Citation
Mass Spectrometry for Structural Proteomic Analysis of Recombinant Human Sialyltransferase and Identification of Nanoparticle Harvested Oligonucleotides

Material Information

Title:
Mass Spectrometry for Structural Proteomic Analysis of Recombinant Human Sialyltransferase and Identification of Nanoparticle Harvested Oligonucleotides
Creator:
TIPTON, JEREMIAH D. ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Amino acids ( jstor )
Biochemistry ( jstor )
Digestion ( jstor )
Ions ( jstor )
Mass spectroscopy ( jstor )
Molecules ( jstor )
Oligonucleotides ( jstor )
Proteomics ( jstor )
Purification ( jstor )
Solvents ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jeremiah D. Tipton. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
4/17/2006
Resource Identifier:
76786549 ( OCLC )

Downloads

This item has the following downloads:

tipton_j ( .pdf )

tipton_j_Page_024.txt

tipton_j_Page_114.txt

tipton_j_Page_119.txt

tipton_j_Page_240.txt

tipton_j_Page_142.txt

tipton_j_Page_084.txt

tipton_j_Page_174.txt

tipton_j_Page_181.txt

tipton_j_Page_023.txt

tipton_j_Page_011.txt

tipton_j_Page_070.txt

tipton_j_Page_002.txt

tipton_j_Page_064.txt

tipton_j_Page_215.txt

tipton_j_Page_166.txt

tipton_j_Page_205.txt

tipton_j_Page_088.txt

tipton_j_Page_165.txt

tipton_j_Page_216.txt

tipton_j_Page_117.txt

tipton_j_Page_071.txt

tipton_j_Page_218.txt

tipton_j_Page_155.txt

tipton_j_Page_091.txt

tipton_j_Page_034.txt

tipton_j_Page_212.txt

tipton_j_Page_019.txt

tipton_j_Page_201.txt

tipton_j_Page_210.txt

tipton_j_Page_047.txt

tipton_j_Page_020.txt

tipton_j_Page_043.txt

tipton_j_Page_078.txt

tipton_j_Page_049.txt

tipton_j_Page_235.txt

tipton_j_Page_122.txt

tipton_j_Page_190.txt

tipton_j_Page_089.txt

tipton_j_Page_173.txt

tipton_j_Page_112.txt

tipton_j_Page_012.txt

tipton_j_Page_080.txt

tipton_j_Page_057.txt

tipton_j_Page_138.txt

tipton_j_Page_156.txt

tipton_j_Page_167.txt

tipton_j_Page_194.txt

tipton_j_Page_056.txt

tipton_j_Page_017.txt

tipton_j_Page_228.txt

tipton_j_Page_098.txt

tipton_j_Page_202.txt

tipton_j_Page_115.txt

tipton_j_Page_045.txt

tipton_j_Page_127.txt

tipton_j_Page_030.txt

tipton_j_Page_051.txt

tipton_j_Page_050.txt

tipton_j_pdf.txt

tipton_j_Page_131.txt

tipton_j_Page_189.txt

tipton_j_Page_232.txt

tipton_j_Page_104.txt

tipton_j_Page_191.txt

tipton_j_Page_211.txt

tipton_j_Page_221.txt

tipton_j_Page_004.txt

tipton_j_Page_105.txt

tipton_j_Page_125.txt

tipton_j_Page_227.txt

tipton_j_Page_090.txt

tipton_j_Page_060.txt

tipton_j_Page_121.txt

tipton_j_Page_029.txt

tipton_j_Page_198.txt

tipton_j_Page_028.txt

tipton_j_Page_170.txt

tipton_j_Page_186.txt

tipton_j_Page_109.txt

tipton_j_Page_200.txt

tipton_j_Page_007.txt

tipton_j_Page_081.txt

tipton_j_Page_038.txt

tipton_j_Page_140.txt

tipton_j_Page_036.txt

tipton_j_Page_069.txt

tipton_j_Page_217.txt

tipton_j_Page_238.txt

tipton_j_Page_196.txt

tipton_j_Page_107.txt

tipton_j_Page_207.txt

tipton_j_Page_143.txt

tipton_j_Page_009.txt

tipton_j_Page_149.txt

tipton_j_Page_124.txt

tipton_j_Page_225.txt

tipton_j_Page_074.txt

tipton_j_Page_001.txt

tipton_j_Page_073.txt

tipton_j_Page_005.txt

tipton_j_Page_134.txt

tipton_j_Page_016.txt

tipton_j_Page_055.txt

tipton_j_Page_097.txt

tipton_j_Page_154.txt

tipton_j_Page_106.txt

tipton_j_Page_139.txt

tipton_j_Page_026.txt

tipton_j_Page_076.txt

tipton_j_Page_025.txt

tipton_j_Page_193.txt

tipton_j_Page_231.txt

tipton_j_Page_077.txt

tipton_j_Page_130.txt

tipton_j_Page_079.txt

tipton_j_Page_027.txt

tipton_j_Page_206.txt

tipton_j_Page_152.txt

tipton_j_Page_159.txt

tipton_j_Page_021.txt

tipton_j_Page_102.txt

tipton_j_Page_072.txt

tipton_j_Page_111.txt

tipton_j_Page_133.txt

tipton_j_Page_150.txt

tipton_j_Page_059.txt

tipton_j_Page_113.txt

tipton_j_Page_054.txt

tipton_j_Page_031.txt

tipton_j_Page_062.txt

tipton_j_Page_053.txt

tipton_j_Page_158.txt

tipton_j_Page_209.txt

tipton_j_Page_116.txt

tipton_j_Page_197.txt

tipton_j_Page_239.txt

tipton_j_Page_008.txt

tipton_j_Page_137.txt

tipton_j_Page_108.txt

tipton_j_Page_094.txt

tipton_j_Page_151.txt

tipton_j_Page_100.txt

tipton_j_Page_093.txt

tipton_j_Page_230.txt

tipton_j_Page_179.txt

tipton_j_Page_153.txt

tipton_j_Page_003.txt

tipton_j_Page_147.txt

tipton_j_Page_126.txt

tipton_j_Page_192.txt

tipton_j_Page_006.txt

tipton_j_Page_223.txt

tipton_j_Page_233.txt

tipton_j_Page_168.txt

tipton_j_Page_086.txt

tipton_j_Page_110.txt

tipton_j_Page_037.txt

tipton_j_Page_195.txt

tipton_j_Page_044.txt

tipton_j_Page_219.txt

tipton_j_Page_162.txt

tipton_j_Page_136.txt

tipton_j_Page_014.txt

tipton_j_Page_099.txt

tipton_j_Page_120.txt

tipton_j_Page_010.txt

tipton_j_Page_039.txt

tipton_j_Page_213.txt

tipton_j_Page_203.txt

tipton_j_Page_052.txt

tipton_j_Page_033.txt

tipton_j_Page_148.txt

tipton_j_Page_061.txt

tipton_j_Page_015.txt

tipton_j_Page_163.txt

tipton_j_Page_183.txt

tipton_j_Page_040.txt

tipton_j_Page_066.txt

tipton_j_Page_042.txt

tipton_j_Page_013.txt

tipton_j_Page_096.txt

tipton_j_Page_234.txt

tipton_j_Page_175.txt

tipton_j_Page_176.txt

tipton_j_Page_083.txt

tipton_j_Page_135.txt

tipton_j_Page_141.txt

tipton_j_Page_085.txt

tipton_j_Page_058.txt

tipton_j_Page_123.txt

tipton_j_Page_063.txt

tipton_j_Page_129.txt

tipton_j_Page_095.txt

tipton_j_Page_184.txt

tipton_j_Page_199.txt

tipton_j_Page_161.txt

tipton_j_Page_032.txt

tipton_j_Page_145.txt

tipton_j_Page_188.txt

tipton_j_Page_128.txt

tipton_j_Page_226.txt

tipton_j_Page_101.txt

tipton_j_Page_041.txt

tipton_j_Page_171.txt

tipton_j_Page_075.txt

tipton_j_Page_022.txt

tipton_j_Page_172.txt

tipton_j_Page_229.txt

tipton_j_Page_220.txt

tipton_j_Page_146.txt

tipton_j_Page_046.txt

tipton_j_Page_035.txt

tipton_j_Page_132.txt

tipton_j_Page_068.txt

tipton_j_Page_048.txt

tipton_j_Page_237.txt

tipton_j_Page_224.txt

tipton_j_Page_180.txt

tipton_j_Page_082.txt

tipton_j_Page_222.txt

tipton_j_Page_144.txt

tipton_j_Page_204.txt

tipton_j_Page_067.txt

tipton_j_Page_160.txt

tipton_j_Page_164.txt

tipton_j_Page_018.txt

tipton_j_Page_214.txt

tipton_j_Page_092.txt

tipton_j_Page_169.txt

tipton_j_Page_185.txt

tipton_j_Page_187.txt

tipton_j_Page_087.txt

tipton_j_Page_236.txt

tipton_j_Page_157.txt

tipton_j_Page_182.txt

tipton_j_Page_103.txt

tipton_j_Page_178.txt

tipton_j_Page_118.txt

tipton_j_Page_208.txt

tipton_j_Page_177.txt

tipton_j_Page_065.txt


Full Text












MASS SPECTROMETRY FOR STRUCTURAL PROTEOMIC ANALYSIS OF
RECOMBINANT HUMAN SIALYLTRANSFERASE AND IDENTIFICATION OF
NANOPARTICLE HARVESTED OLIGONUCLEOTIDES















By

JEREMIAH D. TIPTON


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

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Jeremiah D. Tipton

































This work is dedicated to my family and close friends that have supported
me during this period of growth and to those which have left us too soon.
Bryan LaRose
John P. Collins
Carl Hilley
Crystal Tipton















ACKNOWLEDGMENTS

First and foremost, my sincerest gratitude goes to both of my advisors, Dr. David

Powell and Dr. Nicole Horenstein, for their patience, guidance, and support during the

course of my graduate studies and the course of these projects. I would also like to

recognize and thank the members of my committee, Dr. Jon Stewart, Dr. Weihong Tan,

and Dr. Nancy Denslow, for their suggestions and support.

Between two laboratories, I would like to thank all my new sisters: Jen, Erin,

Fedra, and Mirela from the Horenstein Group and Quan Li, Violeta, Lani, Cris, Daniella

and Joanna from the Powell group for all their support and colorful conversations over

the years. Additionally, I would like to thank Alonso and Josh for their insight on

different interesting areas of science and Scott and Dr. Johnson for their help on

obtaining LCQ-MS data. I would also like to thank Romaine Hughes for her

administrative help and the many breaks spent on the bridge. A special thank you goes

out to Mike, Ben, Travis, Josh, Larry, Brent, and Merve, for all the good times we had

through these interesting years.

In my immediate family, I am deeply indebted to my parents Robin and Dennis for

all their support whenever I was in need and my younger siblings Rachel, Tamara,

Brandon, Derek, and Devin for all their inquiries as to why I am still in college. In my

extended family, I would like to thank Jan and the late John Collins for allowing me to

stay at their house when I needed a quiet place to stay. Lastly, I would like to thank Matt

for all his support over the many years I have known him.

















TABLE OF CONTENTS

page

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

LIST OF TABLES .......... .... ........ .... ............. ............ ix

LIST OF FIGURES ......... ........................................... ............ xi

ABSTRACT .............. .......................................... xvi

CHAPTER

1 INTRODUCTION TO BIOLOGICAL MASS SPECTROMETRY............................1

System s Biology and M ass Spectrom etry ........................................ .....................
M ass Spectrom etry Based Proteom ics...................................... ........................ 3
F u n action al P roteom ics ................................................................ .....................4
Structural Proteomics .................... .... ........ .... ...... ..... ......... 5
Mass Spectrometry and Large Bio-molecules Meet ..........................................5
M ass Spectrometry Analysis of Oligonucleotides......................................................6
Com plete M ass Spectrom etry Analysis.................................... ........................ 8
Separation Science................ ..... ........ ............................. ...... .9
Off-line separation technology ................. ............................................ 10
On-line separation technology ........................................ ............... 12
Ionization Techniques: ESI and M ALDI .................................... ............... 14
M a ss A n aly z e rs ............................................................................................. 19
Quadruple ion trap M S .................................................................... 20
Tim e-of-flight M S ................................................................................. 22
Fourier transform ion cyclotron resonance MS ................. ................24
B ioinform atics ................. ............................................................ 28
Intact peptide MS data analysis with in silico digestion algorithms ..........29
Tandem MS data analysis with in silico digestion algorithm ....................30
Tandem MS data analysis with de novo algorithims...................... ..31
Software packages for analysis of MS and MS/MS data..........................32
Bioinformatics data visualization and interpretation ................................34
S u m m a ry ......................................................................................................3 5





v









2 EXPRESSION AND PURIFICATION OF RECOMBINANT HUMAN ALPHA
2- 3 SIA L TR A N SFE R A SE ........................................................... .....................36

Introduction to Oligosaccharides and Glycosyltransferases ...................................... 36
Complex Carbohydrates in Biological Systems...............................................37
G lycosyltransferase Fam ily ........................................ ........................... 38
Sialyltransferase Sub-Fam ily......................................................... ................ ..... 40
Biological and M edicinal Im portance ....................... .... ....... ...................44
Sialyltransferase Structure and Mechanism Relevant for Structural
P roteom ics .......... ..................................................... ... ...............45
Recombinant Human ao (2- 3) Sialyltransferase IV..............................................48
R results and D discussion ...................... .. .. ............... .................... ............... 50
Baculovirus preparation ......................................... ...............50
Expression and purification of the three sialyltransferase constructs ..........51
Concentration/dilution studies on N-Tag-ST............................................60
K inetic param eters......... ............................................ ........ ... ......... 63
Conclusion............................................ ...............64
M ethods and M materials ..................................................... ............... ... 64
Baculovirus vector preparation for Ins-ST .............................................65
Baculovirus vector preparation for N-Tag-ST ..........................................66
Baculovirus vector preparation for C-Tag-ST ..........................................67
Expression and purification of Ins-ST ...................... ........ .....................68
Expression and purification of N-Tag-ST and C-Tag-ST..........................69
Activity assays for recombinant sialyltransferase.......................................71
N-Tag-ST stability experiments .........................................................71
In-gel digestion of glycoprotein 64 ............................................................ 73
M S an aly sis .................................................................... 74

3 HYDROGEN / DEUTERIUM EXCHANGE HPLC-MS FOR
SIALYLTRANSFERASE SECONDARY STRUCTURE .......................................75

Introduction to Hydrogen/Deuterium Exchange HPLC-MS ....................................75
Monitoring Hydrogen/Deuterium Exchange of Amide Protons ........................76
Hydrogen/Deuterium Exchange Kinetics..................................... ............... 77
Operational Aspects of H/Dx HPLC-ESI-MS Experiments ...............................79
Hydrogen/Deuterium Exchange HPLC-ESI-MS Analysis of Selected Proteins 80
Experimental Design for Analysis of Sialyltransferase..................................82
R results and D iscu ssion ..................... ...... ... .... ..................... .... ... ....... ....84
Optimization of Analytical System for Analysis of Intact Proteins....................84
Dynamic and gas assisted trapping and analysis of intact proteins with
E SI-FTICR M S ...................................................................... .............. 84
Broadband versus heterodyne excitation and detection ............................90
Optimization of HPLC-ESI-FTICR-MS ................................................98
Separation and MS analysis of intact proteins ........................................... 98
H/Dx Direct Infusion and H/Dx HPLC-ESI-FTICR-MS of Intact HEWL....... 101
Separation and MS Analysis of Pepsin Digested Proteins ............................104
M S Analysis of Recombinant Sialyltransferase............................................... 110









EndoHf and PNGase F digestion of intact Ins-ST ............................... 114
MS analysis of pepsin digested sialyltransferase .............. ..................115
Conclusion ..................................... ................................. ......... 116
M material and M methods .......................................................................... ............... 117
FT-ICR/MS Instrumentation ... .. .................... ..................117
HPLC Instrumentation................ ............ ...... ................. 117
Intact Protein Analysis with Broadband Mode ...........................................118
Direct infusion analysis of intact proteins....... .................................. 118
HPLC-ESI-FT-ICR MS analysis of intact proteins...............................119
Pepsin Digestion and HPLC-ESI-FTICR-MS Analysis................................120
H /D x E xperim ents ......... .... .......... ............................... .. ................ .. 12 1

4 MS BASED STRUCTURAL PROTEOMICS WITH SMALL MOLECULE
LABELING AGENTS FOR IDENTIFYING SIALYLTRANSFERASE ACTIVE
SIT E A M IN O A C ID S ................................................................... ..................... 122

Introduction to Bioconjugation Techniques .............................. ................... .... 122
Practical Consideration and Types of Small Molecule Labeling Agents..........123
Practical Considerations for MS Based Structural Proteomics with SMLAs ...124
Literature Review on Ideal Cases................. ............................ ............... 125
Limitations for MS Based Structural Proteomics with SMLAs........................130
Practical Consideration for Labeling of Sialyltransferase with SMLAs ..........131
Bioconjugation of Sialyltransferase with Small Affinity Labels............................132
D ifferential Labeling Schem e....................................... ......................... 133
W orkflow ................................................................... .......... 134
R results and D iscu ssion ........................................................... ..................... 136
Preliminary screening of NAI, EDAC/ETAM, and IOA........................... 136
N AI differential labeling experim ents.................................... ............... 137
Reactivation of NAI derivatized N-Tag-ST with ethanolamine ...............140
SDS-PAGE analysis and in-gel digestion with trypsin or other proteases 141
M S analysis of IGD sam ple ...................................................................... 144
M annual interpretation of M S data ................................... ............... ..152
C o n clu sio n .................................................. ................ 15 2
M ethods and M materials ...... ....... ................... ....... ...... .............. 153
Reaction conditions for preliminary screening of SMLAs ......................153
N A I labeling experim ents ................ .... ....... .... ........................ .... 154
SDS-PAGE analysis and in-gel digestion of recombinant hST3Gal IV....157
M ass spectrom etry analysis................................... ......................... 158

5 MS BASED STRUCTURAL PROTEOMICS WITH SITE-DIRECTED
PHOTOAFFINITY LABELING FOR THE INVESTIGATION OF
SIALYLTRANSFERASE ACTIVE SITE AMINO ACIDS ..................................161

Introduction to Site Directed Photo-affinity Labeling.............................................161
Photoreactive G groups ................................. ... ...................................... 162
Practical Considerations for Site-Directed Photoaffinity Labeling................. 164
Examples of Photoaffinity Labeling.............. ............................ .................164









Practical Consideration for Site Directed Photoaffinity Labeling of
Sialyltransferase ..................................... ........ .... .. .. .... .. ........ .. 64
R results and D discussion ................... ................... ........ .... .. .......... .... 167
CMP-o-azido Mandelate Synthesis and Characterization...............................167
Site-Directed Photoaffinity Labeling of Sialyltransferase with CMP-o-Azido
M andelate................. .. ................... ........................ ............ 169
MS Analysis of In-Gel Digested Irradiated N-Tag-ST/Inhibitor Samples........175
Conclusion ..................................... ................................. ......... 179
M eth od s an d M materials ........................................ .. ............................................ 180
Synthesis and Characterization of CMP-o-azido Mandelate............................. 180
Photoaffinity Labeling Experiments........................................................180
Incubation of N-Tag-ST with and without inhibitor...............................1.80
Photoactivation of inhibitor in the presence of N-Tag-ST......................181
Photoactivation of inhibitor in the presence of commercially available
sialyltransferase .............. ......................................... ...... .. 181
Activity assays for recombinant sialyltransferase.................................... 183
In-G el D igestion w ith Trypsin...................................... ......................... 183
M S A n aly sis ................................................................... 184

6 MS ANALYSIS OF DNA HARVESTED WITH NANOHARVESTING
A G EN T S ................................................................ .... ..... ......... 185

Introduction to Oligonucleotide Analysis with Mass Spectrometry.........................185
N anoharvesting Agents ...................................... .....................................188
On-line HPLC MS of Oligonucleotides ..............................190
MS Instrumentation Considerations ............. ....................... ..............191
E xperim ent D design ......... ................. ...................................... ......................193
Results and Discussion ............. ........ ............... ..... .. .. ............ 193
Solvent System Studies .................... ... .............. ....................... ....194
IP-RP-HPLC ESI-FTICR-MS Analysis of Synthetic Oligonucleotides ..........198
Conclusion ................ ........ ................... ........ .... ........ ..... 199
M methods and M materials ..................................................................... ..................200
FT-ICR/M S Instrum entation ........................................ ......... ............... 200
H PL C Instrum entation ............................................................ .....................20 1

7 CONCLUSIONS AND FUTURE WORK.......................................................202

T he Sialyltransferase P project ....................................................................................203
H/Dx HPLC ESI FT-ICR M S...................................................................... 205
Sm all M olecule Labeling A gents ........................................... ... ............ 207
Site D directed Photoaffinity Labeling............................................. ...............208
MS Analysis of Oligonucleotides Selected with Nanoharvesting Agents..............209
Final Conclusions ................................. .. .... ....... .............. .. 210

LIST OF REFEREN CE S ...... .................................. ......................... ............... 212

B IO G R A PH IC A L SK E T C H ........................................ ............................................223
















LIST OF TABLES


Table page

1-1. Protein Function in Biological O rganism s ........................................ .....................3

1-2. M S E xperim ents on P roteins .......................................................................... ...... 3

1-3. M ass Shift A associated with Com m on PTM s.................................... .....................4

1-4. Solvents Com patible w ith E SI................................ ...................... ............... 17

1-5. Comm on Genomic Databases .............................................................................29

1-6. Common MS and MS/MS Peptide Data Analysis Software ..................................33

2-1. Viral Titers for Different Constructs .............................................. ................ 51

2-2. Sum m ary of an Ins-ST Purification...................................... ......................... 53

2-3. Sum m ary of a N -Tag-ST Purification ............................................ ............... 56

2-4. Sum m ary of C-Tag-ST Purification ........................................ ....... ............... 56

2-5. Summary of Selected Sialyltransferase Cloning and Purification Papers ...............58

2-6. Summary of Kinetic Parameters for Different Sialyltransferase Constructs............63

3-1. Summary of Different Trapping Methods..... ................................................. ...........86

3-2. Relationship Between the Six Important Parameters and Signal Intensity/
Resolution of the Isotopically Resolved +10 Charge State of HEWL ....................88

3-3. Mass Accuracies for Different Protein Standards ............................................. 90

3-4. Mass Accuracies for Different Proteins with HPLC-ESI FTICR-MS Analysis .......99

3-5. Peptides Observed from Bovine Carbonic Anhydrase Digested with Pepsin
Presented in Figure 3-26 ............. ... ......... ..................... 11

3-6. Sialyltransferase Construct Concentration, Buffer, and Molecular Weights ..........112

4-1. Literature Review of Different Bioconjugation Experiments ..............................127









4-2. SMLAs and Activity after 20 Minute Incubation..................... ................. 137

5-1. Literature Review on Case Specific Photoaffinity Labeling ...................................165

5-2. Percent Activity Remaining after Incubation of N-Tag-ST With and Without
Inhibitor P resent. .................................................................... 17 1

5-3. Irradiation of N-Tag-ST at 254 nm (0.65 Amps) with Different Concentrations of
Inhibitor ...................................................................... ........... 172

5-4. Irradiation of Recombinant Rat c2,3-(N)-sialyltransferase with 130 ptM Inhibitor
P re sen t ................................................. ........................................ 17 4

5-5. Irradiation of Recombinant Rat c2,6-(N)-sialyltransferase without Inhibitor
Present and with 130 [tM Inhibitor Present..................................................174

5-6. Mass-to-Charge of +1 Charge State of Theoretical Tryptic Peptides and Four
Possible M ass Shifts .................. ..... .... ........ .......... ............... 177

5-7. Mass-to-Charge of 2 Charge State of Theoretical Tryptic Peptides and Four
Possible M ass Shifts .................................... ...... .. ...... .. ................. 178

6-1. Synthetic Oligonucleotides Prepared for NAH Selection and On-line HPLC-MS
A n aly sis .............................. .......................................................... ............... 19 0
















LIST OF FIGURES


Figure page

1-1. Diagram Representing Components of Total Systems Biology ..............................2

1-2. Common Fragmentation Patterns and Nomenclature...............................................6

1-3. Diagram Describing the Workflow for DNA Sequencing with CE............................7

1-4. Diagram Describing the Workflow for Protein Sequencing with MS.....................7

1-5. W orkflow for Com plete M S A nalysis............................................... ..................... 8

1-6. "Bottom-Up" and "Top-Down" Workflows.........................................................10

1-7. Microbore MudPIT................... ................... ............ 14

1-8. Electrospray Ionization with Dole's Large Molecule Ionization Model .................15

1-9. Diagram Representing Matrix-Assisted Laser Desorption Ionization ...................19

1-10. Actively Shielded 4.7 T Magnet at the University of Florida...............................25

1-11. Cyclotron Motion of Ions in Magnetic Field..................................................25

1-12. Relationship between the Frequency and Mass-to-Charge of an Ion...................26

2-1. Common Sugars Found in Oligosaccharides...................... ..................... 37

2-2. Reaction Catalyzed by Sialyltransferases .......................................................41

2-3. Topology of Sialyltransferase.................. .......... ............................ 42

2-4. Sialylmotifs and Amino Acid Length of Sialyltransferases ....................................46

2-5. Sialyltransferase Constructs Prepared ............................................ ............... 49

2-6. Confirmation of Correct DNA Insertion Product ................................... ........ 51

2-7. Protein Concentration Profile of an Ins-ST Purification.......................................53

2-8. Protein Concentration and Activity Profile of an Ins-ST Purification ....................54









2-9. SDS-PAGE Analysis of Purified Ins-ST and Deglycosylation Reaction..................55

2-10. Protein Concentration and Activity Profile of a C-tag-ST Purification ..................57

2-11. SDS-PAGE Analysis of TCA Precipitated N-Tag-ST and C-Tag-ST ...................57

2-12. Effect on Activity of N-Tag-ST with Dilution of Different Detergent
Concentrations ............. .. .. .. ................ ............. 61

2-13. Effect on N-Tag-ST Activity with Different Dilution Factors and Concentration
with Millipore Ultrafree-MC Centrifugal Filter Units ..........................................62

3-1. Equations Describing H/Dx Kinetics ............................................. ............... 78

3-2. Equation for Calculating the Number of Exchanged Amide Protons ....................79

3-3. Analytical Work Flow for H/Dx Experiment ................ ..................................80

3-4. Six Experiments for Complete Characterization of Recombinant ST3Gal IV
constructs w ith H /D x H PLC-M S ........................................ ........................ 83

3-5. Dynam ic Trapping Experim ent. ........................................ .......................... 86

3-6. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/tiL ofRibonuclease A..........91

3-7. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/tiL of HEWL.........................92

3-8. Expanded +10 Charge State of the Spectra Presented in Figure 3-7.......................93

3-9. Direct Infusion of ESI FTICR-MS Analysis of 15 pmol/[tL of Human Carbonic
A nhy drase I. ....................................................... ................. 94

3-10. Direct Infusion ESI FTICR-MS Analysis of 20 pmol/[tL of Bovine Serum
A lb u m in ................................. ......................................................... ............... 9 5

3-11. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/tiL of HEWL....................96

3-12. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/tiL of HEWL ....................97

3-13. TIC of HEWL (75 pmol) Analyzed with HPLC-ESI-FT-ICR-MS ......................99

3-14. Mass Spectrum at 19.15 mins from the TIC Presented in Figure 3-13. .............100

3-15. TIC of HEWL (50 pmol) and a-Chymotrypsinogen A (50 pmol) Analyzed with
HPLC-ESI-FT-ICR-M S .................................................................... 101

3-16. HEW L Spectrum Observed at 12.30 M minutes ...................................... .......... 102

3-17. a-Chymotrypsinogen Spectrum Observed at 12.93 Minutes ...............................103









3-18. TIC of Unknown Amount of Uridine Kinase Analyzed with HPLC-MS ESI-
F T IC R -M S ........................................... .......................... 104

3-19. Uridine Kinase Spectrum Observed at 12.00 Minutes ........................................105

3-20. TIC of HEWL (25 pmol) Analyzed with HPLC-ESI-FT ICR-MS in Heterodyne
M ode and D ynam ic Trapping ........................................... ......................... 106

3-21. HEWL Spectrum Observed at 14.00 Minutes (Figure 3-20)..............................106

3-22. H/Dx Analysis of HEWEL with Direct Infusion ESI-FTICR-MS..................... 107

3-23. Expanded +10 Charge State Observed in Figure 3-22 ............... .....................108

3-24. TIC of H/Dx-HEWL Sample (10 pmol) Analyzed with HPLC-ESI FTICR MS..109

3-25. Spectra of(1) Non-H/Dx of 10 pmol of HEWL and (2) 15 min H/Dx of 10 pmol
of H EW L ........................................................................ ........ 109

3-26. Sequence Coverage of Bovine Carbonic Anhydrase Digested with Pepsin.......... 110

3-27. Activity of Ins-ST / Endo Hf Digestion............... ....... ...... ...............115

4-1. ICAT Reagent................................. ....... ......... 126

4-2. IC A T W ork-fl ow ........... ............................................................... ............... 127

4-3 IO A Specificity ........ ........................................................................ ........ .. ....... .. 132

4-4. ED C Coupling w ith ETAM .......................................................... ............... 133

4-5. A cetylation of Tyrosine w ith N A I................................................ ........ ....... 133

4-6. D ifferential Labeling Experim ents .................................. ..................................... 134

4-7. Work Flow for Bioconjugation of Recombinant hST3Gal IV and Identification of
D erivatized A m ino A cids............................................... ............................ 136

4-8. L ab eling E xperim ent ....................................................................... .................. 138

4-9. NAI Derivatization of C-Tag-ST without Substrate Present and with 525 [tM
C M P -N euA c P resent ................................................................... ... .................. 138

4-10. NAI Derivatization of C-Tag-ST without Substrate Present, with 1.5 mM CMP-
NeuAc and Saturated a-Lactose Present, and with Saturated a-Lactose Present.. 139

4-11. Reactivation of N -Tag-ST with ETAM .............................................................. 141

4-12. SDS-PAGE Analysis of C-tag-ST and N-tag-ST...............................142









4-13. SDS-Page Analysis of C-Tag-ST and N-Tag-ST After Deglycosylation with
E ndoHf........................................................ ................... ... .... ... ... 143

4-14. SDS-PAGE Analysis of TCA Precipitated C-Tag-ST and N-Tag-ST with
Subsequent IGD with M multiple Proteases.................................... ............... 144

4-15. SDS-Page Analysis of Initial Screen of Ins-ST with EDAC/ETAM, NAI, and
IO A ............................................................................ 1 4 5

4-16. SDS-Page Analysis of C-Tag-ST Labeled with NAI, with and without
Substrate Present ........ ................... ............ .. .. .................. 146

4-17. MALDI-TOF MS Spectrum with On-Spot Water Wash Sample Preparation...... 146

4-18. MALDI-TOF MS Spectrum after ZipTip Sample Preparation..........................147

4-19. Sequence Coverage of C-Tag-ST IGD with Trypsin, On-Spot Water Wash
Cleanup, and MALDI-TOF MS Analysis...................... ...... ............... 147

4-20. Sequence Coverage of C-Tag-ST IGD with Trypsin, Zip-Tip Cleanup, and
MALDI-TOF MS analysis ............. ............... .................. 147

4-21. Typical Total Ion Chromatogram of Trypsin IGD of N-Tag-ST with an LCQ-
DECA. The Retention Times and Base Peaks are Labeled................................148

4-22. MS Spectra at 23.70 Minutes of the Tryptic Peptide LEDYFWVK .................149

4-23. MS/MS Spectra at 23.80 Minutes of the Tryptic Peptide LEDYFWVK .............149

4 -24. Sequence Coverage of Ins-ST Digested with Trypsin / LC MS/MS.................. 149

4-25. Sequence Coverage of N-Tag-ST Digested with Trypsin / LC MS/MS .............150

4-26. Sequence Coverage of C-Tag-ST Digested with Trypsin / LC MS/MS .............150

5-1. D different Photoreactive M oieties ...................................... ................. 162

5-2. Reaction Pathways of Activated Phenylazide ......................................163

5-3. Aromatic Ring Substituted for NeuAc and Ki Values......... ...............................168

5-4 C M P -o-azido M andelate............................................ ........................................ 168

5-5. Negative Ion Mode ESI-FTICR MS Spectrum of CMP-o-azido Mandelate ...........170

5-6. Negative Ion Mode ESI-FTICR MS Analysis of CMP-o-azido Mandelate
D ecom position Products .............................................. .............................. 170

5-7. Activity Profile of N-Tag-ST Incubated With and Without Inhibitor Present........171









5-8. Possible Labile Bonds of CMP-o-Azido Mandelate and Possible Mass Shifts
Associated with Derivatized Tryptic Peptide....................................................... 176

5-9. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/180 [tM
Inhibitor Experim ent .................. .............................. .... .. .. .. ........ .... 177

5-10. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/98 [tM
Inhibitor E xperim ent ................................................................. .................. 178

6-1. Spectrum of 10 pmol/[tL of 3'Modified 19 mer Diluted in Buffer 1 .................195

6-2. Spectrum of 10 pmol/[tL of 3'Modified 19 mer diluted in Buffer 2 ....................196

6-3. Charge State Distribution and Signal Intensity of 10 pmol/[tL of 3'Modifed 19
mer Diluted with the Different Buffer Systems ......................... ..................196

6-4 Charge State Distribution and Signal Intensity of 1 or 10 pmol/ p.L of 5 mer
D iluted in Buffer System 1 or 2 ........................................ ......................... 197

6-5. Linear Dynamic Range of the -2 Charge State of 5 mer Diluted with Buffer 2 ......197

6-6. IP-RP HPLC ESI FTICR MS Analysis of 15 picomoles of 5mer with a PS-DVB
Monolithic Column and 25 mM TEAA as an ion-pairing agent ............................198















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

MASS SPECTROMETRY FOR STRUCTURAL PROTEOMIC ANALYSIS OF
RECOMBINANT HUMAN SIALYLTRANSFERASE AND IDENTIFICATION OF
NANOPARTICLE HARVESTED OLIGONUCLEOTIDES

By

Jeremiah D. Tipton

December 2005

Chair: Nicole A. Horenstein
Major Department: Chemistry

Mass spectrometry (MS) has become one of the most powerful analytical

techniques for investigating proteins, peptides, and oligonucleotides when careful

preparation of biological samples has been performed. The first part of this dissertation

describes the overall goals and analytical tools needed for studying systems biology. As

a subcategory of systems biology, structural proteomics utilizes MS for the analysis of

the three-dimensional structure of proteins with low purification yields. Proteins may be

derivatized with bioconjugation techniques to provide a measurable mass shift which

then can be related structural information. The second part of this dissertation describes

the expression and purification of three recombinant isoforms of human ua-(2-3)

sialyltransferase (ST3GalIV) for bioconjugation experiments. Sialyltransferase was

chosen because it catalyzes the transfer of sialic acid to the terminal ends of

oligosaccharide chains found on glycolipids and glycoproteins. Sialic acids found on the









terminal end of oligosaccharides are involved with several key biological recognition

events such as cell adhesion, cell-cell recognition, and biological masking of disease

states. By describing the 3D-structure of ST3GalIV, rational design of new inhibitors

may proceed. The third part of this dissertation describes three separate bioconjugation

techniques used for the derivatization of ST3GalIV. These techniques include hydrogen

- deuterium exchange, derivatization with small molecule labeling agents, and

derivatization with site-directed photo-affinity labeling. After sample work-up, three

different MS instruments are utilized for mass measurement. The fourth part of this

dissertation describes the MS analysis of oligonucleotides selected with nanoharvesting

agents. Finally, the conclusions and future work for these projects are presented.














CHAPTER 1
INTRODUCTION TO BIOLOGICAL MASS SPECTROMETRY

Systems Biology and Mass Spectrometry

Mass spectrometry (MS) has become one of the most powerful analytical

techniques for investigating the diverse cascade of biological molecules generated by

metabolism in living organisms. MS as an analytical tool measures the mass-to-charge

(m/z) of an ion in the gas phase. Biological MS has created a huge enterprise that has

promised the molecular biologist and biochemists the instrument that will provide

answers to many of their biological questions.1 To be more precise with the opening

statement, MS is the most powerful analytical tool for identification of biological

molecules when careful preparation of biological samples has been performed and

bioinformatics software exists for complex data analysis. With the current state of

biotechnology, many disciplines have aligned to solve the dynamic systems biology

problem, with MS widely accepted as the main analytical tool.2

Systems biology, still in its infancy, is loosely defined as the study of the total

integrated network of all biochemical reactions and interactions between different micro-

and macromolecules within an organism which provide life sustaining processes.2 This is

different than traditional biology which uses the reductionism philosophy towards

studying individual genes, proteins, and cell specific function.3 However, systems

biology would not have a platform for integration of information without first cataloging

the specific function of genes, proteins, and cells. Much like the complex machinery

which creates life, many experts from all the sciences need to work together to produce a










viable picture of an organism which will in turn help increase our understanding of

human health and disease.

Development of technologies in the supporting fields for systems biology,

presented in Figure 1-1, have and will continue to radically transform biomedical and

medicinal research. The left column illustrates the main categories of biological

molecules and the middle column consists of the terms given to an entire complement of

that particular molecule. The right column represents the hierarchical approach for an

integrated view of how the different systems work together to create life, also known as

functional genomics. The term systems biology is used to describe an entire organisms

function as opposed to isolated parts, such as the genome, transcriptome, proteome or

metabolome.



DNA Functional Genomics or
Complete description of "phenomics"
genes in an organism Genome
Genomics
RNA
Complete description of
Complete description of Transcriptome Transcriptomics
mRNA in a genome
Proteins Global Targeted
Proteins
Complete description of
proteins expressed Proteome Proteomics
by a genome Global Targeted

Metabolites
Complete description of r Metabolome --Metabolomics
metabolites expressed
by a genome Global Targeted

Catalog of all protein-DNA,
protein-RNA, and protein-protein Interactome Systems Biology
interactions

Figure 1-1. Diagram Representing Components of Total Systems Biology 2

Although there are many technologies needed for systems biology analysis, this

dissertation is limited to (1) MS based structural proteomics in regards to studying the









protein tertiary structure of sialyltransferase and (2) MS analysis of oligonucleotides in

respect to nanoharvested oligonucleotides. Before describing these projects, the current

state of proteomics, separation technology, mass spectrometry instrumentation, and data

analysis are described. Tables 1-1 and 1-2 present the different functions of proteins and

what MS experiments may be performed to describe useful information.4

Table 1-1. Protein Function in Biological Organisms
Protein Function in Biological System
Catalysis
Defense
Movement
Structure
Storage
Transport
Signaling
Storage

Table 1-2. MS Experiments on Proteins
Mass Spectrometry Experiments
Interactions and Binding
Protein Folding
Protein Dynamics
Higher-order Structure
Primary Structure
Enzyme Kinetics
Post-translational modifications)

Mass Spectrometry Based Proteomics

Proteomics, as termed by Mark R. Wilkins in 1994, is defined as the genome

encoded protein complement at a certain time point.5 It is the goal of proteomics to study

what proteins are present, what the proteins do, and how they interact with each other.6

At different times during an organism's life cycle, different proteins may be present at

different concentrations. The proteomics field is viewed as a high-throughput

technology; however, the workflow and instrumentation are used for studies where very

specific chemical derivatization followed by MS analysis is interpreted to provide









characterization of active-site structure. Furthermore, proteomics may be sub-

categorized into functional and structural proteomics.7

Functional Proteomics

Functional proteomics focuses on identifying the primary sequence of a protein, the

concentration of proteins at different life cycles, the post-translational modifications

(PTMs) in static or dynamic states, and the protein-protein interactions of "mature"

proteins.3 MS Experiments are designed to identify proteins and protein-protein

interactions qualitatively and quantitatively with both high selectivity and high

sensitivity. It has been estimated that the protein concentrations within any living cell or

blood serum ranges between 5 to 15 orders-of-magnitude.8 Also, proteins may be spliced

into other isoforms or be modified with covalent attachment of a variety of PTMs upon

Tablel-3. Mass Shift Associated with Common PTMs
Pyroglutamic acid formed from Gln -17.03
Disulfide bond formation -2.02
Disulfide bond removal +2.02
C-terminal amide formed from Gly -0.98
Deamination of Asn and Gln +0.98
Formylation +28.01
Methylation +14.027
Acetylation +42.04
Carboxylation of Asp and Glu +44.01
Phosphorylation +79.98
Sulfonation +80.06
Cysteinylation +119.14
Myristoylation +210.36
Glycosylation
Deoxyhexoses (Fuc) +146.14
Hexosamine (GlcN, LalN) +161.16
Hexose (Glc, Gal, Man) +162.14
N-Acetylhexosamine (GlcNAc, GalNAc) +203.19
Pentose (Xyl, Ara) +132.12
Sialic acid (NeuNAc) +291.26
Pyridoxal phosphate (Shiff base to lysine) +231.15
ca-N-6-Phosphogluconoylation +258.12









suited for these types of studies.9 Data for functional proteomics are not presented in this

dissertation; however, many other groups have presented excellent data pertaining to the

profiling of certain cell types (i.e., healthy versus disease state),10 the mapping of small

proteomes (i.e., yeast),11J12 or characterizing whole blood serum.13;14

Structural Proteomics

Structural proteomics focuses on the structure-function relationships between

proteins and the mapping of three-dimensional structures of proteins.515 As presented in

Chapters 3 through 5, the mass shift of amino acids derivatized in vitro with specific

labeling agents can provide information on protein structure. Further discussion on the

experimental design for MS based structural proteomics, along with other competing

techniques is presented in Chapters 3 through 5. Obviously, without certain

developments in MS and molecular biology, proteomics would not exist.

