; Witek, R. M.; Rodriguez-Garcia, V.; Mohapatra, P. P.; Rogers, J. W.; Cusido, J.; Abdel-Fattah, A. A. A.; Steel, P. J. Journal of Organic Chemistry 2005, 70, 7866-7881. [286] Hayashi, T.; Mukamel, S. Journal of Physical Chemistry B 2007, 111, 11032-11046. [287] Waldron, R. D. Physical Review 1955, 99, 1727-1735. [288] Feenan, K.; Fowles, G. W. A. Journal of the Chemical Society 1965, 2449-2451. [289] Kyotani, T.; Sato, K.; Mizuno, T.; Kakui, S.; Aizawa, M.; Saito, J.; Ikeda, S.; Ichikawa, S.; Nakane, T. Analytical Sciences 2005, 21, 321-325. [290] Do, J.; Ahn, C. H. Lab on a Chip 2008, 8, 542-549. 209

PAGE 210

[291] VanWiggeren, G. D.; Bynum, M. A.; Ertel, J. P.; Jefferson, S.; Robotti, K. M.; Thrush, E. P.; Baney, D. M.; Killeen, K. P. Sensors and Actuators, B: Chemical 2007, B127, 341-349. [292] Medyantseva, E. P.; Varlamova, R. M.; Gimaletdinova, D. A.; Fattakhova, A. N.; Budnikov, G. K. Journal of Analytical Chemistry 2008, 63, 275-279. [293] Borisov, S. M.; Wolfbeis, O. S. Chemical Reviews 2008, 108, 423-461. [294] Goel, G.; Kumar, A.; Puniya, A. K.; Chen, W.; Singh, K. Journal of Applied Microbiology 2005, 99, 435-442. [295] Zeineldin, R.; Piyasena, M. E.; Sklar, L. A.; Whitten, D.; Lopez, G. P. Langmuir 2008, 24, 4125-4131. [296] Ko, S.; Grant, S. A. Biosensors & Bioelectronics 2006, 21, 1283-1290. [297] Liu, X.; Tan, W. Analytical Chemistry 1999, 71, 5054-5059. [298] Wu, P.; Brand, L. Analytical Biochemistry 1994, 218, 1-13. [299] Kagan, C. R.; Murray, C. B.; Nirmal, M.; Bawendi, M. G. Physical Review Letters 1996, 76, 3043. [300] Cheung, H. C.; Dong, W.-J. Reviews in Fluorescence 2006, 3, 445-462. [301] Schuler, B. Chemphyschem 2005, 6, 1206-1220. [302] Invitrogen. Amine-Reactive Probes, MP00143, Molecular Probes: Eugene, OR, 2007. http://probes.invitrogen.com/media/pis/mp00143.pdf [303] http://probes.invitrogen.com/handbook/boxes/0422.html [304] Piunno, P. A. E.; Krull, U. J.; Hudson, R. H. E.; Damha, M. J.; Cohen, H. Analytical Chemistry 1995, 67, 2635-2643. [305] Wang, J.; Xu, D.; Kawde, A. N.; Polsky, R. Analytical chemistry 2001, 73, 5576-5581. [306] Taton, T. A.; Mirkin, C. A.; Letsinger, R. L. Science 2000, 289, 1757-1760. [307] Drummond, T. G.; Hill, M. G.; Barton, J. K. Nature Biotechnology 2003, 21, 1192-1199. [308] Su, X.; Robelek, R.; Wu, Y.; Wang, G.; Knoll, W. Analytical Chemistry 2004, 76, 489-494. 210

PAGE 211

[309] Maruyama, K.; Mishima, Y.; Minagawa, K.; Motonaka, J. Analytical Chemistry 2002, 74, 3698-3703. [310] Zhao, X.; Tapec-Dytioco, R.; Tan, W. Journal of the American Chemical Society 2003, 125, 11474-11475. [311] Landers, J. P. Analytical Chemistry 2003, 75, 2919-2927. [312] Miao, W.; Bard, A. J. Analytical Chemistry 2004, 76, 5379-5386. [313] Castro, A.; Dalvit, D. A. R.; Paz-Matos, L. Analytical Chemistry 2004, 76, 4169-4174. [314] Wang, R.; Minunni, M.; Tombelli, S.; Mascini, M. Biosensors & Bioelectronics 2004, 20, 598-605. [315] Bui, M.-P. N.; Baek, T. J.; Seong, G. H. Analytical and Bioanalytical Chemistry 2007, 388, 1185-1190. [316] Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 737-738. [317] Marmur, J.; Doty, P. Journal of molecular biology 1961, 3, 585-594. [318] Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science 1991, 254, 1497-1500. [319] Nielsen, P. E. Current opinion in biotechnology 1999, 10, 71-75. [320] Wang, J. Biosensors & Bioelectronics 1998, 13, 757-762. [321] Arlinghaus, H. F.; Kwoka, M. N.; Jacobson, K. B. Analytical Chemistry 1997, 69, 3747-3753. [322] Wang, J.; Palecek, E.; Nielsen, P. E.; Rivas, G.; Cai, X.; Shiraishi, H.; Dontha, N.; Luo, D.; Farias, P. A. M. Journal of the American Chemical Society 1996, 118, 7667-7670. [323] Aoki, H.; Umezawa, Y. Analyst 2003, 128, 681-685. [324] Abel, A. P.; Weller, M. G.; Duveneck, G. L.; Ehrat, M.; Widmer, H. M. Analytical Chemistry 1996, 68, 2905-2912. [325] Methods in Enzymology: Aviding-Biotin Technology, 1990, 184, 1-617. [326] Macdonald, J. R. Impedance Spectroscopy; Emphasizing Solid Materials and Systems; Wiley Interscience publications: NY, 1987. [327] Winkler, B. H.; Baltus, R. E. Journal of Membrane Science 2003, 226, 75-84. 211

PAGE 212

BIOGRAPHICAL SKETCH Hctor Mario Caicedo was born in Cali, Colombia. He graduated with a B.S. in chemistry from the Universidad del Valle, in Cali. For his bachelors thesis he conducted research in analytical chemistry focusing on environmental pollutants under the direction of Dr. Nelly de Palacios (his mentor in chemistry), developing an analytical methodology for the evaluation of chromium in urine and blood by using electrothermal atomic absorption spectroscopy (ETAAS). After that, he worked as a chemist in the industrial field, for a combined time of almost four years, for both a biotechnology and pharmaceutical company. Then, he decided to go back to academia to conduct his Master of Science in analytical chemistry at the Universidad del Valle. In the summer of 2000, he visited the University of Florida as an international exchange student, conducting research in electrochemistry, as a part of his masters research work, using carbon fiber ultramicroelectrodes under the direction of Dr. Anna Brajter-Toth. After returning to the Universidad del Valle, he conducted research in electroanalysis of xanthine by using carbon fiber ultramicroelectrodes with surfaces electrochemically activated and polymer-modified, in the group of Dr. Alonso Jaramillo. Once he graduated with a M.S. in chemistry from the University del Valle, he continued his graduate studies in the Department of Chemistry at the University of Florida to pursue his Ph.D. He worked under the supervision of Dr. Charles R. Martin, where he conducted research in bioanalytical applications of affinity-based nanotube membranes for sensing and separations. He completed his final work in April 2008, when he received a Doctor of Philosophy degree. 212


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20101106_AAAAEG INGEST_TIME 2010-11-06T13:56:21Z PACKAGE UFE0022016_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 30996 DFID F20101106_AAASGA ORIGIN DEPOSITOR PATH caicedo_h_Page_159.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
3c82025b8206f3dc56cebe7ed0279ec5
SHA-1
7ac5b9f2f409bcb67888591d50d58470d0f7de17
5942 F20101106_AAASFM caicedo_h_Page_150thm.jpg
7a7635c3560a96a307a448fe1f50a2a6
2ec153f7db85fede9f8d99c6a15894b33d854871
8981 F20101106_AAASEY caicedo_h_Page_140thm.jpg
b7df140a43a091a72de4bf0861831f68
049f68d7a955293cd73d67299c0e997262fbf4a6
1051922 F20101106_AAARCK caicedo_h_Page_085.jp2
2f94a41ff84fa7644aa6e444131a258c
da6d2246d3fca6914dee1a1432a07a4d11541da2
92251 F20101106_AAAQXE caicedo_h_Page_106.jpg
1703fc77aa49f5dd3dbee5944f8831a3
b72791482a2426f37445be8b75ba9f2074437ff7
1031673 F20101106_AAARBW caicedo_h_Page_064.jp2
6e062d827b4b9e631a6dd0127b63d148
58322b3512ddaf4d4be2eec90f4ef8c04d5ad5ec
35230 F20101106_AAASGB caicedo_h_Page_161.QC.jpg
145c889e35c73e7a10fcc63d85b6b59d
41edadc9dfe506e10d061744df351fc4924f4faa
34331 F20101106_AAASFN caicedo_h_Page_151.QC.jpg
35e79f515ff1f12e7ffb3d81cd94d6ee
7ca30908c7e92e7136b52feb962716752d93bcd1
26687 F20101106_AAASEZ caicedo_h_Page_141.QC.jpg
7a5adc2115acc126404c54a7c6672799
40e2d0cfb94f7a11a6eee0c8759749485ce6d04a
7491 F20101106_AAARZS caicedo_h_Page_042thm.jpg
5a51ceab79f74ef8c3df1aaa382c8ba3
90fe4244b68e7fcb79c0623969dd66b1c8d3ff6f
1051965 F20101106_AAARCL caicedo_h_Page_087.jp2
0fc6eea9e871a81221c74b817358b888
9285d04c949aee5f9b88fe95c8f6fcfb631ad1ff
97332 F20101106_AAAQXF caicedo_h_Page_107.jpg
e0c5417e7265ceb965846019263f56b3
bb29a698ddbee4696c6df17848ba945a08126e3b
924030 F20101106_AAARBX caicedo_h_Page_065.jp2
b870f4af18eb6ec62e479f31d174601d
54f582ac54fdb342dd63a169ba77765a19abac89
104083 F20101106_AAAQWQ caicedo_h_Page_084.jpg
2b861a29e8e2a6e83227830e5ed83d97
df07758b7d531621a74b121125d4ab7382f74a61
8988 F20101106_AAASGC caicedo_h_Page_161thm.jpg
e09adbc45b0bb660fc593e0b15f3d7d6
0acd6a329387bd913c4e0eb336b9a216646e6f88
8422 F20101106_AAASFO caicedo_h_Page_151thm.jpg
fa0e7a874c540bb58ff16093112ef537
614cf6d6189fb3adeeb2a1f428b507a1882a93bb
34393 F20101106_AAARZT caicedo_h_Page_043.QC.jpg
615fcf9f1e815b78117c5c8bac9320d5
57420c10fa33b62c585e36a7a66a11a908b31fc5
1051958 F20101106_AAARCM caicedo_h_Page_089.jp2
45081136d572fa4ac0936d92ff5319f9
b879dc130e8030c8c1b78fc3039e6018c822229f
86481 F20101106_AAAQXG caicedo_h_Page_108.jpg
fb514a3c5fbe598347a7564b6feaeec1
128ef8ae6330495a4bb4143d11aae9513dbb8c08
989477 F20101106_AAARBY caicedo_h_Page_066.jp2
3c8f265026dae4dfe446fe03b2532358
e3235da78ec8ea0e11eebd8df2ac269c4f592b58
63061 F20101106_AAAQWR caicedo_h_Page_086.jpg
f0fd47311800bb02bd5c7c43b6fef78c
d7c30d1baf311aaa70ec0c9c031dfe5299ced577
973507 F20101106_AAARDA caicedo_h_Page_106.jp2
37bbf701fb87409e767fbc8ac4fa0213
a27009952014756d4eaef67a1e926503317b5de0
37164 F20101106_AAASGD caicedo_h_Page_162.QC.jpg
fbe4c962a96a9487869bb54ff65b61e8
51c7b76d112e0978c19b3b23fd95bd91c6eb8897
35229 F20101106_AAASFP caicedo_h_Page_152.QC.jpg
6ec49379f9e3d3d849d9841e1bcb07ba
63216e7e8244b0facc9fa6fa7704cb8f771cba23
9009 F20101106_AAARZU caicedo_h_Page_043thm.jpg
c0ab94c441e70a1a3be869d691f7877a
9e89732740c84d471b118f6f15208778d95a88c1
964800 F20101106_AAARCN caicedo_h_Page_090.jp2
b7c3d6dacedf9ae31aa7eb9cc9e5448c
6ee13842fd7855f91a54f0b95258f76a036b6243
94088 F20101106_AAAQXH caicedo_h_Page_109.jpg
ddc85cf32b7d10340aa7a95c488bda63
d94d08c523f3366764b877ec662f16923e23b914
1051930 F20101106_AAARBZ caicedo_h_Page_067.jp2
6f2e48ddd7773946a5d73bcf69c65e83
6b072d6273a2a5a75ef845831e49073553b9de8a
99619 F20101106_AAAQWS caicedo_h_Page_087.jpg
79fdcc8af64c51abd23b8b70b0ab6867
a03662a9729a0a18ee1a3f280cb58f113f4fbde5
968086 F20101106_AAARDB caicedo_h_Page_109.jp2
9bbb56d293b4933831c91c7d5d992113
2d5842b4f650e2424b94d3961b21856863090ec4
8686 F20101106_AAASGE caicedo_h_Page_162thm.jpg
f450c9ecda3e6bb3b2817d4bad8425d0
63c456b579e2579c6a63d326164f5e2ea875e05f
8503 F20101106_AAASFQ caicedo_h_Page_152thm.jpg
e0ea1d337843b08efc0280c64648b048
694971dc07594ef3cc1fa4497b64995a41daafcd
7750 F20101106_AAARZV caicedo_h_Page_044thm.jpg
503f1a2853ccd42c4c3104f773ea9a21
110f5ccdf6f76adaeb42237094f1d290208ff6b6
112494 F20101106_AAAQXI caicedo_h_Page_110.jpg
658da8b1bbec38cce2fc8a0a035c8f15
246858cdf17454f5e53ee83b2d3719ba0edf253f
92678 F20101106_AAAQWT caicedo_h_Page_091.jpg
404ea0c9f5e2c7bf24d717f4d2a35fee
747ab0b3da9d8b60955d2f8e2b42413cd233ff13
1051936 F20101106_AAARDC caicedo_h_Page_110.jp2
70a8d9250c1d671b0dd4ddc04d027803
199b8552089914f0c7a4469185d1762e0feaced9
990749 F20101106_AAARCO caicedo_h_Page_093.jp2
54c69ac5816e62a34f283438fd698a2c
ffa74017f5c05126d6ff9c0fba8902069fe4e70b
8724 F20101106_AAASGF caicedo_h_Page_163thm.jpg
0b193554d194c5934468b838d8f8f9aa
035faa6483fadb4dcf49d4ab3ce7c3c90adb4f70
36053 F20101106_AAASFR caicedo_h_Page_153.QC.jpg
4df54f944c040b964f776e17f18409d9
9ce33957b27d34d8ea03732bc65ba0ed05266fef
34530 F20101106_AAARZW caicedo_h_Page_045.QC.jpg
56a943e51d496037204140a17e76c35d
fff03442eb5cf9064044693f5208fae5b2c72263
88784 F20101106_AAAQXJ caicedo_h_Page_111.jpg
e91bd760e0e718727b9445c7edae3aaa
b8c8ee54e202cac0b0f7e4fdca8f13f3dd4657a5
85828 F20101106_AAAQWU caicedo_h_Page_092.jpg
414c4274508cb3654ddd518e065f7778
65686ec79dadd48f5a3066e808a3ec1c4011086f
948278 F20101106_AAARDD caicedo_h_Page_111.jp2
95cfeb149baca9682ff8089416aa1ad1
1033fa877cd254d86b22a559b91ca75220817e50
972863 F20101106_AAARCP caicedo_h_Page_094.jp2
64a799766a5544ddb164a22e6e7d4b50
cb3ccaf342de5c2e91c7e33dbb577bf7df526326
32892 F20101106_AAASGG caicedo_h_Page_164.QC.jpg
73d83083b79ed48129d5b9f35652cf8f
95aabcbdf66623cbd2fffc0105370570961e4a39
8708 F20101106_AAASFS caicedo_h_Page_153thm.jpg
db903d3db6a4d183957c176a4412f650
5a08a179ac515aef83fc73f22de2ca40e65990fc
32198 F20101106_AAARZX caicedo_h_Page_046.QC.jpg
b0a62a2f211df2d501da2b0d5076368e
6e8f8617903a5bb4efea1ae904559aa5029d84e3
92088 F20101106_AAAQXK caicedo_h_Page_114.jpg
f8a85aaa7c574a8285cb036cfa1c2b81
7000ed7d693a7882dac2032233c8d198e8875265
90294 F20101106_AAAQWV caicedo_h_Page_094.jpg
29d68f5c1b5d125010d1b0c3dea1085d
7cd5f068b18e53c20aa8c73bb929647ac2a20e35
943509 F20101106_AAARDE caicedo_h_Page_114.jp2
ec0e2e1a921fefdb95db9e488cf3f9d8
b9fe24449319d679c64b24972f1b61987d2bcb71
1051982 F20101106_AAARCQ caicedo_h_Page_095.jp2
ead5d1761651341a6d70d30e506df7b2
cf33e8b1604f88450f944c4aa8962d68d26a7a46
8129 F20101106_AAASGH caicedo_h_Page_164thm.jpg
1e7df245dc3a54c6a11886ef756b0213
f27a15b9ea3568d67bf7fb5ffe5b6e2ce38d0b06
35001 F20101106_AAASFT caicedo_h_Page_154.QC.jpg
a014fbd6292cdb182f63ad983a7ec3cb
a08f412c4236f2177ccac909578f9d55bf9b3507
8159 F20101106_AAARZY caicedo_h_Page_046thm.jpg
eba9c9236fc7ddfc10739d0ecbc82abc
be758b2cad943c3bbee9650639cb1c76be591d8d
106544 F20101106_AAAQXL caicedo_h_Page_115.jpg
9ad31f67322725f764c3e4b5e717cd96
f9026590376719693821439595be36b0016f9534
95792 F20101106_AAAQWW caicedo_h_Page_095.jpg
70ec82e85ebdae27678df3f183f09863
f58caba17eb8a4b255a2d9eda5a9fdfe6aa31383
1051978 F20101106_AAARDF caicedo_h_Page_115.jp2
88a7b7f9761ff2f3dccd78859ccf46f1
2e7c55151222986f9c46c166851885aab5beca23
1003426 F20101106_AAARCR caicedo_h_Page_096.jp2
b68bd3c6b3792eb53a5f3ac1dfae2a6b
0fe9dd763e154e14803258aabc8c9d27eaf8ae07
26324 F20101106_AAASGI caicedo_h_Page_166.QC.jpg
e1c127130f6ffe23cbff6b127cf1bc0d
a02c4861bf041b23eb2e7ecdb698d42aa035744e
8506 F20101106_AAASFU caicedo_h_Page_154thm.jpg
98516b4262269abb5697accd2d9c42b2
f0f87405c6a50b3cf3374d2c3acb5bc641bf290e
29157 F20101106_AAARZZ caicedo_h_Page_047.QC.jpg
5cb79ad59c3273c7737c73fec7d44e37
6e63315894fd308af31af4fe3312425e3cfe80c4
108888 F20101106_AAAQXM caicedo_h_Page_116.jpg
72879365e0542a5a798d2c0c961735f3
91ad844b1a01670d413ed892c709a8d265c1f30e
82781 F20101106_AAAQWX caicedo_h_Page_097.jpg
aa937311c4a5a9a9b75542dc665ec4cc
569cb7afbbeefc252d7d87c928fdf9911ccf461f
1051970 F20101106_AAARDG caicedo_h_Page_116.jp2
01180b5215af92ef7512be177f825d99
04886dd1b54960446b25582f58997cd4bacffdd2
97650 F20101106_AAAQYA caicedo_h_Page_133.jpg
95c7419ede9a1f8db7aa7149f240f08d
9dcbb2211b3863388d492f3afac8847ae5a17bda
839507 F20101106_AAARCS caicedo_h_Page_097.jp2
9f3a2cd65136b13d47890bbc3a50ccd0
fa4e9a74b57826accc4bd270bf919dd700bfa645
7460 F20101106_AAASGJ caicedo_h_Page_167thm.jpg
d831cf313c07194e75834d48f0674f6c
00f54c149447a3a255db7ab5e2dc084be70fb904
34087 F20101106_AAASFV caicedo_h_Page_155.QC.jpg
0f4f1ebdb074da8d488983f3b09d7f4b
0befe955d70943a4cd255fdb266d9834dc2041cc
94682 F20101106_AAAQXN caicedo_h_Page_117.jpg
dfee5a6a10d40e0e6b1b125da02f920d
22f2da6c5dbb91a4774a2114180c5d20cec3b83f
108017 F20101106_AAAQWY caicedo_h_Page_098.jpg
782696f6b85b1182ce13616d8be1f9fa
60657c1627acfd8362a93408517c024a4da1cf08
1001551 F20101106_AAARDH caicedo_h_Page_117.jp2
fc53650e3f031437cb2e035984e5d959
5a3c705f11b0748556dea9aff7debc134bd1cb17
91342 F20101106_AAAQYB caicedo_h_Page_134.jpg
a48172509e45b1f2ab92c236d95ea809
c16c05b13858984def9f1026afb0fb987c76c949
1051972 F20101106_AAARCT caicedo_h_Page_098.jp2
954f0d50ebf612fc2125d2ff0f3abb7b
43944e02e47acae765cd675ab0b501646d3d4adf
25266 F20101106_AAASGK caicedo_h_Page_168.QC.jpg
8e92d24d588fa702771e2677f453494f
a322774f0572510b15ab93c7df724362e3087a23
33419 F20101106_AAASFW caicedo_h_Page_156.QC.jpg
c74d920288e5c262b2d5f1531c1fa0e0
1079a92e9e12b360253a414c7bda3d45aeb78d37
104956 F20101106_AAAQXO caicedo_h_Page_118.jpg
86a062af82688c19c1af46ced6ede62a
29807dd56d0990fe05e1bd4434d21046fc695f44
106044 F20101106_AAAQWZ caicedo_h_Page_099.jpg
3583a075c116371123b852215b64b475
5389576bdebd977d3190d98ed29aefc039f5b230
F20101106_AAARDI caicedo_h_Page_120.jp2
509cd20260955aafe9f5b617140562bb
a6e625f50990212136657c811b4976534b665782
103970 F20101106_AAAQYC caicedo_h_Page_135.jpg
62ab457e6fa5bc78956dfac65161b27c
c43d517131bd4830f1dad3bbfbdd7e33553119ae
1051985 F20101106_AAARCU caicedo_h_Page_099.jp2
5442f84e995dc24c7fee5d8e918cbba8
4848ce79bec3a04af18ebd3c7ac7e131fe99e343
25643 F20101106_AAASGL caicedo_h_Page_170.QC.jpg
cebc3e40cfc034f3add995a946dd5177
ef52c8bb723a1c0a9b3d3f64f0a40f3ad6ff963e
8178 F20101106_AAASFX caicedo_h_Page_156thm.jpg
c258bf011b9af36f441e3c322092f3e7
67e95ffb1a7258ff8484805e15037ee3d744d767
110941 F20101106_AAAQXP caicedo_h_Page_119.jpg
3e95f799a5b481880797fccbd7c1cc7b
da1da3873323c65dc7bc397884cec0f5883ebee7
962263 F20101106_AAARDJ caicedo_h_Page_121.jp2
5a187c83ce7a159af0b2e09594f22eac
42ca1508a2ca10ab0c024c7bbaaa99c129f5a11c
82822 F20101106_AAAQYD caicedo_h_Page_136.jpg
5e7a7c118b5480aaeb3345f79f3b363a
3b95804f0aad4bbba829bb83eaf60da318e7ccc3
961901 F20101106_AAARCV caicedo_h_Page_100.jp2
22c1714e4951dbfd665a2c4143847578
7d6ef967e204fa2c0d523a06d8f78afea6413527
31539 F20101106_AAASHA caicedo_h_Page_178.QC.jpg
6271e53dafed5d02cca9606f5031bfbd
afe0fc3ce5d4c6d74bd5832980529bac3d3bf246
6390 F20101106_AAASGM caicedo_h_Page_170thm.jpg
9f0068b0dccb3b7f2ad38b4c6003432e
1e09e4fc67fa3504614c15ab8888a1fe67811dec
30244 F20101106_AAASFY caicedo_h_Page_157.QC.jpg
db773fee9a8a43ca8d558f645559c18f
7da40e31527bc766bc4a916a4e08e0f7ac188aba
105749 F20101106_AAAQXQ caicedo_h_Page_120.jpg
16fb860cd5c977c0f10260b7cdc3816a
86e74c9226af8008a687bf7c59654aa7ea458799
1051914 F20101106_AAARDK caicedo_h_Page_123.jp2
2286b867dcdb1453082d545d1851a4bf
a8de78b9dcd8360acf594f78b62052362443447e
91010 F20101106_AAAQYE caicedo_h_Page_138.jpg
c006d7b7511246a1a26d7b5123c0f997
80fe4f90ddc71c5e21cbab9031403078ee76533f
1013090 F20101106_AAARCW caicedo_h_Page_101.jp2
3889e1ca13b84e5975f2b08d0473d4c0
6139b01cfeb61ee97d64075c235989c779860677
7813 F20101106_AAASHB caicedo_h_Page_178thm.jpg
8709f4b415b8540e241b2d0bf96a6d8c
c06670178a3b687df61172bba883b70d077fe8b5
36291 F20101106_AAASGN caicedo_h_Page_171.QC.jpg
e45b19c21e158b247e0d885f408da294
cea44c8000c0f16804e2da9c04aee01e89b78402
7687 F20101106_AAASFZ caicedo_h_Page_158thm.jpg
b0a87f8e5db681c03ca5c5c9ef351841
fa70cdbf892b6a220dd21902719181292f2b96eb
1051973 F20101106_AAARDL caicedo_h_Page_124.jp2
bdd14d358d0a3af5fd4b7536ca3ee181
9ba8e9f0f6b078950361f0cbdd77cc50c3b53872
93927 F20101106_AAAQYF caicedo_h_Page_139.jpg
7a488d4ba78c73c4c6a2f422a701bfff
a5f320a6108c6835b51abd9070931e63bed5d7d3
919342 F20101106_AAARCX caicedo_h_Page_102.jp2
1bde563a4f1713ea327e6082c6269864
dd61bbc8892fa32388cc5cff24801c8e3507b4f6
29050 F20101106_AAASHC caicedo_h_Page_179.QC.jpg
7d36af5ef6b6256bd7b2052f04d6d6d5
038babf986d45ad717e0ce10d477a6b95ae7923f
8712 F20101106_AAASGO caicedo_h_Page_171thm.jpg
5e683102a12215d8e7ea432265f46c00
c66437c409244e93901c081cd1f5e155fd93eb92
106823 F20101106_AAAQXR caicedo_h_Page_122.jpg
8695136283f832e34c4699e214466e2e
3402d322756b4cbdf4168072f44f3500932821ad
711877 F20101106_AAAREA caicedo_h_Page_144.jp2
188f2a74e0e52991d0de934738320219
0533d66de957b0371754102ee0b8984fe2ce96e7
1051964 F20101106_AAARDM caicedo_h_Page_125.jp2
ef089f7ef12a8c0520928e0484d51abf
79b0d4de268d6bf410773e5215b4d83710d9fde1
111421 F20101106_AAAQYG caicedo_h_Page_140.jpg
2126984747c19ac6d84190e14ad19f0f
9d8ba22de104d4228d2cfc13351e6be59abbe2ff
926998 F20101106_AAARCY caicedo_h_Page_103.jp2
d6cda5d33251e47f8bec58d5d5fafdb9
19804234fd7dad84fa362c8c09981e31c5e0f60e
7722 F20101106_AAASHD caicedo_h_Page_179thm.jpg
a0ec3281e1939b6ed8392e5551c305e2
855239d4f150a93678bc696802e5b73e745a9e07
28400 F20101106_AAASGP caicedo_h_Page_172.QC.jpg
f97198448a93623dfae63f500b9501b1
252ab5da7516f3c67a222a0a0d48a80940c9ec2f
105520 F20101106_AAAQXS caicedo_h_Page_123.jpg
0f36a70c6c211e4932227b4c46dc181a
bdf8ee9b11ebfbc4ebde3bb819f18e275c24185f
975056 F20101106_AAAREB caicedo_h_Page_145.jp2
66b0f69fd86fc2b81a863e8bd22af9ee
2ed10ff2625b6c8eca02b256ba0f464d8830b9d6
1051937 F20101106_AAARDN caicedo_h_Page_127.jp2
c642358a92f4e3d2578bb7282e806772
ea6dc7705be1306b77f2c491ee02d2c12c215ec0
88568 F20101106_AAAQYH caicedo_h_Page_143.jpg
1f4f99357ae07482234d017896d49d7d
9a6b4ea0647e47e403114f5fb4d550f48dc8c069
1051920 F20101106_AAARCZ caicedo_h_Page_105.jp2
9c6962e1f135ebd90e3b6d38444c731b
e757f6756ddd267b81755ac6451ef07d8bd3c5d3
29182 F20101106_AAASHE caicedo_h_Page_180.QC.jpg
4e1aa9dd6a5b3cc6ad4ac144c7e49d60
9cd7a130a1949a26559fa451b721216a6a6acbb6
7303 F20101106_AAASGQ caicedo_h_Page_172thm.jpg
2b12a2e40b996c8a05ffc67c7226ebe6
2357a6c4599a821576e76abbe3c283b4c171776d
106897 F20101106_AAAQXT caicedo_h_Page_124.jpg
d5365418cddfebd36282c58d089f23be
586b33fb9e68ed4051ad4300c7e7d4c102f19350
856212 F20101106_AAAREC caicedo_h_Page_146.jp2
ac141e6ff398a2743996347046e2f90a
fce56a0a152677e6d9910dc655cba4ad4ce533a9
F20101106_AAARDO caicedo_h_Page_128.jp2
7a354bdd7592d2df9e0432e574bbf267
3b93624a29132b2da309f12259b5ff6dac6971b6
67262 F20101106_AAAQYI caicedo_h_Page_144.jpg
44de5e31885e47c6ad4807f0472775bf
df8e4310c1bc92b1b28a213905219b98ffb6cb76
6853 F20101106_AAASHF caicedo_h_Page_180thm.jpg
8173d316d7ef95e2e81a45da12cfb01d
2eb4a43927a3189b7a7855fb67fbf7b0ff7bc930
37090 F20101106_AAASGR caicedo_h_Page_173.QC.jpg
8e15ed8f863e45306449e230d4e69920
04ed4908e5a6b8bdb96bf9a778261016b80dc0e2
88759 F20101106_AAAQXU caicedo_h_Page_126.jpg
e1597c6b0eb9eabb7b8ade9c73c33f1d
2c5f3b0328db1dc46fb4ab80d12856a07329a480
759950 F20101106_AAARED caicedo_h_Page_147.jp2
4d06d47cea9465329b4e446e17c9d055
bad5087ee3a6d41d082cbea7279575362c21fc7c
F20101106_AAARDP caicedo_h_Page_131.jp2
b32482abe0e07c92697f1d415666a923
ea94b9c9c997b3766de0be1884e7e8dd5165024b
90062 F20101106_AAAQYJ caicedo_h_Page_145.jpg
7908c5552335cedab866dd2d1a1dbf6e
da0c7a295fa8ef92cc3326fa7ef129568f4612bb
27396 F20101106_AAASHG caicedo_h_Page_181.QC.jpg
dc1bdd63c966fda4d36edc7813a2c111
6266a896ee5248ee73747e54ff734b5621dd22e2
8625 F20101106_AAASGS caicedo_h_Page_173thm.jpg
9ebbec780dd51bdf89e21f139498300b
9d6515bba7cad6b48dffb97731a937dc56c1c7cf
