Application of Zirconium Phosphate/Phosphonate Surfaces in Phosphopeptide Enrichment and Protein Immobilization

MISSING IMAGE

Material Information

Title:
Application of Zirconium Phosphate/Phosphonate Surfaces in Phosphopeptide Enrichment and Protein Immobilization
Physical Description:
1 online resource (156 p.)
Language:
english
Creator:
Liu, Hao
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
TALHAM,DANIEL R
Committee Co-Chair:
CAO,YUNWEI CHARLES
Committee Members:
FANUCCI,GAIL E
YOST,RICHARD A
LONG,JOANNA R

Subjects

Subjects / Keywords:
enrichment -- phosphonate -- phosphopeptide -- zirconium
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:
Phosphorylation is one of the most significant post-translational modifications of proteins in eukaryotes, and is involved in numerous cellular events. Therefore the identification and quantification of phosphorylation sites in proteins is of great importance. With its rapid development, mass spectrometry (MS) appears to be the most powerful technology in proteome research. However, the quantitative characterization of phosphoprotein is still a challenging task because of the low abundance of phosphoprotein, the substoichiometric nature of phosphorylation, and some technical limitations, which makes the isolation and concentration of phosphopeptides prior to MS analysis necessary. To date, the ion metal affinity chromatography (IMAC) and metal oxide affinity chromatography (MOAC) are the most commonly used methods. However, nonspecific binding from the acidic side chains of nonphosphorylated peptides results in low specificity and sensitivity in both. The use of zirconium alkyl-phosphonate (ZrP) for capturing molecules containing phosphate groups was first reported by our group and collaborators. The first project applied the ZrP surface in phosphopeptide enrichment by first studying the pH influence on peptide binding and then investigating the ability of the surface in distinguishing phosphopeptide analogs, using surface plasmon resonance enhanced ellipsometry (SPREE) technique. The second project aims to compare the enrichment efficiency of ZrP surface to TiO2, which is the most widely used material in MOAC. A solution deposition method to grow TiO2 thin film on gold slide was tried and atomic layer deposition (ALD) was also performed. Four peptides used in project 1 were chosen for enrichment efficiency comparison using SPREE technique. The objective of the third project is to develop a phosphorylatable protein tag for protein immobilization. 5 peptide candidates were chosen to evaluate the binding affinity to Zr-phosphate and Zr-phosphonate surfaces. The final chosen peptide was fused with a protein to test its ability to immobilize the protein. A bench prepared amine slide was also tested in the same experiments. With SPREE and fluorescence techniques, we can obtain mass quantitative information and details on kinetic binding. Other biophysical techniques such as atomic force microscopy (AFM), X-ray photoelectron spectroscopy (XPS) were applied to characterize the surface.
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 Hao Liu.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: TALHAM,DANIEL R.
Local:
Co-adviser: CAO,YUNWEI CHARLES.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0046311:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 APPLICATION OF ZIRCONIUM PHOSPHATE/PHOSPHONATE SURFACES IN PHOSPHOPEPTIDE ENRICHMENT AND PROTEIN IMMOBILIZATION By HAO LIU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLM ENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Hao Liu

PAGE 3

3 To my parents Jianjiong Li and Zhengjiang Liu To my husband Zheng Zheng and our coming son

PAGE 4

4 ACKNOWLEDGMENTS I would love to give my sincerest thanks to my advisor Prof. Daniel R. Talham. and guidance he gave me during the last five years. He not only provid ed me with inspiring discussions about research, but also, more importantly, taught me to think professionally, allowed me to grow in the group and taught me many many things far beyond the scope of research, from which I believe I will benefit for my enti re lifetime. I would also like to thank all my committee members, Dr. Richard A. Yost, Dr. Y. Charles Cao, Dr. Gail E. Fanucci and Dr. Joanna R. Long. Thank you for agreeing to be on my committee and the time you spent on my oral proposal and dissertation and also all th e precious discussion and advice during my oral qualifying examination. I also want to acknowledge our collaborators at Universit de Nantes in France: Dr. Bruno Bujoli, Dr. Charles Tellier, and Dr. Clmence Quefflec. We had many particu lar discussions towards research and they gave me tremendous help during my total of 8 weeks visit in Nantes. Dr. Clmence Quefflec, who generously helped me settle down when I first visited France and accommodated me during my second visit, did everythin g to make me feel like at home. She showed me around the city and introduced places of history to me in the weekends, and she took good care of me while I was sick. She is a rigorous researcher, a nice collaborator, and a true friend with great personality Our friendship has last for 1.5 years till now and I believe it will be lifetime long. I want to say thank you to my parents. They taught me to be kind, thankful, and to show care and respect for others. They help ed me to find my dream and did what ever t hey could to support my pursuit o f it. They are the ones who told me that

PAGE 5

5 there are so many things in life more important than money, power and fame -love and friendship, integrity and honesty, dreams and persistence. With my deepest love, I also want to s ay thank you to my husband Zheng Zheng, for loving me and supporting me over the years. He shares my happiness, and laughed out. We learned a lot about love and marriage from all those hard time s in the past 5 years pursuing two Ph.D. degree s in a foreign country. I owe thanks to all my friends, for their company during this long journey. They are always ther e when we have good news to share and celebrate, and more impor tantly, when we need help. Friendship is one of the most valuable re sources in I want to express my gratitude to all the Talham group members, for such a relaxed and harmonious working environment. We spent lots of fun times togeth er. Dr. Roxane M. Fabre, who graduated from our group in 2010, helped me a lot with starting up in the lab and the SPREE instrument. Dr. Matthew J. Andrus, who just defended his dis sertation in November taught me how to use the AFM and helped me twice in deep night to clean up the flood when there was a water leakage in the lab. Corey Gros, a current Ph.D. candidate in our group, performed almost all the XPS experiments in my disser tation. And the rest of the group, thank you all for your company and help during this unforgettable period. The past five years at the Chemistry Department, University of Florida was a great pleasant time to be remembered forever. I would like to thank ou r graduate coordinator Dr. Benjamin W. Smith, our graduate records and recruitment assistant

PAGE 6

6 Ms. Lori H. Clark, and analytical program assistant Ms. Antoinette M. Knight, for all their help in academic life and Dr. James C. Horvath for the joyful teachi ng experience. I also appreciate the help of Dr. Eric Lambers from MAIC with the XPS experiment and Dr. Brent Gila from NRF for his help with the ALD experiment. Finally I want to express my appreciation for the research funding provided by NSF. Thank you I love you all.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLED GMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 12 LIST OF ABBREVIATIONS ................................ ................................ ........................... 16 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 1 9 Phosphoproteome ................................ ................................ ................................ .. 19 IMAC ................................ ................................ ................................ ................ 21 MOAC ................................ ................................ ................................ ............... 21 Zr Phosphate/Phosphonate ................................ ................................ .............. 22 Protein Immobilization on Surfaces ................................ ................................ ........ 23 Study Overview ................................ ................................ ................................ ....... 25 2 SURFACE PLASMON RESONANCE ENHANCED ELLIPSOMETRY ................... 33 Ellipsometry ................................ ................................ ................................ ............ 34 Polarization of light ................................ ................................ ........................... 34 Reflection at surfaces: Fresnel equations ................................ ......................... 35 Nulling ellipsometry ................................ ................................ .......................... 36 Modeling of the multilayer system ................................ ................................ .... 38 Surface Plasmon Resonance ................................ ................................ ................. 39 Total Internal Reflection (TIR) ................................ ................................ .......... 40 Surface Plasmons ................................ ................................ ............................ 40 Surface Plasmon Resonance Enhanc ed Ellipsometry(SPREE) ............................. 42 Experimental Set up ................................ ................................ ......................... 43 The SPREE Experiment ................................ ................................ ................... 44 K inetic Model of Phosphopeptide/s urface Interactions ................................ ..... 45 Discussions ................................ ................................ ................................ ............. 49 3 pH DEPENDENCE STUDY OF ZIRCONIUM PHOSPHONATE SURFAC E IN PHOSPHOPEPTIDE ENRICHMENT AND MECHANISM DISCUSSION ............... 56 Background ................................ ................................ ................................ ............. 56 Experimental Section ................................ ................................ .............................. 57 Materials and Solutions ................................ ................................ .................... 57 Substrate Preparation ................................ ................................ ...................... 57 Zirconium Phosphonate Modified Surfaces ................................ ...................... 58 SPREE Experimental Set Up ................................ ................................ ........... 58

PAGE 8

8 Results and Discussions ................................ ................................ ......................... 59 Surface Characte rization ................................ ................................ .................. 59 Maximum binding capacity on zirconium phosphonate surfaces ............... 60 Zirconium phosphonate surfaces in peptide capturing ............................... 61 Selectivity of zirconium phosphonate surface s investigation ..................... 61 Kinetic Analysis of Phosphopeptide/Surface Interactions ................................ 62 Phosphopeptide surface Interaction Mechanism Discussion ........................... 62 Future Work ................................ ................................ ................................ ............ 70 4 SPECIFICITY INVESTIGATION OF ZIRCONIUM PHOSPHONATE SURFACE IN PHOSPHOPEPTIDE ENRICHMENT AND DiFFERENTIAL DESORPTION STUDY ................................ ................................ ................................ .................... 83 Background ................................ ................................ ................................ ............. 83 Experimental Section ................................ ................................ .............................. 84 Results and Discussions ................................ ................................ ......................... 85 Peptide Analogue Kinetic Study ................................ ................................ ....... 85 Differential Desorption Study ................................ ................................ ............ 86 Future Work ................................ ................................ ................................ ............ 87 5 DEPOSITION OF A TITANIUM DIOXIDE THIN FILM ON GOLD AN D ITS PHOSPHOPEPTIDE ENRICHMENT EFFICIENCY COMPARED TO THE ZIRCONIUM PHOSPHONATE SURFACE ................................ ............................. 97 Background ................................ ................................ ................................ ............. 97 Atomic Layer Deposition ................................ ................................ ................... 97 Solution Deposition ................................ ................................ .......................... 99 Experimental Section ................................ ................................ ............................ 100 Solution Deposit ion ................................ ................................ ........................ 100 Atomic Layer Deposition ................................ ................................ ................. 100 SPREE Experiment ................................ ................................ ........................ 101 Zeta Poten tial Measurement ................................ ................................ .......... 101 Results and Discussions ................................ ................................ ....................... 101 Surface Characterization ................................ ................................ ................ 101 SPREE Experiment ................................ ................................ ........................ 104 Zeta Potential Measurement ................................ ................................ .......... 105 Future Work ................................ ................................ ................................ .......... 106 6 DESIGN AND OPTIMIZATION OF A PHOSPHOPEPTIDE ANCHOR FOR SPECIFIC IMMOBILIZATION OF PROTEIN ON ZIRCONIUM PHOSPHATE/PHOSPHONATE MODIFIED AMINE SURFACES ........................ 118 Background ................................ ................................ ................................ ........... 118 Experimental Section ................................ ................................ ............................ 120 Materials and Solutions ................................ ................................ .................. 120 APTES slides preparation ................................ ................................ ........ 121 Zirconium phosphate modified surface ................................ .................... 121

PAGE 9

9 Zirconium phosphate/phosphonate modified surface ............................... 121 Microarray spotting and incubation conditions ................................ ......... 122 Microarray analysis ................................ ................................ .................. 122 Surface Plasmon Resonance Enhanced Ellipsometry (SPREE) experiments ................................ ................................ .......................... 123 Results and Discussion ................................ ................................ ......................... 123 Surface Characterization ................................ ................................ ................ 123 Phosphopeptide Binding on Zr Phosphate/Phosphonate Surfaces ................ 123 Spotting experiments ................................ ................................ ............... 124 SPREE experiments ................................ ................................ ................ 125 Influence of the number of phosphate moieties in phosphopeptides on their binding affinity ................................ ................................ ............... 128 Future Work ................................ ................................ ................................ .......... 129 7 SUMMARY AND CONCLUSIONS ................................ ................................ ........ 144 APPENDIX : pI OF All PEPTIDES ................................ ................................ ............... 149 LIST O F REFERENCES ................................ ................................ ............................. 150 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 156

PAGE 10

10 LIST OF TABLES Table page 1 1 Comparison of p rotein immobilization procedures ................................ .............. 28 3 1 Peptides studied with their sequence, molecular weight and the abbreviat ion in later figures and tables ................................ ................................ ................... 72 3 2 Measured pH for each peptide solution after mixing with 20 mM Tris solution ... 72 3 3 Selectivity of zirconium phosphonate surface toward each peptide set under various pH con ditions ................................ ................................ ......................... 73 3 4 Association and dissociation rate constants of the interactions between phosphopeptides and zirconium phosphonate surface calculated by the Langmuir mode l ................................ ................................ ................................ .. 73 3 5 Strength s of different parameter s under four pH conditions ............................... 73 3 6 Net charge of each peptide under various pH condition s ................................ ... 73 4 1 Peptides studied with their sequence, molecular weight and the abbreviat ion in later figures and tables ................................ ................................ ................... 90 4 2 Association and dissociation rate con stants of the interactions between phosphopeptides and zirconium phosphonate surface c alculated by the Langmuir model ................................ ................................ ................................ .. 90 4 3 Association and dissociation rate constants of the interactions betw een 1 pY/3 pY and zirconium phosphonate surface calculated by the Langmuir model at all four pH conditions ................................ ................................ ........... 91 5 1 Peptides used in SPREE experiment with TiO 2 thin film coated gold slides. .... 108 5 2 Peak position and r elative intensity for O1s peak with various deposition time 108 5 3 Mean surface roughness, film thickness a nd static contact angle of MUO surface and TiO 2 with various deposition time ................................ .................. 108 5 4 Mean surface roughness, film thickness and static contact angle of TiO 2 thin film on gol d and silicon surfac e after ALD ................................ ........................ 109 5 5 Estimated peptide net charge for set 1 a nd 2 peptides at pH 3.0 and 7.4 ........ 109 5 6 Selectivity of TiO 2 surface wi th different depos ition methods at pH 3.0 and 7.4 ................................ ................................ ................................ .................... 109

PAGE 11

11 6 1 Amino acid sequences used for the binding study with zirconium phosphonate substrates. ................................ ................................ .................. 132 A 1 pI values of all peptides used. ................................ ................................ .......... 149

PAGE 12

12 LIST OF FIGURES Figure page 1 1 Structures and abbreviations of amino acid Serine, Threon ine and Tyrosine. .... 29 1 2 Protein kinases and protein phosphatases. Adapted from "The Science Creative Quarterly", scq. ubc.ca. Adapted with permission ................................ 29 1 3 A typical phosp hopeptide enrichment procedure ................................ ................ 30 1 4 Structures of the most commonly used chelating ligands in IMAC ..................... 30 1 5 Scheme of phosphopeptide enrichme nt procedure with IMAC and MOAC ........ 31 1 6 Schematic str zirconium phosphate ................................ ................... 31 1 7 Zirconium phosphon ate surface for DNA microarrays ................................ ........ 32 1 8 Scheme of protein/enzyme immobiliz ation c lassifications and interactions ........ 32 2 1 Interaction of pola rized light on a sample surface ................................ ............... 50 2 2 Interaction of light with material. ................................ ................................ ......... 50 2 3 Interaction of light with a three layer system ................................ ...................... 51 2 4 Schematic of the ellipsometry instrument with th e different polarization states .. 51 2 5 Total internal reflection condition with the light passes from a denser medium (glass) to a less dense one (air), with c being the critical angle ........................ 52 2 6 Surface plasmon resonance on a glass s urface coated with a gold layer .......... 52 2 7 Surface plasmon resonance simulation curve for Cr ( 2 nm)/Au (28. 5 nm) layer system ................................ ................................ ................................ ....... 53 2 8 Schematic of the surface plasmon resonance enhanced ellipsometry technique (SPREE) .. ................................ ................................ ......................... 53 2 9 Typical SPR EE sensorgram ................................ ................................ ............... 54 2 10 Typical SPREE sensorgram of phosphopeptide adsorbing to the zirconium phosphonate surface. ................................ ................................ ......................... 54 2 11 Raw and f itting data of a phosphopeptide binding to the zirconium phosphonate surface in a SPREE experiment and indication of .................. 55 3 1 Procedure of zirconium phosphonate surface preparation ................................ 74

PAGE 13

13 3 2 AFM images of each step in preparing th e zirconium phospho nate surface ...... 74 3 3 X PS spectrum of Zr ODPA on glass ................................ ................................ ... 75 3 4 XPS multiplex of zirconium and phosphorus of Zr ODPA on glass .................... 75 3 5 Fitting data of set 1 peptides binding to zirconium ph osphonate surface at various pH ................................ ................................ ................................ .......... 76 3 6 Fitting data of set 2 peptides binding to zirconium ph osphonate surface at various pH ................................ ................................ ................................ .......... 76 3 7 Fitting data of set 3 peptides binding to zirconium ph osphonate surface at various pH ................................ ................................ ................................ .......... 77 3 8 Fitting data of set 4 peptides binding to zirconium phosphonate surface at variou s pH ................................ ................................ ................................ .......... 77 3 9 4 pe ptides at various pH conditions ................................ ................................ ......................... 78 3 10 Scheme of how to calculate the influence of charged amino acid residues to phosphopeptide bin ding in the simulation function ................................ ............. 78 3 11 Scheme of how to calculate the contribution of two charged amino acid residues to non phosphorylated peptide bin ding in the simulation function ........ 79 3 12 Scheme of how to calculate the contribution of three charged amino acid residues to non phosphorylated peptide bin ding in the simulation function ........ 79 3 13 predicte d by the simulation function vs. experimental value of set 1 4 peptides and linear fit ................................ ................................ .............. 80 3 14 d by simulation function and actual experimental value ................................ .............................. 80 3 15 estimat ed by the simulation function vs. experimental ..... 81 3 16 The zirconium phosphonate surface representation before phos phopeptide binding at low pH and high pH and after phosphopeptide binding . .................... 81 3 17 Scheme illustrating the nonspecif ic binding pr esented by carboxyl side chains ................................ ................................ ................................ ................. 82 4 1 Fitting results of two parallel SP REE experiments of all peptides ....................... 92 4 2 Fitti ng results of two parallel SPREE experiments of 1 pY an d 3 pY at various pH conditions ................................ ................................ ......................... 93

PAGE 14

14 4 3 SPREE raw data with HAc as desorbing solvent at pH 3.0 ................................ 94 4 4 SPREE raw data with pH=1.0 10mM Tris solution as desorbing solvent at pH 3.0 ................................ ................................ ................................ ...................... 94 4 5 SPREE raw data with 300mM NH 4 HCO 3 a s desorbing solvent at pH 3.0 .......... 95 4 6 SPREE raw data with 6% TFA solution and water as desorbing solvent at pH 3.0 ................................ ................................ ................................ ...................... 95 4 7 SPREE raw data with 1% H 3 PO 4 solution as desorbing solven t at pH 3.0 ......... 96 5 1 Scheme of one typical ALD cycle ................................ ................................ ..... 110 5 2 XPS spectrum of TiO 2 deposited on go ld with various deposition time ............ 111 5 3 The O1s peak and peak fitting results of MUO and TiO 2 with various deposition time on gold ................................ ................................ ..................... 112 5 4 AFM image of TiO 2 thin fi l m with various deposition time ................................ 112 5 5 XPS spectrum of TiO 2 thin film on gold surface and silicon surface deposited by ALD ................................ ................................ ................................ .............. 113 5 6 Ti2p and O1s peak and fit result of TiO 2 on silicon and gold surface by ALD ... 114 5 7 XPS spectra of TiO 2 thin film on gold surface at sample tilting angle of 20 degree, 45 degree and 70 degree respectively. ................................ ............... 115 5 8 P rinciples of Angle resolved XPS ................................ ................................ ..... 115 5 9 2D AFM image of TiO 2 thin film on silicon surface and gold surface ................ 116 5 10 Set 1 and 2 peptides binding to TiO 2 surface prepared by soluti on deposition at pH=3.0 and 7.4 ................................ ................................ ............................. 116 5 11 Set 1 and 2 pepti de binding to TiO 2 surface prepared by atomic lay er deposition at pH=3.0 and 7.4 ................................ ................................ ............ 117 5 12 Zeta potential of TiO 2 nanoparticles at pH 3.0 8.0 ................................ ............ 117 6 1 Covalent attachment of proteins on Zr phosphate/phosphonate surf aces via phsophopeptide anchors ................................ ................................ .................. 133 6 2 Comparison between commercially purchased SuperAmine surface and APTES treated glass surface ................................ ................................ ............ 134 6 3 Structure of the organic compound D used to prepare zirconium pho sphonate modified amine surface ................................ ............................... 134

PAGE 15

15 6 4 Scheme of the two amine surfaces after POCl 3 and compound D treatment ... 134 6 5 AFM and XPS surve y spectra of SuperAmine surface ................................ ..... 135 6 6 AFM and XPS survey spectra of APTES surface ................................ ............. 136 6 7 XPS result of Zr and P s ignal on zirconium phosphate a nd zirconium phosphonate modified SuperAmine surface ................................ ..................... 137 6 8 Fluorescence analysis of Zr/phosphonate substrates spotted with peptide 0P, 1P, 2P, 3P and 4P at different pHs ................................ ................................ ... 137 6 9 Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on POCl 3 treated SuperAmine slide 138 6 10 Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on comp ound D treated Superamine slide ................................ ................................ ................................ .................. 139 6 11 Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various co ncentrations (1, 5 and 10 M) on POCl 3 treated APTES slide ......... 140 6 12 Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on compo und D treated APTES slide ................................ ................................ ................................ .................. 141 6 13 Overall signal change 4P binding to zirconium phosphonate surface at pH 3.0 7.4 in super buffer. ................................ ............................... 142 6 14 Raw SPREE data of 0P 4P binding to zirconium phosphonate surface at pH 3.0 in super buffer with a pH 7.4 washing step ................................ ................. 142 6 15 Raw SPREE data of 0P 4P binding to zirconium p hosphonate surface at pH 3.0 in super buffer with addition of Streptavidin afterwards .............................. 143