Mass Spectrometry and Large Bio-molecules Meet

The discovery of the soft ionization techniques, electrospray ionization (ESI) in

1984 by Yamashita and Fenn16 and matrix-assisted laser desorption ionization (MALDI)

in 1988 by Tanaka et al.,17 both recognized with the Nobel Prize in Chemistry 2002,

allowed large non-volatile molecules to be ionized efficiently for MS analysis. 1819

Furthermore, tandem MS (MS/MS) is utilized to identify the amino acid sequence of

proteins or proteolytic peptides. The nomenclature that describes the different

fragmentation patterns observed with MS/MS analysis is shown in Figure 1-2. Figure 1-2

illustrates the basic theoretical peptides cleavages and nomenclature for basic

fragmentation modes.20

The human genome contains at least 40,000 genes expressing about 100,000

functionally unique protein types.2;8;9 After considering polymorphisms and PTMs,










X2 Y2 Z2 XI Yi zI
I I I I I I
I I I I I I
R I I I I I
H
N OH
H2N
I I I I I I
O R2 0


al bi ci a2 b2 c2

Figure 1-2. Common Fragmentation Patterns and Nomenclature

the total number of functionally different proteins could approach 1 million, especially at

different states during the cell cycle. The genome map is very useful in providing

theoretical polypeptides for automated interpretation of MS data. At first, it was

theorized that all the gene products (proteins) of a particular species may be predicted

with the genome map; however, it was quickly discovered that the genome did not

predict the protein's final "mature" structure or temporal production. Although these

three technologies were discovered in the last 10 to 20 years, proteomics is still a young

field of study with dynamic exponential growth in the areas of separation science, mass

spectrometry, and bioinformatics. Before reviewing these aspects in relation to

proteomics, MS analysis of oligonucleotides is introduced.

Mass Spectrometry Analysis of Oligonucleotides

MS analysis of oligonucleotides has not received the same acclaim in the biological

community because of already established protocols and instrumentation. For example,

large-scale genome sequencing has already been made high-throughput with DNA

sequencers. Multiplexed capillary electrophoresis is used to separate fluorophore-tagged









oligonucleotide ladders produced by dideoxy chemistry (Figure 1-3), in a similar manner

to LC MS/MS allowing for separating and sequencing of oligopeptides ladders (Figure 1-

4). The main difference between DNA and proteins is that DNA is static and can be

easily amplified with PCR, whereas proteins are expressed over a large dynamic range

and must be analyzed at the concentration from the particular state of the cell or system.

However, recombinant protein expression technology may be used for production of

proteins for structural proteomics (Chapter 2).



PCR Di Capillary
mDideoxvd Electrophoresis Sequence
~ I method (Sq e Alignment
Oligonucleotides (Sequ
DNA Ladders
Figure 1-3. Diagram Describing the Workflow for DNA Sequencing with CE


Digestion Digestion Mass Database Search
S'//' Spectrometry
(Sequencer)
Protein Peptides (Se ) Identification
Figure 1-4. Diagram Describing the Workflow for Protein Sequencing with MS

In a very similar fashion to proteins and peptides, the development of ESI and

MALDI has provided similar results for ionization of oligonucleotides for MS analysis.

Normally, proteins and peptides are ionized in positive ion mode, thus providing [M +

nH]n+ ions. In contrast, oligonucleotide ionization is performed in negative ion mode due

to the negatively charged phosphate backbone making [M nH]n- ions Also, similar to

proteins and peptides, oligonucleotides may be fragmented in the mass analyzer for

sequencing. Practical ventures for the use of MS for the analysis of oligonucleotides are

presented in Chapter 6.









Complete Mass Spectrometry Analysis

The complete MS analysis for structural proteomics is presented in Figure 1-5. In

the scope of this dissertation, the expression, purification, and characterization of three

recombinant forms of human ac 2-3 sialyltransferase (h23STGalIV) will be described in

Chapter 2. Further derivatization experiments of h23 STGalIV (hydrogen/deuterium

exchange, small affinity labels, and site directed photoaffinity labels) with the goal of

structural proteomic data specific to active site residues will be presented in Chapters 3

through 5. Before describing the wet chemistry preparation (Figure 1-5, Selection 1),

current analytical tools such as separation science, MS instrumentation (ionization to the

gas phase and mass-to-charge analysis), and bioinformatics are reviewed below.

Since the advent of ESI and MALDI and the completion of the human genome, MS

technology has been driven by the biotechnology sector to provide instruments with

improved sensitivity, mass accuracy, and duty cycle. It may take several months to

develop viable molecular biology and biochemical experiments for a one-day analysis

Selection 1: Molecular Biology and
Biochemical Sample Preparation


r ---------------------------------- ,-- -
I Selection2:
Chromatography

CIO
Ionization to Gas Phase C



Mass-to-Charge Analysis
------------------


Bioinformatics
Figure 1-5. Workflow for Complete MS Analysis









with a mass spectrometer. After MS analysis, the data could take days to months to

analyze for useful information unless a viable bioinformatics system is in hand. In

tandem with MS, separation science using 1- or 2-dimensional gel electrophoresis (1D

GE or 2DGE) or high performance liquid chromatography (HPLC) is needed for

chemical noise reduction. Selection or separation may be performed off-line or on-line

with respect to MS analysis. For HPLC on-line with MS, some view the mass

spectrometer as a detector for the chromatograph. Others view the chromatography as

the inlet for MS. Both views are incorrect, because the hyphenation of both sensitive

chromatography and sensitive MS truly creates a powerful tool. The understanding of the

advantages and disadvantages of current analytical tools allows for intelligent design of

biology and biochemistry experiments.

Separation Science

The type of separations needed prior to MS analysis depends on the experiment

performed. As described earlier, this may include a single protein to a set of proteins,

cell or organelle types, or a complete organism with a relatively small proteome.

Proteomics has two main strategies for analyzing proteins, including the "bottom-up" or

"top-down" strategies.21 The "bottom-up" approach includes processing the protein into

peptides via chemical or proteolytic digestion before separation and MS analysis,

whereas the "top-down" approach relies on fragmentation of intact protein within the

mass spectrometer for protein sequence.22;23 Currently, most labs rely on the "bottom-

up" approach because of the high cost associated with instrumentation needed for the

"top-down" approach. The following sections will describe separation science for

traditional and the high-throughput workflows in respect to off-line and on-line with MS

analysis.











"Bottom-Up" Approach "Top-Down" Approach

Traditional High-Throughput
On-Line RP
Separation of Digestion HPLC of Intact Proteins
Proteins of Entire
(SDS-PAGE) Sample


,Off-line
In-l Separation
In-Gel
of Peptides
(SCX, etc.)



Off-line or on-line RP-HPLC
Separation of Peptides

ESI-FTICR-MS
MALDI-TOF ESI-MS (MS/MS)
Analysis Analysis Analysis


Figure 1-6. "Bottom-Up" and "Top-Down" Workflows

Off-line separation technology

The most powerful off-line separation techniques was for many years the use of 1-

dimensinal or 2-dimensional gel electrophoresis (1DGE or 2DGE) with polyacrylamide

gels.2;8 Proteins may be separated based on size, pi, or both size and pi. Traditionally,

proteins are separated with 2DGE and visualized with a variety of different staining

agents. For initial proteomics strategies, 2DGE has strong roots with differential

displaying of proteins expressed in different cell states (i.e., healthy versus disease) with

imaging programs. The most common choices for protein visualization with

polyacrylamide gels includes silver staining (linear dynamic range (LDR) of 4 orders of

magnitude; 500 pg to 5 ng limit-of-detection (LOD), coomassie blue (LDR of 2 orders of

magnitude;10 50 ng LOD), or Sypro Ruby (LDR of 3 orders of magnitude; 100 pg to 1









ng sensitivity).24 To increase sensitivity, the use of Western blotting with specific

antibodies lowers the LOD to 0.25 to 1 ng. However, since the antibody interaction is

only specific to one type of protein, Western blotting is more expensive than basic

staining methods and does not allow for high-throughput.

The main problems with 2DGE is that different proteins react uniquely with

different stains and that protein concentrations varying between 5 to 6 orders-of-

magnitude for different cells and 12 tol3 orders-of-magnitude in extra-cellular fluid.

Also, 2DGE is not very sensitive for proteins of different extremes such as very acidic or

basic, very small or large, and membrane proteins which may exhibit hydrophobic

character.2 In regards to both functional and structural proteomics, proteins separated

with 1D or 2DGE may be digested with common in-gel protease digestion protocols

unless silver staining is used. Furthermore, coomassie blue as a staining agent was

correlated to higher sensitivity with MS analysis after in-gel digestion with proteases.25

The one-protein system described in Chapters 4 and 5 used 1DGE and in-gel digestion

protocols. After 1DGE analysis and in-gel digestion, another separation step is usually

needed to remove salt impurities prior to MS analysis.

Another common off-line separation technique includes the use of Zip Tips.8;26

Zip Tips are pipette tips which have been packed with reverse-phase column material or

metal chelating agents.2 Peptide samples which have been digested are loaded onto a

small volume of particles, washed several times to remove salts or impurities, eluted, and

analyzed with MALDI-MS. Furthermore, particles with chelating agents may be bound

to divalent cations, termed immobilized metal affinity columns (IMAC), and used to trap









certain species, such as phosphorylated peptides. As presented in Chapters 4 and 5, Zip

Tip technology was used to clean up peptide samples after in-gel digestion.

On-line separation technology

Reverse phase (RP) high performance liquid chromatography (HPLC) is most

commonly used for coupling separations with MS because the solvents (mobile phase)

for separations are compatible with ESI (Table 1-3). For effective separation on

common RP materials (C4 or C18), an ion-pairing agent such as acetic acid, trifluoracetic

acid, heptafluorobutyic acid, or formic acid must be present in the mobile phase to

suppress peak broadening. Common solvents for RP chromatography include water,

methanol, acetonitrile, and isopropanol. Coupling to ESI is possible because these ion

pairing agents and solvents are volatile. Solvents may be delivered with a variety of

commercially available pumping systems. Gradients of increasing percent organic may

be delivered at a rate depending on the complexity of the sample. Proteins and peptides

are retained on RP material at high percent aqueous solution, then elute as the percent

organic increases.

Sample introduction is commonly provided with commercially available injectors

or loaded off-line with pressure bombs. Realistically, most of the instrumentation used

for solvent delivery, tubing, zero-dead volume connectors, and sample injectors are

commercially available. The major recent developments for HPLC-MS technology

include the downsizing of volumes and flow rates, systems that can handle very high

backpressure, different types of column materials, and hyphenation of two different

column materials in-line. The tools for separation of intact proteins and peptides were

already in hand to accomplish the work described in this dissertation. The choice of a

separation method depends on sample complexity and cost effectiveness. Strategies have









already been developed to downsize column size and to move away from 2DGE

technology to 2D or multidimensional chromatography. Still, the normal trend in

analytical chemistry is to develop sensitive, faster, and cheaper tools.27

The use of capillary (microbore) columns increased sensitivities as described by

Wilm and Mann in 1994.28 Different commercially prepared column materials may be

packed into capillary columns in-house, thus greatly reducing the cost for commercial

columns, the amount of sample consumed, and the volume injected.29 Furthermore, two

different types of column materials may be packed together to obtain 2D HLPC. In order

to retain chromatography material in the capillary column, sintered silica particles or

silicate-polymerized ceramics have been used as frits.30 Fused silica may be ground

down or pulled for creation of a small orifice tip for generation of ESI. The main

problems with downsizing chromatography systems to 100 nL/min flow rates include

increased backpressure, dead-volume, and leaks. Lastly, a voltage potential must be

applied through a liquid junction found close to the ESI tip. These issues have been

worked out by a variety of labs, thus making capillary HPLC columns common

practice.26;31;32 Since MS is especially well suited to analyze very small amounts of

sample nanoo to femtomoles), downsizing to microbore HPLC has been well received.

"Bottom-up" high-throughput functional proteomics uses two different workflows

depending on sample complexity. If the sample is considered to be of low to moderate

complexity, 2-dimensional high-performance liquid chromatography (2D HPLC) may

achieve the separation needed (Figure 1-7).8;26;27 This is performed by using capillary

electrophoresis (CE), size exclusion (SE), strong anion or cation exchange (SAX or

SCX), or IMAC support phase up-stream from RP separation. The initial column is









loaded with sample, followed by a step gradient elution with the specific salt buffer for

that column material. The solvents normally associated with columns materials other

than RP use buffers that are not compatible with ESI. For each step gradient cycle, a RP

column down-stream retains the peptides eluted from the first separation. Next, a switch

valve changes the flow direction and the mobile phase to elute the peptides from the RP

material.33 Some examples of 2D HPLC include MudPIT34 and "peak parking."35;36

If a sample is too complex, samples are pre-fractionated into small samples prior to

on-line HPLC-MS. As opposed to 2DGE, strongly acidic, strongly basic, extremely

small, extremely large, and hydrophobic proteins and peptides may be analyzed. The

most common analytical set-up for maximum sensitivity for peptide analysis includes the

use of 2D chromatography with a C18 trapping column and a C18 analytical column. The

HPLC-MS system used for analysis of the in-gel digested sialyltransferase samples

presented in Chapters 4 and 5 included a C8i trapping column and a microbore analytical

column.


Liquid-metal interface Capillary opening into MS


Mobile Phases 4
SCX Material RP Material


Figure 1-7. Microbore MudPIT

Ionization Techniques: ESI and MALDI

Although there are many ionization methods available, ESI and MALDI are the

primary methods for analysis of proteins and peptides with MS. ESI is an atmospheric

pressure ionization event, after which ions are moved into low or high vacuum regions.

The mechanism of ion formation is still under discussion; however, the most widely










accepted theory for ionization of large molecules includes Dole's charged residue model

(Figure 1-8).37 First, the voltage is increased at the tip of the ESI needle or the capillary

opening into the MS, thus forming a Taylor cone with droplets of relatively high surface

charge density. At a certain high voltage, the Taylor cone is destroyed, thus destroying

electrospray. Second, the carrier solvent evaporates, causing the droplet to shrink, thus

creating high charge-to-mass character. Next, the droplets eject charged particles

because the critical Rayleigh limit has been reached. Finally, the desolvation process

continues with assistance from a heated capillary inlet until individual ions form.







ESI Needle
(-3000 V)
Droplet at
Critical Rayleigh




Droplet with high
charge-to-mass .............+
character



Figure 1-8. Electrospray Ionization with Dole's Large Molecule Ionization Model

Over the last 15 years, optimization of the flow rate, the applied potential, and the

tip diameter with respect consistent MS signal have proceeded. For higher flow rates (2

[tL/min tol mL/min), a nebulizing (drying gas) is used to assist with ionization; however,

high flow rates may be detrimental to pumping systems. At lower flow rates, termed

nano-ESI (nESI), no drying gas is present. Nano-ESI tips are normally associated with

capillary chromatography and have very small tip diameters (2 to 50 [tm). Greater









sensitivity for analysis of biological molecules is observed experimentally with nESI.38;39

First, the low flow rate associated with nESI produces a smaller Taylor cone which

permits the emitter to be placed much closer to the MS inlet, thus effectively increasing

sampling efficiency. Second, nESI generates smaller charged droplets, therefore

efficiently increasing desolvation rate. Third, much less solvent is delivered to the mass

analyzer thus greatly reducing the solvent load to the vacuum system. Fourth, nESI

consumes much less analyte as compared with conventional ptESI. Finally, micobore

HPLC-nESI-MS increases the sensitivity of the instrument due to concentration of

sample on the RP material.

In respect to on-line HPLC-MS, some species may co-elute. With ESI, multiple

analytes co-eluting will create a competition between each of the analytes. Based on

studies using dyes, hydrophobic molecules will collect on the surface on the ESI droplet,

whereas hydrophilic molecules will be found in a heterogeneous mixture within the

droplet.40 Therefore, an analyte with increased hydrophobicity is more likely to "win-

out" during the ionization event. The problem of co-eluting analytes may be solved with

changing the gradient of the mobile phase delivered to the column, changing ion-pairing

agents in the mobile phase, or changing the scan functions of the MS.

Electrospray has the advantage of imparting multiple charges onto proteins or

peptides, thus bringing high molecular weight molecules within the m/z range exhibited

by modern mass analyzers41;42 In positive ion mode, polypeptides with basic amino acids

will carry positive charge. On average, every 10 amino acids within a polypeptide has an

ionizable group for protonation.41 Furthermore, the protease trypsin cleaves on the C-

terminal end of arginine and lysine, thus leaving an amino acid with an amine for









protonation. Changes in analyte solvent, pH, and temperature produce different charge

states of proteins observed with MS analysis.39;43

Knowing the primary sequence of the protein will allow a maximum number of

protonation sites to be calculated; however, the pi of a protein or peptide does not relate

to ionization efficiency. Furthermore, the fold state of the protein will affect charge state

distribution. Finally, to determine the charge-state of the proteins or peptides, the inverse

of the mass difference between peaks of the isotopic distribution is calculated.41 Table

1-4 represents the signal intensity of three different proteins versus ion-pairing agents.

This table provides the basis for solvent system chosen to analyze intact proteins in

Chapter 3.

Negative ion mode is used more frequently for oligosaccharides, oligonucleotides

and non-standard experiments. For negative ion mode, all of the polarities applied to the

hardware of the instrument must be reversed for ion generation and detection. Often, in

negative ion mode, arcing occurs if the percent aqueous solution is above 40%. Table 1-

4 illustrates the solvents compatible for positive and negative ion mode. Note that most

of these solvents are compatible with RP HPLC.

Tablel-4. Solvents Compatible with ESI
pH acetic acid, formic acid, trifluoroacetic acid* for positive-ion
detection (0.1 % up to 3.0 % v/v)
Buffers, ion Ammonium acetate, ammonium format, triethylamine
pair-reagents heptafluorobutyric acid (HFBA), tetraethyl or tetrabutylammonium
hydroxide. (10- 100mM)
Cation Reagents Potassium or sodium acetate (20 50 uM level)
Solvents Methanol, ethanol, propanol, isopropanol, butanol, acetonitrile,
water, acetic acid, formic acid, acetone, dimethylformamide,
dimethyl sulfoxide, 2-methoxy ethanol, tetrahydrofuran,
dichlormethane, chloroform









MALDI ionization proceeds with N2 laser (337 nm) short pulse irradiation of

44
analytes distributed in a matrix thus creating a plume of neutrals, clusters, and ions.44

Similarly to ESI, the mechanism of ionization is not fully understood. As opposed to

ESI, MALDI predominantly creates singly charged ions. The common matrices are

chromophors, therefore they absorb energy after irradiation and, according to the

prevalent ionization theories, provide a source for protons or other cations to aid in

ionization. According to the current ionization models, the mechanism for ionization is

most likely specific to the type of analyte being investigated. Some current ionization

models suggest that peptides or proteins are pre-protonated in the matrix because of the

low pH preparation protocols.45 Other models suggest that neutral ion-ion reactions

within the post-ablation plume creates ions.46 These ionization reactions occur in the first

tens of nanoseconds after irradiation within the initial desorbing matrix/analyte cloud and

have initial velocities between 300 and 800 m/s.

Nano- to picomoles of proteolytic peptides are routinely analyzed. For the purpose

of proteomics, different studies require different sample protocols found in the literature.

Typically, 1 ptL of 500 nM tolO [tM of protein or proteolytic peptide sample is mixed

with 5 to 10 ptL of 10 mM matrix (2,5-dihydroxybenzoic acid (DHB); 3,5-dimethoxy-4-

hydroxycinnamic acid (sinapinic acid); or a-cyano-4-hydroxycinnamic acid (c-CHCA)

dissolved in 50% ACN, 0.1 % TFA), after which it is spotted onto a sample support (i.e.,

stainless steel plate).

MALDI may be performed at atmospheric pressure; however, ionization is most

commonly performed under vacuum. This increases the duty cycle because the

ionization source must be loaded and pumped down before the experiment can be









performed. Compared to ESI-MS, MALDI-MS is more tolerant to detergents, salts and

buffers; however, there is more ion suppression when a complex sample is present.8 In

the work presented in Chapters 4 and 5, both ionization methods were used with different

mass spectrometers for analysis of intact proteins and peptides.

UV Laser
Pulse = Matrix
0 = Analyte
Desorption Plume


Sample Plate (
01







Figure 1-9. Diagram Representing Matrix-Assisted Laser Desorption Ionization

Mass Analyzers

Mass-to-charge analysis may take place in one of many mass analyzers developed

over the last 108 years. The first mass analyzer was built by J.J. Thompson at

Cambridge's Cavendish Laboratory in 1897. Since then, MS has provided the tools

necessary for characterization of small synthetic molecules to whole viruses.47 Although

many mass spectrometers are manufactured to analyze various types of small and large

molecules, discussion is limited to instruments for which the primary use is analysis of

proteins, peptides, or oligonucleotides. These include quadrupole ion traps (QITs) MS,

time-of-flight (TOF) MS, and Fourier transform ion cyclotron resonance (FT-ICR) MS.48

Over the last ten years, ample commercial instruments were developed to accommodate

ESI and MALDI sources. Also, several instruments which have different combinations of









the individual mass analyzers, i.e. hybrid MS instruments, have been made commercially

available.

Quadruple ion trap MS

The QIT may be considered one of the most robust mass analyzers on the market.

The QIT was originally conceived by Wolfgang Paul (Nobel Prize in Physics, 1989) and

co-workers in the early 1950's and made commercially available in 1983 as a GC

detector. Over the last 50 years, many laboratories have improved cell design and ion

transfer optics to make the instrument available to a variety of investigators.

Operationally, ions may be generated externally, after which they are guided and trapped

inside a cell between two end plates and a ring electrode. Once trapped inside the cell, a

radio frequency is applied to the ring electrode and ramped with increasing amplitude,

thus ejecting the ions in the direction perpendicular to the ring electrode. The ions hit a

conversion dynode, thus ejecting electrons to an electron multiplier for amplification of

signal. Different ions of different m/z have different stabilities within the radio frequency

and electric fields. Ions of lower m/z are ejected first as the radio frequency is ramped.49

The real power of QITs, in terms of proteomics, is the fast scan rates which are

important for detection of peptides after RP-HPLC. Multiple MS scans with QITs differ

from the other instrument types because it has the ability to fragment ions and store them

for several cycles. Scan functions are designed to operate in what is called data-

dependent mode. In data dependent mode, after detection of a relatively abundant m/z, a

subsequent scan isolates that particular m/z prior to MS/MS analysis. Tandem MS is

performed by increasing the pressure in the QIT cell by introducing a pulse of helium,

thus creating what is termed collisionally activated or collisionally induced dissociation

(CAD or CID).49









In respect to proteomics, low energy CID predominantly fragments only the amide

backbone of the polypeptide. Fragmentation is dependent on the polypeptides sequence,

length, and charge-state. In general, different peptides may or may not completely

fragment. Often, intact peptides or fragmented peptides may loose H20, NH3, or CO2.

Furthermore, it has been observed that doubly-charged ions generally provide improved

MS/MS spectra. For QITs, mostly b- and y- type ions are observed with CID (Figure 1-

2). Rearrangements of peptides (intermolecular reactions) in the gas phase been reported

in the literature, however they are not reviewed in this dissertation.20

There are many scan functions available for different MS experiments, specifically

product ion scan, precursor ion scan, and constant neutral loss.8 For example, some

PTMs such as phosphorylation have less stable bonds than the backbone of a polypeptide,

therefore when MS/MS is performed, the major peak observed may be [M 98]+1 or [M -

80]+1. A scan function may be set to select for the neutral loss of phosphate, followed

MS/MS (considered MS/MS/MS or MS3) on the m/z that is missing the phosphate.

The real limitations of the instrument include poor quality data sets due to space

charging events and a limited m/z range of 20 to 2000 m/z. Space charging is an effect

that is observed when too many ions are trapped in the cell, thus causing ions to "leak

out" of the cell by repulsion. To eliminate space charging, automatic gain control (AGC)

is a function that monitors the amount of ions in the cell. Before mass measurement, a

pre-scan measures the amount of ions entering the cell (total ion count). Depending on

the counts, the ion collection time for mass analysis may be adjusted from 10 to 25

microseconds to allow the proper amount of ions to be collected within the linear

dynamic range of the cell.49









QITs are relatively sensitive; however, resolution may be from 1 to 0.1 mass units.

Calibration may be performed externally with common mass errors between 50 to 500

ppm for proteomics. Often, data sets collected for proteomic studies are of poor quality

because of space charging events due to co-eluting peptides. Internal calibration may

improve mass accuracy; however, the internal calibrant may inhibit ionization of a

peptide of interest. Finally, since QITs are scanning instruments, MS and MS/MS data

on certain peptides may be excluded from a particular LC-MS analysis. Currently, QITs

equipped with capillary HPLC nESI have provided the best results in respect to the

capital cost for proteomics. The data presented in Chapters 4 and 5 was collected with

this type of mass spectrometer.

Time-of-flight MS

Time-of-flight (TOF) MS was introduced by William E. Stephens at the University

of Pennsylvania in 1946 and made commercially available in the late 1950's by William

C. Wiley and I. H. McLaren at the Bendix Corporation Detroit, Michigan. TOF MS

separates ions of different m/z based on the kinetic energy as they drift down a tube at

different velosities. For many years, TOF instruments were not considered useful mass

spectrometers because they lacked the resolution needed for routine analysis. The

resolution of TOF mass analyzers was improved by the introduction of delayed extraction

(DE) which normalized the initial velocities of analytes created by a MALDI.49 Ionized

molecules are slowed with DE (-2000 V) and then pulsed (-20,000 V) to a detector at the

other send of the tube.

TOF MS instruments may be operated in two different modes. The first, linear

mode, has a mass range of up to about 1 M Da, but often suffers from poor resolution.

The second, reflectron mode, suffers from a limited mass range (100 6000 m/z), but has









an improved mass accuracy (-5 ppm). Ionized molecules are slowed with DE and then

pulsed down to a series of "focusing" plates (reflectron) which concentrate and redirect

the ions to another detector. Traditionally, a MALDI source is used with TOF

technology for analysis of intact proteins and peptides due to the fact that MALDI

predominately creates singly charged ions. The analysis of singly charged, high mass

proteins is best suited for TOF mass analyzers operated in linear mode if poor resolution

is not an issue. TOF is the easiest MS technology to use due to the fact that it does not

suffer from the space charge effects observed with ion trap MS.

Peptide sequences are observed from the fragmentation of metastable ions during

flight down the TOF tube, thus creating post source decay (PSD). First, a spectrum is

collected in linear mode, thus allowing the identification of the m/z for the peptide of

interest. Next, metastable ions which undergo fragmentation during flight are analyzed in

reflectron mode, thus yielding a spectrum with ions of less m/z than observed in linear

mode. PSD spectra may not be as straight forward to interpret as the MS/MS spectra

obtained from QITs because fragmentation may not be as complete. Therefore, quality

MS/MS data is normally associated with QITs.24

Over the last 10 years, TOF technology has also integrated ESI sources. The

current wave of ESI-TOF-MS instruments has excellent scan rates, mass accuracy, and

mass resolution in comparison to their predecessors. Hybrid instruments which contain a

QIT prior to a TOF (Q-TOF MS) are also commercially available. These types of hybrid

instruments take advantage of the CID capabilities with QITs and the higher mass

resolution capabilities of reflectron TOF MS.









MALDI-TOF has been used in forensics and homeland security to identify possible

peptide or protein mass markers in respect to bioterrorism.50 Also, MALDI-TOF has

been used for imaging different proteins in cells and organelles.51 Furthermore, MALDI-

TOF has strong roots in peptide mass fingerprinting of in-gel digested proteins separated

with 1D and 2DGE.8;52 This instrument was used to monitor peptide digestion as

presented in Chapters 4 and 5.

Fourier transform ion cyclotron resonance MS

The concept of ion cyclotron resonance (ICR) in a magnetic filed was developed by

E.O Lawrence and co-worked in 1931, followed by electrostatic confinement by F. M.

Penning in 1936 in a trap, later known as a "Penning Trap". M. B. Comisarow and A. G.

Marshall revolutionized ICR in 1974 by using Fourier transform (FT) with ICR analysis.

FT allowed several ions to be measured in the cell at one time, instead of only measuring

one ion at a time. Since then, several laboratories have continued to improve FT-ICR MS

instrument design by increasing mass range, mass resolving power, and sensitivity to

produce the highest resolution and sensitivity possible for commercially available mass

spectrometers.

Several optics are needed to accelerate the ion beam past the fringe field of the

magnet (Figure 1-10) which encases the penning trap. Ions are generated externally, after

which they are pulsed into and trapped in the Infinity cell. Once trapped in the cell

with DC voltages applied on the end plates, the ions go into cyclotron and magnetron

motion due to a high magnetic field being present. Figure 1-11 illustrates the cyclotron

motion of a positive or negative ion in a magnetic field where B is the magnetic field in









Tesla, m is the unit mass of an ion, q is the charge of the ion and o, is the angular

velocity in radians/second.

Each m/z has a characteristic cyclotron frequency independent of the radius and

velocity; however, the ion cyclotron frequency is inversely related the m/z, thus

resolution decreases as m/z increases. Figure 1-15 represents the equation governing the

relationship between frequency and m/z where Bis the magnetic field in Tesla, m/z is the

unified mass-to-charge of an ion, o, is the angular velocity in radians/second, and v, is

the velocity in Hertz.
























Figure 1-10. Actively Shielded 4.7 T Magnet at the University of Florida

B B


Figure 1-11. Cyclotron Motion of Ions in Magnetic Field









,m 1.535611 x 107 B
Vc-
S 2 m/z

Figure 1-12. Relationship between the Frequency and Mass-to-Charge of an Ion

Signal cannot be obtain by just measuring the induced magnetron and cyclotron

motion of an ion in a magnetic field. The molecules are excited with a sweeping radio

frequency applied by two opposing plates, thus sending the spatially coherent ion

population into a larger cyclotron orbit. As the orbit decays back to the center, two other

parallel plates measure the differential image current through induction. Following

measurement, the time-domain signal is Fourier transformed to the mass-to-charge

domain. Since the frequency may be measured very precisely, FTICR MS offers very

high resolution. Theoretically, at ultra-low pressures, the observation of signal may be

indefinite for ion measurement and re-measurement. Also, the signal is quantifiable

because transduction responds linearly to the number of ions in the cell. Finally, since

ICR frequencies for ions are observed between 1000 to 1,000,000 Hz, already

commercially available electronics are incorporated into instrumentation.53;54

The main limitations with FTICR MS are the purchase and maintenance costs. To

purchase a basic FTICR MS with a 4.7 T magnet and infrared multiphoton dissociation,

an investigator must relinquish $ 750,000. Including capital cost, maintenance costs are

also high. First, common magnets (4.7, 7 or 9 T) require constant N2 and liquid He re-

filling for maintaining the high-magnetic field. Secondly, sophisticated pumping systems

with 2 to 3 turbomolecular pumps and 3 to 5 roughing pumps are needed to maintain a

cell pressure of at least 10-10 torr for consistent behavior of ions in the cell. If the

pressure is greater than 10-10 torr, gases in the cell dampen the ion cyclotron decay time,

thus destroying the sensitivity and mass resolution provided by the instrument. External









calibration provides mass accuracies of less than 5 ppm as long as the number of ions in

the cell remains relatively consistent between scans. If the number of ions change

dramatically or if there are too many ions within the cell, space charging will shift the

m/z values by repulsion or coalescence of ions packets.

Tandem MS analysis may be performed with one of many different fragmentation

methods.8 These include CAD, electron capture dissociation (ECD), sustained off-

resonance irradiation (SORI) fragmentation, infrared multiphoton dissociation (IRMD),

or blackbody infrared radiative dissociation. Besides CAD, as presented with QITs, these

methods may produce different fragmentation patterns based on the energy associated

with the fragmentation method. Fundamentally, the most important dissociation method

for proteomics includes the use of ECD of intact proteins. As presented in the separation

science section, the "top-down" method only requires separation of intact proteins prior

to MS/MS analysis. ECD, as initially presented by McLafferty et al., fragments intact

proteins though the transfer of electrons.23 The energy released upon transfer of electrons

to intact proteins will fragment the polypeptide backbone to create c- and z-type ions

(Figure 1-10).55 This fragmentation method produces information rich spectra; however,

interpretation has been problematic.

For proteomics, if funding is not an issue, the benefits of FTICR MS are superior to

other types of mass spectrometers. Identification of peptides with FTICR MS often relies

on exact mass; however, mass redundancy due to peptides with the same amino acids in

different sequences may provide unambiguous results. In terms of sensitivity, 1000

proteins in the 2 -100 kDa molecular mass range were identified from 200 -300

nanograms of an E. coli cell extract with LC-ESI-FTICR-MS.56 Furthermore, a study on









cytochrome c reported that 10 zeptomoles (10-21 moles) was efficiently measured.57 For

real world samples, 9 x 10 -18 moles of cytochrome C, representing about 1 % by weight

of proteins in blood cells, was extracted from crude red blood cells and analyzed with

LC-ESI-FTICR-MS. Furthermore, a difference of 9.5 mDa (-200,000 resolution) was

achieved to separate phosphorylation (31.9816 Da) and sulfonation (21.9721 Da)

PTMs.58 As with Q-TOF MS, hybrid instruments with a QIT added before a FTICR MS

(Q-ICR MS) are commercially available, thus providing an added degree of mass

selection. Experiments explained in Chapter 3 use ESI FTICR MS to analyze simple

intact protein samples.

Bioinformatics

As with the growing field of proteomics, bioinformatics is also loosely defined.

Over the years, the term bioinformatics has included software written for molecular

modeling, gene prediction, sequence alignment, bio-molecule array design, mass

spectrometry data analysis, amino acid sequence prediction, and laboratory informatics.

Furthermore, several disciplines such as computational biology, biomathematics,

biometrics, and biostatistics form specialized bioinformatics platforms.7;59

Bioinformatics software are designed for visualization, databases, knowledge

representation, software development, and algorithm development.7

Computer science experts are often hired to aid in the large task of data

management because the high quality interpretation of ample MS data represents a

serious bottleneck in proteomic workflows.7 The development of new MS technologies

for sequencing peptides i.e., ECD with ESI-FTICR MS, created the need for new

algorithms for automated interpretation. Although there are several applications for

bioinformatics, only those pertaining to analysis of proteomic MS data are reviewed.









Furthermore, programs used for qualification of proteins and peptides, not quantitation,

are reviewed.

Currently, there are three main strategies for interpretation of proteomic MS

data.118 The first two includes programs written to compare experimental proteomic MS

data with theoretical in silico digested polypeptides based on the knowledge of a certain

genome. Functionally, in silico mass lists of possible theoretical peptides derived from

theoretical proteins as determined from a known open reading frame from a genome

sequence may be predicted based on the specificity of proteolytic or chemical cleavage.

The experimental peptide m/z list (mass list extracted from experimental MS data) of a

single digested protein is compared with a theoretical peptide mass library generated

from genomic information from an open access databases.7 Positive identification is

possible if the genome of the species is found in a genomic database. Table 1-5

represents a list of common databases containing different species genomes. The third

method, de novo sequencing and identification, relies entirely on mass differences

generated with experimental MS/MS data for the sequencing and identification of

polypeptides.

Table 1-5. Common Genomic Databases
Database Name Wprl-Wide Web Address
EMBL http://www.embl-heidelberg.de/
GenBank http://www.ncbi.nlm.nih.gov/GenBank/
SRS http://www.srs.ebi.ac.uk/
ExPaSy http://www.expasy.ch/
YPD http://www.proteome.com/YPDhome.html
MIPS http://mips.gsf.de/

Intact peptide MS data analysis with in silico digestion algorithms

In silico data analysis, in its simplest form, can find its roots with the peptide mass

fingerprinting algorithm (PMF) originally presented in 1984 with fast atom bombardment









ionization (FAB) MS.1;60 FAB did not produce reliable results and was by no means

high-throughput; however, the trend to perform in-gel digestion followed by MS analysis

also had it's start with FAB MS. Nine years later, after the advent and improvement of

MALDI TOF-MS, Hanzel and co-workers illustrated that PMF may be used to analyze a

variety of gel spots in a semi high-throughput fashion.24 Peptide mass fingerprinting is

useful for the identification of 1 to 3 proteins at a time because MALDI-TOF analysis of

multiple protein digested samples may lead to ion-signal suppression, overlap of signal

corresponding to different peptides, and mass redundancy. A database search is

performed on an experimental peptide m/z lists with one of the programs presented in

Table 1-6 by defining the pi, protease used (cleavage specificity), the number of possible

missed cleavages by the protease defined, the mass tolerance (mass error) of the

instrument, the minimum number of peptide matches required for identify a protein, and

the species from which the protein is extracted.24 Including these parameters, several

common amino acid in vivo PTMs or in vitro chemical modifications may be defined.

Tandem MS data analysis with in silico digestion algorithm

Five different groups in 1993 independently developed software for interpretation

of experimental MS/MS data, including W.J. Hanzel and coworkers, C. Wantanabe, D. J.

Pappin and coworkers, P. James and coworkers, and J. R. Yates III and coworkers.7 The

algorithm developed may be termed expression sequence tag (EST), sequence tag

identification, peptide mass tag (PMT), or peptide fragmentation fingerprinting (PFF).

These strategies were developed due to the huge amounts of information rich data

generated by ESI MS/MS.24 Unlike the PMF method which analyzes intact peptide m/z,

the EST method is used to analyze fragmentation peptide spectra. With EST, the









theoretical in silico peptides are fragmented into possible masses representing the peptide

ladder of that particular peptide for comparison to experimental MS/MS data.