109028 F20101106_AAAQXV caicedo_h_Page_127.jpg
46566d6c41397bf32ae9a1a910667d22
11035468dff7047bbf736b547b2e8956550955d9
992214 F20101106_AAAREE caicedo_h_Page_149.jp2
a52110edbe710601e1734f20aeff536e
674edf7c7e7b6a017cece47521ed83807a234ed8
961039 F20101106_AAARDQ caicedo_h_Page_134.jp2
eeba822a617f6f173841eae1345cb9e7
738eb6927fc200a840fce5834c4dd286140d360c
77435 F20101106_AAAQYK caicedo_h_Page_147.jpg
1fea015167b936184d852b3e4a96f713
048f332cacbd23df6dfc7346ec52712ae674fe76
7084 F20101106_AAASHH caicedo_h_Page_181thm.jpg
27e549d4b9e3fe281157e9ebdba76f0e
21f332a776d651eee14a2f78e17c285450798015
36110 F20101106_AAASGT caicedo_h_Page_174.QC.jpg
f9896cb69b030e8f92435119f2637c08
25ee25fdb65a8f426b6ebb40ed1130394c2f3d44
99663 F20101106_AAAQXW caicedo_h_Page_128.jpg
da86585a6967a67d69fc59c13198e135
bf224bb422dcb61b3c66a9b764f34c31e6ff4f33
821546 F20101106_AAAREF caicedo_h_Page_150.jp2
cd6c8249492d39d5ed9f699bf4473fbe
92fd5a6d6fd976bc4ef23b77de959bf32cef1a59
1051971 F20101106_AAARDR caicedo_h_Page_135.jp2
d11db4bde8bfa30f6088bdc83aeeb67a
758b86fad76b715afadcfeec54079bf88560837b
89424 F20101106_AAAQYL caicedo_h_Page_148.jpg
b4d2be8a6010d3226e3d8244e9892c44
febdac1b681dc7926926314f84524e02db097a97
24123 F20101106_AAASHI caicedo_h_Page_183.QC.jpg
2cca748afee41c4d4b80f1316a6eca8e
012a23791d18209ea981649ac2799b99453bbe11
8597 F20101106_AAASGU caicedo_h_Page_174thm.jpg
f1cca7f86daeb44a64a45fc5688b8af2
b3b640b487e63573df989d06939a13c5dc68153e
90846 F20101106_AAAQXX caicedo_h_Page_129.jpg
412b5d01e90ac1de8eb39508aee41195
c277f97070010dcf39a4761a972a93128694f88b
F20101106_AAAREG caicedo_h_Page_153.jp2
12de8d504343a53581ba9216feff9049
6ddf9593fca8643b1fe14115f0b856491cf68f16
107488 F20101106_AAAQZA caicedo_h_Page_171.jpg
69e9b71ea1a2268af53ba04a73f7654c
21abc8bfe41528d28784e4dba53549aff99c016f
937232 F20101106_AAARDS caicedo_h_Page_136.jp2
9aacef2d7485557e53cdb4fcc0d8e933
b3b2ab30e411f9180628d7431c7233853ac571ac
74747 F20101106_AAAQYM caicedo_h_Page_150.jpg
812907786c6e95b7804109142f8a5453
03c9231066404cb2527090fb96020c2bb733929b
6273 F20101106_AAASHJ caicedo_h_Page_183thm.jpg
9907cb62c5942c08b27b4cfcff1f1e6d
54328c75caf509270c8316de8d91594452d1830a
7883 F20101106_AAASGV caicedo_h_Page_175thm.jpg
532b3c394a94b4178a0af7c98809605f
9fd1c560b0992f2b8b7b0c8fb63f448059c92d28
87569 F20101106_AAAQXY caicedo_h_Page_130.jpg
9607a8ecd2799a7b9a4c83e64e70b639
765ae9e45503ab5cf96359721c3a2653ea6d2733
1051968 F20101106_AAAREH caicedo_h_Page_154.jp2
c69300e44e76196a8d437e6fb81300e8
1af707d2a7df4156e29dbd87907352d977da9749
84613 F20101106_AAAQZB caicedo_h_Page_172.jpg
70652aee632734051fd1b1ceb6d51043
3832d5dd6f82e8da0c527f06dc7339a31e13f565
1051934 F20101106_AAARDT caicedo_h_Page_137.jp2
245ed7fde43784afc33547ea09cb0a2f
4df482b04346aca6fb290fb0e07bb1404ee9949c
106559 F20101106_AAAQYN caicedo_h_Page_151.jpg
6d202775159bf5e8449cf77de586a3c0
2353f9094c7f0fc2f722e573c9e30dd3a68329ef
28375 F20101106_AAASHK caicedo_h_Page_184.QC.jpg
dc9e75f26a7b62e47d59fd5b6d234168
e634e2b39a614fd149b0a691afeb70e3b3cffbb1
29510 F20101106_AAASGW caicedo_h_Page_176.QC.jpg
15de4a6df9a81b7d0ef2879e3894e13c
cbbff9cc1bce0bb3ffeaed0a82c3a5e56954d957
95584 F20101106_AAAQXZ caicedo_h_Page_131.jpg
4c5b5504556bb090c14f96842ce91dd0
4f2366412e2134c14cbe27e0343e13457078d544
1051967 F20101106_AAAREI caicedo_h_Page_155.jp2
a3e4bcc3e9717d7cfad7d1dcc0cb2bc4
e91e7e9f22b3ab0cc3a0121e163fb72017dcdf63
83312 F20101106_AAAQZC caicedo_h_Page_175.jpg
80e2c6f78cddf71d1caf9e75505969e0
573bd23f5ada6bb44a9a5efce3faaf3b4726bb6f
981971 F20101106_AAARDU caicedo_h_Page_138.jp2
e49ba08083e4602a6ee3cd2e484e8378
b8eab22c2888fe886c339784b9a2ccd725b2afcf
106554 F20101106_AAAQYO caicedo_h_Page_152.jpg
5832f23a0c68b9b51f345cff74cc2894
c429960463c7160327594e0867c7dffab14c3bb8
7725 F20101106_AAASHL caicedo_h_Page_184thm.jpg
e6b974adc7266b8eb66850e9f2df8725
7b3f3bb590113d06b91bb51c9243c66fe3828720
7622 F20101106_AAASGX caicedo_h_Page_176thm.jpg
bddfd72a782e20e3c10a8a58d356c98f
5dcca0c4e659b90efbc3fe74781f2e2b109903d7
1051925 F20101106_AAAREJ caicedo_h_Page_156.jp2
aaa3b50042113bc3516592829139cb22
a3770fab4777fdabb69eaa164ca41e306dcf951b
86499 F20101106_AAAQZD caicedo_h_Page_176.jpg
2bb34808c60b50ee0a5193469f75febc
957238aa1e71ded17cc6b0ac2110d219dba8d920
1020721 F20101106_AAARDV caicedo_h_Page_139.jp2
b7060ced320eac0d25b32b7f07fda579
9455abc405aba9b218519904a362f8ebfc09b021
112646 F20101106_AAAQYP caicedo_h_Page_153.jpg
a560f47c3ad671529dd6e34a31e71605
298a8c042a205d7efc8242c768146d0aa0289134
29850 F20101106_AAASIA caicedo_h_Page_195.QC.jpg
d070a9d5f6a2564f4909ccc9d62ea129
d4bdc852a42e2acd324c67b70b298e4701a21ba3
8076 F20101106_AAASHM caicedo_h_Page_185thm.jpg
896291f0bb2027b5cea5195fb4568e01
4141d944f8ff367fce0735d22ce3e6285f82a068
33362 F20101106_AAASGY caicedo_h_Page_177.QC.jpg
3f5e7218472fd889c27362bcd8407858
71dd95d7e852b769b2cbc2b408f8d68484526550
988053 F20101106_AAAREK caicedo_h_Page_157.jp2
3938868a1cac20550cb53350e423bde0
c7eb9c88110a6b6f82defb1b6c6255513e98ed0a
99652 F20101106_AAAQZE caicedo_h_Page_177.jpg
c6b6c664b043acc5e4a0e8c928c30ec7
3dbbe093813d9186ea15cf55157ab1772f91b615
1051980 F20101106_AAARDW caicedo_h_Page_140.jp2
de72518b3b1d0c0c86b9dcdde6a9d980
16385935e1da77df7a92c2081f69c844d48b8b58
103389 F20101106_AAAQYQ caicedo_h_Page_156.jpg
642bd441d277f13731c3321fadf31727
9de16ef0d43e9ab36facac1b4df7d9197f3c4a6e
8364 F20101106_AAASIB caicedo_h_Page_195thm.jpg
1319f729d0ed49142b21b112c3a2f1dc
d7e47ae4adfa6dc79c8ce40ed95f0cf1d9e24838
32212 F20101106_AAASHN caicedo_h_Page_186.QC.jpg
c28c85afa13be7922cebfc243174fe6e
6c17dce62274043e08bbd80eb3f7effe6e3f29c1
8459 F20101106_AAASGZ caicedo_h_Page_177thm.jpg
7a52de1855650fb74bb7dd963a30f8e8
79322158fa03a4cecffd4391fdf3b80938fbe6b2
1051975 F20101106_AAAREL caicedo_h_Page_158.jp2
b44fd7d9f44141542487b4736e1be0fa
fedf3448b472c677ebe251e7f8b28263640e07eb
97447 F20101106_AAAQZF caicedo_h_Page_178.jpg
fae38603dc2eab8f409d3c8a6e1c9c66
f3203fe027691eb6149528bf43d4b460010ddf66
844366 F20101106_AAARDX caicedo_h_Page_141.jp2
cfe6d77448b7db11e115333dc531af7d
3155840ee712e14d399504c2351b4c28fe6883d5
92608 F20101106_AAAQYR caicedo_h_Page_157.jpg
8e4888d9cd108a6ba0b633f17f475882
b9c4e8bf302fa0419eacb528c71697afe87c2097
33651 F20101106_AAASIC caicedo_h_Page_196.QC.jpg
138e312f16eaf8e831465731f0a71fc0
5757cd086a10eb8c2cecf2932c91abf478e35f31
8122 F20101106_AAASHO caicedo_h_Page_186thm.jpg
42e602f902b25e8cd691b49bd392a528
9e630c5c2f25680c1a72ab3e4e6eb609033ca5a8
F20101106_AAAREM caicedo_h_Page_159.jp2
6c090a468ff365aaaaa1b75b0b40dc93
52bd27a3f118a47b8a29b9cc5f9e513359679ed1
88863 F20101106_AAAQZG caicedo_h_Page_179.jpg
60df89d44a7480e57c3b8df4e6d87bba
4928219e9f12e007e5aae7aedd1caacea3bc6613
937905 F20101106_AAARDY caicedo_h_Page_142.jp2
0f3d473aa7c4157ce8e467da95f5c5f1
f203f72d38805181fce7e34c656297eeef763f4a
946854 F20101106_AAARFA caicedo_h_Page_176.jp2
34b427fa832a4ca2fc2f27cba5516f61
acfabdeebfb03685e8f8576357ffef47ff16686d
9056 F20101106_AAASID caicedo_h_Page_196thm.jpg
d0a853e11d0266aaadc652de3582ae7c
329da54d1951606107c0c9aae75ccd1ae9ad0730
33077 F20101106_AAASHP caicedo_h_Page_187.QC.jpg
177521428102e4b8d8b69e9261a9258d
593f248458f5a5fb459eba0c3197630656244f10
F20101106_AAAREN caicedo_h_Page_161.jp2
0132ec63d55badc83c7edf4502195fc2
685f989f775c678d512f7148f3fc25f80cde936e
94308 F20101106_AAAQZH caicedo_h_Page_182.jpg
5a4c5c30585cdeef28dab61d9ed65df2
beff3d0ce53aa99458713f0c163995d6532b7553
937258 F20101106_AAARDZ caicedo_h_Page_143.jp2
9102437c25049167a226851f14b31465
763deb6b355c24658764625bc1534dfb5d2933bf
89782 F20101106_AAAQYS caicedo_h_Page_158.jpg
e0e9faa5cb4c4d80a45abd5045cd14e9
81d6ba91863d39af18984da75bc5ac2c38e64d78
F20101106_AAARFB caicedo_h_Page_177.jp2
e31f69560808d41465a388c06e0ad281
dca2fafb1eb36778c5c425dc5f3bfcf3261f4de7
31953 F20101106_AAASIE caicedo_h_Page_197.QC.jpg
876f9abecec2c7a57c51725440ae30b9
37fb0fe695588b5196a4fd0cb3ac8f1ca89ab406
8414 F20101106_AAASHQ caicedo_h_Page_187thm.jpg
b10d130b46efd94da51b2efff2ce17ed
89f7fb45b25f6da17aea87b96e2612b3a178b418
F20101106_AAAREO caicedo_h_Page_162.jp2
eb273e8131664ad42c58b00e332c162f
9b436892b6a6499537e73e5072708607fde8f404
84459 F20101106_AAAQZI caicedo_h_Page_184.jpg
5e3e86301c4cb68e2bf638790eaf9d54
1009c16ca462cdf2750387c6e9f2e636247d7758
92193 F20101106_AAAQYT caicedo_h_Page_159.jpg
8376990d1766d3e51ce021ca59367721
178bcb643bbfb4457738dd09832fbf694d6b7b3a
996897 F20101106_AAARFC caicedo_h_Page_178.jp2
e45140f806d2800bf247d9163ed3b05a
d6ca720a8b94dea475855a4e7545cb8701efac2b
8910 F20101106_AAASIF caicedo_h_Page_197thm.jpg
95c59a3eaac003f101287eef3df57d7e
9e7cd513cb221c50802ddc42142f6410ffec62f1
15367 F20101106_AAASHR caicedo_h_Page_188.QC.jpg
fd393c5d2199b23528a1b19b7687880f
e75f80d7d522a97463fd10fddd6aafa08c6962ab
1051953 F20101106_AAAREP caicedo_h_Page_164.jp2
ef6d60ec75d4f749b81a126038bc51d3
3eacadda265a6eef68b266af900ee990a68cf109
108830 F20101106_AAAQZJ caicedo_h_Page_185.jpg
699c6d9a9b69920fdbd950e1b41669bf
5eacdb0015400fd869d8c3e9bfa20b681a7d05aa
72784 F20101106_AAAQYU caicedo_h_Page_160.jpg
65692ebfa300be7c05ddb54ad38a99a2
c3cf827060b2f937c8c70b4356a0cfe8136a4bdc
945991 F20101106_AAARFD caicedo_h_Page_179.jp2
29656731405bbc4fbd9814ce53527bbf
d89711d88a662e86730a2324a6f4c7328741c42f
32400 F20101106_AAASIG caicedo_h_Page_198.QC.jpg
0dee3b24a1443a4f22ae4e022f49bed3
048ceee2bbbf1e013b00dc2b4e9d1fc6575854c8
3868 F20101106_AAASHS caicedo_h_Page_188thm.jpg
e6e00293d954728287865ca4c9648b99
16dfe302c898a18cbf49553eb5c6b11819840860
1022232 F20101106_AAAREQ caicedo_h_Page_165.jp2
6177ccfbd207f21f03119584f4376b0d
0660b906dead832bedca0464f032a3db1c40cf10
46919 F20101106_AAAQZK caicedo_h_Page_188.jpg
4f9f222103fe752cdacaea38a7d817b3
67cbaef0ecd25b24a438e37e68433d977695171d
95075 F20101106_AAAQYV caicedo_h_Page_165.jpg
6aa79ab04b44e2bae977c72aae648ae8
678f58506a8b97544c7244df767d7e6a594b0bc2
1022324 F20101106_AAARFE caicedo_h_Page_182.jp2
820f9114f75a9cc512151079b15f1f28
609d2ca319a2bcf2bbc82407391e696cd7148b53
8579 F20101106_AAASIH caicedo_h_Page_198thm.jpg
904300632f632d768665cea785c94195
d358d934145cbbe890fced1b5d0b690866a1c7d8
8711 F20101106_AAASHT caicedo_h_Page_189thm.jpg
9a289a7ee041bb5ae90c82338cbeecc4
7409970264a0298496709f5e4ba00bcd2bc89760
827375 F20101106_AAARER caicedo_h_Page_166.jp2
602044a1e8f5a54640b2d88b34fc4747
c6000f8a07f4a55771623cdf9062cd74c9225afa
111044 F20101106_AAAQZL caicedo_h_Page_189.jpg
af4da55c83e4c0d363f8ec5553b1228e
6db956ce1e4ded0252884d1394038eb876606704
79503 F20101106_AAAQYW caicedo_h_Page_166.jpg
246a1f23c500ce825564836096eff791
7d4d414ec8c14a6fb3e93c747f491e2dfb41784e
790911 F20101106_AAARFF caicedo_h_Page_183.jp2
74f06755f3a8a152afdc278bf4bbd201
9b51aa38dd44ff7403f69ca13c533c60c5b0ee52
30302 F20101106_AAASII caicedo_h_Page_199.QC.jpg
acd74049d899e219b2c3fc6aed96fe52
43de5e11341dfd07ffc895b7b8a36f7e69dd108c
35320 F20101106_AAASHU caicedo_h_Page_190.QC.jpg
0cfa122cbcbb7b56ce915e2ff7bcb77c
ad32e41e7075712571bff2194faea84fd8f07069
916630 F20101106_AAARES caicedo_h_Page_167.jp2
449434b52d549f61830a89244036b0c7
171c61610fa3f01fe69c6bd5658ed04cf1dd734f
107343 F20101106_AAAQZM caicedo_h_Page_190.jpg
d0f9e3f4b44fb939c0c14c2c7cebf682
a2daa2aac715c787f49598b79d91ee991ef5553a
85779 F20101106_AAAQYX caicedo_h_Page_167.jpg
2c6e368aa315b6ddedebeb2c214e183a
fe51fc795bfe41eeb1ba31b828a6e65d9983209f
818002 F20101106_AAARFG caicedo_h_Page_184.jp2
183450167f8123c71ab7b36891644900
09f552e75bd3fbbe3bc980017d8e0aad0606fefc
8329 F20101106_AAASIJ caicedo_h_Page_199thm.jpg
0915767e1171636d5ce37411db0a7074
3b6ae818b22f2707777c47349ca577e4d46876ff
9027 F20101106_AAASHV caicedo_h_Page_191thm.jpg
09d60a0e16a8f3ddceebbfd8ad0a5276
0ee3cb09b78acc85ae84e136c79d06f5746da1b8
825680 F20101106_AAARET caicedo_h_Page_168.jp2
528794495494e8b1242cb2f0acfdb318
5aadb3646b2ab63341844a45f2f0666341f5a2a4
98850 F20101106_AAAQZN caicedo_h_Page_194.jpg
cc7ea3e1a2a74aa0184f2354e955c062
0c3778fefed08821fcbb722e7ada67765f1bfbeb
79414 F20101106_AAAQYY caicedo_h_Page_168.jpg
b08ea020273aca331f0c6e50fc65894f
e7f681ce4d08c456b1210d79b4e43767015fb7f1
1051949 F20101106_AAARFH caicedo_h_Page_187.jp2
db6c8a41faf50258eb0158b53a67cfdd
394ee7df3bf57c4e5166d936596c50639f5f5233
34939 F20101106_AAASIK caicedo_h_Page_200.QC.jpg
37687613ce577a3045fdcb1ec6688334
1d71dfb22f464a3ae1cfe0c23d43cb6500db24bc
8688 F20101106_AAASHW caicedo_h_Page_192thm.jpg
ca0ba78d03672947609466c3ce78cb08
317b1ab407f24b9b4001e70745f100c8a443cd55
F20101106_AAAREU caicedo_h_Page_169.jp2
9cad3451eee3de47f03b9b1ed75603c6
2ee6004aa07851e388e8240d57e210aac6c57fb4
94253 F20101106_AAAQZO caicedo_h_Page_195.jpg
9851f293b8a19f76a5292c19ae8a0edf
ab81131529e12d403aca34bb7fcb49317d3b0bb4
77238 F20101106_AAAQYZ caicedo_h_Page_170.jpg
819149028e9d1bbbbf38f54128d2b889
0401642dc52b6c6ee2f02b6215e80053520e433e
499981 F20101106_AAARFI caicedo_h_Page_188.jp2
50b08b5a3c3fc536b1857d8400016609
d26e6243c1b4bc560fe5488a2070573ed55513d6
31710 F20101106_AAASJA caicedo_h_Page_211.QC.jpg
a682ff47167f7c861c56a8c35b8f5fb3
7285635df360d655d4271e59fb69c968bf597177
9187 F20101106_AAASIL caicedo_h_Page_200thm.jpg
712b89cf0903ded6eada6964d6345b0c
c6708d9ca6d5ab66c8308a27a2ba14d787567e1b
23452 F20101106_AAASHX caicedo_h_Page_193.QC.jpg
688ef42de8b5875b6a07d9a15ee9be33
7e2ad99f314bded4338202e0b42febe1dfeb9e56
829915 F20101106_AAAREV caicedo_h_Page_170.jp2
3545ce35fa1bf9a0b94711c166a8e691
4411f588c0752c6e8f4e8b3ad28257bcd7e6b22e
104111 F20101106_AAAQZP caicedo_h_Page_196.jpg
fab6196692a644c593e81455cdfa1b2f
97cd371ca31e9957d5ff81ed69d537158a682232
F20101106_AAARFJ caicedo_h_Page_189.jp2
584f77755224cf98eac5ff7d940119c9
ccc32518794dd11884e25e8cd49e290407c58f70
30914 F20101106_AAASIM caicedo_h_Page_201.QC.jpg
098b67294598434d605695bf2e0e4b92
2e16d38be1180a638329a87a78406292bc13b697
5682 F20101106_AAASHY caicedo_h_Page_193thm.jpg
c4f800988bdee9cd2aafa2ac283becda
eb58c7b383245eb114031bc988ad8bf189624a29
F20101106_AAAREW caicedo_h_Page_171.jp2
13967041484b432a96de5f37f784bf70
d11d03c5154db23a8ffff0d3276b17fe85dc6a6f
100810 F20101106_AAAQZQ caicedo_h_Page_198.jpg
75614899995438c92aaa339553dc8c3b
a83ece4391f5a0f47252eb1e35665a7b03895b2b
1051957 F20101106_AAARFK caicedo_h_Page_190.jp2
7452628a9f3b20b0ce04b805711025dd
fa737f1dcd2d9b92efd7f6c2c196e34c26a1afdd
8430 F20101106_AAASJB caicedo_h_Page_211thm.jpg
16a34553f6cad817b66ce4a525b6e317
1689f15b5248cfccd2dc3f7af62fbf147b00642e
8071 F20101106_AAASIN caicedo_h_Page_201thm.jpg
d7b4eebe6f95488bb71fa20d438b1359
def407012b7919b1a5fcea0dafd7d0612d318f62
32651 F20101106_AAASHZ caicedo_h_Page_194.QC.jpg
565cefab4eb6d1dc1ce6f7ae2d28400a
67571c27cf7d4a5e8254b22ba449cfb784dbdc82
881368 F20101106_AAAREX caicedo_h_Page_172.jp2
abf1e42362a88cd5d791ef97dd62a5b9
3b95e71327c64d275c028eebe0a47fe46d95000f
95311 F20101106_AAAQZR caicedo_h_Page_199.jpg
f125a9ff96ff2956f1eebd2332ee8a63
68145852a3dbb4ac4785a0f5754c8ffe698a38bc
F20101106_AAARFL caicedo_h_Page_191.jp2
b6bb863cf1067e3a17601b0d76f5d94b
8749fa7de38ef775a39a232316e439888e874b37
7351 F20101106_AAASJC caicedo_h_Page_212thm.jpg
cdcd7eea07e74eb20f408f1960dfa794
51c9ff1b7530663daa35889e5075924eedee7dff
8935 F20101106_AAASIO caicedo_h_Page_202thm.jpg
c3d0de2af55d1145d12ba013a81b0be0
63938523631ad306b1dfe9c75b3e4cfcabfe5a48
1051984 F20101106_AAAREY caicedo_h_Page_174.jp2
2df3924616391c44607595345361a24c
ee099e594b82288da7568de6639b314db2897f05
116507 F20101106_AAAQZS caicedo_h_Page_200.jpg
c293291281a750af329cf167f317885f
156918373645243bd3e9d6e09ff5b83339014a12
F20101106_AAARGA caicedo_h_Page_209.jp2
ba8424c3feaf106a47e693208dc8075e
7021db4f4651e9d1928f8caa2ce61fce2c7b84a8
F20101106_AAARFM caicedo_h_Page_192.jp2
c2138dd011839832c562a13358259d40
9652c7baf3c3c8f7cc19c39e66ec31a90473c0e7
31707 F20101106_AAASIP caicedo_h_Page_203.QC.jpg
ebaaebca9771c0912cf38b7e185c7178
8e4c0b0c8d5db9481762953fd9f24dc92f4efbfb
865110 F20101106_AAAREZ caicedo_h_Page_175.jp2
f9f5a6454b64e29bbfe2d71ebf395f98
79e8e08667732c718c2ddfa81bc5eb1fc83e1ea2
F20101106_AAARGB caicedo_h_Page_210.jp2
42dcce88c7cfd8fb5970862bf7655189
fa82d031429ba5ccda1cd2062c7961ab97f4975e
777648 F20101106_AAARFN caicedo_h_Page_193.jp2
b2044cb246582041b05d304d1ee48e21
408983bbf4a837f5b25e164b3f7862ff1004561e
8683 F20101106_AAASIQ caicedo_h_Page_203thm.jpg
a7d02a796b13a1792ddf553a43022d3e
c61dc4ac5c7961604fe744a631518a9d6ecfb008
106371 F20101106_AAAQZT caicedo_h_Page_202.jpg
53e127d77b6eed3e447614e91d8db61e
db14d2e84926ca2338081a1175940723925bfe5e
1034841 F20101106_AAARGC caicedo_h_Page_212.jp2
773b7efd407491e5f23d0a4fa5d12f59
e9a1ffdb83eb0212db9b1770016e6442b2cb19fe
1011556 F20101106_AAARFO caicedo_h_Page_195.jp2
bcad5eda989aff732f74ed800d5dd50e
017790491c54f0932bece56b2321f6fd34628772
8655 F20101106_AAASIR caicedo_h_Page_204thm.jpg
69dca8fada16ade709f5ce2d11c40fd0
1601e5704753f95160cca96ee17d242e022e76f5
102810 F20101106_AAAQZU caicedo_h_Page_204.jpg
eec571615bcb08495d3746daa3fbbd9e
36ff2e402b185bd2492e8cc614533830e54d41e6
25271604 F20101106_AAARGD caicedo_h_Page_001.tif
3a48c6ab8d71ca63ce4798a4d5e81c2b
6ae4cbc1c5164b589775e8323b2a8de863527f64
F20101106_AAARFP caicedo_h_Page_198.jp2
14a4e170439d41d91b7bded9153a24a0
0d38a0b5654aa3810f1985c91868152bc4886c34
8839 F20101106_AAASIS caicedo_h_Page_205thm.jpg
d733bc91ccc173865b52994f9ca44542
218cd92b80d43c6c8c82f99d04297ef1ecc487e9
96962 F20101106_AAAQZV caicedo_h_Page_205.jpg
74c8a05a21d16ea02c7ed72b293267ca
e2dad73ed7425385c0d9a845951c4168a18a9c8a
F20101106_AAARGE caicedo_h_Page_002.tif
e3f5e6fd469e5ec6e103234d75fbeddc
09b704f1254676eeab7aef067f3ca7bc34856700
1051899 F20101106_AAARFQ caicedo_h_Page_199.jp2
cb740d16eef7fc57ea200601d1c990da
a9e04ece091c126af6ee7b71bde7428dd65e71d6
34232 F20101106_AAASIT caicedo_h_Page_206.QC.jpg
07343b3ccfa724fa810e671d8f976055
0bd38fdacdb4487398e5bf98cd9f7a9ddd4c131c
106947 F20101106_AAAQZW caicedo_h_Page_206.jpg
9b62f2e06cc292117711a7b8af776acd
d0429d9a77150a09b11002497580a5c1fa5050c3
F20101106_AAARGF caicedo_h_Page_003.tif
9a2466a72f907cf41c79fbfe851cebb7
59e993e1c0edcc8ae866e949d31b62d9dfbcefe7
1051923 F20101106_AAARFR caicedo_h_Page_200.jp2
359bd963fdb05472370a88c9804cbba5
7460d7f04c1048a3edc32df55e3e14c874e3e488
31845 F20101106_AAASIU caicedo_h_Page_207.QC.jpg
4288df197095af773e63c1262b5545bf
3cf89d60e5182f4002c4996255574b8869eeff53
100538 F20101106_AAAQZX caicedo_h_Page_207.jpg
fd184e10614479973f16eec55a8aef71
7fe8e2a67e5e3eb0a51c33e85f254f0a5f53ba29
F20101106_AAARGG caicedo_h_Page_004.tif
60ff96641dc9d184ac82bd0082d7d88d
c24ce508e133c4c50a74cdad695941e028f09f30
1051963 F20101106_AAARFS caicedo_h_Page_201.jp2
b2e2d5ffcf4b07d3de1c7cd66b2f14d4
78e96424bddb67649738ea7d30825abde92d7f34
8384 F20101106_AAASIV caicedo_h_Page_207thm.jpg
580692cada8a1b83ed02fe14e1a3a3a1
12d285617204e4400416c67da893b680654bf395
100791 F20101106_AAAQZY caicedo_h_Page_208.jpg
d8228b324db8c5554021a33701c67fc2
84474283d1b253a019ca69fec08942149bb32a3a
F20101106_AAARGH caicedo_h_Page_005.tif
6d0fa3a7b7205ce86c6b1a6d1c642a39
27a5e2516341d1575484f3d03f9b4be979adeeaa
1051889 F20101106_AAARFT caicedo_h_Page_202.jp2
40ddd4db8f0b3cf0973317f954ad7474
0fb9f96a1b412e3c8779ec8f9cdbc6f450d97285
9069 F20101106_AAASIW caicedo_h_Page_208thm.jpg
3098fd23b554bed9674b4e838acb06f7
78d72c5bc2caa494ed909941054ac8116c0c03ab
99065 F20101106_AAAQZZ caicedo_h_Page_210.jpg
606442c5632707087dc1c1ad55efb751
6039b379ab563ecc54827cf45707ba3c80791259
F20101106_AAARGI caicedo_h_Page_006.tif
5f38994609c0506a0cf62f15e9e3c456
2b6e08d82bb8738e0d4db0a83a07842319c4fbd4
F20101106_AAARFU caicedo_h_Page_203.jp2
ffbc9b847fd92f2c9a94e973a55bbf1e
9dd3e3bad9cdf120ce954d53e5d5c7a48ed37bb4
31868 F20101106_AAASIX caicedo_h_Page_209.QC.jpg
5c8e2ba41a35bcfc3507fc954aa33683
977d1fd5973e206b82fc647a4feac02e839a10c8
F20101106_AAARGJ caicedo_h_Page_007.tif
3a7140c8a560e2407630964101092d08
9960b32ba65fc62b117787b59c05efdfbc0eb5a0
1051927 F20101106_AAARFV caicedo_h_Page_204.jp2
8dd129d7abff48584bcf11f770bb838e
89d5c3f0efdb3cdb975d635a732b191a7ddbff81
8234 F20101106_AAASIY caicedo_h_Page_209thm.jpg
ecc6f2ebcf07e5a65a149107a7a7755b
245d10e7573cbbf27ce7d0d6fb1991dd701d27c6
F20101106_AAARGK caicedo_h_Page_010.tif
13c3b8f25e899f02fd942dd0abb6c9dd
3a9ed799e21688e51f38bde30097c423ea5e5e81
F20101106_AAARFW caicedo_h_Page_205.jp2
da8acbc9234373aeebc61bd5aff71225
3c4b95d9a9b748d2eb018994bbaeea6254f1c1ef
31139 F20101106_AAASIZ caicedo_h_Page_210.QC.jpg
4e0c0c6d11a0c83bc4359f06efeac823
88e66c3349e43308b41304d7cbb110a1846ad56a
F20101106_AAARGL caicedo_h_Page_011.tif
ddf37dae15e72cb0dabc62a3892c3dd2
a0ecf770f8269d0fb926ebea8f34eb224cfb41fc
1051941 F20101106_AAARFX caicedo_h_Page_206.jp2
fdd53b5f062f96f3c1a6d36e6911499e
44f74d51f81914a170a529765d8e2e8f4acc0512
F20101106_AAARHA caicedo_h_Page_028.tif
f27bd8b97133159308525cd02d4044ef
e9fee24543f70a3af1f39de068353f899c7a66cd
F20101106_AAARGM caicedo_h_Page_012.tif
b74ee1d4c13cf1623ddc9ef871374662
87701eef66037a44745b3674f4c078de5de457c5
1051909 F20101106_AAARFY caicedo_h_Page_207.jp2
b5e65bca28edcf28ce84efafe331adbe
4e69358b958787494bdb9fcb40a138506200cc65
F20101106_AAARHB caicedo_h_Page_030.tif
240dadc2e0fa2a355133ce0fc41bfc15
79fba408d39fe186534bfb289a707e339525925b
F20101106_AAARGN caicedo_h_Page_014.tif
9ed63f88c5ff0054c13252030bac7954
c4bbb274e7cd1b76547da8aad9470559ddf49141
1051969 F20101106_AAARFZ caicedo_h_Page_208.jp2
7896782f5673658199cef6cbbc4ccebb
f4f325516e4aa4337319c11a12e94d3b76b67229
F20101106_AAARHC caicedo_h_Page_031.tif
584e39834a47437e3b9d3e336ae1f35d
3defc1c8505f0af6557a937439f90822b1d6b196
F20101106_AAARGO caicedo_h_Page_015.tif
2ecebb11ebf946eea700475278d9d3e6
146b1532812b386b2b27589be12888152e366f1e
48009 F20101106_AAAQEA caicedo_h_Page_198.pro
eec4157d3f4aafc0e55398ab08f0d154
a08303ce36224ff5d28b6da0fd949df0a845c38d
F20101106_AAARHD caicedo_h_Page_032.tif
7553dbc657133741413e85e7fd924ef9