PAGE 16

16 LIST OF ABBREVIATIONS AFM Atomic force microscopy ALD Atomic layer deposition AOI Angle of incidence APTES (3 Aminopropyl)triethoxysilane DMAP 4 Dimethylaminopyridine IMAC Ion metal affinity chromatography LB Langmuir Blodgett MES 2 (N morpholino)ethanesulfonic acid MOAC Metal oxide affinity chromatography MSE Mean square error MUO 11 mercapto 1 undecano l ODM Octadecyl mercaptan ODPA Octadecylphosphonic acid OTCC Open tubular capillary columns OTS Octadecyltrichlorosilane ROI Reg ion of incidence SAM Self assembled monolayer SPR Surface plasmon resonance SPREE Surface plasmon resonance enhanced ellipsometry TE Titanium tetraeth oxide TIR Total internal reflection XPS X ray photoelectron spectroscopy

PAGE 17

17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requiremen ts for the Degree of Doctor of Philosophy APPLICATION OF ZIRCONIUM PHOSPHATE/PHOSPHONATE SURFACES IN PHOSPHOPEPTIDE ENRICHMENT AND PROTEIN IMMOBILIZATION By Hao Liu December 2013 Chair: Daniel R. Talham Major: C hemistry Phosphorylation is one of the most significant post translational modifications of proteins in eukaryotes, and is involved in numerous cellular events. Therefore the identification and quantification of phosphorylation sites in proteins is of great importance. With its rapid development, mass spectrometry ( MS ) appears to be the most powerful technology in proteome research. However, the quantitative characterization of phosphoprotein is still a challenging task because of the low abundance of phosphoprotein, the su bstoichiometric nature of phosphorylation, and some technical limitations, which makes the isolation and concentration of phosphopeptides prior to MS analysis necessary. To date, i on metal affinity chromatography (IMAC) and m etal oxide affinity chromatography ( MOAC ) are the most commonly used method s ty and sensitivity in both. The use of zirconium alkyl phosphonate (ZrP) for capturing molecules containing phosphate groups was first reported by our group and collaborators in studies including DNA or protein immobilization The first study described in this thesis applied the ZrP

PAGE 18

18 surface in phosphopeptide enrichment by first studying the pH influence on peptide binding and then investigat ing the ability of the surface in distinguishing phosphopeptide analogs, using surface plasmon resonance enhanced ell ipsometry (SPREE) technique. The second project aims to compare the enrichment efficiency of the ZrP surface to TiO 2 which is the most widely used material in MOAC. A solution deposition method to grow a TiO 2 thin film on gold slide s was tried and atomic layer deposition (ALD) was also performed. Four peptides used in project 1 were chosen for the enrichment efficiency comparison using the SPREE technique. The objective of the third project is to develop a phosphorylatable protein tag for protein immobi liz ation. Five peptide candidates were chosen to evaluate the bi nding affinity to Zr phosphate and Zr phosphonate surface s The final chosen peptide was fused with a protein to test its ability to immobilize the protein. A bench prepared amine slide was also tested in the same experiments. With SPREE and fluorescence techniques, we can obtain mass quantitative information and details on kinetic binding. Other biophysical techniques such as atomic force microscopy (AFM), X ray photoelectron spectroscopy (XPS) were applie d to characterize the surface.

PAGE 19

19 CHAPTER 1 INTRODUCTION Phosphoproteome Phosphorylation is one of the most significant post translational modifications of proteins in which a serine a threonine or a tyrosine residue (Figure 1 1) is phosphorylated by a protein kinase by the addition of a phosphate group. Starting from the 1 950 s, phosphoproteins began to emerge as ke y regulators of cellular life. In 1954, Burnett and Kennedy first reported the biological process of phosphorylation by observing an enzyme activity in which a phosphate was transferred onto a protein. [ 1 ] The protein responsible was a liver enzyme that catalyzed the phosphorylation of casein and bec ame known as a protein kinase (F igure 1 2), the first of its kind to be discovered. A year later, Fischer and Krebs [ 2 ] and Sutherland and Wosilait [ 3 ] showed that an enzyme involved in glycogen metabolism was regulated by the addition or removal of a phosphate, suggesting that reversible phosphorylation could control enzyme activity. Today, with the development of science and technology, it is estimated that protein phosphorylation is involved in about 30% of the proteome and plays crucial regulating roles in numerous cellular events such as cell growth, cell division, cell signaling, metabolism and apoptosis, wh ich is often referred as [ 4 ] [ 5 ] [ 6 ] In these metabolic processes, the protein realizes its regulation functions by switching between a phosphorylated form and an unphosphorylated form, with one being the activate form and the other the inactivate fo rm. Abnormal phosphorylation is the cause of many human diseases like diabetes or ins and pathogens also function by altering the phosphorylation states of proteins. [ 7 ] [ 8 ] Therefore, the identification, location

PAGE 20

20 and quantificati on of phosphorylation sites in proteins is of great imp ortance to help investigate the regulation of cellular metabolic mechanisms and other biological networks. Mass spectrometry ( MS ) is an important technology in proteome re search and is widely applied in the detection and characterization of protein phosphorylation and phosphopeptides due to its ability to identify a large number of p roteins from their corresponding peptides fragments, as well as to identify and locate post translational modifications such as phosphorylation sites. [ 9 ] However, the identification and characterization of phosphoprotein is still a very challenging task because of the low abundance of phosphoprotein, the substoi chiometric nature of phosphorylation, and some technical lim itations, [ 10 ] [ 11 ] which makes a step of isolation and concentration of phosphopeptides prior to MS analysis necessary. In this step, protein digest i s introduced to the phosphopeptide enrichment system, which retains most of the phosphopeptide from the mixture. In the following rinsing step, the unbound peptide from the digest mixture and most of the non specific ally bound non phosphopeptide i s washed out, while the target phosphopeptides remains in the system. Finally in the elution step, a strong eluting solvent would be able to wash all the phosphopeptides off the system, giving concentrated and purified phosphopeptides. A flow chart of the enrichment procedure i s shown in Figure 1 3. Many approaches have been developed to capture phosphopeptide from protein digest mixtures, including immunoprecipitation with phosphoprotein specific antibodie s, immobilized metal affinity chromatography (IMAC), ion exchange chromatography, m etal oxide affinity chromatography ( MOAC ), and other alternative methods such as chemical derivatization and addition of an affinity

PAGE 21

21 tag to phosphorylated amino acids through chemi cal reactions. [ 12 ] To date, IMAC, MOAC are the most commonly used techniques. And zirconium phosphate/phosphonate modified particles or surfaces also becomes more and more popular since the first [ 5 ] [ 13 ] [ 14 ] [ 15 ] [ 16 ] IMAC Immobilized metal affinity chromatogra phy is based on a high affinity binding of an immobilized metal ion by chelating a part of the target protein (Figure 1 5) [ 17 ] It is first introduced by Porath et at. in 1975 under the name of metal chelate affinity chromatograph y a nd is now the most popular tool in phosphopeptide enrichment. [ 18 ] Typical metal ions used in IMAC include Fe 3+ Ga 3+ and Ni 2+ and recently some groups also reported IMAC with Ti 4+ and Zr 4+ [ 6 ] [ 19 ] [ 20 ] In conventional IMAC, the metal ions are usually immobilized onto the adsorption matrix by two types of chelating ligands: iminodiace tic acid (IDA) and nitrilotriacetic acid (NTA), which chelates to the metal ion via carboxylic groups and amino groups respectively(Figure 1 4). However, phospho rylated peptides containing carboxylic group s (indicated in the red box in Figure 1 5) in the side and sensitivity. [ 21 ] [ 19 ] [ 22 ] MOAC MOAC is an upgrade of IMAC, by incorporating metal oxides, mostly titanium dioxide (TiO 2 ), zirconium dioxide (ZrO 2 ), in place of metal ions (Figure 1 5). Compared to IMAC, MOAC was found to have better loading c apacity. [ 23 ] [ 24 ] In MOAC, metal oxides capture phosphopeptides through the formation of bidentate bonds on the metal oxide surface. [ 25 ] [ 26 ] Metal oxides can also be regarded as Lewis acids, the phosphate metal oxide interaction could also be regarded as Lewis acid base interaction. Other

PAGE 22

22 proposed mechanism believed the enrichment is based on the electrostatic interaction between negatively charged peptides and positively charged metal oxide surface at lo w pH. [ 27 ] [ 28 ] Compared to IMAC, MOAC is considered to be time saving, more sel ective and specific. [ 11 ] [ 29 ] [ 17 ] It is suggested that large surface area associated with the nanoscale nature may be the cause of the improvement in enrichment specifici ty and selectivity. However, non specific binding from non phosphorylated peptides containing acidic side cha ins is also observed with MOAC. Zr P hosphate/ P hosphonate zirconium pho ZrP). It is a layered structure consisting of sheets of Zr atoms bound to the O atoms in phosphonic acid (structure shown in Figure 1 6 [ 5 ] ), which has excellent ion exchange capacity, high mechanical strength and biocatalytic performance. [ 30 ] ZrP in phosphopept ide enrichment in 2008. Comparing ZrP to those of IMAC (with Fe 3+ ) and MOAC with TiO 2 ZrP was found to isolate the most phosphopeptide s after phosphoproteome analysis of mouse liver and leukemia cell lysate. [ 31 ] The use of zirconium alkyl phosphonate (ZrP) for capturing molecules containing phosphate groups was first reported by our group and collaborators for its application in oligonucleotide microarrays based on the strong covalent bond formed between the zirconium ion and the phosphate group in single stranded DNA molecules(ss DNA). [ 32 ] Oligonucleotide microarray s were achieved by spotting ss DNA molecules in a certain pattern on the ZrP su rface followed by incubating with complementary DNA with a

PAGE 23

23 fluorophor. Fluorescence could be observed from where the ss DNA molecules were spotted (F igure 1 7). Since then, various types of ZrP modified surfaces were applied in phosphopeptide enrichment, s uch as ZrP coated porous silicon wafer, ZrP coated magnetic nanoparticles, and ZrP coated open tubular capillary columns(OTCC). [ 5 ] [ 14 ] [ 15 ] [ 33 ] All demonstrated the high selectivity and sensitivity of ZrP towards phosphopeptides. As compared to IMAC and TiO 2 ZrP on self assembled monolayers (SAMs) has been shown to bind more phosphopeptides with better selectivit y from different protein digests. [ 16 ] Protein Immobilization on Surfaces Functioning as structural components and playing important roles in numerous metabolic procedures such as immune response, transpor t and storage process, protection events, and signaling processes, proteins are the most significant biopolymers in life. Immobilization of protein s can be defined as attaching selected protein molecules to a support with reduction or loss of its mobility ideally without a loss of activity or function Protein immo bilization is the initial event for its diverse application in affinity chromatography, immunoassays, biosensors, biochips, bioreactors and many clinical analysis and diagnostics. [ 34 ] [ 35 ] One of the most remarkable protein functions is their role as natural biocataly sts enzymes, which control almost all metabolic reactions in human body. Due to their high activity under mild conditions, high selectivity and specificity, immobilized enzymes and microorganisms have found their places in food, medicine and pharmaceutic al industrial and environment al monitoring. [ 36 ] The major concern regarding immobilized prote ins has been the reduction in the biological activity or thermal stability due to immobilization. The loss in activity could possibly be caused by structural modification of the protein during immobilization,

PAGE 24

24 changes in the protein conformation after immob ilization, changes in the protein microenvironment resulting from the interaction between the support and the protein, or limited access of macromolecular substrates to the act ive sites of the enzyme. [ 35 ] Therefore the concept of stabilization has been an important concern when immobilizing proteins/enzymes. The immobilization can be fulfilled by binding the p rotein molecules to a carrier, or via entrapment or enzyme molecules cross linking. Binding to a carrier is currently the most commonly used approach to immobilize proteins, which takes the advantage of the various types of linkages and interactions among the functional groups in side chains of the amino acids and the support. These interactions mainly involve physical adsorption such as hydrophobic and van de r Waals interactions, ionic binding, covalent binding and affinity binding. Figure 1 8 shows a sch eme of various binding interactions [ 37 ] and the comparison of them are listed in Table 1 1. The characteristics of the carr ier matrix are of great importance in determining the performance of the immobilized protein/enzyme system. Ideal support properties include physical resistance to non specific binding, hydrophilicity, inertness toward protein enzyme ease of derivatizatio n, biocompatibility, uniform and oriented functional groups coverage, chemical environment favoring the immobilized protein molecules to retain their native conformation, and avail ability at low cost. [ 38 ] [ 39 ] Carrier matrices can be classified as inorganic and organic according to their chemical composition, from which the organic supports can further be subdivided into natural and synthetic polymers.

PAGE 25

25 In the past decades, protein/enzyme immobilization technology has been rapidly developing and has become a matt er of rational design. To date, many approaches have been developed to immobilize proteins and enzymes, including binding on hollow fiber modules, packed beds, carbon nanotubes, suspended particl es and in hydrogels. [ 40 ] [ 41 ] Nanoporous TiO 2 film was applied in protein immobilization by Topoglidis et. al. in 1998 and then opened the gate for TiO 2 immobilized protein into the electrochemical biosensor application field, in which the direct electron transfer between nan oscale TiO 2 and the immobilized protein help investigate the mechanism of biological redox reactions and develop advanced bioelect ronic devices. [ 42 ] [ 43 ] [ 44 ] ZrP and zirconium zirconium phosphate has been applied in protein immobiliz ations short after its first application in phosphopeptide enrichment, with the mechanism believed to be the intercalation of protein between the layers of crystalline zirconium phosphate. [ 45 ] [ 46 ] [ 47 ] In 2008, after first reported application of ZrP in DNA microarrays, our group and collaborators further applied the same technique in microarrays for probing DNA p rotein interactions, providing new perspective in applying ZrP in protein immobilization. [ 48 ] Study Overview This research aims at appl ying the same surface chemistry, Zr modified gold and amine surface s to biomolecule immobilization, particularly peptides and proteins. Surface plasmon resonance enhanced ellipsometry (SPREE) is th e major technique used in this work. Chapter 2 is an introduction to analytical techniques used in this work with instrumentation and applications, including introduction to fundamentals of surface

PAGE 26

26 plasmon resonance, ellipsometry, and the advantages of th e combination of the two techniques Chapter 3 4 describes the application of ZrP modified gold surface s in a phosphopeptide enrichment study using SPREE technique, including pH dependence, kinetics and selectivity ( C hapter 3). P ossible mechanism s are inv estigated using a mathematical model (C hapter 3 ), and a specifi city test using peptide analogs (C hapter 4 ). As the mechanism of ZrP surface capturing phoshopeptides is believed to be based on the covalent bond formed between the zirconium ion and the phos phate group in the peptides, while the enrichment mechanism of TiO 2 is considered to be involved with electrostatic interaction, a comparison between the two on phosphopeptide selectivity drew our interest. The next chapter C hapter 5, focused on the depos ition of TiO 2 thin films onto gold surface, surface characterization and comparison of efficiency in phosphopeptide enrichment to ZrP surface. The e nrichment mechanism wa s also discussed and compared. This chapter introduce s two approach es to coat a TiO 2 t hin film on gold su r face: solution deposition method and atomic layer deposition method. Certain peptides used in previous experiments in C hapter 3 were again employed to compare with the ZrP surface. C hapter 6 is part of a collaborative research project aims at develop ing a phosphorylatable protein tag that serves as an anchor for protein immobilization on zirconium phosphate and phosphona te modifi ed amine surface. We discussed in detail surface modification, characterization and compared the performance of five protein tag

PAGE 27

27 candidates 5 peptides using fluorescence scanning technique and SPREE technique. And to reduce the cost, this chapter also presented the self prepared amine slides similar to the amine surface used in previous experiments, and implement ed fluorescence experiment using the same peptide candidates was performed to compare the surface chemistry. Finally, C hapter 7 is a summary and conclusion chapter of the overall work.

PAGE 28

28 Table 1 1. Comparison of protein immobilization procedures. Method Advantages Disadvantages Physical Adsorption and ionic binding Simple; mild conditions; Less disruptive to protein molecules Insensitive; binding highly dependent on environment like pH; solvent and temperature Covalent Binding Stable protein support int eraction; ideal for mass production and commercialization Complicated; time consuming; biological activity loss Entrapment Mild procedure; universal for any protein Large diffusional barriers; loss of enzyme activity by leakage; possible denaturation of p roteins Cross linking Simple; strong chemical bonding of the protein molecules Difficult to control the reaction; requires large amount of protein molecules; linked molecules lack of rigidty; low enzyme activity

PAGE 29

29 Figure 1 1. Structures and abbreviation s of amino acid Serine, Threonine and Tyrosine. Figure 1 2. Protein kinases and protein phosphatases. Adapted from "The Science Creative Quarterly", scq.ubc.ca. Adapted with permission.

PAGE 30

30 Figure 1 3. A typical phosphopeptide enrichment procedure. Figure 1 4. Structures of the most commonly used chelating ligands in IMAC: (A) IDA and (B) NTA.

PAGE 31

31 Figure 1 5. Scheme of phosphopeptide enrichment procedure with IMAC and MOAC. Figure 1 zirconium phosphate.

PAGE 32

32 Figure 1 7. Zirconium phosphonate surface for DNA microarrays. Figure 1 8. Scheme of protein/enzyme immobilization classifications and interactions.

PAGE 33

33 CHAPTER 2 SURFACE PLASMON RESONANCE ENHANCED ELLIPSOMETRY The increasing number of methods to synthesize and ext ract biomolecules led to the development of various bio sensors to qualitatively or quantitatively detect interactions among those molecules such as antibodies, DNA, RNA, proteins, and peptides. [ 49 ] Nowadays, the most commonly applied optical biosensors are based on fluorescence technique s However, the demand for label free biosensors as well as time consuming problem associated with attaching fluorescent labels somehow lim its its application. Moreover, layers of bi omolecules are typically in the angstrom to nanometer range, well below optical wavelengths, thus conventional microscopy, which has a resolution limit in the micrometer range, is not sensitive enough Fortunately s ome techniques like interferometry, ellipsometry and surface plasmon resonance are capable of analyzing ultra thin films. [ 50 ] In the presented work, the surface plasmon resonance enhanced ellipsometry (SPREE) technique h as been used to study the biophysics and biochemistry of the binding of biomolecules such as peptides and proteins to zirconium phosphonate modified surfaces Ellipsometry is based on the Fresnel equations to determine layer thicknesses while excitation of surface plasmons in a metallic layer are also sensitive to refractive index and film thickness changes of the surface material. The combination of both highly enhances the sensitivity of this technique due to the propagating plasmon wave on the metal surf ace When used with appropriate metal coatings, SPREE is a powerful tool to study interactions in real time, such as binding events and structural changes, with a high level of sensitivity. [ 49 ] [ 51 ] [ 5 2 ] This method also allows the study of multiple interface interactions such as chemical bond formation, charge transfer

PAGE 34

34 interactions, Van der Waals forces and acid base chemistry. In this chapter the principles of ellipsometry and surface plasmon resonan ce techniques are introduced along with the instrumentation required to perform SPREE experiments. Ellipsometry Ellipsometry is a highly sensitive, non invasive and non destructive technique widely used for investigating dielectric properties of surface s a nd thin film s of different materials such as metals, semiconductors and polymers. This technique measures the change in the polarization of light on a reflecting system at a defined angle of incidence of a polarized light to determine optical constants as well as thicknesses of thin films and bulk materials. The technique measures a relative change in polarization and is therefore not dependent on absolute intensity. This makes ellipsometric measurement very precise and reproducible. In a single measurement film thickness with resolution on the order of angstroms and refractive index to three decimal places can be obtained. E llipsometry principles with fundamental equations were introduced in this section [ 53 ] [ 54 ] Polarization of light A light wave is a transverse electromagnetic wave that has both an electric and a magnetic component. E llipsometry uses only the electric components of the light as it interacts more with matter than the magnetic field does. Natural light is unpolarized, with all planes of propagation being equally probable. Polar ization occurs when the electromagnetic vector oscillates only in certain planes. Light is considered to be linearly polarized when it contains waves that only fluctuate in one specific plane. If light is composed of two plane waves of equal amplitude by d iffering in phase by 90, then the light is said to be circularly polarized, while when the relative phase is other than 90 then the light is elliptically polarized. In fact, the term ellipsometry comes from the word

PAGE 35

35 ellipse, which is the most general sta te of polarization of light. In principle, polarized light is reflected off a sample at an angle of incidence, n Figure 2 1. The incident (E in ) and reflected (E out ) electric vector are connected through the matrix R: (2 1) Reflection at surfaces: Fresnel equations Light interact ing with the surface i s shown on Figure 2 2. A n incident beam from ambient enters the substrate medium with a different refractive index N, and upon contact with the substrate, the incident beam can either reflect or refract. From this simple interaction, the Fresnel reflection coefficients, for t he perpendicular ( s ) and parallel ( p ) components of the electric field can be determined as followed: (2 2) (2 3) with (2 4) where n is the refractive inde x (ratio of the phase velocity in a material to the speed of light in vacuum) and k is the extinction coefficient. The reflectance ratios can directly be obtained from the Fresnel reflection coefficients : R s =| r s | 2 (2 5)