The biggest problem associated with automated data analysis is the incorrect

assignment of b- and y- type ions. Upon fragmentation, peptides may undergo

rearrangements or lose H20, CO2, and/or NH3. Furthermore, a precursor ion may lose

H20, C02, or NH3 or internally fragmented upon ionization. If an intact peptide (parent

ions) undergoes these types of mass loss, fragmentations, or intermolecular

rearrangement, it will not be identified because the precursor or MS/MS ion mass does

not match the theoretical in silico m/z.60 For programs using the EST algorithm,

experimental MS/MS m/z lists, protease specificity, number of missed cleavages,

instrument mass tolerance (mass error), taxonomy, and variable and fixed amino acid

modifications are defined.60 Currently, Mascot and Sequest are the main packages used

to aid in interpretation of proteomic MS data.

Tandem MS data analysis with de novo algorithims

If the genome of a species is unknown, de novo sequencing may provide improved

results over in silico algorithms; however, de novo sequencing algorithms will not always

provide correct assignment of peptide sequences.7;24 Older de novo sequencing programs

include the generation of all theoretically possible sequences. Experimental MS/MS m/z

lists are compared against theoretical spectra for the best match. Using this

methodology provides many possible solutions because the number of sequences grows

exponentially as the length of the polypeptide increases.

The newer programs simply derive a theoretical sequence based on the

experimental MS/MS m/z list. Currently, de novo peptide sequencing from MS/MS data

is computationally laborious. Often MS/MS spectra lack enough high quality peaks for









assignment of a complete peptide sequence. A new dissociation method, electron transfer

dissociation (ETD) with linear ion traps (LIT), produce c- and z- type ions which appear

to have well defined fragmentation patterns.61 ETD may provide improved MS/MS

spectra for de novo sequencing; however, two main problems still exist for this type of

sequence alignment. First, several amino acids and amino acid pairs, i.e.

leucine/isoleucine or glutamine/lysine, have the same or similar nominal masses.

Second, cleavage does not always occur at every peptide bond. For example, some

fragmentation ions are below the noise level, the C-terminal side of proline is often

resistant to cleavage, mobile protons are often absent, and peptides with free N-termini

often lack fragmentation between the first and second amino acids.

Software packages for analysis of MS and MS/MS data

Table 1-6 lists many of the programs available for MS and MS/MS data analysis,

with Sequest and Mascot being the most common.62;63 Sequest was the first

commercially available software package for of proteomic MS/MS data analysis as

developed by J. Yates III and co-workers in early 1994. Sequest does not rely on a data

preprocessing step because the algorithm is based on cross correlation between

experimental MS/MS data and an in silico MS/MS spectrum. Mascot was developed in

collaboration by D. Pappin at Imperial Cancer Research Fund, UK and A. Bleasby at the

SERC Daresbury Laboratory, UK under the name MOWSE. After further development

with the search algorithms, Matrix Science (Boston, MS) purchased the MOWSE

software to distribute it under the new name Mascot. In contrast to Sequest, Mascot

requires a pre-interpretation step to create 'dta' files which contain the experimental

MS/MS m/z lists. Mascot relies on the probability of random matches between

theoretical and experimental m/z lists. Sequest and Mascot utilize the known specificity









of the protease, the parent ion mass, and the partial sequence information from

experimental MS/MS data for searching against the known genomic databases.

Furthermore, several common amino acid in vivo PTMs or in vitro chemical

modifications may be defined.

Table 1-6. Common MS and MS/MS Peptide Data Analysis Software
Method Name Web-Site
PMF Mascot http://matrixscience.com
MS-Digest http://prospector.ucsf.edu/ucsfhtml3.4msdigest.htm
MS-Fit http://prospector.ucsf.edu/ucsfhtml3.4msfit.htm
ProFound http://prowl.rockerfeller. edu/cgi-bin/ProFound
Pepldent http://www.exaspy.ch/tools/peptident/
PMT Mascot http://matrixscience.com
PepFrag http://prowl.rockerfeller.edu/prowl/
MOWSE http://hgmp/mrc.ac.uk/Bioinformatics/Webapp/mowse/
MS-Seq http://prospector.ucsf.edu/ucsfhtml3.4msseq.htm
MS-Tag http://prospector.ucsf.edu/ucsfhtml3.4mstagfd.htm
Multildent http://www.exaspy.ch/tools/multident/
PepMAPPER http://wolf.bi.unmist.ac.uk.mapper
Sequest http://fields.scripps.edu/sequest/
De Novo DeNovoX not available as shareware on the world wide web
SEQPEP not available as shareware on the world wide web
SeqMS http://www.protein.osaka-
u.ac.jp/rcsfp/profiling/SeqMS.html
Sherenga not available as shareware on the world wide web
PEAKS http://www.bioinformaticssolutions.com/products/inde
x.ph

Including the open access software to analyze proteomics data, different MS

vendors have created their own proteomics solutions software packages.60 The quality of

the data analysis software varies from instrument to instrument as does the quality of

instrument performance.31 Due to the diversity and size of the MS industry,

standardizing data files has been difficult. Currently, the 'dta' file format has provided

somewhat of a standard for the open access platforms. Most MS vendors provide ways to


extract data into 'dta' files so they can be analyzed with Mascot.









Bioinformatics data visualization and interpretation

Once experimental MS/MS data have been processed with the chosen algorithm,

protein identification candidates are generated with an associated score that represents the

likelihood that an experimental precursor-ion and partial MS/MS m/z list or spectrum

match the theoretical in silico digested peptide mass and peptide ladder. The higher the

score, the more likely the protein is a positive identification. Also generated with the

output are the numbers of peptides identified which pertain to a particular protein's

identification, also know as percent sequence coverage. Typically, the higher the percent

sequence coverage, the higher the likelihood that a particular protein may be considered a

real candidate.60 For the applications presented in Chapters 4 and 5, Sequest and Mascot

outputs are not used to identify a protein. The MS and MS/MS data in these cases are

used to identify amino acids which have been derivatized in vitro. Furthermore, the

sequence of the protein presented in this dissertation is already known.

The main differences between the programs presented in Table 1-6 are the way the

protein candidates are scored and the way the data is visualized. After bioinformatics

analysis of proteomic MS/MS data, it is the investigator's job to provide realistic

interpretation of what statistics has provided. After receiving an output from automated

analysis with bioinformatics software, the data requires manual MS/MS peak verification

to avoid false identification of proteolytic peptides. Furthermore, the instrument's mass

accuracy is very important for correct identification. Positive identification of peptide

candidates will have a mass error of less than 20 ppm, with an upper threshold for good

statistics around 200 ppm. For example, if the mass error thresholds are increased during

bioinformatic analysis of MS data, there will be an increased chance of false positives.60

Depending on the instrument, the mass tolerance of the instrument should be taken into









consideration. FT ICR-MS data should have very high mass accuracy as compared to a

QIT-MS data. Statistics is very useful for making a decision, but should never

overshadow careful thinking.

Summary

The tools available for analysis of proteins and oligonucleotides have been

presented. In the scope of this dissertation, the following chapters present data related to

the wet chemistry prior to and including MS analysis. First, Chapter 2 describes the

expression and purification of sialyltransferases needed for structural proteomic studies.

Second, Chapter 3 describes the use of hydrogen-deuterium exchange high-performance

liquid chromatography electrospray ionization Fourier transform ion cyclotron resonance

mass spectrometry analysis with sialyltransferase for studying secondary structure.

Third, Chapter 4 describes the use of small molecule labeling agents for derivatization of

sialyltransferase, followed by mass spectrometry analysis for identifying active-site

amino acids. Fourth, Chapter 5 describes the site-directed photoaffinity labeling of

sialyltransferase followed by mass spectrometry analysis for revealing the identity of

active-site amino acids. Fifth, Chapter 6 describes ion-pairing reverse-phase high-

performance liquid chromatography negative ion mode electrospray ionization Fourier

transform ion cyclotron resonance mass spectrometry for identification of

oligonucleotides selected with nanoharvesting agents. Finally, Chapter 7 describes the

conclusions and future work. The entire work is performed with an interdisciplinary

mind set towards structural proteomics as a small component of systems biology.














CHAPTER 2
EXPRESSION AND PURIFICATION OF RECOMBINANT HUMAN ALPHA 2- 3
SIALTRANSFERASE

Introduction to Oligosaccharides and Glycosyltransferases

Including proteins and oligonucleotides as presented in Chapter 1, oligosaccharides

compose another important class of bio-molecules. Compared to oligonucleotide and

protein linear structure, oligosaccharides may be covalently linked in a great diversity of

branched patterns. During the recent 'omics' nomenclature revolution, the glycome has

been defined as the whole set of carbohydrate molecules in an organism.64 Complex

oligosaccharides provide many functions ranging from energy storage to information

transfer systems that control cell function.65 Oligosaccharides may be attached to a

variety of different proteins and lipids through selective glycosidic bonds. Even when

including phosphorylation, glycosylation is the most common PTM found in living cells.

The complicated glycome has added an extra dimension to the already difficult

proteomics problem. To study Glycomics, investigators have focused on the structures of

the glycans themselves, the binding partners to the glycans, or the enzymes needed for

building and degrading glycan structure. This chapter presents background on the

glycosyltransferase family and the sialyltransferase sub-family, followed by the

description of recombinant human a (2-3) sialyltransferase expression, purification, and

characterization. Purified sialyltransferase will be bioconjugated prior to proteomic MS

analysis as described in Chapters 3 through 5.











H OH

OHo
HO
HO H
H
H OH
Mannose
(Man)


H

H OH 0
HO K
H H
CH3 H
OH OH
Fucose
(Fuc)


OH OH
HOH

H HO
HO OH
OH
H H


H OH



HO H
NHAc
H OH
N-Acetylglucosamine
(GlcNAc)


H

HAc H H.-O

HO CO2H

H OH


N-Acetylgalactoseamine Galactose Sialic Acid
(GalNAc) (Gal) (Sia)
Figure 2-1. Common Sugars Found in Oligosaccharides

Complex Carbohydrates in Biological Systems

Complex carbohydrates, also know as oligosaccharides or glycans, may decorate


different proteins or lipids to generate glycoproteins or glycolipids. The glycosidic


linkage to proteins may be O-linked through serine, threonine, or hydroxylysine or N-


linked to asparagines at specific amino acid sequons.52;66 O-linked glycans may be found


at Cys-X-X-Gly-Gly-Ser/Thr-Cys sequons with X representing any amino acid, however,


other as yet defined sequons may be possible. It has also been found that a proline +1 or


-3 from the Ser/Thr make glycan attachment favorable. N-linked glycan sites are well


defined and more abundant than O-linked glycans. N-linked glycans are attached to


asparagines contained in AsnXSer sequons where X can be any amino acid except


proline; however, an AsnXSer sequon does not necessarily mean that glycosylation will


be present. A third linkage for glycans includes attachment to Cys, Asp, Asn, Gly, and


Ser found on cell surface proteins inserted into lipid bilayers. In comparison to 0- and









N-linked glycans, the sequon specificity for attachment of glycans to proteins in lipid

bilayers has yet to be defined well.66

Glycomics attempts to define how proteins are affected by the attachment of

saccharides and what are the recognition events associated with saccharides and their

binding partners.67 Normally, glycans attached to proteins affect solubility, protease

resistance, and quaternary structure. Also, glycoproteins and glycolipids are the most

abundant species found on the outer walls of mammalian cells. Oligosaccharides are

important for mediation of cell growth, cell-cell adhesion, fertilization, immune defense

response, inflammation, viral replication, parasitic infection, masking of tumors, and

degradation of blood clots. 66;68 Further complexity of the glycome is revealed when

deciphering temporal states during an organisms life cycle i.e., embryonic

developmentt.6

MS analysis of oligonucleotide structures have been useful for identifying glycan

structure in respect to different cell and disease states in a similar manner to functional

proteomics.50;52;55;68 In the scope of this dissertation, the enzymes which build glycan

chains, glycosyltransferases (specifically sialyltransferases), are reviewed. The study of

the glycan structures themselves are not within the scope of this dissertation.

Glycosyltransferase Family

Glycosyltransferases are a family of enzymes that are membrane bound in the

Golgi apparatus and endoplasmic reticulum and synthesize glycans in an "assembly-line"

manner.66 Extensive studies on the N-and O-glycosylation pathways in mammalian cells

have revealed that glycans are synthesized by an ordered series of sugar transfer and

cleavage reactions.69 These enzymes are responsible for the transfer of different

monosaccharides from sugar-nucleotide donor substrates to form covalent linkages with









other simple and complex oligonucleotides.65 Interestingly, the 1970 Nobel Prize in

Chemistry was awarded to L. F. Leloir for his discovery of UDP-Glucose as a substrate

for glycogen biosynthesis. Almost all reactions which involve the synthesis of

oligosaccharide chains involve sugar-nucleotide donor substrates.70

Four main factors are used to classify and name the different glycosyltransferase

sub-families. The first includes the type of sugar transferred and the type of sugar-

nucleotide donor substrate used. The second includes the stereo- and regio-selectivity

exhibited by the acceptor substrate.65 The third involves the description of a retaining or

inverting mechanism with respect to the anomeric hydroxyl upon catalytic transfer.67

The fourth includes the dependence upon divalent metal cations for catalysis.66 In

humans, it was calculated that there are over 250 different glycosyltransferases, classified

into 47 different sub-families. Overall, from completed genomes, it is estimated that

there are over 7200 glycosyltransferases identified from bioinformatic studies.

Furthermore, of all the open reading frames deposited in genomic databases, about 1%

are dedicated to glycosyltransferase activity based on know primary structure motifs.71

Even though the catalytic transfer of monosaccharides by glycosyltransferases are

performed using similar donor and acceptor substrates, the sequence homology and

structure homology do not readily transfer between the sub-families. However, within

each sub-family, there are several structural motifs with high percent conservation.

Although there are many important glycosyltransferases needed for production of the

extremely complex glycans chains observed in nature, the focus of this dissertation now

turns to sialyltransferases. Sialyltransferases are of particular interest because they









transfer sialic acid, a charged sugar, which provides the "finishing touches" on large

oligosaccharides.

Sialyltransferase Sub-Family

Sialyltransferases catalyze the transfer of sialic acid from cytidine 5'-

monophosphate N-acetylneuraminic acid (CMP-NeuAc) to the terminal non-reducing

positions of oligosaccharides, glycoproteins, or glycolipids.72 NeuAc is an interesting

saccharide because it has a carboxylate, instead of a hydroxyl, attached to the anomeric

carbon. Interest in sialyltransferases arose due to the biochemistry of N-acetylneuraminic

acid (NeuAc) 'capped' carbohydrate moieties which bear a negative charge from the

carboxylate group attached to the anomeric carbon. Furthermore, NeuAc is the most

common naturally occurring sialic acid in human glycoconjugates.73 Since sialic acids

are attached to the termini of large oligosaccharides, they are implicated in important

physiological phenomena and disease states.73;74 Since sialyltransferases are the

enzymes involved with the transfer of sialic acid, there are large efforts to characterize

this family of enzymes.

As with all glycosyltransferases, the type of linkage formed is described by the

nomenclature of sialyltransferases. The linkages formed by sialyltransferases may be

through an c(2-3)- or an c(2-6)-bond to Gal (ST3Gal or ST6Gal); through an

c(2-6)-bond to GalNAc (ST6GalNAc); or through an a(2-8)-bond to Sia (ST8Sia).75

The c(2-3)- linkage is found to be most predominate than the a(2-6) linkage.76

Figure 2-2 represents the reaction catalyzed by c(2-3) sialyltransferases, where R is

Gal.































Figure 2-2. Reaction Catalyzed by Sialyltransferases

Sialyltransferases are type II transmembrane glycoproteins that have a short NH2-

terminal cytoplasmic tail, an amino acid signal anchor domain, and an extended stem

region with a large C-terminal catalytic domain (Figure 2-3).77 Sialyltransferases are

anchored in the trans cisternae of the Golgi apparatus and the trans Golgi network. 78

The stem region may be from 20 to 200 amino acids long and may mediate acceptor

specificity.79;80 As with all glycosyltransferase genes, sialyltransferase genes are

differentially expressed in a cell-type, tissue-type, and stage specific manner.72 In-vivo

proteolysis may occur at the stem region, thus making the enzyme soluble.79

Sialyltransferases have been found mainly in the deuterostome lineage such as

mammals, birds, amphibians, boney fish, and insects. Also, viruses produce similar

sialyltransferases to mammals for pathological infection.81 Furthermore,

sialyltransferases have been cloned from bacteria, however they lack any sequence










Catalytic Domain








Stem Region
Golgi Lumen



Cytoplasmic Cytoplasm
Tail NH2

Figure 2-3. Topology of Sialyltransferase

homology or identity to the deuterostome forms. Lastly, sialic acid has been detected in

plants and fungi, but characterization of sialyltransferases from these organisms have

lagged further behind than mammalian, virus, or bacteria.82 Funding for medicinal

chemistry based studies and completion of human, different virus, and different bacterial

genomes has made mammal, virus, and bacterial sialyltransferases characterization more

prevalent. Currently, over 22 unique sialyltransferase members have been cloned from
79
mammalian sources.7

According to the current Henrissat family classification of glycosyltransferases,

sialyltransferases are split into 3 evolutionary families. The first, termed GT 29, consists

of 94 sequences from mammals, plants, and viruses (ST3Gal, ST6Gal, ST6GalNAc,

ST8Sia). The second and third, termed GT42 and GT52, consist of 16 and 28 sequons

from different bacteria.7 Most insect cells have similar glycosylation pathways, and

until very recently, were believed to lack endogenous sialyltransferase activity.









Studies on Autographa california nuclear polehedosis virus with Spodoptera

frugiperda (fall armyworm) did not provide evidence for sialyltransferase activity.83

However, it was speculated that insects produce sialyltransferases during embryonic

development. In 2004, one sialyltransferase gene from the Drosophila (fruit fly) genome

was aligned against members of the vertebrate ST6Gal family.80;81 Cloning, expression,

and characterization of the ST6Gal from Drosophila melanogaster solidified that

sialyltransferases are expressed in insects at a certain stages of development.

Furthermore, the ST6Gal gene was aligned with a sialyltransferase gene found in the

mosquito genome (47% sequence identity).80

Many viruses have sialyltransferase genes which are similar to the mammalian

sialyltransferase genes and are used to invade a host cell. For example, the myxoma

virus, Leporipoxvirus Poxivirdae, uses a sialyltransferase gene as one of the enzymes to

enhance virulence towards rabbits. The sialyltransferase characterized in this study has

43% sequence identity and 60 % sequence similarity to vertebrate ST3Gal family

members.84 Another example consists of a sialyltransferase encoded by the Hepatitis B

virus which infects the liver to induce malignant tumors.85

Although bacterial sialyltransferases utilize the same sugar-nucleotide donor

substrate for catalysis, they show no sequence homology or identity to mammalian, plant,

virus, or insect sialyltransferases. However, sialyltransferases have been cloned and

characterized from the pathogenic Neisseria meningitides, Neisseria gonorhoeae, and the

marine bacteria Photobacterium damsela. In comparison to other taxons, bacterial

sialyltransferases however, they exhibit much broader acceptor specificities. It was

speculated that bacterial sialyltransferases evolved differently than invertebrate









sialyltransferase because they were not exposed to the same constrains.86;87 Furthermore,

the only crystal structure of a recombinant sialyltransferase, cloned form Campylobacter

jejuni, was published in early 2004.88 Since all sialyltransferases share the same sugar-

nucleotide donor substrates, some three-dimensional structure may be shared.

Biological and Medicinal Importance

Sialic acids are involved with numerous biological processes, such as the regulation

of glycoproteins in the blood stream, cell-cell interaction, cellular regulation, facilitation

or prevention of aggregation, immune system function, and nervous system development.

Of particular interest are the roles that sialic acid plays in the masking of tumors and the

inflammatory/immune response for pathogen-host recognition of bacteria or viruses73;80;84

The masking of tumors by sialic acid has been linked to brain, colon, breast, and

prostate cancers.69 Often, the link between incorrect glycan structure and malignant

tumors are not always understood. However, it is well accepted that the increase in

branching of glycan chains associated with certain cancers will facilitate an increase in

sialic acid attachment, thus an increase in the sialyltransferase activity is observed.

Furthermore, due to the complexity and the microheterogeneity of glycosylation, the

specifics of a certain glycan involved with cancer are often difficult to determine. Even

though there are many questions with the sialic acid/sialyltransferase cancer relationship,

four generalities are described.6875

First, the negative charge buildup from increased sialic acid residues may prevent

cell-cell interactions by repulsion events. Second, selections or siglecs (cell adhesion

molecules) bind to irregularly sialylated oligosaccharides and propagate malignant

transformation. Third, the complete abnormal oligosaccharide may be masked by sialic

acids, thus not allowing galectins to bind for repair. Fourth, specific cell signaling









pathways might be the target for malignant transformation with viruses. It may be

implied that by simply removing the sialic acids would address malignant repression;

however, this is not always the case.68;74;75

Immune system response often uses oligosaccharides to discriminate between

friendly and enemy structures. The failure to determine between self and non-self

structures may lead to autoimmune reactions.74 For example, the binding of terminal

sialic acid residues are the initial step for influenza virus infection. Hemagglutinin, a

surface protein on the influenza virus, mediates attachment and fusion of the cell with the

virus membrane. Obviously, there have been several studies to inhibit and characterize

this sort of binding by modifying CMP-NeuAc at the C5 and C9 positions of sialic acids.
89-91 Furthermore, studies on Hepatitis B virus show that insertional mutagenesis of

sialyltransferase and other enzymes cause proliferation of human liver cancer.85 To

facilitate rational drug design for inhibition of disease, several investigators have studied

sialyltransferase primary structure.

Sialyltransferase Structure and Mechanism Relevant for Structural Proteomics

Vertebrate sialyltransferases share the same topographical features; however they

show very little sequence identity except 4 consensus motifs termed L, S,VS, motifs and

motif 3. Figure 2-4 represents the conserved amino acids between the different types of

human sialyltransferase, where the sialylmotifs are labeled with the amount of amino

acids and the numbers in parentheses correspond to the number of amino acids between

the motifs or the N- and C-termini.81 Interestingly, the identification of structural

sialylmotifs was aided with peptide mass fingerprinting in 1992 with liquid secondary ion

and double focusing MS. Primers were designed based on the common sequence of









membrane bound sialyltransferase and known sialylmotifs to clone new

sialyltransferases.92

20 Different Sialyltransferases May Contain to 302 600 Amino Acids

(9-17 AAs) (8-21 AAs)
(73-293 AAs) 42 AAs (89 -109 AAs) 24 AA 4 AAs 6 AAs (22-30 AAs)

L motif S motif motif 3 VS motif
Figure 2-4. Sialylmotifs and Amino Acid Length of Sialyltransferases

The significance of these 4 sialylmotifs has been investigated using ST6Gal I and

ST3Gal I as model enzymes.72;79 First, mutagenesis studies on the L-motif affected

binding to the donor substrate.78 Second, mutagenesis of the S-motif altered the binding

properties of both the donor and acceptor subtrates93 Third, mutagenesis of motif 3 was

shown to be important for functional and/or structural roles.79 Fourth, mutagenesis of

VS-motif revealed the complete loss of activity. In these studies, amino acids were

mutated which may be related to catalytic amino acids on the acceptor side of a binding

cleft.79;80;94 Fifth, mutagenesis studies on cysteines have revealed a disulfide bond which

may bring the L and S-motifs within close proximity of each other.95 Finally,

glycosyltransferases normally have a divalent metal dependence and a common DXD

motif. Sialyltransferase do not show a metal dependence and lack the DXD motif.94

Using rat liver ST6Gal as a model, sialyltransferases were shown to catalyze a

highly dissociative transition state with substantial oxocarbenium ion character96-10.

These kinetic studies also revealed that the mechanism was steady-state random and that

the enzyme had a bell-shape pH versus rate profile.97 Interestingly, kinetic studies on

sialyltransferase from the bacteria C. jejuni 8and the insect D. melanogaster 80 both show

similar bell-shape pH versus rate profiles to the mammalian sialyltransferases.94;97 Also,

recent data presented by Dr. Erin Burke suggest that protonation of the non-bridging









oxygen from a general acid aids in CMP bond breaking. Therefore, a general base and a

general acid are the targets for the structural proteomic studies presented in Chapter 4.

In February 2004, the first crystal structure of a bacterial sialyltransferase was

published.88'101 In this report, it was suggested that a histidine acts as a general base to

assist in the nucleophilic attack of a hydroxyl and a tyrosine acts as a general acid to

assist the departure of CMP. The loss of activity during the previously described

mutagenesis study on mammalian sialyltransferases on the tyrosine residue (motif 3),

taken in light of the bacterial crystal structure, may indicated that this amino acid is

involved with active-site catalysis. However, since this sialyltransferase from bacteria

does not have sequence homology or identity to mammalian sialyltransferases, it cannot

be ruled out that another amino acid other than tyrosine acts as the general acid needed

for catalysis in the mammalian sialyltransferases. Another interesting feature of the

bacterial sialyltransferase is that a loop of 12 amino acids becomes ordered upon binding

to inhibitor.88 This 'lid' may shield the oxocarbenium transition state reaction from

nucleophilic solvent attack.

The availability of recombinant sialyltransferase will facilitate our MS based

structural proteomics study to answer the question as to the identity and location of the

general acid and/or general base.88 The absolute identification of the general acid and/or

general base in tandem with molecular mechanics computational chemistry will provide

the basis for intelligent design of selective sialyltransferase inhibitors.102-105 Furthermore,

once a labeling strategy has been developed for our model system, recombinant human

placental ST3Gal IV, it may be applied to all sialyltransferases with possible insight into

other motifs or important structural properties.









Recombinant Human a (2- 3) Sialyltransferase IV

Traditionally, mammalian sialyltransferases are difficult to express and purify in

large quantities due in part to their membrane bound protein properties and that they

themselves are glycosylated. Because these enzymes are complex, eukaryotic cell lines

are used for reliable expression of recombinant mammalian sialyltransferases. Different

cell lines such as yeast, insect, and mammalian origins were used to produce soluble,

catalytically active mammalian sialyltransferases. Attempts to produce catalytically

active mammalian sialyltransferases in E. coli in our lab and others have not resulted in

active enzyme.106 Several glycosyltransferases have been successfully cloned and

overexpressed in active form with yeast such as Saccharomyces cerevisiae (full-

length)107'108 and Pichiapastoris (secreted).109 Mammalian cell lines have endogenous

sialyltransferase activity and have been useful for expression of glycosyltransferases and

sialyltransferases108;110

Since insect cell lines such as Spodopterafrugiperda (Sf9) lack endogenous

sialyltransferase activity, they are used to express soluble glycosyltransferases and

sialyltransferases with baculovirus vectors. 79;111-113 Through the beauty of modern

molecular biology, soluble sialyltransferase was produced in our lab with a cleavable

insulin signal peptide cloned upstream from the catalytic domain, replacing the N-

terminal anchor and stem region, thus facilitating expressed construct secretion into the

media.92;114-116 Also, to improve purification yields, fusion tails were placed between the

signal peptide and the catalytic domain.79;108;111-113;117 Fusion tag/affinity purification

pairs may include the mouse IgM signal peptide and the IgG binding domain of the

protein A, the streptavidin fusion tag with avidin binding, GST with anti-GST antibodies,

or polyhistidine tags (His6xTags)with nickel agrose beads (Ni+2-NTA).118 Of these fusion










tags, a polyhistidine tag was the first choice because of the ease of incorporation and the

minimal amount of enzyme structure perturbation.

We produced three recombinant forms of human placental a-(2-3)

sialyltransferase (hST3Gal IV; EC# 2.4.99.4) 110;119 which lacked the first 40 amino

acids coding for the NH2-signal anchor and stem region. The three recombinant forms

were overexpressed in Spodoptrafrugiperda (Sf9) insect cells using a baculovirus vector

with Invitrogen's (Carlsbad, CA) Bac-to-Bac Baculovirus Expression System.120 The

first recombinant form consists of a canine insulin cleavable signal peptide cloned in

front of the catalytic domain of hST3Gal IV (Ins-ST). The second recombinant form

consists of the canine insulin signal peptide followed by a His6x-tag and two glycines in

front of the catalytic domain of hST3Gal IV (N-Tag-ST). The third recombinant form

consists of the canine insulin signal peptide followed by the catalytic domain of hST3Gal

IV with two glycines and a His6x-tag placed at the C-terminus (C-Tag-ST). The inclusion

of the His6x-tag will aid in ease of purification. The aim of this experiment was to

produce

-A 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain (Ins-ST)

Secretion Peptide C-Terminal Catalytic Domain


-A 40 AAs + Canine Insulin Secretion Peptide + Polyhistidine Fusion Tag + C-Terminal Catalytic Domain (N-Tag-ST)

Secretion Peptide HHHHHHGG C-Terminal Catalytic Domain


-A 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain + Polyhistidine Fusion Tag (C-Tag-ST)

Secretion Peptide C-Terminal Catalytic Domain GGHHHHHH

Figure 2-5. Sialyltransferase Constructs Prepared









relatively large amounts of sialyltransferase at a low price for multiple studies on

sialyltransferase secondary structure. The assumption was made that cleaving the stem

region will not perturbate activity.79

Results and Discussion

Baculovirus preparation

Dr. Nicole Horenstein originally prepared the pFastBacHTa vector that harbored

the cDNA encoding the Ins-ST gene lacking the first 40 amino acids. The two other

plasmid constructs containing N- and C-terminal His6xtag, in addition to the insulin signal

peptide, were prepared by Bronson Anatao. Correct construction of cDNA clones with

canine insulin secretion peptide and His6x-tags were confirmed by DNA sequence

analysis performed at the University of Florida Protein Core Facility.

All steps prior to affinity chromatography followed the Bac-to-Bac Baculovirus

Expression System manual from Invitrogen (www.invitrogen.com) with no deviation

from the described protocol. The three pFastBacHTa plasmids containing the particular

genes were separately transformed into competent DH1OBacTM E. coli cells. PCR

analysis (Figure 2-6) of mini-prepped recombinant backmid indicated an insertion

product of proper length.

Spodopterafrugiperda (Sf9) insect cells were transfected in 6 well plates with

recombinant bacmid DNA with the aid of Cellfectin Reagent. After transfection, the

viral titers were below the level needed for sialyltransferase expression. To produce

higher viral titers (amplification), 50 to 75 mL of cell cultures containing 2-3 x 106

cells/mL were infected with 1 to 4 mL of viral stocks. After 48 to 55 hours post

infection, the cell supernatant was clarified and sterile filtered. Subsequent amplification

proceeded until a viral titer of about lxl08 plaque forming units/mL (pfu/mL) was









achieved. Table 2-1 represents the viral titers used for infection of insect cells for

expression of sialyltransferase.

1 2 3 4 5


9461 bp

6557 bp :
4361 bp




2322 bp
2027 bp


1353 bp
1078 bp

Figure 2-6. Confirmation of Correct DNA Insertion Product where: Lanes 1 and 5 are
Base Pair Markers, and Lanes 2 through 4 Correspond to Different Colonies
Selected Based on Blue/White Screening after Liquid Culture and Mini-prep
of the Recombinant Bacmid.

Table 2-1. Viral Titers for Different Constructs
Construct Titer in pfu
Ins-ST 1.7x 108
N-Tag-ST 2.1 x 108
C-Tag-ST 2.7 x 108

Expression and purification of the three sialyltransferase constructs

Purification of the Ins-ST construct presented several different technical issues.

First, a sepharose cytidine diphosphate-hexanolamine affinity (sCDP-Hex) material was

synthesized by Dr. Erin Burke and Dr. Nicole Horenstein according to the literature.64

Second, the support phase of the sCDP-Hex resin was easily compressed, thus not

allowing for large pressure gradients. If large volumes of buffer were added to the sCDP-

Hex column, flow reduced dramatically. Third, the addition of 300 mM p-lactose to the









initial supernatant before the loading of clarified supernatant (harvested cells 72 hours

post infection) resulted in higher purity. Although the purity improved, the flow rate was

reduced due to the increased viscosity. A contaminant protein, glycoprotein 64 (GP64),

was expressed along with sialyltransferase in the baculovirus system. Using trypsin in-

gel digestion of a band excised from SDS-PAGE analysis of an up-pure sample, GP 64

was positively identified as the contaminate protein with MS analysis. GP 64 was

believed to bind to the N-glycans of sialyltransferase, thus creating the need for the 3-

lactose wash after loading clarified supernatant. Finally, because of the low flow rates

associated with the sCDP-Hex column, the initial cell culture supernatant needed to be

concentrated 5 fold.

In an attempt to create constant and increased flow rate through the column, a post-

column pump was added in-line, however, this compressed the resin, thus producing no

flow. Several attempts were made to purify the Ins-ST construct, often with no retention

of enzyme on the column; however, some purifications led to positive results. Table 2-4

summarizes the purification of the Ins-ST construct. The definition of a unit of activity

was the number of [tmols of CMP-NeuAc converted to sialyl-lactose per minute.

Presented in the Table 2-4, cell culture supernatant corresponds to pooled supernatant 72

hours post infection of 2 to 2.5 x 106 cells/mL with 3 mL of Ins-ST recombinant

baculovirus. Ultrafiltration corresponds to the concentration of clarified supernatant with

an Amicon Ultrafiltration device with a polyethersulfone membrane with a 10 kDa

molecular weight cutoff (MWCO). Load buffer addition corresponds to the addition of

MES, pH 6.8 to 50 mM, p-lactose to 300mM, glycerol to 20% v/v, and triton CF 54 to

0.05% v/v to the concentrated supernatant. Pooled fractions correspond to fractions







53


eluted with high specific activity. Concentration corresponds to ultrafiltration of pooled

fractions with an Amicon Ultrafiltration device with a polyethersulfone membrane with a

10 kDa MWCO.

Table 2-2. Summary of an Ins-ST Purification
Volume Activity Yield SA
Ins ST Purification mL mU % U U /L U/mg
Cell Culture Supernatant 370 820 100 2.20 0.029
Ultrafiltration 80 760 93 9.60 0.022
Load Buffer Addition 100 710 87 7.10 0.025
Pooled Fractions 28 230 28 8.12 0.161
Concentration 5 160 20 31.9 0.132

Figure 2-7 represents the elution profile for a steep step salt gradient applied to the

column after loading the concentrate and washing with the p-lactose buffer. The elution

of Ins-ST from sCDP-Hex was performed by increasing the salt concentration in a step

gradient fashion. For the experiment presented in Figure 2-7, activity assays were not

initially available and fractions corresponding to either 250 mM or 350 mM NaCl were

pooled separately. Later, activity assays reveled that only the 350 mM NaCl pooled

fractions contained sialyltransferase activity and SDS-PAGE analysis revealed

sialyltransferase with large amounts of contaminating proteins. This data suggested the

need for more than 5 column equivalents per step for effective purification.


500 5
S450 --- mMNaC1 -4.5
S400 o mg/mLofProtein 4
350 3.5
Z 300 3 3
250 2.5
200 2
150 1.5
S100o 1
50 0.5

10 12 14 16 18 20 22 24 26 28 30 32 34 36

Fraction Number

Figure 2-7. Protein Concentration Profile of an Ins-ST Purification










Figure 2-8 represents the elution profile for Ins-ST with a shallow step gradient and

sialyltransferase activity for fractions corresponding to 350 mM and 450 mM NaC1.

Activity was determined as described in the Methods and Materials section To save

radiolabeled substrates used to determine activity, full activity profiles were not

collected. On average, 30 to 35 % of the initial activity was not retained on the column.

Attempts were made to collect sialyltransferase from the flow through by reloading it

onto the column; however, sialyltransferase activity was not retained. Also, fractions

eluted with 100mM and 250 mM NaCl did not contain activity. Furthermore, if 20 %

glycerol and 0.01 % triton CF 54 were not present, activity was completely lost during

purification. Finally, after collection and concentration of pooled fractions, these

constructs were sensitive to dilution, dialysis, and further ultrafiltration, with loss of 90%

of the initial activity.



400 45.0
350 40.0
300 35.0
30.0
250 -
25.0
200 -
20 20.0
150 15.0
S100 10.0
50 5.0
0 -. 0.0
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72
Fraction

Figure 2-8. Protein Concentration and Activity Profile of an Ins-ST Purification

SDS-PAGE analysis of purified sialyltransferase revealed a step ladder of possible

glycoforms (Figure 2-9, Lane 2). According to the known consensus sequence for









glycosylation, hST3Gal IV has four possible sites for glycosylation. To confirm that the

multiple bands are due to glycosylation, Endo Hf, a recombinant protein fusion of

Endoglycosidase H and maltose binding protein, was used to deglycosylate Ins-ST.

Figure 2-9, Lane 3 corresponds to the deglycosylation product of sialyltransferase. The

observed molecular weight (-34,000 Da), after the deglycosylation reaction, corresponds

to the theoretical molecular weight of the primary sequence (-34,005 Da) free of glycans.

1 2 3




66 kDa

45 kDa
36 kDa
29 kDa
26 kDa

20 kDa

14 kDa
Figure 2-9. SDS-PAGE Analysis of Purified Ins-ST and Deglycosylation Reaction
where: Lane 1 = Molecular Weight Markers; Lane 2 = Purified Ins-ST; and
Lane 3 = Ins-ST Deglycosylated with EndoHf

To possibly improve overall yields and ease of purification, Ni 2+-NTA with

His6xTag technology provided somewhat different results. The Ni2+-NTA resin does not

compress easily, thus increasing the flow rate dramatically in comparison to the CDP-

Hexanolamine-Agrose resin used for the purification of Ins-ST. This eliminated the need

for supernatant concentration. Elution of N- and C-Tag-ST was performed by increasing

the imidazole concentration in a step gradient fashion (5 mM, 50 mM, 135 mM imidizole

in a buffer containing 20 % glycerol, 0.05 % triton CF 54, 100 mM KC1, 50 mM MES,

Ph 6.8). The purification ofN- and C-Tag-ST took 11 hours (harvest to concentrated










pooled fractions), thus providing another positive quality of Ni2+-agrose over the sCDP-

Hex affinity column. Tables 2-5 and 2-6 summarize the purification of the N-Tag-ST

and C-Tag-ST constructs respectively.

Table 2-3. Summary of a N-Tag-ST Purification
Volume Activity Yield SA
N-Tag-ST Purification mL mU % U U/L U/mg
Cell Culture Supernatant 600 570 100 0.95 0.0009
Load Buffer Addition 660 540 95 0.82 0.0003
Ni-NTA Pooled Fractions 50 190 33 3.80 0.045
Concentration 4.5 155 27 34.5 0.216

Table 2-4. Summary of C-Tag-ST Purification
Volume Activity Yield SA
C-Tag-ST Purification mL mU % U mU / L U/mg
Cell Culture Supernatant 470 560 100 1.20 0.008
Load Buffer Addition 580 420 75 0.72 0.007
Ni-NTA Pooled Fractions 50 180 32 3.60 0.120
Concentration 5.5 240 43 43.6 0.162

Presented in the Tables 2-3 and 2-4, cell culture supernatant corresponds to

collected supernatant 72 hours post infection of 2.5 x 106 cells/mL with 3 mL of N-Tag-

ST or C-Tag-ST recombinant baculovirus stocks. Load buffer addition corresponds to

activity after bringing the supernatant concentration to 20 % glycerol, 0.01 % triton CF

54, and 50mM MES, pH 6.8. Without the addition of both 20% glycerol 0.01% triton CF

54, the resulting yields were less than 1%. Ni2+-NTA pooled fractions correspond to

fractions eluted with high specific activity. Concentration corresponds to ultrafiltration

of pooled fractions with an Amicon Ultrafiltration device with a polyethersulfone

membrane with 10 kDa MWCO. These constructs were sensitive to dilution, dialysis,

and further ultrafiltration, with loss of 90% of the initial activity.

Compared to sCDP-Hex resin, the Ni2+-NTA column retained all of the activity.