54a09ecc7dbc6797079b456d06e8a5d927d1055e
F20101106_AAARGP caicedo_h_Page_016.tif
1f0143913b3f50953537fa8ff4ff4f63
dd4bb5afacb67867d83f64cf569055ea7f9b8357
96402 F20101106_AAAQEB caicedo_h_Page_112.jpg
80567e9ac2a3160cfeedfb5510977dff
164b09748fd03abf15b5182a9dd2e58dbd49dc05
F20101106_AAARHE caicedo_h_Page_033.tif
a84b1989c5c1d252a5e8f59c30eb811a
265f7971a26cca3b72093a32f92cca587cc1c2af
F20101106_AAARGQ caicedo_h_Page_017.tif
b16ef1432f2f7f943407bfab53306f1e
53877e1f5858a54921b02140fe1da1ed52053cca
2109 F20101106_AAAQEC caicedo_h_Page_099.txt
faf081c58756746007754a4a403d9606
4cf42110c77eb3c755c5e1233b16bafeb979e99a
F20101106_AAARHF caicedo_h_Page_035.tif
0eb2c97ec2743b17d7682ed452137dbb
7a5ecf384a2764c42746a3405f8b68ecde2598a4
F20101106_AAARGR caicedo_h_Page_019.tif
7bd0f97eaaac2d0df7db688f0016df0d
4f3d88c398370a8b26cac671e5a2b46dfdcf2ed9
88316 F20101106_AAAQED caicedo_h_Page_090.jpg
da2e7d684acd8f080c4fdedcaa283178
ba15351e74d69e324067f03ccbc4d1932fe3a174
F20101106_AAARHG caicedo_h_Page_039.tif
636a9617cb506f777082a7f5e5a36c1c
5b365470b9c1b1c6f2c70e7dc7a4849aa9f3ab2b
F20101106_AAARGS caicedo_h_Page_020.tif
c7dc8204422cf78d92d8721841b29b04
8e7cc148bac0649336e402f5453590564beef9e3
1051979 F20101106_AAAQEE caicedo_h_Page_107.jp2
2a789d9e29d9659b93425e7b0aba160e
37cb8bbdddeb24ba16866492fdf98ae152e4db40
F20101106_AAARHH caicedo_h_Page_040.tif
38ae7a2cfa7fe6586229d194cac455d1
357565a39e8b9e8e4b501eacc82d649032c7deae
F20101106_AAARGT caicedo_h_Page_021.tif
5f042ff3a817883e86b8dab08ac26e10
2dfd9ada89d11529b7692bc0c39c485bb978187e
46462 F20101106_AAAQEF caicedo_h_Page_066.pro
5391e07c40e75127dc4783f14014b74c
2bf602d4c33ddc053e38a1102d1a3ea8e8857acc
F20101106_AAARHI caicedo_h_Page_042.tif
2bb65e29777608832647328003e0ae0d
150d4f3c099094c45411ddd32f6a5c2c82902c0e
F20101106_AAARGU caicedo_h_Page_022.tif
ef29ccaebb76ab07b094b6e3c7b29015
848feb315b201421bc7a4a97777f30cfd3681ea5
112941 F20101106_AAAQEG caicedo_h_Page_191.jpg
d753ac6adb4f98edbe31bdc007190f83
c7d9d4efba63b39a50b3b18c4c7f6fd4e65b24bc
F20101106_AAARHJ caicedo_h_Page_043.tif
1c3ab7ec31237bba674b754f71a79369
b8f9105cff29a060ba24a3b052dc5ec62cc34ef6
F20101106_AAARGV caicedo_h_Page_023.tif
3384cc24fc19ba6d9a9a0964166efe4a
c669820e3896efc6b945a137eaa88c9a8d0caf98
98213 F20101106_AAAQEH caicedo_h_Page_197.jpg
f3678f7d74b522e169a6693d34916100
c5d2645b5d1c275d8945c6ed2d12152c60e71992
F20101106_AAARHK caicedo_h_Page_045.tif
1ae9bff3798b749bb2d0ec3d01a53ba3
638f1c932bf7346f9123f1b1de2ef986ac6f27e5
F20101106_AAARGW caicedo_h_Page_024.tif
faf5a4faea2e75b183b7d39ca7969a79
28ee3f1e03b2dfd470ef0f5a156799e861024f24
32241 F20101106_AAAQEI caicedo_h_Page_208.QC.jpg
4e4033964a55866105255cef5d2f7ff9
0ab663445823a5aaf5dad91a9ed3dbe877bb7a3f
F20101106_AAARGX caicedo_h_Page_025.tif
e0bc53287231a3e7727a9d2138d6c0ce
1798bf3bb9f4db40aeeb8cf7662fda5924381c9d
F20101106_AAARIA caicedo_h_Page_068.tif
2bf9462890849ba2ebaeac5490cc6629
27f4001f0e1aad9541cd58f062904bd183a78229
1051986 F20101106_AAAQEJ caicedo_h_Page_122.jp2
4adaae99d6bdfbb3a64ef65e26785f4e
8ecc3e66b401f426bfc02d2d6aeadeec53293a73
F20101106_AAARHL caicedo_h_Page_048.tif
7ce904d999800b5c839704d256ee10d1
18aa1059961f7ff8e0190b0886bef2832b78d22d
852462 F20101106_AAAQDW caicedo_h_Page_108.jp2
08102d38869d37fabef3c21e2abd0827
7790e7559e162d6a0b44d10d3c463aa1d4f5cf1d
F20101106_AAARGY caicedo_h_Page_026.tif
7918b2d09eadb5122041f01e287e28ca
7d7656676be32021c7ea072141c08ac68e9a9b9a
103449 F20101106_AAAQEK caicedo_h_Page_187.jpg
5b5eff51df60e3a764084be4c6e04ea8
a7ca888831fb27f9fa8cbe80e700e5dc42d3257c
F20101106_AAARHM caicedo_h_Page_050.tif
4644886c436b315f8d417082695c539c
3a6abe4312f10884982539ac50990999e60bf0fb
1039544 F20101106_AAAQDX caicedo_h_Page_082.jp2
ffc5d9397cefb49b19321c7a1508682a
6ffd691ce2c6dc3e01a0a82798ec89d217248dd3
F20101106_AAARGZ caicedo_h_Page_027.tif
3320c120d76f28f4eaa40a769fae4bb3
03af313c04e2412f13b6c04d886c0a2f39c0b100
F20101106_AAARIB caicedo_h_Page_069.tif
4bf200c6d8ac759016790c409c6facde
2ff1097e7ee54df9ed01c8a9a70f27ce27743592
2272 F20101106_AAAQEL caicedo_h_Page_143.txt
43fb3b0817f1f34e85580abcbb726918
ac90b50bb44bd15a1f8d941305d176f0ee6055bb
F20101106_AAARHN caicedo_h_Page_052.tif
edf6e44f91df30c9c555b2fbb7404c7d
38ea19b39a7285a515ba4de32975386e1fe740d2
F20101106_AAARIC caicedo_h_Page_070.tif
d152acb8a9c5f16c36b8f01de5e6db44
0817af2982035f0e5415609df34ac7403028bb78
161595 F20101106_AAAQDY caicedo_h_Page_014.jpg
0ffbe40f19ae5793c75978aa5a611e32
a1c03cd9b2e69c69b1c57c7d8ebcfbe671188e58
991644 F20101106_AAAQFA caicedo_h_Page_044.jp2
ea2a52fbc96fb9c7ae912eb6bf2a2f17
f9ef8c8750a44f4696cc550bd739f5e033946f08
112077 F20101106_AAAQEM caicedo_h_Page_173.jpg
02f191b2dc1998b1139a94e95f107211
4b0a305c2430b051ba80fb48ece14f08463e0e72
F20101106_AAARHO caicedo_h_Page_054.tif
7857977dbe332026bd052a55aeeb6172
266ef1fe087760f5d42562b0248ddbe57c26f8f1
F20101106_AAARID caicedo_h_Page_071.tif
6ac7ad577493a3adf8443f7f501e4b2c
f24565ba467397ba44a782e76e977a7f0c3f33aa
7671 F20101106_AAAQDZ caicedo_h_Page_073thm.jpg
87e446d943e393e247078e7341bf7360
d89a3d7e0c2fd0e93f529fe9591cc18650d75a69
1051956 F20101106_AAAQFB caicedo_h_Page_088.jp2
7106495113e317e0e4bfc2cb9149f229
52c8a6ebd86b0515b1044313d452dcb55c87249c
83166 F20101106_AAAQEN caicedo_h_Page_028.jpg
c075e1bcbe99b3f06a8aeccc0e45d2f0
e5d9de51df081647042983f154c6d2dd91c2431e
F20101106_AAARHP caicedo_h_Page_056.tif
e22e8a4c502f360ee8bd472317427ddf
27d030355053a5110142045afcb80f5b9c45d892
F20101106_AAARIE caicedo_h_Page_073.tif
b408dc2ea010fa8ee5bc58c4ce75ba8f
daf029d216a1c10e04f0aa13d6a1cac07ef4d454
43531 F20101106_AAAQFC caicedo_h_Page_175.pro
0d7d07819e18731d2bafb11d62c8ec10
53ece3aaa0f849c20707f3dc8416db7121b6e1e1
112400 F20101106_AAAQEO caicedo_h_Page_009.jpg
06d60afd5f31c1c50cc01b3932160a0a
18b55365cac581eafac5202b0bb2aa1458ea9a9c
F20101106_AAARHQ caicedo_h_Page_057.tif
a0f69daff698f37a0cde3e7a97eb7045
ab230969a5c77bc207077782cb7c558c2ea51cb8
F20101106_AAARIF caicedo_h_Page_074.tif
ddc3d0e9ea70d3e1f46302678810c5bf
17751e1ea4c99eece71975713efeeb929d3669ff
103350 F20101106_AAAQFD caicedo_h_Page_155.jpg
4e6f5f5db187faa79b6b69855880ccc6
848fe4d67d5b414abc9babe1a8e82eac0b10ee44
F20101106_AAAQEP caicedo_h_Page_125.tif
5ce22c248ab343210968f0f8bf7243b2
e4d7b26101fa5260f077e05dc06aaf72356e603d
F20101106_AAARHR caicedo_h_Page_058.tif
551ebdbf8fc874f3f1e384cd20bedbf2
dab8ddc6d79515a7727a197bd7cf3e1ebfee9688
F20101106_AAARIG caicedo_h_Page_078.tif
26f611a412f0f3c92ef1d5dc3288b888
fb71ca14ee2ed9882172f0043cb4340894af9423
F20101106_AAAQFE caicedo_h_Page_199.tif
90635caaa0403919af204f2b27cf0a52
45f229429cef692851fa98716a1d153c70d7a7c5
7108 F20101106_AAAQEQ caicedo_h_Page_147thm.jpg
956b958e46375250d32547262d8cb7eb
47509dec91d2dc69297e0e980792e1ed74e6000b
F20101106_AAARHS caicedo_h_Page_059.tif
a3ea3fc695c10b84c41f9002e0697626
a320a32411224775b75984cd0268f3251ed722e0
F20101106_AAARIH caicedo_h_Page_079.tif
a1e7333f128fc1728c324d472ce2b2bc
2b5b28425d8a15429bd99958a3de78fa499ae0a2
F20101106_AAAQFF caicedo_h_Page_186.jp2
d3e195318979402e5f5226847eba01a8
be0a001dc07439ecacf56ffded2f90591812a971
1847 F20101106_AAAQER caicedo_h_Page_211.txt
5e79aa14f62bfdaa6d8bd5b2d10e779d
a894d6931772d6da9811bbcd9230591b0f15e9bd
F20101106_AAARHT caicedo_h_Page_060.tif
035a9427fa90df4e227ecbf56e6c0646
abf4f925cba62727efd042155d5d33398750301e
F20101106_AAARII caicedo_h_Page_080.tif
927a92ebfb859ea4e6972b8f16041180
394b4019cd51790beab1dfcc97efde2f49313b5a
7005 F20101106_AAAQFG caicedo_h_Page_168thm.jpg
a27a1bb9995711a2e3fb57f7ff9970bd
b5a8d56d354626ce475ea0a8be806a9a6c24afb1
42146 F20101106_AAAQES caicedo_h_Page_139.pro
ae75e36ce5edb7357f50edb8ac9000f1
d34c1435792c103ffac14fd13aa7627497df865b
F20101106_AAARHU caicedo_h_Page_061.tif
85a9ae9b48b105ca39f7f22b1a9bf931
0f6b21d8a5950766f8b07f74c9451236e9b896d1
F20101106_AAARIJ caicedo_h_Page_081.tif
744423c46d7805828c81794a47eae498
88e250d06e1ff28b03b3162bf27fb8403673cc3d
107647 F20101106_AAAQFH caicedo_h_Page_174.jpg
7542eb8584e102d98dd8b56fb1122485
f4fbf323ca3bdded6a47f97024ce381e909a0d8d
1002336 F20101106_AAAQET caicedo_h_Page_181.jp2
31802ca0ff462ede2caf9b6a23741e1c
baba4bbfeda61fd2e6b6dc6a90b2fd868c7d5fbb
F20101106_AAARHV caicedo_h_Page_062.tif
2deca829e93d240fbf82db8678dfc211
d8ac573e6e77708afebaf5c16f490944d1748ce8
F20101106_AAARIK caicedo_h_Page_083.tif
49f6c9ee38e33077a2cd27e4abc0de05
5dc1207c48a67ed83a50a13efdde58424a06c54d
7436 F20101106_AAAQFI caicedo_h_Page_103thm.jpg
cc5e441d3ea7cefbf1178acd9a55c05f
195f622994d6a02de9f9c7099e3c1953d1d44fad
32877 F20101106_AAAQEU caicedo_h_Page_064.QC.jpg
233bd7b0dd838d287f63179306c46b29
75260f2c48db30f89fb83c6183c7769fe1906a46
F20101106_AAARHW caicedo_h_Page_063.tif
4c896f10c3a2f1746d9c0fa129fc8344
d5228caae78b3a8506bb2c04ec3baa99e0a7bd71
F20101106_AAARJA caicedo_h_Page_102.tif
6b633659ea5fa1405aa3af07ae0ab99c
05ffaa6c40688a08b733c3b12dda57b173a7451b
F20101106_AAARIL caicedo_h_Page_084.tif
aa5a007ce78f795421e394f609a72419
dae9d80fb748ed5b567cdba60ac32646cb3e00a0
997778 F20101106_AAAQFJ caicedo_h_Page_112.jp2
6f62dea52fd07798a000c0bc381d1ccb
111530f5e31a42fa6c57beef75a61c3730bcad93
2587 F20101106_AAAQEV caicedo_h_Page_035.txt
140e705a279749a6971cf83a09a38b6c
d9b5274436643b0210d52afcf3c874fb24acf042
F20101106_AAARHX caicedo_h_Page_064.tif
45e94a2cab533abe8d70d195347f82f2
4ca67ea06c6d0cc48267f572701795fe26e40273
F20101106_AAARJB caicedo_h_Page_103.tif
4d6b43af3c9d488a97169d9aae8b44ef
2c00e79e8992259c4f4ca6c508bc5c48d9ec7b23
F20101106_AAARIM caicedo_h_Page_086.tif
149ea2bfeefb1cccfdbadb75411a8e09
a47e8690ef66e15671ebad407332453811b4a3ee
F20101106_AAAQFK caicedo_h_Page_160.pro
d5d1143f7f89a3ccee9371cfa3ec036e
b32a498c09198a0ef204b92f5e88cfe32a931869
8460 F20101106_AAAQEW caicedo_h_Page_074thm.jpg
e11936e5f466bef5b13e5e6930aebbcd
e25976f2ce6d79a7c72a1dbdf4788306931852e6
F20101106_AAARHY caicedo_h_Page_066.tif
9088f5e5e27e63f8982382074770dd2c
24c2a2ff6c1bb767fb6ebafe023703c4c7a9e6c5
F20101106_AAARIN caicedo_h_Page_088.tif
17bf317ba0534779814aa0ea76b3632e
85c5acaf6a1f9aeb63a1c00aeab482341939915b
53707 F20101106_AAAQFL caicedo_h_Page_120.pro
8bc1e2876bbdf4a0c1fb522b5d8cffb2
9524ed0724b776c3663f9a142e2d6d99b0af6d9e
108546 F20101106_AAAQEX caicedo_h_Page_192.jpg
1da702d17bd07f7c73ec9a524a69ec12
550d64fcd6bc73a341464c16cdaf0ea9a7f2333d
F20101106_AAARHZ caicedo_h_Page_067.tif
09b572dcd49d5a3ab6bb9804a8a4ee50
2c1f26f7e134dc7d799f33e9be05a3fe9fefab4c
F20101106_AAARJC caicedo_h_Page_104.tif
ebc78b75f54c75d310e69154e3463d0a
b16c51b048d72c8f4e4d3e3a436a65003cdda0f8
F20101106_AAARIO caicedo_h_Page_089.tif
1f441e9ac97e796b4fcd3f0c0a29a1a5
5371c0207c6fb5322aa004efc883995273020180
F20101106_AAAQGA caicedo_h_Page_144.tif
d135de8d3f6f6d2cfb02005d46c707bf
55d13ebe0222491b215886fa9ef7da28138dd4d5
925667 F20101106_AAAQFM caicedo_h_Page_076.jp2
d1182d9eaff6ad30835c3986962a8199
77ddd1b514bdd90581887ba258d30acc845aeb6d
7697 F20101106_AAAQEY caicedo_h_Page_034thm.jpg
8f15fbd04d3f7343b2bc8178e96e3664
3be4e8687f9c2ca5c30157fdfba9d9ef301aaf09
F20101106_AAARJD caicedo_h_Page_107.tif
69c7616ad71ab780d4004def4986cd26
51d791ef0a4b2ee4ccdfa09ea31e760cc499fa83
F20101106_AAARIP caicedo_h_Page_090.tif
431eb5a088ffa577d7af39285dbe52e7
ec04a43d6568e0baf1bec4323fcd4ca97ece4d8d
30976 F20101106_AAAQGB caicedo_h_Page_143.QC.jpg
c35c7665278a856775ea2b01db25acad
5368d79559cb7767956294e2b42368c1d2e29c06
7506 F20101106_AAAQFN caicedo_h_Page_134thm.jpg
93d09e7d028a97a7fe2b89541cee63ae
f22be2e3d6993dad81496828f68170dd773134ff
F20101106_AAAQEZ caicedo_h_Page_082.tif
37e44080973e2b7ab2fb511d5a749ffa
05f0b35d6871f9e832155c2d46fea3bbcfadd534
F20101106_AAARJE caicedo_h_Page_108.tif
e3ff5ff28a321ba62b4b67997bbfdf50
654dd6e8f9158f1aa6ae12ab4677091db9af667e
F20101106_AAARIQ caicedo_h_Page_091.tif
66212a746de73bf29c55efc62006e09f
6cb17c25a3108efb63fff7a50f99306d3b4ab90e
1051981 F20101106_AAAQGC caicedo_h_Page_173.jp2
7f61d15c97c3f5e2567c68d0a55db1b3
6b7fbfb65f8a4e7145ce5905e3faea49aed22ac6
51916 F20101106_AAAQFO caicedo_h_Page_105.pro
4b1dfedd7d1ab51f43ce66ab56ed3ec7
62a36ee2ef9d5a91cbd30dc8caa2174eb1abcf67
F20101106_AAARJF caicedo_h_Page_109.tif
f50963a40d6e64a0e0aadcab755a4cbe
6e107d7d1723eda36b313bd14a7cf6dd385f0e5b
F20101106_AAARIR caicedo_h_Page_092.tif
bfdb1861f3a52087131fea0c65b27bd4
081d7dd4a6b4c1f215f0262d8b214ec6929b7566
7934 F20101106_AAAQGD caicedo_h_Page_032thm.jpg
d7c50224554c6e168193801a2b2933e7
c689fc4560d4470926cedf63a2af6998b34df739
F20101106_AAAQFP caicedo_h_Page_133.tif
b36b128d2915b1fd326324bb87031916
6b8657cc1bd2342f38fca4b0132469a700972bf7
F20101106_AAARJG caicedo_h_Page_110.tif
f45ffb22047d9f92321e3a06e24246ed
b950b6c4a626cfeccc1f8307cf41391933c05e9a
F20101106_AAARIS caicedo_h_Page_093.tif
964dd7d7ad6182b2717ffa7695695acb
837aa8ea0bd85c64ed788f3960121561da550ed5
1051948 F20101106_AAAQGE caicedo_h_Page_151.jp2
3a599f729a4f131b43b01337d8c064cd
832f728000f8c96fdb283346c6258d7ad4fda412
1051960 F20101106_AAAQFQ caicedo_h_Page_045.jp2
8bb27f7aa510b3eea413912a46d3bc47
4ca1855f9ec78d7f4d557ed23e77cdae9e88a524
F20101106_AAARJH caicedo_h_Page_111.tif
7831860dc27d8820295089ff4b2bd0de
3b39762be280407b7805a5cd1d218353643f42af
F20101106_AAARIT caicedo_h_Page_094.tif
254d71d30d54b1f599a2ecd616376c0b
18763146e55c64c4d05cfb4308aa7051975694cd
75497 F20101106_AAAQGF caicedo_h_Page_183.jpg
a02b05cdc01d92169a1c03d08b855e88
1a112fc36f29b5520fc2c79f086e2ceea56e99aa
F20101106_AAAQFR caicedo_h_Page_169.tif
1eecf66dfc71ba792260f4ba562ddbba
080ebb007bcadb6f701b2cfeca1341d846d3d1b3
F20101106_AAARJI caicedo_h_Page_112.tif
f3339039938690195299d686ba29c68e
dc7088994763a0e32947e26a1e3cdc48287b29ad
F20101106_AAARIU caicedo_h_Page_095.tif
6d6edcc99dea879bcd8df85dc3dc6349
c2486de4e984d167b3edc489b78f77109dd270d9
2177 F20101106_AAAQGG caicedo_h_Page_055.txt
5148949e4676f78c8da1f68257099255
bc174821e401735fcc84de4170fd42bec0a26255
1039013 F20101106_AAAQFS caicedo_h_Page_040.jp2
fcc129eb8489eeb7e2218e82947af4e3
cb5c67d109e442d812e99ca62b768af7746e3a70
F20101106_AAARJJ caicedo_h_Page_113.tif
f81a9d288debf444f55cf278ff55d1d0
462752f49023efa6fceff2ee4974fbe3e588c4dd
F20101106_AAARIV caicedo_h_Page_096.tif
7952f1a282f30234b23b208697ba5dee
af5d44b7a76c3776b8734bd08ae02bf087215c58
32044 F20101106_AAAQGH caicedo_h_Page_204.QC.jpg
c8a3ae485acb7d4684d62fdf7821d637
2df697f7d72b2d0cb8a8009f0f70ceccd694ddb0
2216 F20101106_AAAQFT caicedo_h_Page_085.txt
c747fe302962731244aa53e7b1706065
64b95aa4b19d59b397f1d0ad7022f5ec5da293a3
F20101106_AAARJK caicedo_h_Page_114.tif
2e958c052d9c4891f117537fe68fd1a5
7f2a4261d8456e1cff53f34415226f1038a32308
F20101106_AAARIW caicedo_h_Page_098.tif
99062d54fae509315a8e69fab2d014f3
e20a051a5e3b1ca8b017bf5cd22aef1006001983
45835 F20101106_AAAQGI caicedo_h_Page_212.pro
2ff587e5b6e58690cf36f477d080b566
cdf18851185c6ee7f0cf279162dd63899f5aa855
1963 F20101106_AAAQFU caicedo_h_Page_108.txt
bcd533916e0839250c37fc917524e9cd
8ed3bfcd503b1841857d63de951b1f1d7f35da73
F20101106_AAARKA caicedo_h_Page_141.tif
66c5416f4735ecc8015c8d83267a2a04
8f07e0ff9d96f86b429b95ae45a34562bc0a8729
F20101106_AAARJL caicedo_h_Page_116.tif
e600075c303e8778dc393d6b77050bbc
ba027167e308c638e96fc5ad00ed2aea7e7567f6
F20101106_AAARIX caicedo_h_Page_099.tif
0514aa11b708b4bd81f8d683efe4fc27
897aa83bf2600e8be70ce1bd0cc3bd8931da4ac5
1051983 F20101106_AAAQGJ caicedo_h_Page_119.jp2
5c97edb29671b507794502657719e347
3ab64022d2ad99c8b9b4d774d5f72364b4563d00
678635 F20101106_AAAQFV caicedo_h_Page_086.jp2
6b3ba1f4e3000db41ceeb94963168e18
6cac73818a004dc3d6160044aec5a2f78b21d65b
F20101106_AAARKB caicedo_h_Page_142.tif
5675957f17468945a35f33b0f65aaa4a
a7954d9563d4b7b1fef59e171f1b3c2f283df792
F20101106_AAARJM caicedo_h_Page_117.tif
94c10ae5f0b2c23542a7412d19d728bf
4337668ace2bf73d3e39b2ab36ddcec79be86ddf
F20101106_AAARIY caicedo_h_Page_100.tif
29706bd3823a04e1d8aee448b682d4e4
734c0e7071d786def0eaab13397d82151f1b2fdf
106941 F20101106_AAAQGK caicedo_h_Page_061.jpg
fe0ef503a89b75d486f5ff9829774658
5ddb255600ff5b34204511f280409d39de530fe1
7732 F20101106_AAAQFW caicedo_h_Page_093thm.jpg
f367d5774fbbff75a95484c874861ec6
db1db2e49929a1aa18659042f39e3727127e40fe
F20101106_AAARKC caicedo_h_Page_145.tif
0b991f7778fd02955b7c63329f3e4787
a3e911c99cc465f1b0b9b201f79075e4bea26d50
F20101106_AAARJN caicedo_h_Page_118.tif
65db476275d176aacb1a60766a30e593
58a583af005ee57ca10460595188ce19ef791432
F20101106_AAARIZ caicedo_h_Page_101.tif
54dd427028b745c4b3a9d48996917b4c
2453d5a7977cff9cd36e753da14f3be337d23c22
F20101106_AAAQGL caicedo_h_Page_185.jp2
4352aa98fd40a2537eab892925125fdc
828314e839122112c342470c327842e4d3e75a24
2182 F20101106_AAAQFX caicedo_h_Page_113.txt
aaaca0bb3b175e7c50f6d472f58ee7c8
ed2c34dfadddfb0cab990b7c41eedc171fb4aa52
F20101106_AAAQHA caicedo_h_Page_051.tif
cb0d5774cd47e1ddd5be31249ff5ff52
ce5f984601bbdc672ce24624d893dabcfe56c3d8
F20101106_AAARJO caicedo_h_Page_120.tif
2f53ed117b0a0808b6683161767a1d0a
544f9351c65426bd8b44803fed760c408c35ea46
F20101106_AAAQGM caicedo_h_Page_077.tif
3abb992205520c474e2f1617b06bc46c
5349a911edd3d7c2c830fb93e0dda100a9c5bd56
7579 F20101106_AAAQFY caicedo_h_Page_009thm.jpg
c0e2df6c4ce2e7750dbebb44eb4a8245
42052c4cc715ae369625120c4146fb78616c9c18
F20101106_AAARKD caicedo_h_Page_146.tif
131e0254f31d4f82df6b250803fc64f9
dd15b8e1d8063f7bf0556ba32d4aef7986fe088f
F20101106_AAARJP caicedo_h_Page_121.tif
d64ed2d0e3440b747896376abc8ec01a
f373157d0b24a180ed3c77cedba12fd7ac894060
2157 F20101106_AAAQGN caicedo_h_Page_155.txt
6854802d3b7995c0423d8671f5c823dc
9a38a14fcbdcc6ab43abd6dcebd8d60136753d75
8131 F20101106_AAAQFZ caicedo_h_Page_051thm.jpg
4d17e4808de4d65c0c267d794a66ed73
ad924cca467a4b4a4054a696655d06ac132c28c2
497 F20101106_AAAQHB caicedo_h_Page_001.txt
4b517c7303404ea3c476cb3817737405
001f74c1cf461abc902224a9af5c80514db6e5e7
F20101106_AAARKE caicedo_h_Page_148.tif
f9a9dbdd210a03e66f8e589c1d6569e2
75bccf22017a1d5c4159d9e7ff6916b41162ad48
F20101106_AAARJQ caicedo_h_Page_122.tif
c603b1db98e4186b93dccb6359e94a1c
5bfaa6bb296982729a04fb92b20376a6757292ff
163908 F20101106_AAAQGO caicedo_h_Page_013.jpg
6388242e32b5633ab99df99b6ae39996
cf7de8f74c9036c6637ff0a42760b7cf7e7521d5
110275 F20101106_AAAQHC caicedo_h_Page_163.jpg
70fe72b2ebb04c97c3697a6b0021252d
2fa999c887246009cb2669926e31a9041ca473d6
F20101106_AAARKF caicedo_h_Page_149.tif
90f41eb0640c2b2beaf6503a6fd6b57f
e66dfe4c9bc20039daba81e6c5d4eb930672c6e5
F20101106_AAARJR caicedo_h_Page_128.tif
d65b78adc14b006b54ada4409abd1477
aa9a056a333fb9670b4ce470ac26e7b1ec0a385d
1051921 F20101106_AAAQGP caicedo_h_Page_113.jp2
00cbdacfb4915058e04a82491a3c5cb9
47256c46f78a780c81ec0133f24620be41975143
8781 F20101106_AAAQHD caicedo_h_Page_083thm.jpg
87db45c71435ff47de0caceae37a6bea
8b5bb7fd3a6fe3387709eb02e85e9d83766ce7ca
F20101106_AAARKG caicedo_h_Page_150.tif
ff13aa18880d29d4e334b22975fc7e80
4c3a1f8b6d86fedb89c2f300e76c0deaa439c652
F20101106_AAARJS caicedo_h_Page_130.tif
a16d3891eddd8d88247d32058ed597a7
74a7b8608a2ed17ec61db4a982aa80c505529874
2296 F20101106_AAAQGQ caicedo_h_Page_064.txt
4eebc203ed3c51e5ea0e8da05d16af52
d2bbd4d2c53c3a51ac18012c8b56cf4e1c6d7ab9
F20101106_AAAQHE caicedo_h_Page_195.tif
a8befa5bee0f4b8023c0762601cd95b9
6fac9f2d22cadfeaff01e241c972d009f2c846ea
F20101106_AAARKH caicedo_h_Page_151.tif
2f2ce6da4f4f22ab00691c9b68b975e2
20fe53b47a8fb3f8bf06de3296988c68178359b3
F20101106_AAARJT caicedo_h_Page_131.tif
fa55b62a1094b4f819bb0dc9b8c2fa96
2459adef0fd5a832a3ce6c9e13416e54d0492047
F20101106_AAAQGR caicedo_h_Page_211.jp2
c96d3e159acf19e02b20725293a9a4b6
a2f0d673db5b5c65e31b4b8ab9d7294f4ccf1df7
1692 F20101106_AAAQHF caicedo_h_Page_180.txt
fbe00bb09b1a6ce8c03988a0f3e455f8
b20769a0367ce5191d6b4c19264a3ce642c86eef
F20101106_AAARKI caicedo_h_Page_152.tif
b8084b0c70bfc596085947c7e12b19bf
3d6979ec0a9cc9885a04271b4561e5bd051f691d
F20101106_AAARJU caicedo_h_Page_132.tif
d0a26251c8c936b049af8b552f46ce0f
8689d6aa20aa5a3d277ac8715449aa17cae62760
51568 F20101106_AAAQGS caicedo_h_Page_171.pro
d0717597f89164143a756504ab0c28d1
a8b8c3694ad410c555d96de3323b8db3f3e08688
8739 F20101106_AAAQHG caicedo_h_Page_105thm.jpg
d6c72f3ed5017bde67e6fde85c695254
bbb2925b266d4d40e7f232255947cd59d95fd946
F20101106_AAARKJ caicedo_h_Page_153.tif
4a2a1568d663fdd1269d74c1af51e2aa
29d483ef13a482472f26b08266aaee8148e64604
F20101106_AAARJV caicedo_h_Page_134.tif
f3fe9c8a4598e725c26a5c29b6b03140
6b94d28cd0c5e599f8a489945e488cd5df3327c4
F20101106_AAAQGT caicedo_h_Page_193.tif
bff60ae1d9335f471c0c25653d2811c2
7f7918aa597aa30e909b7c596cabed2d13f67f22
F20101106_AAAQHH caicedo_h_Page_154.tif
79f3638e64b1448457ce2e0ef182a734
07ccff712305b8af2236e39a9ffcca970380b462
F20101106_AAARKK caicedo_h_Page_155.tif
aee79f498719b635c1b9a2377f2479d0
ff3dadece99f8137444ef2fc80dc40cdfb0e0e11
F20101106_AAARJW caicedo_h_Page_136.tif
d7523dc329f4035491ffd7e255960605
838e48d02fdccf22cfc92d37b580b6a513ec4c04
1002098 F20101106_AAAQGU caicedo_h_Page_080.jp2
517380c0071e3391d981165e2955e2e2
75eed542748b63c70bbffddddd4eedcd38df09ce
F20101106_AAAQHI caicedo_h_Page_087.tif
79dca91275e356282367ef6e58526066
82afbde86665c1a439fa44ca710b99066c65e1d7
F20101106_AAARLA caicedo_h_Page_176.tif
37bd1d9c72326562923ace0d2df65063
4796a4443642ba1144ae424beff391966580bab9
F20101106_AAARKL caicedo_h_Page_156.tif
474c17679d45b30ca0921557a92e289e
f407ada7c7f9fa5c08be7c3a0fb44ff48e789112
F20101106_AAARJX caicedo_h_Page_138.tif
cb89e455901800e290e1f57e651edf5c
1bb60c540606de4510865d2cf5a6108ac22edf60
F20101106_AAAQGV caicedo_h_Page_008.tif
48aa387f267379e754cd3ea0d9d698c7
6e8989c758d2095dee5638e754879e70bc5357b7
43281 F20101106_AAAQHJ caicedo_h_Page_039.pro
aef26baddc2a028a5152922ab88b85e0
bec08cf1e799beabf7837b6097d8566051f85f5e
F20101106_AAARLB caicedo_h_Page_177.tif
873070a6908833cc485c05fd7f902145
d68b2847228bf3db55c5ad08210ae20c3e9366e1
F20101106_AAARKM caicedo_h_Page_157.tif
04316412e5cfa8ba09c3b4e9adbc3ece