PAGE 36

36 R p =| r p | 2 (2 6) T h e interaction of light with a multi layer system is show n in Figure 2 3. Film thickness measurements utilize the interference between multiple light beams. The light traveling through the film rejoins the light reflected from the film surface. The various components of light will have different phases, depending on the additional optical distances they travel. The interference between multiple light beams will depend on the wavelength (due to different phase velocities) and angle of incidence (due to differ ent path lengths). Using the three phase formalism, the total reflection coefficients and the film phase thickness, can be determined as follow s : (2 7) (2 8) with (2 9) where r 12 and r 23 are the Fresnel reflection coefficients for the interface between media 1 and 2, an d 2 and 3, respectively. For a sample with multiple layers, the N phase model is employed and the total reflection coefficients are calculated by the Airy formulas, the scattering matrix formalism and the 4*4 matrix formalism. Nulling ellipsometry When lin early polarized light with an axis pointing anywhere but in the s or p direction is incident on a sample, the reflected light will in general exhibit an elliptical

PAGE 37

37 state of polarization. The other way around, the same elliptical state of polarization (but with a reversed sense of rotation) incident on a surface will generate a linearly polarized reflection. More g enerally, using a linear polarizer and a quarter wave compensator combination (PC combination) one can always find an elliptical polarization that produces a perfec t linearly polarized reflection with a sample which is not depolarizing The consequence involves : one can easily detect this particular state by using a second polarizer as an analyzer in the reflected beam. For a linearly polarized beam it is possible to extinguish the beam by setting the analyzer to a 90 position with respect to the axis of the linear polar ization. This is called the n tice this is equivalent to finding a minimum in the signal of a pho to detector. Therefore the recipe for a nulling ellipsometer will be: 1. Let light pass through a PC combination, while recording the angular setting of P and C; 2. Change P and C in such a way that the reflection fro m the sample S is linearly polarized; 3. Use a photo detector behind an analyzer A to detect this as a minimum in the signal. [ 55 ] [ 56 ] The instrument used in this res earch is the Imaging EP 3 system (Nanofilm, Germany), which is a 532 nm single wavelength ellipsometer capable of multilayer a nalysis by the variation of the angle of incidence. It is a powerful instrument with a lateral resolution of 1 m and of 1 in the z direction. The imaging ellipsometer operates on the principle of classical null imaging ellipsometry and real time ellipsometric contrast imaging. The laser beam is elliptically polarized after it passes

PAGE 38

38 through the linear polarizer and the quarter wave compensator (Figure 2 4). The elliptically polarized light is then reflected off the sample onto an analyzer and imaged onto a CCD camera through a long working distance objective. I n this configuration, th e orientation of the angles of polarizer and compensator is chosen in such a way that the elliptically polarized light is completely linear polarized after it is reflected off the sample. The ellipsometric null condition is obtained when the absolute minimum of light flux is detected at the CCD camer a. The angles of polarizer, compensator, and analyzer that obtained the null condition are related to the ellipsometric parameters delta ( ) and psi Modeling of the multilayer system In ellipsometry, incident monochromatic light is reflected or transm itted at a surface. For anisotropic materials, where the matrix R is diagonal ( Rsp = Rps =0), the two complex reflection coefficients Rp (= Rpp ) and Rs (= Rss ) describes the two parameters a nd measured by the ellipsometer: (2 10) (2 11) (2 12) gives the ratio of amplitude change for the polarization components, while denotes the relative phase shift of these polarization components upon reflec tion. Reduction of those measured data with computerized optical modeling leads to a deduction of film thickness and the complex refractive indices. The optical model

PAGE 39

39 simulates those two parameters as a function of the optical properties of the sample. The model can also calculate optical properties such as refractive index and thicknesses to simulate in order to compare with the measured Surface Plasmon Resonance Surface plasmon resonance (SPR) is the collective oscillation of electrons in a solid or liquid stimulated by incident light. It occurs when polarized light strikes an electrically conducting surface at the interface between two media. The resonance condition is established when the frequency of light photons matches the natural frequency of surface electrons oscillating against the restoring force of positive nuclei. This generates elec tron charge density waves, namely plasmons, reducing the intensity of reflected light at a specific angle known as the resonance angle, in proportion to the mass on a sensor surface. [ 57 ] [ 50 ] [ 58 ] The optical technique based on SPR phenomenon is widely used in biological applications including biosensors, immunodiagnostics, bioth erapeutic and drug discovery research kinetic analysis of antibody antigen interaction as well as protein activity and stability analysis in biopharmaceutical production The SPR phenomenon was discovered in the early 20 th century and occurs u n der total i nternal reflection condition on the surface of thin layers of noble metals, mostly Au and Ag In principle, the SPR technique measure s the change in refractive index occurring at the surface of a thin metal film. The advantage of this technique involves it s high sensitivity, real time monitoring feature and no need for fluorescent or other labeling agents. For example, t he technique can monitor multistep adsorption and the measurements are made in real time providing dynamic kinetic information in a very se nsitive and label free biochemical

PAGE 40

40 experiment. Refractive index change less than 10 5 can be detected with time resolution of few seconds. [ 59 ] [ 60 ] Total Internal R eflection (TIR) Total internal reflection is a phenomenon that happens when a light beam propagating in a medium with higher refractive index strikes the boundary of a medium with lower refractive index at an angle larger than a particular critical angle with respect to the normal to the surface. The critical angle is the angle of incidence above whi ch the total internal reflection occurs (Figure 2 5) In SPR, typically, the incident light is pass ed through a glass in a form of a prism, whose refractive index is greater than unity. As the angle of incidence increases, the transmitted beam gradually approaches tangency with the interface. At one minimum angle, the critical angle, the transmitted bea m is parallel to the interface thus no energy can be transmitted across the interface. W hen the incident angle equals or is greater than the critical angle, a ll the energy from the incident beam appears in the reflected beam, causing total internal reflect ion. Surface Plasmons However, i f the total internal reflection interface is coated with a thin metallic layer, the p polarized component of the incident light may penetrate the metal layer as a and excite electromagnetic surfac e plasmon waves propagating within the surface of the metallic layer (Figure 2 6). Only the p polarized light component can interact with the plasmons as the s polarized component is not in the same oscillation plane as the plasmons. For a non magnetic lay er, this surface plasmon wave will also be p polarized and will create an enhanced evanescent wave.

PAGE 41

41 The particular configuration shown in Figure 2 6 is called the Kretschmann configuration, which is used to create an evanescent field on the gold surface wh ere the polarized light is directed through a prism with high refractive index ( n = 1.72) to the thin layer of gold in contact with the sample solution in the flow cell with the low refractive index ( n = 1.33). The purpose of using high refractive index me dium is to modify the wave vector of the light by decreasing the phase velocity of the photons. At a specific incident angle, greater than the TIR angle, the vector of the evanescent field matches the wave vector of the plasmons oscillations at the metal/d ielectric interface. If the reflectance R p of p polarized light at a fixed wavelength is measured as a function of the incidence angle then a sharp minimum will be observed where this frequency matching condition is satisfied and SPR occurs (Figure 2 7). S PR occurs at an incident angle of ~65. The minimum reflected light corresponds to the photon energy being coupled to the electrons in the film The observed minimum depends on different parameters of the reflecting system: refractive index (n), extinction coefficient (k) and thickness (d) of the different layers. According to above introduction, the wave of electron density and its associated electromagnetic field generated on a metal surface upon incidence of light is what is referred to as a surface plas mon. The term "plasmon" is chosen both to sugges t that it is an excitation of an electron plasma and also to indicate its quantum nature -just as a "photon" is a quantum of light, a "plasm on" can be a quantum excitation of an electron plasma. Plasmons, although composed of many electrons, behave as if they were a single charged particle. Part of their energy is expressed as oscillation in the plane of the metal surface. Their movement, like the movement of any electrically charged

PAGE 42

42 particles, generates an electrical field. They propagate along the surface with amplitude decaying exponentially in the direction perpendicular to the interface, and can interact with molecules up to a 100 nm range It is important to note that not all metals can support surface plasmons. They exist only in metals with the right material properties, and typically only over narrow frequency ranges. For visible light, silver and gold have plasmon resonances and are ty pically materials used in applications T he frequency of the surface plasmon depends on the bulk plasma frequency and on the dielectric constant of the medium in contact with the metal. Surface Plasmon Resonance Enhanced Ellipsometry (SPREE) The first work utilizing ellipsometry for surface plasmon analysis appeared in 1976 by Abeles. [ 61 ] After then the SPREE method has been more and more applied in detecting genomic DNA adsorption, low molecular weight toxins, pesticides and herbicide s interactions between biomolecules and functionalized polymer surfaces, an d anti body interactions [ 62 ] When the SPR effect is combined with ellipsometry under total internal reflection condition s it yields a hig hly sensitive sensor technique because of the sensitivity enhancement due to SPR waves propagating at the metal surface. The resulting surface plasmon resonance enhanced ellipsometry ( SPREE ) technique greatly improves the precision and sensitivity of t he measurement. When the laser light of the ellipsometer shines through the prism and the glass slide onto the gold film at an angle near the resonance angle the optical reflectivity of the gold changes very sensitively with the presence of biomolecules o n the gold surface. This high sensitivity optical response comes from the efficiency of the collective excitation of conducting electrons

PAGE 43

43 on the gold surface. This configuration can achieve a resolution of 5 10 7 refractive index units for the SPREE techni que compared to 10 5 for the ellipsometric measurements and in the order of 2 10 6 to 1 10 5 refractive index units for SPR only. 40 This method is also less sensitive to external light and intensity fluctuations of the in cident light source compare to the SPR technique because it measures a relative ratio instead of absolute intensity The adsorption and desorption of biomolecules on the gold surface can be observed and quantified by monitoring the change as the increase in the resonance angle is proporti onal to the refractive index and thickness (surface coverage) of the adsorbed layer. as a function of the angle of incidence (AOI) spectra resemble typical SPR curves, since both dependencies represent the Fresnel R p amplitude. The minimum detectable sig nal change of the instrument used in this work was about 10 millidegrees in which results in a thickness precision of 0.1 nm. Therefore, in our SPREE experiment, adsorption of biomolecules that changes the refractive index in the interfacial region can b e sensitively monitored by SPREE. Experimental S et up The scheme of the experimental set up is shown in Figure 2 8a. In our experiment, the substrate used is SF10 glass slides sputtered with a layer of 2 nm chromium as the adhesive layer followed by a 28.5 nm gold layer. The substrate is assembled with a 100 L sample cell, and a 60 SF10 prism ( n =1.72) is mounted with the glass slide using diodomethane as an index matching fluid. The cell, as seen on Figure 2 8b, has inlet and outlet tubes allowing injecti on of different solutions into the cell via a peristaltic pump with digital flow rate control A Peltier temperature control is linked to the cell and the temperature is set at 24 .5 C for all experiments. The cell was sealed against the sample through a r ubber O ring. A monochromic laser beam ( = 532

PAGE 44

44 nm) is elliptically polarized after it passes through a linear polarizer and a quarter wave plate. The elliptically polarized light is then reflected off the sample onto an analyzer and imaged onto a CCD camera trough a 10x working distance objecti ve. The angles of the polarizer and the analyzer are chosen in such a way that the elliptically polarized light is completely linear polarized after it reflects off the sample. The ellipsometric null condition is obtained when the minimum of light intensit y is detected at the CCD camera. The angles of the different components are related to the optical properties of the sample. The software AnalysisR from the company Nanofilm in Germany is used to fit the experimental data to a particular model by which we can then obtain the parameters of the system (thickness, refractive index and extinction coefficient of the layers). Real time measurements consist of recording the value at a fixed angle, a few degrees before the resonance angle to get the highest resol ution, as a function of time. Figure 2 9 shows this specific angle. The experimental set up was optimized to maximize sensitivity and to increase the signal to noise ratio. SPREE detection limits depend on the noise and baseline drift. Long time measuremen ts degrade the detection limit since baseline drifts becomes increasingly problematic with increasing time. Potential sources of baseline drift are temperature fluctuations, c hanges in the flow rate and air bubbles In our experiment, sufficient time was a llowed to obtain a stable signal thus getting the best signal to noise ratio. The SPREE E xperiment Figure 2 10 shows a typical adsorption procedure for a solution of phosphopeptide flowing over a zirconium phosphonate modified signal is plo tted as a function of time. The experiment starts with the buffer in contact with the

PAGE 45

45 surface (A). The phosphopeptide solution is then flowed over the surface a t point B. W hen phosphopeptides reach the surface and begin to adsorb or bind to the surface, ca using an increase in At point C, the adsorption reaches equilibrium and at point D, a buffer wash is performed to monitor any desorption. Kinetics information can be obtained by usi signal in SPREE is not a simple report of association and dis sociation at the surface. It is a combination of all chemical processes and transport processes of diffusion and flow. Kinetic M odel of P hosphopeptide/surface I nteractions The adsorption process of phosphopeptides to the zirconium phosphonate surface depe nds on mass transport (the transport of phosphopeptides to the surface) and intrinsic adsorption kinetic rate (the binding of phosphopeptides with immobilized zirconium ions). Kinetics information can be obtained by using an appropriate model. Several meth ods have been reported to calculate the association and dissociation rate constants from SPREE data, including linearization, analytical integration [ 63 ] and numerical integration [ 64 ] In our experiment, we will adapt the linearization method to calculate the rate constants of phosphopeptide binding to zirconium phosphonate surfaces in a simple Langmuir interaction [ 65 ] [ 66 ] The fitting model we applied to fit our data here is the Langmuir interaction model. The simple binding of an analyte A, in our case the phosp hopeptide, to a receptor B, namely the immobilized zirconium ions, can be based on the law of mass action, which can be represented by the following equation: (2 13) where k on is the association rate constant and k off is the dissociation rate constant.

PAGE 46

46 As indicated in equation 2 13, the procedure is considered reversible. The analyte can either be brought to the receptor by the moving buffer solution and diffusion, or taken away from the surface by a continuous buffer flow. Equilibrium is reached when the rate of complex AB formation equals the rate of the complex dissociation and the dissociation equilibrium constant K D equals When the concentration of analyte equals K D half the receptors will be occupie d at equilibrium. In this model, the reaction between the analyte and the immobilized receptor follows first order reaction kinetics, therefore during the association process, the increase in concentration of the complex AB as a function of time can be rep resented by: (2 14 ) At time t the concentrations of the analyte A and the unoccupied immobilized receptor B is: and (2 15 ) in which it is assumed that the concentr ation of free analyte remains uniform in space and constant in time (= c A (0) ). However, this assumption has some limits [ 67 ] [ 68 ] [ 69 ] By substituting eqn. 2 15 in eqn. 2 14 we have: (2 16 ) At time t=0, the concentration of the bound analyte is zero and the unbound receptor has its highest co ncentration. Solving the above differential equation:

PAGE 47

47 (2 17 ) Since c B (0) = c AB (max) eqn. 2 17 becomes: (2 18 ) Eqn. 3 6 shows the concentration of complex AB increasing with time. The ellipsometric signal relates to complex AB concentration as shown in eqn. 2 19 : (2 19 ) Hence we can build the equation describing the changing ellipsometric signal with time. Before the whole association process, the signal baseline shoul d be stable. The baseline value corresponds to: (2 20 ) solution. For the association process, changing ellipsometric signal with time can b e described in eqn. 2 21 (2 21 )

PAGE 48

48 Because the conce ntration of complex AB follows a quasi first order rate reaction, ellipsometric signal can also be formulated using a quasi first order rate reaction equation. (2 22 ) During the dissociation process, the concentration of the complex AB decreases as a function of time according to a first order decay, given by: (2 23 ) Intergration of eqn. 2 23 gives: (2 24 ) Combining eqn. 2 19 and 2 24 gives: (2 25 ) The d issociation equation is : (2 26 ) The derived equations are used to fit the raw data obtained by SPREE. A fit to a raw SPREE experi ment is shown in Figure 2 1 1 with between the average of starting baseline signal and the average of signal after the final washing step, and is proportional to the surface mass density change This will be discussed in details in Chapter 3 In each of our S PREE experiment, 9 regions of interest were chosen and the final result shown is the averaged value of all 9 regions.

PAGE 49

49 Discussion s In the research work presented further on, characterization of biosensors is shown to be fundamental for the understanding of biological phenomena. The SPREE technique makes rapid, sensitive, specific detection of chemical and biological analytes possible by allowing the study of binding interactions between peptides and modified surfaces while kinetic information is obtained to better understand the interaction mechanism. Imaging techniques such as atomic force microscopy and quantitative technique such as X ray photoelectron spectroscopy were also used to characterize the biosensors in more detail

PAGE 50

50 Figure 2 1. Interaction o f polarized light on a sample surface. Figure 2 2. Interaction of light with material.

PAGE 51

51 Figure 2 3. Interaction of light with a three layer system. Figure 2 4. Schematic of the ellipsometry instrument with the different polarization states.

PAGE 52

52 Fi gure 2 5. Total internal reflection condition with the light passes from a denser medium (glass) to a less dense one (air), with c being the critical angle. Figure 2 6. Surface plasmon resonance on a glass surface coated with a gold layer.

PAGE 53

53 Figure 2 7. Surface plasmon resonance simulation curve for Cr (2 nm)/Au (28.5 nm) layer system. Figure 2 8. Schematic of the surface plasmon resonance enhanced ellipsometry technique (SPREE): (A ). Image of the SPR cell composed of the SF10 prism in the Kretschm ann configuration ( B ). The sample is situated under the prism and has a volume of 100 L.

PAGE 54

54 Figure 2 9. Typical SPREE sensorgram. The dip in the psi signal is associated with the resonance phenomenon. The red curve shows from the shift in the angle of re sonance after adsorption of molecules on the surface. By selecting an angle before the angle of resonance, maximum signal change can be obtained. Figure 2 10. Typical SPREE sensorgram of phosphopeptide adsorbing to the zirconium phosphonate surface.

PAGE 55

55 Figure 2 11. Raw and fitting data of a phosphopeptide binding to the zirconium phosphonate surface in a SPREE experiment and indication of

PAGE 56

56 CHAPTER 3 p H DEPENDENCE STUDY OF ZIRCONIUM PHOSPHONATE SURFACE IN PHOSPHOPEPTIDE ENRICHMENT AND MECHANISM DISCUSSION Background C ompared to metal oxides and IMAC, zi rconium phosphonate surfaces have been shown to bind more phosphopeptides with better selectivity from certain protein digests. The improved enrichment performance of the ZrP method is possibly du e to the coordination mode of the phosphate gr oup which causes strong affinity interactions between the phosphate group and the zirconium layer on the surface. Our group has been working with zirconiu m phosphonate and phosphonate surfaces for years. Recently, Dr. Williams studied the phosphopeptides enrichment capacity of zirconium phosphate based nanoparticles by mass spectrometry and Dr. Fa b r e studied membrane protein immobilization by zirconium phos phonate surface supported lipid bilayers using SPRE E technique. [ 70 ] Among all the phosphopeptide enrichment techniques, there has not been any one fo cused on pH dependence study of the process. Most of the enrichment techniques work un der very acidic pH conditions. However, pH is a crucial factor in phosphopeptide enrichment because it can alter the peptide charge and the state of ionization of charged side chains in peptides, also the charge that the enrichment resin carries. Therefore varying the pH may change the ability of the technique in capturing phosphpeptides. In this work, the pH dependence of phosphopeptide enrichment using zirconium alkyl p hosphonate surface w as studied by SPREE technique. This technique was chosen due to the fact that it is very sensitive to surface refractive index change, which could be use d to monit or the peptide binding process. M eanwhile kinetics information can be obt ained, which may be used in discussing binding mechanism s.