Furthermore, to increase specific activity, an imidazole step gradient was used to elute

proteins which may non-specifically absorb to the column. For example, fractions eluted










with the 50 mM imidazole buffer appeared to be yellowish, but did not contain activity.

Figure 2-10 illustrates the typical elution profile for either His6x-tag constructs. The His6x

tag enzymes eluted at 135 mM imidazole over 10 fractions (5 mL) with constant specific

activity. Figure 2-11 represents the SDS-PAGE of both constructs.


Increasing Imidazole Concentration


800

700

S600

. 500

400

S300

S200

100

0


* Activity in mU
0 |tg protein / mL


16

14

12

10

8

6 C

4

2

0


0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96
Fraction

Figure 2-10. Protein Concentration and Activity Profile of a C-tag-ST Purification

1 2 3




66 kDa


45 kDa
36 kDa

29 kDa
26 kDa

20 kDa

14 kDa

Figure 2-11. SDS-PAGE Analysis of TCA Precipitated N-Tag-ST and C-Tag-ST where:
Lane 1 = C-Tag-ST; Lane 2 = Molecular Weight Markers; and Lane 3 = N-
Tag-ST.









Overt the last 25 years several laboratories have cloned and purified different

sialyltransferases. Table 2-5 corresponds to the trends on sialyltransferase cloning and

purification as reported in the literature.

Table 2-5. Summary of Selected Sialyltransferase Cloning and Purification Papers
Enzyme and Source Final Other Notes and
Reference Product (Acceptor Substrate)
1) ST6Gal Extracted from 2.77 U Started with 20 L of cell
(1979) porcine 2.4 % Yield extract (asialo-ovine
porcine submaxillary 62 ug submaxillary mucin)
submaxillary glands in intact
glands121 form
2) ST3Gal IV Extracted from 1.0 [tmol/min (U) Started with 17 L extract.
(1982) rat liver in 7 % Yield (asialo-ca -acid glycoprotein)
rat liver122 intact form 36 ug Total ST
3) ST6Gal I Extracted from 27 [tmol/min (U) Started with 17 L extract
(1982) rat liver in 16 % Yield (lacto-N-tetraose)
rat liver122 intact form 3.3 mg
4) ST3Gal IV Extracted from 0.140 [tmol/min (U) Extracted from 2 kg of
(1985)119 human placenta 10 % Yield human placenta (Galpl-
human in intact form 90 ug Total ST 4GlcNAc)
placenta (original clone
for our studies)
5. ST3Gal IV Expressed in Did not fully Main purpose was to identify
(1992) COS-1 cells characterize and clone new ST (Galpl-
porcine with secretion 5mU/mL after 3GlcNAc)
liver116 peptide harvest and
concentration
6. ST3Gal V Extracted from Did not characterize Characterization of acceptor
(1993) human placenta substrate and detergents
human intact form (Gal3 1-4GlcNAc3 1-3 Gal)
placenta123
7. ST6Gal I Expressed in Only reported U/L Reported specific activity as
(1995) Saccharomyces expression over resonstituted yeast
human Cerevisiae time lypophilisate
placenta107 Yeast in intact 0.31 U / L in 150 L 0.8mU / mg protein
form bioreactor (asialofetuin or asialo-ovine
submaxillary mucin or
LacNAc)









Table 2-5. Continued
Enzyme and Source Final Other Notes and
Reference Product (Acceptor Substrate)
8. ST3Gal III Expressed in Reported as Immobilized on
(2003) Sf9 insect cells immobilization nickel column
rat liver113 with insulin yield: 59% for synthesis of
secretion sialylated oligosaccharide
peptide and (3-D-Galp(1-3)-3-D-
His6x tag GlcNAcp-OR)
9. ST3Gal III Expressed in Did not fully Main purpose was to show
(2003) Sf9 insect cells characterize, that fusion of IgG protein A
human"1 with IgM Reported Specific can be used with affinity
secretion activity to be 1.6 separation (asialofetuin)
peptide and IgG U/mg.
binding domain 6.4 mg / mL
of IgG
10. ST3Gal I Expressed in Only reported Did not report all purification
(2004) Sf9 insect cells specific activity parameters (Galp1-
human79 with insulin 400 nU / mg 3GalNAca-sp-biotin)
secretion
peptide and
His6x tag

Originally, sialyltransferases were purified from selected organs (Table 1-5,

numbers 1-4, and 6) with difficult procedures needed for the extraction of activity. With

modern day molecular biology, the yields associated with a 1 L expression are

comparable with early purification methods. Compared to other protein families,

membrane bound proteins have always been difficult to purify in high yield because of

the extraction protocols used and their hydrophobic nature.

For the rest of the purifications described in Table 2-5 (numbers 5, and 7-10), the

full characterization of the purification was not as important because the clone of interest

was used for mutagenesis studies, kinetic studies, specificity studies, or synthesis of

sialylated oligosaccharides. For these types of publications, authors present relative

activity differences and partial kinetic parameters to validate the importance of their

application. Furthermore, sialyltransferases have specific acceptor substrates; however, a









glycan found on a glycoprotein versus a glycan not bound to a glycoprotein will exhibit a

different Km. There are other papers published on sialyltransferase cloning and

purification which not presented; however, they follow the same trends observed in Table

2-5.

Compared to the literature values presented in Table 2-5, our purification tables

reflect similar values for expression and final yields. Our specific activities may be lower

than expected because a non-standard acceptor substrate (a-lactose) was used with the

activity assays. Furthermore, it appears that other investigators had problems

concentrating the enzyme. This observation is based on the lack of full characterization.

Sialyltransferases are difficult enzymes to clone and purify, but modern biochemistry and

MS analysis are geared for analysis of low concentrations of enzymes. The fact that is it

difficult to purify large quantities of sialyltransferases (and most glycosyltransferases)

has translated into a lack of crystal structures and the lack of complete characterizations.

Concentration/dilution studies on N-Tag-ST

To investigate the stability of recombinant hST3 Gal IV, the N-Tag-ST construct

was used as a model. The main purpose of these experiments was to determine if the N-

Tag-ST could be desalted with a minimal loss of activity. Furthermore, for the labeling

experiments presented in Chapters 4 and 5, it would be beneficial to remove excess

labeling agents. Unfortunately, dialysis resulted in recovery of 10 % of the initial

activity. The next strategy was to use different micro-ultrafiltration devices after dilution

with an exchange buffer (20% glycerol, 0.05 % triton X-100, 100 mM KC1, and 50 mM

MES, pH 6.8). Dilution of recombinant hST3Gal IV by a factor of 10 with one addition

of exchange (dilution) buffer, followed by ultrafiltration with a 50 mL capacity Amicon










ultrafiltration stir cell with an amicon polyethersulfone membrane with a 10 kDa

MWCO, led to the loss of 95 % of the initial activity. This result was surprising because

it was the same filtration unit and membrane used for the concentration of the pooled

fractions after purification. The use of a Millipore micron centrifugal filter device

(MICRON YM-10; Reg. Cellulose; 10,000 MWCO) for concentration resulted in the loss

of 99% of the initial activity. For the best results, concentration of N-Tag-ST was

performed with a Millipore ultrafree-MC centrifugal filter unit (Durapore PVDF

membrane; 10,000 MWCO) with 400 ptL capacity.

Figure 2-12 represents the 1:1 dilution of purified N-Tag-ST with different

concentrations of triton X-100 or triton CF 54 present in the dilution buffer described

earlier. The purified enzyme was dilute in half with either triton X-100 (red bar graphs)

or triton CF 54 (blue bar graphs). The percent original activity was normalized based on

the dilution factor. The data revealed that dilution to 0.025 % triton X-100 retains higher

percent original activity than samples diluted with triton CF 54 at any concentration.



) 100%oo
80oo
600%
S40%

O 20%
0%

Oo "Ooso
Sample to ito o


Figure 2-12. Effect on Activity of N-Tag-ST with Dilution of Different Detergent
Concentrations









Figure 2-13 represents the ultrafiltration of the N-Tag-ST with Millipore ultrafree-

MC centrifugal filter units (Durapore PVDF membrane; 10,000 MWCO) after dilution

with 20% glycerol, 0.01 %, triton X-100, 100 mM KC1, 50 mM MES, pH 6.8. On the x-

axis, the first value corresponds to the volume of purified N-Tag-ST added to the device,

with the second value corresponding to the number of 200 pIL dilutions with the buffer

described above. For example, 400 x 0 corresponds to 400 pIL of purified N-tag-ST

concentrated to 100 ILL; 200 x 3 corresponds to 200 [IL of purified N-tag-ST diluted to

400 IpL with dilution buffer, then concentrated to about 200 ILL, followed by the addition

of 200 IpL of dilution buffer, then concentrated to 200 pIL again, followed by the addition

of 200 IpL of dilution buffer again, then concentrated to 100 [LL. The dilution of N-Tag-

ST was performed in the device in between centrifugation cycles. Percent activity was

normalized based on the activity prior to concentration and the final volume (-10Q[L). It

appears that the N-Tag-St construct is not stable under the conditions described for

removal of high salt concentrations.


70%
.4 60%
S50%
S40%
30%
S20%
10%
0%

o +/ 0 +/ -

Sample
Figure 2-13. Effect on N-Tag-ST Activity with Different Dilution Factors and
Concentration with Millipore Ultrafree-MC Centrifugal Filter Units









To ensure that imidazole did not affect the activity assay, a blank consisting of 135

mM imidazole did not produce activity above background. It appears that there was not a

useful protocol for removal of imidazole salt. For the labeling studies presented in

Chapters 4 and 5, the constructs were derivatized after purification with no further

desalting.

Kinetic parameters

Kinetic parameters were determined in collaboration with Dr. Erin Burke. Table 2-

6 presents the experimental kinetic parameters for our recombinant hST3Gal IV

constructs and the wild-type human placental ST3Gal IV.119 In comparison to our

expressed hST3Gal IV constructs, the WT ST3Gal IV was purified with the N-terminal

domain and the stem region intact. The kinetic data suggests that the inclusion of these

domains with the WT ST3Gal IV effectively lowers the affinity for a-lactose. The

second important data point presented in Table 2-6 is the marginal increase of the C-tag-

ST Km for CMP-NeuAc. It is suggested that a His6xtag added to the N-terminal end of

the catalytic domain will not change the structure of the protein, whereas a His6xtag

added to the C-terminal may change the structure of the enzyme. Since the WT, Ins, and

N-Tag ST have similar Km values for CMP-NeuAc, and the C-Tag-ST has an increased

Km value, it is suggested that the C-terminal region is important for binding of CMP-

NeuAc.

Table 2-6. Summary of Kinetic Parameters for Different Sialyltransferase Constructs
CMP NeuAc a-lactose
Recombinant Enzyme Km ( .M) Km (mM)
WT ST3Gal IV119 63 220 + 40
Ins ST 82 +5 171 +18
N Tag ST 74 + 8 155 + 14
C Tag ST 267 + 20 158+ 11









Conclusion

The addition of a His6xtag to the N-terminal or C-terminal end of the catalytic

domain of ST3Gal IV (N-Tag-ST and C-Tag-ST) provided constructs that were much

easier to purify than Ins-ST. All three sialyltransferase constructs had similar Km values

for donor and acceptor substrates in comparison to the wild-type h23 STGal IV, except

the C-Tag-ST Km for CMP-NeuAc. Normally, the N-terminal end of the catalytic

domain is extended through a stem region to an anchor which ligates the protein to the

golgi apparatus; therefore, changes in the in the N-terminal region of the catalytic domain

of ST3Gal IV would not necessarily perturbate structure. Since the C-Tag-ST Km is

larger for the donor substrate, it is proposed that the His6xtag slightly alters the secondary

structure. Lastly, since Ins-ST, C-Tag-ST, and N-Tag-ST were not stable to desalting

procedures, the construct will be derivatized with bioconjugation techniques in the

purification buffers with no further sample clean up.

Methods and Materials

All restriction enzymes, E. coli strains JM109 & ER2925, DNA Polymerase I

(Klenow, Large Fragment), and T4 DNA ligase were purchased from New England

BioLabs (Beverly, MA). Wizard Plus Minipreps Kit and dNTPs were purchased from

Promega (Madison, WI). Shrimp alkaline phosphatase (SAP) was purchased from Roche

(Indianapolis, IN). QIAquick Nucleotide Removal Kit and QIAquick Gel Extraction

Kit were purchased from Qiagen (Valencia, CA). The BCA Protein Assay Kit was

purchased from Pierce (Rockford, IL). The BAC-TO-BAC Baculovirus Expression

System (BBES), Spodopterafrugiperda (Sf9) insect cells, BACPACKTM Baculovirus

Rapid Titer Kit, and DH10 BAC E. coli competent cells were purchased from Invitrogen

(Carlsbad, CA). Primers for cloning and PCR analysis were obtained from Integrated









DNA Technologies (Coralville, IA). The protocol for recombinant virus preparation was

found in Invitrogen's instruction manual for BBES version D April6, 2004. The

sepharose CDP-hexanolamine affinity column (sCDP-Hex) was prepared as per the

literature.121;124 The [6-3H] N-acetyl D-mannosamine isotopomer used in the synthesis of

[9-3H] NeuAc was purchased from Moravek. N-acetyl neuraminic acid (NANA) aldolase

[EC 4.1.3.3] used with the synthesis of [9-3H] neuraminic acid (NeuAc) was cloned,

overexpressed in E. coli and purified according to literature procedures.125;126 The E. coli

expression plasmid pWV200B harboring the E. coli CMP-NeuAc synthetase gene [EC

2.7.7.43] used for the synthesis of all CMP-NeuAc substrates was a generous gift from

Dr. W. F. Vann at the National Institutes of Health. Radioactive samples from

sialyltransferase activity assays were analyzed with a Packard 1600 TR liquid

scintillation analyzer. Trypsin for in-gel digestion was purchased from Promega and

Sigma-Aldrich (St. Louis, MO). All other chemicals for purification were obtained from

Sigma-Aldrich in the highest-purity.

Baculovirus vector preparation for Ins-ST

The original expression construct derived from pFastBackHTa contained an N-

Terminal polyhistidine tag, but lacked a secretion signal (N. Horenstein, unpublished

work). The sequence coding for the His-tag was removed from the construct with RsrII

and BamHI. A 114 bp insert containing the sequence coding for the canine insulin leader

peptide with flanking RsrII and BamHI sites was prepared in the following way. Two

complementary single-stranded 70 bp oligonucleotide primers (Forward Ins-ST: 5'-

CGCGCGGTCCGAAATGATGGCCCTCTGGATGCGCCTCCTGCCCCTGCTGGCCC

TGCTGGCCCTCTGGGCG-3' and Reverse Ins-ST: 5'-

GCGGATCCGCCCCGGGAATCAACGAAGGCTCGGGTGGGCGCGGGCGCCCAG









AGGGCCAGCAGGGCCAGCA-3') spanning the insulin leader peptide were obtained

commercially. The underlined bases denote the location of the RsrlI and BamHI

restriction sites for the Forward Ins-ST and Reverse Ins-ST primers, respectively. The

oligonucleotides were annealed and then filled in with Klenow fragment to produce the

114 bp insert which was doubly-digested with RsrlI and BamHI. The insert was then

ligated into the RsrlI and BamHI restriction sites of the expression vector using T4 DNA

ligase. The new construct (Ins-ST) was sequenced at the University of Florida ICBR

Protein Core Facility to confirm insertion of the Ins-ST gene. Isolation ofbacmids and

generation of baculovirus followed the BBES protocols.

Baculovirus vector preparation for N-Tag-ST

The Ins-ST plasmid was digested for 12 hours at 37 C with BamHI. Clean-up of

the enzymatic digestion was performed using the QIAquick Nucleotide Removal Kit.

The plasmid was digested with Stul (6h at 37 C), thermally inactivated (20 min at 65

C), agarose gel purified, and extracted using the QIAquick Gel Extraction Kit. The 5'

ends generated by digestion were dephosphorylated using SAP. The entire

Nterm6XHisTag insert was created by allowing two complimentary primers to anneal

and was then extended with Klenow fragment. Primers were mixed in equal volumes

(2.5 pL) of NtermHISForUpper (5'-

GGTAGGCCCTGGCCATTAAGCGGATGCTGGAGATGGGAGCTATCAAGAACCT

CACGTCC-3'), (50 [iM, 0.80 tg/tiL) and NtermHisRev_Lower (5'-

AGCAGGCCTTGCTCTCTGCCTCACCCTGGAGGAGGCACGGCTCCTTCTTCTCG

CC-3'), (50 iM, 0.81 gg/4L) and heated at 90 oC for 10 minutes. After allowing the

mixture to equilibrate to room temperature, it was brought up to a total volume of 20 [L

containing 11.2 [L deionized water, 420 [M of a dNTP mixture, 10 mM Tris-HCl pH









7.5, 5 mM MgC12, 7.5 mM dithiothreitol, and 5 U of Klenow fragment. This reaction

mixture was then allowed to incubate at 25 C for 80 minutes. After inactivation (20

mins at 75 C) the mixture was cooled to room temperature and then doubly digested

with BamHI and Stul (9 hours at 37 C). Following heat inactivation, a clean-up of the

enzymatic reaction was performed using the QIAquick Nucleotide Removal Kit. The

Nterm6XHisTag insert was then ligated into the BamHI/Stul sites of the

pFastBacHTalnsulin/ST vector. The new construct,

pFastBacHTalnsulin/Nterm6XHisTag/ST (N-Tag-ST), was sequenced at University of

Florida ICBR Protein Core Facility to confirm the presence of the Nterm6XHisTag

insert. Isolation of bacmids and generation of baculovirus stocks followed the BBES

protocol.

Baculovirus vector preparation for C-Tag-ST

The pFastBacHTalnsulin/ST plasmid was purified from E. coli strain ER2925

(dcm-) using the Wizard Plus Minipreps kit. This step was necessary to produce

plasmid that could be restricted with Dam or Dcm- sensitive restriction enzymes such as

Eco01091. The isolated plasmid was digested with Eco0109I for 10 hours at 37 C.

Clean-up of the enzymatic digestion was performed using the QIAquick Nucleotide

Removal Kit. The Eco0109I digested plasmid was digested for 8 hours at 37 C with

Xhol. After thermal inactivation (20 min at 65 C), the doubly digested plasmid was

agarose gel purified then extracted with the QIAquick Gel Extraction Kit. The 5' ends

generated by digestion were dephosphorylated using SAP.

The C-Tag-ST insert was created following a similar procedure to that of the

Nterm6XHisTag insert. Primers were mixed in equal volumes (2.3 gL) of

CtermHISForUpper (5'-









GGTAGGCCCTGGCCATTAAGCGGATGCTGGAGATGGGAGCTATCAAGAACCT

CACGTCC-3'), (50 [iM, 0.88 [tg/tL) and CtermHisRev_Lower (5'-

AGCCTCGAGTTAGTGATGGTGATGGTGATGACCGCCGAAGGACGTGAGGTTC

TTGATAGC-3'), (50 [iM, 0.92 gg/4L) and heated at 90 oC for 10 minutes. After

allowing the mixture to equilibrate to room temperature, it was brought up to a total

volume of 20 [L containing 11.5 [L deionized water, 462 [M of a dNTP mixture, 10

mM Tris-HCl pH 7.5, 5 mM MgC12, 7.5 mM dithiothreitol, and 5 U of Klenow fragment.

This reaction mixture was then allowed to incubate 25 C for 2 hrs. After inactivation,

the mixture was cooled to room temperature and then digested with Eco0109I (16 h at 37

C). The plasmid was purified using the QIAquick Nucleotide Removal Kit, digested

with Xhol (8 h at 37 C), then thermally inactivated. The Cterm6XHisTag insert was

ligated into the Eco01091/XhoI sites of the pFastBacHTalnsulin/ST vector. The new

construct, pFastBacHTa/ST/6XHisTag/Ctermlnsulin (C-Tag-ST), was sequenced at

University of Florida ICBR Protein Core Facility to confirm the presence of the

Cterm6XHisTag insert. Isolation of bacmids and generation of baculovirus stocks

followed BBES protocol.

Expression and purification of Ins-ST

Recombinant baculovirus stocks, as prepared following the common BBES

protocols, were amplified four times before a sufficient titer was obtained. For Ins-ST

baculovirus amplification, Sf9 insect cell cultures (50 75 mL) containing 2 x 106

cells/mL were infected with 4 mL of viral stocks, collected 42-50 hours post-infection,

and centrifuged at 4000 rpm for 15 mins at 40C Prior to subsequent amplifications or

expression, viral stocks were filtered with 0.2 |tm sterile filters. For Ins-ST expression,









cell cultures (50 75 mL) containing 2 x 106 cells/mL in 250 mL polycarbonate shaker

flasks (Corning, NY) were infected with 3 mL of viral stock with a titer of 1.7 x 108

pfu/mL. After 72 hours post infection, the following purification protocols were

followed.

Cultures were collected and centrifuged at 14000 RPM (4 C), after which, the

supernatant was concentrated using a 200 mL capacity Amicon ultrafiltration stir cell

model 8200 (Billerica, MA) with an amicon ultrafiltration polyethersulfone membrane

with a 10 kDa MWCO (product number PBGC04310). The concentrate was brought to

20% glycerol, 0.01 triton CF 54, 300 mM p-lactose and loaded onto a sepharose CDP-

hexanolamine affinity column (1.7 x 12 cm) equilibrated with 10 mL of a buffer

containing 300 mM p-lactose, 20% glycerol, 0.01% triton CF 54, and 50 mM MES, pH

6.8 at 4 C. This buffer was also used to wash the column after loading the sample.

Elution of enzyme proceeded using a KC1 step gradient of 150 mM, 250 mM, 350 mM

and 450 mM. These KC1 buffers also include 20% glycerol, 0.01% triton CF 54, 50 mM

MES, pH 6.8. Fractions (5 mL) containing activity were pooled and concentrated with a

50 mL Amicon ultrafiltration stir cell model 8050 with an Amicon polyethersulfone

membrane with alO kDa MWCO (product number: PBGC09005). Activity was

determined by the method described below. Bradford and bicinchoninic acid (BCA)

assays were used to estimate protein concentration.

Expression and purification of N-Tag-ST and C-Tag-ST

Recombinant baculovirus stocks, as prepared following the common BBES

protocols, were amplified four times before a sufficient titer was obtained. For N-Tag-ST

and C-Tag-ST baculovirus amplification, Sf9 insect cell cultures (50 75 mL) containing

2 x 106 cells/mL were infected with 4 mL of viral stocks, collected 42 50 hours post-









infection, and centrifuged at 4000 rpm for 15 mins at 40C Prior to subsequent

amplifications or expression for purification, viral stocks were filtered with 0.2 |tm sterile

filters. For N-Tag-ST and C-Tag-ST expression, cell cultures (50 75 mL) containing 2

x 106 cells/mL in 250 mL polycarbonate shaker flasks (Corning, NY) were infected with

3 mL of 2.1 x 108 pfu/mL viral stock (N-Tag-ST) or 3 mL of 2.7 x 108 pfu/mL viral

stock (C-Tag-ST). After 72 hours post infection, the following purification protocols

were followed.

Cultures were collected and centrifuged at 14000 RPM (4 C), after which, the

clarified supernatant was brought to 20% glycerol, 0.01 % Triton CF 54. Novagen Ni-

NTA-His Bind resin was packed into a column (2 x 10 cm) at 40C and was equilibrated

with 10 mL of buffer containing 5 mM imidazole, 100 mM KC 20% glycerol, 0.01%

triton CF 54, 50 mM MES pH 6.8. After loading, the column was washed with 10

column equivalents of a buffer containing 300 mM p-lactose, 5 mM imidazole, 100 mM

KC 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8, followed by a 5 column

equivalent wash with a buffer containing 5 mM imidazole, 100 mM KC1, 20% glycerol,

0.01% triton CF 54, 50 mM MES pH 6.8. Next, the column was washed with a buffer

containing 50 mM imidazole, 100 mM KC1, 20% glycerol, 0.01% triton CF 54, 50 mM

MES pH 6.8 to remove any proteins which non-specifically bind to the column. Finally,

N-Tag-ST or C-Tag-ST was eluted with a buffer consisting of 35 mM imidazole, 100

mM KC1, 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8. Fractions (5 mL)

containing activity were pooled and concentrated with a 50 mL Amicon ultrafiltration stir

cell model 8050 with an Amicon polyethersulfone membrane with alO kDa MWCO

(product number: PBGC09005). Activity was determined by the method described









below. Bradford and bicinchoninic acid (BCA) assays were used to estimate protein

concentration.

Activity assays for recombinant sialyltransferase

All activity assays reported on sialyltransferase were performed in the following

way: a 10 ptL aliquot from a selected sample was incubated with a 10 [tL mixture of 100-

170 [LM [9' 3H] CMP-NeuAc (final concentration of 50 to 85 [LM; 10,000 20,000

counts-per-min; specific activity of 10 20 [tCi/[tmol) and 235 mM a-lactose (final

concentration of 118 mM) for the appropriate amount of time to limit the consumption of

CMP-NeuAc to less than 10% at 25 C. The reaction mixture was quenched with 500

mL of 5 mM inorganic phosphate buffer, pH 6.8 and then applied to a Dowex 1 X 8, 200

mesh (P042-) mini-columns pH 6.8.114;122 Reactions were eluted with 3.5 mL of 5 mM Pi

buffer pH 6.8 into liquid scintillation vials. The definition of a unit of activity is the

amount of enzyme that produces 1 [tmol of sialyl-lactose per minute under the assay

conditions described. Because the activity was not obtained under saturating conditions,

the activity reported was corrected with experimentally determined kinetic parameters,

known substrate concentrations, and the following bi-substrate equation.

Vadjusted = (Vmax*[A]*[B]) / (KmA* KmB + KmA* [B] + KmB[A] + [A]*[B])

N-Tag-ST stability experiments

All samples and dilution buffers were kept at 4 C during these experiments. N-

Tag-ST sample starting buffer consisted of 20% glycerol, 0.01 % triton CF 54, 100 mM

KC1, 135 mM imidazole (i.e. purification buffer). The following conditions were used to

generate the graph illustrating the activity of N-Tag-ST after dilution with different









detergent concentrations (Figure 2-12). Percent activity was normalized based on the

activity prior to dilution and the final volume.

* 0.025% Triton X-100: N-Tag-ST (20 [LL) was diluted to 40 p.L with a buffer
consisting of 20% glycerol, 0.05 % triton x 100, 100 mM KC1, 50 mM MES pH
6.8.

* 0.050% Triton X-100: N-Tag-ST (20 [LL) was diluted to 40 p.L with a buffer
consisting of 20% glycerol, 0.1 % triton x 100, 100 mM KC1, 50 mM MES pH 6.8.

* 0.010% Triton CF 54: N-Tag-ST (20 [tL) was diluted to 40 ptL with a buffer
consisting of 20% glycerol, 0.01 % triton CF 54, 100 mM KC1, 50 mM MES pH
6.8.

* 0.025% Triton CF 54: N-Tag-ST (20 [tL) was diluted to 40 ptL with a buffer
consisting of 20% glycerol, 0.05 % triton CF 54, 100 mM KC1, 50 mM MES pH
6.8.

* 0.050% Triton CF 54: N-Tag-ST (20 [tL) was diluted to 40 ptL with a buffer
consisting of 20% glycerol, 0.1 % triton CF 54 / 100 mM KC1 / 50 mM MES pH
6.8.

For the dilution\concentration experiments, ultrafiltration was performed with

Millipore's Ultrafree-MC centrifugal filter units (Durapore PVDF membrane; 10,000

MWCO; 400 ptL capacity) at 3000 RPM. The following conditions were used to examine

enzyme activity after dilution and concentration with different dilution factors (Figure 2-

13). For nomenclature, the first value corresponds to the volume of purified N-Tag-ST

added to the device, with the second value corresponding to the number of dilutions with

dilution buffer (20% glycerol, 0.01 %, triton X-100, 100 mM KC1, 50 mM MES pH 6.8).

Percent original activity was normalized based on the activity prior to concentration and

the final volume (-100 |tL).

* 400 x 0: N-Tag-ST (400 [tL) was concentrated to a final volume of 102 [tL.

* 200 x 1: N-Tag-ST (200 [tL) was first diluted to 400 ptL with dilution buffer prior to
being concentrated to a final volume of 132 [tL.









* 200 x 3: N-Tag-ST (200 [tL) was first diluted to 400 ptL with dilution buffer. After
concentration to 200 [tL, another 200 ptL of dilution buffer was added to the
concentrate while still in the unit. The solution was then concentrated one more
time to 200 [tL, when upon the sample was diluted to 400 p.L again with dilution
buffer prior to being concentrated to a final volume of 140 [tL.

* 100 xl: N-Tag-ST (100 [tL) was first diluted to 400 ptL with dilution buffer prior
to being concentrated to a final volume of 70 [tL.

* 200 x 5: N-Tag-ST (200 [tL) was first diluted to 400 ptL with dilution buffer. After
concentration to 200 [tL, another 200 ptL of dilution buffer was added to the
concentrate while still in the unit. Dilution and concentration was performed 3
more times until a final volume of 120 itL was achieved.

* 400 x 5: N-Tag-ST (200 [tL) was first concentrated to 200 ptL before being diluted
to 400 ptL with dilution buffer. Dilution and concentration was performed 4 more
times until a final volume of 120 ptL was achieved.

In-gel digestion of glycoprotein 64

Glycoprotein 64 (GP 64) was identified by SDS-PAGE analysis, in-gel digested

with Trypsin Gold (Promega), and LCQ-MS analysis of extracted peptides. The band

corresponding to GP 64 was cut into 1mm cubes and washed 2 times with water in 500

ptL microfuge tubes. The cubes were then destined with 100 to 150 ptL of a buffer

containing 25 mM ammonium bicarbonate and 50 % acetonitrile at 27 C for 4 to 8 hours

with removal and re-application 3 times. After all the coomassie blue stain was extracted

from the gel slices, they where again washed with water. To digest recombinant

sialyltransferase, 50 ptL of 20 ng/ptL trypsin in 20 mM ammonium bicarbonate was added

and allowed to absorb into the gel cubes for 30 minutes at 27 C. Next, the appropriate

volume of 20 mM ammonium bicarbonate was added to cover the gel cubes and allowed

to incubate at 27 C overnight. Supernatant was collected with subsequent extraction of

peptides from the gel cubes with 100 ptL of a buffer containing 30 % acetonitrile and 5%

acetic acid (v/v) three times. Collected supernatant and extracted peptides were









combined and vacuumed to dryness with a Labconco CentriVap Concentrator (Model

number 7810000) at 37 C. For LCQ MS analysis, samples were re-suspended in 5%

methanol, 0.5 % acetic acid (v/v).

MS analysis

LCQ MS was performed on a Thermo Electron LCQ Deca (San Jose, CA) operated

in positive ion mode. A spray voltage of 1.8 kV +/- 0.2 was applied at the liquid junction

before a capillary column. An Eldex MicroPro pump delivered solvent at 6 [tL/min to a

LC Packings C18 PepMap Nano-Prepcolumn (0.3 x 1 mm). Five microliters of sample

was loaded onto the nano-prep column and washed for 5 mins at 0 % B. After washing,

the direction of flow was reversed with a new flow rate of 200 nL/min after pre-column

splitting from 8 [tL/min. Elution of peptides from a New Objective (0.75mm x 5 mm)

Proteopep capillary column included a gradient of 3-60% B over 30 minutes and 60-90%

B over 1 minute with solvent A as 0.1% acetic acid, 3% acetonitrile and solvent B as

0.1% acetic acid, 95% acetonitrile.














CHAPTER 3
HYDROGEN / DEUTERIUM EXCHANGE HPLC-MS FOR SIALYLTRANSFERASE
SECONDARY STRUCTURE

Introduction to Hydrogen/Deuterium Exchange HPLC-MS

The full characterization of a protein requires the understanding of function,

structure, and dynamics as investigated with a variety of techniques. A protein's

functional characterization is normally performed with biochemical, molecular, and

cellular types of analysis (Chapter 2), whereas protein tertiary structure and dynamics are

investigated with physical methods.127 Over the last 10 years, many groups have applied

hydrogen-deuterium exchange (H/Dx) coupled with on-line HPLC-MS as an analytical

technique to describe protein secondary structure in regards to protein topology, protein-

protein interactions, protein folding dynamics, and protein-ligand interactions.128 H/Dx

provides information on the solvent accessibility of amide protons by the exchange of a

proton to a deuterium. Since there is a mass shift associated with this exchange, MS is

used to determine the new molecular weight, thus yielding information on solvent

accessible sites. Because MS is geared towards analyzing very small amounts

(picomoles) of proteins and mass shifts associated with protein/peptide derivatization,

many investigators have chosen H/Dx HPLC-ESI-MS to study the secondary structure

and the folding dynamics of a particular protein.

Including H/Dx HPLC-ESI-MS, X-ray crystallography and nuclear magnetic

resonance (NMR) may describe protein three dimensional structure. X-ray

crystallography and NMR are proven techniques for locating all atoms in a protein to









within 2-3 A.127;129 NMR spectroscopy has improved, but analysis is still limited to

proteins under 35,000 Da and a minimal amount of posttranslational modifications

(PTMs). In regards to X-ray crystallization, the art of crystallizing proteins has limited

the number of protein crystal structures reported.130 Also, proteins with PTMs, such as

glycosylation, normally hamper crystallization.127 We evaluate H/Dx HPLC-ESI-MS to

study the active-site amino acids and dynamics of recombinant hST3Gal IV because it's

molecular weight is about 37,000 Da and it contains four N-linked glycan.

Monitoring Hydrogen/Deuterium Exchange of Amide Protons

H/Dx of protein amide protons may be monitored with a variety of methods;

however, NMR and MS are currently the main analytical methods for analysis because of

high through-put demands. The first discussion of H/Dx was presented by Linderstrom-

Lang in the mid 1950's after the description of ca-helixes and P-sheets.131 They

speculated that hydrogens not involved with H-bonding should be observable with

exchange experiments. In the 1960's, hydrogen-tritium exchange with liquid scintillation

analysis demonstrated that different exchange rates were observable between amide

protons involved with H-bonding and those which were solvent accessible.

NMR spectroscopy has also been used to study H/Dx; however, NMR spectroscopy

has the limitations described earlier. Johnson and co-workers were the first to use H/Dx

HPLC-ESI-MS to investigate the differences between apo- and holomyoglobin in

1994.131 H/Dx monitored with NMR spectroscopy is considered to be high-resolution in

comparison to H/Dx HPLC-MS because NMR may yields spatial resolution (exact sites

of exchange) and MS yields segment resolution (mass shift of a exchanged proteolytic

peptide in comparison to the mass of a the same non-exchanged proteolytic peptide).









Furthermore, the H/Dx of HEWL (-14.3 kDa) with NMR and MS analysis resulted in

similar descriptions of solvent assessable amide protons.132

Theoretically, MS/MS may provide the same spatial resolution as NMR; however,

hydrogen and deuterium may scramble upon peptide fragmentation to provide useless

information.129 Currently, it is accepted that MS/MS spectra are not useful for locating

individually exchanged amide protons. Thus, the length of a peptide as analyzed with

MS is inherent to the spatial details of solvent accessible amide protons. The digestion of

exchanged proteins is performed at low pH with pepsin. Pepsin is the main low pH

protease readily available for H/Dx HPLC-MS work flows. Since pepsin does not have

strict specificity for polypeptide cleavage, peptides with different cleavages within the

same region may overlap to provide spatial resolution.133

Hydrogen/Deuterium Exchange Kinetics

Proteins have protons with different exchange kinetics. For amide protons, the

kinetics are determined by protein structure, pH, and adjacent amino acid side chains.

These three characteristics affect solvent accessibility, backbone flexibility, and hydrogen

bonding. If an amide proton is exposed to D20 at pH 7.0 (25 C), the exchange rate will

be about 10-s. In comparison, the exchange rates of amide protons involved with H-

bonding are about 108 s.128

Proteins in solution exhibit highly dynamic behavior such as native flexing, also

known as "breathing." Furthermore, proteins may undergo allosteric or local changes

("hinge movement") upon protein-protein or protein ligand binding.127;134-136 The

changes in structure may be from local fluctuations to large conformational perturbations,

thus changing the intermolecular and solvent interactions of amide protons.136 The

difference between exchange rates of amide protons in unfolded and folded proteins









provides a probe for studying conformational changes and dynamic process. The model

for exchange kinetics of local or global protein structure is described with Equations 1

through 3 in Figure 3-1, where kex is the observed exchange rate for amide protons, kl is

the unfolding rate of the protein, k2 is the exchange rate of the solvent accessible amide

protons, and k-1 is the refolding rate of the protein.

Equation 1: kex = koki/(ko+k+ki)

Equation 2: EX1 ki >> ko then kex = ko

Equation 3: EX2 ko>> ki then kex = koki/ki
Figure 3-1. Equations Describing H/Dx Kinetics128;135

The link between measurable H/Dx rates and flexing dynamics is described by

Equation 1. From Equation 1, two extremes of kinetics, EX 1 for Equation 2 and EX 2

for Equation 3 are described. EX 1 represents the pseudo-first order kinetics in which

any un-folded event will result in complete exchange of the region which becomes

solvent accessible, thus the rate limiting step for EX1 is the protein unfolding. EX 2

represents the second-order kinetics in which protein-refolding is faster than H/Dx.

Portions of the proteins must unfold and refold several times before all portions of the

particular section are completely exchanged. Normally, proteins in solution follow the

EX 2 mechanism, whereas H/Dx under denaturing conditions will follow the EX 1

mechanism. The EX 2 mechanism is used to calculate protein folding energies (free-

energy of folding) and the El mechanism is used to map individual conformation states

that become populated upon different denaturing conditions.135

For the purpose of structural proteomics, it is beneficial to study proteins which

have an EX 2 mechanism. If a protein has an EX 1 mechanism, a bimodal distribution

i.e., two separate mass envelopes, is observed with MS analysis. This means that the









protein being studied is not stable under the conditions used for H/Dx. For structural

proteomic studies, it is important to maintain the tertiary structure of proteins in solution

during exchange.127

Operational Aspects of H/Dx HPLC-ESI-MS Experiments

To initiate exchange, the protein is diluted with 99.9 % D20 at pH 7.0 (with and

without substrate), quenched with formic or acetic acid at 0 oC, digested with pepsin, and

analyzed with HPLC ESI-MS. Compared are peptides with the same primary sequence

but different molecular weights due to deuterium incorporation. The equation in Figure

3-2 is applied to calculate deuterium incorporated (D) into a protein or peptide, where mt

is the protein or peptide mass at a designated exchange/quench time, mo/o is the mass of

the protein or peptide without deuterium incorporation, mloo% is the theoretical mass of a

fully deuterated protein or peptide, and N is the total number of amide protons available

for exchange.



D =N( mt mo%
mloo% mo%

Figure 3-2. Equation for Calculating the Number of Exchanged Amide Protons 131

One of the most important practical aspects of H/Dx HPLC-ESI-MS is back-

exchange of deuterium to hydrogen upon sample analysis. Back-exchange is a pH,

temperature, and time dependent process, thus the time needed for pepsin digestion and

separation are optimized to minimize back-exchange.133 Furthermore, digestion and

separation are performed at low pH and 0 C.









Global exchange and local exchange experiments are designed to obtain maximum

amounts of data (Figure 3-3).131 A global exchange experiment refers to a non-digested

protein, whereas a local exchange experiment refers to a digested protein.

SH/D exchange of sample and quench with
low pH buffer at different time intervals

r----------------------


SPepsin digest for 30 to 45
1:1 enzyme:substrate
I 4C




(DII HPLC

----------- -------------

S ESI FTICR MS
Figure 3-3. Analytical Work Flow for H/Dx Experiment

For a global exchange experiment, three steps are needed. First, the protein is

exchanged then quenched with a low pH buffer. Second, HPLC provides a quick desalt

and allows fast exchanging protons found on amino acid side chains to exchange back to

protons. Lastly, proteins are analyzed with MS. The total mass increase between non-

exchanged and exchanged protein yields the total number of solvent accessible amide

protons. The local exchange protocol follows the same steps as the global exchange

experiment; however, digestion with pepsin is added after the quench step. Peptic

peptides are mapped against the primary sequence of the selected protein to determine the

percent coverage and possible local resolution.