150a042cb9f90fd639dd7710cd869c5daaa55fba
F20101106_AAARJY caicedo_h_Page_139.tif
65aa213c0272fa452b7bdcb7eb53607f
267be814015409ed0237883f2b7913e44b9d136a
7629 F20101106_AAAQGW caicedo_h_Page_096thm.jpg
0e2bf35dca716b6fa599ace643b43dca
1e954207f54b1eb1a5e5f533937b6e3843c9fd97
1902 F20101106_AAAQHK caicedo_h_Page_197.txt
fa5cb6d00c0aa0cad896a8fbfe78ef50
f65553bdbc336fde839bb943d647b465f61385a0
F20101106_AAARLC caicedo_h_Page_178.tif
f1ee9d1474c4d1157998aae279c48231
5fafbef43c10241c0aed4c5c1289330e36e0275c
F20101106_AAARKN caicedo_h_Page_158.tif
9c099dd1617c230dd73b06804ce89ac3
d19395e65820e8b34aebd4be9d86fff7c8df07b2
F20101106_AAARJZ caicedo_h_Page_140.tif
20747b1192648e35997d1c4e2553d060
aa12da7236ddc0a23315013e7cf53315e49a419e
1051906 F20101106_AAAQGX caicedo_h_Page_196.jp2
790163992c5cf5af8d21ca70ba695182
47b43b466cff3450f01a6ff71d8c32eb22464245
28758 F20101106_AAAQIA caicedo_h_Page_102.QC.jpg
c4599b0d8dc1a4c89eccf7901ccb8080
afcfe24ebd47e39a3ba6004501860d87aa1252c4
2076 F20101106_AAAQHL caicedo_h_Page_202.txt
5edd0218b94a1b0ab53832f048a11fbd
91795782ccab5b20fb3539d604c7134cff84d616
F20101106_AAARLD caicedo_h_Page_180.tif
e97548166a063007adbd5ca5da8ae06d
10787b4ea32eaee644554bb7240351c5b9a89b5d
F20101106_AAARKO caicedo_h_Page_160.tif
1dc07c254741e3862953e927457ee87d
5a4ba0967c73f14294b02e493d7122ebe1b054e2
25783 F20101106_AAAQGY caicedo_h_Page_147.QC.jpg
0e995220afa812aca5e14a0a438221da
36913b9b06a23120f1008cde3ecc340bf0b8c353
F20101106_AAAQIB caicedo_h_Page_053.tif
e1411b85212ff2056bd9a5f08ee3b1c6
e75884863f51b9160e812f242c54afa472da0a51
2055 F20101106_AAAQHM caicedo_h_Page_123.txt
3eae2c7d1b8f531cf149ab2097a56b97
a5f595b9cbc7dc1be7aceea8c162645f2f86339e
F20101106_AAARKP caicedo_h_Page_162.tif
6ee43c91d282384728308e7ee0ccfe95
19271b8bd5ac9e248190006c1846fb44ef245535
8492 F20101106_AAAQHN caicedo_h_Page_190thm.jpg
eb78c76c2e70c2cc4913b2d19773df48
a06f27dcbf1c9e5e81effa1365a55e1fd469fce8
940114 F20101106_AAAQGZ caicedo_h_Page_126.jp2
04035e31d5256e9cef4e691148043441
e7b2c93ba3d85c39ce194ed295c1fd459b2b8535
F20101106_AAARLE caicedo_h_Page_181.tif
9dccc6e6a338fa66539b511c92fbb3d2
7b46313296672b4f5945999c96b14ef1de23e688
F20101106_AAARKQ caicedo_h_Page_163.tif
dc5d6f50ae947cf1f74408cc46d25851
3bf30b5bea42a4d86476bf9cb5b684956414edf6
116494 F20101106_AAAQIC caicedo_h_Page_007.pro
b86481c9cca39266e84941bee1d64182
97098e792e1fb1a64cc43ead8e46055357a75724
51602 F20101106_AAAQHO caicedo_h_Page_072.pro
c38efb48ff1d5043ebebe344ea0842e4
006800923aa77b3f7a7e1c252260fb677a68c3b3
F20101106_AAARLF caicedo_h_Page_183.tif
90c8f3d37278b57827b1e24c7e19693d
fa830fdb7f50fe401ea087f6e8f26ce2d9b6681e
F20101106_AAARKR caicedo_h_Page_164.tif
69c85d7aaa290901263ad6310b43ff00
0a64191149d2854a49e8cfbbdf46ab753c7e4609
32765 F20101106_AAAQID caicedo_h_Page_021.QC.jpg
e51e33974857bf0aaf85bce72412598a
6b7a21266b02377e8d6de05b1404dfcf4b09607d
8064 F20101106_AAAQHP caicedo_h_Page_114thm.jpg
2d7e4a0bbf071548fc1cc63a2a0e66b1
e5589d0e49fb01603b0281979c8e640e4f4ccb8d
F20101106_AAARLG caicedo_h_Page_184.tif
3e0113cc0c0b14b5f688cf55318ede4e
c738cd768ecb6aa6449de4ac0431a603057254ab
F20101106_AAARKS caicedo_h_Page_165.tif
abd5fe25689a9a2911366274804f6296
da226644a70baaa0ba0850ea95936dbfd6179c4d
9132 F20101106_AAAQIE caicedo_h_Page_010thm.jpg
fc16f5e31bd692c67ab6f75ae05fedbf
fdfe1edd16a7c17ebb56523bfcc1d870efd3005e
977271 F20101106_AAAQHQ caicedo_h_Page_129.jp2
1ba48d7dadfd72a62c49591b7f5f560b
4e7f52abce4ba4a0b911a20e81011b1004c3e60b
F20101106_AAARLH caicedo_h_Page_185.tif
50453e20b97b18478af08da991d36ef0
6b354042e74701e6493317736fa79426802a9f47
F20101106_AAARKT caicedo_h_Page_167.tif
da80de88b241a6688fb405d24d1d76db
e905b3435fe0ac122ce55f41f3bcb18e46c94ab2
F20101106_AAAQIF caicedo_h_Page_123.tif
57f8c768cf9c983a2f5b20669f1e7968
6ab4c94c581e1821d919baf5d2931b6244ce8f29
35015 F20101106_AAAQHR caicedo_h_Page_192.QC.jpg
c969cea50be1565f477d636ea4b3c916
eaa46593d00efadf359250b5c1b19acad38f7d8b
F20101106_AAARLI caicedo_h_Page_186.tif
72109e78d7ed7611cc0ddc1d696aadc8
90f54a1d1764365363f2fb8875b8b6b061abacca
F20101106_AAARKU caicedo_h_Page_170.tif
63a793944403a801d6ef0efd96b768fb
6d2a9f7857876b2bf94c72fa2d6cae16e78cb73b
8372 F20101106_AAAQIG caicedo_h_Page_155thm.jpg
a7cdc889521bfe2013f0a05bfe97da29
628027aa77dc5cf6590837581abcd9632e9e8d2b
7896 F20101106_AAAQHS caicedo_h_Page_131thm.jpg
4d965547d51a364100fa3ff665a769a2
ced1305679d16c0751f0b43724248c98dbcb9945
F20101106_AAARLJ caicedo_h_Page_187.tif
538e05785c000afab0e515b48bbe157a
e6b0aa689231ceae0deac651dc228ee47a402d00
F20101106_AAARKV caicedo_h_Page_171.tif
49bdd99ceabe9abe97c8a8dd3fc60fea
67e89d978bbc0b3df0090eccf3bea327f3dfb7d4
2186 F20101106_AAAQIH caicedo_h_Page_003.pro
ff804c68b0828c49c8e7960078d828a2
1671d581e10f251b4b0c84c66263449ded0d7bae
94179 F20101106_AAAQHT caicedo_h_Page_080.jpg
3af69a3fa6e9152744538d353cd08d22
d34c88f97c14e399bf2c0197bad0fd47c438b6ec
F20101106_AAARLK caicedo_h_Page_188.tif
a3006d0494b03e51255404cb8b6d5eeb
d8ff12c8558ae21442bb184fad0fd73715bf918d
F20101106_AAARKW caicedo_h_Page_172.tif
101e6cc2107ece67987f3307d09ed9b1
e69ed97589600a8c952b032153693350fd672392
31290 F20101106_AAAQII caicedo_h_Page_137.QC.jpg
3d507f927fa98d2e54d66e368205d925
9c4b4dd1b3bb6b5040d53f3781e6ca4a8f89424a
2210 F20101106_AAAQHU caicedo_h_Page_189.txt
aedec3aa444f458ba5eb250981a1d51a
4d80592dc6773d5dd2cf0a7d8e08f4e3fd002c68
96690 F20101106_AAARMA caicedo_h_Page_005.pro
3f87d58c3853ef13d646ae840c87ecd6
9c97549b9dd26ae1a1d89b777fe6daaef68f4f36
F20101106_AAARLL caicedo_h_Page_189.tif
4ab7891b7c60bf5f823bd6ea364f2286
3ca439d9f080a4d5c72fe52e417ba715de8101b2
F20101106_AAARKX caicedo_h_Page_173.tif
d5d12f26e42f58e1d6f0e91e7a38dcee
250bda1bc892e5538561b268ea09f826a1c90764
30831 F20101106_AAAQIJ caicedo_h_Page_182.QC.jpg
fff065ecef1c3eeb8632db679412dec2
603a8804240ae1a60829c2d459c13a24c102f548
2338 F20101106_AAAQHV caicedo_h_Page_016.txt
b19c2d1bfd74e7a0eeecbffcc8ff947a
ddef2520c408a496b9bdcc735f5b4582d63e0c76
111669 F20101106_AAARMB caicedo_h_Page_006.pro
0c117d1a7b3a0d8e99690083fc40f9f7
99fc516276716e42d6461f0b28f7a2b83e73040f
F20101106_AAARLM caicedo_h_Page_192.tif
79d27803d425c82719fd181f5ddf4ed7
2ec032fa297ab5b5f293ed762422b301326d202e
F20101106_AAARKY caicedo_h_Page_174.tif
3120ac9f83800f9141af73a15e64960d
2975c794815a7883347c57e88b631dccb431a93e
8154 F20101106_AAAQIK caicedo_h_Page_104thm.jpg
7930775f6988e35f46d7ef6dd92fd19f
ca4b24c6b8fcb5093003f90d5deb78f229fb2ab1
40219 F20101106_AAAQHW caicedo_h_Page_028.pro
29996f6195f8cfeb47d416a10d77a8a3
5a95b5c5f3cac7b9d2be4661fddeb4ea6db69e2f
26316 F20101106_AAARMC caicedo_h_Page_008.pro
522fe8719073acb995daeec8194895e0
aa7eeef3b79785ec4559425692c2840ee214f096
F20101106_AAARLN caicedo_h_Page_194.tif
a1c4398bf6d303248d80af0bf37a30ea
2807149b46fd8892b9b0853df45d122285196b94
F20101106_AAARKZ caicedo_h_Page_175.tif
18f560219b129345eed2f7c1e44aa446
a913184ed7288aa47b7cb40d126821a7c1e4d3d8
870591 F20101106_AAAQIL caicedo_h_Page_132.jp2
85b205d9b22d304737ffcfe145904009
0f1773c53fcfb1d159a1e839590fba0bd824b5e3
106449 F20101106_AAAQHX caicedo_h_Page_125.jpg
7173c13cb41f98d4e7c7231d8ef95692
b94724e02cde61c03256ad6ec5fa4222b24a84b3
2217 F20101106_AAAQJA caicedo_h_Page_079.txt
32244206fbefb4b0488c97010f366927
50bb695b71e872a172926bc17c39f95069bcc9bf
57373 F20101106_AAARMD caicedo_h_Page_009.pro
c519a832713cf49182b59f67b17efb99
64ab375b090dcdc5f39bd192c28c7b5a85ad6bcd
F20101106_AAARLO caicedo_h_Page_196.tif
b912ef9ccced79b9368442283dcefff0
7b46e33399f8c9ab9c41207feed214649d1ccc5d
98639 F20101106_AAAQIM caicedo_h_Page_164.jpg
d03ffce9a878a297553952e9329b4440
17a7f675fb762bef5fccaefb70a0395f47403f34
98518 F20101106_AAAQHY caicedo_h_Page_186.jpg
8f8e072fe79edda076c968f83390f7ac
8321c7c551dba9619d2b099164eae0c7d7b94b88
F20101106_AAAQJB caicedo_h_Page_105.tif
af508bb5bb9a9e29c9cd713caa74d54e
0c8a20eb657657e4f977c4ce7c28b33d0d6aa4a8
76320 F20101106_AAARME caicedo_h_Page_010.pro
8478cc932fb17de595b06623173d9e86
3a212ca3169f1b1d252427c04d6963518784fb5a
F20101106_AAARLP caicedo_h_Page_198.tif
7db81f65f12bb5bb142bd858df38c855
d42c41283749f6d13c21618f7f7208cb97c56c0b
35101 F20101106_AAAQIN caicedo_h_Page_052.QC.jpg
22504e77e414f506824b67cf2df3d105
d81ef963d4252f25a3e44e77cdb5297563c226e7
30875 F20101106_AAAQHZ caicedo_h_Page_026.QC.jpg
b9b011dc787488f98658acf1a2117ba4
22508f8fef3105cb573314328469642770d00753
1172 F20101106_AAAQJC caicedo_h_Page_054.txt
457b186bde6ed94b81a65af980ac6c64
85c5e8d802490254eb6877e0a1ded93c32d01d9b
F20101106_AAARLQ caicedo_h_Page_200.tif
9787e085997575dbcb746b8e66ac5c81
214050d8f2199f9a85334ec6e1451d87625aa9e8
F20101106_AAAQIO caicedo_h_Page_143.tif
db41b144c29b0c44f4a528d0d433fd98
ccd8081a0e9c5d2e8a2a598421a8d99cbc552bae
83156 F20101106_AAARMF caicedo_h_Page_011.pro
bb956acf2a5a094ec96ca373d2fac134
78d526e37150365f34e06fec9aa35171e8853a95
F20101106_AAARLR caicedo_h_Page_201.tif
5c2ae73d37c8ef33c9c72ebd0ca97967
d306019bbb980c1a91c953d48022c404b83eb730
1726 F20101106_AAAQIP caicedo_h_Page_176.txt
9c15b9ad1a502cf3016dffb1b6decd28
4147c65850eed9e4a67b0c0fa1f94c1fe9b1bd75
2481 F20101106_AAAQJD caicedo_h_Page_081.txt
8e630b0d812602edee6f3586952f28b9
78527983b231c7fdddc5985a96275e63da4aecee
83787 F20101106_AAARMG caicedo_h_Page_012.pro
5269be3fb75e998cd36a03c4504b4d6f
64a0f6198a3c899a83ce35ec1350b6cba476c1ac
F20101106_AAARLS caicedo_h_Page_204.tif
46a185a646bc6c82018a578c1dc55984
d5458e550b56e18378ce10bb17cb6fc2226cfdbd
30630 F20101106_AAAQIQ caicedo_h_Page_165.QC.jpg
4ec11ee8687283afbea559e08ceebad6
2fc82c07d5f8e27c414c6afd6c4242af741de698
2030 F20101106_AAAQJE caicedo_h_Page_101.txt
9640bb6f6a54428218782e230202913c
773742ccbe07338f03588309cc89f475140ecb30
80570 F20101106_AAARMH caicedo_h_Page_013.pro
72a7d17bc2507e86d936b9f77437005f
2c5c1c2df45c2918fc85e55c3725ef423b3e3cbf
F20101106_AAARLT caicedo_h_Page_206.tif
295c7957da1ff60b0e0a0992c4bc7275
59fc2de5a49c2e735686190c54c7caab5e58ddb6
48934 F20101106_AAAQIR caicedo_h_Page_049.pro
e7eefc7c0d629d146aa895d311b5c320
f2f8bed0a93dad754ea1a04b36306b22c824b6a2
920311 F20101106_AAAQJF caicedo_h_Page_042.jp2
25a82067b372e53dd88343c1dff7e643
1217ddc0c10127e8b92603481eb767fa3109da1f
81591 F20101106_AAARMI caicedo_h_Page_014.pro
e50c6c9e88d882ccb2be887c7a1742f0
ce6e2abb9b1a927a8d7b6f1654cd6642cad0564a
F20101106_AAARLU caicedo_h_Page_208.tif
018b81dc85546dbfcae709f405e0e2c1
964af8fecb2a185d403c9942a0ead41e46c64108
8044 F20101106_AAAQIS caicedo_h_Page_049thm.jpg
dc611933722a83729c82e3cf69d0955d
0598ee133a83dcd9d11280931d4a36d7fabf4b21
F20101106_AAAQJG caicedo_h_Page_013.tif
fcf49f779c0dd4138ab1724e8f5db1be
bf670ea5199e3c06395ade1d8161b80faa160b83
80392 F20101106_AAARMJ caicedo_h_Page_015.pro
abd5a57599b7fe2e83a0d8a07bd406d9
b1cbca07cb758cb62c8e7ff435e310550b4966c2
F20101106_AAARLV caicedo_h_Page_209.tif
a320aba112db33ff48e6df8dda30c21d
6785ffeb79c77088b2e642b1cf2be24d059c73b5
82788 F20101106_AAAQIT caicedo_h_Page_042.jpg
c9fce7246fb73f0a1d85eae77599adc4
d48f9698c1fe0ea14472aaa8f5067f3d6ef862a1
36075 F20101106_AAAQJH caicedo_h_Page_058.QC.jpg
f50a2654838cf8317b9d43cb8a8e8f50
90b75c73cc5ff5cded6e8f615b714089b8e3c5e1
56559 F20101106_AAARMK caicedo_h_Page_016.pro
9ad605e6bc2d7af0ba5da016e6c41b01
9f0639eca37778397bb05fb56d64aaa04c4169d7
F20101106_AAARLW caicedo_h_Page_212.tif
a25de987ae7ab82f1cce20ae5c2393ca
a87e2271ce633557f81a9edf334266b8281b43f3
1051935 F20101106_AAAQIU caicedo_h_Page_077.jp2
e5c6ed98cbc5be188427d4c5cc42ad5c
0b0654ea1eb4959da99899b77d585256786fd7c1
2851 F20101106_AAAQJI caicedo_h_Page_161.txt
727801d9bcb80bed6f9f51aa950deb8d
1dcb7e413af5b55ff3853451ab45cb7398b066c6
48957 F20101106_AAARNA caicedo_h_Page_034.pro
15d7701f221a4665d5d8cadb6eb03b53
b36a0de614e6389645852ca2963cdcf8f2bb43e2
48419 F20101106_AAARML caicedo_h_Page_017.pro
d37ba491a8040b1837b254a2969b2090
9150dc45edd53c6ab193e16e08ae5ba077553e3f
F20101106_AAARLX caicedo_h_Page_001.pro
c1cc5512c0e5b4ee7f96cd3e109f109e
7fe40093020f114d625de26b343ca1f307d26738
1051966 F20101106_AAAQIV caicedo_h_Page_010.jp2
11de8e9460d5a8ef442ff65b09163e0a
48e34c949323f3c930547dc3128de636cf5a693c
2121 F20101106_AAAQJJ caicedo_h_Page_173.txt
fdd734ed20b429686e06c834b1ca8c0a
734e845ea7804cbf682c775508e2fa02831df6ba
50266 F20101106_AAARNB caicedo_h_Page_035.pro
7ed97eb087dde869aef7b9554c7ee04b
e0de557ab6f2c18525b58b6f1202bef413c623be
31071 F20101106_AAARMM caicedo_h_Page_018.pro
72d81d927edbf0ee3d68603eca7e21fb
fe7bafe28fee50ec0f1257b72e0fac68a644bebc
992 F20101106_AAARLY caicedo_h_Page_002.pro
c811b7978bd458b8e80958f44ef8833c
61be789604f95eca073e4fe37d31d220a6aed8f7
3732249 F20101106_AAAQIW caicedo_h.pdf
c3771374911364c241e913de3b742c72
5d391f06abeac66bd993e43ed18d5552159f8b26
33173 F20101106_AAAQJK caicedo_h_Page_169.QC.jpg
2a6035650c034244effdbc3f614f0020
1daab8a1c6e426f348f3385cc71421767bebb616
52774 F20101106_AAARNC caicedo_h_Page_036.pro
70c0c7ebc9976c74db02b0f147432c1a
03981f11c53da459b2e7aa2135e871ee3f3703a5
53386 F20101106_AAARMN caicedo_h_Page_019.pro
996299b88d414286c5e803dd8fe07e65
626228716dc0930542236045c061430feb2e7f2d
31536 F20101106_AAARLZ caicedo_h_Page_004.pro
93fea0326c1bf1d8eda28c8ba6b984fb
48076ccd11a60b36759b452d1a846b313e03a8db
8428 F20101106_AAAQIX caicedo_h_Page_135thm.jpg
765ca433fb90c21ecc5d3fa0a34d0433
fcaf1d7689907a4f3522090e4a03c70b2759a46b
28619 F20101106_AAAQKA caicedo_h_Page_038.QC.jpg
45cb78a29f26cf58bd02d9272ccc4090
ef3ef3e7ea6828b5a4491935a839df440317b1cd
7769 F20101106_AAAQJL caicedo_h_Page_039thm.jpg
b80f048245e1e097392f3166bea77095
8530c7428290d688e13e5e95de4cb1e5fe608be3
51392 F20101106_AAARND caicedo_h_Page_037.pro
f2ad9e8afb34527361fb0de34b2c3825
f9382d805651d92ee8e46cde1fe492bcc6dfc576
54970 F20101106_AAARMO caicedo_h_Page_020.pro
3d99cff6c8906ac2e5951289fcf652c4
543adaecb384648cd4dd46c277bdedb742a05ce2
2091 F20101106_AAAQIY caicedo_h_Page_187.txt
8fe9fdbca867ffc1689a8b0531e98d7d
585f9caca4c06e598724796ac67640cfed74bddd
80017 F20101106_AAAQKB caicedo_h_Page_141.jpg
0eceea31920d5fc018d09479f499db24
91e38d30332bd74a8282e3d1905781054fac39d5
F20101106_AAAQJM caicedo_h_Page_047.tif
9905b6588d64f227d6dac7ce5d39a19d
e1284af75f9ec991db3b615ae70125815a76c1e6
40626 F20101106_AAARNE caicedo_h_Page_038.pro
c13380d9770a22eb80b55cc070751bac
f5f7904f9a23408e6704de3924062a29418c75f5
49298 F20101106_AAARMP caicedo_h_Page_021.pro
8451f511b79a99c6f79d60c2f6d12c9b
4495b9364340f3a7e53935a5a516911a919231ca
87024 F20101106_AAAQIZ caicedo_h_Page_142.jpg
6f68cc6e78deae606310bfaa9cd20260
b1ea55fce5f34c527b699888cba26b8f3d18cd69
1048388 F20101106_AAAQKC caicedo_h_Page_075.jp2
4fd03150587103cdd74b2035e2edbff0
5276ea1a2c19fb356e17d83607db89aa762a97b2
2176 F20101106_AAAQJN caicedo_h_Page_039.txt
09a835d65698f05194fcd78ccab8613b
e59af8d3c9e6cdbed67b5dfd678f4dd8a57c5678
47123 F20101106_AAARNF caicedo_h_Page_040.pro
24d4322adc6da64436cf5080b2c93e0d
844bb63f5fdd132cc735ac83dc18b2339be0eeaf
52939 F20101106_AAARMQ caicedo_h_Page_022.pro
950565eedfd41897df61c14b1490a061
10b270accfcf17900bae8b0cc4a2950da60d267f
44424 F20101106_AAAQKD caicedo_h_Page_096.pro
79a942005b7310b3f97313e5975fae2d
f75de0456d600296ce3576ad6dbe926380bfaecd
F20101106_AAAQJO caicedo_h_Page_135.tif
fe512135883ee51d26803e80f21685cc
4f45c95c5cfbafbcd57a65269ec1f836be50d55d
51286 F20101106_AAARMR caicedo_h_Page_023.pro
427be24539813b808a7c6b028c6543c0
edb9fae90cda6aab3006bc4fa0c9cd2468aca99e
32027 F20101106_AAAQJP caicedo_h_Page_080.QC.jpg
cab805e9a8fd160ff33769dc07199ffd
008080f37446bcdadd6d80fa749365af59b8b30c
53703 F20101106_AAARNG caicedo_h_Page_041.pro
9232713751e99e3d2efe0f4433854402
cd0698f0360e3646b8f68366076932df768147ec
52297 F20101106_AAARMS caicedo_h_Page_024.pro
d4b822b11fce7af03f7573028887861c
666af5c798773e7608ea196856025973841ea6c0
44607 F20101106_AAAQKE caicedo_h_Page_063.pro
25127f14c73febdfdef30a539a158d94
8f3769e65be2e2d6b3a7ee690029cf82778c7acc
100504 F20101106_AAAQJQ caicedo_h_Page_203.jpg
d12ca639e2121d1d7ddd503c15772bc3
502d0f45d4d4d0eb6ecf6cab8775b1ccf5503f45
36509 F20101106_AAARNH caicedo_h_Page_042.pro
3b77681553d1ca80a9f9803ad323f446
ff977572fcce0fb13536bb4d62f080cf0ddeb661
36819 F20101106_AAARMT caicedo_h_Page_025.pro
c6975b6ae1f5dc806fede21d3a274ef3
e6916763360fb49f660ddc9f225ade86a8f7532c
F20101106_AAAQKF caicedo_h_Page_044.tif
aeb54a79bb720069bf2e063053c75e7c
3329b44fc483de3f37036f63f6a80794ea8b555a
105107 F20101106_AAAQJR caicedo_h_Page_036.jpg
3880e1834a9c68eebb587576956536b0
c146d4faae4af2aa7ff124beb359ff16be9721e6
52932 F20101106_AAARNI caicedo_h_Page_043.pro
c63b47abb0722eac58e08f49ae69b4f6
14e459439179f090a7e3a0cf5532a45afe9fbc94
42908 F20101106_AAARMU caicedo_h_Page_026.pro
a3b59dbd9ce86933f6050fc2b720aa34
8086c7ee0944060268f0c036133f1732196f4bdb
763795 F20101106_AAAQKG caicedo_h_Page_160.jp2
12a60e5e46ec2f44f35179e7de59f118
19926c59bc8973ad87b38bd2139be93714c7696a
F20101106_AAAQJS caicedo_h_Page_126.tif
185879ce07ad6d87e28f0f4306b54563
bfb52a1ab30e39d829a1936bf48782f12c80f589
43652 F20101106_AAARNJ caicedo_h_Page_044.pro
6021bcbca717c98051eb3987deac425f
049a5c5f63b182be91bed354f8f4a1c2c889dd64
56968 F20101106_AAARMV caicedo_h_Page_027.pro
53868a31d20d93b4b5d3ddba763a78bf
ce854076e0b74caa05831bbf10fddebc81a32b79
96620 F20101106_AAAQKH caicedo_h_Page_096.jpg
0b539f818ab0c699c42048acc9b4ea94
1bf78133acecb122cd95bd52d750f8e41803e915
4013 F20101106_AAAQJT caicedo_h_Page_005.txt
971abd2173bfd9898ff29c6fdf472191
e35c749027c4da92ec727049efce24f959f241bf
53439 F20101106_AAARNK caicedo_h_Page_045.pro
ad12c1f48875b167974f62b2748e2f23
b22bb08d22f9aa7306e65fa4533d8830d2290566
43397 F20101106_AAARMW caicedo_h_Page_029.pro
84e35520db0d79f57a1302df57c85fd0
613ec1653ebcd3479d58299865661fca6e789c4e
36010 F20101106_AAAQKI caicedo_h_Page_119.QC.jpg
adff0a094fe7a1e76460c3ba4925ad88
fde1af4cfdbcb6907f29adf65515746749757026
35603 F20101106_AAAQJU caicedo_h_Page_170.pro
49f948d8dc3d0b75477b46daf9eb2a80
eaf22e3672824c81da80cb5f72022ab493708c4d
50206 F20101106_AAAROA caicedo_h_Page_074.pro
4e1741b2fb69eb03d36faecfb99cb7ea
d95353640ad13de521045e2f76fb2429b9933405
49355 F20101106_AAARNL caicedo_h_Page_046.pro
8767913f8f3f59a387bf8c02189ce9c2
93878c23c027264932915b4ad7e0f32b72a40a97
27901 F20101106_AAARMX caicedo_h_Page_030.pro
cab8b7284397aede6ac211a8417c47b4
f3745fb5cf36d8e3f7f129e0c506795c518adf75
1989 F20101106_AAAQKJ caicedo_h_Page_206.txt
591f5b777ab09eaedf883c011abfd618
09ba56584d7685094f504b83d2b27a9482ada33c
88415 F20101106_AAAQJV caicedo_h_Page_103.jpg
5f35a5ce8d37a2e70b275c775060d4ea
d709a05bdf793f4b6338e41e24708848bc4e4220
40379 F20101106_AAAROB caicedo_h_Page_075.pro
c4ee0e52c9333bc2b6d3c33e2f2435f4
e0078af31efdf8d767ade87b821cb6af15e4a0b3
48051 F20101106_AAARNM caicedo_h_Page_047.pro
b0427bf7f4074ae5e8eb0d308d9ed75a
48fdb4afb842d45dcdb04a88e36a621019e240a5
56233 F20101106_AAARMY caicedo_h_Page_031.pro
dae3d08732e3ad82d9f74a88fabdd9d0
032bc5d55b5cd6a484a9086c28a83abac522c86b
32580 F20101106_AAAQKK caicedo_h_Page_205.QC.jpg
3a1c754d1219c7f382a7e04c1c83dd56
2cd349e6ec508a30aa0d0bd6dfc622da5a32cf78
34560 F20101106_AAAQJW caicedo_h_Page_024.QC.jpg
69327bbebc1c84a11b99ab7bd05c7884
91c768b2a147d80c0086f23fd15544df4dd55bce
41395 F20101106_AAAROC caicedo_h_Page_076.pro
3e08f8d024e0bee2b0309c9e284d7927
abfb8860d87a51aff1f1cca2c89829ec9b95a01a
48179 F20101106_AAARNN caicedo_h_Page_048.pro
7282072c54f03c11927cbf115bc3f2ff
38fbe057b6c8131f5baa4404ca7165de9a909df2
56232 F20101106_AAARMZ caicedo_h_Page_033.pro
2f5ea64851b4f4cd0a87c409bc7fe93e
7f15fe0b5cb0295ce4d4b28f0ddcd4b46139bde3
94927 F20101106_AAAQLA caicedo_h_Page_040.jpg
d3bd736588bf4c766116c0e1e3955e05
6d0f004bde85b993490a36e51019c08678578c7c
F20101106_AAAQKL caicedo_h_Page_043.jp2
f69e0035b4c0a38c20ba0c138f797acc
105ce22fe51105e523f15040b36f9b7cb6ec11b9
F20101106_AAAQJX caicedo_h_Page_106.tif
adeac69d718f5ae29bfb1b05a1c15cc4
ca2dc6dfc87651f6e5022ad5e3dd78205479f47c
54370 F20101106_AAAROD caicedo_h_Page_077.pro
6a39c2c2d588067d8d8315d5fb0745fd
6b0cfe96bcd7e565b107f0a38e23e5a14c9db3a4
54348 F20101106_AAARNO caicedo_h_Page_053.pro
d444e7f72f39852b566e7ac7e3d3a8b2
90c19e4802b2a7ca1a33ed37ff1df1d5f83db5f9
33177 F20101106_AAAQLB caicedo_h_Page_202.QC.jpg
3ab0eae210f2d8440ea76905db8ac221
6d888ca7fb9b44c502ce04026640c31d1d38fa85
7498 F20101106_AAAQKM caicedo_h_Page_132thm.jpg
ad30299cf33cf6003037c82b3b075337
771516a4923571544ccba7dbb5cd85ad9ab894bd
F20101106_AAAQJY caicedo_h_Page_182thm.jpg
446c6335fd54633e92cf35b6f9999821
7c7d7f7dcb37f2ff630c30ff92692107d0db4fcd
46414 F20101106_AAAROE caicedo_h_Page_078.pro
344d5118f3eef42661b0505c0016b7ce
faf2ad0a25775e4eab9baf70633c93c2a379cf81
48993 F20101106_AAARNP caicedo_h_Page_056.pro
b1801ab8192d95fb4ba536d06af319ce
614f0bf8407ed6e93baf5aa2007feaec13350a4c
85474 F20101106_AAAQLC caicedo_h_Page_102.jpg
58ee1f655c66443aeed471f0b05dd5be
ec44f286b67958fe0289f65b7e0c8f41cf835bd3
71640 F20101106_AAAQKN caicedo_h_Page_193.jpg
4612bcfe4492cad1106858e109c80515
f5d799c71d6fd314bd6be4b58c91597c0488e577
48161 F20101106_AAAQJZ caicedo_h_Page_177.pro
eaf9370f1f98da16c33b7b813ec8d2ab
1640f926ba2611f02379bc4c378a2a870e96606d
43879 F20101106_AAAROF caicedo_h_Page_081.pro
6bfc93abd5d03a864709f9981a122d1d
42854b44be7b1d98b253debf9efb372dd757edf8
38641 F20101106_AAARNQ caicedo_h_Page_059.pro
036c515b1a8a65d322e3cc20b1fa5483
73b22d2e43a75c565cdb3f2736bb62d85e7fe52d
101900 F20101106_AAAQLD caicedo_h_Page_089.jpg
2974b94f71f7708f0362ac7e5149e943
0471794045ae6043ba6ea169d945b13c996b856e
8305 F20101106_AAAQKO caicedo_h_Page_112thm.jpg
ce1a0dbc41bcfab04b4f1513ea286303
da25b585dc03f4cabde148396c325e68fc9d8061
48749 F20101106_AAAROG caicedo_h_Page_082.pro
4e313e7fe6602f5a5f16619e91b15689
331ab44bf4f9000091a25e7564c35c5dcab6ff2c
51108 F20101106_AAARNR caicedo_h_Page_061.pro
d4de364402e4b0310f56f8bc98dd30d4
3728365fc99b47f660b09db1cc1c98129c021dc5
F20101106_AAAQLE caicedo_h_Page_057.jp2