PAGE 57

57 Four pairs of peptides varying in the charged side chains, each consisting of one phosphorylated peptide and one non phosphorylated peptide with the sam e sequence, were studied under four different pH conditions. Then a possible mechanism was proposed and a mathematical model was built to test the rationale of the mechanism and two test peptides were evaluate d to inspect the ability of the model to p redict peptide binding behavior Experimental Se ction Materials and Solutions All peptide s used in this work were obtained from Anaspec. Inc. (California). The sequence and molecular weight and the abbreviations of the chosen peptides were listed in Table 3 1. All peptides were used as received (HPLC pu rity>95%). Peptides were dissolved in Milli 1mg/mL and made into 100 L aliquots. All aliquots were stored at 20C for later use. Four pH values were chosen: 3.0, 4.5, 6.0 and 7.4. Tris base solution was used as the solution system, concentration being 10 mM with 100mM NaCl. pH was carefully adjusted to target value s with 1M, 0.1M and 0.01M HCl or NaOH. The final measured pH values are 3 .09, 4.46, 6.18 and 7.40. Tris solutions with concentration of 20 mM a nd 200 mM NaCl were also prepared with pH from 3.0 8.0. Substrate Preparation Slides used for SPREE experiments were made of 28.5 nm of gold evaporated on a 2 nm chromium adhesion layer on a clean SF10 glass slide (Schott Glass). Gold slides were cleaned with a solution containing 15% ammonium hydroxide, 15% hydrogen peroxide and 70% Milli Q water at 60 80C for 5 min, rinsed with Milli Q water and ethanol, and dried with a nitrogen flow. Cleaned gold slides were immersed in a 1

PAGE 58

58 mM octadecylmercaptan(ODM) solution in ethanol for 16h to make the surface hydrophobic. The hydrophobicity was analyzed by contact angle measurement. The slides were rinsed with ethanol and dried with nitrogen gas afterwards. Zirconium Phosphonate Modified Surfaces A monolayer of o ctadecylphosphonic acid (ODPA) which was synthesized by our collaborating lab in Nantes, France was transferred to the gold slide surface using a KSV 3000 Teflon coated Langmuir Blodgett (LB) trough ( Connecticut ). The surface pressure was monitored with a filter paper Wilhelmy plate. The aqueous subphase was 2.6 mM CaCl 2 solution with pH 7.8, adjusted with a potassium hydroxide solution. ODPA was spread from a 0.30 mg/mL chloroform: ethanol=9:1(v:v) solution with a Halmiton syringe in the LB trough. The s olvent was allowed to evaporate for 10 min, and the monolayer was compressed at a rate of 8 mNmin 1 to reach a surface pressure of 20 mNm. Hydrophobic substrates were then dipped down through the monolayer surface at a rate of 5 mmmin 1 during which pr ocess the monolayer was transferred from the subphase onto the substrates. The substrates were then incubated with a 3 mM zirconyl chloride (Sigma Aldrich) solution for 4 days to bind a monolayer of Zr 4+ ions on the surface. A scheme of the substrate prepa re procedure is shown in F igure 3 1. The slides were then rinsed with and stored in Milli Q water for future use. SPREE Experimental Set Up The SPREE experiments were performed with a Nanofilm imaging ellipsometer EP3 (Germany) coupled with a SPR flow ce ll with Kretschmann configuration. The angle of incidence is set at 64 for all experiments, and the temperature of all the experiments were kept at 24 C by a digital temperature controller (Jemo, Germany). Changes in the surface refraction index caused by the peptide surface interactions are

PAGE 59

59 detected as changes in the psi signal. A SF10 prism was used in all experiments and was coupled together with the gold substrate using diiodomethane as the immersion oil. During the experiments, solutions were pumped t hrough the flow cell using a peristaltic pump (Rainin, California) with digital control of the flow speed varying from approximate ly 0.30 L/min to 1.50 mL/min. In any experiment, a solution baseline was first established by flowing the Tris solut ion of ce rtain pH for 15 20 min. To eliminate the possible change in surface refractive index caused by the introducing of peptide, one aliquot of peptide water solution was mixed with 20mM Tris solution (with 200mM NaCl) of pH close to target pH in order to obtain a pepti de solution of corresponding pH, and then pumped through the cell. The pH of each peptide s olution measured is listed in T able 3 2. After the peptide solution was pumped through the cell the peristaltic pump was stopped, and enough time was given for the peptide to react with the surface. After the psi signal satur ated, a rigorous Tris solution was again applied to wash off the non bound and some loosely bound peptide and monitor the desorption process. All experiments were repeated at least once t o test the reproducibility of the data. Results and Discussions Surface Characterization The surface morphology and chemistry of the substrate was characterized with AFM and XPS. The surface morphology of the gold surface, gold with ODM and with the final Zr ODPA layer was shown in F igure 3 2. AFM was on tapping mode with a measuring scale of 5 m 2 The height of roughness measurement was on a 20nm scale. It is found that the both ODM and zirconium phosphonate films formed uniform films after deposition o n the gold surface.

PAGE 60

60 XPS spectrum and the multiplex i s shown in F igure 3 3 and 3 4. XPS was done on glass slides to save the cost. The P:Zr ratio calculated is 0.96:1.0 close to their stoichiometri c ratio 1:1 according to our previous publications [ 71 ] [ 70 ] [ 72 ] Maximum b inding c apacity on z irconium p hosphonate s urfaces The area of the elliptic O ring of the SPR flow cell can be roughly calculated by the equation of ellipse area : (3 1) where a and b is the length of the major axis and the minor axis respectively. The length measured is 1.1 cm and 0.6 cm for a and b, therefore the area of the O ring will be T he cross sectional area of each zirconium ion will be approximately Therefore a complete coverage of zirconium atoms on the surface will be estimated to be: Within the area of the O ring, the number of zirconium bindi ng sites could be calculated as: However, the zirconium was bound to a monolayer of ODPA with a stoichiometric ratio of 1:1. The mean molecular area of ODPA prior to transferring is about 15 times compared to c ross sectional area of zirconium ion, so theoretically the number of available zirconium binding sites should be 15 times less, which is

PAGE 61

61 The minimum amount of peptide used in the experiments is 5.1110 8 mol, which has phosphate groups, which is an excess compared to the number of available binding sites, therefore the surface should be saturated with peptide in all experiments. Zirconium p hosphonate s urface s in p eptide c apturing In SPREE experiment, peptide binding onto the zirconium phosphonate surface is monitored in real time. The specificity of the zirconium phosphonate surfaces for phosphorylated peptides was studied with peptide set 1 4 under all four pH conditions. Raw data was fit using the L angmuir model introduced in Chapter 2. The fitting results of two parallel experiments are shown in Figure 3 5 to 3 8. with error bar s representing the standard deviation of each data point of the raw data at each stage (baseline stage, peptide adsorption s tage, peptide saturation stage, peptide desorptio n stage) to the fitting data Selectivity of zirconium phosphonate surface s investigation It is obvious that for all sets of peptides, the phosphorylated binds more to the surface than the non phosphorylated signal change is proportional to surface refractive index change, which is further proportional to mass density change on the surface, hereby we define selectivity as

PAGE 62

62 for each pept ide set at certain pH. The selectively at each pH wa s then calculated and shown in T able 3 3. N/A means the surface has 100% selectivity of phosphorylated peptide over non phosphorylated peptide since there is no non specific binding of the n on phosphoryl ated peptide. W e observed that the selectivity change over pH is not the same for all the peptide sets. For set 1 and set 4, the selectivity increases as pH increases, however for set 3 it is completely the opposite. For set 2 the selectivity d e according to pH. Kinetic Analysis of Phosphopeptide/Surface Interactions The transport of the peptide molecules to the surface is an important process in determining the binding kinetics. The rate of adsorption is determined by mass transport and kineti cs at the interface. To minimize the influence of mass transport on the binding kinetics, the peptide Tris mixture was pumped through the flow cell within 5s using the highest pumping speed of the peristaltic pump to mak e sure the psi signal change respect s binding kinetic instead of diffusion control The value of k on k off and K D of the phosphopeptide from each set was then calculated from the fitting data and listed in T able 3 4. There is not a very obvious trend that the dissociation constant K D with p H increases. Phosphopeptide surface Interaction Mechanism Discussion To better compare the difference in peptide binding at each pH, the overall signal change calculated from the difference between the stabilized signal after peptide desorption and the bu ffer baseline in the fitting data of each peptide set is calculated and plotted in Figure 3 9.

PAGE 63

63 Noticing the fact that set 1 and set 4 peptides are similar since they only have acidic side chains in the sequence (D and E), while set 2 peptides have only ba sic side chains (K and R), and set 3 peptides have equal amount of acidic and basic side chains, and they show different dependence trends toward pH, we want to further investigate the interaction mechanisms. We b elieve phosphorylated peptides attach to t he surface mostly because of the c oordinate c ovalent bond ing between the peptide phosphate and the zirconium In addition charges on the peptide side chains and Zr phosphonate surface vary under different buffer pH, which may cause electrostatic interacti ons, attractive or repulsive, between peptide and the surface Zr layer, either help ing bring the peptide molecules down to the surface, or prevent ing the peptide molecules from access ing the surface and thus hinder ing the adsorption process. It is assumed that these two types of interaction are the most significant between peptide and Zr phosphonate surface, and a mathematical model was built to simulate their binding affinity with these two terms. For each phosphopeptide used in the experiments, there is a non phosphopeptide version with the same sequence. Thus, coordination and electrostatic forces can be compared between phos and non phos peptide pairs. Meanwhile, we can also get the idea how the strength of covalent bond formed and different electrostati c forces change with pH. We suppose that for the phosphopeptide to coordinate to Zr phosphate group must orient toward the Zr surface. It can be imagine d that under the same buffer pH, the ability of phosphate to form a coordinate covalent bond for differ ent phosphopeptides would be almost the same, while electrostatic forces act in a totally different w ay. Non phosphopeptide would adopt different conformations when getting close to Zr surface in

PAGE 64

64 order to let its favored charged residues face the charged s urface. The amino acids Aspartic acid (D) and Glutamic acid (E) contains carboxyl groups as nega tively charged residues in favo red pH values, while Argi nine (R) and Lysine (K) contain amino groups as positively charged residues. All charged residues may in fluence binding no matter facing directly to the surface or stretching towards other directions. A common idea can be raised that the contributions of charged residues to the electrostatic force depend s on the distance they are from the charged surface. Be cause it is very difficult to know the actual position of each charged residues when peptides bind to the surface, we build an empirical function to simulate the experimental data in order to evaluate how different charged residues in a peptide contribute to the binding affinity. We first define a "central amino acid", which is the amino acid facing directly toward the surface when binding. Electrostatic forces fade away as the charged amino acid s move along peptide sequence away from the "center amino acid ". (3 2 ) This equation describes the electrostatic force between a charged amino acid and the Zr surface. F e o is the electrostatic force between the peptide and the surface if the charged residue is at the "central amin o acid", n is the number between the "central amino acid" and the amino acid where the charged residue locates. For a phosphopeptide, the "central amino acid" would always be the one attached to phosphate group. Electrostatic forces for all charged residue s can be calculated with e qn.3 16 and the total electrostatic force can be calculated as:

PAGE 65

65 (3 3 ) where i indicates each charged amino acid and n i is the number of amino acids between the "central amino acid" and the i th ch arged amino acid (Figure 3 10 ). In p60. phos., tyrosine with phosphate acts as "central amino acid" when binding to Zr surface. The two g lutamic acids are 3 and 4 amino acids away from the "central amino acid", s o their contributions to electrostatic force are different. T he t otal electrostatic force would be the sum of their partial contributions. However, the case f or a non phosphopeptide is more complicated. Because there is no dominant interaction like co valency between the peptide and the Zr surface, t he peptide can choose as many conformations as possible when getting closed to the surface. All amino acids have chances to face the surface but only those with charged residues contribute to binding. So we only take into account those conformations contri bute significantly to binding. We made two assumptions and simplified this problem as follows: (1) For each conformation, there is one amino acid facing dir ectly toward the Zr surface, and o nly when a charged amino acid is facing directly toward the surfac e, is the corresponding conformation counted. (2) Each conformation taken into account has equal weight. Hence the total electrostatic force is modeled as: (3 4 ) Total electrostatic force is defined by eqn.3 4, in which th e first summation indicates all conformations that contribute to binding are considered, and each

PAGE 66

66 conformation has equal weight as The second summation indicates that in each conformation, t he t otal electrostatic force is the su mmation of partial electrostatic forces from a ll charged residues. Figure 3 11 and 3 12 introduced two different cases of non phosphopeptide electrostatic force calculation. The binding signal of all the peptides from the four peptide sets was calculated with this method and compared with the SPREE experimental results. Covalency ( C ), negatively charged residue Zr electrostatic force ( Fn ), and possitively charged residue Zr electrostatic force ( Fp ) are derived by fitting to the binding signal. Total surfac e binding ( Sb ) is derived from the following equation: (3 5 ) Based on the mathematical function the calculated parameters in e qn. 3 5 are listed in Table 3 5 and the actual experimental binding signal change vs. calculated bind ing signal change at all pH conditions for the four peptide sets are plotted in Figure 3 13 From the calculated parameter values in eqn 3 19, we found that the covalency contribution decreases by more than half as the pH increases from 3.0 to 7.4. Negati vely charged residues contributes positively to peptide binding, and this contribution also decreases as pH increases due to electrostatic interaction, while positively charged residues contributed negatively to peptide binding at lower pH(3.0 and 4.5) and positively at higher pH(6.0 and 7.4) as effected based on the pH dependence of the surface charge Ideally, the plot of predicted by the simulation function vs. actual experimental sho uld correlate lin early with the function y = x After fitting the plot of

PAGE 67

67 our experimental value f rom all four set s of peptides vs. the calculated from the simulation function we have a linear fit of y=0.7 9x+0.076 and R 2 =0.91 showing acceptable c orrelation between the experimental value and calculated value using our mathematical model. Additionally, if we take out the factor of covalency from the simulation function, perform the fitting with only electrostatic interaction factors, the same plot o f calculated value vs. experimental value shows a correlation of R 2 =0.63 poorer compared to the model with covalency as the main interaction factor, proving the necessity of covalency in the interaction mechanism. T wo test peptide s pTP1 and pTP2 were use d to demonstrate the ability of the simulation function to predict peptide binding under certain pH condition s SPREE experiments were performed with these two peptides at all four pH conditions and the overall is plotted for both experimental value and th e predicted value in F igure 3 14 It is obvious that the experimental value is consistent with the predicted value for both, and they follow the same trend over pH. Especially for pTP2, a neutral peptide without any charged side c hains in the sequence, the binding trend is a reflection of the dependence of coordination interaction with pH. The fact that the binding signal decreases as pH increases corresponds to the fitting results by the simulation function in T able 3 6. Similarl y, t he correlation between the calculated value and the actual experimental value at each pH for pTP1 and pTP2 shown in Figure 3 15 gives a linear fit equation of y=1.01 x+0.0056 and y=0.89x 0.032 with a R 2 =0.99 and R 2 =0.96 respectively, while after taking out the covalency factor from the simulation function only results in an R 2 =0.83 for pTP1 and R 2 =0.7 6 for pTP2. These examples also prov ed that

PAGE 68

68 the main interaction force between the phosphopeptide and the zirconium phosphonate surface is the coordinate c ovalent interaction. However, there are also limitations go with the simulation function, as the simulation result highly depends on the tra ining sets. F or example, it can only be applied to simulate and predict binding results of simple systems with the s ame pH range and peptides sequence similar to the training sets. Application of the simulation function to other pH systems and peptides with greatly different structures than the training peptides may lead to less inaccurate results. After testing the re liability of the simulation function and the proposed interaction mechanism, we can use it to explain the pH dependence of different peptides on pH. Since the factors contributing to peptide binding are a combination of covalency and electrostatic interact ion between the peptide and the surface, it is important for us to figure out the surface charge and peptide charge under each pH condition. Calculated peptide charge and anticipa ted surface charge is shown in Table 3 6 and Figure 3 16 Peptide net charge is calculated based on the equation : (3 6 ) where Z is the net charge of the peptide sequence; Ni is the number of arginine, lysine, and histidine residues and the N terminus; pK ai is the pK a values of the N terminus and the arginine, lysine, and histidine residues; Nj is the number of aspartic acid, glutamic acid, cysteine, tyrosine residues, phosphate groups and the C terminus; pK aj is the pK a values of the C terminus and the aspartic acid, glutamic acid, cysteine, tyrosine residues and the phosphate group. [ 73 ]

PAGE 69

69 At lower pH, the surface zirconium ion coordination sites are completed by hy droxide groups in the solution and at higher pH the species occu pying the coordination sites are oxides. Upon the addition of phosphopeptide, the phosphate group is basic enough to compete for the coordination sites of zirconium and form a zirconium phosphonate bilayer structure. [ 74 ] Therefore we suppose at acidic pH, the surface carries positive charge, which decreases as pH increases, until in basic solution, the surface is close to neutral. As for acidic peptide set No. 1 an d 4, there is electrostatic attraction between the surface and the peptide molecules, and the peptide net charge increases while surface charge decreases with pH. More important ly the covalency contribution also decreases with pH, because of the fact that hydroxide is a better leaving group than oxide and is easier to be replaced by phosphate group. Therefore we observed less peptide binding as pH goes higher. For the non phosphopeptide, we believe that the non specific binding is caused by electrostatic i nteraction and weak non covalent linkage between the carboxyl groups in the side chains to the zirconium surface (Figure 3 17 ). For basic peptide set 2, the peptide is positively charged under all pH condition and the net charge is about the same. Therefor e there are repulsive interaction s between the peptide and the surface most of the time. The unfavorable influence weakens as pH goes up, but so does the covalent bond between the phosphate and the zirconium on the surface. This might be the reason that th e phosphopeptides have the same level of binding under all pH conditions. There is no acidic side chain to interact with the positively charged surface so no non specific binding is observed under all pH conditions. This set of peptide s further supports o ur interaction mechanism that the coordination force is the major and

PAGE 70

70 most significant interaction between peptide molecules and zirconium io n s at the surface. P eptide set 3, is more complicated compared to the other sets, as the peptide has both acidic an d basic residues and the influence to binding is an overall effect. At pH 3.0, both the peptide and surface is relatively highly positively charged, and the influence from the acidic residues and basic residues in the peptide sequence almost cancel ou t eac h other (T able 3 5). The repulsion between the peptide and the surface thus hinders the peptide binding. At pH 4.5, peptide turns to be negatively charged and the electrostatic interaction switches to attraction in this case and the overall influence of th e side chains favors the binding process. Thus we observed more binding at pH 4.5. At pH 6.0, though the overall effect of the side chains highly favors binding, the covalency strength weakens, resulting in less binding compared to that at pH 4.5. At pH 7. 4 the peptide binding is both unfavorable in covalency and electrostatic interaction, so even less binding occurs on the surface. The non specific binding in this case is also a result of the acidic side chain as in set 1 and 4. Future Work In this work a zirconium alkyl phosphonate surface was prepared and characterized. The binding affinity of different sets of phosphopeptides and non phosphopeptides with this surface was studied by SPREE. The results indicated that this surface showed good selectivity t owards phosphopeptide s over non phosphopeptide s with the same sequence and can therefore be used as an enrichment medium for phosphopeptides. Moreover, pH dependence of the binding affinity w as observed, with kinetics studies of the rate constant at each stage it was found that low a possible interaction mechanism between the phosphopeptide and the zirconium

PAGE 71

71 phosphonate surface was proposed and a mathematical simulation fun ction was built based on possible factors which could potentially influence the binding and used to fit the experimental raw data to determine the variable parameters in the equ ation. The equation was then applie d to predict the binding behavior of two pho sphopept ides to test its reliability. It turned out that the function simulates experimental data well and proved that the factors we took into consideration is reasonable. However, there are also some limitations about the simulation function, for example it can only be applied to simple systems with the same pH range and peptides sequence similar to the training sets. T he whole proposed mechanism was used to explain the pH dependence of binding behavior of the phosphopeptides and to determine the better pH condition for phosphopeptide enrichment based on the peptide sequence. For example, higher pH will work better for phosphopeptide enrichment when the mixture has a lot of acidic non phosphopeptides, since the non specific binding caused by acidic residu es is minimized at pH 7.4. However, if we want to enrich a peptide out from a basic non phosphopeptide mixture, pH 3.0 will be the best choice because at pH 3.0, the covalency between the peptide and the surface is the strongest and will lead to highest bi nding, recovering the most phosphopeptide, while no non specific binding will occur from basic peptid e at pH 3.0. Future work will further investigate the specificity of the zirconium phosphonate surface towards phosphopeptide an alogs at optimal pH conditi on.

PAGE 72

72 Table 3 1. Peptides studied with their sequence, molecular weight and the abbreviation in later figures and tables. Peptide Abbreviation Molecular Weight Set 1 TSTEPQ p Y QPGENL p60. phos. 1544.5 TSTEPQYQPGENL p60. 1464.5 Set 2 FFKNIVTPR p T PPPSQGK NH 2 pMBP 1913.3 Ac FFKNIVTPRTPPPSQGK NH 2 MBP 1955.3 Set 3 TRDI p Y ETDYYRK 1 pY 1702.7 TRDIYETDYYRK 3 Y 1622.8 Set 4 WAGGDA p S GE pDSIP 928.8 WAGGDASGE DSIP 848.3 Test peptide 1 E ND p Y INASL pTP1 1119.1 Test peptide 2 QGISF p S QPTC pTP2 1148.2 Table 3 2. Measured pH for each peptide solution after mixing with 20 mM Tris solution. Anticipated pH Set 1 Set 2 Set 3 Set 4 Test Peptide p60. phos. p60. pMBP MBP 1 p Y 3 Y pDSIP DSIP pTP1 pTP2 3.0 3.13 3.19 3.21 3.22 3.14 3.09 3.14 3.20 3.12 3.12 4.5 4.47 4.38 4.48 4.78 4.35 4.41 4.47 4.39 4.46 4.31 6.0 5.73 5.53 5.97 6.19 5.72 5.81 5.76 5.55 5.77 6.19 7.4 7.42 7.27 7.31 7.30 7.24 7.28 7.39 7.31 7.38 7.40

PAGE 73

73 Table 3 3. Selectivity of zirconium phosphonate surface toward each peptide set under various pH conditions. Selectivity pH Set 1 Set 2 Set 3 Set 4 3.0 3.3 N/A 24.7 6.5 4.5 2.8 N/A 5.1 5.4 6.0 2.2 N/A 4.8 3.7 7.4 3.7 N/A 3.7 N/A Table 3 4. Association a nd dissociation rate constants of the interactions between phosphopeptides and zirconium phosphonate surface calculated by the Langmuir model. p60. phos. pMBP pH K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off (10 2 mi n 1 ) K D (10 5 M 1 ) 3.0 1.8 0.1 6.50.1 3.70.1 1.20.1 6.30.1 5.50.2 4.5 2.00.1 6.40.1 3.20.1 1.70.1 7.30.1 4.30.3 6.0 1.70.1 8.00.1 4.60.1 1.60.1 9.70.1 6.20.3 7.4 1.50.1 5.20.1 3.40.2 1.60.1 9.90.1 6.20.2 1 pY pDSIP pH K on ( 10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) 3.0 2.80.1 4.20.1 1.50.1 2.30.1 6.30.1 2.70.1 4.5 3.50.1 5.00.1 1.50.1 1.90.1 7.40.1 4.00.2 6.0 3.10.1 6.90.1 2.20.1 1.70.1 9.70.1 5 .70.2 7.4 1.30.1 6.40.1 5.00.3 1.40.1 9.60.1 7.10.6 Table 3 5. Strength s of different parameter s under four pH conditions. pH Co valent Negatively Charged Residue Zr Electrostatic Interaction Positively Charged Residue Zr Electrostatic Intera ction 3.0 0.51 0.1 1 0.1 1 4.5 0.41 0.12 0.059 6.0 0.2 5 0.094 0.086 7.4 0.20 0.020 0.10 Table 3 6. Net charge of each peptide under various pH condition s Net Charge pH p60. phos. p60. pMBP MBP 1 pY 3 Y pDSIP DSIP 3.0 0.839 0.073 3.088 3.000 1 .972 2.883 0.907 0.005 4.5 2.343 1.273 2.929 3.000 0.331 0.739 2.517 1.447 6.0 3.681 1.965 2.285 2.999 1.684 0.031 3.691 1.975 7.4 3.988 2.006 2.009 2.998 1.994 0.012 3.989 2.004

PAGE 74

74 Figure 3 1.Procedure of zirconium phosphonate surfac e preparation. Figure 3 2. AFM images of each step in preparing the zirconium phosphonate surface: (A) plain gold; (B) self assembled ODM monolayer; and (C) the Zr ODPA surface.