Hydrogen/Deuterium Exchange HPLC-ESI-MS Analysis of Selected Proteins

It is the goal of structural proteomics to provide useful answers to biological

questions based on well-designed experiments. Many laboratories have reported H/Dx









HPLC-ESI-MS methodologies for describing protein-protein and protein-ligand

interactions. Protein-protein interactions are not described in this dissertation; however,

the experimental design is similar to experiments used to solve protein-ligand interactions

and dynamics. Often, H/Dx data are used to map regions of solvent accessibility on

solved crystal structures. H/Dx data strongly complements crystal structure data;

however, X-ray crystallography only presents a "snap-shot" of a highly dynamic protein

structure. Also, the H/Dx data may describe new motifs without a solved crystal

structure. Finally, the requirement exists that the protein under investigation is

amendable to the experimental conditions.

The following examples provided the basis for experiential design for the analysis

of recombinant ST3Gal IV. First, dihydrodipicolinate reductase (28.7 kDa) was

investigated in global and local experiments to describe catalytically open and closed

forms upon binding of NADH and inhibitors. Four peptic peptides were localized to a

binding region and a hinge region needed for "closing" upon binding of inhibitor. The

results clearly described the difference between catalytically inactive and active forms

not described by the crystal structure.137 Second, diaminopimelate dehydrogenase exists

as a homodimer (70.4 kDa) that binds to either NADPH, diaminopimelate (DAP), or

both. Using localization experiments, the exchange of several different peptide regions

under different conditions (binding of NADPH, DAP, or both) resulted in different amide

proton solvent accessibility. The final results characterized a region that changes

conformation based on which set of substrates were present.138 Third, purine nucleoside

phosphorylase exists as a homotrimer (31.5 kDa) that has three separate active sites.

Interestingly, when only one site was occupied by the inhibitor, Immucillin-H, the









homotrimer was completely deactivated. Globular studies show that less amide protons

are exchanged in the protein-inhibitor complex. Furthermore, localization studies with

peptic peptides illustrated that 10 different peptides from all three subunits had different

exchange rates when completed to one equivalent of inhibitor. This study illustrated that

binding of one molecule of inhibitor to one site in the homotrimer created allosteric

effects in respect to the other subunits. Once again, this could not have be described with

X-ray crystallography, although the data was mapped onto the crystal structure for

conformation of solvent accessible sites.139 Although there are many other systems

described in the literature, they all follow the same H/Dx HPLC-MS protocols.

Experimental Design for Analysis of Sialyltransferase

ESI-FTICR-MS provides high resolution and the ability to trap and analyze large

proteins, thus providing functional advantages over other MS instruments. Exact mass

determination facilitates correct assignment of intact protein and peptides from pepsin

digestion. Figure 3-4 provides the experimental design for a complete study on

recombinant hST3Gal IV secondary structure.

First, hST3Gal IV is denatured and exchanged, thus allowing the percent back

exchange to be calculated. Second, non-denatured ST3Gal IV is exchanged and

quenched in thirty-second intervals. As opposed to the denatured ST3Gal IV case, only a

certain number of amide protons are solvent accessible, thus the number of solvent

accessible amide protons of non-substrate, non-denature ST3Gal IV is determined.

Third, ST3Gal IV is digested with pepsin prior to LC analysis. At this point, the peptide

coverage, digestion efficiency, and reproducibility of retention times are evaluated.

Fourth, H/Dx data on the

































Data Analysis-Bioinformatics
Figure 3-4. Six Experiments for Complete Characterization of Recombinant ST3Gal IV
constructs with H/Dx HPLC-MS

substrate-ST3Gal IV complex is collected using the natural substrates. The number of

amide sites protected (or de-protected) from solvent is calculated, thus leading to the final

two experiments.

The last two experiments include digestion of H/Dx protein, with and without

substrate present. As stated before, these experiments identify peptides with changes of

solvent accessibility. Data collected on important peptides are compared to known

biological data on hST3Gal IV. Before analyzing ST3Gal IV, conditions are optimized

with standard proteins. Important structural motifs and structural dynamics of

recombinant hST3Gal IV may be determined by comparing H/Dx data sets of globular

enzyme, global enzyme-substrate, pepsin digest of enzyme, and pepsin digest of enzyme-

substrate complex.




Full Text

PAGE 1

MASS SPECTROMETRY FOR STRUCTURAL PROTEOMIC ANALYSIS OF RECOMBINANT HUMAN SIALYLTRANSFE RASE AND IDENTIFICATION OF NANOPARTICLE HARVESTED OLIGONUCLEOTIDES By JEREMIAH D. TIPTON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Jeremiah D. Tipton

PAGE 3

This work is dedicated to my family and close friends that have supported me during this period of growth and to those which have left us too soon. Bryan LaRose John P. Collins Carl Hilley Crystal Tipton

PAGE 4

ACKNOWLEDGMENTS First and foremost, my sincerest gratitude goes to both of my advisors, Dr. David Powell and Dr. Nicole Horenstein, for their patience, guidance, a nd support during the course of my graduate studies and the course of these projects. I would also like to recognize and thank the members of my co mmittee, Dr. Jon Stew art, Dr. Weihong Tan, and Dr. Nancy Denslow, for their suggestions and support. Between two laboratories, I would like to thank all my new sisters: Jen, Erin, Fedra, and Mirela from the Horenstein Group and Quan Li, Violeta, Lani, Cris, Daniella and Joanna from the Powell group for all thei r support and colorful conversations over the years. Additionally, I would like to thank Alonso and Josh for their insight on different interesting areas of science a nd Scott and Dr. Johnson for their help on obtaining LCQ-MS data. I would also like to thank Romaine Hughes for her administrative help and the many breaks spen t on the bridge. A special thank you goes out to Mike, Ben, Travis, Josh, Larry, Brent, and Merve, for all the good times we had through these interesting years. In my immediate family, I am deeply inde bted to my parents Robin and Dennis for all their support whenever I was in need and my younger siblings Rachel, Tamara, Brandon, Derek, and Devin for all their inquiries as to why I am still in college. In my extended family, I would like to thank Jan a nd the late John Collins for allowing me to stay at their house when I needed a quiet place to stay. Lastly, I would like to thank Matt for all his support over the ma ny years I have known him. iv

PAGE 5

1 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................................................................................. iv LIST OF TABLES ............................................................................................................. ix LIST OF FIGURES ........................................................................................................... xi ABSTRACT ..................................................................................................................... xvi CHAPTER INTRODUCTION TO BIOLOGI CAL MASS SPECTROMETRY ............................ 1 Systems Biology and Mass Spectrometry .................................................................... 1 Mass Spectrometry Based Proteomics .......................................................................... 3 Functional Proteomics ........................................................................................... 4 Structural Proteomics ............................................................................................ 5 Mass Spectrometry and Larg e Bio-molecules Meet ............................................. 5 Mass Spectrometry Analysis of Oligonucleotides ........................................................ 6 Complete Mass Spectrometry Analysis ........................................................................ 8 Separation Science ................................................................................................. 9 Off-line separation technology ..................................................................... 10 On-line separation technology ..................................................................... 12 Ionization Techniques: ESI and MALDI ............................................................ 14 Mass Analyzers ................................................................................................... 19 Quadruple ion trap MS ................................................................................. 20 Time-of-flight MS ........................................................................................ 22 Fourier transform ion cy clotron resonance MS ............................................ 24 Bioinformatics ..................................................................................................... 28 Intact peptide MS data analysis with in silico digestion algorithms ............ 29 Tandem MS data analysis with in silico digestion algorithm ...................... 30 Tandem MS data analysis with de novo algorithims .................................... 31 Software packages for analysis of MS and MS/MS data ............................. 32 Bioinformatics data visua lization and interpretation ................................... 34 Summary ..................................................................................................................... 35 v

PAGE 6

2 EXPRESSION AND PURIFICATION OF RECOMBINANT HUMAN ALPHA 2 3 SIALTRANSFERASE ...................................................................................... 36 Introduction to Oligosaccharide s and Glycosyltransferases ....................................... 36 Complex Carbohydrates in Biological Systems .................................................. 37 Glycosyltransferase Family ................................................................................. 38 Sialyltransferase Sub-Family ...................................................................................... 40 Biological and Medicinal Importance ................................................................. 44 Sialyltransferase Structure and Mechanism Relevant for Structural Proteomics ........................................................................................................ 45 Recombinant Human (2 3) Sialyltransferase IV ................................................. 48 Results and Discussion ........................................................................................ 50 Baculovirus preparation ............................................................................... 50 Expression and purification of the th ree sialyltransferase constructs .......... 51 Concentration/dilution studies on N-Tag-ST ............................................... 60 Kinetic parameters ........................................................................................ 63 Conclusion ........................................................................................................... 64 Methods and Materials ........................................................................................ 64 Baculovirus vector preparation for Ins-ST ................................................... 65 Baculovirus vector prep aration for N-Tag-ST ............................................. 66 Baculovirus vector prep aration for C-Tag-ST ............................................. 67 Expression and purification of Ins-ST ......................................................... 68 Expression and purification of N-Tag-ST and C-Tag-ST ............................ 69 Activity assays for recombinant sialyltransferase ........................................ 71 N-Tag-ST stability experiments ................................................................... 71 In-gel digestion of glycoprotein 64 .............................................................. 73 MS analysis .................................................................................................. 74 3 HYDROGEN / DEUTERIUM EXCHANGE HPLC-MS FOR SIALYLTRANSFERASE SE CONDARY STRUCTURE ........................................ 75 Introduction to Hydrogen/Deuterium Exchange HPLC-MS ...................................... 75 Monitoring Hydrogen/Deuterium Exchange of Amide Protons ......................... 76 Hydrogen/Deuterium Exchange Kinetics ............................................................ 77 Operational Aspects of H/Dx HPLC-ESI-MS Experiments ............................... 79 Hydrogen/Deuterium Exchange HPLC-ESI-M S Analysis of Selected Proteins 80 Experimental Design for Analys is of Sialyltransferase ....................................... 82 Results and Discussion ............................................................................................... 84 Optimization of Analytical System for Analysis of Intact Proteins .................... 84 Dynamic and gas assisted trapping and analysis of intact proteins with ESI-FTICR MS ......................................................................................... 84 Broadband versus heterodyne excitation and detection ............................... 90 Optimization of HPLC-ESI-FTICR-MS ...................................................... 98 Separation and MS analysis of intact proteins ............................................. 98 H/Dx Direct Infusion and H/Dx HPLC-ESI-FTICR-MS of Intact HEWL ....... 101 Separation and MS Analysis of Pepsin Digested Proteins ................................ 104 MS Analysis of Recombinant Sialyltransferase ................................................ 110 vi

PAGE 7

EndoHf and PNGase F diges tion of intact Ins-ST ..................................... 114 MS analysis of pepsin digested sialyltransferase ....................................... 115 Conclusion ................................................................................................................ 116 Material and Methods ............................................................................................... 117 FT-ICR/MS Instrumentation ............................................................................. 117 HPLC Instrumentation ....................................................................................... 117 Intact Protein Analysis with Broadband Mode ................................................. 118 Direct infusion analysis of intact proteins .................................................. 118 HPLC-ESI-FT-ICR MS analysis of intact proteins .................................... 119 Pepsin Digestion and HPLC-ESI-FTICR-MS Analysis .................................... 120 H/Dx Experiments ............................................................................................. 121 4 MS BASED STRUCTURAL PROT EOMICS WITH SMALL MOLECULE LABELING AGENTS FOR IDENTIFYI NG SIALYLTRANSFERASE ACTIVE SITE AMINO ACIDS .............................................................................................. 122 Introduction to Bioconjugation Techniques ............................................................. 122 Practical Consideration and Types of Small Molecule Labeling Agents .......... 123 Practical Considerations for MS Based Structural Proteomics with SMLAs ... 124 Literature Review on Ideal Cases ...................................................................... 125 Limitations for MS Based Struct ural Proteomics with SMLAs ........................ 130 Practical Consideration for Labeling of Sialyltransferase with SMLAs ........... 131 Bioconjugation of Sialyltransferas e with Small Affinity Labels .............................. 132 Differential Labeling Scheme ............................................................................ 133 Workflow ........................................................................................................... 134 Results and Discussion ...................................................................................... 136 Preliminary screening of NAI, EDAC/ETAM, and IOA ........................... 136 NAI differential labeling experiments ........................................................ 137 Reactivation of NAI derivatized N-Tag-ST with ethanolamine ................ 140 SDS-PAGE analysis and in-gel digestion with trypsin or other proteases 141 MS analysis of IGD sample ....................................................................... 144 Manual interpretation of MS data .............................................................. 152 Conclusion ......................................................................................................... 152 Methods and Materials ...................................................................................... 153 Reaction conditions for prelim inary screening of SMLAs ........................ 153 NAI labeling experiments .......................................................................... 154 SDS-PAGE analysis and in-gel dige stion of recombinant hST3Gal IV .... 157 Mass spectrometry analysis ........................................................................ 158 5 MS BASED STRUCTURAL PROT EOMICS WITH SITE-DIRECTED PHOTOAFFINITY LABELING FO R THE INVESTIGATION OF SIALYLTRANSFERASE ACTI VE SITE AMINO ACIDS ................................... 161 Introduction to Site Directed Photo-affinity Labeling .............................................. 161 Photoreactive Groups ........................................................................................ 162 Practical Considerations for Site -Directed Photoaffinity Labeling ................... 164 Examples of Photoaffinity Labeling .................................................................. 164 vii

PAGE 8

Practical Consideration for Site Di rected Photoaffinity Labeling of Sialyltransferase ............................................................................................. 164 Site-Directed Photoaffinity Labeling of Sialyltransf erase with CMP-o-Azido Photoactivation of inhibitor in th e presence of commercially available Results and Discussion ............................................................................................. 167 CMP-o-azido Mandelate Synthesis and Characterization ................................. 167 Mandelate ....................................................................................................... 169 MS Analysis of In-Gel Digested Irra diated N-Tag-ST/Inhibitor Samples ........ 175 Conclusion ................................................................................................................ 179 Methods and Materials ............................................................................................. 180 Synthesis and Characterizati on of CMP-o-azido Mandelate ............................. 180 Photoaffinity Labeling Experiments .................................................................. 180 Incubation of N-Tag-ST with and without inhibitor .................................. 180 Photoactivation of inhibitor in the presence of N-Tag-ST ......................... 181 sialyltransferase ...................................................................................... 181 Activity assays for recombinant sialyltransferase ...................................... 183 In-Gel Digestion with Trypsin ........................................................................... 183 MS Analysis ...................................................................................................... 184 6 MS ANALYSIS OF DNA HARVESTED WITH NANOHARVESTING AGENTS ................................................................................................................... 185 Introduction to Oligonucleotide Analysis with Mass Spectrometry ......................... 185 Nanoharvesting Agents ..................................................................................... 188 On-line HPLC MS of Oligonucleotides ............................................................ 190 MS Instrumentation Considerations .................................................................. 191 Experiment Design ............................................................................................ 193 Results and Discussion ............................................................................................. 193 Solvent System Studies ..................................................................................... 194 IP-RP-HPLC ESI-FTICR-MS Analysis of Synthetic Oligonucleotides ........... 198 Conclusion ................................................................................................................ 199 Methods and Materials ............................................................................................. 200 FT-ICR/MS Instrumentation ............................................................................. 200 HPLC Instrumentation ....................................................................................... 201 7 CONCLUSIONS AND FUTURE WORK ............................................................... 202 The Sialyltransferase Project .................................................................................... 203 H/Dx HPLC ESI FT-ICR MS ............................................................................ 205 Small Molecule Labeling Agents ...................................................................... 207 Site Directed Photoaffinity Labeling ................................................................. 208 MS Analysis of Oligonucleotides Se lected with Nanoharvesting Agents ................ 209 Final Conclusions ..................................................................................................... 210 LIST OF REFERENCES ................................................................................................. 212 BIOGRAPHICAL SKETCH ........................................................................................... 223 viii

PAGE 9

LIST OF TABLES Table page 1-1. Protein Function in Biological Organisms .................................................................. 3 1-2. MS Experiments on Proteins ....................................................................................... 3 1-3. Mass Shift Associated with Common PTMs ............................................................... 4 1-4. Solvents Compatible with ESI .................................................................................... 17 1-5. Common Genomic Databases ................................................................................... 29 1-6. Common MS and MS/MS Peptide Data Analysis Software ..................................... 33 2-1. Viral Titers for Different Constructs ......................................................................... 51 2-2. Summary of an Ins-ST Purification ........................................................................... 53 2-3. Summary of a N-Tag-ST Purification ....................................................................... 56 2-4. Summary of C-Tag-ST Purification .......................................................................... 56 2-5. Summary of Selected Sialyltransf erase Cloning and Purification Papers ................. 58 2-6. Summary of Kinetic Parameters for Different Sialyltran sferase Constructs ............. 63 3-1. Summary of Different Trapping Methods ................................................................. 86 3-2. Relationship Between the Six Impor tant Parameters and Signal Intensity/ Resolution of the Isotopically Resolved + 10 Charge State of HEWL ...................... 88 3-3. Mass Accuracies for Di fferent Protein Standards ..................................................... 90 3-4. Mass Accuracies for Di fferent Proteins with HPLC -ESI FTICR-MS Analysis ....... 99 3-5. Peptides Observed from Bovine Carbonic Anhydrase Digested with Pepsin Presented in Figure 3-26 ........................................................................................ 111 3-6. Sialyltransferase Construct Concentr ation, Buffer, and Molecular Weights ........... 112 4-1. Literature Review of Diff erent Bioconjugation Experiments ................................. 127 ix

PAGE 10

4-2. SMLAs and Activity af ter 20 Minute Incubation .................................................... 137 5-1. Literature Review on Case Specific Photoaffinity Labeling ................................... 165 5-2. Percent Activity Remaining after In cubation of N-Tag-ST With and Without Inhibitor Present. .................................................................................................... 171 5-3. Irradiation of N-Tag-ST at 254 nm ( 0.65 Amps) with Different Concentrations of Inhibitor .................................................................................................................. 172 5-4. Irradiation of Recombinant Rat 2,3-(N)-sialyltransferase with 130 M Inhibitor Present .................................................................................................................... 174 5-5. Irradiation of Recombinant Rat 2,6-(N)-sialyltransferase without Inhibitor Present and with 130 M Inhibitor Present ............................................................ 174 5-6. Mass-to-Charge of + 1 Charge State of Theoretical Tryptic Peptides and Four Possible Mass Shifts ............................................................................................... 177 5-7. Mass-to-Charge of + 2 Charge State of Theoretical Tryptic Peptides and Four Possible Mass Shifts ............................................................................................... 178 6-1. Synthetic Oligonucleotides Prepared for NAH Selection and On-line HPLC-MS Analysis .................................................................................................................. 190 x

PAGE 11

LIST OF FIGURES Figure page 1-1. Diagram Representing Compone nts of Total Systems Biology ................................. 2 1-2. Common Fragmentation Patterns and Nomenclature .................................................. 6 1-3. Diagram Describing the Workflow for DNA Sequencing with CE ............................ 7 1-4. Diagram Describing the Workflow for Protein Sequencing with MS ......................... 7 1-5. Workflow for Comp lete MS Analysis .......................................................................... 8 1-6. Bottom-Up and Top-Down Workflows .............................................................. 10 1-7. Microbore MudPIT .................................................................................................... 14 1-8. Electrospray Ionization with Dole s Large Molecule Ionization Model ................... 15 1-9. Diagram Representing Matrix-Assi sted Laser Desorption Ionization ..................... 19 1-10. Actively Shielded 4.7 T Magnet at the University of Florida ................................. 25 1-11. Cyclotron Motion of Ions in Magnetic Field ........................................................... 25 1-12. Relationship between the Freque ncy and Mass-to-Charge of an Ion ..................... 26 2-1. Common Sugars Found in Oligosaccharides ............................................................. 37 2-2. Reaction Catalyzed by Sialyltransferases ................................................................... 41 2-3. Topology of Sialyltransferase .................................................................................... 42 2-4. Sialylmotifs and Amino Acid Length of Sialyltransferases ...................................... 46 2-5. Sialyltransferase Constructs Prepared ....................................................................... 49 2-6. Confirmation of Correct DNA Insertion Product ..................................................... 51 2-7. Protein Concentration Profile of an Ins-ST Purification ........................................... 53 2-8. Protein Concentration and Activity Profile of an Ins-ST Purification ...................... 54 xi

PAGE 12

2-9. SDS-PAGE Analysis of Purified Ins-ST and Deglycosylation Reaction .................. 55 2-10. Protein Concentration and Activity Profile of a C-tag-ST Purification .................. 57 2-11. SDS-PAGE Analysis of TCA Pr ecipitated N-Tag-ST and C-Tag-ST ................... 57 2-12. Effect on Activity of N-Tag-ST with Dilution of Different Detergent Concentrations .......................................................................................................... 61 2-13. Effect on N-Tag-ST Activity with Di fferent Dilution Factors and Concentration with Millipore Ultrafree-MC Centrifugal Filter Units ............................................. 62 3-1. Equations Describing H/Dx Kinetics ........................................................................ 78 3-2. Equation for Calculating the Nu mber of Exchanged Amide Protons ...................... 79 3-3. Analytical Work Flow for H/Dx Experiment ............................................................ 80 3-4. Six Experiments for Complete Char acterization of Recombinant ST3Gal IV constructs with H/Dx HPLC-MS ............................................................................. 83 3-5. Dynamic Trapping Experiment. ................................................................................ 86 3-6. Direct Infusion ESI FTI CR-MS Analysis of 10 pmol/ L of Ribonuclease A .......... 91 3-7. Direct Infusion ESI FTI CR-MS Analysis of 10 pmol/ L of HEWL......................... 92 3-8. Expanded +10 Charge State of th e Spectra Presented in Figure 3-7 ......................... 93 3-9. Direct Infusion of ESI FT ICR-MS Analysis of 15 pmol / L of Human Carbonic Anhydrase I. ............................................................................................................. 94 3-10. Direct Infusion ESI FTICR-MS Analysis of 20 pmol/ L of Bovine Serum Albumin .................................................................................................................... 95 3-11. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of HEWL ...................... 96 3-12. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of HEWL ...................... 97 3-13. TIC of HEWL (75 pmol) Anal yzed with HPLC-ESI-FT-ICR-MS ........................ 99 3-14. Mass Spectrum at 19.15 mins from th e TIC Presented in Figure 3-13. .............. 100 3-15. TIC of HEWL (50 pmol) and -Chymotrypsinogen A (50 pmol) Analyzed with HPLC-ESI-FT-ICR-MS ......................................................................................... 101 3-16. HEWL Spectrum Observed at 12.30 Minutes ...................................................... 102 3-17. -Chymotrypsinogen Spectrum Observed at 12.93 Minutes ............................... 103 xii

PAGE 13

3-18. TIC of Unknown Amount of Uridine Kinase Analyzed with HPLC-MS ESIFTICR-MS ............................................................................................................. 104 3-19. Uridine Kinase Spectrum Observed at 12.00 Minutes ......................................... 105 3-20. TIC of HEWL (25 pmol) Analyzed with HPLC-ESI-FT ICR-MS in Heterodyne Mode and Dynamic Trapping ................................................................................ 106 3-21. HEWL Spectrum Observed at 14.00 Minutes (Figure 3-20) ................................. 106 3-22. H/Dx Analysis of HEWEL w ith Direct Infusion ESI-FTICR-MS ........................ 107 3-23. Expanded +10 Charge Stat e Observed in Figure 3-22 .......................................... 108 3-24. TIC of H/Dx-HEWL Sample (10 pmol ) Analyzed with HPLC-ESI FTICR MS .. 109 3-25. Spectra of (1) Non-H/Dx of 10 pmol of HEWL and (2) 15 min H/Dx of 10 pmol of HEWL. ............................................................................................................... 109 3-26. Sequence Coverage of Bovine Ca rbonic Anhydrase Digested with Pepsin .......... 110 3-27. Activity of Ins-ST / Endo H f Digestion ................................................................. 115 4-1. ICAT Reagent .......................................................................................................... 126 4-2. ICAT Work-flow ..................................................................................................... 127 4-3. IOA Specificity ........................................................................................................ 132 4-4. EDC Coupling with ETAM ..................................................................................... 133 4-5. Acetylation of Tyrosine with NAI ........................................................................... 133 4-6. Differential Labeling Experiments .......................................................................... 134 4-7. Work Flow for Bioconjugation of Reco mbinant hST3Gal IV and Identification of Derivatized Amino Acids ....................................................................................... 136 4-8. Labeling Experiment ............................................................................................... 138 4-9. NAI Derivatization of C-Tag-ST without Substrate Present and with 525 M CMP-NeuAc Present .............................................................................................. 138 4-10. NAI Derivatization of C-Tag-ST wit hout Substrate Present, with 1.5 mM CMPNeuAc and Saturated -Lactose Present, and with Saturated -Lactose Present .. 139 4-11. Reactivation of N-Tag-ST with ETAM ................................................................. 141 4-12. SDS-PAGE Analysis of C-tag-ST and N-tag-ST .................................................. 142 xiii

PAGE 14

4-13. SDS-Page Analysis of C-Tag-ST a nd N-Tag-ST After Deglycosylation with Endo Hf ..................................................................................................................... 143 4-14. SDS-PAGE Analysis of TCA Preci pitated C-Tag-ST and N-Tag-ST with Subsequent IGD with Multiple Proteases ............................................................... 144 4-15. SDS-Page Analysis of Initial Scr een of Ins-ST with EDAC/ETAM, NAI, and IOA ........................................................................................................................ 145 4-16. SDS-Page Analysis of C-Tag-ST Labeled with NAI, with and without Substrate Present .................................................................................................... 146 4-17. MALDI-TOF MS Spectrum with On-S pot Water Wash Sample Preparation ...... 146 4-18. MALDI-TOF MS Spectrum after ZipTip Sample Preparation .......................... 147 4-19. Sequence Coverage of C-Tag-ST IGD with Trypsin, On-Spot Water Wash Cleanup, and MALDI-TOF MS Analysis .............................................................. 147 4-20. Sequence Coverage of C-Tag-ST IGD with Trypsin, Zip-Tip Cleanup, and MALDI-TOF MS analysis ..................................................................................... 147 4-21. Typical Total Ion Chromatogram of Trypsin IGD of N-Tag-ST with an LCQ DECA. The Retention Times and Base Peaks are Labeled ................................... 148 4-22. MS Spectra at 23.70 Minutes of the Tryptic Peptide LEDYFWVK ..................... 149 4-23. MS/MS Spectra at 23.80 Minutes of the Tryptic Peptide LEDYFWVK .............. 149 4 -24. Sequence Coverage of Ins-ST Digested with Trypsin / LC MS/MS .................... 149 4-25. Sequence Coverage of N-Tag-ST Digested with Trypsin / LC MS/MS ............... 150 4-26. Sequence Coverage of C-Tag-ST Digested with Trypsin / LC MS/MS ............... 150 5-1. Different Photoreactive Moieties ............................................................................. 162 5-2. Reaction Pathways of Activated Phenylazide ......................................................... 163 5-3. Aromatic Ring Substituted for NeuAc and K i Values ............................................. 168 5-4. CMP-o-azido Mandelate .......................................................................................... 168 5-5. Negative Ion Mode ESI-FTICR MS Spectrum of CMP-o-azido Mandelate ........... 170 5-6. Negative Ion Mode ESI-FTICR MS Analysis of CMP-o-azido Mandelate Decomposition Products ........................................................................................ 170 5-7. Activity Profile of N-Tag-ST Incuba ted With and Without Inhibitor Present. ....... 171 xiv

PAGE 15

5-8. Possible Labile Bonds of CMP-o-Az ido Mandelate and Possible Mass Shifts Associated with Derivatized Tryptic Peptide ......................................................... 176 5-9. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/180 M Inhibitor Experiment .............................................................................................. 177 5-10. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/98 M Inhibitor Experiment .............................................................................................. 178 6-1. Spectrum of 10 pmol/ L of 3Modified 19 mer Diluted in Buffer 1 ...................... 195 6-2. Spectrum of 10 pmol/ L of 3Modified 19 mer diluted in Buffer 2 ....................... 196 6-3. Charge State Distribution and Signal Intensity of 10 pmol/ L of 3Modifed 19 mer Diluted with the Different Buffer Systems ..................................................... 196 6-4 Charge State Distri bution and Signal Intensity of 1 or 10 pmol/ L of 5 mer Diluted in Buffer System 1 or 2 ............................................................................. 197 6-5. Linear Dynamic Range of the 2 Charge State of 5 mer Diluted with Buffer 2 ...... 197 6-6. IP-RP HPLC ESI FTICR MS Analysis of 15 picomoles of 5mer with a PS-DVB Monolithic Column and 25 mM TE AA as an ion-pairing agent ............................ 198 xv

PAGE 16

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MASS SPECTROMETRY FOR STRUCTURAL PROTEOMIC ANALYSIS OF RECOMBINANT HUMAN SIALYLTRANSFE RASE AND IDENTIFICATION OF NANOPARTICLE HARVESTED OLIGONUCLEOTIDES By Jeremiah D. Tipton December 2005 Chair: Nicole A. Horenstein Major Department: Chemistry Mass spectrometry (MS) has become one of the most pow erful analytical techniques for investigating proteins, pe ptides, and oligonucleotides when careful preparation of biological samples has been perf ormed. The first part of this dissertation describes the overall goals and analytical tools needed for studying systems biology. As a subcategory of systems biology, structural pr oteomics utilizes MS for the analysis of the three-dimensional structure of proteins with low purificat ion yields. Proteins may be derivatized with bioconjuga tion techniques to provide a measurable mass shift which then can be related structural information. The second part of this dissertation describes the expression and purification of th ree recombinant isoforms of human -(2 3) sialyltransferase (ST3GalIV) for bioconjuga tion experiments. Sialyltransferase was chosen because it catalyzes the transfer of sialic acid to the terminal ends of oligosaccharide chains found on glycolipids and glycoproteins Sialic acids found on the xvi

PAGE 17

terminal end of oligosaccharides are invol ved with several key biological recognition events such as cell adhesion, cell-cell recognition, and biological masking of disease states. By describing the 3D -structure of ST3GalIV, rati onal design of new inhibitors may proceed. The third part of this dissert ation describes three separate bioconjugation techniques used for the derivatization of ST3GalIV. These techniques include hydrogen deuterium exchange, derivatization with small molecule labeling agents, and derivatization with si te-directed photo-affinity labeli ng. After sample work-up, three different MS instruments are utilized for ma ss measurement. The fourth part of this dissertation describes the MS analysis of oligonucleotides selected with nanoharvesting agents. Finally, the conclusions and future work for these projects are presented. xvii

PAGE 18

CHAPTER 1 INTRODUCTION TO BIOLOGICAL MASS SPECTROMETRY Systems Biology and Mass Spectrometry Mass spectrometry (MS) has become one of the most powerful analytical techniques for investigating the diverse cas cade of biological molecules generated by metabolism in living organisms. MS as an analytical tool measures the mass-to-charge (m/z) of an ion in the gas phase. Biological MS has created a huge enterprise that has promised the molecular biologist and bioc hemists the instrument that will provide answers to many of their biological questions. 1 To be more precise with the opening statement, MS is the most powerful analytical tool for identification of biological molecules when careful preparation of bi ological samples has been performed and bioinformatics software exists for complex da ta analysis. With the current state of biotechnology, many disciplines have aligne d to solve the dynamic systems biology problem, with MS widely accepted as the main analytical tool. 2 Systems biology, still in its infancy, is l oosely defined as the study of the total integrated network of all biochemical reactions and interactions betw een different microand macromolecules within an organism which provide life sustaining processes. 2 This is different than traditional biology which uses the reductionism philosophy towards studying individual genes, protei ns, and cell specific function. 3 However, systems biology would not have a platform for integration of information w ithout first cataloging the specific function of gene s, proteins, and cells. Much like the complex machinery which creates life, many experts from all the sc iences need to work together to produce a 1

PAGE 19

2 viable picture of an organism which will in turn help increase our understanding of human health and disease. Development of technologies in the supporting fields for systems biology, presented in Figure 1-1, have and will continue to radica lly transform biomedical and medicinal research. The left column illustrates the main categories of biological molecules and the middle column consists of th e terms given to an entire complement of that particular molecule. The right column represents the hierarchical approach for an integrated view of how the different systems work together to create life, also known as functional genomics. The term systems biology is used to describe an entire organisms function as opposed to isolated parts, such as the genome, transcriptome, proteome or metabolome. DNA Complete description of genes in an organism RNA Complete description of mRNA in a genome Proteins Complete description of proteins expressed by a genome Metabolites Complete description of metabolites expressed by a genome Catalog of all protein-DNA, protein-RNA, and protein-protein interactions Genome Transcriptome Proteome Metabolome Interactome Genomics Transcriptomics Proteomics Metabolomics Functional Genomics or phenomics Global Targeted Global Targeted Global Targeted Systems Biology DNAComplete description of genes in anorganismRNAComplete description of mRNA in a genomeProteinsComplete description of proteins expressed by a genomeMetabolitesComplete description of metabolites expressed by a genome Catalog of all protein-DNA, protein-RNA, and protein-protein interactions DNAComplete description of genes in anorganismRNAComplete description of mRNA in a genomeProteinsComplete description of proteins expressed by a genomeMetabolitesComplete description of metabolites expressed by a genome Catalog of all protein-DNA, protein-RNA, and protein-protein interactions Genome Transcriptome Proteome Metabolome Interactome Genomics Transcriptomics Proteomics Metabolomics Functional Genomics or phenomicsGlobalTargeted GlobalTargeted GlobalTargeted Systems Biology Figure 1-1. Diagram Representing Co mponents of Total Systems Biology 2 Although there are many technologies need ed for systems biology analysis, this dissertation is limited to (1) MS based struct ural proteomics in regards to studying the

PAGE 20

3 protein tertiary structure of si alyltransferase and (2) MS anal ysis of oligonucleotides in respect to nanoharvested oligonucleotides. Be fore describing these projects, the current state of proteomics, separation technology, mass spectrometry instrumentation, and data analysis are described. Tables 1-1 and 1-2 present the differe nt functions of proteins and what MS experiments may be performed to describe useful information. 4 Table 1-1. Protein Functi on in Biological Organisms Protein Function in Biological System Catalysis Defense Movement Structure Storage Transport Signaling Storage Table 1-2. MS Experiments on Proteins Mass Spectrometry Experiments Interactions and Binding Protein Folding Protein Dynamics Higher-order Structure Primary Structure Enzyme Kinetics Post-translational modification(s) Mass Spectrometry Based Proteomics Proteomics, as termed by Mark R. Wilkins in 1994, is defined as the genome encoded protein complement at a certain time point. 5 It is the goal of proteomics to study what proteins are present, what the proteins do, and how they intera ct with each other. 6 At different times during an organisms life cy cle, different proteins may be present at different concentrations. The proteomics field is viewed as a high-throughput technology; however, the workflow and instrumentation are used for studies where very specific chemical derivatization followed by MS analysis is interpreted to provide

PAGE 21

4 characterization of active-site structure. Furthermore, proteomics may be sub categorized into functional and structural proteomics. 7 Functional Proteomics Functional proteomics focuses on identifyi ng the primary sequence of a protein, the concentration of proteins at different life cycles, the pos t-translational modifications (PTMs) in static or dynamic states and th e protein-protein inte ractions of mature proteins. 3 MS Experiments are designed to id entify proteins a nd protein-protein interactions qualitatively and quantitativ ely with both high selectivity and high sensitivity. It has been estimated that the pr otein concentrations w ithin any living cell or blood serum ranges between 5 to 15 orders-of-magnitude. 8 Also, proteins may be spliced into other isoforms or be modified with c ovalent attachment of a variety of PTMs upon Table1-3. Mass Shift Asso ciated with Common PTMs Pyroglutamic acid formed from Gln -17.03 Disulfide bond formation -2.02 Disulfide bond removal +2.02 C-terminal amide formed from Gly -0.98 Deamination of Asn and Gln +0.98 Formylation +28.01 Methylation +14.027 Acetylation +42.04 Carboxylation of Asp and Glu +44.01 Phosphorylation +79.98 Sulfonation +80.06 Cysteinylation +119.14 Myristoylation +210.36 Glycosylation Deoxyhexoses (Fuc) +146.14 Hexosamine (GlcN, LalN) +161.16 Hexose (Glc, Gal, Man) +162.14 N-Acetylhexosamine (GlcNAc, GalNAc) +203.19 Pentose (Xyl, Ara) +132.12 Sialic acid (NeuNAc) +291.26 Pyridoxal phosphate (Shiff base to lysine) +231.15 -N-6-Phosphogluconoylation +258.12

PAGE 22

5 suited for these types of studies. 9 Data for functional proteomi cs are not presented in this dissertation; however, many other groups have presented excellent da ta pertaining to the profiling of certain cell types (i.e., healthy ve rsus disease state), 10 the mapping of small proteomes (i.e., yeast), 11;12 or characterizing whole blood serum. 13;14 Structural Proteomics Structural proteomics focuses on the st ructure-function relationships between proteins and the mapping of three-di mensional structures of proteins. 5;15 As presented in Chapters 3 through 5, the mass shif t of amino acids derivatized in vitro with specific labeling agents can provide information on pr otein structure. Further discussion on the experimental design for MS based structur al proteomics, along with other competing techniques is presented in Chapters 3 through 5. Obviousl y, without certain developments in MS and molecular bi ology, proteomics would not exist. Mass Spectrometry and Large Bio-molecules Meet The discovery of the soft ionization tec hniques, electrospray ionization (ESI) in 1984 by Yamashita and Fenn 16 and matrix-assisted laser desorption ionization (MALDI) in 1988 by Tanaka et al., 17 both recognized with the N obel Prize in Chemistry 2002, allowed large non-volatile molecules to be ionized efficiently for MS analysis. 18;19 Furthermore, tandem MS (MS/MS) is utilized to identify the amino acid sequence of proteins or proteolytic peptides. The no menclature that describes the different fragmentation patterns observed with MS/MS an alysis is shown in Figure 1-2. Figure 1-2 illustrates the basic theoretical peptides cleavages and nomenclature for basic fragmentation modes. 20 The human genome contains at least 40,000 genes expressing about 100,000 functionally unique protein types. 2;8;9 After considering polymorphisms and PTMs,

PAGE 23

6 H2N H N N H R1 O R2 O O OH R3 x2 y2 z2 x1 y1 z1 H2N H N N H R1OR2O O OH R3 x2y2z2x1y1z1 a1b1 c1a1b1c1 a2b2 c2a2b2c2 Figure 1-2. Common Fragmentati on Patterns and Nomenclature the total number of functionally different prot eins could approach 1 million, especially at different states during the cell cycle. Th e genome map is very useful in providing theoretical polypeptides for automated interpretation of MS data. At first, it was theorized that all the gene pr oducts (proteins) of a particul ar species may be predicted with the genome map; however, it was qui ckly discovered that the genome did not predict the proteins final mature struct ure or temporal prod uction. Although these three technologies were discove red in the last 10 to 20 years, proteomics is still a young field of study with dynamic e xponential growth in the areas of separation science, mass spectrometry, and bioinformatics. Before reviewing these aspects in relation to proteomics, MS analysis of oligonucleotides is introduced. Mass Spectrometry Analysis of Oligonucleotides MS analysis of oligonucleotides has not re ceived the same acclaim in the biological community because of already established pr otocols and instrumentation. For example, large-scale genome sequencing has alr eady been made high-throughput with DNA sequencers. Multiplexed capillary electrophores is is used to separate fluorophore-tagged

PAGE 24

7 oligonucleotide ladders produced by dideoxy chem istry (Figure 1-3), in a similar manner to LC MS/MS allowing for separating and se quencing of oligopeptides ladders (Figure 1 4). The main difference between DNA and proteins is that DNA is static and can be easily amplified with PCR, whereas proteins are expressed over a large dynamic range and must be analyzed at the concentration from the particular state of the cell or system. However, recombinant protein expression te chnology may be used for production of proteins for structural proteo mics (Chapter 2). CapillaryCapillary Sequence Sequence DNA PCR DNA PCR Dideoxy Dideoxy Electrophoresis Alignment Electrophoresis Alignment methodmethod (Sequencer)(Sequencer) Oligonucleotides Oligonucleotides LaddersLadders Figure 1-3. Diagram Desc ribing the Workflow for DNA Sequencing with CE DigestionDigestion Digestion Mass Digestion Mass Database Search Database Search Spectrometry Spectrometry Protein Peptides ProteinPeptides (Sequencer) Identification (Sequencer) Identification Figure 1-4. Diagram Descri bing the Workflow for Protein Sequencing with MS In a very similar fashion to proteins and peptides, the deve lopment of ESI and MALDI has provided similar results for ionization of oligonucleotides for MS analysis. Normally, proteins and peptides are ionized in positive ion mode, thus providing [M + nH] n+ ions. In contrast, oligonucleotide ionizatio n is performed in negative ion mode due to the negatively charged phosphate backbone making [M nH] nions Also, similar to proteins and peptides, oligonucleotides may be fragmented in the mass analyzer for sequencing. Practical ventures for the use of MS for the an alysis of oligonucleotides are presented in Chapter 6.