c23600131d7f010e8ec8b78d91a5ccd9
fb9b7804ca9b6fe2c85e5600be5edc77be934bbe
1927 F20101106_AAAQKP caicedo_h_Page_139.txt
902b62bec6fe6e54643fc6767040bc8f
401c1330ec06ad8b132917ef33f9f3a1fbc9e025
54880 F20101106_AAARNS caicedo_h_Page_062.pro
c582efbd18430cbcb8aca16083e6ee18
e38f1b4175c59cc7c85a7623811d300ca4a36b80
2027 F20101106_AAAQKQ caicedo_h_Page_023.txt
5820c980ffbf2d98e9076f0b244fb280
145a57bd74387c9a68c1dd436218a090aa0459d8
56599 F20101106_AAAROH caicedo_h_Page_083.pro
73223afa56375f4a190f5f409cb51ffb
3e80b3bd40e9daf36b2efc764a74dadd93e13655
49670 F20101106_AAARNT caicedo_h_Page_064.pro
12c3c7b54b86d9689118382804bd4973
a57275d004c12d87cbe91c2745ab165c82872732
1873 F20101106_AAAQLF caicedo_h_Page_141.txt
4727b0a8c1684743d289ce765c968c89
ec43c7c9fc86891d6ad88e8977f03df06b7b34ab
50562 F20101106_AAAQKR caicedo_h_Page_154.pro
7b796f7383604ef5114e9282ab4ea04f
0f2aeae49af767eaf0ed432450e206d14994c058
55186 F20101106_AAAROI caicedo_h_Page_084.pro
02d03f7c0519ff503b89456acd69e363
47d51749f64b8db42389adce8c97555c4c802bf7
39461 F20101106_AAARNU caicedo_h_Page_065.pro
69948e83f3c7be97a692c92ac8a6c534
613cd32ed95ae964c511a421f421842beb676175
7877 F20101106_AAAQLG caicedo_h_Page_035thm.jpg
2a5f1d060b7b87707c630e9158321377
6027228729234c2179a67e30089d728e2ffe5e3e
52113 F20101106_AAAQKS caicedo_h_Page_060.pro
b0b0318072e0c9c7b85e1cea40400681
eecd52d714a3fed0b95bf13d6c9715b8ef8088dd
55748 F20101106_AAAROJ caicedo_h_Page_085.pro
c733bdc8b9cd5f416b10d1d9b11d0306
ee3b313848a6574e5e7adc4991d15362b0181ac1
47946 F20101106_AAARNV caicedo_h_Page_067.pro
4a665c656928426af601bdfcff137479
2d26290767a7174c60ebb1073b4bbf28af13138d
8734 F20101106_AAAQLH caicedo_h_Page_116thm.jpg
5765f6493bf3ea917697c643338f5fff
e2789e87a02670b18e29d1481d9de9931c52f125
28446 F20101106_AAAQKT caicedo_h_Page_005.QC.jpg
1f255333fffd6baedfa8dbaccc04ff52
cfd460b3e77f53cc0d789800717a9c59b4b51559
30187 F20101106_AAAROK caicedo_h_Page_086.pro
6f1437d0c2cdfae99406052af1a6bfe6
66abd3d2eabe5b0d6690f0d0191da330638cd25e
41492 F20101106_AAARNW caicedo_h_Page_069.pro
4b6a01f14f32341ef94edd1e4bbfcfb7
fc4d3df44f7c3e5b2697ddbdcdab2dd33f1762e9
F20101106_AAAQLI caicedo_h_Page_019.jp2
b936f6defea8854f4dabdf608df1f9e0
ab851700124abd435456a4cb7784104c6c91225e
30222 F20101106_AAAQKU caicedo_h_Page_044.QC.jpg
f6631e297610967eb34adbbd682e2ec1
d3ba59423aadfe9628f55c7c74f532493a291162
42151 F20101106_AAARPA caicedo_h_Page_106.pro
530d91233012e4d0529e454631bc9790
9443273227466dc6ac28ac60ae12ea868dfc908c
46773 F20101106_AAAROL caicedo_h_Page_087.pro
0ed529aa78550db5ac1e1079e3eb0fd7
f105e55eda89df4894d1f7c0d6e78d50ccaa127b
42952 F20101106_AAARNX caicedo_h_Page_070.pro
86702548a07ed28b17a45fb6be197a15
891f7d571e6de333d254c8e1463eb4761f5cc521
2058 F20101106_AAAQLJ caicedo_h_Page_175.txt
e3e92d04db56240728c512948a81047e
db1854ad5cd1a9f7a18ca8df1de45582fd460dd9
28368 F20101106_AAAQKV caicedo_h_Page_108.QC.jpg
c6e8b0cc1722e9f1c844007e823f02a3
aa17b35c24a43dcb85105eecdd4fba9949ac4f26
38856 F20101106_AAARPB caicedo_h_Page_108.pro
cd87881e87272221fdaa32e7ad6ac827
cb31ab1a0ad597b3ca992503239b483e9c7df907
46968 F20101106_AAAROM caicedo_h_Page_088.pro
4e5f9303336f1f0dfac9a2358de59131
876f813d36e4d017e1b9afed1dc8bd904669eefb
40484 F20101106_AAARNY caicedo_h_Page_071.pro
245e007f40e35cc9e5907e65c2122564
eb03bab5eaf3b991039967a546de90c46d7a18c9
54041 F20101106_AAAQLK caicedo_h_Page_098.pro
ea95cc35f5351237d6ebffa1eabcdc10
a4cf2f9b035acb809b46f462896f5bb61a993c59
50599 F20101106_AAAQKW caicedo_h_Page_156.pro
2fba83120155a5bca4c94de54419584c
ca8781de3d361a850748863fad748ff6f53dc061
45276 F20101106_AAARPC caicedo_h_Page_109.pro
3daf1f43881abeabaddea6ecfe2a8914
a198437772039c0d897f6dbc2d87b6c25c19883c
51022 F20101106_AAARON caicedo_h_Page_089.pro
aac3231675084a7cfcfa7d9430b52311
7a656daebd3f167991e2f428aac24db31e633489
43498 F20101106_AAARNZ caicedo_h_Page_073.pro
5070856a230fa9eb84eb7d88b341e7eb
faf9427612516419281e0d32708c1663fd1eb070
8047 F20101106_AAAQLL caicedo_h_Page_210thm.jpg
87468e3c1243faa39ababea09d9a2299
ab77d421ed7e97ae5a41d4e76b7b0c55801e569f
7213 F20101106_AAAQKX caicedo_h_Page_166thm.jpg
c8a7254f21e72b189d5ccec8dbebccb0
0d04b073715d4d4cfe9f014f346b25032033196d
F20101106_AAAQMA caicedo_h_Page_097.tif
e5912c2faf6c41a414421b8b94aa4931
e6162ca42bb515c47608bc76ae1842b9e7fce154
55221 F20101106_AAARPD caicedo_h_Page_110.pro
fc415a3f041d59b68710b87002fa96b0
d17a007097f631388987a5ec5e1de5c5a463f207
43845 F20101106_AAAROO caicedo_h_Page_090.pro
3c220ed5b7ebe4584d3c9cb597093c46
019dc57beb08f6f385917e835c41626f13c586de
F20101106_AAAQLM caicedo_h_Page_055.tif
e4c020d49186f9440f347cf1087b4c6d
91dd9569df233c25121c9883f2d3bff8f761dd28
33829 F20101106_AAAQKY caicedo_h_Page_105.QC.jpg
4aef2e766434f1317b68891d9a4f7dd8
2068c8a56f79b18f145c383aed6c5c5a018b3404
932198 F20101106_AAAQMB caicedo_h_Page_071.jp2
693bf8c52a416d1d1153b59f5f72eda5
22133261f1dfe9199aaccf1974802132a87aca24
42946 F20101106_AAARPE caicedo_h_Page_111.pro
82f706ca0c5c52629dbaea43d80fcb14
70050d3f75bde271780ae02ecc1858bb0841bbaa
47326 F20101106_AAAROP caicedo_h_Page_091.pro
98982180007883164f818ab715b91444
4fb7680d95d04ea22c0d8c4d80bb7f3df28b11ea
52495 F20101106_AAAQLN caicedo_h_Page_113.pro
223394a8a82a93d131a33096cba41a78
e94155d927eb7871e5d2e53b729b4d67f359ad18
F20101106_AAAQKZ caicedo_h_Page_119.tif
7dd76b781bf71ae2d3d3f58ac0439548
5b2bd64aa58ca6adb5e5a67b7a6a931e34447396
36536 F20101106_AAAQMC caicedo_h_Page_020.QC.jpg
558c18d93c0fff1481eb9f2324d7e7ad
e457e703aabf33a402c52ea36ebd4651e3ffa11b
54274 F20101106_AAARPF caicedo_h_Page_112.pro
1c2e0c946a563f8a23c565e2d60e6957
21802d5ba7f2bec2b867d8c13ba7e4493aab814c
44748 F20101106_AAAROQ caicedo_h_Page_092.pro
276f3542172874b56a6d57a2f8584913
1218d4981bd103f62f90b682a883ce1cda731b29
28659 F20101106_AAAQLO caicedo_h_Page_090.QC.jpg
82c03892c6c79e6bac6b05eabae580ae
34d44cfbbd06e1332b66117e70dff56ee493de37
54771 F20101106_AAAQMD caicedo_h_Page_003.jp2
fdede6b8683a7c975947225df0879145
4cea734c809f07d5867cd7478c0e1cfdea7746dc
45278 F20101106_AAARPG caicedo_h_Page_114.pro
11c6ecaf5200516cb347d4368c168872
a8e14195c88a66857a7e32fb04c2c6c7a027082e
46318 F20101106_AAAROR caicedo_h_Page_093.pro
20e674d2efb41e1ec04be4f18b7f60b6
c2da0adb07f84aaab00aa5022232c17d0ee6dc37
36093 F20101106_AAAQLP caicedo_h_Page_063.QC.jpg
319bb0e0a4d98e522ebae9a8ea9a3f05
1ad1916f43b89fe19909d911aa301e5703858d59
8435 F20101106_AAAQME caicedo_h_Page_169thm.jpg
2b033aecbd6211df23637b0dd2d71af5
d3e0ec3afa463a35728b9927e182a00257066465
52341 F20101106_AAARPH caicedo_h_Page_115.pro
19ae0ed3a55317ba53036329fbe51e6d
7567ec7ff71cbb48c01a78069455e6518b3f568b
45516 F20101106_AAAROS caicedo_h_Page_094.pro
8b28a28e219e8faa0dd9d52edf2b6166
d2d0d3e36057ba10160b73456f23795a304b9f2f
4985 F20101106_AAAQLQ caicedo_h_Page_086thm.jpg
cb7a48aa8c9d5225417eace9d2fdf7fa
2f0ae13627c7e35cc37010373bd3a65d323adc85
F20101106_AAAQMF caicedo_h_Page_129.tif
e247c55a3b2569d2ea9f3613af1b59c5
3ed7b9d7f39e631f55d1b316d90ceee90e947766
47829 F20101106_AAAROT caicedo_h_Page_095.pro
59544a4a4c5e9cd3563e7b43ab6cca3c
02a077816c309cc261c2a00b7acba403c631f2cb
F20101106_AAAQLR caicedo_h_Page_046.tif
164d0013e1956fcd1f38118fbff80335
244de3e8f04073a5ee00b1dd2c8bedf7710fa0ec
53063 F20101106_AAARPI caicedo_h_Page_117.pro
41095c7eaaeb27ff6b0d7572fc44d36b
ed5872c8b1d3fbb042d5345469cad1094bbfb25b
39439 F20101106_AAAROU caicedo_h_Page_097.pro
087d74842e5c71e430f05419959e3e86
655eac220ff1fa174830c0a0f1d94c08f03d389d
1795 F20101106_AAAQLS caicedo_h_Page_063.txt
88da9230ba117b5ac6236cf725e3f7b0
d44eb589d5dbe891d719d524ac133f18e3b68ba7
7471 F20101106_AAAQMG caicedo_h_Page_038thm.jpg
1c067090f998cb68a8eaa39cd0d25c09
84918f3d62ffa3baefc56803d4491861bb8fbb02
52252 F20101106_AAARPJ caicedo_h_Page_118.pro
88a4d054cacceb1a76bf2f4b60308e73
83f42cdff9bc7f9da6bd4dd7fabe8a8b4cb9b15f
53363 F20101106_AAAROV caicedo_h_Page_099.pro
93fe3ae6c2dc457060c4318d8fadc328
f40d5760cdecaee8f2d64b7cdd3d05fda2bec5a9
8307 F20101106_AAAQLT caicedo_h_Page_107thm.jpg
97f98ebc1db1c819d770dccbaa4623bf
bdfa7bf27327a5185d9523426b5148daa07bded1
1639 F20101106_AAAQMH caicedo_h_Page_160.txt
55a5ff78cfd7c84d59c4aab9796ed359
3981c7673e69dfed46268dac17c3f4b54651483a
43404 F20101106_AAARPK caicedo_h_Page_121.pro
a1ac4a468abf9c30fda2e99a15698ee6
ca23d2366124cce8464664d707c6751666cd5c81
44032 F20101106_AAAROW caicedo_h_Page_101.pro
f01eb9d9bb5e0e657aa872711ea00d83
e94d46a8b06f9fec5a349c4b0e402f76f4194e61
30678 F20101106_AAAQLU caicedo_h_Page_016.QC.jpg
8e004dbda7552130cbfb4e55603e5548
fe0ffd2a6b951a97384e7c2d2b6c09db6280f3db
30241 F20101106_AAAQMI caicedo_h_Page_109.QC.jpg
1af4d941d10230e9a83b8e0808269e38
18f670b0dc5697ef88e5f4efc088d52b5f9c1bcf
36858 F20101106_AAARQA caicedo_h_Page_146.pro
849bdcdfab21e65301b2047cde5a7f9f
5af36c88248f606a31fdc09cf6d108e0392417d6
51801 F20101106_AAARPL caicedo_h_Page_122.pro
0418521f788f38227a55e201f4e6309a
0889f313c24fb62c3b42b8327ab664b4edf3b189
40515 F20101106_AAAROX caicedo_h_Page_102.pro
c259daefc296121082118eee5c4178e5
330f9f966d3dd3377e33cc24462647f5cc9ca5c3
107406 F20101106_AAAQLV caicedo_h_Page_045.jpg
5b8a1fbfdada2e6afbec5e350360313f
f5e799ce82b519555eccbae0a48fe0493e775894
F20101106_AAAQMJ caicedo_h_Page_065.tif
ecd076b08b2f9d801de99ea36072e7eb
f1e72da6a1946f1b88b203aa981d7fe6a5c8a221
31483 F20101106_AAARQB caicedo_h_Page_147.pro
993a70202abd70a2b37eb4434d0b0e2d
70f17b87892e5b58c67295bdd937a3bee611a9c7
52671 F20101106_AAARPM caicedo_h_Page_124.pro
7e0da89d0c95fcd417109a971135d553
87c7b99b38662bb6541d111767dc9ff32ffa3c42
38613 F20101106_AAAROY caicedo_h_Page_103.pro
a531e6028b9182ec89d64de32eb70247
2cd7c953dfe9c5ed02678685aaeb3310550423ed
35241 F20101106_AAAQMK caicedo_h_Page_060.QC.jpg
4d0b86defeb51376f01bfd8b60725567
84c0b5d83a414683071041856d7c563d6becccb2
F20101106_AAAQLW caicedo_h_Page_075.tif
593544cd0eff96ba4a26dc5747763cf8
b6aefad5e5de0e3e959c61622ab9ced7e402d3c7
41271 F20101106_AAARQC caicedo_h_Page_148.pro
0b2dddab37ab02f82ca26520dc15d3a3
b8e63d927e0216e206d80c12805057c9235a3e60
53158 F20101106_AAARPN caicedo_h_Page_127.pro
fbd8daba16286d36a868edc8525508e7
2731aad719299e1da536bca3a59b83ce487faf97
44383 F20101106_AAAROZ caicedo_h_Page_104.pro
863dc2ae75fe9a189a873417e33f87c7
aa1cbb87c7a5194b01b8901484811af476848cba
726876 F20101106_AAAQNA caicedo_h_Page_008.jp2
f891634e3f6be9da0e40fa889169d972
6f0094c97fe8b9adcb0b31fc870da042d067fb68
F20101106_AAAQML caicedo_h_Page_009.tif
e64bdc34846c607016a9f55a572d0642
5577a5b1a5aa315db41b72328a90737c4328482c
50944 F20101106_AAAQLX caicedo_h_Page_107.pro
c17bce69b0cf372a57115e33e1e87114
7ce1631324f00b1e9d1c02c261dea4e3e4ff17b7
47178 F20101106_AAARQD caicedo_h_Page_149.pro
0701056ded9bfb50d6f3da460f157127
070efd3e37169a40ef601c655dd41263360f592b
38017 F20101106_AAARPO caicedo_h_Page_129.pro
4c10aa56f2b34a8e8ea1c7aa2142de98
ba2672b4d689a71a80919ca6081f6296fb497191
F20101106_AAAQNB caicedo_h_Page_049.tif
cdb4f657e68de68d3b6ae8ee50402fae
a53dd186443f9efc9fff63883c8be6981f5fbf25
104873 F20101106_AAAQMM caicedo_h_Page_113.jpg
fb9eb1e5ded74bfefcefb3cb0dcc1f5f
dcf6e67b524def500b67e0fc3a2986bdcbca9d79
2191 F20101106_AAAQLY caicedo_h_Page_066.txt
7d692c251117bff347766a6a8143932f
ebbb4265bb40042a4b2d4a9eb9101fe6d196fb71
36171 F20101106_AAARQE caicedo_h_Page_150.pro
471d014a2d57827d98bd0452247ead3e
483784dbcee7c9706b5c1b3724cb2dbf1a11c81e
41134 F20101106_AAARPP caicedo_h_Page_130.pro
0ae19c7a9d8132b602ce03cb3459de5c
d0c478532126060989b1cc83ad32d34c989228c7
1051943 F20101106_AAAQNC caicedo_h_Page_180.jp2
ada16566fb767ce4e485bcf17fc160db
5b62031b8e2bb44319ded48cbda625a2dc315f51
10063 F20101106_AAAQMN caicedo_h_Page_008.QC.jpg
736a23fd1d9cc1ffdbdbed28cdf8a47b
f8bdbc1a33a2700c9a08f4b11bf07ad40341abbf
109182 F20101106_AAAQLZ caicedo_h_Page_161.jpg
e19a30ec8bfdd0ad240ede8ca7a77453
8d70353ee23c21e4136b7cc5aa8dfa62c43756ec
F20101106_AAARQF caicedo_h_Page_151.pro
aee50579f2403668a1c2cb1f667ad35d
40c27da4d1f78f334903597e600c8611c02adb1f
43320 F20101106_AAARPQ caicedo_h_Page_131.pro
78a7ce9b7878c65d74be9127088fcd35
09a5f46940b2d94130432076467e3d0b6dae5459
2211 F20101106_AAAQND caicedo_h_Page_056.txt
98e28e5b032cb75720546b7f53544b45
629e5f8137d7f64e9210b9b04ce3bb1a1aeefec2
F20101106_AAAQMO caicedo_h_Page_041.tif
dcccb5aca38cfa0c7306cd2cd3c614ed
ddd50b9a98f85a56963aff4efe6f6e65be69ec4d
51921 F20101106_AAARQG caicedo_h_Page_152.pro
3c7608dd0cbad4ddd011e39c80c185b2
a0dce2b14f915b241055192077b128872f6e0839
39642 F20101106_AAARPR caicedo_h_Page_132.pro
514fba515e219778692c9f6d1e7d7d6c
5d5a7dd776fa55a01250b3349ddc765ba5f722a4
926274 F20101106_AAAQNE caicedo_h_Page_130.jp2
afe95ec63bfb27a0f6f8d3bce4175654
534fb1ba8eb1b5e20005dca3d71636f929035c35
925179 F20101106_AAAQMP caicedo_h_Page_148.jp2
a2bf59ee9652aa2de09290bafa76bd6c
b0a6c5d25fec0efccafef7bc28c8fde995dae81e
50393 F20101106_AAARQH caicedo_h_Page_155.pro
7b3e9febb8a91614001857768b0ba3b0
b322a8feb61b9e876cabeda3d487cacb96ea8ecd
44474 F20101106_AAARPS caicedo_h_Page_134.pro
efc0cb2dfb93ccecc97121b5989ae7c0
421f4722ddfc2bf7e320f14b5de6739a5df7f60f
25006 F20101106_AAAQNF caicedo_h_Page_160.QC.jpg
ce319b9c0d0aecf82d40f85e82afdb85
a2b6a337d5d3bc9160a5921747321d49766be4a1
98548 F20101106_AAAQMQ caicedo_h_Page_211.jpg
fff1a73e5b48d39b62e9e2fd2f6f4526
8eb3fdcc5ec128bfec5ce9a6bdd0e5615803407a
41379 F20101106_AAARQI caicedo_h_Page_157.pro
a2349fcf3d46945527f3796950768a05
d2a23e7a03692b2c0cf1a0f37f9becf94585de0f
51202 F20101106_AAARPT caicedo_h_Page_135.pro
f4a89d3a7744ec4688d6b31eaaa9bdc2
3318d718c616a4e860030bba14b7814e2ee6e250
34592 F20101106_AAAQNG caicedo_h_Page_099.QC.jpg
6a15fc0d66aa8cb5c11d73124ca94cc1
cbd7045241d21b7136c07753095087837a3dbb99
F20101106_AAAQMR caicedo_h_Page_147.tif
4a92da8516e61e91d76ef240ba2a2710
6339009bf44db12745f531dfba98f53c26762761
34831 F20101106_AAARPU caicedo_h_Page_136.pro
c5e3a12216e8d3c16af16d0bb89d6a73
4ea06d025baae61bee6dba73080183f17c897838
29251 F20101106_AAAQMS caicedo_h_Page_054.pro
1ea8fad654a108c33041b58fb0b9b04a
8a3e0dd00879f05166d8842ae8101ed1dcd93320
41431 F20101106_AAARQJ caicedo_h_Page_158.pro
295cb3fb4fe417c90d4a298e820ce781
3b74da3f96cbafc4777a1f8fd4fef7366f4dcf03
45106 F20101106_AAARPV caicedo_h_Page_137.pro
a99fae24aede527465d7da2a6da886f5
6fdc22f7861fd1c9b0d0a1937ab091498e524ad8
F20101106_AAAQNH caicedo_h_Page_205.tif
66022cc72d4a67a757b98faea254410d
4f7d51fbbc0261e27b320785391c9d1e4934385d
26232 F20101106_AAAQMT caicedo_h_Page_136.QC.jpg
10b180fc6784a434254e51605632fa80
fc15b29696f888b146bb2f8298f5505db472471c
59402 F20101106_AAARQK caicedo_h_Page_161.pro
1c6f524ad07f9f540903becef3663bd3
e1b211d441f51b0e836d7244a0f1faa98d24f437
46144 F20101106_AAARPW caicedo_h_Page_138.pro
7ba80d0ca2f22f024277cd6d40a40fd7
91edca65dea8a072b7d0b995b94895f47fdf9081
45175 F20101106_AAAQNI caicedo_h_Page_187.pro
2ab27d025d7c73a881fae943563d780e
89dc963b2ed1501c8864a4bad5734c44884f70b9
92603 F20101106_AAAQMU caicedo_h_Page_093.jpg
c47ab70aba43c574a5cd0492aaed371f
c64b2fd1325b3b7e36f25b10cab612095b6cd0f4
54369 F20101106_AAARRA caicedo_h_Page_189.pro
52f0aa0b34b9639ae4ca015a79ea077d
a8afe50b81ef7e6e813d48668ac546b6d8f2a547
55410 F20101106_AAARQL caicedo_h_Page_162.pro
db094116694b4ec33c7aa9b8cdb0fbc3
c4825e766a38062f191db186811627a46795b094
39428 F20101106_AAARPX caicedo_h_Page_142.pro
dc7432c6fec4472f8a85f30cb4ecb9ab
290d9e22a1587a775f3177856112a347c632be03
1690 F20101106_AAAQNJ caicedo_h_Page_172.txt
4bb4ded6c10b59669f620cd4c6406141
32ff728570fe1d2690fd531d3f35bc33bfb83a43
34493 F20101106_AAAQMV caicedo_h_Page_118.QC.jpg
21e88867079d9d3bf061a3595c265ace
e452eed6b66850db44dca475998c0996591c2d78
55521 F20101106_AAARRB caicedo_h_Page_191.pro
7cde414b65bbeaf14cabfe1cde20ac44
7d6738c265388dae8ed7c694b8d98d71a9518325
56174 F20101106_AAARQM caicedo_h_Page_163.pro
c236c9a4a8eecf744a3f8c60d72a10c0
155938b52ff870920a870feaa9157a817f904d37
28215 F20101106_AAARPY caicedo_h_Page_144.pro
8d55eebcb5de31dbfddaa405812213dd
a48b14def6cffe620ca1425576aaa0c19a11b6e9
41449 F20101106_AAAQNK caicedo_h_Page_179.pro
b79560c4e63ad83f5f1daa6e5ec5b0a7
7d2a0dff0c5c6c37b8b0275226642585f7a0fd1f
6867 F20101106_AAAQMW caicedo_h_Page_160thm.jpg
8e398d0fd70b68857b00f072a45639f3
cb0550041d8150fc90befc394a528fef9c7ee8e4
53150 F20101106_AAARRC caicedo_h_Page_192.pro
845db701a27d038b77e16d1cc0a278a9
49efa28b945ad585c69f836ab04038423c233b3b
48387 F20101106_AAARQN caicedo_h_Page_164.pro
73638b215bc70d3d69ccacef2fdc40a1
549fe5a656c15071e7c3f61fc2c849a7be481f6d
40701 F20101106_AAARPZ caicedo_h_Page_145.pro
dd9e10c4f969c7f340d0053d8b756cbb
ff8d5f6564069fb44bab21f97b338379f418625a
36423 F20101106_AAAQNL caicedo_h_Page_191.QC.jpg
fbd2c71df7164ea20768fc0da05396dd
8ac4b91b925524d147d15b5cc49aa1036f3cd889
7790 F20101106_AAAQMX caicedo_h_Page_139thm.jpg
a508b2ae7050c75ad9a804f13cf87cf8
d6ff017a400db2ecad1a0ed8a8239b4b69702ec4
F20101106_AAAQOA caicedo_h_Page_137.tif
6d332bf1e0882ffb12aba9b8dfd96a62
a585ba5de59a67c0e96ed3c57581420b51e9898e
33764 F20101106_AAARRD caicedo_h_Page_193.pro
7023f852d792d8c0ca51423d16ceb831
5807c55e1a6eaa980b65dbbef022d05aa6a52e2b
44292 F20101106_AAARQO caicedo_h_Page_165.pro
3efb87337fd47d1882dcb8a62fc70fd3
e1b963a2ad537effc36a9893a2f487b803e8b2af
9216 F20101106_AAAQNM caicedo_h_Page_062thm.jpg
61b0e8fd4ebd0cc4d12205c103f608b8
54a5af2e0d5a9572870a9d192d6d6d4e36d8ca16
7452 F20101106_AAAQMY caicedo_h_Page_028thm.jpg
f726d8afbee4226ba5bffc20991b7f34
1686bca80b284c40ba9037d14cadcfe335cd0280
32227 F20101106_AAAQOB caicedo_h_Page_185.QC.jpg
8363648a2a563720e960a0ced526b978
6f53d00645c948f9eb56cf44c668b5d1841b739b
47589 F20101106_AAARRE caicedo_h_Page_194.pro
dd9f8815f4fdeb42038a08cffba0ba8a
b947af42a04c8a026df4c3bc9acd009e83d3eb8c
38643 F20101106_AAARQP caicedo_h_Page_168.pro
618fa241650be897e2599eca2e16a444
755694d7f333b3f1ef20aceb880a45c09b091409
47912 F20101106_AAAQNN caicedo_h_Page_128.pro
ec286f39f1cc7ab69bf65b7cf46bba3b
06bd1aada9ee2f02670a1664fa596f2026ce7306
28073 F20101106_AAAQMZ caicedo_h_Page_167.QC.jpg
fbe9418cd8a090bc894140459948396e
6a295e7dd01575cb478be2c14a3abf4377668d60
34001 F20101106_AAAQOC caicedo_h_Page_074.QC.jpg
ee606fe368dafbb588d71be4b40ac99a
9c2cbec8b492c661e22fdb83e5d6d7556cfabe67
44927 F20101106_AAARRF caicedo_h_Page_195.pro
fd174ad4dfb8e6edb74e089e60dbed86
6c2cc903b35e07ca261e31c59f588eff9a22ac7f
54671 F20101106_AAARQQ caicedo_h_Page_169.pro
db8e951a8e8f638aed4b1dc8b64bd949
bdc53088b9eb7dee915942b77a3a6c31807c0153
51379 F20101106_AAAQNO caicedo_h_Page_052.pro
821411842143ab724276092cb0c4ff28
3ec4b15c3d947a964f81b351d107f7740563f6af
103995 F20101106_AAAQOD caicedo_h_Page_154.jpg
506c23428ba0923a9ec1b10e0e34f818
a0fe513459557b70acec25c8e5e958784c3a46ee
51364 F20101106_AAARRG caicedo_h_Page_196.pro
fead9a44d911c2b0ecbe78e66dce40e3
2e522295650fe966d8dcddc026fb75a7d73867ac
37884 F20101106_AAARQR caicedo_h_Page_172.pro
46305f512daeb791ea4276d3cab7bd56
dc65e096fbb56ac892c379532a32ac814686f3ab
2082 F20101106_AAAQNP caicedo_h_Page_091.txt
f774315d49b968a4a97f66a203c2f438
191cef6f63595107108bf929c26c496ece41e2b2
87009 F20101106_AAAQOE caicedo_h_Page_047.jpg
307688a90059ccce992bf1c28f89e0f1
40936e36e589f4604163a2e71f89a2faecba00f1
47337 F20101106_AAARRH caicedo_h_Page_197.pro
5854419a93c560020d2563b30663936a
3dcb751b2636b0e0c856448ad6df567bc1eeab91
53095 F20101106_AAARQS caicedo_h_Page_173.pro
a12184f8955154d4ec6f4f715d0a21fe
645ccac0d02cd57312d742d1c0f20863bdaa4334
53039 F20101106_AAAQNQ caicedo_h_Page_190.pro
2fff2832ace9d5b9879089f970ee65f7
217375c8e14eb411f0b1d751e2bc89bdd88ac34a
111631 F20101106_AAAQOF caicedo_h_Page_033.jpg
36fcd9265ccdab7ded967296552e88be
30e1c4cabba47a64429590f336f5fe17c8650db7
45797 F20101106_AAARRI caicedo_h_Page_199.pro
f7676ff5981c93f979d150b7414350fc
09dcad88a05b48ddf6cb410acf8c8e4c20c2044a
53208 F20101106_AAARQT caicedo_h_Page_174.pro
e5e3d8caba62a3a0133e4c84bb9bf555
1047c579501db34b0c74eb801f05b078449ef9fe
51896 F20101106_AAAQNR caicedo_h_Page_202.pro
1db1674c4c472fbbe46a7bae8bf0531a
2179340611acfdcd42ff3c0e6934d400b0dc287a
35775 F20101106_AAAQOG caicedo_h_Page_163.QC.jpg
abc4bdc2915a3d7a6802a2e194551061
e995fdb1d2db295eab33268d20784c14b1646f17
55924 F20101106_AAARRJ caicedo_h_Page_200.pro
175645ac2cf50a492a5bf381e4c4b6f1
7381e0311d415d6ca4ba2d5acffcd4599b1a1e71
39611 F20101106_AAARQU caicedo_h_Page_176.pro
798bc8698c5891df91f472f767e9b4ba
6d6b0057cf5c21c3b4daf91f77d1d8d3397c6151
2137 F20101106_AAAQNS caicedo_h_Page_094.txt
5d340e4689a9885867f9837c44b8bcc7
af046cc10c2d9cb5a8b4136c38c4c02da23f8515
112374 F20101106_AAAQOH caicedo_h_Page_085.jpg
ff2b950316f370afa0b55df3bd51a09f
d3def5a29a09da2c5fa9b63dad1acddd0d88a164
44106 F20101106_AAARQV caicedo_h_Page_178.pro
c7392f9015f8a381832b2735b81e544a
e669213c28ef0cc0bcbec8266fc88aa9a2bc267c
F20101106_AAAQNT caicedo_h_Page_182.tif
00f90a9838e49bf84ad9f5b089ddaabe
9c8bf72e867e332794ab3f58530e17f663061305
48639 F20101106_AAARRK caicedo_h_Page_201.pro
5e16c3c830e6cefa2d3691140ce8c2d8
76d2566ae0b16baa7921f75cf0ef357210ca8bdb
40383 F20101106_AAARQW caicedo_h_Page_181.pro
a71fe189cb2397e318da4198af7d0dd8
559bf2ea2edd98062afb5549646f6f21c9b5a6d3
7493 F20101106_AAAQNU caicedo_h_Page_090thm.jpg
3966a9207b9a42b029d028523144f460
44615bf72bca2ddf4a2563ccdff7079a7742fa93
89562 F20101106_AAAQOI caicedo_h_Page_181.jpg
c91ea0dc97e6c64b6fbbca4467dba70a
ded62d926e270f8a329a2a4655b314226b1a8e50
3070 F20101106_AAARSA caicedo_h_Page_010.txt
0e41f856021c68d63c6e2750fc4813ca
781b0735f083739cb5402bca76a56141b517a4bf
47654 F20101106_AAARRL caicedo_h_Page_203.pro
a74dfcab9d7d7a16e137abb49b450b81
2fa0f96c79240404a6c930e447fd5a13a4c3a05e
30233 F20101106_AAARQX caicedo_h_Page_183.pro
cc8a78489f1d58be25cd7c0d5d02aada
c561d0b076bcf3170c746ecb0c9b46fae5f119ea
52018 F20101106_AAAQNV caicedo_h_Page_123.pro
9f4d661c167239f611b2d2db0ece7de1
b9907c6b686441931e5504980b9de6a8321974e9
F20101106_AAAQOJ caicedo_h_Page_191.tif