PAGE 75

75 Figure 3 3. XPS spectrum of Zr ODPA on glass. Figure 3 4. XPS mult iplex of zirconium and p hosphorus of Zr ODPA on glass.

PAGE 76

76 Figure 3 5. Fitting data of set 1 peptides binding to zirconium phosphonate surface at various pH: (A) 3.0; (B) 4.5; (C) 6.0 and (D) 7.4. Figure 3 6. Fitting data of set 2 peptides binding to zir conium phosphonate surface at various pH: (A) 3.0; (B) 4.5; (C) 6.0 and (D) 7.4.

PAGE 77

77 Figure 3 7. Fitting data of set 3 peptides binding to zirconium phosphonate surface at various pH: (A) 3.0; (B) 4.5; (C) 6.0 and (D) 7.4. Figure 3 8. Fitting data of set 4 peptides binding to zirconium phosphonate surface at various pH: (A) 3.0; (B) 4.5; (C) 6.0 and (D) 7.4.

PAGE 78

78 Figure 3 9. Overall signal change of in SPREE experiment of set 1 4 peptides at various pH conditions. Figure 3 10. Scheme of how to calculate the influence of charged amino acid residues to phosphopeptide binding in the simulation function.

PAGE 79

79 Figure 3 11. Scheme of how to calculate the contribution of two charged amino acid residues to non phosphorylated peptide binding in the simulation function. Figure 3 12. Scheme of how to calculate the contribution of three charged amino acid residues to non phosphorylated peptide binding i n the simulation function.

PAGE 80

80 Figure 3 13. Plot of value predic te d by the simulation function vs. experimental value of set 1 4 peptides and linear fit. Figure 3 function and actual experimental value: (A) pTP1 and (B) pTP2

PAGE 81

81 Fi gure 3 15. Correlation of estimat ed by the simulation function vs. experimental : (A) pTP1 and (B) pTP2. Figure 3 16. The zirconium phosphonate surface representation before phosphopeptide binding at low pH (A) and high pH (B) and after phosphopept ide binding (C).

PAGE 82

82 Figure 3 17. Scheme illustrating the nonspecific binding presented by carboxyl side chains.

PAGE 83

83 CHAPTER 4 SPECIFICITY INVESTIGATION OF ZIRCONIUM PHOSPHONATE SURFACE IN PHOSPHOPEPTIDE ENRICHMENT AND DiFFERENTIAL DESORPTION STUDY Backgrou nd Among the existing phosphopeptide enrichment techniques, the specificity has not been given enough attention. Phosphopeptides can possibly differ in sequence length, number of phosphorylation sites and also site of phosphorylation. For instance, a short peptide may have faster binding kinetics compared to long peptide due to the fact that the phosphate group and charged residues are easier to be exposed, and bi phosphorylated or tri phosphorylated peptides may have different binding affinities from singl y phosphorylated peptides. The sensitive real time monitoring feature of the SPREE technique allows us to get an idea of the total mass that would adsorb on the surface and the kinetics information associated with the adsorption process during the enrichme nt of different type of peptides, which will help determine the specificity of the technique if there is any. After studying the pH dependence and the enrichment mechanism, we applied the SPREE technique to investigate the kinetic binding process of 3 pair s of phosphopeptide analogs, in order to further explore the specificity of zirconium phosphonare surface in phosphopeptide enrichment. The 3 pairs of peptides include phosphopeptides of the same amino acid sequence yet varying in length, number of phospho rylation sites and location of phosphorylation sites, re spectively. SPREE was conducted for each peptide pair at specific pH value which was determined based on previous pH dependence study. Also, to achieve optimal enrichment the desorbing solvent used t o wash off non specific ally bound peptide after the enrichment was also studied aiming at discovering a solvent which could ideally wash off all the bound non phosphopeptide while doing no harm to the phosphopeptide on the surface.

PAGE 84

84 Five solvent systems wer e chosen to learn their ability to wash off non phosphopeptide as well as phosphopeptide from the surface at pH 3.0. Set 1 peptide from Chapter 3, p 60. phos. and p60. non phos. were chosen as peptide s of interest. Experimental Section All peptides used in this work were also obtained from Anaspec. Inc. (California) and prepared in the same way as previous. The sequence and molecular weight and the abbreviations of the chosen peptides were listed in Table 4 1. For pair 1 peptides, there are only aci dic sid e chains in the sequence. B ased on the earlier pH dependence study, acidic phosphopeptide s the highest binding at pH 3.0 so this value was chosen for the SPREE experiment. For pair 2 peptides, peptide 1 pY showed highest binding level at pH 4.5 in the pr evious experiments which was also the pH chosen in this case. However, this peptide group is the most complicated one because they have both acidic and basic residues in the sequence. To better test the reliability of the experimental data, SPREE experime nts were performed also under 4 pH conditions as in early work. For peptide pair 3, it has 7 basic residues in the sequence with only one acidic residue, and the phosphate group is surrounded mainly by basic residues, thus we expect the peptide to show beh avior similar to a basic In this case, pH 7.4 was chosen as peptide molecules All Tris solution and substrate preparatio n was the same as used in Chapter 3 work. In the desorbing solvent study, five solvents systems used were: 0.1% acetic acid (HAc) solution, 300mM ammonium bicarbonate (NH 4 HCO 3 ) solution, 10mM Tris solution(pH=1.0), 1% phosphoric acid (H 3 PO 4 ) solution and 6% trifluoroacetic acid

PAGE 85

85 (TFA) solution + water. The peptides used were the p60. phos. and p60. set from Chapter 3 and the SPREE experiments were conducted at pH 3.0. The desorbing solvent was introduced into the system after the peptides saturated the surf ace and original pH 3.0 Tris solution wash. After the wash signal stabilized, pH 3.0 Tris solution was run again to eliminate the signal change caused by surface refractive index change due to the addition of the new materials in the desorbing solvent. R esults and Discussions Peptide Analogue Kinetic Study SPREE experiments were then performed with peptides in Table 4 1 at chosen pH and the fitting results to exp eriment raw data is plotted in F igure 4 1. Kinetics parameters k on k off and K D was calcul ated and listed in T able 4 2. From the result shown in F igure 4 1, for pair 1 peptides, the mass of 6AA bound to the surface is about half of that of 13AA. Since the peptides are relatively in excess amount compared to available surfaced zirconium ions, the su rface should be mass saturated and the signal change depends on the surface mass density change As the molecular weight of 13AA is about twice of the molecular weight of 6AA, after saturation of surface zirconium coordination sites, the mass density left on the surface for 13AA should also be twice of 6AA. Also there is not much difference in the kinetic constants of the two. For peptide pair 2, almost no desorption is observed for 3 pY, a consequence of more than one phosphate group involved in the bindin g. For pair 3 peptide, no significant difference in s urface mass density is observed. M ore experiments should be done to certify if the surface can distinguish peptide with the same sequence and differ ent phosphorylation locations.

PAGE 86

86 Figure 4 2 shows that a t each pH, the amount of 1 pY and 3 pY bound onto the surface in the adsorption step is almost the same, however, in the desorption step, there was only little 3 pY was washed off from the surface and the amount of 1 pY washed off was obviously more. This is due to the fact that there is not much difference in the molecular weight of 1 pY and 3 pY, however, in 3 pY, which has 3 phosphorylation sites, it is very likely that more than one phosphate group contributed to the binding. The kinetic information sho wn in Table 4 3 indicates a greater K on and a smaller K off and K D for 3 pY compared to1 pY, which may be caused by mult iple phosphate groups binding. Differential Desorption Study A differential desorption study was performed by introducing solvent s other than the experiment buffer to investiga te their ability to wash off non specifically bound peptides. Raw data of SPREE experiments with five desorption solvents is shown in F igure 4 3 to 4 7. From the results shown in F igure 4 3 (a) and (b), by comparing the signal level before and after the wash, we found that 0.1% HAc solution barely washed off any non specifically bound non phosphopeptide, as well as little covalently bound phosphopeptide. After increasing the concentration from 0.1% to 1.0%, there was still hardly any change in non phosphopeptide binding signal (Figure 4 3 (c)), which indicated that HAc is not a good desorbing solvent candidate in this case. From the results shown in Figure 4 4, it is clear that pH 1.0 Tris solution could wash off abo ut 50% of non specifically bound non phosphopeptide, which is better than the HAc solution but still not perfect.

PAGE 87

87 Figure 4 5 showed that 300mM NH 4 HCO 3 solution not only washed off all the non phosphopeptide but also a ll the bound phosphopeptide, which weakens its capability of being a good desorbing solvent. Figure 4 6 presented the results of a two step wash: 6% TFA solution followed by water. We found that this combination could wash off all the bound non phosphopep tide from the surface while leaving about half of the phosphopeptide on the surface. From the results shown in Figure 4 7, we found that 1% H 3 PO 4 solution works very well as the desorbing solvent, for its ability to wash off most of the non specifically bo und non phosphopeptide (about 80%) while only washed off little of phosphopeptide (about 15%) from the surface. This is due to the stronger coordinating ability of phosphate which competes for the zirconium sites, bringing non phosphopeptide down from the surface. Future Work In this work 3 pairs of peptide analogs were used to study the specificity of binding to the zirconium phosphonate surface in model phosphopeptide enrichment studies Peptides used include peptide analogs with different lengths (6AA an d 13AA), different number of phosphorylation sites (1 pY and 3 pY) and different location of phospho rylation site (p S and p Y). According to the results of the SPREE experiment and kinetic analysis, we found that for the 6AA and 13AA pair, the amount of 6 AA bound to the surface is almost half that 13AA bound to the surface The observation supports our calculation in C hapter 3 that the peptide is in excess relative to the number of zirconium sites on the surface With all the available zirconium sites on t he surface occupied the signal change, reflect s the surface mass density change, as the molecular weight of 6AA is about half of that of 13AA. For the 1 pY and 3 pY pair, 3 pY follows the

PAGE 88

88 same pH dependence as 1 pY. However, t he kinetic information indica ts that 3 pY b inds more tightly compared to 1 pY, which was also reflected by the amount washed off in the desorption step. From this we can conclude that there was more than one phosphate group involved in binding to the surface. I n the desorbing solvent study, five solvent systems were investigated in order to find a desorbing solvent that could wash off all the non specifically bound non phosphopeptide while doing no harm to the cova lently bound phosphopeptide, us ing the p60. and p60. phos. pair. These f ive solvents were chosen according to the washing solvent in relative phosphopeptide enrichment with mass spectrometry analysis. From the SPREE results we obtained, acetic acid is not a good desorbing solvent candidate for the reason that it could not wash off even the non phosphopeptide. Ammonium bicarbonate solution works well for washing off the non phosphopeptide, however, all bound phosphopeptide was also washed off, making ammonium bicarbonate also not a good choice. pH 1.0 Tris solution was able to w ash off almost 50% bound non phosphopeptide from the surface and no desorption of phosphopeptide was observed under the same condition. A combination of trifluoroacetic acid and water could wash off all the non phosphopeptide and about 50% of covalently bo und phosphopeptide. Finally, the best desorbing solvent we found was 1% phosphoric acid solution, which washed off about 80% of non phosphopep tide due to the strong competition of the phosphate group for surface zirconium sites and left around 85% phosphop eptide on the surface. Future work will apply the result s of the pH and desorbing solvent study to phosphopeptide enrichment from a protein digest and mass spectrometry analysis. In

PAGE 89

89 addition, a comparison of the selectivity to the most commonly used MOAC material TiO 2 which is believed to have a different phosphopeptide enrichment mechanism, will also be explored.

PAGE 90

90 Table 4 1. Peptides studied with their sequence, molecular weight and the abbreviation in later figures and tables. Peptide Abbreviation Molecular Weight Pair 1 DADE p Y LIPQQGFF 13AA 1622.7 DADE p Y L NH 2 6AA 803.7 Pair 2 TRDI p Y ETDYYRK 1 pY 1703.8 TRDI p Y ETD p Y p Y RK 3 pY 1863.8 Pair 3 KRREILSRRPS p Y RK p Y 1926.2 KRREILSRRP p S YRKC p S 2029.3 Table 4 2 Association and dissociation rate constants of the interactions between phosphopeptides and zirconium phosphonate surface calculated by the Langmuir model. 13AA 6AA pH K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off ( 10 2 min 1 ) K D (10 5 M 1 ) 3.0 2.40.1 7.10.1 3.00.1 2.40.1 7.00.1 2.80.1 1 pY 3 pY pH K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) 4.5 3.50.1 5.00.1 1.40.1 5.20.1 2.1 0.1 0.360. 1 p S p Y pH K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) 7.4 1.60.1 4.80.1 2.90.2 1.60.1 7.00.1 4.50.3

PAGE 91

91 Table 4 3. Association and dissociation rate constants of the interactions between 1 pY/3 pY and zirconium phosphonate surface calculated by the Langmuir model at all four pH conditions. 1 pY 3 pY pH K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) K on (10 3 M 1 min 1 ) K off (10 2 min 1 ) K D (10 5 M 1 ) 3.0 2.80. 1 4.20.1 1.50.1 2.90.1 2.30.1 0.800.1 4.5 3.50.1 5.00.1 1.40.1 5.20.1 2.10.1 0.400. 1 6.0 3.10.1 6.90.1 2.20.1 3.80.1 1.70.1 0.450.1 7.4 1.30.1 6.40.1 5.00.3 1.40.1 1.60.1 1.20.1

PAGE 92

92 Figure 4 1. Fitting results of two parallel SPREE experiments of all peptides: (A) pair 1 peptides; (B) pair 2 peptides; and (C) pair 3 peptides binding to zirconium phosphonate surface at pH= (A) 3.0; (B) 4.5; and (C)7.4.

PAGE 93

93 Figure 4 2. Fitting results of two parallel SPREE experiments of 1 pY and 3 pY at various pH conditions: (A) pH 3.0; (B) pH 4.5; (C) pH 6.0; and (D) pH 7.4.

PAGE 94

94 Figure 4 3. SPREE raw data with HAc as desorbing solvent at pH 3.0: (A) p60. non phos. with 0.1% HAc solution; (B) p60. phos. with 0.1% HAc solution and (C) p60. non phos. with 1% HAc solution. Figure 4 4. SPREE raw data with pH=1.0 10mM Tris solution as desorbing solvent at pH 3.0: (A) p60. non phos. with pH=1.0 10mM Tris solution and (B) p60. phos. with pH=1.0 10mM Tris solution.

PAGE 95

95 Figure 4 5. SPREE raw data with 300mM NH 4 HCO 3 as desorbing solvent at pH 3.0: (A) p60. non phos. with 300mM NH 4 HCO 3 solution and (B) p60. phos. with 300mM NH 4 HCO 3 solution. Figure 4 6. SPREE raw data with 6% TFA solution and water as desorbing solvent at pH 3.0: (A) p60. non phos. with 6% TFA solution + water and (B) p60. phos. with 6% TFA solution + water.

PAGE 96

96 Figure 4 7. SPREE raw data with 1% H 3 PO 4 solution as desorbing solvent at pH 3.0: (A) p60. non phos. with 1% H 3 PO 4 solution and (B) p60. phos. with 1% H 3 PO 4 solution.

PAGE 97

97 CHAPTER 5 DEPOSITION OF A TITANIUM DIOXIDE THIN FILM ON GOLD AND ITS PHOSPHOPEPTIDE ENRICHMENT EFFICIENCY COMPARED TO THE ZIRCONIUM PHOSPHONATE SURFACE Background After studying the selectivity of our zirconium phosphonate surface in phosphopeptide enri chment and elucidating the enrichment mechanism as mainly based on the coordinate covalent interaction between the phosphate group in the peptide and the surface zirconium ion, it will be interesting to compare the phosphopeptide enrichment efficiency to t hat of a metal oxide, the mechanism of which is believed more likely to be electrostatic interaction s between negative charges on peptide phosphate s and side chains and positively charged surface at low pH. T he most commonly used metal oxide in phosphopept ide enrichment, TiO 2 was chosen to compare with the zirconium phosphonate surface. There have been several approaches to deposit TiO 2 thin films on solid substrates, including chemical vapor deposition (CVD) [ 75 ] [ 76 ] pulsed laser deposition [ 77 ] [ 78 ] reactive sputtering [ 79 ] [ 80 ] sol gel deposition [ 80 ] [ 81 ] electrophoresis deposition [ 82 ] [ 83 ] and atomic layer deposition (ALD) [ 84 ] [ 85 ] In our work, we adopted the ALD method and a direct solution deposition. Atomic Layer Deposition Atomic layer deposition (ALD) is a thin film deposition technique that is based on the sequential use of a gas phase chemical process Most of the ALD reactions use two chemic als, so called precursors The two precursors react with a surface one at a time in a sequential manner. By exposing the precursors to the growth surface repeatedly, a thin f ilm is deposited. A typical ALD procedure is shown in Figure 5 1.