PAGE 25

8 Complete Mass Spectrometry Analysis The complete MS analysis for structural pr oteomics is presented in Figure 1-5. In the scope of this dissertation, the expressi on, purification, and char acterization of three recombinant forms of human 2 3 sialyltransferase (h23STGalIV) will be described in Chapter 2. Further derivatiz ation experiments of h23STG alIV (hydrogen/deuterium exchange, small affinity labels, and site di rected photoaffinity labels) with the goal of structural proteomic data specific to active si te residues will be presented in Chapters 3 through 5. Before describing the wet chemistry preparation (Figur e 1-5, Selection 1), current analytical tools such as separation science, MS instrumenta tion (ionization to the gas phase and mass-to-charge analysis), and bioinformatics are reviewed below. Since the advent of ESI and MALDI and the completion of the human genome, MS technology has been driven by the biotechnolo gy sector to provide instruments with improved sensitivity, mass accuracy, and duty cycle. It may take several months to develop viable molecular biology and biochemical experiments for a one-day analysis Selection 1: Molecular Biology and Biochemical Sample Preparation Selection2: Chromatography Ionization to Gas Phase Mass-to-Charge Analysis Bioinformatics Instrumentation F e e d B a c k Selection 1: Molecular Biology and Biochemical Sample Preparation Selection2: Chromatography Ionization to Gas Phase Mass-to-Charge Analysis Bioinformatics Selection 1: Molecular Biology and Biochemical Sample Preparation Selection2: Chromatography Ionization to Gas Phase Mass-to-Charge Analysis Bioinformatics Instrumentation F e e d B a c k Figure 1-5. Workflow for Complete MS Analysis

PAGE 26

9 with a mass spectrometer. After MS analysis, the data could take days to months to analyze for useful information unless a viab le bioinformatics system is in hand. In tandem with MS, separation science using 1or 2-dimensional gel electrophoresis (1D GE or 2DGE) or high performance liquid chromatography (HPLC) is needed for chemical noise reduction. Selection or sepa ration may be perfor med off-line or on-line with respect to MS analysis. For HPLC on-line with MS, some view the mass spectrometer as a detector for the chromat ograph. Others view the chromatography as the inlet for MS. Both views are incorrect, because the hyphenation of both sensitive chromatography and sensitive MS truly creates a powerful tool. The understanding of the advantages and disadvantages of current analytical tools allo ws for intelligent design of biology and biochemistry experiments. Separation Science The type of separations needed prior to MS analysis depends on the experiment performed. As described earlier, this may incl ude a single protein to a set of proteins, cell or organelle types, or a complete or ganism with a relatively small proteome. Proteomics has two main strategies for anal yzing proteins, includi ng the bottom-up or top-down strategies. 21 The bottom-up approach includes processing the protein into peptides via chemical or proteolytic dige stion before separation and MS analysis, whereas the top-down approach relies on fragmentation of intact protein within the mass spectrometer for protein sequence. 22;23 Currently, most labs rely on the bottom up approach because of the high cost associ ated with instrumentation needed for the top-down approach. The following secti ons will describe separation science for traditional and the high-throughput workflows in respect to off-line and on-line with MS analysis.

PAGE 27

10 Bottom-Up Approach Top-Down Approach Separation of Proteins (SDS-PAGE) In-Gel Digestion Digestion of Entire Sample Off-line Separation of Peptides (SCX, etc.) Off-line or on-line RP-HPLC Separation of Peptides ESI-MS Analysis MALDI-TOF Analysis Traditional High-Throughput On-Line RP HPLC of Intact Proteins ESI-FTICR-MS (MS/MS) Analysis Bottom-Up Approach Top-Down Approach Separation of Proteins (SDS-PAGE) In-Gel Digestion Digestion of Entire Sample Off-line Separation of Peptides (SCX, etc.) Off-line or on-line RP-HPLC Separation ofPeptides ESI-MS Analysis MALDI-TOF Analysis Traditional High-Throughput On-Line RP HPLC of IntactProteins ESI-FTICR-MS (MS/MS) Analysis Figure 1-6. Bottom-Up and Top-Down Workflows Off-line separation technology The most powerful off-line separation tec hniques was for many years the use of 1 dimensinal or 2-dimensional gel electrophoresis (1DGE or 2DGE) with polyacrylamide gels. 2;8 Proteins may be separated based on size, pI, or both size and pI. Traditionally, proteins are separated with 2DGE and visual ized with a variety of different staining agents. For initial proteomics strategies 2DGE has strong roots with differential displaying of proteins expressed in different cell states (i.e., healthy versus disease) with imaging programs. The most common c hoices for protein visualization with polyacrylamide gels includes silver staining (linear dynamic range (LDR) of 4 orders of magnitude; 500 pg to 5 ng limit-of-detection (LOD), coomassie blue (LDR of 2 orders of magnitude;10 50 ng LOD), or Sypro Ruby (LDR of 3 orders of magnitude; 100 pg to 1

PAGE 28

11 ng sensitivity). 24 To increase sensitivity, the use of Western blotting with specific antibodies lowers the LOD to 0.25 to 1 ng. However, since the antibody interaction is only specific to one type of protein, Western blotting is more expensive than basic staining methods and does not allow for high-throughput. The main problems with 2DGE is that different proteins react uniquely with different stains and that pr otein concentrations varying between 5 to 6 orders-ofmagnitude for different cells and 12 to13 orders-of-magnitude in extra-cellular fluid. Also, 2DGE is not very sensitiv e for proteins of different extr emes such as very acidic or basic, very small or large, and membra ne proteins which may exhibit hydrophobic character. 2 In regards to both functional and stru ctural proteomics, proteins separated with 1D or 2DGE may be digested with common in-gel protease digestion protocols unless silver staining is used. Furthermore, coomassie blue as a staining agent was correlated to higher sensitivity with MS analysis after in-g el digestion with proteases. 25 The one-protein system described in Chapters 4 and 5 used 1DGE and in-gel digestion protocols. After 1DGE analys is and in-gel digestion, anothe r separation step is usually needed to remove salt impurities prior to MS analysis. 8;26 Another common off-line separation tec hnique includes the use of Zip Tips Zip Tips are pipette tips which have been packed with reverse-phase column material or metal chelating agents. 2 Peptide samples which have been digested are loaded onto a small volume of particles, wash ed several times to remove salts or impurities, eluted, and analyzed with MALDI-MS. Furthermore, pa rticles with chelating agents may be bound to divalent cations, termed immobilized metal affinity columns (IMAC), and used to trap

PAGE 29

12 certain species, such as phosphorylated peptid es. As presented in Chapters 4 and 5, Zip Tip technology was used to clean up peptide samples after in-gel digestion. On-line separation technology Reverse phase (RP) high performance liquid chromatography (HPLC) is most commonly used for coupling separations with MS because the solvents (mobile phase) for separations are compatible with ESI (T able 1-3). For effective separation on common RP materials (C 4 or C18), an ion-pairing agent such as acetic acid, trifluoracetic acid, heptafluorobutyic acid, or formic acid mu st be present in the mobile phase to suppress peak broadening. Common solvents for RP chromatography include water, methanol, acetonitrile, and isopropanol. Coupling to ESI is possible because these ion pairing agents and solvents are volatile. Solvents may be delivered with a variety of commercially available pumping systems. Gr adients of increasing percent organic may be delivered at a rate depending on the complex ity of the sample. Proteins and peptides are retained on RP material at high percent aqueous solution, then elute as the percent organic increases. Sample introduction is commonly provided with commercially available injectors or loaded off-line with pressure bombs. Rea listically, most of the instrumentation used for solvent delivery, tubing, zero-dead volum e connectors, and sample injectors are commercially available. The major recent developments for HPLC-MS technology include the downsizing of volumes and flow rates, systems that can handle very high backpressure, different types of column materials, and hyphenation of two different column materials in-line. The tools for separa tion of intact protei ns and peptides were already in hand to accomplish the work descri bed in this dissertation. The choice of a separation method depends on sample complexity and cost effectiveness. Strategies have

PAGE 30

13 already been developed to downsize colu mn size and to move away from 2DGE technology to 2D or multidimensional chro matography. Still, the normal trend in analytical chemistry is to develop sensitive, faster, and cheaper tools. 27 The use of capillary (microbore) columns increased sensitivi ties as described by Wilm and Mann in 1994. 28 Different commercially prepar ed column materials may be packed into capillary columns in-house, t hus greatly reducing the cost for commercial columns, the amount of sample consumed, and the volume injected. 29 Furthermore, two different types of column materials may be p acked together to obtain 2D HLPC. In order to retain chromatography material in the capillary column, sintered silica particles or silicate-polymerized ceramics have been used as frits. 30 Fused silica may be ground down or pulled for creation of a small orific e tip for generation of ESI. The main problems with downsizing chromatography sy stems to 100 nL/min flow rates include increased backpressure, dead-volume, and leak s. Lastly, a voltage potential must be applied through a liquid junction found close to the ESI tip. These issues have been worked out by a variety of labs, thus making capillary HPLC columns common practice. 26;31;32 Since MS is especially well suite d to analyze very small amounts of sample (nano to femtomoles), downsizing to microbore HPLC has been well received. Bottom-up high-throughput functional prot eomics uses two different workflows depending on sample complexity. If the sample is considered to be of low to moderate complexity, 2-dimensional high-perform ance liquid chromatography (2D HPLC) may achieve the separation needed (Figure 1-7). 8;26;27 This is performe d by using capillary electrophoresis (CE), size exclusion (SE), strong anion or cation exchange (SAX or SCX), or IMAC support phase up-stream from RP separation. The initial column is

PAGE 31

14 loaded with sample, followed by a step gradie nt elution with the sp ecific salt buffer for that column material. The solvents normally associated with columns materials other than RP use buffers that are not compatible with ESI. For each step gradient cycle, a RP column down-stream retains the peptides eluted from the first separation. Next, a switch valve changes the flow direction and the mobile phase to elute the peptides from the RP material. 33 Some examples of 2D HPLC include MudPIT 34 and peak parking. 35;36 If a sample is too complex, samples are pre-fractionated into small samples prior to on-line HPLC-MS. As opposed to 2DGE, strongly acidic, strongly basic, extremely small, extremely large, and hydrophobic protei ns and peptides may be analyzed. The most common analytical set-up for maximum sensitivity for peptide analysis includes the use of 2D chromatography with a C 18 trapping column and a C18 analytical column. The HPLC-MS system used for analysis of the in-gel digested sialyltransferase samples presented in Chapters 4 and 5 included a C 18 trapping column and a microbore analytical column. SCX Material RP Material Capillary opening into MS Liquid-metal interface Mobile Phases SCX MaterialRP Material Capillary opening into MS Liquid-metal interface Mobile Phases Figure 1-7. Microbore MudPIT Ionization Techniques: ESI and MALDI Although there are many ionization methods available, ESI and MALDI are the primary methods for analysis of proteins and peptides with MS. ESI is an atmospheric pressure ionization event, after which ions are moved into low or high vacuum regions. The mechanism of ion formation is still unde r discussion; however, the most widely

PAGE 32

15 accepted theory for ionization of large molecu les includes Doles charged residue model (Figure 1-8).37 First, the voltage is increased at th e tip of the ESI needle or the capillary opening into the MS, thus forming a Taylor c one with droplets of re latively high surface charge density. At a certain high voltage, th e Taylor cone is dest royed, thus destroying electrospray. Second, the carrier solvent evap orates, causing the droplet to shrink, thus creating high charge-to-mass character. Ne xt, the droplets eject charged particles because the critical Rayleigh limit has been reached. Finally, the desolvation process continues with assistance from a heated capillary inlet until individual ions form. Figure 1-8. Electrospray Ioni zation with Doles Large Molecule Ionization Model Over the last 15 years, optimization of th e flow rate, the applied potential, and the tip diameter with respect consistent MS signa l have proceeded. For higher flow rates (2 L/min to1 mL/min), a nebulizing (drying gas) is used to assist with ionization; however, high flow rates may be detrimental to pumping systems. At lower flow rates, termed nano-ESI (nESI), no drying gas is present. Nano-ESI tips are normally associated with capillary chromatography and have very small tip diameters (2 to 50 m). Greater ESI Needle (~3000 V) + + + + + + + + + + + + Droplet with high charge-to-mass character + + + + Droplet at Critical Rayleigh ESI Needle (~3000 V) + + + + + + + + + + + + Droplet withhigh charge-to-mass character + + + +Droplet at Critical Rayleigh ESI Needle (~3000 V) + + + + + + + + + + + + Droplet withhigh charge-to-mass character + + + +Droplet at Critical Rayleigh ESI Needle (~3000 V) + + + + + + + + + + + + Droplet withhigh charge-to-mass character + + + +Droplet at Critical Rayleigh

PAGE 33

16 sensitivity for analysis of biological molecules is observe d experimentally with nESI. 38;39 First, the low flow rate associated with nESI produces a smaller Taylor cone which permits the emitter to be placed much closer to the MS inlet, thus effectively increasing sampling efficiency. Second, nESI generates smaller charged droplets, therefore efficiently increasing desolvation rate. Thir d, much less solvent is delivered to the mass analyzer thus greatly reducing the solvent lo ad to the vacuum system. Fourth, nESI consumes much less analyte as compared with conventional ESI. Finally, micobore HPLC-nESI-MS increases the sensitivity of the instrument due to concentration of sample on the RP material. In respect to on-line HPLC-MS, some species may co-elute. With ESI, multiple analytes co-eluting will create a competiti on between each of the analytes. Based on studies using dyes, hydrophobic molecules will collect on the surface on the ESI droplet, whereas hydrophilic molecules will be found in a heterogeneous mixture within the droplet. 40 Therefore, an analyte with increase d hydrophobicity is more likely to winout during the ionization event. The problem of co-eluting analytes may be solved with changing the gradient of the mobile phase de livered to the column, changing ion-pairing agents in the mobile phase, or changing the scan functions of the MS. Electrospray has the advantage of impart ing multiple charges onto proteins or peptides, thus bringing high molecular weight molecules within the m/z range exhibited by modern mass analyzers 41;42 In positive ion mode, polypep tides with basic amino acids will carry positive charge. On average, every 10 amino acids within a polypeptide has an ionizable group for protonation. 41 Furthermore, the protease trypsin cleaves on the Cterminal end of arginine and lysine, thus leaving an amino acid with an amine for

PAGE 34

17 protonation. Changes in analyte solvent, pH, and temperature produce different charge states of proteins obser ved with MS analysis. 39;43 Knowing the primary sequence of the prot ein will allow a maximum number of protonation sites to be calculated; however, the pI of a prot ein or peptide does not relate to ionization efficiency. Furthermore, the fold state of the protein will affect charge state distribution. Finally, to determin e the charge-state of the prot eins or peptides, the inverse of the mass difference between peaks of the isotopic distribution is calculated. 41 Table 1-4 represents the signal intens ity of three different proteins versus ion-pairing agents. This table provides the basis for solvent syst em chosen to analyze intact proteins in Chapter 3. Negative ion mode is used more frequently for oligosaccharides, oligonucleotides and non-standard experiments. For negative ion mode, all of the polarities applied to the hardware of the instrument must be reversed for ion generation and detection. Often, in negative ion mode, arcing occurs if the per cent aqueous solution is above 40%. Table 1 4 illustrates the solvents compatible for posit ive and negative ion mode. Note that most of these solvents are compatible with RP HPLC. Table1-4. Solvents Compatible with ESI pH acetic acid, formic acid, trifluoroacetic acid* for positive-ion detection (0.1 % up to 3.0 % v/v) Buffers, ion pair-reagents Ammonium acetate, ammonium formate, triethylamine heptafluorobutyric acid (HFBA), te traethyl or tetrabutylammonium hydroxide. (10 100mM) Cation Reagents Potassium or sodium acetate (20 50 uM level) Solvents Methanol, ethanol, propanol isopropanol, butanol, acetonitrile, water, acetic acid, formic acid, acetone, dimethylformamide, dimethyl sulfoxide, 2-methoxy ethanol, tetrahydrofuran, dichlormethane, chloroform

PAGE 35

18 MALDI ionization proceeds with N 2 laser (337 nm) short pulse irradiation of analytes distributed in a matrix thus crea ting a plume of neutrals, clusters, and ions. 44 Similarly to ESI, the mechanism of ionizat ion is not fully understood. As opposed to ESI, MALDI predominantly creates singly charged ions. The common matrices are chromophors, therefore they absorb ener gy after irradiation and, according to the prevalent ionization theories, provide a source for protons or other cations to aid in ionization. According to the current ionization models, the mechan ism for ionization is most likely specific to the type of analyte be ing investigated. So me current ionization models suggest that peptides or proteins ar e pre-protonated in the matrix because of the low pH preparation protocols. 45 Other models suggest that neutral ion-ion reactions within the post-ablation plume creates ions. 46 These ionization reacti ons occur in the first tens of nanoseconds after irradi ation within the in itial desorbing matr ix/analyte cloud and have initial velocities between 300 and 800 m/s. Nanoto picomoles of proteolytic peptides are routinely analy zed. For the purpose of proteomics, different studies require different sample prot ocols found in the literature. Typically, 1 L of 500 nM to10 M of protein or proteolyti c peptide sample is mixed with 5 to 10 L of 10 mM matrix (2,5-dihydroxybe nzoic acid (DHB); 3,5-dimethoxy-4hydroxycinnamic acid (sinapinic acid); or -cyano-4-hydroxycinnamic acid ( -CHCA) dissolved in 50% ACN 0.1 % TFA), after whic h it is spotted onto a sample support (i.e., stainless steel plate). MALDI may be performed at atmospheric pressure; however, ionization is most commonly performed under vacuum. This increases the duty cycle because the ionization source must be loaded and pump ed down before the experiment can be

PAGE 36

19 performed. Compared to ESI-MS, MALDI-MS is more tolerant to detergents, salts and buffers; however, there is more ion suppre ssion when a complex sample is present. 8 In the work presented in Chapters 4 and 5, both ionization methods were used with different mass spectrometers for analysis of intact proteins and peptides. Sample Plate UV Laser Pulse Desorption Plume = Matrix = Analyte Sample Plate UV Laser Pulse Desorption Plume = Matrix = AnalyteFigure 1-9. Diagram Representing Matr ix-Assisted Laser Desorption Ionization Mass Analyzers Mass-to-charge analysis may take place in one of many mass analyzers developed over the last 108 years. The first mass analyzer was built by J.J. Thompson at Cambridges Cavendish Labor atory in 1897. Since then, MS has provided the tools necessary for characterization of small synthetic molecules to whole viruses. 47 Although many mass spectrometers are manufactured to analyze various types of small and large molecules, discussion is limited to instrument s for which the primary use is analysis of proteins, peptides, or oligonucleotides. Th ese include quadrupole ion traps (QITs) MS, time-of-flight (TOF) MS, and Fourier transf orm ion cyclotron res onance (FT-ICR) MS. 48 Over the last ten years, ample commercial instruments were developed to accommodate ESI and MALDI sources. Also, several instrume nts which have different combinations of

PAGE 37

20 the individual mass analyzers, i.e. hybrid MS instruments, have been made commercially available. Quadruple ion trap MS The QIT may be considered one of the mo st robust mass analyzers on the market. The QIT was originally conceived by Wolfga ng Paul (Nobel Prize in Physics, 1989) and co-workers in the early 1950s and made commercially available in 1983 as a GC detector. Over the last 50 years, many la boratories have improved cell design and ion transfer optics to make the instrument av ailable to a variety of investigators. Operationally, ions may be generated externa lly, after which they ar e guided and trapped inside a cell between two end pl ates and a ring electrode. Once trapped inside the cell, a radio frequency is applied to the ring electrode and ramped with increasing amplitude, thus ejecting the ions in the direction perpe ndicular to the ring el ectrode. The ions hit a conversion dynode, thus ejecting electrons to an electron multiplier for amplification of signal. Different ions of different m/z have different stab ilities within th e radio frequency and electric fields. Ions of lower m/z are ejected first as the radio frequency is ramped. 49 The real power of QITs, in terms of proteomics, is the fast scan rates which are important for detection of peptides after RP -HPLC. Multiple MS scans with QITs differ from the other instrument types because it has the ability to fragment ions and store them for several cycles. Scan functions are designed to operate in what is called datadependent mode. In data dependent mode, af ter detection of a relatively abundant m/z, a subsequent scan isolates that particular m/z prior to MS/MS analysis. Tandem MS is performed by increasing the pressure in the QIT cell by introducing a pulse of helium, thus creating what is termed collisionally ac tivated or collisionally induced dissociation (CAD or CID). 49

PAGE 38

21 In respect to proteomics, low energy CI D predominantly fragments only the amide backbone of the polypeptide. Fragmentation is dependent on the polypeptides sequence, length, and charge-state. In general, different peptides may or may not completely fragment. Often, intact peptides or fragmented peptides may loose H 2O, NH3, or CO2. Furthermore, it has been observed that doubl y-charged ions generally provide improved MS/MS spectra. For QITs, mostly band ytype ions are observed with CID (Figure 1 2). Rearrangements of peptides (intermolecu lar reactions) in the ga s phase been reported in the literature, however they are not reviewed in this dissertation. 20 There are many scan functions available for different MS experiments, specifically product ion scan, precursor ion sc an, and constant neutral loss. 8 For example, some PTMs such as phosphorylation have less stable bonds than the backbone of a polypeptide, therefore when MS/MS is performed, th e major peak observed may be [M 98] +1 or [M 80] +1 A scan function may be set to select for the neutral loss of phosphate, followed MS/MS (considered MS/MS/MS or MS 3 ) on the m/z that is missing the phosphate. The real limitations of the instrument incl ude poor quality data sets due to space charging events and a limited m/z range of 20 to 2000 m/z. Space charging is an effect that is observed when too many ions are tra pped in the cell, thus causing ions to leak out of the cell by repulsion. To eliminate space charging, automatic gain control (AGC) is a function that monitors the amount of ions in the cell. Before mass measurement, a pre-scan measures the amount of ions entering the cell (total ion count). Depending on the counts, the ion collection time for mass analysis may be adjusted from 10 to 25 microseconds to allow the proper amount of ions to be collected within the linear dynamic range of the cell. 49

PAGE 39

22 QITs are relatively sensitive; however, reso lution may be from 1 to 0.1 mass units. Calibration may be performed externally wi th common mass errors between 50 to 500 ppm for proteomics. Often, data sets collect ed for proteomic studies are of poor quality because of space charging events due to co-e luting peptides. Internal calibration may improve mass accuracy; however, the internal calibrant ma y inhibit ionization of a peptide of interest. Finally, since QITs ar e scanning instruments, MS and MS/MS data on certain peptides may be excluded from a pa rticular LC-MS analysis. Currently, QITs equipped with capillary HPLC nESI have pr ovided the best results in respect to the capital cost for proteomics. The data presen ted in Chapters 4 and 5 was collected with this type of mass spectrometer. Time-of-flight MS Time-of-flight (TOF) MS was introduced by William E. Stephens at the University of Pennsylvania in 1946 and made commercially available in the late 1950s by William C. Wiley and I. H. McLaren at the Bendix Corporation Detroit, Michigan. TOF MS separates ions of different m/z based on the kinetic energy as they drift down a tube at different velosities. For many years, TOF in struments were not considered useful mass spectrometers because they lacked the reso lution needed for routine analysis. The resolution of TOF mass analyzers was improve d by the introduction of delayed extraction (DE) which normalized the initial veloci ties of analytes created by a MALDI. 49 Ionized molecules are slowed with DE (~2000 V) and then pulsed (~20,000 V) to a detector at the other send of the tube. TOF MS instruments may be operated in two different modes. The first, linear mode, has a mass range of up to about 1 M Da but often suffers from poor resolution. The second, reflectron mode, suffers from a limited mass range (100 6000 m/z), but has

PAGE 40

23 an improved mass accuracy (~5 ppm). Ionized molecules are slowed with DE and then pulsed down to a series of focusing plates (reflectron) which conc entrate and redirect the ions to another detect or. Traditionall y, a MALDI source is used with TOF technology for analysis of intact proteins and peptides due to the fact that MALDI predominately creates singly charged ions. The analysis of singly charged, high mass proteins is best suited for TOF mass analyzers operated in linear mode if poor resolution is not an issue. TOF is the easiest MS tec hnology to use due to the fact that it does not suffer from the space charge effects observed with ion trap MS. Peptide sequences are observed from the fragmentation of metastable ions during flight down the TOF tube, thus creating post s ource decay (PSD). First, a spectrum is collected in linear mode, thus allowing the identification of the m/z for the peptide of interest. Next, metastable ions which undergo fragmentation during flig ht are analyzed in reflectron mode, thus yielding a spectrum with ions of less m/z than observed in linear mode. PSD spectra may not be as straight forward to interpret as the MS/MS spectra obtained from QITs because fragmentation ma y not be as complete. Therefore, quality MS/MS data is normally associated with QITs. 24 Over the last 10 years, TOF technology has also integrated ESI sources. The current wave of ESI-TOF-MS instruments ha s excellent scan rates, mass accuracy, and mass resolution in comparison to their predece ssors. Hybrid instruments which contain a QIT prior to a TOF (Q-TOF MS) are also comm ercially available. These types of hybrid instruments take advantage of the CID capabilities with QITs and the higher mass resolution capabilities of reflectron TOF MS.

PAGE 41

24 MALDI-TOF has been used in forensics and homeland security to identify possible peptide or protein mass markers in respect to bioterrorism. 50 Also, MALDI-TOF has been used for imaging different pr oteins in cells and organelles. 51 Furthermore, MALDI TOF has strong roots in peptide mass fingerprint ing of in-gel digested proteins separated with 1D and 2DGE. 8;52 This instrument was used to monitor peptide digestion as presented in Chapters 4 and 5. Fourier transform ion cyclotron resonance MS The concept of ion cyclotron resonance (I CR) in a magnetic filed was developed by E.O Lawrence and co-worked in 1931, followed by electrostatic c onfinement by F. M. Penning in 1936 in a trap, later known as a Pen ning Trap. M. B. Comisarow and A. G. Marshall revolutionized ICR in 1974 by using Four ier transform (FT) with ICR analysis. FT allowed several ions to be measured in the cell at one time, instead of only measuring one ion at a time. Since then, several labor atories have continued to improve FT-ICR MS instrument design by increasing mass range, mass resolving power, and sensitivity to produce the highest resolution and sensitivity possible for commercially available mass spectrometers. Several optics are needed to accelerate th e ion beam past the fringe field of the magnet (Figure 1-10) which encases the penning trap. Ions are generated externally, after which they are pulsed into and trapped in th e Infinity cell. Once trapped in the cell with DC voltages applied on the end plates, the ions go into cycl otron and magnetron motion due to a high magnetic field being present. Figure 1-11 illustrates the cyclotron motion of a positive or negative ion in a magnetic field where B is the magnetic field in

PAGE 42

25 Tesla, m is the unit mass of an i on, q is the charge of the ion and c is the angular velocity in radians/second. Each m/z has a characteristic cyclotron frequency independent of the radius and velocity; however, the ion cyclotron freque ncy is inversely related the m/z, thus resolution decreases as m/z increases. Figur e 1-15 represents the equation governing the relationship between frequency and m/z where is the magnetic field in Tesla, m/z is the unified mass-to-charge of an ion, c is the angular velocity in radians/second, and c is the velocity in Hertz. Figure 1-10. Actively Shielded 4.7 T Ma gnet at the University of Florida + B B + B B Figure 1-11. Cyclotron Motion of Ions in Magnetic Field

PAGE 43

26 c 1.535611 x 107 c1.535611 x 107 c= c= = m/z = m/z Figure 1-12. Relationship between the Fr equency and Mass-to-Charge of an Ion Signal cannot be obtain by just measuri ng the induced magnetron and cyclotron motion of an ion in a magnetic field. The molecules are excited with a sweeping radio frequency applied by two opposing plates, t hus sending the spatially coherent ion population into a larger cyclotron orbit. As the orbit decays back to the center, two other parallel plates measure the differential image current through induction. Following measurement, the time-domain signal is F ourier transformed to the mass-to-charge domain. Since the frequency may be measur ed very precisely, FTICR MS offers very high resolution. Theoretically, at ultra-low pr essures, the observat ion of signal may be indefinite for ion measurement and re-measur ement. Also, the signal is quantifiable because transduction responds linearly to the nu mber of ions in the cell. Finally, since ICR frequencies for ions are observed between 1000 to 1,000,000 Hz, already commercially available electronics are incorporated into instrumentation. 53;54 The main limitations with FTICR MS are the purchase and maintenance costs. To purchase a basic FTICR MS with a 4.7 T ma gnet and infrared multiphoton dissociation, an investigator must relinquish $ 750,000. In cluding capital cost, maintenance costs are also high. First, common magnets (4.7, 7 or 9 T) require constant N 2 and liquid He re filling for maintaining the high-magnetic field. Secondly, sophisticated pumping systems with 2 to 3 turbomolecular pumps and 3 to 5 roughing pumps are needed to maintain a cell pressure of at least 10 -10 torr for consistent behavior of ions in the cell. If the pressure is greater than 10 -10 torr, gases in the cell dampen the ion cyclotron decay time, thus destroying the sensitivity and mass resolution provided by the instrument. External

PAGE 44

27 calibration provides mass accuracies of less than 5 ppm as long as the number of ions in the cell remains relatively consistent between scans. If the number of ions change dramatically or if there are too many ions within the cell, space charging will shift the m/z values by repulsion or coal escence of ions packets. Tandem MS analysis may be performed w ith one of many different fragmentation methods. 8 These include CAD, electron capture dissociation (ECD), sustained offresonance irradiation (SORI) fragmentation, infrared multiphoton dissociation (IRMD), or blackbody infrared radiative dissociation. Besides CAD, as presented with QITs, these methods may produce different fragmentati on patterns based on the energy associated with the fragmentation method. Fundamenta lly, the most important dissociation method for proteomics includes the use of ECD of int act proteins. As presented in the separation science section, the top-down method only requires separati on of intact proteins prior to MS/MS analysis. ECD, as initially pres ented by McLafferty et al., fragments intact proteins though the tr ansfer of electrons. 23 The energy released upon transfer of electrons to intact proteins will frag ment the polypeptide backbone to create cand z-type ions (Figure 1-10). 55 This fragmentation method produces information rich spectra; however, interpretation has been problematic. For proteomics, if funding is not an issue, the benefits of FTICR MS are superior to other types of mass spectrometers Identification of peptides with FTICR MS often relies on exact mass; however, mass redundancy due to peptides with the same amino acids in different sequences may provide unambiguous results. In terms of sensitivity, 1000 proteins in the 2 -100 kDa molecular ma ss range were identified from 200 -300 nanograms of an E. coli cell extract with LC-ESI-FTICR-MS. 56 Furthermore, a study on

PAGE 45

28 cytochrome c reported that 10 zeptomoles (10 -21 moles) was efficiently measured. 57 For real world samples, 9 x 10 -18 moles of cytochrome C, repr esenting about 1 % by weight of proteins in blood cells, was extracted from crude red blood cells and analyzed with LC-ESI-FTICR-MS. Furthermore, a diffe rence of 9.5 mDa (~200,000 resolution) was achieved to separate phosphorylation (31.9816 Da) and sulfonation (21.9721 Da) PTMs. 58 As with Q-TOF MS, hybrid instruments with a QIT added before a FTICR MS (Q-ICR MS) are commercially available, thus providing an added degree of mass selection. Experiments explained in Chapte r 3 use ESI FTICR MS to analyze simple intact protein samples. Bioinformatics As with the growing field of proteomics, bioinformatics is also loosely defined. Over the years, the term bioinformatics ha s included software written for molecular modeling, gene prediction, sequence alignm ent, bio-molecule array design, mass spectrometry data analysis, amino acid sequence prediction, and laboratory informatics. Furthermore, several disciplines such as computational biology, biomathematics, biometrics, and biostatistics form specialized bioinformatics platforms. 7;59 Bioinformatics software are designed for visualization, databases, knowledge representation, software developm ent, and algorithm development. 7 Computer science experts are often hire d to aid in the la rge task of data management because the high quality interp retation of ample MS data represents a serious bottleneck in proteomic workflows. 7 The development of new MS technologies for sequencing peptides i.e., ECD with ESI-FTICR MS, created the need for new algorithms for automated interpretation. A lthough there are several applications for bioinformatics, only those pert aining to analysis of proteo mic MS data are reviewed.

PAGE 46

29 Furthermore, programs used for qualification of proteins and peptid es, not quantitation, are reviewed. Currently, there are three main strategi es for interpretation of proteomic MS data. 1;18 The first two includes programs written to compare experimental proteomic MS data with theoretical in silico digested polypeptides based on the knowledge of a certain genome. Functionally, in silico mass lists of possible theore tical peptides derived from theoretical proteins as determined from a known open reading frame from a genome sequence may be predicted based on the specificity of proteoly tic or chemical cleavage. The experimental peptide m/z list (mass list extracted from experime ntal MS data) of a single digested protein is compared with a theoretical peptide mass library generated from genomic information from an open access databases. 7 Positive identification is possible if the genome of the species is found in a genomic database. Table 1-5 represents a list of common databases contai ning different species genomes. The third method, de novo sequencing and identification, relie s entirely on mass differences generated with experimental MS/MS data for the sequencing and identification of polypeptides. Table 1-5. Common Genomic Databases Database Name Wprl-Wide Web Address EMBL http://www.embl-heidelberg.de/ GenBank http://www.ncbi.nlm.nih.gov/GenBank/ SRS http://www.srs.ebi.ac.uk/ ExPaSy http://www.expasy.ch/ YPD http://www.proteome.com/YPDhome.html MIPS http://mips.gsf.de/ Intact peptide MS data analysis with in silico digestion algorithms In silico data analysis, in its simplest form, can find its roots with the peptide mass fingerprinting algorithm (PMF) originally presented in 1984 with fast atom bombardment

PAGE 47

30 ionization (FAB) MS. 1;60 FAB did not produce reliable results and was by no means high-throughput; however, the trend to perform in-gel di gestion followed by MS analysis also had its start with FAB MS. Nine years later, after the advent and improvement of MALDI TOF-MS, Hanzel and co -workers illustrated that PM F may be used to analyze a variety of gel spots in a semi high-throughput fashion. 24 Peptide mass fingerprinting is useful for the identification of 1 to 3 protei ns at a time because MALDI-TOF analysis of multiple protein digested samples may lead to ion-signal suppression, overlap of signal corresponding to different peptides, and mass redundancy. A database search is performed on an experimental peptide m/z lis ts with one of the programs presented in Table 1-6 by defining the pI, protease used (cleavage specificity), the number of possible missed cleavages by the protease defined, the mass tolerance (mass error) of the instrument, the minimum number of peptide matches required for identify a protein, and the species from which the protein is extracted. 24 Including these parameters, several common amino acid in vivo PTMs or in vitro chemical modifications may be defined. Tandem MS data analysis with in silico digestion algorithm Five different groups in 1993 independently developed software for interpretation of experimental MS/MS data, including W.J. Hanzel and coworkers, C. Wantanabe, D. J. Pappin and coworkers, P. James and cowork ers, and J. R. Yates III and coworkers. 7 The algorithm developed may be termed expr ession sequence tag (EST), sequence tag identification, peptide mass tag (PMT), or pe ptide fragmentation fingerprinting (PFF). These strategies were developed due to the huge amounts of information rich data generated by ESI MS/MS. 24 Unlike the PMF method which analyzes intact peptide m/z, the EST method is used to analyze fragmentation peptide spectra. With EST, the

PAGE 48

31 theoretical in silico peptides are fragmented into po ssible masses representing the peptide ladder of that particular peptide for comp arison to experimental MS/MS data. The biggest problem associated with auto mated data analysis is the incorrect assignment of band ytype ions. Upon fragmentation, peptides may undergo rearrangements or lose H 2O, CO2, and/or NH3. Furthermore, a precursor ion may lose H 2O, CO2, or NH3 or internally fragmented upon ionizat ion. If an intact peptide (parent ions) undergoes these types of mass loss, fragmentations, or intermolecular rearrangement, it will not be identified becau se the precursor or MS/MS ion mass does not match the theoretical in silico m/z. 60 For programs using the EST algorithm, experimental MS/MS m/z lists, protease specificity, number of missed cleavages, instrument mass tolerance (mass error), ta xonomy, and variable and fixed amino acid modifications are defined. 60 Currently, Mascot and Sequest are the main packages used to aid in interpretation of proteomic MS data. Tandem MS data analysis with de novo algorithims If the genome of a species is unknown, de novo sequencing may provide improved results over in silico algorithms; however, de novo sequencing algorithms will not always provide correct assignmen t of peptide sequences. 7;24 Older de novo sequencing programs include the generation of all theoretically possible sequences. Experimental MS/MS m/z lists are compared against theoretical spectra for the best matchs. Using this methodology provides many possible solutions b ecause the number of sequences grows exponentially as the length of the polypeptide increases. The newer programs simply derive a theoretical sequence based on the experimental MS/MS m/z list. Currently, de novo peptide sequencing from MS/MS data is computationally laborious. Often MS/MS spectra lack enough high quality peaks for

PAGE 49

32 assignment of a complete peptide sequence. A new dissociation method, electron transfer dissociation (ETD) with linear i on traps (LIT), produce cand ztype ions which appear to have well defined fragmentation patterns. 61 ETD may provide improved MS/MS spectra for de novo sequencing; however, two main proble ms still exist for this type of sequence alignment. First, several am ino acids and amino acid pairs, i.e. leucine/isoleucine or glutamine/lysine, ha ve the same or similar nominal masses. Second, cleavage does not always occur at every peptide bond. For example, some fragmentation ions are below the noise level, the C-terminal side of proline is often resistant to cleavage, mobile protons are of ten absent, and peptides with free N-termini often lack fragmentation between the first and second amino acids. Software packages for analysis of MS and MS/MS data Table 1-6 lists many of the programs avai lable for MS and MS/MS data analysis, with Sequest and Mascot being the most common. 62;63 Sequest was the first commercially available software package for of proteomic MS/MS data analysis as developed by J. Yates III and co-workers in early 1994. Sequest does not rely on a data preprocessing step because the algorithm is based on cross correlation between experimental MS/MS data and an in silico MS/MS spectrum. Mascot was developed in collaboration by D. Pappin at Imperial Canc er Research Fund, UK and A. Bleasby at the SERC Daresbury Laboratory, UK under the name MOWSE. After further development with the search algorithms, Matrix Sc ience (Boston, MS) purchased the MOWSE software to distribute it under the new name Ma scot. In contrast to Sequest, Mascot requires a pre-interpretation step to create dta files which contain the experimental MS/MS m/z lists. Mascot relies on the probability of random matches between theoretical and experimental m/z lists. Sequest and Mascot utilize the known specificity

PAGE 50

33 of the protease, the parent ion mass, a nd the partial sequence information from experimental MS/MS data for searching against the known genomic databases. Furthermore, several common amino acid in vivo PTMs or in vitro chemical modifications may be defined. Table 1-6. Common MS and MS/MS Pe ptide Data Analysis Software Method Name Web-Site PMF Mascot MS-Digest http://matrixscience.com http://prospector.ucsf.edu/ucsfhtml3.4msdigest.htm MS-Fit http://prospector.ucsf.edu/ucsfhtml3.4msfit.htm ProFound http://prowl.rocke rfeller.edu/cgi-bin/ProFound PepIdent http://www.ex aspy.ch/tools/peptident/ PMT Mascot http://matrixscience.com PepFrag http://prowl.rockerfeller.edu/prowl/ MOWSE http://hgmp/mrc.ac.uk/Bioinformatics/Webapp/mowse/ MS-Seq http://prospector.ucsf.edu/ucsfhtml3.4msseq.htm MS-Tag http://prospector.ucsf.edu/ucsfhtml3.4mstagfd.htm MultiIdent http://www. exaspy.ch/tools/multident/ PepMAPPER http://wolf.bi.unmist.ac.uk.mapper Sequest http://field s.scripps.edu/sequest/ De Novo DeNovoX not available as shar eware on the world wide web SEQPEP not available as shareware on the world wide web SeqMS http://www.protein.osakau.ac.jp/rcsfp/profiling/SeqMS.html Sherenga not available as shareware on the world wide web PEAKS http://www.bioinformaticssolutions.com/products/inde x.ph Including the open access software to analyze proteomics data, different MS vendors have created their own proteo mics solutions software packages. 60 The quality of the data analysis software varies from inst rument to instrument as does the quality of instrument performance. 31 Due to the diversity a nd size of the MS industry, standardizing data files has been difficult. Currently, th e dta file format has provided somewhat of a standard for the open access plat forms. Most MS vendors provide ways to extract data into dta files so they can be analyzed with Mascot.