12a092f4bdb548589ae327700004ab41
037bf1fb248c653654be04a515e75397526aee50
3367 F20101106_AAARSB caicedo_h_Page_011.txt
40a8ecdc94c62780ed2edea0d634d68b
0b4518ede09827f07aa5bc4107323f11588ea455
48760 F20101106_AAARRM caicedo_h_Page_204.pro
6d59ed93d175353434c57550f0ec3f34
6fa9639b3f3ddccecd38ac5e2b4aaf358f6b819f
53096 F20101106_AAARQY caicedo_h_Page_185.pro
db93b1d3a6de19fb000b5c2dc6410dad
de4812d67bbff022d4f3ccec3a5f479ec5bc7048
F20101106_AAAQNW caicedo_h_Page_168.tif
14685e4788c9eec1be9fa926d773c016
1dd877d2ab2bd0e0fd8d8a3a18375314e92781aa
28391 F20101106_AAAQOK caicedo_h_Page_175.QC.jpg
89f562d06dd27fcaec165bc90bcbc3a8
08cfd240c016ceacc800dfbea09d784fbc34fc22
3377 F20101106_AAARSC caicedo_h_Page_012.txt
d1136b23fa032e058ae27f1ea760388d
99866699fcc000553b98b26f0434ee2ac347a064
45771 F20101106_AAARRN caicedo_h_Page_205.pro
dfe70fc591a6ea9079e29fbc3a361f6b
378cdbee1e3135b168bc81a033f19743896e079c
21390 F20101106_AAARQZ caicedo_h_Page_188.pro
31b6b26207bc7f459c5d9a4e49128963
a58b8c66b64fb4bd9c398e20d0726b49a0f41c5a
7587 F20101106_AAAQNX caicedo_h_Page_040thm.jpg
b2c2e096dc4de959eba0447d4c226ae3
e6cc1c49ff25479c55715f2a873ba684451f74c3
F20101106_AAAQPA caicedo_h_Page_127.tif
bb157b5cfd19b364cd35f6487723cae9
7a8e16f070b2ea83451d47427f28cf96f46bfe69
53184 F20101106_AAAQOL caicedo_h_Page_058.pro
d827c049050122f7e19ab57333011cbb
f5e8c837012941ea352080c65a7eecdc008e5dd6
3239 F20101106_AAARSD caicedo_h_Page_013.txt
4aab5f78a9dbcd6006d71a0ce281ab58
748b1e6fe4a1aeac4ae498b991bbb211c397d2bb
49831 F20101106_AAARRO caicedo_h_Page_206.pro
93a3241c08f285a771909b8398ddef55
c2e20ea12f66007efa5e72e30dd45a7bf3ecd07e
1981 F20101106_AAAQNY caicedo_h_Page_074.txt
aacfa0a7a6559fe64fac288b93931b50
c08a5c9df05349ee78f9be84ba0f3a18b01b6a72
2650 F20101106_AAAQPB caicedo_h_Page_163.txt
59573769f3307e1154424e57350f20ab
70eb5cb7f869bb2bed6e7d7be0df87370515684f
107425 F20101106_AAAQOM caicedo_h_Page_060.jpg
d77612cce615196e133f195b9199c350
074e052453fe1fa4f4c3d5f581d78bdad6f67422
3273 F20101106_AAARSE caicedo_h_Page_014.txt
0448193fd926bc450b4f5b4521c7a09a
885686fc743b0a5d0de1dd823ef62c525cc32a63
47641 F20101106_AAARRP caicedo_h_Page_207.pro
e6ff28e4fbe5ed57b70f52915ca782d8
ccf981218d04c41afaca55d56dc911e1da2c1f23
29168 F20101106_AAAQNZ caicedo_h_Page_148.QC.jpg
b72a0a830dfe5d9c7650843bf3522c4a
2b6d65c11e127dcec948f486e9384fe25333ee51
52925 F20101106_AAAQPC caicedo_h_Page_153.pro
d90aca7d92a965de89386e7e80db4928
da4e0748d8eeb22c1edc387eda18268417f648dd
1051952 F20101106_AAAQON caicedo_h_Page_061.jp2
25a51ea2ed01c048cbeac10c6d537c32
e2b171c96aa5796a2e45eb326035564cbb88372f
3237 F20101106_AAARSF caicedo_h_Page_015.txt
42ee3525e7426f65c7b3a7b4db307aad
8a794d1d56af31e0dd314a5c998e171eeccbd981
48383 F20101106_AAARRQ caicedo_h_Page_208.pro
e1f23e74674d1ecc0b892eddf8e5e45d
6e3cb84abba061e13a7de7ab56825bded24e4c57
F20101106_AAAQPD caicedo_h_Page_190.tif
a89996da858a413a4375adb898677359
3ab6fcac883f82883f5a585890953029c02b917d
55089 F20101106_AAAQOO caicedo_h_Page_140.pro
4b46b711808d37f64db99897542b8a36
0b5dc36e904c1db7542c374aeb5c8df547177cee
2097 F20101106_AAARSG caicedo_h_Page_017.txt
bfc56ddd461bf20a74942016010a935a
7e81cf50e41c6dcfbef74302bddc184997a9e3ac
48671 F20101106_AAARRR caicedo_h_Page_209.pro
fa89d66fe21cbbd33c387ccb1ca7cc11
b96f755e19483ba3963cd30f7ad9e3ba0dee9bc2
29247 F20101106_AAAQPE caicedo_h_Page_158.QC.jpg
2d6093286c09bdcfe3412a2ca2bbc6a9
498a46fc81f4a963c713f593f0921273736a2fbd
8598 F20101106_AAAQOP caicedo_h_Page_020thm.jpg
078d73a19b10c68a1434c05133eecc1a
57b98a4dd1c418e9ae5940900678b94f70c623bd
1230 F20101106_AAARSH caicedo_h_Page_018.txt
746d16caf33ab7aa071b623c4faf31c6
b31298231310e7d446c14b61827ce8e0f17982f2
46239 F20101106_AAARRS caicedo_h_Page_210.pro
39cf640aadaf1030920c310ecfc87803
366a4f0d097d8704c079de2bc59e69342227c43a
141324 F20101106_AAAQPF caicedo_h_Page_010.jpg
0ef0be693a04863ae5935fc0558e6411
ec44ca965dc317b4861560455bf299e9b923e763
8626 F20101106_AAAQOQ caicedo_h_Page_119thm.jpg
f566a8ebc589619ef8959ffe5179bfb0
e054e4d75803e6a697b8c3b27b25985a4cb4e8a8
F20101106_AAARSI caicedo_h_Page_019.txt
4963a26b547cf354de3225a81ca03be3
5389ec6064a75595a1697fededee09952796ad54
46178 F20101106_AAARRT caicedo_h_Page_211.pro
a319903bfc8d9bbccaf54786731b9d13
0f0fdb2ee6dafde6736b5044fd4e3128e98225b6
93353 F20101106_AAAQPG caicedo_h_Page_149.jpg
22608b64ab3f3088135e80824142b813
b7fa805c6a040753e26d82b282e5001ff17633d0
7660 F20101106_AAAQOR caicedo_h_Page_106thm.jpg
712ceb8cb826514524643b13f1d995a2
fed2de37940325bc95c0616005fefeb9e63f4284
2196 F20101106_AAARSJ caicedo_h_Page_020.txt
7fbd2de45d25932df6549c3f947f2803
6a74beed21cb5b109d7e99a5adfbe4758d6e7f97
136 F20101106_AAARRU caicedo_h_Page_003.txt
e386a7a744a64998f98738167e538000
0f5fb1ae08855ce8c6ccd1f124e152eca94727d2
F20101106_AAAQPH caicedo_h_Page_203.tif
59e163669d6738bb2f37938398b38b93
ef323b7e52ea5d2ffc6a7d05a7affb0776ff79f0
99076 F20101106_AAAQOS caicedo_h_Page_137.jpg
85198640e9654c081d8964733382e99a
621ed22e36b597856ec23d0da4235d9853085682
1957 F20101106_AAARSK caicedo_h_Page_021.txt
dc3d9a7d8545cd310bd7c2951bd4c2b5
b5add7727eab65d503197e087f73e90b69e7c2d3
1327 F20101106_AAARRV caicedo_h_Page_004.txt
d176ac9c548fa6d953e79a82cbb4e4ab
eef5f1dda2e271d347475263d243fcbbaddf3631
5895 F20101106_AAAQPI caicedo_h_Page_004thm.jpg
5ec0f05d4fba1547e633f1eab047e73a
9a67a4ba2fc4a9b62a80629e5855477b2c508628
2199 F20101106_AAAQOT caicedo_h_Page_158.txt
47fae28a016ca3a31a32c0d0a785d314
4001af0f82942606ce07078b0f644a5db46bc68c
4598 F20101106_AAARRW caicedo_h_Page_006.txt
b8843aa744d921b3b44735ad3cd38609
8de478b2d80f033469b2d529eae8713b399c3f97
8457 F20101106_AAAQOU caicedo_h_Page_115thm.jpg
bfabaec4c317e8dfb80cd090aab97e05
dbe6810481dde6aa5a6b86b1c548630420fdf4c7
2010 F20101106_AAARTA caicedo_h_Page_040.txt
4c4dba39cf9fd3896eb38c82f65655ed
24491a37c265b5255b43283af501e69ef688f6ef
2094 F20101106_AAARSL caicedo_h_Page_022.txt
7d6f81d3d41426d178c523d6d08726c0
90202427063ddf2ddd3f856adc02e635ab26ac34
4753 F20101106_AAARRX caicedo_h_Page_007.txt
555d6dd32003900ddfe2b033cdbfd4c1
fe75baa3a1b670848c464b0dd114792709ba0806
88739 F20101106_AAAQPJ caicedo_h_Page_059.jpg
530d1b96944e5266f6933e54ce5b9abb
98e8cdfd3f628052cbed155b198f0fb1d70c0a97
30488 F20101106_AAAQOV caicedo_h_Page_106.QC.jpg
cba31443725516643719f7f5bfcf2a9a
114b3d0dbef7fb23f8dd6dc8e4058702306bdce2
2112 F20101106_AAARTB caicedo_h_Page_041.txt
4c9431b4f5bece898e47aa5fa4e80e78
15c062bdb75ce448dd4dd762813e20fc08ae0a42
2059 F20101106_AAARSM caicedo_h_Page_024.txt
354824ab220ef947f57aafc068357013
071324d071feb19501e03a4f056c7ce677373c7b
1075 F20101106_AAARRY caicedo_h_Page_008.txt
8187db178cd1d21a98938715f950cf9c
2aa87e094fe368a35930e4cbb9e0868b9d71c07c
92113 F20101106_AAAQPK caicedo_h_Page_070.jpg
4a1f7e395c65c1e845e0d74f70789950
37f88b4d6939c4109dd5d75bd777667744fe8fcf
35760 F20101106_AAAQOW caicedo_h_Page_041.QC.jpg
9c17925340a07ba723791d124516f0db
1165dc9772a50df36904e576340b67362880959b
1673 F20101106_AAARTC caicedo_h_Page_042.txt
bdb5eb17eebf5ad514bc93c7d1db298c
d04504c5ab6f3341dbeb300dfc8e2a0297a17c92
1507 F20101106_AAARSN caicedo_h_Page_025.txt
bff6c90627ceab1b98aeeaf6d82d0928
f963cb49dfdef4d113119daf01ccd2f454780757
2364 F20101106_AAARRZ caicedo_h_Page_009.txt
05405f38f49173cd749ab052c90889e7
225054ab033749b4f7400330f7f6c596965958bb
1755 F20101106_AAAQQA caicedo_h_Page_146.txt
418ad23131c498ba4dfcd22f3fb53176
a8512e2be8a63d9d9d641aeadd5e744d7cacad2c
36035 F20101106_AAAQPL caicedo_h_Page_184.pro
4b7b9dbdbfd9cad2470d1279463f16e6
abc3e94ebc8a080972c9eaeb3f940c471067647e
F20101106_AAAQOX caicedo_h_Page_197.jp2
d81d5773459f42bdfac5f58141873da3
fa51e2c782b79b8b2abef8158b344fdceb6e40f7
2125 F20101106_AAARTD caicedo_h_Page_044.txt
105d762d91f0cc673a6593b7aa58283f
80febb09d7959d079580a178ceba0e27b76e7448
1733 F20101106_AAARSO caicedo_h_Page_026.txt
36842cba500795353cf4384cbe3cd9e5
0a2df3fefef83c10a0df89520343f20f009774f3
89136 F20101106_AAAQQB caicedo_h_Page_121.jpg
2ef763ed099872e3293e7a63a6895f4a
c9d61f81b1bedccd5b5fa9c258851b843059c254
F20101106_AAAQPM caicedo_h_Page_085.tif
76cce10a1a7e6d0fcd1a7f1c877de6ad
6089e08e9a65a1e98eaff8ec1ab3d4efde2c7db7
44263 F20101106_AAAQOY caicedo_h_Page_133.pro
4591f7ac869326216d4dd056fe60896e
f2f7a05fb229eb8bc01bdc136efab46bbbf3bc3c
2104 F20101106_AAARTE caicedo_h_Page_045.txt
a1b71c327ba8d12f38271ae99421f373
1fc7070483328bec583dc0f96c3f263771e36d22
2236 F20101106_AAARSP caicedo_h_Page_027.txt
ba1fcb87be16b1b95727728ecfdd886c
ac29d2b41516d0a1b34f1796121ef63a50ca3a92
2357 F20101106_AAAQQC caicedo_h_Page_072.txt
19176b0b120c953630843d2454e7842a
5d2d48abae5cbd495dce9c51cbbf28d904b628e2
9041 F20101106_AAAQPN caicedo_h_Page_033thm.jpg
7915b9027b4e140617d283ae26447d43
89b000093a4434182cc0ab614747639b513466fc
96 F20101106_AAAQOZ caicedo_h_Page_002.txt
a430b945231498f7d8b0a72f49c3955b
1cad3ed05c5c72534a7992eb301b2ec7717f746b
F20101106_AAARTF caicedo_h_Page_046.txt
376a18ec910feac6f61e32380851b1f0
7cb8f16e051c55c6df2bf78d89657200a0534a39
2048 F20101106_AAARSQ caicedo_h_Page_028.txt
bdbe54984124ca322919818f18855da3
a6188b058d58cb5fb7f79bb83a91ed934a91a8b7
1051961 F20101106_AAAQQD caicedo_h_Page_194.jp2
ad317c06ce657fc00330c3d777f8fda8
b536edb8646a829f96b9c294e98c531c99b86e24
31064 F20101106_AAAQPO caicedo_h_Page_095.QC.jpg
2cedf94da5e6fdf73337960b31bbd9d2
ae279860d039770546710cd1610d23b749e57ef9
2535 F20101106_AAARTG caicedo_h_Page_047.txt
49029be04c44248f172608eae5215619
2aa55bdaee2f5d43fec84c6c90b5a94f07a022fd
1742 F20101106_AAARSR caicedo_h_Page_029.txt
269f843e8dc016f247307ae6414ca68e
1a2ac50014ff57adbcd5014ddd744ddba42440fe
52552 F20101106_AAAQQE caicedo_h_Page_125.pro
43bd5ec5a4a6609bc31da2a4528e7613
bb40448b69df7875b7eaf57dc56ff59d0665ebfd
F20101106_AAAQPP caicedo_h_Page_163.jp2
dcf708be9b01dbd3bf4dab0c1d362f66
9b0cc96cf24c2f1fcedc0f6a318a6c0e97f60dde
2036 F20101106_AAARTH caicedo_h_Page_048.txt
187f67f1c862d0d5f85c3217b62a7c90
21f42ae15cf6b8f7716d7cec40eef32c9930933e
1313 F20101106_AAARSS caicedo_h_Page_030.txt
3f3592e2d9b455e86f85de886bcab2cd
a68a4df5d0f9b86860d357e1c5cbd86133a8c148
F20101106_AAAQQF caicedo_h_Page_034.tif
5c07fed96922c506c96bb2d81d20dc96
984a0cf3c712cab5c652b2d4f869eedb752acb54
8046 F20101106_AAAQPQ caicedo_h_Page_101thm.jpg
92f28700d9495f7685f7c38063b399ed
0e8b7072aeee28f540c44d5ad7c55a0c05a9afad
2098 F20101106_AAARTI caicedo_h_Page_049.txt
6c30f7497bb223ee04bd3af9b61a2d9e
2912d726e240d5a145b3acdb959e902fb6fd3bb3
2200 F20101106_AAARST caicedo_h_Page_031.txt
0bcc13bf5670a27cf30da106302c0dd6
4d36cfdc2a4f2bd37bf0bb6a0dc8e2a47a70823d
F20101106_AAAQQG caicedo_h_Page_115.tif
d2c03716af43b86a9321863d07bbfc8f
8e2c87777b2998afd2697e127c45b147afbb4eea
F20101106_AAAQPR caicedo_h_Page_207.tif
800f24a856f6e33a53cdb93cfe116de5
a44cbf8e1f13bb059581674b9f085a80d33da241
1891 F20101106_AAARTJ caicedo_h_Page_050.txt
743b894988c3d183b723f66820ef0ac2
cf3e3dd98c31762212b487484036d31260ac3987
2139 F20101106_AAARSU caicedo_h_Page_032.txt
50e4f3de5d73be7ed9e7def492fe112e
90c3081d76d4af1c45a5fb32bb410bae5c6c9a27
F20101106_AAAQQH caicedo_h_Page_049.jp2
ec8e512c807737cec00a193d4132030b
79b22c12144c87a4133b3d6a6073d74ac867fcd6
F20101106_AAAQPS caicedo_h_Page_210.tif
6d24fe794747c7be111504098619da36
d9ea3e94d41c03d2cdf09be9ee90643c2a5a7ab7
1936 F20101106_AAARTK caicedo_h_Page_051.txt
b73bd0eb8dd0f3fa54ecf7117dc4e3e3
61308c41a3d82196aeb1f75872c18dcdc4a2fe35
2223 F20101106_AAARSV caicedo_h_Page_033.txt
5e53805595526104953d40c46162660f
d0a2bff814c6b42cfe2d270496a5612adccb8d31
101671 F20101106_AAAQQI caicedo_h_Page_209.jpg
7c5db49d57d9bc6d7ab4b2e40207fd14
c02706dcd824d3dfa566b707090dc15bb56da63a
2133 F20101106_AAAQPT caicedo_h_Page_114.txt
4e8c4ad0c2f917e337c4e885c678313e
cb1ee2e57a62b7f8a100b76a1df644a805657209
2033 F20101106_AAARTL caicedo_h_Page_052.txt
8b6ab97c8ba2df2b0c5f3c682e670984
08636f7a00a0e7b1d517a6eb0deabad546c9fe15
2330 F20101106_AAARSW caicedo_h_Page_034.txt
8015c921b2bf331b0dc691ff68caf178
68675d8760fb45c1ea7ac1b2f954e1b2afcfe94f
49474 F20101106_AAAQQJ caicedo_h_Page_057.pro
36cc4c5ae9709cdb5d9845fa7fd1303e
6771e684cb393f88ada1baedff1eba39bde3aec9
51076 F20101106_AAAQPU caicedo_h_Page_182.pro
71c1763f04e5d30616449a56c8b3e534
c308c9278c8f6b5449b6dab03cff46e425148210
F20101106_AAARUA caicedo_h_Page_080.txt
aa870f7092af87b5b31bda1e6dfcddfa
4a892eb66637bfd65988c8f8bd76b2db7c536cdb
2085 F20101106_AAARSX caicedo_h_Page_036.txt
db5987b19f4b4de3fb33ea8420a50200
e42ae8855092f9c50dea5d1ed728fa6f95aa1dfc
8894 F20101106_AAAQPV caicedo_h_Page_127thm.jpg
e636d48fbce37f3ccb4000854d668be5
b34fecff8fa8f778dd3f57542a473ec844004ecc
2222 F20101106_AAARUB caicedo_h_Page_083.txt
ea9c3f69807dbb379fa168eb046f3e1d
105c058fec66f8c108864116054032043fed255e
2128 F20101106_AAARTM caicedo_h_Page_053.txt
8939ebe1d130c4c17e50da970134eb17
18006b6187185284b6ae76a24e8b1d36d32afc17
2141 F20101106_AAARSY caicedo_h_Page_037.txt
f6c2311152ea5670a1f925af07806e58
4777fc5bf803f9cc1dfe7cc75ec20036ef7099d0
7453 F20101106_AAAQQK caicedo_h_Page_079thm.jpg
898e3e4b88d6a9e66f8eec4b8132165d
4c1ec3879108d8b279918623d62d8c0dda4784ab
84872 F20101106_AAAQPW caicedo_h_Page_132.jpg
c345c0460b326fbc7f497bb0540512e0
d5ab572f21c32bd5c98dd4cc163a85bf16d95921
2499 F20101106_AAARUC caicedo_h_Page_084.txt
2a50d01c1b5d3cdb1df627f28cfeba79
e85c2874f4e75645460f529336bd0803a77d0a9e
1987 F20101106_AAARTN caicedo_h_Page_057.txt
8a2458e0ebcb8a37753ce8ab045073b2
78d542efdeab14b212fcebe4124ee72293d84c2d
1838 F20101106_AAARSZ caicedo_h_Page_038.txt
61f8f7da8b5409743fec5d8ea42f0c57
f98299c3ac81143c93bf4a925e4954ed29c12ea2
7492 F20101106_AAAQQL caicedo_h_Page_026thm.jpg
51117320253dec009decc4eff0a95afb
6df00de3bab3002ba6aa80f20766d5e210d79155
54793 F20101106_AAAQPX caicedo_h_Page_119.pro
5213b0e482a6ea3453b66867b95e78e9
044555657e7ec12fdfef9d36b246aad38a26b432
938843 F20101106_AAAQRA caicedo_h_Page_092.jp2
2792e9a1253f205faccd99115d7eddb8
b26221cfea4bf8017ba3b01e8ec02dc28c6cecc4
1196 F20101106_AAARUD caicedo_h_Page_086.txt
9ff14cbf821f556194301508002187fa
91842643fcad0d565f9b53cffd8374092fd61cce
2092 F20101106_AAARTO caicedo_h_Page_058.txt
b0c856953acd9e572ba8923330547dea
ea1cbcf8cac2dd97b725f75b157596283f3eed30
F20101106_AAAQQM caicedo_h_Page_038.tif
47b892315b5452e9f5a5167883efb2bb
1c9fe73f331fda4e8f27a60c0525504beb3a98ec
29393 F20101106_AAAQPY caicedo_h_Page_121.QC.jpg
b60f55f6e38f7e6958a465db3bbe6db9
6d0c15d2431edb46581067c80383d6713d813970
F20101106_AAAQRB caicedo_h_Page_058.jp2
3c2c39f9865d8a32bf16cada358453a3
a64c36c19ae331085647fa5b34248a991b35c63d
1966 F20101106_AAARUE caicedo_h_Page_087.txt
ac52f5160fe2891628b401d20b803ecb
eab0aeb66a0bff12558360f8c6003bb04242165b
2184 F20101106_AAARTP caicedo_h_Page_062.txt
dbe7ffeaea589f3b7d22d49b6f445997
2aadd135c35969bc5623fb19e4029efa97abbdb3
1051977 F20101106_AAAQQN caicedo_h_Page_118.jp2
2409b00282b3805a798b84dbc25446ce
82f4e4416e0069aa37ac4c7acffcf7c8c3959283
42685 F20101106_AAAQPZ caicedo_h_Page_100.pro
d29bbe0d054b2ce3734ce1ee55a9716c
cdbb2e8baaaede6b58d3974129f702df1f011db1
8496 F20101106_AAAQRC caicedo_h_Page_159thm.jpg
00cdd0be8dbc9af4325ce8df44adc7de
b5159c193513f6fc9f3a8de11d64ea7308b400c2
2355 F20101106_AAARUF caicedo_h_Page_088.txt
d12f58a81ced4a603e82d26ee803b5fc
59ca3baf7567860166531f673bd13af7c9ee370b
F20101106_AAARTQ caicedo_h_Page_065.txt
d436a791e84c32ada19f508b2a119797
3df7c9ada2aa0d388e02c95779babf22f438d7dc
8565 F20101106_AAAQQO caicedo_h_Page_001.QC.jpg
d9e219f6cfe3e462fb55bc2338ea2258
d6c9ba899e1c1d8075df94341e9afb012bfa21c6
1431 F20101106_AAAQRD caicedo_h_Page_150.txt
2647d89d6dce2af7e3a98921eb8857a5
64c25581326d792741f31f9f42034d192cb7d0cf
7594 F20101106_AAASAA caicedo_h_Page_047thm.jpg
c9c879b66c56bb888ae022b6e73afeed
c16dd1fafec2dc00c38fc5ab0fc14caff6588e50
2107 F20101106_AAARUG caicedo_h_Page_089.txt
ac2d96562a2281075bed582834643c55
1e5ec1fe637456025bda22baaa6755cd2b171b7a
2512 F20101106_AAARTR caicedo_h_Page_068.txt
d7fff9ceaa3e5ea83d12e728be43ea3c
86062a54e95873bf92a51d1e3b2bfab701d6c165
F20101106_AAAQQP caicedo_h_Page_124.tif
817c3f6d6e6162f8f9a638df7a6e8d67
cd3880e05221668089e42a6daf3883530d8d92ff
F20101106_AAAQRE caicedo_h_Page_124.txt
bfcc87104128d744e3dcf4a95ea12ce3
ce2ac7774c0abd3fb5dc08362f161a6d590239fc
30477 F20101106_AAASAB caicedo_h_Page_048.QC.jpg
939909bcfe485d7bf1eebcf4805204c5
d032679af9234d5fdf2b3b110358735f77d45f14
1949 F20101106_AAARUH caicedo_h_Page_090.txt
36ae422fed78d7363016b8b85debf531
c29b3230d5ed92b3c120661e4d9d1d64977da0a4
1950 F20101106_AAARTS caicedo_h_Page_069.txt
dd7a9b73a097cf27b0ac70227a8bf1cb
3876fbb5213d3c6322bd69a4424c3b1a518f1c84
52701 F20101106_AAAQQQ caicedo_h_Page_055.pro
092985bb006cb4d96713434ad3a384a1
4f825f3af08a983f4a38846548adbe0f13dcc782
101609 F20101106_AAAQRF caicedo_h_Page_088.jpg
03e6abcc0d316867a81efba9f1cb90fd
e6df996d1e10bc29a2d70571948ea6ea5ce10b15
32540 F20101106_AAASAC caicedo_h_Page_049.QC.jpg
392de12fd7e52d6c9b4efee0c127b0f7
1ef94ce8eb1656c2d2ba2f392fd0a8caa99efbdf
2192 F20101106_AAARUI caicedo_h_Page_092.txt
098b036e54b15fce50eb42577b4d3753
dbb0435a1275b7c07d18d29181bfaff51000bb59
2003 F20101106_AAARTT caicedo_h_Page_070.txt
c81ce5ea6cf55324af4982655abce500
d7362f4dc6044fba23edea882f8809ce04d3d1e1
F20101106_AAAQQR caicedo_h_Page_029.tif
eb6fba90880b28cec0632204f47a441a
2116f7fb38c396f25339f3c9cc17fe010ea08e40
43775 F20101106_AAAQRG caicedo_h_Page_186.pro
73b0a6be3d0845855dbb62d89b62d611
8f170113fbd5e719b9229cf3fbf5f55f450c321c
30301 F20101106_AAASAD caicedo_h_Page_050.QC.jpg
4760e285c3a81381db6e81c4ff2ab039
9fd9d1367cc2307a831837a67867b0632a863ef8
2261 F20101106_AAARUJ caicedo_h_Page_093.txt
f1524cbf4b32bdb662baa27988ba6c66
98678d70f01cb66b579a3a0c66ad5c1a886f687f
1924 F20101106_AAARTU caicedo_h_Page_071.txt
5c641c4f7c44ec7003cc4584ff0fad9f
e02025bd06392b6dd6a03593c026cb4c3786a32d
2095 F20101106_AAAQQS caicedo_h_Page_043.txt
43d8a260a4402e1b81d5a4b610a5523d
2489661c96eeebf295b6ec8174a65058023a48e0
37969 F20101106_AAAQRH caicedo_h_Page_010.QC.jpg
9dad7131ec108cf8bec39cab6dcd26a8
dbd8ffa3b688057f52fe586e6163097d368a732b
7855 F20101106_AAASAE caicedo_h_Page_050thm.jpg
0adcdd04001cf35645ff101db352b846
545b087bfa556e420ace463257a9afd713dbc46a
2042 F20101106_AAARUK caicedo_h_Page_095.txt
dede38c87343bf20cf6b68b57bb62c19
5ff3afbfe6c6dc06b232ff2de29b8aa880b4816a
2108 F20101106_AAARTV caicedo_h_Page_073.txt
f329508980b079f2171510a4ef77dd7d
911a5b19c2e569b75941fdbaacef4e9b125f3ee9
F20101106_AAAQRI caicedo_h_Page_036.tif
018a090bca60f9d03c20ff847b228228
6e845cc86d13a7305f51a3e8777fc4a1e8e051a5
2101 F20101106_AAAQQT caicedo_h_Page_082.txt
99029af031f933495216a96b01aeab4b
f008082519f4e7bae2e47e28a8a7ff250f7e7ad3
32743 F20101106_AAASAF caicedo_h_Page_051.QC.jpg
f6d505b5a47016c05da5c13cf7e02aae
d7eaa745f1a3a16841c81d3c1f47a69a6592f703
2065 F20101106_AAARUL caicedo_h_Page_096.txt
7f2d74b7ca164dfb84c549b2abec5884
3815beed501caf11bd38235b1893e67e45133778
1980 F20101106_AAARTW caicedo_h_Page_075.txt
0aedb49369833b4d411f4411bd2e6c41
56c413c3106f3a67c77e6638d25440f79b04fc8a
37720 F20101106_AAAQRJ caicedo_h_Page_180.pro
3bda05c912239e78d73eb9835ad60020
b34ff30c416156d933a7e338006815249fb9f415
44049 F20101106_AAAQQU caicedo_h_Page_050.pro
a6589b75c6d4146d22697d23d970dbcb
9e048285ba43d07ffcbfa940d0b24e987ae951df
8478 F20101106_AAASAG caicedo_h_Page_052thm.jpg
1dfac5f2410fbf877d9f74cf683404e6
9ed41db07dbe66c0014fdb543f8c39d8ce97e415
2116 F20101106_AAARVA caicedo_h_Page_116.txt
cf2c2bba218cdb5ebe422dc1f69aa0f0
ec4772f89441be6fcb729f2591a1054fe87964d9
2127 F20101106_AAARUM caicedo_h_Page_097.txt
29de104ea0918feb0e7d165d4ea744ee
928921be6a2653b1cc36236a7eea15b229c0ddf2
2012 F20101106_AAARTX caicedo_h_Page_076.txt
29a2690f1b9bab8bd6d78125a201efaf
f3d6d04e769c4dba6c9d1c14b0d7bcf241b8e921
35082 F20101106_AAAQRK caicedo_h_Page_098.QC.jpg
17b458f96ab76bddea146ae4d0273340
13f836d10ee10020de184c64d624ba90a9943d0b
8814 F20101106_AAAQQV caicedo_h_Page_194thm.jpg
8168b553f69ee914434c1fb50ef60444
21393bdf31b421466b0bace1924cfa487d5a8bbf
37026 F20101106_AAASAH caicedo_h_Page_053.QC.jpg
dd4cc72310dbb98a1d8fc24d699314ae
54a88983aeb6abd51bf412f4d52b2af481f1fa3d
2521 F20101106_AAARVB caicedo_h_Page_117.txt
a731d4739b247c918c525db02df19f4f
0cf27962ff26735232b3f7e67009ab331f3567e1
2142 F20101106_AAARTY caicedo_h_Page_077.txt
1288d4a89512e193d35452d6f604962d
194849ea99329a3dc141839b450d475f2d29bf26
2021 F20101106_AAAQQW caicedo_h_Page_061.txt
73743faafe903a1e2daf211b0889b220
a9235dbca8f0c7d0abca2afe7234d3427b5e8f15
8641 F20101106_AAASAI caicedo_h_Page_053thm.jpg
d52f5aa2fd56b0e0744ff45fdad4797d
7cb20757dc489835e944ec770c6b330acd5b7837
F20101106_AAARVC caicedo_h_Page_118.txt
be4e61cc3b57917b809903fdc6e4c1de
d5bd6b9d0921e56617ce5c1f543b7be7a7e8b6d6
F20101106_AAARUN caicedo_h_Page_098.txt
1ddea2e05622d9327342edb5044045b6
5b27bbc9797bafc6c98364f58760eba72e7f141f
2444 F20101106_AAARTZ caicedo_h_Page_078.txt
fc0573d8f4916328138500e0cc9216eb
bbc10b7fd685b453865d02176573cf1732a4abff
F20101106_AAAQSA caicedo_h_Page_179.tif
439ef089ab8b178f8b340868d2e31dc9
fb875f6041a60df98279ea500bdaa909de059738
F20101106_AAAQRL caicedo_h_Page_056.jp2
20d068a79c0fbb56137dea1bf37d4772
32259fe31a88cdd3a015e4b5c642b1390e2a196c
F20101106_AAAQQX caicedo_h_Page_206thm.jpg
662ff5575a4237e9919768a5a0c4ed0b
65bf3b18d0a1b935d20fe4f633c1828a335b09b0
19997 F20101106_AAASAJ caicedo_h_Page_054.QC.jpg
e5a8071e1260a7b01a2e50fbaf2874f3
83b2fca9fc3f45a7e5b78baa9b2eb2a66503c51b
2156 F20101106_AAARVD caicedo_h_Page_119.txt
bbf039a1a320634f5eba8c91fe4889b1
e692fe1c50654ac30d8bd36d848f0da20c14fd74
1951 F20101106_AAARUO caicedo_h_Page_100.txt
f1853a5c4e30d5df3d91d4a6387c3e96
d9f2c427ff4ae2a891eb574a78a7a1d1696ab872
99210 F20101106_AAAQSB caicedo_h_Page_049.jpg
4bbd6a0bd10d05468a4a75bc357b6244
e332441b882c1aa1c1520e81c5eb268ce2215fa2
7529 F20101106_AAAQRM caicedo_h_Page_165thm.jpg