PAGE 98

98 ALD is a self limiting technique due to the fact that the amount of film material deposited in each reaction cycle is constant. Because of the self limiting character and surface reactions, ALD film growth makes atomic scale deposition control possible. The film growth can be obtained as fine as ~0.1 per cycle by keeping the precursors separate throughout the coating process. Separa tion of the precursors is usually accomplished by pulsing a purge gas (typically nitrogen or argon ) after each precursor pulse to remove excess precursor from the process chamber. The concept of the ALD process was first proposed by Prof. V.B. Aleskovskii in his Ph.D. thesis in 1952. In the mid 1970s the work of Suntola and coworkers in Finland took the ALD technique into an industrial use and worldwide awareness. Interest in ALD has increased in the mid 1990s and 2000s, with the interest focused on silicon based microelectronics ALD is considered as one deposition method with the greatest potential for producing very thin, conformal films with control of the thick ness and composition at the atomic level. A major driving force for the recent interest is the prospect of scalin g down microelectronic devices. [ 86 ] [ 87 ] ALD can be used to deposit several types of thin films, including various oxides (e.g. Al 2 O 3 TiO 2 SnO 2 ZnO, HfO 2 ), metal nitrides (e.g. TiN, TaN, WN, NbN), metals (e .g. Ru, Ir, Pt), and metal sulfides (e.g. ZnS). In TiO 2 ALD deposition, the Ti precursor used is mostly titanium tetrachloride (TiCl 4 ), titanium tetramethoxide (Ti(OCH 3 ) 4 ) and titanium tetra ethoxide (Ti(OC 2 H 5 ) 4 ), [ 85 ] [ 84 ] [ 88 ] [ 89 ] while the O precursor is usually water. Atomic layer deposition and direct solution deposition were both used to coat TiO 2 thin films on gold surface. Surface characterization was done by AFM, XPS, contact an gle measurement and ellipsometry SPREE experiment s were conducted with

PAGE 99

99 set 1(p60. phos. and p60. non phos.) and set 2 (pMBP and MBP) peptides used in Chapter 3 at pH 3.0 and 7.4. By comparing the peptide binding behavior we would be able to compare the phosphopeptide enrichment efficiency and the interaction mechanism t o that of our zirconium phospho n a te surface. Solution Deposition TiO 2 dep osited directly from solution generally use s silicon and glass substrates. The substrates are often treated with either self assembled monolayers (SAMs) in order to obtain a patterned thin film, or sputtered with a seed layer of TiO 2 to increase film adhesion [ 90 ] The Kunihito Koumoto group at Nagoya University in Japan has published several methods to deposit TiO 2 thin film s on patterned SAMs. In 2001 and 2002 they reported selective and micropat terning of TiO 2 thin film s on SAM using various titanium precusors [ 91 ] [ 92 ] and in the year 2003 they published their method to deposit TiO 2 thin film s at room temperature from aqueous peroxotitanate solution and site selective deposition of TiO 2 in aqueous solution usi ng a seed layer of TiO 2 [ 93 ] [ 94 ] Another Japanese group found that the crystal phase of TiO 2 films deposited directly from aqueous solution can be controlled by initial pH values and precursor solution concentration [ 95 ] Gregory K. L. Goh and coworkers in Singapore applied an amorphous TiO 2 seed layer to glass substrate s and deposit ed TiO 2 from a TiCl 4 and HNO 3 mixture solution [ 96 ] [ 97 ] [ 90 ] Earlier this year, Tredici and et al. reported solution deposition of several micropatterned metal oxide thin films including TiO 2 through a soft lithographic process based on metal loaded hydrogels [ 98 ]

PAGE 100

100 Experimental Section Solution D eposition T he d eposition procedure mostly followed that published by Masuda and el al. in 2002. [ 92 ] The Ti precursor chosen was titanium tetraethoxide (TE) (Sigma) due to the slowest film growth speed, 0.14 nm/min reported in the literature. A 5 nm thick TiO 2 thin film is our goal because the thickness is comparable to our ODPA Zr layer, which will reduce change s in the SPR signal caused by film thickness difference. A 0.1M TE solution was made with anhydrous toluene as the solvent in a glove box under nitr ogen atmosphere. Gold slides were cleaned as describled previously in C ha pter 3 and then immersed in a 2.5 mM 11 mercapto 1 undecanol (MUO) (Sigma) solution in ethanol for 3h to make the surface hydrophilic and rich with OH groups, like the silicon surface used in the literature. The slides were rinsed and ultrasonically washe d in ethanol to wash off the excess amount of MUO. Then the slides were dried with nitrogen gas and brought into the glove box immediately. Slides were soaked in 0.1M TE solution with a time gradient to study the film growing speed and surface morpho logy a t various incubating times of 5 min, 15 min, 30 min, 1 h, 2 h and 3 h. XPS, AFM and ellipsometry experiments were conducted to study surface chemistry, surface morphology and film thickness, which indicate the film growing speed. SPREE experiments were the n performed to investigate the binding behavior of phosphopeptide and non phosphopeptide onto TiO 2 thin film coated gold slides. Atomic L ayer D eposition An atomic layer deposition followed the procedure of Seo et al. in 2004. [ 99 ] The gold slides were also immersed in MUO solution to generate a h ydrophilic surface. The ALD system used in our experiment is a Cambridge Nanotech Fiji 200

PAGE 101

101 equipped with a loadlock. The method of growth was thermal ALD with a chamber and substrate temperature of 200 C. The precursors used for the TiO 2 growth are t et rakis(dimethylamido)titanium (TDMAT) and H 2 O. The cycle times were 0.06 sec H 2 O pulse, 10 sec wait, 0.1 sec Ti pulse, 5 sec wait, and this pattern was repeated for 136 cycles to provide approximately 5nm of TiO 2 Chamber pressure during the growth is appr oximately 200 mTorr. A silicon slide was used as a control. Again, XPS, AFM, ellipsometry and SPREE experiments were done to evaluate the deposition efficiency. SPREE E xperiment In the following SPREE experiments, two sets of peptide, set 1 and 2 fr om C ha pter 3 (see T able 5 1) were used to test the phosphopeptide enrichment efficiency and to compare with the results of the zirconium phosphonate surface at pH 3.0 and 7.4. Zeta P otential M easurement Zeta potential measurement s of TiO 2 nanoparticles were per formed by a pH titration from pH 3.0 to 8.0 on a Malvern Zetasizer (Malvern, UK) in order to investigate the surface charge under certain pH condition. A 1 mg/mL suspension of commercially purchased TiO 2 nanoparticle s (Sigma) was made with Milli Q water an d the auto pH titration was run from pH 3.0 to 8.0 with 0.1 M HCl solution, 0.1 M KOH solution and 0.01 M KOH solution. One data point was taken at every 0.5 unit change in pH. Results and Discussions Surface Characterization An XPS spectrum of TiO 2 thin film on gold with 15min and 3h deposition time is shown in Figure 5 2.

PAGE 102

102 Moreover, by comparing the XPS s pectra taken at various deposition time s we found that the O1s peak obviously split into two peaks at and after 15 min and the relative peak intensity also changes with time. The O1s peak and peak fitting results of MUO coated slide and TiO 2 thin film coated slides with deposition time from 15 min to 3h was shown in Figure 5 3. The results indicate that the TiO 2 thin film is gradually formed on the MUO s urface. However, the Ti:O ratio is nearly 1:3 instead of 1:2. A future condensation step with high temperature treatment may help improve the Ti :O ratio. Peak position and relative intensity for O1s peak in Figure 5 3 is listed in Table 5 2. The O1s p eak indicates that TiO 2 starts to present on the surface after 15 min of deposition time The peak contributed by the OH group of the MUO molecule appears at around 5 30 eV and the Ti O peak appears at about 528 eV. The peak area showed that the relative inte n sity of the Ti O peak increases as longer deposition time ado pted. AFM image s of the MUO surface and TiO 2 surface s with various deposition time s are shown in Figure 5 4. Mean surface roughness calculated by AFM,film thickness achieved with ellipsometry a nd static contact angle measured by contact angle goniometer i s shown in T able 5 3. Figure 5 4 together with Table 5 3 showed that a s deposition time goes longer, TiO 2 first forms little islands on the surface and then clusters of these little islands are formed. XPS result s of a 5nm TiO 2 thin film deposited on gold and a control silicon surface are shown in Figure 5 5.

PAGE 103

103 The Ti2p and O1s peak and fit of TiO 2 on the two surfaces are magn ified and shown in Figure 5 6. The Ti:O element ratio is 1:2.31 and 1:2. 28 respectively, close to the actual ratio 1:2. Both the O1s peak on gold and on the silicon surface showed a single peak and no peak split was observed. The fitted O1s peak yield binding energies of 528. 39 eV on gold surface and 528.27 eV on silicon surf ace. Together with the peak position in solution deposition, we can conclude that the peak is contributed by oxygen of the Ti O bond. However, we see the Au peak for the TiO 2 thin film deposited on the gold surface but no Si peak observed for the silicon s urface. Therefore to see if the film dep osited on gold is homogeneous and uniforml y covered, angle resolved XPS was performed. Angle resolved XPS is a non destructive method for measuring the depth distribution of species present in XPS spectra. By tilting the sample away from or towards the analyzer changing the photoelectron escape depth, the sensitivity to the outermost atomic layers can also be changed. [ 100 ] Angle resolved XPS experiments were conducted at the sample tilting angle of 20 degrees, 45 degree and 70 degree respectively. Figure 5 7 shows the X PS spectra taken at the three angles The principle is shown in Figure 5 8. The angle between the surface and analyzer is usually 45. When the sample tilting angle increases /decreases the mean free escape path for photoelectron becomes shorter /longer making the technique more /less surface sensitive The result indicated that the surface is more likely to be covered by a uniform homogeneous TiO 2 thin film. The AFM image of TiO 2 on the two surfa ces after ALD is shown in Figure 5 9 Mean surface roughness film thickness and contact angle was listed in Table 5 4.

PAGE 104

104 Surface mean roughness shows that the TiO 2 thin film s are very smooth. And th e reason that the film on gold is rougher than the silicon surface is possibly because t hat the film on gold has poorer adhesion compared to silicon surface due to surface character. SPREE E xperiment SPREE experiments were performed in the same way as in C hapter 3. pH 3.0 and 7.4 were chosen to study the binding behavior of peptide set s 1 and 2. As the simple Langmuir mo del we used to fit the raw data in C hapter 3 and C fit the raw data well in this work, raw data is shown here. SPREE results of peptide set 1 and 2 at the TiO 2 surface prepared by solution deposition and atomic layer deposition at pH 3.0 an d 7.4 are shown in F igure 5 10 and 5 11 respectively. We found that for peptide set 1, binding occurs for both phosphopeptide and nonphosphopeptide at pH 3.0 but no binding is observed at pH 7.4 for both peptides. However, peptide set 2 showed the exactly surface, while at pH 7.4, the phosphopeptide and nonphosphopeptide exhibits the same level of binding. The same behavior was observed for TiO 2 deposited with both solution and atomic layer de position methods. And overall binding level is higher for TiO 2 deposited with solution deposition compared to the surface deposited with ALD. This is possibly because the thin film surface by solution deposition is rougher therefore the specific surface ar ea is larger. The observation that set 1 and 2 peptides show different pH dependence when binding to TiO 2 surface s com pared to zirconium phosphonate indicates that the binding mechanism is different tha n the mechanism we proposed in C hapter 3 for zirconi um phosphonate surface. Adsorption to TiO 2 appears to be based primarily on electrostatic

PAGE 105

105 i nteractions. To further test this assumption, we calculated the peptide net charge using the peptide net charge calculation [ 73 ] equation listed in C hapter 3 and the result s are shown in T able 5 5. Zeta P otential M easurement The zeta potential of TiO 2 nanoparticles is shown in F igure 5 12 We found that the isoelectric point for TiO 2 nanopar ticles is around pH 6.0, meaning that at pH lower than 6.0, the surface carries a positive charge; while at pH higher than 6.0, the surface is negatively charged. Together with the peptide net charge calculation results listed in T able 5 5, we conclude tha t the mechanism of TiO 2 surface in phosphopeptide enrichment is likely to be based on electrostatic interactions. At pH 3.0, the surface is positively charged, meanwhile for peptide set 1, the peptides are negatively charged, therefore there is electrostat ic attraction between the surface and the peptide, resulting in peptide binding to the surface. At pH 7.4, both the surface and the peptides are negatively charged, and the electrostatic repulsion between the two prevents the peptides accessing and binding to the surface. For the same reason, with positively charged peptide set 2, when the surface is positively charged at pH 3.0, no peptide binding is observed because of the electrostatic repulsion interaction, while when the surface carries negative charge at pH 7.4, electrostatic attraction favors peptide binding. The selectivity was calculated as the same way in C hapter 3 by and listed in T able 5 6. The selectivity for TiO 2 deposited with solution deposition and w much difference and are both lower compared to the selectivity calculated for the zirconium phosphonate surface at the corresponding pH, further

PAGE 106

106 pointing this attributes to different mechanisms the two surfaces adapt when binding phosphopeptides. Future Work In this work two approaches, solution deposition and atomic layer deposition were applied to deposit TiO 2 thin film onto gold surface in order to compare the phosphopeptide enrichment efficiency of TiO 2 which is the most commonly used material in MOA C for phosphopeptide enrichment, to our zircounium phosphonate surface. After the deposition surface characterization was achieved with XPS, AFM, ellipsometry and contact angle measurement to study the surface chemistry, surface morphology, film thickness and deposited thin film property, respectively. With the help of the surface characterization, we are able to confirm the successful deposition of TiO 2 thin film on gold surface, and the thin film obtained by atomic layer deposition is ultra smooth. Two pa irs of peptide analogs were used to study the phosphopeptide enrichment efficiency of the TiO 2 surface. Set 1 and 2 peptides from Chapter 3, (p60. phos. and p60. set and pMBP and MBP set) were chosen and SPREE experiments were performed at pH 3.0 and 7.4. The SPREE results were quite different compared to the results of the same peptide binding to zirconium phosphonate surface under the same pH condition. We assumed this is due to the different mechanism they adapted in phosphopeptide binding Peptide net c harge was calculated at each pH and zeta potential measurement was conducted to investigate the surface charge of TiO 2 Results showed that when both the peptide and the surface carries like charges, there is no binding to the surface observed. However, wh en the peptides and the surface carries opposite charge, binding to the surface occurs. Therefore we conclude that the mechanism for TiO 2 in phosphopeptide enrichment is based on the electrostatic

PAGE 107

107 interaction between the charged peptide and the surface. A nd the selectivity calculation showed that the TiO 2 surface has poorer selectivity towards phosphopeptide compared to the selectivity of the zirconium phosphonate surface. Future work will include apply high temperature annealing to recover the Ti O bond in the solution deposition method in or der to improve the Ti:O ratio.

PAGE 108

108 Table 5 1. Peptides used in SPREE experiment with TiO 2 thin film s coated gold slides. peptide Abbreviation Molecular Weight Set 1 TSTEPQ p Y QPGENL p60. phos. 1544.5 TSTEP QYQPGENL p60. 1464.5 Set 2 FFKNIVTPR p T PPPSQGK NH 2 pMBP 1913.3 Ac FFKNIVTPRTPPPSQGK NH 2 MBP 1955.3 Table 5 2. Peak position and relative intensity for O1s peak with various deposition time Deposition Time O1s Peak Position(eV) O1s Peak Are a 0 min (MUO surface) 530.61 1752 15 min 530.30 528.31 3744 976 30 min 530.60 528.39 3076 1146 1h 530.51 528.77 3584 4486 2h 530.12 528.51 4131 5421 3h 530.21 528.49 3509 5942 Table 5 3. Mean surface roughness, film thickne ss and static contact angle of MUO surface and TiO 2 with various deposition time. Deposition Time Mean Surface Roughness (nm) Film Thickness (nm) Static Contact Angle ( ) 0 min (MUO surface) 1.072 0.90.2 12.22.0 15 min 1.166 1.20.1 19.71.9 30 min 1.429 2.40.1 26.42.5 1h 2.793 4.60.3 41.62.7 2h 5.818 7.30.4 64.52.4 3h 5.967 6.90.2 68.33.1

PAGE 109

109 Table 5 4. Mean surface roughness, film thickness and static contact angle of TiO 2 thin film on gold and silicon surface after ALD. Surface Mean S urface Roughness (nm) Film Thickness (nm) Static Contact Angle ( ) Si 0.472 5.00.3 51.22.1 Au 0.816 5.20.3 54.51.6 Table 5 5. Estimated peptide net charge for set 1 and 2 peptides at pH 3.0 and 7.4. Net Charge pH p60. phos. p60. pMBP MBP 3.0 0.839 0.073 3.088 3.000 7.4 3.988 2.006 2.009 2.998 Table 5 6. Selectivity of TiO 2 surface with different deposition methods at pH 3.0 and 7.4. Selectivity(TiO 2 Solution) Selectivity(TiO 2 ALD) pH Set 1 Set 2 Set 1 Set 2 3.0 1.22 N/A 0.92 N/A 7 .4 N/A 1.28 N/A 0.87

PAGE 110

110 Figure 5 1. Scheme of one typical ALD cycle.

PAGE 111

111 Figure 5 2. XPS spectrum of TiO 2 deposited on gold with various deposition time: (A) 15 min and (B) 3h and (C) comparison of the Ti 2p3 peak over the time.

PAGE 112

112 Figure 5 3. The O1 s peak and peak fitting results of MUO and TiO 2 with various deposition time on gold. Figure 5 4. AFM image of TiO 2 thin film with various deposition time of (A) 0 min (MUO surface); (B) 15 min; (C) 30 min; (D) 1h; (E) 2h and (F) 3h.

PAGE 113

113 Figure 5 5. XP S spectrum of TiO 2 thin film on gold surface (A) and silicon surface (B) deposited by ALD.

PAGE 114

114 Figure 5 6. Ti2p(A) and O1s(B) peak and fit result of TiO 2 on silicon and gold surface by ALD.

PAGE 115

115 Figure 5 7. XPS spectra of TiO 2 thin film on gold surface a t sample tilting angle of 20 degree, 45 degree and 70 degree respectively. Figure 5 8. Principles of Angle resolved XPS.

PAGE 116

116 Figure 5 9. 2D AFM image of TiO 2 thin film on silicon surface(A) and gold surface(B). Figure 5 10. Set 1 and 2 peptides bin ding to TiO 2 surface prepared by solution deposition at pH=3.0 and 7.4.

PAGE 117

117 Figure 5 11. Set 1 and 2 peptides binding to TiO 2 surface prepared by atomic layer deposition at pH=3.0 and 7.4 Figure 5 12. Zeta potential of TiO 2 nanoparticles at pH 3.0 8.0

PAGE 118

118 CHAPTER 6 DESIGN AND OPTIMIZATION OF A PHOSPHOPEPTIDE ANCHOR FOR SPECIFIC IMMOBILIZATION OF PROTEIN ON ZIRCONIUM PHOSPHATE/ PHOSPHONATE MODIFIED AMINE SURFACE S Background Major challenges have to be addressed related to the design of new concepts in bio materials engineering. This is the case for example in the field of biological microarrays, which are now widely used for the high throughput analysis of biological media for applications in genomics, proteomics or diagnosis among others. [ 101 ] [ 102 ] [ 103 ] [ 104 ] Following the success of oligonucleotide microarrays, t here is no doubt that the use of these biological tools will increase dramatically and become a routine procedure in the near future. One key point in the development of such biosensors is the control and optimization of the binding of the biological probes on solid supports to obtain integrated analytical and miniaturized devices. [ 105 ] Hence, the binding of the probes needs to be efficient and the ir orientation needs to be homogeneous to facilitate the capture of the analyte (i.e. the target) and increase the sensitivity of detection. [ 106 ] To make th ese devices of practical use, st rong binding between the probes and the support is desirable, using chemical processes which are the simplest and cheapest as possible, but robust enough to ensure repeatability of biological experiments. A variety of methods are commonly used for immobil izing proteins for the design of microarrays most often based on nonspecific adsorption (for example via H bonding or electrostatic interactions) of the protein to a solid support. [ 107 ] Simple chemical couplings of reactive groups present on proteins, typically amine or carboxylic acid residues, have also been used to immobilize p roteins onto surfaces suitably functionalized by complementary reactive groups. [ 108 ] Both methods, which require

PAGE 119

119 highly purified proteins, often result in randomly oriented and partially denatured proteins. More recently, the use of recombinant tags allowed proteins to attach to a substrate in a defined orientation, but these more specific interactions used for immobilization of the probes (e.g., glutathione S transferase, oligohistidine) although strong are however reversible [ 109 ] In that context, we were interested in developing a general method where very little chemistry would be used to prepare the substrate for microarrays and where no chemical modification would be needed on proteins or biomolecules to be immobilized in order to avoid their d enaturation. A few years ago, we reported the use of zirconium phosphonate monolayers coated on glass slides as reactive surfaces able to provide covalent attachment of phosphate terminated biological probes. This approach is very attractive since phosphor ylation can be achieved using enzymes without affecting the functionality of the modified biological probes. This original concept was first validated in phosphate terminated oligonucleotides for which specific binding to the zirconium phosp honate surface occurred This was then successfully extended to the immobilization of phosphate terminated ds DNA More recently our goal was to investigate the potential of this technology for the design o f protein microarrays, which is a more challengin g issue needing original solutions to allow stable surface attachment of proteins while controlling their orientation. In order to preserve the high affinity binding of protein probe s, once immobilized, an innovative solution consists in the attachment of the probes onto the surface via a short phosphorylated peptide sequence fused at the N or C terminus of the proteins that would make it able to bind to the active surface (OPO 3 (Figure 6 1)

PAGE 120

120 This a pproach has great potential since fusion of such peptide tags should in principle be feasible for any type of proteins expressed by genetic engineering and this was confirmed by us in a recent exploratory study. [ 109 ] In this work, the best conditions to provide high binding a ffinity of the phosphopeptide for the zirconium phosphonate support was investigated, in particular spotting conditions (pH and nature of the buffer) and the number of phosphorylated amino acid units present in the phosphopeptide tag. The resulting phospho rylated peptide tags, either pure or fused to affinity proteins were found to specifically bind on the zirconium phosphonate surface, and the resulting microarray was shown to be highly efficient in terms of both sensitivity and signal to noise ratio. Exp erimental Section Materials and S olution s Zirconyl chloride octahydrate, 2,4,6 collidine, phosphorous oxytrichloride (POCl 3 ), octadecyltrichlorosilane (OTS) (3 a minopropyl)triethoxysilane(APTES) 4 d imethylaminopyridine (DMAP), casein and s treptavidin we re used as received from Aldrich. Reagents were of analytical grade and were used as r eceived from commercial sources Glass slides were purchased from Goldseal (Horses thickness 0.93 1.05 mm). Aminopropylsilane glass slides (SuperAmine 2) were purchased from Arrayit (Sunnyvale, CA). Phosphorylated peptides were purchased from ProteoGenix (Schiltighem, France) with the name and sequence listed in Table 6 1 with the bold letter p indicating the location of the phosphorylation site The purity of the peptides was verified by the mass spectroscopy spectra that the manufacturer provided. The buffer used throughout the experiments consisted of 10 mM glycine, 10 mM Tris base, 10 mM 2 (N morpholino)ethanesulfonic acid(MES) and 10 mM acetic aci d, which