PAGE 51

34 Bioinformatics data visualization and interpretation Once experimental MS/MS data have been processed with the chosen algorithm, protein identification candidates are generated with an associat ed score that represents the likelihood that an experimental precursor-ion and partial MS/MS m/z list or spectrum match the theoretical in silico digested peptide mass and peptide ladder. The higher the score, the more likely the pr otein is a positive identification. Also generated with the output are the numbers of peptides identifie d which pertain to a particular proteins identification, also know as percent sequence coverage. Typically, the higher the percent sequence coverage, the higher the likelihood that a particular protein may be considered a real candidate. 60 For the applications presented in Chapters 4 and 5, Sequest and Mascot outputs are not used to identif y a protein. The MS and MS/MS data in these cases are used to identify amino acids which have been derivatized in vitro. Furthermore, the sequence of the protein presented in this dissertation is already known. The main differences between the programs presented in Table 1-6 are the way the protein candidates are scored and the way the data is visualized. After bioinformatics analysis of proteomic MS/MS data, it is th e investigators job to provide realistic interpretation of what statistics has provide d. After receiving an output from automated analysis with bioinformatics software, the data requires manual MS/MS peak verification to avoid false identification of proteolytic pe ptides. Furthermore, the instruments mass accuracy is very important for correct identif ication. Positive identification of peptide candidates will have a mass erro r of less than 20 ppm, with an upper threshold for good statistics around 200 ppm. For example, if th e mass error thresholds are increased during bioinformatic analysis of MS data, there will be an increased chan ce of false positives. 60 Depending on the instrument, the mass tolerance of the instrument should be taken into

PAGE 52

35 consideration. FT ICR-MS data should have very high mass accuracy as compared to a QIT-MS data. Statistics is very usef ul for making a decision, but should never overshadow careful thinking. Summary The tools available for analysis of pr oteins and oligonucle otides have been presented. In the scope of this dissertation, the following chapters present data related to the wet chemistry prior to and including MS analysis. First, Chapter 2 describes the expression and purification of si alyltransferases needed for st ructural proteomic studies. Second, Chapter 3 describes the use of hydr ogen-deuterium exchange high-performance liquid chromatography electrospr ay ionization Fourier transf orm ion cyclotron resonance mass spectrometry analysis with sialyltransf erase for studying secondary structure. Third, Chapter 4 describes the use of small mo lecule labeling agents for derivatization of sialyltransferase, followed by mass spectrome try analysis for identifying active-site amino acids. Fourth, Chapter 5 describes th e site-directed photoaffinity labeling of sialyltransferase followed by mass spectrometry analysis for reveali ng the identity of active-site amino acids. Fifth, Chapter 6 describes ion-pairin g reverse-phase highperformance liquid chromatography negative ion mode electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry for identification of oligonucleotides selected with nanoharvesting agents. Fina lly, Chapter 7 describes the conclusions and future work. The entire wo rk is performed with an interdisciplinary mind set towards structural proteomics as a small component of systems biology.

PAGE 53

CHAPTER 2 EXPRESSION AND PURIFICATION OF RECOMBINANT HUMAN ALPHA 2 3 SIALTRANSFERASE Introduction to Oligosaccharides and Glycosyltransferases Including proteins and oligonuc leotides as presented in Chapter 1, oligosaccharides compose another important class of bio-mol ecules. Compared to oligonucleotide and protein linear structure, oligosaccharides may be covalen tly linked in a great diversity of branched patterns. During th e recent omics nomenclature revolution, the glycome has been defined as the whole set of carbohydrate molecules in an organism. 64 Complex oligosaccharides provide many functions ra nging from energy storage to information transfer systems that control cell function. 65 Oligosaccharides may be attached to a variety of different proteins and lipids through selective glycosidic bonds. Even when including phosphorylation, glycosylation is the most common PTM found in living cells. The complicated glycome has added an ex tra dimension to the already difficult proteomics problem. To study Glycomics, invest igators have focused on the structures of the glycans themselves, the binding partners to the glycans, or the enzymes needed for building and degrading glycan structure. This chapter presents background on the glycosyltransferase family and the sial yltransferase sub-family, followed by the description of recombinant human (2 3) sialyltransferase expression, purification, and characterization. Purified sial yltransferase will be bioconjug ated prior to proteomic MS analysis as described in Chapters 3 through 5. 36

PAGE 54

37 O H HO H HO OH OH H H H OH O H HO OH H OH OH H CH3 H H O H HO H HO H OH H H OH NHAc O H HO H HO OH OH H H H OH O H HO OH H OH OH H CH3H H O H HO H HO H OH H H OH NHAc Mannose Fucose N-Acetylglucosamine Mannose Fucose N-Acetylglucosamine (Man) (Fuc) (GlcNAc) (Man) (Fuc) (GlcNAc) O OH HO H H OH NHAc O OH HO H H OH NHAc O OH HO H OH H OH O OH HO H OH H OH HH HAcHAc HH HH NN HH OH H OH H CO2H CO2H H OHH OH H H H H O H HO H OH H R H O H HO H OH H R H N-Acetylgalactoseamine Galactose Sialic Acid N-Acetylgalactoseamine Galactose Sialic Acid (GalNAc) (Gal) (Sia) (GalNAc) (Gal) (Sia) Figure 2-1. Common Sugars Found in Oligosaccharides Complex Carbohydrates in Biological Systems Complex carbohydrates, also know as oligosa ccharides or glycans, may decorate different proteins or lipids to generate glycoproteins or glycolipids. The glycosidic linkage to proteins may be O-linked thr ough serine, threonine, or hydroxylysine or Nlinked to asparagines at specific amino acid sequons. 52;66 O-linked glycans may be found at Cys-X-X-Gly-Gly-Ser/Thr-Cys sequons with X representing any amino acid, however, other as yet defined sequons may be possible. It has also been found that a proline +1 or -3 from the Ser/Thr make glycan attachment favorable. N-linked gl ycan sites are well defined and more abundant than O-linked gly cans. N-linked glycans are attached to asparagines contained in AsnXSer sequons where X can be any amino acid except proline; however, an AsnXSer sequon does not necessarily mean that glycosylation will be present. A third linkage for glycans includes attachment to Cys, Asp, Asn, Gly, and Ser found on cell surface proteins inserted into lipid bilayers. In comparison to Oand

PAGE 55

38 N-linked glycans, the sequon specificity for at tachment of glycans to proteins in lipid bilayers has yet to be defined well. 66 Glycomics attempts to define how protei ns are affected by the attachment of saccharides and what are the recognition events associated with saccharides and their binding partners. 67 Normally, glycans attached to pr oteins affect solubility, protease resistance, and quaternary stru cture. Also, glycoproteins and glycolipids are the most abundant species found on the outer walls of mammalian cells. Oligosaccharides are important for mediation of cell growth, cel l-cell adhesion, fertilization, immune defense response, inflammation, viral replication, pa rasitic infection, masking of tumors, and degradation of blood clots. 66;68 Further complexity of the glycome is revealed when deciphering temporal states during an organisms life cycle i.e., embryonic developement. 66 MS analysis of oligonucleotid e structures have been usef ul for identifying glycan structure in respect to different cell and dis ease states in a simila r manner to functional proteomics. 50;52;55;68 In the scope of this dissertation, the enzymes which build glycan chains, glycosyltransferases (s pecifically sialyltransferases), are reviewed. The study of the glycan structures themselves are not within the scope of this dissertation. Glycosyltransferase Family Glycosyltransferases are a family of enzymes that are membrane bound in the Golgi apparatus and endoplasmic reticulum and synthesize glycans in an assembly-line 66 manner. Extensive studies on the N-and O-glycosylation pathways in mammalian cells have revealed that glycans are synthesized by an ordered series of sugar transfer and cleavage reactions. 69 These enzymes are responsible for the transfer of different monosaccharides from sugar-nucleotide donor subs trates to form covalent linkages with

PAGE 56

39 other simple and complex oligonucleotides. 65 Interestingly, the 1970 Nobel Prize in Chemistry was awarded to L. F. Leloir for his discovery of UDP-Glucose as a substrate for glycogen biosynthesis. Almost all re actions which involve the synthesis of oligosaccharide chains involve s ugar-nucleotide donor substrates. 70 Four main factors are used to classify and name the different glycosyltransferase sub-families. The first includes the type of sugar transferred and the type of sugarnucleotide donor substrate used. The second includes the stereoa nd regio-selectivity exhibited by the acceptor substrate. 65 The third involves the description of a retaining or inverting mechanism with respect to the anomeric hydroxyl upon catalytic transfer. 67 The fourth includes the dependence upon di valent metal cations for catalysis. 66 In humans, it was calculated that there are over 25 0 different glycosyltran sferases, classified into 47 different sub-families. Overall, from completed genomes, it is estimated that there are over 7200 glycosyltr ansferases identified from bioinformatic studies. Furthermore, of all the open reading frames deposited in genomic databases, about 1% are dedicated to glycosyltr ansferase activity based on know primary structure motifs. 71 Even though the catalytic tr ansfer of monosaccharides by glycosyltransferases are performed using similar donor and accepto r substrates, the sequence homology and structure homology do not readily transfer be tween the sub-families. However, within each sub-family, there are several structur al motifs with high percent conservation. Although there are many important glycosyltr ansferases needed for production of the extremely complex glycans chains observed in nature, the focus of this dissertation now turns to sialyltransferases. Sialyltransferases are of par ticular interest because they

PAGE 57

40 transfer sialic acid, a charged sugar, which provides the finishing touches on large oligosaccharides. Sialyltransferase Sub-Family Sialyltransferases catalyze the transfer of sialic acid from cytidine 5monophosphate N-acetylneuraminic acid (CMP-NeuAc) to the terminal non-reducing positions of oligosaccharides, glycoproteins, or glycolipids. 72 NeuAc is an interesting saccharide because it has a carboxylate, instead of a hydroxyl, attached to the anomeric carbon. Interest in sialyltransf erases arose due to the bioche mistry of N-acetylneuraminic acid (NeuAc) capped carbohydrate moieties which bear a negative charge from the carboxylate group attached to th e anomeric carbon. Furtherm ore, NeuAc is the most common naturally occurring sialic acid in human glycoconjugates. 73 Since sialic acids are attached to the termini of large oligosaccharides, they are implicated in important physiological phenomena and disease states. 73;74 Since sialyltransferases are the enzymes involved with the transfer of sialic acid, there are large efforts to characterize this family of enzymes. As with all glycosyltransfer ases, the type of linkage formed is described by the nomenclature of sialyltransferases. The li nkages formed by sialyltransferases may be through an 2 3)or an 2 6)-bond to Gal (ST3Gal or ST6Gal); through an 2 6)-bond to GalNAc (ST6GalNAc); or through an 2 8)-bond to Sia (ST8Sia). 75 The 2 3)linkage is found to be most predominate than the 2 6) linkage. 76 Figure 2-2 represents th e reaction catalyzed by 2 3) sialyltransferases, where R is Gal.

PAGE 58

41 O N N O OH HO O P O O NH2 O O CO2HO H N OH OH HO O O HO H N OH OH HO O CO2O R R OH CMP CMP-NeuAc O N N O OH HO O P O O NH2OO CO2HO H N OH OH HO O O HO H N OH OH HO O CO2O R ROH CMP CMP-NeuAc Figure 2-2. Reaction Catalyze d by Sialyltransferases Sialyltransferases are type II transmembr ane glycoproteins that have a short NH 2 terminal cytoplasmic tail, an amino acid signal anchor domain, and an extended stem region with a large C-terminal catalytic domain (Figure 2-3). 77 Sialyltransferases are anchored in the trans cisternae of the Golgi apparatus and the trans Golgi network. 78 The stem region may be from 20 to 200 am ino acids long and ma y mediate acceptor specificity. 79;80 As with all glycosyltransferase genes, sialyltransferase genes are differentially expressed in a cell-type tissue-type, and stage specific manner. 72 In-vivo proteolysis may occur at the stem re gion, thus making the enzyme soluble. 79 Sialyltransferases have b een found mainly in the deut erostome lineage such as mammals, birds, amphibians, boney fish, and insects. Also, viruses produce similar sialyltransferases to mammals for pathological infection. 81 Furthermore, sialyltransferases have been cloned from bacteria, however they lack any sequence

PAGE 59

NH2Cytoplasmic TailCytoplasm Golgi LumenStem RegionCatalytic DomainSignal Anchor 42 NH2 Tail Figure 2-3. Topology of Sialyltransferase Cytoplasmic Cytoplasm Golgi Lumen Stem Region Catalytic Domain Signal Anchor homology or identity to the deuterostome forms. Lastly, sialic acid has been detected in plants and fungi, but characte rization of sialyltransferases from these organisms have lagged further behind than ma mmalian, virus, or bacteria. 82 Funding for medicinal chemistry based studies and completion of huma n, different virus, a nd different bacterial genomes has made mammal, virus, and bacteria l sialyltransferases characterization more prevalent. Currently, over 22 unique sialyl transferase members have been cloned from mammalian sources. 79 According to the current Henrissat family classification of glycosyltransferases, sialyltransferases are split into 3 evolutionary families. The first, termed GT 29, consists of 94 sequences from mammals, plants, a nd viruses (ST3Gal, ST6Gal, ST6GalNAc, ST8Sia). The second and third, termed GT42 and GT52, consist of 16 and 28 sequons from different bacteria. 71 Most insect cells have sim ilar glycosylation pathways, and until very recently, were believed to lack e ndogenous sialyltransferase activity.

PAGE 60

43 Studies on Autographa california nuclear polehedos is virus with Spodoptera frugiperda (fall armyworm) did not provide evid ence for sialyltransferase activity. 83 However, it was speculated that insects produce sialyltransferases during embryonic development. In 2004, one sialyltransferase gene from the Drosophila (fruit fly) genome was aligned against members of the vertebrate ST6Gal family. 80;81 Cloning, expression, and characterization of the ST6Gal from Drosophila melanogaster solidified that sialyltransferases are expressed in insect s at a certain stages of development. Furthermore, the ST6Gal gene was aligned with a sialyltransferase gene found in the mosquito genome (47% sequence identity). 80 Many viruses have sialyltransferase gene s which are similar to the mammalian sialyltransferase genes and are used to inva de a host cell. For example, the myxoma virus, Leporipoxvirus Poxivirdae, uses a sialyltransferase gene as one of the enzymes to enhance virulence towards rabbits. The sialyl transferase characterized in this study has 43% sequence identity and 60 % sequence similarity to vertebrate ST3Gal family members. 84 Another example consists of a sial yltransferase encoded by the Hepatitis B virus which infects the liver to induce malignant tumors. 85 Although bacterial sialyltran sferases utilize the same sugar-nucleotide donor substrate for catalysis, they show no sequen ce homology or identity to mammalian, plant, virus, or insect sialyltransferases. Howeve r, sialyltransferases have been cloned and characterized from the pathogenic Neisseria meningitides Neisseria gonorhoeae, and the marine bacteria Photobacterium damsela In comparison to other taxons, bacterial sialyltransferases however, they exhibit much broader acceptor specificities. It was speculated that bacterial sialyltransferases evolved differently than invertebrate

PAGE 61

44 sialyltransferase because they were not exposed to the same constrains. 86;87 Furthermore, the only crystal structure of a recomb inant sialyltransferase, cloned form Campylobacter jejuni, was published in early 2004. 88 Since all sialyltransfer ases share the same sugarnucleotide donor substrates, some threedimensional structure may be shared. Biological and Medicinal Importance Sialic acids are involved with numerous biol ogical processes, su ch as the regulation of glycoproteins in the blood stream, cell-cell interaction, cellular regulation, facilitation or prevention of aggregation, immune system function, and nervous system development. Of particular interest are the roles that sialic acid plays in the mask ing of tumors and the inflammatory/immune response for pathogen-hos t recognition of bacteria or viruses 73;80;84 The masking of tumors by sialic acid ha s been linked to brain, colon, breast, and prostate cancers. 69 Often, the link between incorrect glycan structure and malignant tumors are not always understood. However, it is well accepted th at the increase in branching of glycan chains associated with certain cancers will facilit ate an increase in sialic acid attachment, thus an increase in the sialyltransferase activity is observed. Furthermore, due to the complexity and th e microheterogeneity of glycosylation, the specifics of a certain glycan involved with cancer are often di fficult to determine. Even though there are many questions with the sialic acid/sialyltransfera se cancer relationship, four generalities are described. 68;75 First, the negative charge buildup from in creased sialic acid residues may prevent cell-cell interactions by repulsion events. S econd, selectins or siglecs (cell adhesion molecules) bind to irregularl y sialylated oligosaccharid es and propagate malignant transformation. Third, the complete abnorma l oligosaccharide may be masked by sialic acids, thus not allowing galectins to bind for repair. Fourth, specific cell signaling

PAGE 62

89-91 45 pathways might be the target for malignant transformation with viruses. It may be implied that by simply removing the sialic acids would address ma lignant repression; however, this is not always the case. 68;74;75 Immune system response often uses oli gosaccharides to discriminate between friendly and enemy structures. The failure to determine between self and non-self structures may lead to autoimmune reactions. 74 For example, the binding of terminal sialic acid residues are the initial step for influenza viru s infection. Hemagglutinin, a surface protein on the influenza virus, mediates attachment and fusion of the cell with the virus membrane. Obviously, there have been several studies to inhibit and characterize this sort of binding by modifying CMP-NeuAc at the C5 and C9 positions of sialic acids. Furthermore, studies on Hepatitis B viru s show that insertional mutagenesis of sialyltransferase and other enzymes cause proliferation of human liver cancer. 85 To facilitate rational drug design fo r inhibition of disease, severa l investigators have studied sialyltransferase primary structure. Sialyltransferase Structure and Mechanism Relevant for Structural Proteomics Vertebrate sialyltransferas es share the same topographical features; however they show very little sequence iden tity except 4 consensus motifs termed L, S,VS, motifs and motif 3. Figure 2-4 represents the conserved amino acids between the different types of human sialyltransferase, where the sialylmo tifs are labeled with the amount of amino acids and the numbers in parentheses corres pond to the number of amino acids between the motifs or the Nand C-termini. 81 Interestingly, the iden tification of structural sialylmotifs was aided with peptide mass finge rprinting in 1992 with liquid secondary ion and double focusing MS. Primers were de signed based on the common sequence of

PAGE 63

46 membrane bound sialyltransferase and known sialylmotifs to clone new sialyltransferases. 92 20 Different Sialyltransferases May Contain to 302 600 Amino Acids 20 Different Sialyltransferases May Contain to 302 600 Amino Acids 42 AAs L motif S motif 24 AAs (89 -109 AAs) (73-293 AAs) motif 3 4 AAs VS motif 6 AAs (9-17 AAs) (8-21 AAs) (22-30 AAs) 42 AAs L motif Smotif 24 AAs (89 -109 AAs) (73-293 AAs) motif 3 4 AAs VS motif 6 AAs (9-17 AAs)(8-21 AAs) (22-30 AAs) Figure 2-4. Sialylmotifs and Amino Ac id Length of Sialyltransferases cleft. The significance of these 4 sialylmotifs ha s been investigated using ST6Gal I and ST3Gal I as model enzymes. 72;79 First, mutagenesis studies on the L-motif affected binding to the donor substrate. 78 Second, mutagenesis of the S-motif altered the binding properties of both the donor and acceptor subtrates 93 Third, mutagenesis of motif 3 was shown to be important for func tional and/or structural roles. 79 Fourth, mutagenesis of VS-motif revealed the complete loss of activ ity. In these studies, amino acids were mutated which may be related to catalytic amino acids on the acceptor side of a binding 79;80;94 Fifth, mutagenesis studies on cysteines have revealed a disulfide bond which may bring the L and S-motifs within close proximity of each other. 95 Finally, glycosyltransferases normally have a di valent metal dependence and a common DXD motif. Sialyltransferase do not show a metal dependence and lack the DXD motif. 94 Using rat liver ST6Gal as a model, sial yltransferases were shown to catalyze a highly dissociative transition state with substantial oxocarbenium ion character 96-100 These kinetic studies also revealed that th e mechanism was steady-state random and that the enzyme had a bell-shape pH versus rate profile. 97 Interestingly, kinetic studies on sialyltransferase from the bacteria C. jejuni 88 and the insect D. melanogaster 80 both show similar bell-shape pH versus rate prof iles to the mammalian sialyltransferases. 94;97 Also, recent data presented by Dr. Erin Burke s uggest that protonation of the non-bridging

PAGE 64

47 oxygen from a general acid aids in CMP bond br eaking. Therefore, a general base and a general acid are the targets for the structural proteomic studies presen ted in Chapter 4. In February 2004, the first crystal structur e of a bacterial si alyltransferase was published.88;101 In this report, it was suggested that a histidine acts as a general base to assist in the nucleophilic attack of a hydroxyl and a tyrosine acts as a general acid to assist the departure of CMP. The loss of activity during the pr eviously described mutagenesis study on mammalian sialyltransferas es on the tyrosine residue (motif 3), taken in light of the bacterial crystal struct ure, may indicated that this amino acid is involved with active-site catalysis. However, since this sialyltransferase from bacteria does not have sequence homology or identity to mammalian sialyltr ansferases, it cannot be ruled out that another amino acid other than tyrosine acts as th e general acid needed for catalysis in the mammalian sialyltransferases. Another interesting feature of the bacterial sialyltransferase is that a loop of 12 amino acids becomes ordered upon binding to inhibitor.88 This lid may shield the oxocarbe nium transition state reaction from nucleophilic solvent attack. The availability of recombinant sialyltr ansferase will facilitate our MS based structural proteomics study to answer the question as to the identity and location of the general acid and/or general base.88 The absolute identification of the general acid and/or general base in tandem with molecular mechanics computational chemistry will provide the basis for intelligent design of selective sialyltransferase inhibitors.102-105 Furthermore, once a labeling strategy has been developed for our model system, recombinant human placental ST3Gal IV, it may be applied to all si alyltransferases with possible insight into other motifs or important structural properties.

PAGE 65

48 Recombinant Human (2 3) Sialyltransferase IV Traditionally, mammalian sialyltransferases are difficult to express and purify in large quantities due in part to their me mbrane bound protein prope rties and that they themselves are glycosylated. Because thes e enzymes are complex, eukaryotic cell lines are used for reliable expre ssion of recombinant mammalian si alyltransferases. Different cell lines such as yeast, insect, and mamma lian origins were used to produce soluble, catalytically active mammalian sialyltransferases. Attempts to produce catalytically active mammalian sialyltransferases in E. coli in our lab and others have not resulted in active enzyme.106 Several glycosyltransferases have been successfully cloned and overexpressed in active form with yeast such as Saccharomyces cerevisiae (fulllength)107;108 and Pichia pastoris (secreted).109 Mammalian cell lines have endogenous sialyltransferase activity and ha ve been useful for expression of glycosyltransferases and sialyltransferases108;110 Since insect cell lines such as Spodoptera frugiperda ( Sf 9) lack endogenous sialyltransferase activity, they are used to express soluble glyc osyltransferases and sialyltransferases with baculovirus vectors. 79;111-113 Through the beauty of modern molecular biology, soluble sialyltransferase was produced in our lab with a cleavable insulin signal peptide cloned upstream from the catalytic domain, replacing the Nterminal anchor and stem regi on, thus facilitating expressed construct secretion into the media.92;114-116 Also, to improve purif ication yields, fusion tails were placed between the signal peptide and the catalytic domain.79;108;111-113;117 Fusion tag/affinity purification pairs may include the mouse IgM signal pept ide and the IgG binding domain of the protein A, the streptavidin fusion tag with avidin binding, GST with anti-GST antibodies, or polyhistidine tags (His6xTags)with nickel agrose beads (Ni+2-NTA).118 Of these fusion

PAGE 66

49 tags, a polyhistidine tag was the first choice be cause of the ease of incorporation and the minimal amount of enzyme structure perturbation. We produced three recombinant forms of human placental -(2 3) sialyltransferase (hST3Gal IV; EC# 2.4.99.4) 110;119 which lacked the first 40 amino acids coding for the NH2-signal anchor and stem region. The three recombinant forms were overexpressed in Spodoptra frugiperda ( Sf 9) insect cells usi ng a baculovirus vector with Invitrogens (Carlsbad, CA) Bacto-Bac Baculovirus Expression System.120 The first recombinant form consists of a canin e insulin cleavable signal peptide cloned in front of the catalytic domain of hST3Gal IV (Ins-ST). The second recombinant form consists of the canine insulin signal peptide followed by a His6x-tag and two glycines in front of the catalytic domain of hST3Gal IV (N-Tag-ST). The third recombinant form consists of the canine insulin signal peptide followed by the catalytic domain of hST3Gal IV with two glycines and a His6x-tag placed at the C-terminus (C-Tag-ST). The inclusion of the His6x-tag will aid in ease of purification. The aim of this experiment was to produce 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain (Ins-ST) 40 AAs+ Canine Insulin Secretion Peptide + PolyhistidineFusion Tag + C-Terminal Catalytic Domain (N-Tag-ST) 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain + PolyhistidineFusion Tag (C-Tag-ST) Secretion PeptideC-Terminal Catalytic Domain Secretion Peptide C-Terminal Catalytic Domain C-Terminal Catalytic Domain Secretion Peptide HHHHHHGG GGHHHHHH 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain (Ins-ST) 40 AAs+ Canine Insulin Secretion Peptide + PolyhistidineFusion Tag + C-Terminal Catalytic Domain (N-Tag-ST) 40 AAs+ Canine Insulin Secretion Peptide + C-Terminal Catalytic Domain + PolyhistidineFusion Tag (C-Tag-ST) Secretion PeptideC-Terminal Catalytic Domain Secretion Peptide C-Terminal Catalytic Domain C-Terminal Catalytic Domain Secretion Peptide HHHHHHGG GGHHHHHH Figure 2-5. Sialyltransfer ase Constructs Prepared

PAGE 67

50 relatively large amounts of sialyltransferas e at a low price for multiple studies on sialyltransferase secondary st ructure. The assumption was made that cleaving the stem region will not perturbate activity.79 Results and Discussion Baculovirus preparation Dr. Nicole Horenstein originally prepared the pFastBacHTa vector that harbored the cDNA encoding the Ins-ST gene lacki ng the first 40 amino acids. The two other plasmid constructs containi ng Nand C-terminal His6xtag, in addition to the insulin signal peptide, were prepared by Bronson Anatao. Correct construction of cDNA clones with canine insulin secretion peptide and His6x-tags were confirmed by DNA sequence analysis performed at the University of Florida Protein Core Facility. All steps prior to affinity chromatography followed the Bac-to-Bac Baculovirus Expression System manual from Invitrogen (www.invitrogen.com) with no deviation from the described protocol. The three pFas tBacHTa plasmids containing the particular genes were separately transformed into competent DH10BacTM E. coli cells. PCR analysis (Figure 2-6) of mini-prepped recombinant backmid indicated an insertion product of proper length. Spodoptera frugiperda ( Sf 9) insect cells were transf ected in 6 well plates with recombinant bacmid DNA with the aid of Ce llfectin Reagent. Af ter transfection, the viral titers were below the level needed for sialyltransf erase expression. To produce higher viral titers (amplification), 50 to 75 mL of cell cultures containing 2-3 x 106 cells/mL were infected with 1 to 4 mL of viral stocks. After 48 to 55 hours post infection, the cell supern atant was clarified and sterile filt ered. Subsequent amplification proceeded until a viral titer of about 1x108 plaque forming units/mL (pfu/mL) was

PAGE 68

51 achieved. Table 2-1 represents the viral titers used for in fection of insect cells for expression of sialyltransferase. 123 45 9461 bp 6557 bp 4361 bp 2322 bp 2027 bp 1353 bp 1078 bp 123 45 9461 bp 6557 bp 4361 bp 2322 bp 2027 bp 1353 bp 1078 bp Figure 2-6. Confirmation of Correct DNA In sertion Product where: Lanes 1 and 5 are Base Pair Markers, and Lanes 2 thro ugh 4 Correspond to Different Colonies Selected Based on Blue/White Screeni ng after Liquid Culture and Mini-prep of the Recombinant Bacmid. Table 2-1. Viral Titers for Different Constructs Construct Titer in pfu Ins-ST 1.7 x 108 N-Tag-ST 2.1 x 108 C-Tag-ST 2.7 x 108 Expression and purification of the three sialyltransferase constructs Purification of the Ins-ST construct presen ted several different technical issues. First, a sepharose cytidine di phosphate-hexanolamine affinity (sCDP-Hex) material was synthesized by Dr. Erin Burke and Dr. Nicole Horenstein according to the literature.64 Second, the support phase of the sCDP-Hex resin was easily compressed, thus not allowing for large pressure grad ients. If large volumes of buffer were added to the sCDPHex column, flow reduced dramatica lly. Third, the addition of 300 mM -lactose to the

PAGE 69

52 initial supernatant before the loading of cl arified supernatant (harvested cells 72 hours post infection) resulte d in higher purity. Although the pur ity improved, the flow rate was reduced due to the increased viscosity. A contaminant protein, glycoprotein 64 (GP64), was expressed along with sialyl transferase in the baculoviru s system. Using trypsin ingel digestion of a band excised from SDS-PA GE analysis of an up-pure sample, GP 64 was positively identified as the contaminate protein with MS analysis. GP 64 was believed to bind to the N-glycans of sialyltr ansferase, thus creating the need for the lactose wash after loading clarified supernatan t. Finally, because of the low flow rates associated with the sCDP-Hex column, the in itial cell culture supern atant needed to be concentrated 5 fold. In an attempt to create constant and incr eased flow rate thro ugh the column, a postcolumn pump was added in-line, however, th is compressed the resin, thus producing no flow. Several attempts were made to purify the Ins-ST construct, often with no retention of enzyme on the column; however, some purif ications led to positive results. Table 2-4 summarizes the purification of th e Ins-ST construct. The de finition of a unit of activity was the number of mols of CMP-NeuAc converted to sialyl-lactose per minute. Presented in the Table 2-4, ce ll culture supernatant correspond s to pooled supernatant 72 hours post infection of 2 to 2.5 x 106 cells/mL with 3 mL of Ins-ST recombinant baculovirus. Ultrafiltration co rresponds to the concentration of clarified supernatant with an Amicon Ultrafiltration device with a polyethersulfone membrane with a 10 kDa molecular weight cutoff (MWCO). Load bu ffer addition corresponds to the addition of MES, pH 6.8 to 50 mM, -lactose to 300mM, glycerol to 20% v/v, and triton CF 54 to 0.05% v/v to the concentrated supernatant. Pooled fractions correspond to fractions

PAGE 70

53 eluted with high specific activity. Concentr ation corresponds to ultrafiltration of pooled fractions with an Amicon Ultrafiltration device with a polyethersulfone membrane with a 10 kDa MWCO. Table 2-2. Summary of an Ins-ST Purification Volume Activity Yield SA Ins ST Purification mL mU % U U / L U/mg Cell Culture Supernatant 370 820 100 2.20 0.029 Ultrafiltration 80 760 93 9.60 0.022 Load Buffer Addition 100 710 87 7.10 0.025 Pooled Fractions 28 230 28 8.12 0.161 Concentration 5 160 20 31.9 0.132 Figure 2-7 represents the elut ion profile for a steep step salt gradient applied to the column after loading the con centrate and washing with the -lactose buffer. The elution of Ins-ST from sCDP-Hex was performed by increasing the salt conc entration in a step gradient fashion. For the experiment presen ted in Figure 2-7, activity assays were not intially available and fractions correspondi ng to either 250 mM or 350 mM NaCl were pooled separately. Later, activity assays reveled that only the 350 mM NaCl pooled fractions contained sialyltransferase ac tivity and SDS-PAGE analysis revealed sialyltransferase with large amounts of contam inating proteins. This data suggested the need for more than 5 column equivalents per step for effectiv e purification. 0 50 100 150 200 250 300 350 400 450 500 1012141618202224262830323436Fraction 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 [NaCl] [P] ---mM NaCl mg/mL of Protein Fraction Number Concentration of protein in mg/mL Concentration of NaCl in mM 0 50 100 150 200 250 300 350 400 450 500 1012141618202224262830323436Fraction 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 [NaCl] [P] ---mM NaCl mg/mL of Protein Fraction Number Concentration of protein in mg/mL Concentration of NaCl in mM Figure 2-7. Protein Concentration Pr ofile of an Ins-ST Purification

PAGE 71

54 Figure 2-8 represents the elut ion profile for Ins-ST with a shallow step gradient and sialyltransferase activity fo r fractions corresponding to 350 mM and 450 mM NaCl. Activity was determined as described in th e Methods and Materials section To save radiolabeled substrates used to determ ine activity, full activity profiles were not collected. On average, 30 to 35 % of the initial activity was not retained on the column. Attempts were made to collect sialyltran sferase from the flow through by reloading it onto the column; however, sialyltransferase ac tivity was not retained. Also, fractions eluted with 100mM and 250 mM NaCl did not contain activity. Furthermore, if 20 % glycerol and 0.01 % triton CF 54 were not pres ent, activity was completely lost during purification. Finally, after collection and concentrati on of pooled fractions, these constructs were sensitive to dilution, dialysis and further ultrafiltration, with loss of 90% of the initial activity. 0 50 100 150 200 250 300 350 400 04812162024283236404448525660646872Fraction ug/mL of Protein0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0mU Figure 2-8. Protein Concentr ation and Activity Profile of an Ins-ST Purification SDS-PAGE analysis of purifie d sialyltransferase revealed a step ladder of possible glycoforms (Figure 2-9, Lane 2). Acco rding to the known consensus sequence for

PAGE 72

55 glycosylation, hST3Gal IV has f our possible sites for glycosyl ation. To confirm that the multiple bands are due to glycosylation, Endo Hf, a recombinant protein fusion of Endoglycosidase H and maltose binding protein, was used to deglycosylate Ins-ST. Figure 2-9, Lane 3 corresponds to the deglycos ylation product of sialyltransferase. The observed molecular weight (~34,000 Da), afte r the deglycosylation reaction, corresponds to the theoretical molecular weight of the primary sequence (~34,005 Da) free of glycans. 123 66 kDa 45 kDa 36 kDa 29 kDa 26 kDa 20 kDa 14 kDa 123 66 kDa 45 kDa 36 kDa 29 kDa 26 kDa 20 kDa 14 kDa Figure 2-9. SDS-PAGE Analys is of Purified Ins-ST a nd Deglycosylation Reaction where: Lane 1 = Molecular Weight Mark ers; Lane 2 = Purified Ins-ST; and Lane 3 = Ins-ST Deglycosylated with EndoHf. To possibly improve overall yiel ds and ease of purification, Ni 2+-NTA with His6xTag technology provided somewhat different results. The Ni2+-NTA resin does not compress easily, thus increasing the flow rate dramatically in comparison to the CDPHexanolamine-Agrose resin used for the purification of Ins-ST. This eliminated the need for supernatant concentration. Elution of Nand C-Tag-ST was performed by increasing the imidazole concentration in a step grad ient fashion (5 mM, 50 mM, 135 mM imidizole in a buffer containing 20 % glycerol, 0.05 % triton CF 54, 100 mM KCl, 50 mM MES, Ph 6.8). The purification of Nand C-TagST took 11 hours (harvest to concentrated

PAGE 73

56 pooled fractions), thus providing another positive quality of Ni2+-agrose over the sCDPHex affinity column. Tables 2-5 and 2-6 summarize the purification of the N-Tag-ST and C-Tag-ST constructs respectively. Table 2-3. Summary of a N-Tag-ST Purification Volume Activity Yield SA N-Tag-ST Purification mL mU % U U / L U/mg Cell Culture Supernatant 600 570 100 0.95 0.0009 Load Buffer Addition 660 540 95 0.82 0.0003 Ni-NTA Pooled Fractions 50 190 33 3.80 0.045 Concentration 4.5 155 27 34.5 0.216 Table 2-4. Summary of C-Tag-ST Purification Volume Activity Yield SA C-Tag-ST Purification mL mU % U mU / L U/mg Cell Culture Supernatant 470 560 100 1.20 0.008 Load Buffer Addition 580 420 75 0.72 0.007 Ni-NTA Pooled Fractions 50 180 32 3.60 0.120 Concentration 5.5 240 43 43.6 0.162 Presented in the Tables 2-3 and 2-4, cell culture supernatant corresponds to collected supernatant 72 hour s post infection of 2.5 x 106 cells/mL with 3 mL of N-TagST or C-Tag-ST recombinant baculovirus st ocks. Load buffer addition corresponds to activity after bringing the supernatant con centration to 20 % gly cerol, 0.01 % triton CF 54, and 50mM MES, pH 6.8. Without the addi tion of both 20% glyc erol 0.01% triton CF 54, the resulting yields were less than 1%. Ni2+-NTA pooled fractions correspond to fractions eluted with high sp ecific activity. Concentration corresponds to ultrafiltration of pooled fractions with an Amicon Ultraf iltration device with a polyethersulfone membrane with 10 kDa MWCO. These constr ucts were sensitive to dilution, dialysis, and further ultrafiltration, with loss of 90% of the ini tial activity. Compared to sCDP-Hex resin, the Ni2+-NTA column retained all of the activity. Furthermore, to increase specific activity, an imidazole step gradient was used to elute proteins which may non-specifically absorb to the column. For example, fractions eluted

PAGE 74

57 with the 50 mM imidazole buffer appeared to be yellowish, but did not contain activity. Figure 2-10 illustrates the typical elution profile for either His6x-tag constructs. The His6x tag enzymes eluted at 135 mM imidazole over 10 fractions (5 mL) with contant specific activity. Figure 2-11 represents the SDS-PAGE of both constructs. 0 100 200 300 400 500 600 700 800 06121824303642485460667278849096Fraction ug/mL of Protei n 0 2 4 6 8 10 12 14 16mUIncreasing Imidazole Concentration Activity in mU g protein / mL Concentration of protein in g/mL Activity in mU 0 100 200 300 400 500 600 700 800 06121824303642485460667278849096Fraction ug/mL of Protei n 0 2 4 6 8 10 12 14 16mUIncreasing Imidazole Concentration Activity in mU g protein / mL Activity in mU g protein / mL Concentration of protein in g/mL Activity in mU 0 100 200 300 400 500 600 700 800 06121824303642485460667278849096Fraction ug/mL of Protei n 0 2 4 6 8 10 12 14 16mUIncreasing Imidazole Concentration Activity in mU g protein / mL Concentration of protein in g/mL Activity in mU 0 100 200 300 400 500 600 700 800 06121824303642485460667278849096Fraction ug/mL of Protei n 0 2 4 6 8 10 12 14 16mUIncreasing Imidazole Concentration Activity in mU g protein / mL Activity in mU g protein / mL Concentration of protein in g/mL Activity in mU Figure 2-10. Protein Concen tration and Activity Profile of a C-tag-ST Purification 66 kDa 45 kDa 36 kDa 29 kDa 26 kDa 20 kDa 14 kDa 123 66 kDa 45 kDa 36 kDa 29 kDa 26 kDa 20 kDa 14 kDa 123 Figure 2-11. SDS-PAGE Analysis of TCA Pr ecipitated N-Tag-ST and C-Tag-ST where: Lane 1 = C-Tag-ST; Lane 2 = Molecula r Weight Markers; and Lane 3 = NTag-ST.