14c1ce30f72235cc5bd3ce7e3ad649b7
49fe2450a28dae38590946c74f41e791a275b4e6
48909 F20101106_AAAQQY caicedo_h_Page_032.pro
414f130127b997c412d21717508c6032
b3c0c66ddb4b88a6be5c92c312c13760ff2b5354
5052 F20101106_AAASAK caicedo_h_Page_054thm.jpg
1c097903b5b95799b13b1d8a10458951
80314bafa0d3d98a1cf4c0bcc927d32d094b6018
2337 F20101106_AAARVE caicedo_h_Page_120.txt
ffd9b9de817ad28571e080785517236b
3ae06b8ad9fa7e6127bbad419dcc6d6823a5bab9
1863 F20101106_AAARUP caicedo_h_Page_102.txt
af5d21a07ecc61ae9a185fb84e1f2a9e
5b6e4eb8000316a0010a37bd565f75359e303658
27317 F20101106_AAAQSC caicedo_h_Page_132.QC.jpg
4b6e8b97ebade9bf7fdc9257b02bf2fa
64f8dab4c406fdef0618a6ba15b22b03e7d2180e
31021 F20101106_AAAQRN caicedo_h_Page_009.QC.jpg
63941009fdca393d671fc6964d3ff741
4dc920d24fa0a8163ef64efd2084e259a60a0507
1029881 F20101106_AAAQQZ caicedo_h_Page_104.jp2
9c47d0011f6a6ae46641f1cd678bd536
2c8a3d732d26138cb37e003e25074e5a44f23462
35766 F20101106_AAASAL caicedo_h_Page_055.QC.jpg
0821272631472296933166a1d67a1bc7
ab1229b7544cb7e6a2dadd9e4e156f829486ab82
1860 F20101106_AAARVF caicedo_h_Page_121.txt
314204da9dad475bc19f957274352b36
3c20f1b316805c37fd53d7d32835f4dab4f23858
1834 F20101106_AAARUQ caicedo_h_Page_103.txt
656f1c0df56011f737402e00225de471
bdf822ea1124ad221ae17bfc306b8386d17924fb
37775 F20101106_AAAQSD caicedo_h_Page_141.pro
583e3019609316e769dd049543c0cebc
9782b293b1c0a7b78bd61c0e9159b0da371885f6
28069 F20101106_AAAQRO caicedo_h_Page_092.QC.jpg
0dfc7e40a3530f668fdeaf1389d539ad
35da837dde80beffec852e91320d5f5458dc4a5f
7951 F20101106_AAASBA caicedo_h_Page_065thm.jpg
5b3a53734d2799ca261d011c13d8c19c
51b4382dde70bfcc3f9c1f7ee06849ef580e585a
F20101106_AAASAM caicedo_h_Page_055thm.jpg
10e96968909853c8e683d3aaddc69142
65b0a032d57c8bd89b8c136837d651bc1d9b4705
2152 F20101106_AAARVG caicedo_h_Page_122.txt
67443211e8de0377e2d843ce1a7ba5ad
d8af3a9b196bdc70c00dc88a76167c78113c350a
2158 F20101106_AAARUR caicedo_h_Page_104.txt
57ff37d5d355cd9a2e578d9bdf274165
48b5d3217a4df616a3b6a7977ca8591280b321fb
35539 F20101106_AAAQSE caicedo_h_Page_189.QC.jpg
1059a05f98932536ba9d0e918434a72f
95eaad4d2326f892a7caf131a0d947ccca32d732
32960 F20101106_AAAQRP caicedo_h_Page_087.QC.jpg
84ad2607934bbce4ea5757d210d424b3
4331c6a6a2e544859c492e5c22e7cf1c3dfc2cd2
29638 F20101106_AAASBB caicedo_h_Page_066.QC.jpg
11ec9bfe8e06a7bb99dcccd3a2743dfd
7a9ba2a3a193ad2816d3f93095394cb98172e4c0
31730 F20101106_AAASAN caicedo_h_Page_056.QC.jpg
bd6dac828edc5786bafb1f547eee7226
d624b6e860eebcff8975eda2dbaf37e7cd74c64a
2062 F20101106_AAARVH caicedo_h_Page_125.txt
84c59f121267b2f0ac6885fe85bc25f8
7c08e145fffb00ba5215b59d02d14a51b676f26a
2057 F20101106_AAARUS caicedo_h_Page_105.txt
2bba1405e08643741c8719f9a3227c3f
5412eba3845efbafe4c681ca1c5829c88e0c4ba1
7982 F20101106_AAAQSF caicedo_h_Page_126thm.jpg
ec045e1a5f4ce49b5b6ce70c51b6a722
871bd6849aa0f1fc5627135378c375d1e98516a3
F20101106_AAAQRQ caicedo_h_Page_072.jp2
ac48934c4ef61b89931e6cc7e924864c
9020cb6ac80cd1a29b556f5d5b3fec06ae5445bc
7747 F20101106_AAASBC caicedo_h_Page_066thm.jpg
4d3d47c5851e9b240ee099e49663fe82
e5cd9d1202ea9e2c8e410fbfce6d83dfae543f3b
7582 F20101106_AAASAO caicedo_h_Page_056thm.jpg
cdd76ba5468dbebf1eb383d898ed45f7
24cad3dd39b3132afc974c0b6807e4fd0391d8d9
1983 F20101106_AAARVI caicedo_h_Page_126.txt
b89d79d9bede8ee4f8030f2778551113
3b9b37b70e1e53531dc5efbf9e4cbc6d08c3bee9
1897 F20101106_AAARUT caicedo_h_Page_106.txt
cf15ffff9f88dcaf7d0e8d2f1ae3862a
96a4566a725681f412826d796db595ab6dc0ccd1
37832 F20101106_AAAQSG caicedo_h_Page_166.pro
e38bba51939e937687004062b55e2136
3174732fea3827fa5fd7301a0c6e6c3b90b8ac3b
8434 F20101106_AAAQRR caicedo_h_Page_045thm.jpg
7729eceee965ed6f772f05fc4761f68f
6f8fc3b310b62e7c952d25c150b37af5ef98de4e
33037 F20101106_AAASBD caicedo_h_Page_067.QC.jpg
af94dc7124a9a0144bb035cdccb88344
fed73472ca129677a05dcfa1446649b8178bca41
34169 F20101106_AAASAP caicedo_h_Page_057.QC.jpg
a2a5e58f026f95bc068b0b112d052eae
fbef12bd7eb98555888477a3bbceede784af0d74
F20101106_AAARVJ caicedo_h_Page_127.txt
6874d651bd7b091889d56c06bd13ef96
fd482b7ba3aa3501dd4507e380881f30021583e4
2124 F20101106_AAARUU caicedo_h_Page_107.txt
8097ae703dd63f68699fcc68569fe11d
d3eae47cbf7e695b785aad9a91adc3199e5e10f7
F20101106_AAAQSH caicedo_h_Page_190.txt
b9dede05224f668d81c2e9d49e29ecf2
7bd296eb9e2e3d31c90c0d5c3dca019e111847c4
7551 F20101106_AAAQRS caicedo_h_Page_157thm.jpg
f4d79daeea9c120f86483873f61efc41
65cef3a235191a36d7ce17b135605bc2b0eb7cec
8123 F20101106_AAASBE caicedo_h_Page_067thm.jpg
aa10816d15bc5d54d338401e1407f4e0
95fb9596ec8e48abe17a8cf7befd3cbe104212e9
8158 F20101106_AAASAQ caicedo_h_Page_057thm.jpg
308f3d898ad549cbff49e6eb2176a032
8ee19c6d7fe22a9798135387962580357c4abcb4
1909 F20101106_AAARVK caicedo_h_Page_128.txt
95eea41d8f3bfc2c0b12391e7711fd7d
77609074f293fee12397f8af62b54c7d9830f566
2189 F20101106_AAARUV caicedo_h_Page_109.txt
4607fa2f6eee14d28a14b5ac1b342234
d25ccf763b589541c5b32c5759184ab1041ebfbe
53899 F20101106_AAAQSI caicedo_h_Page_116.pro
114ede2ecd401eaed6474a1d4f389db7
d67a9279345a444db108f211fb04e39ce1f56c67
F20101106_AAAQRT caicedo_h_Page_037.jp2
5e77a9122e7c684435dd577f924fb024
e45568e925f44ffaa40d1c80ac3fbb9fd886b26d
31073 F20101106_AAASBF caicedo_h_Page_068.QC.jpg
daf086f6173cb352c0c65a99c3daf7a3
10d83dabc0a1df48119e27be402dd119efd7edbd
8732 F20101106_AAASAR caicedo_h_Page_058thm.jpg
4feca65506d8c9832e3b042bcbe59518
d87e126d58cd94f4a3e70e1efc941a5ab4eea9b5
1566 F20101106_AAARVL caicedo_h_Page_129.txt
67f05def870dc7f5ab57c4ba99081d34
ec7aae5e8bfb5549882343aef088ebb5a4a3ab85
2169 F20101106_AAARUW caicedo_h_Page_110.txt
3de79bccfc8740ce3dcecfe15b7bed53
88273b5abb775f0e5a2d8331a78f9df79d6af05b
F20101106_AAAQSJ caicedo_h_Page_072.tif
1f42633848559bf3606b8fa236a8ed47
c3afe4ed15373770b91099e8b672748b4a09e4c2
F20101106_AAAQRU caicedo_h_Page_152.txt
bd5a5edaf13a86a1bfdeb636abce3351
8691c13c62f9da219095ca4b7402d38cdd45fb3b
8209 F20101106_AAASBG caicedo_h_Page_068thm.jpg
4d8e4abc1a36f24bfc9b68263d548cdb
dcb7681a43570cef536323c4502acadb0e868cc0
29993 F20101106_AAASAS caicedo_h_Page_059.QC.jpg
4b97a7a2f3585e6037eaa09e8dc61129
c82b7fcf97673e900a8a266380a9138e92893d12
F20101106_AAARWA caicedo_h_Page_153.txt
cb09f4019116f055f11b2952e41c6328
b4c9d2bb30025d81538eeec17b2d5d5a7b4efded
1986 F20101106_AAARVM caicedo_h_Page_130.txt
b1ac2af34d555c5cd64fb2cb82b3cd26
20038b1a6c20571accf98dd9bc1555276551fc14
2001 F20101106_AAARUX caicedo_h_Page_111.txt
0dbd848b1f73b8784b5726e5c0c3756e
ce323193ba0aa39abd015622c625043d629d95e7
41901 F20101106_AAAQSK caicedo_h_Page_167.pro
0c4547ecba8734656286ed4f1070a09c
967dadc32da6a7f62ff253ce2978028a787ca939
47539 F20101106_AAAQRV caicedo_h_Page_143.pro
f7f6c4a4592fa8f40142c2320bf7d21b
f6dc598042e3cff4c77bf7986720e4ace77dcb50
33022 F20101106_AAASBH caicedo_h_Page_069.QC.jpg
831085491fb6530bf28ae11d5e97a8d7
9327387c28014d7a3013f7aecb61124f58bb5ad3
7787 F20101106_AAASAT caicedo_h_Page_059thm.jpg
34b4c7da878d83dd022e39f95068c62d
1b8b6e17d7d0e01fd77496da8443376a370d5228
1996 F20101106_AAARWB caicedo_h_Page_154.txt
2653c20e91a3dab96a7b775a53d4ac31
ac4fca2795c1af16d5ffe0eaa50c087d1c7b70f4
1992 F20101106_AAARVN caicedo_h_Page_131.txt
046801797dfdf6a8d240ee4630588435
e8656e2271b50b72dbd92eb27cc617671bcdfe79
2723 F20101106_AAARUY caicedo_h_Page_112.txt
d13058e88a283e20962c855c35c42b3d
a94c26aa1576072087ddc9c6df45078de1f76cbf
F20101106_AAAQSL caicedo_h_Page_202.tif
e556b443d4f823cd93d534d68b6f4df7
358ef094fd4009247b690b7af724aba4abac5cbd
1935 F20101106_AAAQRW caicedo_h_Page_194.txt
258bd34102e47b505dc12f0cb93c7b47
010b1d4a5aed69602dd801b06cdb43b6571b4bc8
8588 F20101106_AAASBI caicedo_h_Page_069thm.jpg
b86d970abfc82b1033361765289e3e19
c54446bccad5f20ff98ec684e04f70c1a7a49cee
34729 F20101106_AAASAU caicedo_h_Page_061.QC.jpg
c3d32e9e8cb3f73d3b8ca8cfcb3b1f49
98ee03edb6640d034aada26183ee0e9a69fd5a93
F20101106_AAARWC caicedo_h_Page_157.txt
5a8a9ac5a5469dc33ad000e9f7c2f0be
f347af8cc4fe107f9e7fccfcb362d79add42223b
2073 F20101106_AAARUZ caicedo_h_Page_115.txt
f53f0c932f1409544f74ff949b1a2bb7
324ec696890e9bef24f9f355cf33ab89cce4a77b
1635 F20101106_AAAQRX caicedo_h_Page_059.txt
d9d49b03acfdde12478a28995a29e17a
d356e8ccf998b1ebd16eda896c4735d99eb0f12a
42027 F20101106_AAAQTA caicedo_h_Page_126.pro
f350abd4b7f46bc2ec1d18742827ff48
e924250ab0f39150df18b1eb77300c9cca506ed3
31187 F20101106_AAASBJ caicedo_h_Page_070.QC.jpg
9a4c97f0d3fadacb05daab4a56fa8b39
99f2f06f42bfce26f6a1bef4fb61d889fc605a86
8519 F20101106_AAASAV caicedo_h_Page_061thm.jpg
fe80f6f04c77e4d04d4c11058d709b45
2290e02c333ab5878553820da65fa3dce5796321
1762 F20101106_AAARWD caicedo_h_Page_159.txt
8d7106c5ab0445cf5d15493882063a61
d212337716ae55db2bb9b003a1ab3dc7ca1e9e71
1898 F20101106_AAARVO caicedo_h_Page_132.txt
acfeea41bd5d36e52e3c6b94661b3e26
682aa82170ee92c7b281c25fdffe5605be7a9ca9
33726 F20101106_AAAQSM caicedo_h_Page_079.pro
07b466f1cdffcced6f72163f4993c5a7
6efe747f3060dd06d4676060db857c48e58f6337
105173 F20101106_AAAQRY caicedo_h_Page_169.jpg
33ef8b232c0ddef97c15d4301391860b
b9ca18b464015ffc9fc76dc994af1625570faf99
F20101106_AAAQTB caicedo_h_Page_060thm.jpg
0cda31b03fb36508cb844bd7230832d8
999f112db0fe05637b4ffab80f8cf74dc4874f8e
F20101106_AAASBK caicedo_h_Page_070thm.jpg
2bdec50f2507c95b6852d82885703d4c
695ad9e6d87e2f161ce5406028f5ea809381928c
36854 F20101106_AAASAW caicedo_h_Page_062.QC.jpg
e84d7859d7aecf51c9db40c55676bdd3
d673cba1dbd34daed0b06c072646e43b62e333ca
2178 F20101106_AAARWE caicedo_h_Page_162.txt
0a2bd5fb7940a6e85bf999c61f46b665
f9b59086ca425c2dcc80e8bdd15dff8eb0163a2a
2068 F20101106_AAARVP caicedo_h_Page_133.txt
705aa1a9e706af0b51946a745689f03b
bf1946d1e20087430461736a84e05006150db808
112735 F20101106_AAAQSN caicedo_h_Page_162.jpg
3e9ac8399de2777945cf5e20397b22e4
49cbc29b7e571dcf9efcc9e7bda4a83d8d26b8e3
F20101106_AAAQRZ caicedo_h_Page_197.tif
c5b2d5b6349d7077c311d8492c0cbd4d
77fe9788096abfe95d0100c43569c2823a4bbe65
49775 F20101106_AAAQTC caicedo_h_Page_068.pro
660b4dfefad6fb499c7d296129755654
c340b83408b77abab25219a5a0393d4d399a9ba6
27919 F20101106_AAASBL caicedo_h_Page_071.QC.jpg
dadef83f6afac8bc3fc5b0f2efde05cc
de2df702c903c081f7a118bf381c09759aed1d5e
9197 F20101106_AAASAX caicedo_h_Page_063thm.jpg
797b7b6bf8e8bf9dcd432954d1a3cb10
16834bfae1016562d17e57c43ca8b9b9f9af41a5
F20101106_AAARWF caicedo_h_Page_164.txt
3051b59ac4548191f44ffd121d640106
c51c1d148fcda0ffaa9bdcff1569a9e52042c58b
2014 F20101106_AAARVQ caicedo_h_Page_135.txt
ce56bc1cbbf1e50a186db7c1368740e3
578f009b59f2c35c39274ee4ac4db55642329756
F20101106_AAAQSO caicedo_h_Page_037.tif
735b19e800d5467ca2a574f5edc65483
c0983e66130661e1bf7771ee2eb9849c29d978a7
F20101106_AAAQTD caicedo_h_Page_211.tif
2da20d431e13495c274177f073c85b53
fe85d7836b036f0dc1ae26d08442bb171a57c587
29673 F20101106_AAASCA caicedo_h_Page_081.QC.jpg
639da437b994291f9e4e81742b5da637
3e3b986d77ab927384233635a7222a90c18a2e0d
7462 F20101106_AAASBM caicedo_h_Page_071thm.jpg
c66aa5fa741f0c5de0f57c46f8976b6f
426104a9fa6ef72ce0c4d7f7c1de8eaf959e4abf
F20101106_AAARWG caicedo_h_Page_165.txt
a2e77f5e0da5129fc28e3c49363a6501
60dd2683f75a82548d54df6e92d113d4cc6ac9bc
1875 F20101106_AAARVR caicedo_h_Page_136.txt
f3d5aeae42c945b7702f268b71aa3eb7
89b9ffb1861faa36573186651eea5766a7322752
1011096 F20101106_AAAQSP caicedo_h_Page_091.jp2
f9afcb76e876a6b83ca9153b7495256e
6075ecdf27e2b32e69fd596ccd30a9ae118dee52
45427 F20101106_AAAQTE caicedo_h_Page_080.pro
82d5b132a624b1402ae4fdbbf4f97895
42e5719dbff5d4bb0d7bed404a85ef8f4e68d936
8092 F20101106_AAASCB caicedo_h_Page_081thm.jpg
8bb18296fc7bd8298b1c1156d96e610d
de048e3d83534de377f80f6700534ab63dd5f657
33338 F20101106_AAASBN caicedo_h_Page_072.QC.jpg
7efcb942aa4854ad93eb7885eed88cfc
af96384d90b1d0d319380568edfd39e0ac181210
F20101106_AAASAY caicedo_h_Page_064thm.jpg
5f22811c6f8f06f66119e15fbe71ee07
a6b3cf406de56c5f4e41ce9f0eb4ace9400cba0e
1772 F20101106_AAARWH caicedo_h_Page_166.txt
37ec9c9ba11bd6601485eb58b0146138
07754d5559a19f9c2c46dc8441b6f3aa713db5c7
F20101106_AAARVS caicedo_h_Page_137.txt
9a80f1e2281890bcfb5ce1771f6ae3e2
6ddddbdb04b1a2a6f93a7ecd5654afca46c8a38b
2060 F20101106_AAAQSQ caicedo_h_Page_060.txt
292f7ee1427a325631d9fe01494743bb
eb667102e3b4c170eb6fe85c0c409558d46c81fc
7895 F20101106_AAAQTF caicedo_h_Page_143thm.jpg
e91423ddad80f3d52ac10ae023ec33d7
0655280edbb666f526ae8f723933d51f174b01ac
30692 F20101106_AAASCC caicedo_h_Page_082.QC.jpg
a99d3871c8dded47af44af99ec038f6e
3a9482a4338bf195e344e0e006f5076b374c865f
8282 F20101106_AAASBO caicedo_h_Page_072thm.jpg
6fb17ed2649683b35deb7fd79ab6aaf4
08d01dca9dd21577b16e1adda14c5be3cb1adc11
29888 F20101106_AAASAZ caicedo_h_Page_065.QC.jpg
2c188180002438cbbc91bf8ffb676336
cfe4adeeca2248bfb1a8eaff4e91c453381112ed
2161 F20101106_AAARWI caicedo_h_Page_167.txt
374e940658bf660b3b812b4bbb43d0cd
1c56e5e4ad94c85964b8f54c748c7698fdeb19af
2040 F20101106_AAARVT caicedo_h_Page_138.txt
c6c915519c3822f4945dba301462db1b
432ef3c979ceebd72b4c96346149fd9a6d5f0a33
8272 F20101106_AAAQSR caicedo_h_Page_122thm.jpg
98fb1162e650b7724fcb269cc44b01e8
a6e4f0f03c581c17190f933ffdd9a1f6f54a31de
91318 F20101106_AAAQTG caicedo_h_Page_180.jpg
fc0a1bc033c75866c6b600ccdb5d0fb9
302c51ffcf1cecd137f13b0da2a9e4e20243e07f
37435 F20101106_AAASCD caicedo_h_Page_083.QC.jpg
c94c98dbe561fa7ae8a1616677ec8e2c
a8e4c6b45f8ad37d4f51d316aa0824de93827d0b
29813 F20101106_AAASBP caicedo_h_Page_073.QC.jpg
6c35ae740f2545d971e386c166bdd4e5
05a7e6d02c04efecc2176af83e3db31c84339eb9
1922 F20101106_AAARWJ caicedo_h_Page_168.txt
9b243cbdd82aeb43940287a75bf72688
b6e93f2440d1f631217650aa2887857ebf045ac6
F20101106_AAARVU caicedo_h_Page_140.txt
4b6ab50fa602b6fa865078d6f0ce2664
9ce1c28e21dd67976cb843e504c7c47cc23e4188
79646 F20101106_AAAQSS caicedo_h_Page_146.jpg
fe0d4f755c3a92188afdef655500d216
5b816b2d1cf0733a3ffcc696e8ea3f7702d021f4
981119 F20101106_AAAQTH caicedo_h_Page_047.jp2
80e45576757577cd196c9d3bfe6f6864
f37d7ea3be99e9b2ea2d0f5b10a0cb7c5db30272
34057 F20101106_AAASCE caicedo_h_Page_084.QC.jpg
dcd8c35662d9b007449e4ab08987ed15
fb894724ec7b1b393d52072ef22282f810b39d77
33010 F20101106_AAASBQ caicedo_h_Page_075.QC.jpg
f1d86c41befe902b8cdae8135df92bc0
1f9e02e60244d21f4c01d7607b26eff3f463b476
F20101106_AAARWK caicedo_h_Page_169.txt
fc7e29224ad8f7a0f6d8c3e6d5a5e678
6cdb5f4118cc088ba45f30b7fab0a6b5b07b51b6
F20101106_AAARVV caicedo_h_Page_142.txt
67a15293131030e1d5b21abaedd4407c
8245a7e261b28273fc88652c7665072a7abf21a9
F20101106_AAAQST caicedo_h_Page_161.tif
bf5d1a5ba34ed05d79f32c798c8163d6
baa6ca6a05378c396f1fda0bf657af733d093ec5
F20101106_AAAQTI caicedo_h_Page_134.txt
93c2125eda14a0498c0684b3c941b284
ca5e10d11fb9c785360d6194f1a13e2c272c3084
F20101106_AAASCF caicedo_h_Page_084thm.jpg
ba4e8801fe7d9b8dc80ef46a18af9854
c2a034ac0657cf782d1218bfdd1c9ab1d97ddf29
8792 F20101106_AAASBR caicedo_h_Page_075thm.jpg
8e908faa4ed220bb0b754f4ee2dc65fa
1b31deb3b8d1542b9c26e1f8784789d6bcd6da12
1459 F20101106_AAARWL caicedo_h_Page_170.txt
73500ddc580a365660630d497a71cbfb
0d3f2c2e5e2305c140e1044a46339d9bdbed74ca
1272 F20101106_AAARVW caicedo_h_Page_144.txt
51b9a85b602aa1b30d1c89e293cb35aa
c70d85b8372d632c4a7b24c7f9fe6f2c15bcdc4c
1998 F20101106_AAAQSU caicedo_h_Page_067.txt
ec448c20ead6a2d6dec11c0c48e397fd
d4bd2f535ea7a3a5251b80459b336f53908e0f9a
1027470 F20101106_AAAQTJ caicedo_h_Page_133.jp2
28a0552b10b7ac617cecc9cf876116c8
a4556968303d39f9a0b9ae0c99c86aad2dfa1499
37299 F20101106_AAASCG caicedo_h_Page_085.QC.jpg
e0f4dbedf7dba701b3344dda31793794
a9305bc7023b9ebabcd6f2131a849e2a04b5efe8
29419 F20101106_AAASBS caicedo_h_Page_076.QC.jpg
c14e33f46e194c7493622a1931dc9778
edccd60f576e46d741784f4cdbd6586e75f040c4
1809 F20101106_AAARXA caicedo_h_Page_195.txt
d5e82f9421ba41090f17bcafee5607f6
3d247dd9c2876807c528eb1a97013976685b0ce0
2153 F20101106_AAARWM caicedo_h_Page_171.txt
a217f2d5cbe533fcb575e42d4907ff62
14557b41e6ef359159e9c2f25acc880200872579
F20101106_AAARVX caicedo_h_Page_145.txt
ea9d37a1adf4c9be8d64700cd8c07365
4e7b86933f60b5e4741572778ab77903fbc8251f
F20101106_AAAQSV caicedo_h_Page_076.tif
037befff1a64da5f70de0f4e46ba87c5
176911befc7368ab615d76a930152bbd16d7445f
1928 F20101106_AAAQTK caicedo_h_Page_179.txt
4c032a5ce8308a7eb55c1488ab6cbf3d
9653ca129f88e47da2c71f1b98a5b0279fbadc54
8914 F20101106_AAASCH caicedo_h_Page_085thm.jpg
ff32bbb156d27c6784fd668d96d1a9dc
1ea03a3708f0d1e832afe36020844b0c27bee4bc
7497 F20101106_AAASBT caicedo_h_Page_076thm.jpg
a2ff2b1135e47ae33a5fbc6d2974d546
fe2c697da35cbc1ad272436f5378d27031753cd2
F20101106_AAARXB caicedo_h_Page_196.txt
7dddc108d49c4ea8a7134e4d6911baaf
06d0ec0dffbdf0553156acf6151bb22f2e1dec74
F20101106_AAARWN caicedo_h_Page_174.txt
b996231f9ba060c4de0afc893f6e22e6
c562137dad1009ef34476ec747e114c774e9318d
1994 F20101106_AAARVY caicedo_h_Page_148.txt
a07b3d14284d093a0eea22a7f1bf71f2
4ef168ac25e1468ca7599a08ad3d80cafbfe27ec
29766 F20101106_AAAQSW caicedo_h_Page_134.QC.jpg
52159e54bfd489c094ad8fbcb1d22ba7
36417d4252d0e8201662d557817889c776eb9b3d
2007 F20101106_AAAQTL caicedo_h_Page_156.txt
faf74826f3c80f822ee97e966c754729
453bb3c24958b47e7149404f36630ae8398d41a7
21042 F20101106_AAASCI caicedo_h_Page_086.QC.jpg
886a22323db4c33e4f307be75a82f8df
3baeb23620e334d2c52612fcafabadc528e50cc8
36431 F20101106_AAASBU caicedo_h_Page_077.QC.jpg
3801a0a134d4ffba9df559e45e722993
8a6b3d5eb3ed0ebbcee475e9651e28dab20b7b86
1915 F20101106_AAARXC caicedo_h_Page_198.txt
6f5fa1507990204ace100613ad413efb
3c6b691a5afbdae1d9692a8e165cc7ad407f5bfa
1913 F20101106_AAARWO caicedo_h_Page_177.txt
f19c31c55cb8e681ce65606520cf0e06
0d8e6da334097be29089567827eb6b739699766f
F20101106_AAARVZ caicedo_h_Page_151.txt
7ea198680e830a8cd066fee1f9aea618
425136ba7adc4011cf19b4c21ba33daf6caee4e2
28091 F20101106_AAAQSX caicedo_h_Page_028.QC.jpg
ebbbbc771b3a3e7d009d344e7708c1b0
ca06f535a2292982b815915b8fe4fed0d2e0a3c3
6976 F20101106_AAAQTM caicedo_h_Page_141thm.jpg
c6c207ae5fa4c75c6d092d10000aaaf8
c99289692379cf2c67327f673f448b3e9c5df020
7841 F20101106_AAASCJ caicedo_h_Page_087thm.jpg
a8da85f3f6c68f095ca47128ae6b0b6d
81176e193ea902c7ed14a001e4e9a01a6a42c396
8752 F20101106_AAASBV caicedo_h_Page_077thm.jpg
1e33a2dc1df2101de96e7ab62b6f15e9
306c0436eab2b0a748712770f433eba1b75f669b
1843 F20101106_AAARXD caicedo_h_Page_199.txt
b30c18c8b5a9a3eb88fa7e01e0a56505
01c531ab8b3a5eb3fe35b266509f519fde73f0f8
101888 F20101106_AAAQSY caicedo_h_Page_201.jpg
b93fcd86284da4a8908301663efc4642
18b987f6da36209334f63d946e85b6237e58d76f
33780 F20101106_AAASCK caicedo_h_Page_088.QC.jpg
294d6cb5bad8d6331d22a3b1b23d4b34
0636842e9af564ba60b3e91f66b0320b15de4cc9
33781 F20101106_AAASBW caicedo_h_Page_078.QC.jpg
4a8fb6d34da5077e4cb877a79a3048cd
92a59087b3ea28b1c2a210496d0b55f03b8b7943
2250 F20101106_AAARXE caicedo_h_Page_200.txt
80759a3ea144c7339cb54293ca2fea20
819ba2894b5c833bd55e0853cd88a8787b3299bd
1881 F20101106_AAARWP caicedo_h_Page_178.txt
f3a171532a049805cadcc234c606e249
138f542b57baddde07fc3618da3978de0ff688b5
31054 F20101106_AAAQSZ caicedo_h_Page_032.QC.jpg
4abc6c64b9a513ff1803008a908be7f4
2f1a5f4957a04cf54bdb0531788f877aa96e85a1
27804 F20101106_AAAQUC caicedo_h_Page_001.jpg
94c2b8c3240dc43baee5b4cb36057cf3
3e07243c3222460ca87be54a7fc364963b28e165
31341 F20101106_AAAQTN caicedo_h_Page_212.QC.jpg
f8dee138b998ff76b63836ee66ef2fb0
4794f264d18c77ac45e33f14b293a80f1f7c2fc5
8309 F20101106_AAASCL caicedo_h_Page_088thm.jpg
fe146748ebfe618c209abcd6ff49261d
46f6444fafe2535b2c76fb3bec136f380066f598
8728 F20101106_AAASBX caicedo_h_Page_078thm.jpg
af2b2e517059c02cf593fac3a355c2a4
354d2f96a32b954771b31299dcffb6f622db88c2
F20101106_AAARXF caicedo_h_Page_201.txt
b1d78adfd94ae154cc0ee82edf0fd231
c6396985d8f2d6cf8853aac3c097a8c2eb84b246
2084 F20101106_AAARWQ caicedo_h_Page_181.txt
8f437189a507ccf3a0b0835a1db6e5c3
c5bc62d0d8a197e48a7fb3cb5c1e5c8b4843d8c0
4224 F20101106_AAAQUD caicedo_h_Page_002.jpg
038f2e19b99ce13f758a2a5946fab7fa
79a5e9e71b79c84e5adb1c1443ca64538aabbaeb
1654 F20101106_AAAQTO caicedo_h_Page_147.txt
4e2c8194964aa23bbcb97f5c32bba722
b1d11c77581ae7c05a61c1b521ee52ab1e3483a0
28830 F20101106_AAASDA caicedo_h_Page_100.QC.jpg
593af2b92127b679808b850229213198
d2070b096b36a0f75d7d07111b7b917a753e4b9d
33259 F20101106_AAASCM caicedo_h_Page_089.QC.jpg
89047f9630fda10e2843a09da5518dc2
4f3d4456ce0d9daf7a2267e542b54406042f6877
26099 F20101106_AAASBY caicedo_h_Page_079.QC.jpg
8dd16531567baf69c587287a3fcaed0c
5e1c245b85a15906c75dea09dd1e9017665faf34
1910 F20101106_AAARXG caicedo_h_Page_203.txt
2cca3c0f00e894290b1d3ea8656d6c5c
be36fdfdf1aff5a04f534762028d5cfaddaef6e5
2225 F20101106_AAARWR caicedo_h_Page_182.txt
38405d3f36bad420414a85f2fdac8e41
d90411e55d8d5223f44159c38ff2cb4d98808c08
7114 F20101106_AAAQUE caicedo_h_Page_003.jpg
c510451531ca0a8c5147ce1516a3194b
b9826c0d91db6d1c39da2c1f1a0d2fd9a56156af
F20101106_AAAQTP caicedo_h_Page_166.tif
21a1d74cfe37813971368c0210584ac6
f4afc197c6532ae140c039cfe52bf7ccc47b0026
7537 F20101106_AAASDB caicedo_h_Page_100thm.jpg
183d1562f12ac6248e3407eacf359ae6
dbdf62499a18f77e0eca6671b9311fe8a634fb84
8376 F20101106_AAASCN caicedo_h_Page_089thm.jpg
14ddd56d858561f90c6a259089d5de30
521bf1c65cc8318392ec61e2345ee0559acf3148
F20101106_AAARXH caicedo_h_Page_204.txt
e2375d1a255cf63bf25a3a8d58e29646
cdaab46e1b22d3ac20adc148c2d64e00583868f2
1589 F20101106_AAARWS caicedo_h_Page_183.txt
fd9456aa9eae0bc3a3d7502c2f7afdba
432ac8bdb87e1c9f8948be11c14c16137bb0c985
67931 F20101106_AAAQUF caicedo_h_Page_004.jpg
21fd367c6211f12e6a748a193a15c676
a1839ded18ca6ec9b84ef63031e0e4d9df33c860
8150 F20101106_AAAQTQ caicedo_h_Page_082thm.jpg
91ad5e16ddae68bbdb39883d3b07ec38
f59ac2e50baaa53830b707c92c930e3a47807c75
32189 F20101106_AAASDC caicedo_h_Page_101.QC.jpg
8e131039bc3c7ccce57d9d140753a217
d1e30334d0e4e46210b312c835929d42479a578a
29799 F20101106_AAASCO caicedo_h_Page_091.QC.jpg
f60cc0e90e7920b5c5d0e0c5a70141e2
ebb9f8e1168eaba9b9216a2ab5ec04739da6432a
7853 F20101106_AAASBZ caicedo_h_Page_080thm.jpg