PAGE 121

121 has a pH range of 2 10. Four pH conditions were chosen according to our previous experiments in Chapter 3: pH 3, 4.5, 6, and 7.4) APTES slides preparation APTES was stored in the glove box under N 2 atmosphere after purchase Glass slides were cle aned with HNO 3 : HCl=3:1 solution, followed by five times for 1 min with Milli Q water in an ultrasonic bath (50 kHz) and dried with a N 2 flow Cleaned slides were etched with trifluoroacetic acid for 90 min and stored in vacuum for at least 10 h. Silanizat ion of cleaned slides w as then carried out in Petri dishes that had been washed in "piranha bath" (3.5% H 2 O 2 in 18 M H 2 SO 4 ), followed by rinsing with water and acetone. Slide s were then silanized with APTES (2% in 95% aqueous acetone) for 3 min followed by washing with acetone (12 times, 5 min each). Curing of the silane linkages was carried out in an oven at 110C for 1 h [ 110 ] Surface characterization was done by AFM and XPS and compared to that of the SuperAmine surface. The comparison of the SuperAmine surface and APTES surface is shown in Fi gure 6 2. Zirconium phosphate modified surface SuperAmine 2 slides ( used as received ) and prepared APTES slides were soaked in a 1:1 solution of 2,4,6 collidine (20 mM) and POCl 3 (20 mM) in anhydrous acetonitrile and gently rocked for 16 h. The slides wer e rin s ed with acetonitrile and dried with Argon The phosphonate terminated samples were then exposed to a 25 mM ZrOCl 2 8H 2 O solution in Milli Q water overnight. The slides were rins ed with Milli Q water and stored under argon in Milli Q water. Zirconium p hosphate/ phosphonate modified surface SuperAmine 2 slides ( used as received ) and prepared APTES slides were soaked in a 1:1 solution of DMAP and an organic compound with a terminal phosphate

PAGE 122

122 compound D (structure shown in Figure 6 3 ) synthesized by our c ollaborators in France in anhydrous dichloromethane and gently rocked for 16 h. The slides were rin s ed with dichloromethane and then were again annealed at 20 0 C for 24 h The phosphonate terminated samples were then exposed to a 25 mM ZrOCl 2 8H 2 O solutio n in Milli Q water overnight. The slides were rin s ed with Milli Q water and stored under argon in Milli Q water. The two amine surfaces after POCl 3 and compound D treatment is compared and shown in Figure 6 4. Microarray spotting and incubation conditions The modified slides were washed once with ultrapure water and dried by centrifugation (1500 rpm, 1 min). The slides were then spotted with a non contact spotter (sciFlex Arrayer S3, Scienion). The spotted slides were placed 1 h in an incubation chamber at 37 C. To passivate the unspotted areas, slides were treated after spotting with a solution of 0.3 wt casein in a TBS (Tris buffered saline) solution of TrisHCl (20 mM) and NaCl (150 mM) at pH 7.4. Incubation was performed by applying an Alexa 647 labe led Streptavidin solution ( 1 M in TBS 0.3 wt; casein) to the substrate for 1 h at room temperature in the dark. Microarrays were washed 3 times with TBS for 5 min, then once with ultrapure water. Finally, the slides were spun dry by centrifug ation at 1 500 rpm for 1 min. All washes and incubations were performed in small staining jars at room temperature on an oscillating shaker. Microarray analysis All microarrays were scanned on a Scanarray Gx apparatus (Perkin Elmer) with a laser power and gain value of 60. Suitable excitation wavelength and emission filter were used to detect Alexa 647: 650 nm (excitation), 665 nm (emission). The location of

PAGE 123

123 each analyte spot on the array and measurement of the fluorescence intensities was performed using the Genepix mapping software (Axon laboratories, Palo Alto, CA). Surface Plasmon Resonance Enhanced Ellipsometry (SPREE) experiments SPREE measurements were performed with the same experiment al setup as in C hapter 3. The purpose is to compare the results to that of fl uorescence microarray Gold slides were modified with zirconium phosphonate as described in C hapter 3. The c oncentration of the peptide solution is still 1mg/mL and the volume applied is 100 L per experiment. The b uffe rs used in this case are the same buffer s used for microarray experiments. After the surface was saturated with the peptides follo wed by a vigorous buffer wash, s treptavidin was introduced into the system to investigate any furt her binding. Results and D iscussion Surface Characterization AFM and XPS characterization results for the SuperAmine surface and APTES slides a re shown in Figure 6 5 and 6 6 The surface roughness for the SuperAmine surface and APTES are 1.698 nm and 1.83 2 nm respectively and the XPS results shown similar results, indicating ou r bench prepared APTES slides are similar to the commercially purchased SuperAmine slides XPS multiplex results of zirconium and phosphorous on the zirconium phosphate and phosphona te modified SuperAmine surface s are shown in Figure 6 7 The Zr and P ratio is 1.6:1 and 1.3:1 respectively. Phosphopeptide B inding on Zr Phosphate/Phosphonate S urfaces Surface affinity of all five peptide candidates is evaluated by fluorescence spotting experiments as well as SPREE experiments. Results are listed and discussed below.

PAGE 124

1 24 Spotting experiments Zirconium phosphonate monolayer s coated on glass slides were made by self assembly from an adapted procedure. [ 111 ] Briefly, commercially available amine coated slides ( SuperAmine 2 ) and bench prepared APTES slides were treated with POCl 3 and compound D and subsequently dipped into an aqueous solution of zirconium oxychloride. Then the resulting zirconium phosphonate surfaces were used to investigate their interaction with short phosphopeptide segments made of 12 amino acids versus a non phosphorylated analo gue For that matter, the number of phosphate groups on the peptide sequence was varied (from 0 to 4 phosphate serine moieties (pS) see Table 6 1 ) and the influence of spotting conditions (pH) was also studied All peptides were biotinylated on the C ter minal to allow quantification of the immobilized probes up on incubation with streptavidin labelled with a fluorophore (Alexa Fluor 647) The interaction of the selected peptides (Table 6 1) with the zirconium phosphonate substrates was compared by spotting a solution of the peptides at fixed concentrations in a buffer at different pH Two identical sets of microarrays were spotted on each slide to account for any surface inhomogeneities P rior to incubation with labeled streptavidin, casein, which has a high phosphate content was used to passivate the non spotted areas and prevent nonspecific protein binding. [ 48 ] Figure 6 8 depicts the spotted pattern at the end of the incubation process The fluorescence signal was quantified on a fluorescence reader. The quantified signal of 0P 4P spotted at pH 3.0 7.4 on POCl 3 and compound D treated two amine slides, SuperAmine slide and APTES slide ar e shown in Figure 6 9 to 6 12 It is important to note that over the whole pH range, the non phosphorylated peptide (0P) binds very poorly to the Zr/phosphonate surface, in sharp contr ast to

PAGE 125

125 phosphorylated peptides This observation strongly suggests that immobilization of the peptide s (OPO 3 bonds Moreover, none of the peptides were found to bind when the same experiment was performed on the bare superamine slide or on SuperA mine slide s only treated with POCl 3. In addition, t he amount of phospho peptide bound to the surface decreased when the spotting concentration varied from 10 M to 1 M Interestingly, t he binding of the phosphopeptides 1P to 4P decreased as the pH value increased. At pH 6 .0 and 7.4, weak fluorescent spots could be observed after incubation with labeled streptavidin indicating relatively poor immobilization on the zirconium phosphonate surface ( F igure 6 8 ). By contrast, a strong binding of the four phosphopeptides was present at pH 3 .0 and to a lesser extent at pH 4.5, evidenced by highly fluorescent spots with a ca. 15 fold intensity increase compared with pH 6 .0 and 7.4 This result is consistent with the pH dependence we observed in C hapter 3 with set 1 peptides, containing only acidic amino acid side chains ( g lutamic a cid ( E ) and a spartic a cid ( D ) ) in the sequence. The peptides used in this project (see Table 6 1) all have several D a nd E residues in the sequence and is very acidic, it is not surprised to find that it follows the same pH dependence as set 1 peptides. Occasionally, peptide 1P was found to have more surface affinity than 2P, leading us to question whether the additional phosphate group in 2P, which is relative ly close to the b iotin end, would pull the b iotin close to surface, limi ting the s treptavidin molecule access to b iotin. This requires further experiment to prove. SPREE experiments In order to complement spotting experiments binding of the peptides on the Zr/phosphonate surface at pH 3.0 7.4 conditions were also probed using the SPREE

PAGE 126

126 technique ( F igure 6 11) SPREE was also used to monitor the binding of s treptavidin to the peptide at pH 3.0 for all the peptides From the overall signal change for all peptides under various pH conditions shown in F igure 6 13 we can again observe a pH dependence, the amount of peptide binding to the surface decreases as the pH increases for a given peptide. M eanwhile at the same p H, 4P binds slightly more to the surface compared to other peptides. The n on phosphorylated peptide 0P has poor surface binding under all pH conditions. We expect a signal increase after introducing the phosphopeptide into the system. Since the peptide was modified with b iotin at the N terminal, another signal increase may be observed when flowing s treptavidin through the cell after the traditional buffer washing step. When trying to dissolve s treptavidin in water and pH 3.0 super buffer, it was found that the protein tends to aggregate and deactivate under such condition s Therefore we dissolve s treptavidin in pH=7.4 super buffer and in each experiment with pH 3.0, after the buffer wash ing step, before adding s treptavidin solution, we run another washing st ep by flowing pH 7.4 super buffer through the cell first and give enough time for the signal to reach a stable level again, in order to minimize possible surface refractive index change caused by changing solution. This is also similar to the pH 7.4 washin g step in the spotting experiment, making the final SPREE signal change The poor binding affinity of the non phosphorylated peptide (0P) was also observed since the amount of peptide bound to the surface after the rinsing ste p was found to be pretty low, with a final signal change o 10 0.25 (Figure 6 1 4 ) Regarding the four phosphopeptides 1P, 2P, 3P and 4P at pH 3 .0 the Zr/phosphonate

PAGE 127

127 surface exhibits a high er and specific binding than 0P as evidenced by the amplitude change of 0.85 Also, the amount of peptide washed off in the pH 7.4 washing step was larger for 0P compared to the other 4 phosphopeptides, indicating that the binding of 0P is mostly non specific binding caused by electrostatic interaction between the positively charged surface and negatively char ged peptide side chain at pH 3.0. Th e SPREE results are in good agreement with the spotting experiments and confirm that working at pH 3 .0 result s in the specific anchoring of phosphopeptides on to zirconium phosphonate surfaces while at higher pH, no sign ificant binding occur s The reason why 1P 4P exhibits similar level of surface binding is that the surface is mass saturated by these peptides. However, when adding s treptavidin to the system, interesting behavior was observed. The protein binds to 0P and 4P is less compared to the amount binds to 1P, 2P and 3P (see Figure 6 1 5 the arrow indicate s the addition of s treptavidin ). It is easy to understand there is no s treptavidin bi nding for peptide 0P since before adding the protein, the pH 7.4 buffer washin g step already washed off most non specifically bound peptide. However, for 4P, the reason why there is almost no s treptavidin binding could be due to the fact that all the phosphorylation sites contributes in surface binding and therefore the biotin is pu lled down to the surface, making it difficult for a big protein molecule, namely the s treptavidin to access and bind. Although for 2P and 3P, there is also a phosphorylation sit e close to biotin, there are three and two amino acid residues in between the p hosphorylation site and other phosphorylation sit es, while for 4P there is only one amino acid residue separating the

PAGE 128

128 phosphorylation site close to biotin and other phosphorylation sites, so we guess the amino acid residues in between the phosphorylation s ites could cause some steric hindrance for phosphate groups binding to the surface. The more amino acid residues in between the phosphorylation sites, the more difficult it is for all phosphate groups to bind to the surface. This may be proved by experimen ts with new 2P and 3P peptide s with all of the phosphorylation sites together, for example the sequence could be Biotin REEDDDD pS pS EDE for 2P and Biotin REEDDD pS pS pS EDE for 3P respectively. Influence of the number of phosphate moieties in phosphope ptides on their binding affinity Results from the spotting and SPREE experiments were also analyzed to investigate whether the number of serine phosphate blocks in the phosphopeptide sequence has a significant influence on their affinity for the Zr/phosph onate supports. Taking account of the previous observations, comparison of the binding ability of the five peptides (Table 6 1) was limited to higher sp otting concentration (5 and 10 M, respectively ) For all pH, the spotting experiments show that the 3P and 4P phosphopeptides bind more strongly onto the surface since the fluorescence intensities are ca. 2 or 3 times higher than for the 1P and 2P phosphopeptides. The poor binding ability of the non phosphorylated analogue (0P) was again clearly apparent. From the SPR EE data recorded at pH 3 .0 ( Figure 6 14 ), the amount of phosphopeptide bound to the Zr/phosphonate support was quite similar for the four phosphopeptides 1P to 4P due to surface mass saturation and the kinetics of the fixation is rapid. L ow bi nding of the non phosphorylated peptide was observed 0. 35 versus 6 0.8 for 1P to 4P) as a result of n on specific weak interactions (electrostatic interaction H bonding) between 0P and the support. While the binding

PAGE 129

129 properties of the four phosphopeptides were found to be similar in SPREE expe riments, the best results obtained in the spotting experiments for 3P and 4P by comparison with 1P and 2P could reflect the higher binding stability of the two former peptides in the presence of competitors, in particular during the casein mediated block ing step, due to a higher number of phosphate binding sites. Future Wor k In this work inorganic and organic treatment s were applied to coat a phosphate monolayer on the surface of commercially available SuperAmine slide and bench prepared amin e slide The n the surface was incubated with zirconium solution to generate a zirconium phosphate (inorganic treatment)/phosphonate (organic treatment) surface. Surface characterization was done by AFM and XPS to compare the bench prepared amine slide to purchased Su perAmine slide s Five peptide candidates with the number of phosphorylati on sites varying from 0 to 4 were chosen to study their binding affinity to the zirconium phosp hate/phosphonate surface under four pH conditions: 3.0, 4.5, 6.0 and 7.4. Fluorescence s potting experiment s and SPREE experiment s were performed to investigate the surface binding affinity. Surface characterization and fluorescence spotting experim ents both proved that the bench prepared amine slide surface is similar to the surface of the c ommercial ly purchased SuperAmine slide. F rom the spotting expe riment results we can conclude that phosphorylation of peptides makes them able to bind strongly to Zr phosphonate surfaces via the formation of P O Zr bond. The surface affinity of 0P is very poor under all pH conditions compared to other phosphorylated peptides. For 1P 4P, the surface binding affinity increases as the number of phosphorylation sites increases, and decreases as the pH increases.

PAGE 130

130 The SPREE experiment s for all peptides under fo ur pH conditions further confirmed that the surface binding affinity for a given peptide decreases as the pH increases, and non phosphorylated peptide 0P has poorer binding under all pH conditions. However, at any pH, the surface binding difference among d ifferent phosphorylated peptides was not as obvious as what was observed in the spotting experiments. The reason is that SPREE is a surface mass density sensitive technique and the surface is mass saturated with the peptides. The SPREE experiment at pH 3.0 showed almost no s treptavidin binding for 4 P. The reason might be that all 4 phosphate groups are involved in surface binding thereby pulling the b iotin close to surface, preventing streptavidin access and bind ing The reason why this behavior is not obse rved with 2P and 3P is probably the steric hindrance effect caused by the amino acid residues in between the phosphate group close to the b iotin end and other phosphate groups. This assumption requires further experiments with peptides of different sequenc es to prove. We conclude that the presence of multiple phosphate groups leads to improved results, though perhaps not mandatory providing efficient and stable covalent binding of phosphopeptides. The appropriate peptide candidate for protein immobilization was chosen to be 4P based on the fluorescence spotting experiment due to the fact that the amount of bound peptide remain ing on the surface is most for 4P after several vigorous washing steps. Though based on SPREE experiments there is less s treptavidin b inding observed, the reason is more likely to be the steric hindrance and kinetic issues, instead of surface binding ability. pH 3.0 appears to be the best pH condition to work with in future protein immobilization experiment s

PAGE 131

131 Future work of this project includes more spotting and SPREE experiments with 2P and 3P with new sequences, for example, with all phosphate groups together and relative ly far away from the b iotin end to investigate the relationship between the location of phosphorylation sites and t he s treptavidin binding behavior We are also waiting for the protein spotting experiment results with 4P at pH 3.0, which will be done by our collaborators at Universit de Nantes in France.

PAGE 132

132 Table 6 1. A mino acid sequences used for the binding study wi th zirconium phosphonate substrates. Name Amino Acid Sequence Molecular Weight 0P Biotin REEDSDSDSEDE 1638.57 1P Biotin REEDEDDD p S EDE 1788.6 2P Biotin REED p S DDD p S EDE 1826.54 3P Biotin REED p S DD p S p S EDE 1878.51 4P Biotin REED p S D p S p S p S EDE 1930.48

PAGE 133

133 Figure 6 1 : Covalent attachment of proteins on Zr phosphate /phosphonate surfaces via phsophopeptide anchors

PAGE 134

134 Figure 6 2. Comparison between commercially purchased SuperAmine surface(A) and APTES(B) treated glass surface. Figure 6 3. Structure of the organic compound D used to prepare zirconium phosphonate modified amine surface. Figure 6 4. Scheme of the two amine surfaces after POCl 3 (A) and compound D (B) treatment.

PAGE 135

135 Figure 6 5. AFM (A) and XPS (B) survey spectra of SuperAmine surface.

PAGE 136

136 Figure 6 6. AFM (A) and XPS (B) survey spectra of APTES surface.

PAGE 137

137 Figure 6 7. XPS result of Zr and P signal on zirconium phosphate (A) an d zirconium phosphonate (B) modified SuperAmine surface. Figure 6 8 : Fluorescence analysis of Zr/phosphonate substrates spotted with peptide 0P, 1P, 2P, 3P and 4P at different pHs. For each condition, three spotting concentrations were used (from the l eft to the right: 10, 5 and 1 M). After spotting saturation and rinsing, the substrates were incubated with streptavidine labelled with a fluorophore ( Alexa Fluor 647 )

PAGE 138

138 Figure 6 9. Fluoresence quantification as a function of pH for peptides 0P 4P sp otted at various concentrations (1, 5 and 10 M) on POCl 3 treated SuperAmine slide.

PAGE 139

139 Figure 6 10. Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on compound D treated Superamine slide

PAGE 140

140 Figure 6 11. Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on POCl 3 treated APTES slide.

PAGE 141

141 Figure 6 12. Fluoresence quantification as a function of pH for peptides 0P 4P spotted at various concentrations (1, 5 and 10 M) on compound D treated APTES slide.

PAGE 142

142 Figure 6 13 : Overall signal change of 0P 4P binding to zirconium phosphonate surface at pH 3.0 7.4 in super buffer. Figure 6 14 : Raw SPREE data of 0P 4P binding to zir conium phosphonate surface at pH 3.0 in super buffer with a pH 7.4 washing step (plots are vertically offset for better comparison).