aebb3fc7b7e515391e3341aed587cbae
3e54b4ab20074bf6858dfc4ba934d32e09bcf207
F20101106_AAARXI caicedo_h_Page_205.txt
3123e0b4aad1a7d295c1455a498d7e5b
32f7c25b91c01812f05f4dab2804c7fac0979a92
1776 F20101106_AAARWT caicedo_h_Page_184.txt
00f64f2fb530b2c1949e77b03aab857a
7ffae97db9185a33a2c2f191009136dc867211e8
124200 F20101106_AAAQUG caicedo_h_Page_005.jpg
72793911f19539fa34b5221188ba5294
272f17aa3058f201c09f06d3ae84292888087f76
40329 F20101106_AAAQTR caicedo_h_Page_159.pro
4a24891a23cbb3ed550a609cf9882a23
8678ec08580883394b4d621747608c3160122341
96114 F20101106_AAARAA caicedo_h_Page_212.jpg
94db7b04388138b1c202e4684bf79e01
745a503e02eb90972e2fe7fd0ab948650f080392
7280 F20101106_AAASDD caicedo_h_Page_102thm.jpg
0fe8e91dc0614a1f0ce19e557654390d
82c9ad30c0f09a36a2a98c6e4306dc87711d96ee
7608 F20101106_AAASCP caicedo_h_Page_091thm.jpg
1b2789f1b6a36a6034df233f51662742
6ae3d9f0dc970908110360ec33e9547c9c73d027
1912 F20101106_AAARXJ caicedo_h_Page_207.txt
95c86989e869e1e812da3d8502b333d3
3152ca615185a8ed26adc2d409aee8ce770bd98f
2524 F20101106_AAARWU caicedo_h_Page_185.txt
37553efb294d7524d40a2eb9cd1c6648
248cdcfafb616149e9422c3cabec37162378c210
173487 F20101106_AAAQUH caicedo_h_Page_006.jpg
9fd134e4e2708d1402edf0834dd93b27
7f219719b037bb359276d815591decfb4d674702
1051947 F20101106_AAAQTS caicedo_h_Page_152.jp2
b11f38dddb209df7a42372158d410fe0
4ffd0fdd46d0b993c75629be473f41da54413782
264741 F20101106_AAARAB caicedo_h_Page_001.jp2
80b4d0c01fcf4a7541ba809d6ef2e7f5
f6fee85819fdb680da093869b16e7d6e476eb194
27392 F20101106_AAASDE caicedo_h_Page_103.QC.jpg
37064cdda00969a6cc62d1daa1078348
81156fd4601e33d374755fd857a8ea6cc954c1c3
7703 F20101106_AAASCQ caicedo_h_Page_092thm.jpg
68e08091fff9a2491dc0b408324b1403
6690978baab9966325814f0f9d02bbf38ee3b4b6
F20101106_AAARXK caicedo_h_Page_208.txt
235dd78a64d2638b026600925dda19b8
231ccaf12a7cbbc2828beb1cc964213a04a601a8
F20101106_AAARWV caicedo_h_Page_186.txt
be5f55f58a5323ec84d0cde92092d8db
0187d944a45264758f6792d06a0b5c04e61b508b
165843 F20101106_AAAQUI caicedo_h_Page_007.jpg
a5f2287a350c7432489f71da769724b4
cb3583f9a6c6f9f6ea955489b436c9e7e5c93a90
48006 F20101106_AAAQTT caicedo_h_Page_051.pro
9a85de597536cf358839bb805a5fcdc2
ee568f2025f8fcedb9ba35368f9cc2ad9ab8af1e
29722 F20101106_AAARAC caicedo_h_Page_002.jp2
435fef0e05d89c71991791c9790231a3
711003fe52b29e29fc2bb508441f0d79a182dd62
30739 F20101106_AAASDF caicedo_h_Page_104.QC.jpg
3569686e038e14ca4d72ae218e1c227a
8dbdc90df967db773c9afbc10982c8dd38fafa86
30925 F20101106_AAASCR caicedo_h_Page_093.QC.jpg
c96d7a5af60de71545a11c4f5bf648ce
d3956cb6e1449bee21951d48995acf61891812e6
1947 F20101106_AAARXL caicedo_h_Page_209.txt
208e173909d933240523e61baa78e149
8ec1706394bfdfd9d5eb78e3122f321135a30773
848 F20101106_AAARWW caicedo_h_Page_188.txt
40d3d0a58d6c79712140c168cebd614b
0e64f4fa009ff4eaccef253a0ec2eb47637fb894
40403 F20101106_AAAQUJ caicedo_h_Page_008.jpg
2c4515d16729e25b88c807e35c2aa020
663ea1ebdc616783fe8f7e9b6ac618e9c41b4f70
2143 F20101106_AAAQTU caicedo_h_Page_149.txt
bde2622eb62ea0702698e2b5531f0f27
85dc53a316ed522068a5444e533cd32528e5b13a
722063 F20101106_AAARAD caicedo_h_Page_004.jp2
abecee9b302c18791608cdd1c005d92a
b9ada76381e158b8adec03a09d7ff2ad7c91c7e3
31240 F20101106_AAASDG caicedo_h_Page_107.QC.jpg
c4f7f761a903fbd5e710e50644081cb9
10d252d4de54700820768bff1fbde29b58ddaf2c
29884 F20101106_AAASCS caicedo_h_Page_094.QC.jpg
78618117bb8fe033a7b48912a88dce2d
f3b6f45d30e9bf2291ffa51138ab8734e135820a
8611 F20101106_AAARYA caicedo_h_Page_007thm.jpg
506cd6c75b551493cd51d25c5714d430
f8c7a468e0e9efc1c4a9bb8da69e08f2b23f0a97
1854 F20101106_AAARXM caicedo_h_Page_210.txt
129e451a77df7fb387fd5e221d51a8ce
4f78e2e18d760a5a0a8f8eb4094c19bd2dbbdfa5
F20101106_AAARWX caicedo_h_Page_191.txt
dcbcc387b9f884831fd5df359ab3a1fc
3a60fb60eeebbcfb6d599edc7a2a33fbd0dd1176
168580 F20101106_AAAQUK caicedo_h_Page_011.jpg
3114e15da881812f6b93df50d14e263a
1797b2fcde5e55447fef8bf72743ae8f750bf9ad
F20101106_AAAQTV caicedo_h_Page_159.tif
10bcd389891fe1de3b878bdbe2ab8ff2
804d798b3e9c936c28ee78911325c99cccc186f2
F20101106_AAARAE caicedo_h_Page_005.jp2
f90b12fc558cfabaadbbf90624fccd3c
a35c55e961ee3dfb5a3fdb422c04933f4bbf1a70
7451 F20101106_AAASDH caicedo_h_Page_108thm.jpg
93441fe4ab886884828421c5775c74ab
b6327e340befa099ec367c433cfef4dfb9993d91
7917 F20101106_AAASCT caicedo_h_Page_094thm.jpg
aff3378419465d44d7ddd2a5f03f83d9
28d500e1804971c78990d99ba18111d060d9b99a
2696 F20101106_AAARYB caicedo_h_Page_008thm.jpg
1b7a705180004f7cdbe6bf1594fa0a9d
f41c45954fe8f916ce69d2524b37dcd6d76eea8b
1853 F20101106_AAARXN caicedo_h_Page_212.txt
b1111d0427b3680074a0f72f3536a6cd
c3a09211d28a12d5b3cb4d7c251d0125e726683c
2088 F20101106_AAARWY caicedo_h_Page_192.txt
564adefbd50c71ad91195ffca2535de3
cfdb358269df31a24b456cf057736357220f7978
161175 F20101106_AAAQUL caicedo_h_Page_012.jpg
c4f21ce41400ead905eea1395aae49bb
62faffff92cac76479f422da7ebb9df805c4406e
21120 F20101106_AAAQTW caicedo_h_Page_018.QC.jpg
042c09a1b9be2ee0b55c4f87a75f4c01
dd53f0d30de5ee6485dc00dbce7e3ea304b68f33
F20101106_AAARAF caicedo_h_Page_006.jp2
110781e337c3da897f6ec94fa60b615e
94e52cda06b6173ae299c066f11f12c6bd12dd09
7627 F20101106_AAASDI caicedo_h_Page_109thm.jpg
a461b9e44a68a879d95acea4efbe7565
4cffdc8d9d8442d6cd5c237412fcf8f58ea0ed0d
8097 F20101106_AAASCU caicedo_h_Page_095thm.jpg
f7ac4c28fea96141cd34fa62ba4c9847
32f7a76a688994f28dbab9b217a3e2209b863e64
42788 F20101106_AAARYC caicedo_h_Page_011.QC.jpg
c72b0098095c69a1acc569e44757f384
24aa86fb33a6ae939ebb8650978694015837e8e9
2146 F20101106_AAARXO caicedo_h_Page_001thm.jpg
b78f91a4c9e480e45bb4320526fa8ad3
276885e12c5529e5f81125a2488f99c45f52ad09
1336 F20101106_AAARWZ caicedo_h_Page_193.txt
8cc157cd91f666dfd821ea364843fd0e
1bdd935bac484dedeb88b956ba479a09ec376188
85870 F20101106_AAAQVA caicedo_h_Page_030.jpg
806ad1a9aef9f37cbe5edf20d23654e4
ffb57695e85c4e06f58dc3998387bac5c48dde15
160766 F20101106_AAAQUM caicedo_h_Page_015.jpg
e083d096c8c9c146f19dd294a0b01370
d527f38202f8c65f649b4334c12409c3395a15ad
8177 F20101106_AAAQTX caicedo_h_Page_048thm.jpg
6ebb754296ee59899a3cd81a8a4578ac
4a392ef80dc4412dcf8200720dccfe166890000c
1051974 F20101106_AAARAG caicedo_h_Page_007.jp2
2455fb29f198c0875fc5181fda22fc3b
532d491f926a4b8f90279f28c693a732b9a543e6
37316 F20101106_AAASDJ caicedo_h_Page_110.QC.jpg
5d603dbad845254d315f0094c22752f9
8873fbd9239fbfa8e1f87b4c17329679c72e1173
30687 F20101106_AAASCV caicedo_h_Page_096.QC.jpg
5b341403e612bea9a7528f7526ee473d
876b8e7fd0cf03b725ef2442f75a2ff11573402b
9986 F20101106_AAARYD caicedo_h_Page_011thm.jpg
fe04b818a3cd91b7b3a60165521252de
f7cc3e4910b66b91dd5399f56a13edb76ca33727
27874 F20101106_AAARXP caicedo_h_Page_042.QC.jpg
0fd730896b0f43a0e601819a2632654f
3b2ab0e0d3ca287dad7d3d74b95c1df3ce3802e6
110818 F20101106_AAAQVB caicedo_h_Page_031.jpg
c951c697ead0c0c2e0156a9f1125aafd
bfad68656252fbfa05e86bf8e9fbd3a312ec4271
123845 F20101106_AAAQUN caicedo_h_Page_016.jpg
eeee3f0d36af63f12a37d252ee6b97b4
be909692afa8fb61bb7b3ed37eaa8551f1022676
F20101106_AAAQTY caicedo_h_Page_018.tif
d3c6a33b0613fd1e7e98d4e33b67722f
507edde3ce808657dac6a6c2dbb6bc9d4fb3d65c
F20101106_AAARAH caicedo_h_Page_009.jp2
60d88e92a90a7d51c2144adfc9041132
575798f348c6b708a2e2504e529ecefda1ac6665
8646 F20101106_AAASDK caicedo_h_Page_110thm.jpg
aa4d7542e85b7013aa907e0e57a155bb
a23e3db9b4878b2347c8f06420b169f4f76a2bb8
26137 F20101106_AAASCW caicedo_h_Page_097.QC.jpg
943e638bfb026da51779c67f5e1b4e80
6b1d2c5cf24d602d489a83d1889f343e1973e3d2
42875 F20101106_AAARYE caicedo_h_Page_012.QC.jpg
3b586c1153549e9415be15cdcf2b131d
20e546c667e185c91e0e5c84a185059a311eb33d
96048 F20101106_AAAQVC caicedo_h_Page_032.jpg
2a51b5dce4e0b17f1b94776f1703fd8c
93b89786d559b3a17ff1a5db3cb254d0a336df23
244901 F20101106_AAAQTZ UFE0022016_00001.mets FULL
3cb16d2bab1bebf07cfbc8370f4831bb
0691e8f85c6d104c9b3bb916731908242d202a05
F20101106_AAARAI caicedo_h_Page_011.jp2
70fc24bf32ec8387a4daa3c5e9b259d9
aef5f272c48d4c9ec66f0d98984ee37db92f08f1
28750 F20101106_AAASDL caicedo_h_Page_111.QC.jpg
c3b33c98e9f6424c01ad68200040308d
dca54998d418843d1496e787ade9348abdb7f912
7078 F20101106_AAASCX caicedo_h_Page_097thm.jpg
95ff8b13cca8445b7deea3d6804143f9
62c8976a917e18cf6d2290a9c80044e96dbaf716
9985 F20101106_AAARYF caicedo_h_Page_012thm.jpg
a3eccf24e7ebb3e0d0fd4ea0ba6dba5c
cf3ca873b62136c3330e6bd79da3e3a6de381796
317298 F20101106_AAARXQ UFE0022016_00001.xml
908870f7555bce09f205760c4d947897
f6db003c858d7366bea649eb302233f9afc1f1a2
93444 F20101106_AAAQVD caicedo_h_Page_034.jpg
8931982c3b41a0d0f2445353642b4764
adbf4ec1ffe8d5aaa1134c101a935e7afc0946fe
101336 F20101106_AAAQUO caicedo_h_Page_017.jpg
85c61a44a71ed926041d7a132fba6e00
de0b97971a2645afc2fcac7b024b8eefad7a811c
F20101106_AAARAJ caicedo_h_Page_012.jp2
2c325601b15797cfd4a959cb2b2e0361
78264bbeba13a6e128b3943f7feebc911208dbea
33443 F20101106_AAASEA caicedo_h_Page_123.QC.jpg
65befedbe1fdf18e8f487bd4102d9879
3016bd08bd9745d58ec79dd1dc62bac1eef0d9b8
7437 F20101106_AAASDM caicedo_h_Page_111thm.jpg
5f76d724fde73ae5fffaf6610a55f4c5
cdef69e5136e52e669ee3ba1ca0580846b63f761
8918 F20101106_AAASCY caicedo_h_Page_098thm.jpg
ebff7829804a581dbfbe4401614e5af7
bae83a28e0a7687f1f0cae4ba88740980e1d432f
42871 F20101106_AAARYG caicedo_h_Page_013.QC.jpg
1faf016291e344162c5669828380f0b7
322fd27d0fd31e951754b4a443d0d5b89927fdb5
1237 F20101106_AAARXR caicedo_h_Page_002.QC.jpg
8d3412f9ba7c9185970564551ccef725
9fa6b251e26f540d79d70ad06030f51f438b62cd
94065 F20101106_AAAQVE caicedo_h_Page_035.jpg
981ad81336c7830e8b1a430d4a5e2c16
995200c1f828fea2d14b98cf23af0de7285506ae
65351 F20101106_AAAQUP caicedo_h_Page_018.jpg
89e850af9fdef6303c0c132f9f89e7be
e0b20acc16bd59d884c37e4eb17c910fd5764af3
F20101106_AAARAK caicedo_h_Page_013.jp2
ee25e82f631e49b509a40da0fb2325e7
788be49ea085104538f24a081b03064c91259e7c
8549 F20101106_AAASEB caicedo_h_Page_123thm.jpg
c43e6bd5bd10adec545cd7ebf10d8f4b
f3b6f25e5d2b96009b8ce58bed21d24531c0aa98
31322 F20101106_AAASDN caicedo_h_Page_112.QC.jpg
06a343fd70529a9a4911957f197f3684
049b96e0b4f745a509c1e26b12790b2275cdd16b
8675 F20101106_AAASCZ caicedo_h_Page_099thm.jpg
17b661e2459992b5a44aa6affd45cc5c
cc4ccccaf8c7c3fde208169a5c06dd8066823d8d
10164 F20101106_AAARYH caicedo_h_Page_013thm.jpg
c31ddc169f899d394dd2c139e695c120
8077dc98c2ec3321856bbe05db5187c8abf4b980
639 F20101106_AAARXS caicedo_h_Page_002thm.jpg
e1ac75878733875a811ae832734f6f62
f318eacda8603db18fe5138371c4a0aa06005de0
109427 F20101106_AAAQVF caicedo_h_Page_037.jpg
ff3b467eb2dafb49bb0447dcce146ebb
536720edab6efc6a329ca72fee7815a785d09e12
110302 F20101106_AAAQUQ caicedo_h_Page_019.jpg
127559d7e85af35039315862605ae3fc
8ba169ce5787db2e1df2c525070b54ff6064d4cc
F20101106_AAARAL caicedo_h_Page_014.jp2
0aa93fe86bb010bb312c7761987eb299
b4f09fe271c1b810afc3d2c15af0d4ce48f606c6
34903 F20101106_AAASEC caicedo_h_Page_124.QC.jpg
7010ad3ad389b93431c7da96eab2a055
ba35c77f6c237f3c3e0261dd7aeff3cf85c36437
35053 F20101106_AAASDO caicedo_h_Page_113.QC.jpg
9b9315ec0cecb30e089529e375096dc4
4a1ee1ddb1578fa8f2876d3447b223a9cda37c2c
41852 F20101106_AAARYI caicedo_h_Page_014.QC.jpg
5c4cb0e73e492a854fcac52a5da6e287
9b53e12b22876a2dd02f42810a35f308acf3f74e
F20101106_AAARXT caicedo_h_Page_003.QC.jpg
2afe788b96a52859aee5dce92cf387f9
7ce0b3a89d14cf62c95389a01769e81b2ef07cf2
87558 F20101106_AAAQVG caicedo_h_Page_038.jpg
d54c84d4dc20d146747db45cc8919ad4
05a5666199b949ea5b09dbff221e75f1437e80de
109684 F20101106_AAAQUR caicedo_h_Page_020.jpg
f924155852eec7c91812d55798daa37d
93331ef1c3f70bd352638bbd64fbf643c1676159
918966 F20101106_AAARBA caicedo_h_Page_030.jp2
e57b3cdc4d27fe04127849302b7e286f
53866ead777a076b201b6d26da812f3782122c80
F20101106_AAARAM caicedo_h_Page_015.jp2
688f8600d9ec3ad5f365ca55f342abc6
0089caabd113400e19c538fdf754f802624660c1
8442 F20101106_AAASED caicedo_h_Page_124thm.jpg
e6f9acf363a06423f403d308e66fd3a1
d50c7d27f5946d94632e1222844490108320d429
8604 F20101106_AAASDP caicedo_h_Page_113thm.jpg
683bbcb98fb1470b3b9ff844fcbb8be2
ce01f78782d36303efd0ef71453e0e5e151f43c3
9753 F20101106_AAARYJ caicedo_h_Page_014thm.jpg
e1642ffe0e6570ae982fc88e6bd1f0ba
020ec9dfedd4ad24d2e94de328b8f3b5a71cbc87
769 F20101106_AAARXU caicedo_h_Page_003thm.jpg
b07fa71046cc2075b9169441054cd828
c13248f2a02d195383ed34e9d6a880318ed495ad
90293 F20101106_AAAQVH caicedo_h_Page_039.jpg
33c4b44ac49a9376404ffb61643b8d75
b10c45b0050792b194628d419df2e50a80ea3d72
102016 F20101106_AAAQUS caicedo_h_Page_021.jpg
c14957045a7e0a13e2390f57791817f0
0cb409a6c0d84e3bc33ec8854b61de48dfb423ff
F20101106_AAARBB caicedo_h_Page_031.jp2
fa3267c8d1b39283402d6972bfcc9ae9
c26cd23e925d0a949d965af5481a1202fc4ce488
F20101106_AAARAN caicedo_h_Page_016.jp2
126f1a28052f53c29d856edcd64adf1c
2a4d8d8f9d471afc7932b5836666973236963e4f
35738 F20101106_AAASEE caicedo_h_Page_125.QC.jpg
91ae229f590fce5092461f1b780f0323
a76f7d717d6204a303b685bf03e8105eacbaf7a7
31438 F20101106_AAASDQ caicedo_h_Page_114.QC.jpg
451a50ea7dac70a9d4abb3cdd81e79bf
130aea65b825723cb469888aa7d73c31dd899fbb
42347 F20101106_AAARYK caicedo_h_Page_015.QC.jpg
b2219662b75485b16fe419d738db4e37
c321705765601d136ee8ce08111cc8adb25594ac
21784 F20101106_AAARXV caicedo_h_Page_004.QC.jpg
e0629e8e42cacf48e754946957c68c61
b54ae2d7e9fff8f57cbe03500d5cd641391584fb
109019 F20101106_AAAQVI caicedo_h_Page_041.jpg
7983aca80b6ebaa1d01495e90c96259a
72e0e8ad02d782c05cdfbe175a97396e85edac3f
107654 F20101106_AAAQUT caicedo_h_Page_022.jpg
fe07af948510cde0df9796462cdd3377
5911eb4bf1452d7f2882d001e0c7032862523cdc
1036406 F20101106_AAARBC caicedo_h_Page_032.jp2
7e4f3f668e6062b78dc0aa6ea61dfe65
72b272b010484e0b47361006f125541985a82574
1051950 F20101106_AAARAO caicedo_h_Page_017.jp2
179155aa0ff2b71d4626baf124bc8da3
89bb8929d22e09c215d8bd7c43c472fbfdb649ae
8632 F20101106_AAASEF caicedo_h_Page_125thm.jpg
324ef5f78abdbfa1d4ac488a7544af8a
2a3ef2961b23a3fbfe932a54402eac6a2aee8038
35063 F20101106_AAASDR caicedo_h_Page_115.QC.jpg
d98f66c8cb1f9502616f41ba12853d89
60e32c39918281c586ece435e8dbb55c87900635
9837 F20101106_AAARYL caicedo_h_Page_015thm.jpg
55f39639904a78678b61b71fe29954ff
285af231cb664984a5217ea3e8fba5bc523b9d2f
7061 F20101106_AAARXW caicedo_h_Page_005thm.jpg
5c03183247a9fa0f2635cc5af8b4687f
d5ecf8b63c336a0fd1114e1cc7fd77ebb76765ab
106745 F20101106_AAAQVJ caicedo_h_Page_043.jpg
842e11527384b113b3253613448eb400
b5cde178dbd472d381e3fcca0f10a9b1b4f4f740
103021 F20101106_AAAQUU caicedo_h_Page_023.jpg
f1608b6de1b9deff3502555305d92e97
ab499d8c939c252717e864516ddcf654ca3eefd1
F20101106_AAARBD caicedo_h_Page_033.jp2
9eadd088b7dc53c32ba815c6c2afc03c
64f095d4d9bc1c354663aac30061a4ca48056151
707126 F20101106_AAARAP caicedo_h_Page_018.jp2
fffb6d1124b74f97bf565cdcb72bf6e9
270b41985f341ef9b347362b773e0074e51cfc4d
29253 F20101106_AAASEG caicedo_h_Page_126.QC.jpg
3b9f5293432a9495a17e8ec01bb07122
4023679620b5d638d1bcd9a4f533f870713439e4
35522 F20101106_AAASDS caicedo_h_Page_116.QC.jpg
c95a5e4accc961e8e45475d5c314a8cf
af93c9702b1a549548b397e8a962beb2c5859bb2
37746 F20101106_AAARZA caicedo_h_Page_027.QC.jpg
f2428736129bf4d33cc1e26e84099d77
e81dc0e1cecdc688983273c9f8c497f70a572fec
7058 F20101106_AAARYM caicedo_h_Page_016thm.jpg
e7b0695fd5577e2dc2c21fcb7d563fa9
35f43104a930da3d239a26c64ccb942775871e4b
39563 F20101106_AAARXX caicedo_h_Page_006.QC.jpg
db84ab5c960baa9127943513a634ea1f
532a4912423f7a00b1d5955708174dbf243e7250
90962 F20101106_AAAQVK caicedo_h_Page_044.jpg
16dde5df3983b3e14dd064a0cee22cb3
8bc5fb3c42ba5577793c8675d28a44817cdd9f80
106114 F20101106_AAAQUV caicedo_h_Page_024.jpg
1ea9826e2345cd25631d54828499c522
056429774ddf848f9b1e6f33b320a29d4a08160b
1007328 F20101106_AAARBE caicedo_h_Page_034.jp2
81ccc3dd0e18a95f2354974468b6bea1
29fd0989e10424ea9bbf974926650dc5a263eaf5
1051955 F20101106_AAARAQ caicedo_h_Page_020.jp2
7b25b1fdf36fb5cb52bf21a408f4af86
b2a48845aac1fb67b4636bd1f48f3bf72265e81b
36394 F20101106_AAASEH caicedo_h_Page_127.QC.jpg
6a9d1b3d5ae745538236d120cde1ecb5
2ecbffc2f8efca9ef0c722ab6b92f3db1ae9a2f7
31015 F20101106_AAASDT caicedo_h_Page_117.QC.jpg
802ab58d43bcf34433974807cf2e861a
0a6895690d600b69808d79a22158458ba50591f3
8667 F20101106_AAARZB caicedo_h_Page_027thm.jpg
43f2b9d972c52eaa289c8df57c4d11f1
48d7f7720a88cdf9046a41a3f4580743d3caba8e
30992 F20101106_AAARYN caicedo_h_Page_017.QC.jpg
1659e11ac415cb58643401c6589d7aec
04b53fbf754496f5a72510064610a065fa98a78c
9158 F20101106_AAARXY caicedo_h_Page_006thm.jpg
40cb98e7ee345da802d84de0a86f655b
fc4a9b57fba4f17c02225ddbca785f465fd84145
99262 F20101106_AAAQVL caicedo_h_Page_046.jpg
cd4fce82ea532b5dbf055f15144bca84
5c209ae85a298dc04668f93fdd06e3893ec4d672
85866 F20101106_AAAQUW caicedo_h_Page_025.jpg
45f63d0a958125e6016e274fc1de780d
3127beda5b8968fb4b449d808cb3b481687c5882
F20101106_AAARBF caicedo_h_Page_035.jp2
58bda8e95b757ab33e1fdb9cc553ea1c
5b026d0679456a99ec4446f31c3c4f0ebda54a20
F20101106_AAARAR caicedo_h_Page_021.jp2
b5f7a7f7206cc4f59b62d64d42d40f1b
f8a14e97ef47e2c2b27dd1cb539c413871cd41c5
33364 F20101106_AAASEI caicedo_h_Page_128.QC.jpg
dfdb3207b2b612e94b683b2d3e69819c
b1ab171356e9bc4c12af3fb9d1ebfaa93ab7a700
8295 F20101106_AAASDU caicedo_h_Page_117thm.jpg
1027aa68b4e69637a6a7d1db5470e9d6
981a24a3e4a6d3f003aa2d437a1dcefdf24b9079
34185 F20101106_AAARZC caicedo_h_Page_029.QC.jpg
5282b7e0b8bf03928ea1c21c494935b0
89f9d53e6774278c422348f531d2de9929599f17
7942 F20101106_AAARYO caicedo_h_Page_017thm.jpg
d638ecc30775955fa2951872e6b78cf7
d1e30955b02a26fdb459c9252601ff1bc948d80c
36812 F20101106_AAARXZ caicedo_h_Page_007.QC.jpg
96ba92868a55e0b782ceaf640b75dbe1
67f998e1bfcbb0ca4fa8e75563704bbefa1c23f0
96441 F20101106_AAAQVM caicedo_h_Page_048.jpg
f2c0948ca61231c33aeb40e1f41702a2
31a3f2389d8bb3a9b9452430841effb5bfda5588
93432 F20101106_AAAQUX caicedo_h_Page_026.jpg
754df327679e397af90668a9a6549ff1
8ef4078332c690e27e3ceae2b1bee81087fa84a1
F20101106_AAARBG caicedo_h_Page_036.jp2
c3150a56063cd8fdc08183ac03866a6e
bb4f594036e742a3fbb0c6e8d5301e14a2998bda
92396 F20101106_AAAQWA caicedo_h_Page_066.jpg
d0dbe9170a3631e6f0d8285f54254a4d
69701088424f14e1b5872593c93cb0247ed4360f
F20101106_AAARAS caicedo_h_Page_022.jp2
2a16aa06e40b3da216d795d033e20179
3c9f66c1d92fee467f32a3535c4f4921487bddb7
7965 F20101106_AAASEJ caicedo_h_Page_128thm.jpg
9882d56668eb411030564a78d009fbfa
9671d274e012df40e6a2e6ac4cd22a0555acad6b
8784 F20101106_AAASDV caicedo_h_Page_118thm.jpg
f7d97916c5d5d91af68327328b60ecfc
bf2e9aa258a1058ac9d0e084ece567b995638cb9
F20101106_AAARZD caicedo_h_Page_029thm.jpg
af45386cbbbd9f4544d943e91974d142
e5c14f63f692d00f8490072ad3babb6126b56070
5112 F20101106_AAARYP caicedo_h_Page_018thm.jpg
a9d850a44d2db6c6e3a3b48735792abe
ab159719f0ac98effebc87a95dc3b949d72de4e3
89235 F20101106_AAAQVN caicedo_h_Page_050.jpg
6da8e12734468483756dce26033471d5
efb9286a6c1cc7c9aa7487c45253824da7c8c05d
113912 F20101106_AAAQUY caicedo_h_Page_027.jpg
25e2d800109a26b2f8ca8104267d893b
1944bf8fa612df31b6c9559e6d29a06cc6d05a31
974918 F20101106_AAARBH caicedo_h_Page_038.jp2
9fe7021c6fdccdd0ae3d59c5d5d5bce6
6b644bc439e1646ec106f628aefbc4e57626c12f
F20101106_AAAQWB caicedo_h_Page_067.jpg
462adfc6dd86ff1b7b06135401f10dfe
ea871c460df208cd5c33e6ec4dcebdbe4f30d107
1051904 F20101106_AAARAT caicedo_h_Page_023.jp2
8b20a983b7b7d357fb624782f0e087ca
6136ad37a27a3ad806a25f558da621e42b75ee43
29854 F20101106_AAASEK caicedo_h_Page_129.QC.jpg
03e7cecc24a21beb688e38727366f866
63bc0e1babc7920b74c43b35839e80f6bdffac95
33818 F20101106_AAASDW caicedo_h_Page_120.QC.jpg
b9f6c9a630dd9f9a0c60a5565124a593
5bcdaa97df44da267cf7021d7e9c7a70ea5bbbac
F20101106_AAARZE caicedo_h_Page_030.QC.jpg
fbdad59344680251df93590c36b6c5ef
b3f7bd929b8c08a3dc00193f147b0491d0ba5982
35925 F20101106_AAARYQ caicedo_h_Page_019.QC.jpg
c63ae7da2d2a46d0678c0e33219c9f5b
d5199a9ed222248cc164f81012fd23a5a846b14b
98143 F20101106_AAAQVO caicedo_h_Page_051.jpg
d6367f17323c28d47b26e2c3ca883a73
dcd03e453658fc1fd527bcb50652e35d41be98ce
105670 F20101106_AAAQUZ caicedo_h_Page_029.jpg
045c8386f5403354b77caa1a27ddecd2
a1ee7b4340128034aaf3cb41e59a2dfd489ff399
971875 F20101106_AAARBI caicedo_h_Page_039.jp2
8701e1912671b5846fc9a41155bc0100
388d800dda2c1b3217b0fc75a304ac2ddbbb36bb
94876 F20101106_AAAQWC caicedo_h_Page_068.jpg
65543e46ab572e2b52e14bd67c46b312
0cf57ed1a4c9d3ff301c80af1bc5657c58abff3d
F20101106_AAARAU caicedo_h_Page_024.jp2
3f9a60a06aebe415a192ee5a21914b6c
de6dee712109de103b275552985ab79bc44a63b3
7761 F20101106_AAASEL caicedo_h_Page_129thm.jpg
b84e91ce7a07b7f9e82272f9427e6af6
187d10205f3b8c6aaba0ae8b32ce835c8c917fb9
8630 F20101106_AAASDX caicedo_h_Page_120thm.jpg
43e5437fb8af96d50fdd0430e5ab9c0f
c261e249f920941dd4b69b0778fb0c3aea57f8e6
7249 F20101106_AAARZF caicedo_h_Page_030thm.jpg
c85d8005a6db5ae9b62b09a2b96108c7
32d7b136870c16e375584ccad8d178184528da4a
F20101106_AAARBJ caicedo_h_Page_041.jp2
9363f080e8751655a8a25710faaadf78
c10ae52960deede43431d2b046436e2490b179d3
105465 F20101106_AAAQWD caicedo_h_Page_069.jpg
c2c3b7fc8e063bb750a9b1b55524d8fb
687c0b51f71ced8130bb544166108c5e19626ab1
951186 F20101106_AAARAV caicedo_h_Page_025.jp2
c2a28febdf513b954bdd39afd31e22b1
fe44ab87b0102d0ded2d4b4b85ac5a35bff9aaf6
F20101106_AAASFA caicedo_h_Page_142.QC.jpg
d2b287c7c0d12e5461637e475ded0a55
8466e5248bf715ed08935eb316a74a6656384842
29311 F20101106_AAASEM caicedo_h_Page_130.QC.jpg
897f66655ea97dc1beba79e13f0df4f3
7d6b3be9329dbc0525d58825bc356c9342321d94
7361 F20101106_AAASDY caicedo_h_Page_121thm.jpg
dacbe90b5250c5f8b4c8a1ddde5dff8b
3609d159d720c8cd9333c82414d6367977038af9
35537 F20101106_AAARZG caicedo_h_Page_031.QC.jpg
420e9f552d44319546adb6e6f2c4de9c
f0c3871f58d40d81cd2aa7e72a5d6f31b107e89d
8657 F20101106_AAARYR caicedo_h_Page_019thm.jpg
be3483ae70f0de635d38b8e20043a8b8
50c6edaa40da881881ba2a7388e9825ad6c0a48b
105110 F20101106_AAAQVP caicedo_h_Page_052.jpg
f204eaffe773a2cea0aca092492b095f
dea9f71779e3adfb42a843ea6b14bab0817b288e
F20101106_AAARBK caicedo_h_Page_046.jp2
653df6cf518810ebf071a65e2be8d