PAGE 143

143 Figure 6 15 Raw SPREE data of 0P 4P binding to zirconium phosphonate surface at pH 3.0 in super buffer with addition o f Streptavidin afterwards (plots are vertically offset for better comparison)

PAGE 144

144 CHAPTER 7 SUMMARY AND CONCLUSIONS The focus of the research presented in this dissertation was the development of biosensors for the immobilization of phosphopeptides and pro teins. For the systems discussed in different c hapters, zirconium phosphonate surface was used as a support for biomolecules immobilization. This surface was first reported by our group and collaborators and has been a popular use for biomolecules immobili zation such as DNA and proteins. In C hapter 3, efficiency of zirconium phosphoante surface in phosphopeptide enrichment is investigated It is becoming obvious that protein phosphorylation plays an important role in the regulation and metabolism of many ce ll processes. Ongoing research on phosphopeptide enrichment approaches helps study protein phosphorylation quantitatively and protein phosphorylation regulation mechanisms. Therefore, our zirconium phosphonate surface was applied to enrich phosphopeptides. pH dependence was first studied by observing the binding behavior of various peptide analog pair s consisting of one phosphopeptide and one non phosphopeptide with the same amino acid sequence. pH dependence is of great importance to us as it could be a guide for certain peptide enrichment. Selectivity of the zirconium phosphonate surface towards phosphopeptide over non phosphopeptide and kinetic information for phosphopeptide binding process has been looked into and a peptide surface interaction mechanism based on different pH dependence of the peptides was proposed. The rationality of the mechanism was testified by a mathematical model. It turned out to be that the surface binding is mostly cont ributed by covalency between the phosphate group and the surface zirconium ion electrostatic interaction between charged residue side chains in the sequence and the surface at

PAGE 145

145 certain pH also influences the binding As well as the design of biosensors, ch aracterization by many techniques was an important aspect of this research work. As demonstrated in this work, it was possible to gain insight in phosphopeptide enrichment efficiency and pH dependence of zirconium phosphonate surface in phosphopeptide enri chment. After figuring out pH dependence and phosphopeptide surface interaction mechanism, more phosphopeptide analog pairs were studied at chosen optimal pH based on the peptide sequence ( C hapter 4) including phosphopeptides with the similar sequence bu t different lengths (one is a fragment of the other), phosphopeptides with the same sequence but different number of phosphorylation sites, as well as phosphopeptides with similar sequence but different phosphorylation site, in order to further study the a bility of the zirconium surface in distinguishing those phosphopeptide analogs. Also, to find a desorbing solvent which is able to wash off the non specifically bound non phosphopeptide is also our interest and goal. Five different desorbing solvent was st udied with a phosphopeptide and non phosphopeptide analog pair. The results were also shown in C hapter 4. The most effective desorbing solvent is found to be 1% phosphoric acid solution, which could wash off almost 80% of non specifically bound non phospho peptides while not affecting covalently bound phosphopeptides much. This solvent can be used in other phosphopeptide enrichment procedures as well as future mass spectrometry analysis of phosphoproteomics. Titanium dioxide is the most commonly used materia l in metal oxide affinity chromatography in phosphopeptide enrichment, with the mechanism believed to be the electrostatic interaction between the negatively charged phosphate groups and the

PAGE 146

146 positively charged metal oxide surface at low pH. This leads to o ur idea to compare the phosphopeptide enrichment efficiency of titanium dioxide with our zirconium phosphonate surface, since the two methods are based on different interaction mechanisms. Due to the fact that we want to apply the SPREE technique for the c omparison, the most important step is to coat a TiO 2 thin film on gold surface. In C hapter 5, t wo approaches, solution deposition and atomic deposition were adapted to deposit TiO 2 thin film onto gold surface. Surface characterization approved the success of thin film deposition SPREE experiments were then performed with two phosphopeptide and non phosphopeptide pairs at two selected pH, at which condition the two peptide set are believed to carry opposite net charge Zeta potential measurement of TiO 2 nan oparticles was done to determine the surface charge at certain pH. The peptide binding behavior together with peptide and surface charge at given pH showed that the mechanism for TiO 2 in phosphopeptide enrichment is truly based on electrostatic interaction and the calculated selectivity of this approach towards phosphopeptide is lower compared to our zirconium phosphonate surface. Recently, proteins immobilization has become of high interest in the understanding of biological mechanisms. Chapter 6 describe d the design of peptide anchors to immobilize proteins onto solid supports Two amine surfaces, commercially purchased SuperAmine surface and bench prepared APTES slides, were used as the support. Organic and inorganic treatment was applied to generate a z irconium phosphate and pho sphonate surface respectively. Four pH conditions were chosen according to our pH dependence study in C hapter 3. Five peptide candidates, with number of phosphorylation sites varying from 0 4, w ere utilized to investigate their

PAGE 147

147 af finity to the zirconium phosphate/phosphonate surface under various conditions. Fluorescence spotting experiment as well as SPREE experiment was performed to examine the surface affinity of all peptides. Peptide candidate with the best surface affinity and the corresponding pH was chosen for future protein immobilization test. To date, all the phosphopeptide enrichment work is conducted under acidic pH conditions and no pH influence was reported. Our pH dependence study allows us to investigate the influenc e of pH on phosphopeptide enrichment using zirconium phosphonate surfaces, providing a new perspective in guiding future phosphopeptide enrichment. For example, if we want to isolate certain acidic peptides from a peptide mixture, we can work at acidic pH like pH 3.0 since that the most phosphopeptides are recovered at this pH Meanwhile if basic peptides are our interest, basic pH such as pH 7.4 is recommended because there is not much difference in phosphopeptide affinity to the surface while least non sp ecific binding from acidic side chains and no non specific binding from basic side chains occurs at this pH. The simulation function based on the possible factors that could influence peptide binding to the surface simulates the experimental results well a nd proves that the factors we use to build the function is reasonable. Following experiments with peptide analogues of various lengths, numbers of phosphorylation sites and locations of phosphorylation sites further help us to evaluate the surface in disti nguishing peptide analogues under certain p H conditions, which may also provides guidance to future phospho proteomic study Since we propose the phosphopeptide and zirconium phosphonate surface interact with each other via the coordinate covalent bond for med between the peptide phosphate and the surface zirconium ion, comparing the phosphopeptide enrichment

PAGE 148

148 behavior of our zirconium phosphonate surface to that of metal oxide surface, the mechanism of which is believed to be based on the electrostatic inter action between the negatively charged peptide phosphate and positively charged surface under acidic pH condition, will be significant. The different binding behaviors under certain pH conditions shown by the same peptide onto zirconium phosphonate surface and TiO 2 surface, the most commonly used metal oxide for phosphopeptide enrichment, indicates that the metal oxide surface does adopt a different mechanism in capturing phosphopeptides. This explains from the mechanism the reason why zirconium phosphonate surfaces have better selectivity towards phosphopeptide compared to TiO 2 surfaces in many cases. When further apply the zirconium phosphonate surface in protein immobilization, using a phosphorylatable peptide tag enables the feature of immobilizing prot ein in certain orientations on a solid surface while retaining their activities. Investigating the surface affinity of peptide tag candidates with various numbers of phosphorylation sites helps us find out how many phosphates are necessary to efficiently i mmobilize protein on the surface. It turned out that the coordinate covalent bond between the phosphate group and the surface zirconium ion is so strong that even one phosphate group in the peptide sequence provides good surface affinity. Presence of mult iple phosphate groups leads to higher surface affinity, though not mandatory to provide efficient and stable covalent binding of phosphopeptides.

PAGE 149

149 APPENDIX pI OF All PEPTIDES The isoelectric point of the peptide pI is the pH at which the net charge of the peptide is zero. The pI value s calculated for all the peptides used in this work are listed in Table A 1 Table A 1 pI values of all peptides used. Peptide Abbreviation pI p60. phos. 2.05 p60. 2.92 pMBP 10.6 MBP 14.0 1 pY 4.28 3 Y 6.89 3 pY 3 .09 pDSIP 2.04 DSIP 2.90 pTP1 2.04 pTP2 2.06 6AA 2.61 13AA 2.02 p S 10.7 p Y 11.8 0P 2.64 1P 2.07 2P 1.76 3P 1.58 4P 1.44

PAGE 150

150 LIST OF REFERENCES [1] G. Burnett, E. P. Kennedy, J. Biol. Chem. 1954 211 969 980. [2] E. H. Fischer, E. G. Krebs, J. Biol. Chem. 1955 216, 121 132. [3] E. W. S utherland W. D. W osilait Nature 1955 175 169 170 [ 4] J. Ptacek, G. Devgan, G. Michaud, H. Zhu, X. Zhu, J. Fasolo, H. Guo, G. Jona, A. Breitkreutz, R. Sopko, Nature 2005 438 679 684 [5] H. Zhou, S. Xu, M. Ye, S. Feng, C. Pan, X. Jiang, X. Li, G. Han, Y. Fu, H. Zou, J. Proteome Res. 2006 5 2431 2437 [6] S. Feng, M. Ye, H. Zhou, X. Jiang, X. Jiang, H. Zou, B. Gong, Mol. Cell. Proteomics 2007 6 1656 1665 [7] G. Lee, R. Thangavel, V. M. Sharma, J. M. Li tersky, K. Bhaskar, S. M. Fang, L. H. Do, A. Andreadis, G. Van Hoesen, H. Ksiezak Reding, J. Neurosci. 2004 24 2304 2312 [8] J. Avila, FEBS letters 2006 580 2922 2927 [9] J. Lu, Y. Li, C. Deng, Nanoscale 2011 3 1225 1233 [10] H. K. Kweon, K. Hka nsson, Anal. Chem. 2006 78 1743 1749 [11] H. Zhou, R. Tian, M. Ye, S. Xu, S. Feng, C. Pan, X. Jiang, X. Li, H. Zou, Electrophoresis 2007 28 2201 2215 [12] I. L. Batalha, C. R. Lowe, A. C. A. Roque, Trends Biotechnol 2012 30 100 110 [13] J. Wei, Y Zhang, J. Wang, F. Tan, J. Liu, Y. Cai, X. Qian, Rapid Commun. Mass Spectrom. 2008 22 1069 1080 [14] L. Zhao, R. Wu, G. Han, H. Zhou, L. Ren, R. Tian, H. Zou, J. Am. Soc. Mass Spectrom. 2008 19 1176 1186 [15] Y. Xue, J. Wei, H. Han, L. Zhao, D. Cao J. Wang, X. Yang, Y. Zhang, X. Qian, J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2009 877 757 764 [16] T. Hoang, U. Roth, K. Kowalewski, C. Belisle, K. Steinert, M. Karas, Anal. Chem. 2009 82 219 228 [17] T. E. Thingholm, O. N. Jensen, M. R. Larsen, Proteomics 2009 9 1451 1468. [18 ] J. Porath, J. A. N. Carlsson, I. Olsson, G. Belfrage, Nature 1975 258 598 599

PAGE 151

151 [19 ] H. Zhou, M. Ye, J. Dong, G. Han, X. Jiang, R. Wu, H. Zou, J. Proteome Res. 2008 7 3957 3967 [20 ] Z. Yu, G. Han, S. Sun, X. Jiang, R. Chen, F. Wang, R. a. Wu, M. Ye, H. Zou, Anal. Chim. Acta 2009 636 34 41 [21 ] F. Wolschin, S. Wienkoop, W. Weckwerth, Proteomics 2005 5 4389 4397 [22 ] J. D. Dunn, G. E. Reid, M. L. Bruening, Mass Spectrom. Rev. 2010 29 29 54 [23 ] C. T. Chen, Y. C. Chen, Anal. Chem. 2005 77 5912 5919 [24 ] S. S. Liang, H. Makamba, S. Y. Huang, S. H. Chen, J. Chromatogr. A 2006 1116 38 45 [25 ] C. Klemm, S. Otto, C. Wolf, R. F. Haseloff, M. Beyermann, E. Krause, J. Mass Spectrom. 2006 41 1623 1632 [26 ] A. Leitner, Trends Anal. Chem. 2010 29 177 185 [27 ] H. F. Wu, K. Agrawal, K. Shrivas, Y. H. Lee, J. Mass Spectrom. 2010 45 1402 1408 [28 ] S. S. Park, S. Maudsley, Anal Biochem 2011 409 81 88 [29 ] S. B. Ficarro, M. L. McCleland, P. T. Stukenb erg, D. J. Burke, M. M. Ross, J. Shabanowitz, D. F. Hunt, F. M. White, Nat Biotech 2002 20 301 305 [30 ] T. E. Thingholm, O. N. Jensen, M. R. Larsen, Proteomics 2009 9 1451 1468 [31 ] A. Clearfield, J. A. Stynes, J. Inorg. Nucl. Chem. 1964 26 117 129 [32 ] S. Xu, J. C. Whitin, T. T. S. Yu, H. Zhou, D. Sun, H. J. Sue, H. Zou, H. J. Cohen, R. N. Zare, Anal. Chem. 2008 80 5542 5549 [33 ] G. Nonglaton, I. O. Benitez, I. Guisle, M. Pipelier, J. Lger, D. Dubreuil, C. Tellier, D. R. Talham, B. Bujoli, J. Am. Chem. Soc. 2004 126 1497 1502 [34 ] L. Zhao, R. Wu, H. Zou, Methods Mol. Biol. 2011 790 215 222 [35 ] H. Chen, M. Lee, S. Choi, J. H. Kim, H. J. Choi, S. H. Kim, J. Lee, K. Koh, Sensors 2007 7 1091 1107 [36 ] S. V. Rao, K. W. Anderson, L. G. Bachas, Mikrochim. Acta 1998 128 127 143 [37 ] K. Hernandez, R. Fernandez Lafuente, Enzyme Microb. Technol. 2011 48 107 122 [38 ] T. Cha, A. Guo, Y. Jun, D. Pei, X. Y. Zhu, Proteomics 2004 4 1965 1976

PAGE 152

152 [39 ] W. H. Scouten, J. H. T. Luong, R. Stephen Brown, Trends Biotechnol. 1995 13 178 185 [40 ] R. J. Chen, Y. Zhang, D. Wang, H. Dai, J. Am. Chem. Soc. 2001 123 3838 3839 [41 ] K. Jiang, L. S. Schadler, R. W. Siegel, X. Zhang, H. Zhang, M. Terrones, J. Mater. Chem. 2004 14 37 39 [42 ] E. Topoglid is, A. E. G. Cass, G. Gilardi, S. Sadeghi, N. Beaumont, J. R. Durrant, Anal. Chem. 1998 70 5111 5113 [43 ] S. Liu, A. Chen, Langmuir 2005 21 8409 8413 [44 ] P. Si, S. Ding, J. Yuan, X. W. Lou, D. H. Kim, ACS Nano 2011 5 7617 7626 [45 ] C. V. Kumar, A. Chaudhari, J. Am. Chem. Soc 2000 122 830 837 [46 ] L. N. Geng, N. Li, N. Dai, X. F. Wen, F. L. Zhao, K. A. Li, Colloids Surf B Biointerfaces 2003 29 81 88 [47 ] L. Geng, X. Wang, N. Li, M. Xiang, K. a. Li, Colloids Surf B Biointerfaces 2004 34 231 238 [48 ] J. Monot, M. Petit, S. M. Lane, I. Guisle, J. Lger, C. Tellier, D. R. Talham, B. Bujoli, J. Am. Chem. Soc 2008 130 6243 6251 [49 ] P. Westphal, A. Bornmann, Sens. Actuators B Chem. 2002 84 278 282 [50 ] J. Homola, S. S. Yee, G. Gauglitz Sens. Actuators B Chem. 1999 54 3 15 [51 ] T. Iwata, S. Maeda, Appl. Opt. 2007 46 1575 1582 [52 ] A. V. Nabok, A. Tsargorodskaya, A. K. Hassan, N. F. Starodub, Appl. Surf. Sci. 2005 246 381 386 [53 ] J. A. De Feijter, J. Benjamins, F. A. Veer, Biop olymers 1978 17 1759 1772 [54 ] D. E. Aspnes, J. B. Theeten, F. Hottier, Phys. Rev. 1979 20 3292 3302 [55 ] J. Benesch, A. Askendal, P. Tengvall, Colloids Surf B Biointerfaces 2000 18 71 81 [56 ] R. M. Plocinik, G. J. Simpson, Anal. Chim. Acta 2003 496 133 142 [57 ] P. Pattnaik, Appl Biochem Biotechnol 2005 126 79 92 [58 ] B. Liedberg, C. Nylander, I. Lunstrm, Sensor Actuator. 1983 4 299 304

PAGE 153

153 [59 ] U. Jnsson, L. Fgerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, S. Lfs, B. Persso n, H. Roos, I. Rnnberg, Bio Techniques 1991 11 620 627 [60 ] J. Homola, Anal. Bioanal. Chem. 2003 377 528 539 [61 ] F. Abels, Surf. Sci. 1976 56 237 251 [62 ] Z. Ustundag, M. O. Caglayan, R. Guzel, E. Piskin, A. O. Solak, Analyst 2011 136 1464 1471 [ 63 ] D. J. Oshannessy, M. Brighamburke, K. K. Soneson, P. Hensley, I. Brooks, Anal. Biochem. 1993 212 457 468 [64 ] R. J. Fisher, M. Fivash, J. Casas Finet, J. W. Erickson, A. Kondoh, S. V. Bladen, C. Fisher, D. K. Watson, T. Papas, Protein Sci. 1 994 3 257 266 [65 ] J. Madsen, K. S. Pedersen, M. L. Michelsen, Ind. Eng. Chem. Res. 2000 39 1111 1114 [66 ] I. C. R. Holford, R. W. M. Wedderburn, G. E. G. Mattingly, Eur J Soil Sci 1974 25 242 255 [67 ] P. Schuck, Biophys. J. 1996 70 1230 12 49 [68 ] M. L. Yarmush, D. B. Patankar, D. M. Yarmush, Mol Immun 1996 33 1203 1214 [69 ] R. W. Glaser, Anal. Biochem. 1993 213 152 161 [70 ] M. Cinier, M. Petit, M. N. Williams, R. M. Fabre, F. D R Pecorari, D. R. Talham, B. Bujoli, C. Tellier, Bi oconjugate Chem. 2009 20 2270 2277 [71 ] H. Byrd, S. Whipps, J. K. Pike, J. Ma, S. E. Nagler, D. R. Talham, J. Am. Chem. Soc. 1994 116 295 301 [72 ] R. M. Fabre, D. R. Talham, Langmuir 2009 25 12644 12652 [73 ] D. S. Moore, Biochem. Educ. 1985 13 10 11 [ 74 ] B. Bujoli, S. M. Lane, G. Nonglaton, M. Pipelier, J. Leger, D. R. Talham, C. Tellier, Chem. Eur. J. 2005 11 1980 1988 [75 ] D. Byun, Y. Jin, B. Kim, J. Kee Lee, D. Park, J. Hazard. Mater. 2000 73 199 206 [76 ] H. Yoshitake, T. Sugihara, T. Tatsumi, Chem. Mater. 2002 14 1023 1029 [77 ] F. Torta, M. Fusi, C. S. Casari, C. E. Bottani, A. Bachi, J. Proteome Res. 2009 8 1932 1942

PAGE 154

154 [78 ] A. P. Caricato, S. Capone, G. Ciccarella, M. Martino, R. Rella, F. Romano, J. Spadavecchia, A. Taurino, T. Tunno, D. Valerini, Appl. Surf. Sci. 2007 253 7937 7941 [79 ] S. Ohno, D. Sato, M. Kon, Y. Sato, M. Yoshikawa, P. Frach, Y. Shigesato, Jpn. J. Appl. Phys. 2004 43 8234 [80 ] A. Yasumori, H. Shinoda, Y. Kameshima, S. Hayashi, K. Okada, J. Mater. Chem. 2001 11 1253 1257 [81 ] M. Grtzel, J. Sol Gel Sci. Technol. 2001 22 7 13 [82 ] W. Sugimoto, O. Terabayashi, Y. Murakami, Y. Takasu, J. Mater. Chem. 2002 12 3814 3818 [83 ] A. Boccaccini, P. Karapappas, J. Marijuan, C. Kaya, J. Mater. Sci. 2004 39 851 859 [84 ] J. S. King, E. Graugnard, C. J. Summers, Adv. Mater. 2005 17 1010 1013 [85 ] V. Pore, A. Rahtu, M. Leskel, M. Ritala, T. Sajavaara, J. Keinonen, Chem. Vap. Deposition 2004 10 143 148 [86 ] V. B. Aleskovskii, Russ. J. Gen. Chem. 2002 7 2 569 574 [87 ] R. L. Puurunen, J. Appl. Phys. 2005 97 121301 121352 [88 ] M. Kemell, V. Pore, M. Ritala, M. Leskela, M. Linden, J. Am. Chem. Soc. 2005 127 14178 14179 [89 ] M. A. Cameron, I. P. Gartland, J. A. Smith, S. F. Diaz, S. M. George, Langmu ir 2000 16 7435 7444 [90 ] G. K. L. Goh, C. P. K. Liew, J. Kim, T. J. White, J. Cryst. Growth 2006 291 94 99 [91 ] Y. Masuda, T. Sugiyama, H. Lin, W. S. Seo, K. Koumoto, Thin Solid Films 2001 382 153 157 [92 ] Y. Masuda, Y. Jinbo, T. Yonezawa, K. Ko umoto, Chem. Mater. 2002 14 1236 1241 [93 ] Y. Gao, Y. Masuda, Z. Peng, T. Yonezawa, K. Koumoto, J. Mater. Chem. 2003 13 608 613 [94 ] Y. Masuda, S. Ieda, K. Koumoto, Langmuir 2003 19 4415 4419 [95 ] S. Yamabi, H. Imai, Chem. Mater. 2002 14 609 6 14

PAGE 155

155 [96 ] G. K. L. Goh, S. K. Donthu, P. K. Pallathadka, Chem. Mater. 2004 16 2857 2861 [97 ] G. K. L. Goh, X. Q. Han, C. P. K. Liew, C. S. S. Tay, J. Electrochem. Soc. 2005 152 532 536 [98 ] I. G. Tredici, F. Maglia, A. Resmini, U. Anselmi Tamburini, T hin Solid Films 2013 542 52 59 [99 ] E. K. Seo, J. W. Lee, H. M. Sung Suh, M. M. Sung, Chem. Mater. 2004 16 1878 1883 [100 ] P. J. Cumpson, Appl. Surf. Sci. 1999 144 145 16 20 [101 ] L. Berrade, A. Garcia, J. Camarero, Pharm Res 2011 28 1480 1499 [102 ] W. Hu, C. Peng, M. Lv, X. Li, Y. Zhang, N. Chen, C. Fan, Q. Huang, ACS Nano 2011 5 3693 3700 [103 ] Z. L. Pianowski, N. Winssinger, Chem. Commun. 2007 3820 3822 [104 ] Y. Choi, C. S. Lee, Methods Mol. Biol. 2011 671 207 217 [105 ] W. Kusnezow, J. D. Hoheisel, J. Mol. Recogn. 2003 16 165 176 [106 ] P. Peluso, D. S. Wilson, D. Do, H. Tran, M. Venkatasubbaiah, D. Quincy, B. Heidecker, K. Poindexter, N. Tolani, M. Phelan, K. Witte, L. S. Jung, P. Wagner, S. Nock, Anal. Biochem. 2003 312 113 124 [107 ] P. Arenkov, A. Kukhtin, A. Gemmell, S. Voloshchuk, V. Chupeeva, A. Mirzabekov, Anal. Biochem 2000 278 123 131 [108 ] C. Doyle, B. James, S. Daniels, D. E. Williams, Va cuum 2012 86 547 555. [109 ] M. Cinier, M. Petit, F. Pecorari, D. Talham, B. Bujoli, C. Tellier, J Biol Inorg Chem 2012 17 399 407 [110 ] S. Karrasch, M. Dolder, F. Schabert, J. Ramsden, A. Engel, Biophys. J. 1993 65 2437 2446 [111 ] A. M. Massari R W. Gurney, M. D. Wightman, C. H. K. Huang, S. T. Nguyen, J. T. Hupp, Polyhedron 2003 22 3065 3072

PAGE 156

156 BIOGRAPHICAL SKETCH Hao Liu was born in Yuncheng, China in 1986. She graduated from Yuncheng high school in 2004, after which she entered Xiamen Un iversity in Fujian Province, southern China. the same year. Hao was qualified in the analy tical chemistry division as a PhD Hao defended her work and graduated in the fall of 2013 from the University of Florida with a PhD in chemistry, specialized in bioanaly tical chemistry.