PAGE 75

58 Overt the last 25 years several laborato ries have cloned and purified different sialyltransferases. Table 2-5 corresponds to the trends on sialyltransferase cloning and purification as reported in the literature. Table 2-5. Summary of Sel ected Sialyltransferase Cloni ng and Purification Papers Enzyme and Reference Source Final Product Other Notes and (Acceptor Substrate) 1) ST6Gal (1979) porcine submaxillary glands121 Extracted from porcine submaxillary glands in intact form 2.77 U 2.4 % Yield 62 ug Started with 20 L of cell extract (asialo-ovine submaxillary mucin) 2) ST3Gal IV (1982) rat liver122 Extracted from rat liver in intact form 1.0 mol/min (U) 7 % Yield 36 ug Total ST Started with 17 L extract. (asialo1-acid glycoprotein) 3) ST6Gal I (1982) rat liver122 Extracted from rat liver in intact form 27 mol/min (U) 16 % Yield 3.3 mg Started with 17 L extract (lacto-N-tetraose) 4) ST3Gal IV (1985)119 human placenta Extracted from human placenta in intact form (original clone for our studies) 0.140 mol/min (U) 10 % Yield 90 ug Total ST Extracted from 2 kg of human placenta (Gal 14GlcNAc 5. ST3Gal IV (1992) porcine liver116 Expressed in COS-1 cells with secretion peptide Did not fully characterize 5mU/mL after harvest and concentration Main purpose was to identify and clone new ST (Gal 13GlcNAc 6. ST3Gal V (1993) human placenta123 Extracted from human placenta intact form Did not characterize Char acterization of acceptor substrate and detergents (Gal 1-4GlcNAc 1-3Gal) 7. ST6Gal I (1995) human placenta107 Expressed in Saccharomyces Cerevisiae Yeast in intact form Only reported U/L expression over time 0.31 U / L in 150 L bioreactor Reported specific activity as resonstituted yeast lypophilisate 0.8mU / mg protein (asialofetuin or asialo-ovine submaxillary mucin or LacNAc)

PAGE 76

59 Table 2-5. Continued Enzyme and Reference Source Final Product Other Notes and (Acceptor Substrate) 8. ST3Gal III (2003) rat liver113 Expressed in Sf 9 insect cells with insulin secretion peptide and His6x tag Reported as immobilization yield: 59% Immobilized on nickel column for synthesis of sialylated oligosaccharide ( -D-Gal p (1-3)-DGlcNAc p -OR) 9. ST3Gal III (2003) human111 Expressed in Sf 9 insect cells with IgM secretion peptide and IgG binding domain of IgG Did not fully characterize. Reported Specific activity to be 1.6 U/mg. 6.4 mg / mL Main purpose was to show that fusion of IgG protein A can be used with affinity separation (asialofetuin) 10. ST3Gal I (2004) human79 Expressed in Sf 9 insect cells with insulin secretion peptide and His6x tag Only reported specific activity 400 nU / mg Did not report all purification parameters (Gal 13GalNA -sp-biotin) Originally, sialyltransferases were puri fied from selected organs (Table 1-5, numbers 1-4, and 6) with difficult procedures needed for the extraction of activity. With modern day molecular biology, the yields associated with a 1 L expression are comparable with early purification methods Compared to other protein families, membrane bound proteins have always been di fficult to purify in high yield because of the extraction protocols used and their hydrophobic nature. For the rest of the purifica tions described in Table 25 (numbers 5, and 7-10), the full characterization of the purification was not as important because the clone of interest was used for mutagenesis studies, kinetic st udies, specificity studi es, or synthesis of sialylated oligosaccharides. For these type s of publications, aut hors present relative activity differences and partial kinetic parame ters to validate the importance of their application. Furthermore, sialyltransferases have specific acceptor substrates; however, a

PAGE 77

60 glycan found on a glycoprotein versus a gly can not bound to a glycoprotein will exhibit a different Km. There are other papers published on sialyltransferase cloning and purification which not presented; however, they follow the same trends observed in Table 2-5. Compared to the literature values presen ted in Table 2-5, our purification tables reflect similar values for expr ession and final yields. Our sp ecific activities may be lower than expected because a non-standard acceptor substrate ( -lactose) was used with the activity assays. Furthermor e, it appears that other i nvestigators had problems concentrating the enzyme. This observation is based on the lack of full characterization. Sialyltransferases are difficult enzymes to cl one and purify, but modern biochemistry and MS analysis are geared for analysis of low con centrations of enzymes. The fact that is it difficult to purify large quantities of sialy ltransferases (and most glycosyltransferases) has translated into a lack of crystal structures an d the lack of complete characterizations. Concentration/dilution studies on N-Tag-ST To investigate the stability of recombin ant hST3Gal IV, the N-Tag-ST construct was used as a model. The main purpose of these experiments was to determine if the NTag-ST could be desalted with a minimal lo ss of activity. Furtherm ore, for the labeling experiments presented in Chapters 4 and 5, it would be beneficial to remove excess labeling agents. Unfortunately, dialysis resu lted in recovery of 10 % of the initial activity. The next strategy was to use different micro-ultrafi ltration devices after dilution with an exchange buffer (20% glycerol 0.05 % triton X-100, 100 mM KCl, and 50 mM MES, pH 6.8). Dilution of recombinant hS T3Gal IV by a factor of 10 with one addition of exchange (dilution) buffer, followed by ultrafiltration with a 50 mL capacity Amicon

PAGE 78

61 ultrafiltration stir cell with an amicon polyethersulfone membrane with a 10 kDa MWCO, led to the loss of 95 % of the initial activity. This result was surprising because it was the same filtration unit and membrane used for the concentration of the pooled fractions after purification. The use of a Millipore micr on centrifugal filter device (MICRON YM-10; Reg. Cellulose; 10,000 MWCO) for concentration resulted in the loss of 99% of the initial activ ity. For the best results, c oncentration of N-Tag-ST was performed with a Millipore ultrafree-MC centrifugal filter unit (Durapore PVDF membrane;10,000 MWCO) with 400 L capacity. Figure 2-12 represents the 1:1 dilution of purified NTag-ST with different concentrations of triton X-100 or triton CF 54 present in the dilution buffer described earlier. The purified enzyme was dilute in ha lf with either triton X-100 (red bar graphs) or triton CF 54 (blue bar gra phs). The percent original activity was normalized based on the dilution factor. The data revealed that dilution to 0.025 % triton X-100 retains higher percent original activity than samples dilu ted with triton CF 54 at any concentration. 86% 76% 74% 13% 17% 0% 20% 40% 60% 80% 100% 0 0 2 5 % T r i t o n X 1 0 0 0 0 5 0 % T r i t o n X 1 0 0 0 0 1 0 % T r i t o n C F 5 4 0 0 2 5 % T r i t o n C F 5 4 0 0 5 % T r i t o n C F 5 4 Percent Original ActivitySample 86% 76% 74% 13% 17% 0% 20% 40% 60% 80% 100% 0 0 2 5 % T r i t o n X 1 0 0 0 0 5 0 % T r i t o n X 1 0 0 0 0 1 0 % T r i t o n C F 5 4 0 0 2 5 % T r i t o n C F 5 4 0 0 5 % T r i t o n C F 5 4 Percent Original ActivitySample Figure 2-12. Effect on Activity of N-TagST with Dilution of Different Detergent Concentrations

PAGE 79

62 Figure 2-13 represents the ul trafiltration of the N-Tag-ST with Millipore ultrafreeMC centrifugal filter units (Durapore P VDF membrane;10,000 MWCO) after dilution with 20% glycerol, 0.01 %, triton X-100, 100 mM KCl, 50 mM MES, pH 6.8. On the xaxis, the first value corresponds to the volume of purified N-Tag-ST added to the device, with the second value corres ponding to the number of 200 L dilutions with the buffer described above. For exampl e, 400 x 0 corresponds to 400 L of purified N-tag-ST concentrated to 100 L; 200 x 3 corresponds to 200 L of purified N-tag-ST diluted to 400 L with dilution buffer, then concentrated to about 200 L, followed by the addition of 200 L of dilution buffer, then concentrated to 200 L again, followed by the addition of 200 L of dilution buffer again, then concentrated to 100 L. The dilution of N-TagST was performed in the device in between centrifugation cycles. Percent activity was normalized based on the activity prior to concentration and the final volume (~100 L). It appears that the N-Tag-St construct is not stable under the conditions described for removal of high salt concentrations. 50% 34% 21% 17% 7% 16% 0% 10% 20% 30% 40% 50% 60% 70%Percent Original Activity 400 x 0 200 x 1 200 x 3 1 00 x 1 200 x 5 400 x 5 Sample Figure 2-13. Effect on N-Tag-ST Activity with Different Dilution Factors and Concentration with Millipore Ultra free-MC Centrifugal Filter Units

PAGE 80

63 To ensure that imidazole did not affect the activity assay, a blank consisting of 135 mM imidazole did not produce activity above ba ckground. It appears that there was not a useful protocol for removal of imidazole sa lt. For the labeling studies presented in Chapters 4 and 5, the constructs were deri vatized after purifica tion with no further desalting. Kinetic parameters Kinetic parameters were determined in co llaboration with Dr. Erin Burke. Table 26 presents the experimental kinetic para meters for our recombinant hST3Gal IV constructs and the wild-type human placental ST3Gal IV.119 In comparison to our expressed hST3Gal IV constructs, the WT ST3G al IV was purified with the N-terminal domain and the stem region intact. The kineti c data suggests that the inclusion of these domains with the WT ST3Gal IV effectively lowers the affinity for -lactose. The second important data point presented in Table 2-6 is the marginal increase of the C-tagST Km for CMP-NeuAc. It is suggested that a His6xtag added to the N-terminal end of the catalytic domain will not change the structure of the protein, whereas a His6xtag added to the C-terminal may change the struct ure of the enzyme. Since the WT, Ins, and N-Tag ST have similar Km values for CMP-NeuAc, and the C-Tag-ST has an increased Km value, it is suggested that the C-terminal region is important for binding of CMPNeuAc. Table 2-6. Summary of Kinetic Parameters for Different Si alyltransferase Constructs CMP NeuAc -lactose Recombinant Enzyme Km ( M) Km (mM) WT ST3Gal IV119 63 220 + 40 Ins ST 82 + 5 171 + 18 N Tag ST 74 + 8 155 + 14 C Tag ST 267 + 20 158 + 11

PAGE 81

64 Conclusion The addition of a His6xtag to the N-terminal or C-terminal end of the catalytic domain of ST3Gal IV (N-Tag-ST and C-TagST) provided constructs that were much easier to purify than Ins-ST. All three sialyltransferase constructs had similar Km values for donor and acceptor substrates in compar ison to the wild-type h23STGal IV, except the C-Tag-ST Km for CMP-NeuAc. Normally, the N-terminal end of the catalytic domain is extended through a stem region to an anchor which ligates the protein to the golgi apparatus; therefore, cha nges in the in the N-terminal region of the catalytic domain of ST3Gal IV would not necessarily perturbate structure. Since the C-Tag-ST Km is larger for the donor substrate, it is proposed that the His6xtag slightly alters the secondary structure. Lastly, since Ins-ST, C-Tag-ST, and N-Tag-ST were not stable to desalting procedures, the construct will be derivatized with bioconjugation techniques in the purification buffers with no further sample clean up. Methods and Materials All restriction enzymes, E. coli strains JM109 & ER2925, DNA Polymerase I (Klenow, Large Fragment), and T4 DNA lig ase were purchased from New England BioLabs (Beverly, MA). Wizard Plus Minipreps Kit and dNTPs were purchased from Promega (Madison, WI). Shrimp alkaline pho sphatase (SAP) was purchased from Roche (Indianapolis, IN). QIAquick Nucleotide Removal Kit and QIAquick Gel Extraction Kit were purchased from Qiagen (Valenci a, CA). The BCA Protein Assay Kit was purchased from Pierce (Rockford, IL). The BAC-TO-BAC Bacu lovirus Expression System (BBES), Spodoptera frugiperda ( Sf 9) insect cells, BACPACKTM Baculovirus Rapid Titer Kit, and DH10 BAC E. coli competent cells were purchased from Invitrogen (Carlsbad, CA). Primers for cloning and P CR analysis were obtai ned from Integrated

PAGE 82

65 DNA Technologies (Coralville, IA). The prot ocol for recombinant virus preparation was found in Invitrogens instruction manua l for BBES version D April6, 2004. The sepharose CDP-hexanolamine affinity column (sCDP-Hex) was prepared as per the literature.121;124 The [6-3H] N-acetyl D-mannosamine isotopomer used in the synthesis of [9-3H] NeuAc was purchased from Moravek. N-acetyl neuraminic acid (NANA) aldolase [EC 4.1.3.3] used with the synthesis of [9-3H] neuraminic acid (NeuAc) was cloned, overexpressed in E. coli and purified according to literature procedures.125;126 The E. coli expression plasmid pWV200B harboring the E. coli CMP-NeuAc synthetase gene [EC 2.7.7.43] used for the synthesis of all CMP-NeuA c substrates was a generous gift from Dr. W. F. Vann at the National Institute s of Health. Radioactive samples from sialyltransferase activity assays were analyzed w ith a Packard 1600 TR liquid scintillation analyzer. Trypsin for in-gel digestion was purchased from Promega and Sigma-Aldrich (St. Louis, MO). All other chemicals for purification were obtained from Sigma-Aldrich in the highest-purity. Baculovirus vector preparation for Ins-ST The original expression c onstruct derived from pFastBackHTa contained an NTerminal polyhistidine tag, but lacked a secretion signal (N. Horenstein, unpublished work). The sequence coding for the His-ta g was removed from the construct with RsrII and BamHI A 114 bp insert containing the sequence coding for the canine insulin leader peptide with flanking RsrII and BamHI sites was prepared in the following way. Two complementary single-stranded 70 bp oligonuc leotide primers (Forward Ins-ST: 5CGCGCGGTCCG AAATGATGGCCCTCTGGATGCGCCTCCTGCCCCTGCTGGCCC TGCTGGCCCTCTGGGCG-3 and Reverse Ins-ST: 5GCGGATCC GCCCCGGGAATCAACGAAGGCTCGGGTGGGCGCGGGCGCCCAG

PAGE 83

66 AGGGCCAGCAGGGCCAGCA-3) spanning the in sulin leader peptide were obtained commercially. The underlined ba ses denote the location of the RsrII and BamHI restriction sites for the Forward Ins-ST a nd Reverse Ins-ST primers, respectively. The oligonucleotides were annealed and then filled in with Kl enow fragment to produce the 114 bp insert which was doubly-digested with RsrII and BamHI The insert was then ligated into the RsrII and BamHI restriction sites of the e xpression vector using T4 DNA ligase. The new construct (Ins-ST) was seque nced at the Univers ity of Florida ICBR Protein Core Facility to conf irm insertion of the Ins-ST ge ne. Isolation of bacmids and generation of baculovirus followed the BBES protocols. Baculovirus vector preparation for N-Tag-ST The Ins-ST plasmid was digested for 12 hours at 37 C with BamHI Clean-up of the enzymatic digestion was performed usi ng the QIAquick Nucleo tide Removal Kit. The plasmid was digested with StuI (6h at 37 C), thermally inactivated (20 min at 65 oC), agarose gel purified, and extracted using the QIAquick Gel Extraction Kit. The 5 ends generated by digestion were dep hosphorylated using SAP. The entire Nterm6XHisTag insert was created by allowi ng two complimentary primers to anneal and was then extended with Klenow fragment Primers were mixed in equal volumes (2.5 L) of NtermHISFor_Upper (5GGTAGGCCC TGGCCATTAAGCGGATGCTGGAGAT GGGAGCTATCAAGAACCT CACGTCC-3), (50 M, 0.80 g/L) and NtermHisRev_Lower (5AGCAGGCCT TGCTCTCTGCCTCACCCTGGAG GAGGCACGGCTCCTTCTTCTCG CC-3), (50 M, 0.81 g/L) and heated at 90 C for 10 minutes. After allowing the mixture to equilibrate to room temperature, it was brought up to a total volume of 20 L containing 11.2 L deionized water, 420 M of a dNTP mixture, 10 mM Tris-HCl pH

PAGE 84

67 7.5, 5 mM MgCl2, 7.5 mM dithiothreitol, and 5 U of Klenow fragment. This reaction mixture was then allowed to incubate at 25 C for 80 minutes. After inactivation (20 mins at 75 oC) the mixture was cooled to room temperature and then doubly digested with BamHI and StuI (9 hours at 37 C). Following h eat inactivation, a clean-up of the enzymatic reaction was performed using the QIAquick Nucleotide Removal Kit. The Nterm6XHisTag insert was then ligated into the BamHI/StuI sites of the pFastBacHTaInsulin/ST vect or. The new construct, pFastBacHTaInsulin/Nterm6XHisTag/ST (N-Tag -ST), was sequenced at University of Florida ICBR Protein Core Facility to c onfirm the presence of the Nterm6XHisTag insert. Isolation of bacmids and generati on of baculovirus stocks followed the BBES protocol. Baculovirus vector preparation for C-Tag-ST The pFastBacHTaInsulin/ST plasmid was purified from E. coli strain ER2925 (dcm-) using the Wizard Plus Minipreps kit. This st ep was necessary to produce plasmid that could be rest ricted with Dam or Dcmsensitive restriction enzymes such as Eco01091 The isolated plasmid was digested with Eco0109I for 10 hours at 37 C. Clean-up of the enzymatic digestion was performed using the QIAquick Nucleotide Removal Kit. The Eco0109I digested plasmid was digested for 8 hours at 37 C with XhoI After thermal inactivation (20 min at 65 oC), the doubly digested plasmid was agarose gel purified then extracted with the QIAquick Gel Extraction Kit. The 5 ends generated by digestion were dephosphorylated using SAP. The C-Tag-ST insert was created followi ng a similar procedure to that of the Nterm6XHisTag insert. Primers were mixed in equal volumes (2.3 L) of CtermHISFor_Upper (5-

PAGE 85

68 GGTAGGCCC TGGCCATTAAGCGGATGCTGGAGAT GGGAGCTATCAAGAACCT CACGTCC-3), (50 M, 0.88 g/L) and CtermHisRev_Lower (5AGCCTCGAG TTAGTGATGGTGATGGTGATGAC CGCCGAAGGACGTGAGGTTC TTGATAGC-3), (50 M, 0.92 g/L) and he ated at 90 C for 10 minutes. After allowing the mixture to equilibrate to room temperature, it was brought up to a total volume of 20 L containing 11.5 L deionized water, 462 M of a dNTP mixture, 10 mM Tris-HCl pH 7.5, 5 mM MgCl2, 7.5 mM dithiothreitol, and 5 U of Klenow fragment. This reaction mixture was then allowed to incubate 25 C fo r 2 hrs. After inactivation, the mixture was cooled to room te mperature and then digested with Eco0109I (16 h at 37 C). The plasmid was purified using the QI Aquick Nucleotide Removal Kit, digested with XhoI (8 h at 37 C), then thermally inac tivated. The Cterm6XHisTag insert was ligated into the Eco01091/XhoI sites of the pFastBacHTaInsulin/ST vector. The new construct, pFastBacHTa/ST/6XHisTag/CtermInsulin (C-Tag-ST) was sequenced at University of Florida ICBR Protein Core F acility to confirm the presence of the Cterm6XHisTag insert. Isolation of bacm ids and generation of baculovirus stocks followed BBES protocol. Expression and purification of Ins-ST Recombinant baculovirus stocks, as prepared following the common BBES protocols, were amplified four times before a sufficient titer was obtained. For Ins-ST baculovirus amplification, Sf 9 insect cell cultures ( 50 75 mL) containing 2 x 106 cells/mL were infected with 4 mL of vira l stocks, collected 42-50 hours post-infection, and centrifuged at 4000 rpm for 15 mins at 4oC Prior to subseque nt amplifications or expression, viral stocks were filtered with 0.2 m sterile filters. For Ins-ST expression,

PAGE 86

69 cell cultures (50 75 mL) containing 2 x 106 cells/mL in 250 mL polycarbonate shaker flasks (Corning, NY) were infected with 3 mL of viral stock w ith a titer of 1.7 x 108 pfu/mL. After 72 hours post in fection, the following purif ication protocols were followed. Cultures were collected a nd centrifuged at 14000 RPM (4o C), after which, the supernatant was concentrated using a 200 mL capacity Amicon ultrafiltration stir cell model 8200 (Billerica, MA ) with an amicon ultrafiltration polyethersulfone membrane with a 10 kDa MWCO (product number PBGC 04310). The concentrate was brought to 20% glycerol, 0.01 triton CF 54, 300 mM -lactose and loaded onto a sepharose CDPhexanolamine affinity column (1.7 x 12 cm ) equilibrated with 10 mL of a buffer containing 300 mM -lactose, 20% glycerol, 0.01% tr iton CF 54, and 50 mM MES, pH 6.8 at 4 oC. This buffer was also used to wash the column after loading the sample. Elution of enzyme proceeded using a KC l step gradient of 150 mM, 250 mM, 350 mM and 450 mM. These KCl buffers also incl ude 20% glycerol, 0.01% triton CF 54, 50 mM MES, pH 6.8. Fractions (5 mL) containing ac tivity were pooled and concentrated with a 50 mL Amicon ultrafiltration stir cell model 8050 with an Amicon polyethersulfone membrane with a10 kDa MWCO (produc t number: PBGC09005) Activity was determined by the method described below. Bradford and bicinchoninic acid (BCA) assays were used to estima te protein concentration. Expression and purification of N-Tag-ST and C-Tag-ST Recombinant baculovirus stocks, as prepared following the common BBES protocols, were amplified four times before a sufficient titer was obtained. For N-Tag-ST and C-Tag-ST baculovirus amplification, Sf 9 insect cell cultures (50 75 mL) containing 2 x 106 cells/mL were infected w ith 4 mL of viral stocks, collected 42 50 hours post-

PAGE 87

70 infection, and centrifuged at 4000 rpm for 15 mins at 4oC Prior to subsequent amplifications or expression for purifica tion, viral stocks were filtered with 0.2 m sterile filters. For N-Tag-ST and C-Tag-ST expre ssion, cell cultures (50 75 mL) containing 2 x 106 cells/mL in 250 mL polycarbonate shaker fl asks (Corning, NY) were infected with 3 mL of 2.1 x 108 pfu/mL viral stock (N-Tag -ST) or 3 mL of 2.7 x 108 pfu/mL viral stock (C-Tag-ST). After 72 hours post inf ection, the following purification protocols were followed. Cultures were collected a nd centrifuged at 14000 RPM (4o C), after which, the clarified supernatant was brought to 20% gl ycerol, 0.01 % Triton CF 54. Novagen NiNTA-His Bind resin was packed in to a column (2 x 10 cm) at 4oC and was equilibrated with 10 mL of buffer containing 5 mM im idazole, 100 mM KCl 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8. After lo ading, the column was washed with 10 column equivalents of a buffer containing 300 mM -lactose, 5 mM imidazole, 100 mM KCl 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8, followed by a 5 column equivalent wash with a bu ffer containing 5 mM imidazole 100 mM KCl, 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8. Next, the column was washed with a buffer containing 50 mM imidazole, 100 mM KCl, 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8 to remove any proteins which non-specifically bind to the column. Finally, N-Tag-ST or C-Tag-ST was eluted with a buffer consisting of 35 mM imidazole, 100 mM KCl, 20% glycerol, 0.01% triton CF 54, 50 mM MES pH 6.8. Fractions (5 mL) containing activity were pooled and concentrated with a 50 mL Amicon ultrafiltration stir cell model 8050 with an Amicon polyethersu lfone membrane with a10 kDa MWCO (product number: PBGC09005). Activity was determined by the method described

PAGE 88

71 below. Bradford and bicinchoninic acid (BCA ) assays were used to estimate protein concentration. Activity assays for recombin ant sialyltransferase All activity assays reported on sialyltran sferase were performed in the following way: a 10 L aliquot from a selected sample was incubated with a 10 L mixture of 100170 M [9 3H] CMP-NeuAc (final concentration of 50 to 85 M; 10,000 20,000 counts-per-min; specific activity of 10 20 Ci/ mol) and 235 mM -lactose (final concentration of 118 mM) for th e appropriate amount of time to limit the consumption of CMP-NeuAc to less than 10% at 25 oC. The reaction mixture was quenched with 500 mL of 5 mM inorganic phosphate buffer, pH 6.8 and then applied to a Dowex 1 X 8, 200 mesh (PO4 2-) mini-columns pH 6.8.114;122 Reactions were eluted with 3.5 mL of 5 mM Pi buffer pH 6.8 into liquid scinti llation vials. The definition of a unit of activity is the amount of enzyme that produces 1 mol of sialyl-lactose pe r minute under the assay conditions described. Because the activity was not obtained under saturating conditions, the activity reported was corrected with expe rimentally determined kinetic parameters, known substrate concentrations, and the following bi-substrate equation. adjusted = ( max*[A]*[B]) / (KmA* KmB + KmA* [B] + KmB[A] + [A]*[B]) N-Tag-ST stability experiments All samples and dilution buffers were kept at 4 oC during these experiments. NTag-ST sample starting buffer consisted of 20% glycerol, 0.01 % triton CF 54, 100 mM KCl, 135 mM imidazole (i.e. purification buffe r). The following conditions were used to generate the graph illustrating the activity of N-Tag-ST after dilution with different

PAGE 89

72 detergent concentrations (Figure 2-12). Percent activity was normalized based on the activity prior to dilution and the final volume. 0.025% Triton X-100 : N-Tag-ST (20 L) was diluted to 40 L with a buffer consisting of 20% glycerol, 0.05 % tr iton x 100, 100 mM KCl, 50 mM MES pH 6.8. 0.050% Triton X-100 : N-Tag-ST (20 L) was diluted to 40 L with a buffer consisting of 20% glycerol, 0.1 % trit on x 100, 100 mM KCl, 50 mM MES pH 6.8. 0.010% Triton CF 54 : N-Tag-ST (20 L) was diluted to 40 L with a buffer consisting of 20% glycerol, 0.01 % trit on CF 54, 100 mM KCl, 50 mM MES pH 6.8. 0.025% Triton CF 54 : N-Tag-ST (20 L) was diluted to 40 L with a buffer consisting of 20% glycerol, 0.05 % trit on CF 54, 100 mM KCl, 50 mM MES pH 6.8. 0.050% Triton CF 54 : N-Tag-ST (20 L) was diluted to 40 L with a buffer consisting of 20% glycerol, 0.1 % trit on CF 54 / 100 mM KCl / 50 mM MES pH 6.8. For the dilution\concentration experiment s, ultrafiltration was performed with Millipores Ultrafree-MC centrifugal filter uni ts (Durapore PVDF membrane;10,000 MWCO; 400 L capacity) at 3000 RPM. The following conditions were used to examine enzyme activity after dilution and concentrati on with different dilu tion factors (Figure 213). For nomenclature, the fi rst value corresponds to the volume of purified N-Tag-ST added to the device, with the second value corresponding to the number of dilutions with dilution buffer (20% glycerol, 0.01 %, trit on X-100, 100 mM KCl, 50 mM MES pH 6.8). Percent original activity was normalized based on the activity prior to concentration and the final volume (~100 L). 400 x 0 : N-Tag-ST (400 L) was concentrated to a final volume of 102 L. 200 x 1 : N-Tag-ST (200 L) was first diluted to 400 L with dilution buffer prior to being concentrated to a final volume of 132 L.

PAGE 90

73 200 x 3 : N-Tag-ST (200 L) was first diluted to 400 L with dilution buffer. After concentration to 200 L, another 200 L of dilution buffer was added to the concentrate while still in the unit. The solution was then concentrated one more time to 200 L, when upon the sample was diluted to 400 L again with dilution buffer prior to being concentrat ed to a final volume of 140 L. 100 x1 : N-Tag-ST (100 L) was first diluted to 400 L with dilution buffer prior to being concentrated to a final volume of 70 L. 200 x 5 : N-Tag-ST (200 L) was first diluted to 400 L with dilution buffer. After concentration to 200 L, another 200 L of dilution buffer was added to the concentrate while still in the unit. Di lution and concentration was performed 3 more times until a final volume of 120 L was achieved. 400 x 5 : N-Tag-ST (200 L) was first concentrated to 200 L before being diluted to 400 L with dilution buffer. Dilution and concentration was performed 4 more times until a final volume of 120 L was achieved. In-gel digestion of glycoprotein 64 Glycoprotein 64 (GP 64) was identified by SDS-PAGE analysis, in-gel digested with Trypsin Gold (Promega), and LCQ-MS analysis of extracted peptides. The band corresponding to GP 64 was cut into 1mm cubes and washed 2 times with water in 500 L microfuge tubes. The cubes we re then destained with 100 to 150 L of a buffer containing 25 mM ammonium bicarbona te and 50 % acetonitrile at 27 oC for 4 to 8 hours with removal and re-application 3 times. Afte r all the coomassie blue stain was extracted from the gel slices, they where again wash ed with water. To digest recombinant sialyltransferase, 50 L of 20 ng/ L trypsin in 20 mM ammonium bicarbonate was added and allowed to absorb into th e gel cubes for 30 minutes at 27 oC. Next, the appropriate volume of 20 mM ammonium bicarbonate was added to cover the gel cubes and allowed to incubate at 27 oC overnight. Supernatant was collect ed with subsequent extraction of peptides from the gel cubes with 100 L of a buffer containing 30 % acetonitrile and 5% acetic acid (v/v) three times. Collected supernatant and extracted peptides were

PAGE 91

74 combined and vacuumed to dryness with a Labconco CentriVap Concentrator (Model number 7810000) at 37 oC. For LCQ MS analysis, samples were re-suspended in 5% methanol, 0.5 % acetic acid (v/v). MS analysis LCQ MS was performed on a Thermo Elect ron LCQ Deca (San Jose, CA) operated in positive ion mode. A spra y voltage of 1.8 kV +/0.2 was applied at the liquid junction before a capillary column. An Eldex Mi croPro pump delivered solvent at 6 L/min to a LC Packings C18 PepMap Nano-Prepcolumn (0.3 x 1 mm). Five microliters of sample was loaded onto the nano-prep column and wash ed for 5 mins at 0 % B. After washing, the direction of flow was reve rsed with a new flow rate of 200 nL/min after pre-column splitting from 8 L/min. Elution of peptides fr om a New Objective (0.75mm x 5 mm) Proteopep capillary column included a grad ient of 3-60% B over 30 minutes and 60-90% B over 1 minute with solvent A as 0.1% ace tic acid, 3% acetonitrile and solvent B as 0.1% acetic acid, 95% acetonitrile.

PAGE 92

75 CHAPTER 3 HYDROGEN / DEUTERIUM EXCHANGE HPLC-MS FOR SIALYLTRANSFERASE SECONDARY STRUCTURE Introduction to Hydrogen/Deuterium Exchange HPLC-MS The full characterization of a protein requires the understa nding of function, structure, and dynamics as investigated w ith a variety of techniques. A proteins functional characterization is normally perf ormed with biochemical, molecular, and cellular types of analys is (Chapter 2), whereas protein te rtiary structure and dynamics are investigated with physical methods.127 Over the last 10 years, many groups have applied hydrogen-deuterium exchange (H/Dx) coupled with on-line HPLC-MS as an analytical technique to describe protein secondary structure in regard s to protein topology, proteinprotein interactions, protein folding d ynamics, and protein-ligand interactions.128 H/Dx provides information on the solvent accessibil ity of amide protons by the exchange of a proton to a deuterium. Since there is a mass shift associated with this exchange, MS is used to determine the new molecular wei ght, thus yielding information on solvent accessible sites. Because MS is geared towards analyzing very small amounts (picomoles) of proteins and mass shifts asso ciated with protein/peptide derivatization, many investigators have chosen H/Dx HPLC -ESI-MS to study the secondary structure and the folding dynamics of a particular protein. Including H/Dx HPLC-ESI-MS, X-ray crystallography and nuclear magnetic resonance (NMR) may describe protein three dimensional structure. X-ray crystallography and NMR are proven techniques for locating all atoms in a protein to

PAGE 93

76 within 2-3 .127;129 NMR spectroscopy has improved, but analysis is still limited to proteins under 35,000 Da and a minimal am ount of posttranslational modifications (PTMs). In regards to X-ray crystallizati on, the art of crystallizi ng proteins has limited the number of protein crys tal structures reported.130 Also, proteins with PTMs, such as glycosylation, normally hamper crystallization.127 We evaluate H/Dx HPLC-ESI-MS to study the active-site amino acids and dynamics of recombinant hST3Gal IV because its molecular weight is about 37,000 Da a nd it contains four N-linked glycan. Monitoring Hydrogen/Deuterium Exchange of Amide Protons H/Dx of protein amide protons may be monitored with a variety of methods; however, NMR and MS are currently the main an alytical methods for analysis because of high through-put demands. The first discussi on of H/Dx was presen ted by LinderstromLang in the mid 1950s after the description of -helixes and -sheets.131 They speculated that hydrogens not involved w ith H-bonding should be observable with exchange experiments. In the 1960s, hydrogentritium exchange with liquid scintillation analysis demonstrated that different ex change rates were observable between amide protons involved with H-bonding and those which were solvent accessible. NMR spectroscopy has also been used to study H/Dx; however, NMR spectroscopy has the limitations described earlier. Johnson an d co-workers were the first to use H/Dx HPLC-ESI-MS to investigate the differenc es between apoand holomyoglobin in 1994.131 H/Dx monitored with NM R spectroscopy is considered to be high-resolution in comparison to H/Dx HPLC-MS because NMR ma y yields spatial resolution (exact sites of exchange) and MS yields segment resolu tion (mass shift of a exchanged proteolytic peptide in comparison to the mass of a the same non-exchanged prot eolytic peptide).

PAGE 94

77 Furthermore, the H/Dx of HEWL (~14.3 kDa) with NMR and MS analysis resulted in similar descriptions of solvent assessable amide protons.132 Theoretically, MS/MS may provide the same spatial resolution as NMR; however, hydrogen and deuterium may scramble upon pept ide fragmentation to provide useless information.129 Currently, it is accepted that MS/MS spectra are not useful for locating individually exchanged amide protons. Thus, the length of a peptid e as analyzed with MS is inherent to the spatial details of solv ent accessible amide prot ons. The digestion of exchanged proteins is performed at low pH with pepsin. Pepsin is the main low pH protease readily available for H/Dx HPLC-MS work flows. Since pepsin does not have strict specificity for polypeptide cleavage, pe ptides with different cleavages within the same region may overlap to provide spatial resolution.133 Hydrogen/Deuterium Exchange Kinetics Proteins have protons with different exchange kinetics. For amide protons, the kinetics are determined by protein structure, pH, and adjacent amino acid side chains. These three characteristics affect solvent acc essibility, backbone flexibility, and hydrogen bonding. If an amide proton is exposed to D2O at pH 7.0 (25 oC), the exchange rate will be about 10-1s. In comparison, the exchange ra tes of amide protons involved with Hbonding are about 108 s.128 Proteins in solution exhibit highly dynamic behavior such as native flexing, also known as breathing. Furthermore, proteins may undergo allosteric or local changes (hinge movement) upon protein-pr otein or protein ligand binding.127;134-136 The changes in structure may be from local fluctu ations to large confor mational perturbations, thus changing the intermolecular and so lvent interactions of amide protons.136 The difference between exchange rates of amide protons in unfolded and folded proteins

PAGE 95

78 provides a probe for studying conformational changes and dynamic process. The model for exchange kinetics of local or global pr otein structure is desc ribed with Equations 1 through 3 in Figure 3-1, where kex is the observed exchange rate for amide protons, k1 is the unfolding rate of the protein, k2 is the exchange rate of the solvent accessible amide protons, and k-1 is the refolding rate of the protein. Equation 1: kex= k0ki/(k0+kc+ki) Equation 2: EX1 ki>> kcthen kex= k0Equation 3: EX2 kc>> kithen kex= k0ki/ki Figure 3-1. Equations De scribing H/Dx Kinetics128;135 The link between measurable H/Dx rate s and flexing dynamics is described by Equation 1. From Equation 1, two extremes of kinetics, EX 1 for Equation 2 and EX 2 for Equation 3 are described. EX 1 represen ts the pseudo-first order kinetics in which any un-folded event will result in comple te exchange of the region which becomes solvent accessible, thus the rate limiting step for EX1 is the protein unfolding. EX 2 represents the second-order ki netics in which protein-refold ing is faster than H/Dx. Portions of the proteins must unfold and refold several times before all portions of the particular section are completely exchanged. Normally, proteins in solution follow the EX 2 mechanism, whereas H/Dx under denaturing conditions will follow the EX 1 mechanism. The EX 2 mechanism is used to calculate protein fo lding energies (freeenergy of folding) and the E1 mechanism is used to map individual conformation states that become populated upon di fferent denaturing conditions.135 For the purpose of structural proteomics, it is beneficial to study proteins which have an EX 2 mechanism. If a protein ha s an EX 1 mechanism, a bimodal distribution i.e., two separate mass envelopes, is observe d with MS analysis. This means that the

PAGE 96

79 protein being studied is not stable under the conditions used for H/Dx. For structural proteomic studies, it is important to maintain th e tertiary structure of proteins in solution during exchange.127 Operational Aspects of H/Dx HPLC-ESI-MS Experiments To initiate exchange, the protein is diluted with 99.9 % D2O at pH 7.0 (with and without substrate), quenched with formic or acetic acid at 0 oC, digested with pepsin, and analyzed with HPLC ESI-MS. Compared are peptides with the same primary sequence but different molecular weight s due to deuterium incorpora tion. The equation in Figure 3-2 is applied to calculate de uterium incorporated (D) into a protein or peptide, where mt is the protein or peptide mass at a designated exchange/quench time, m0% is the mass of the protein or peptide without deuterium incorporation, m100% is the theoretical mass of a fully deuterated protein or peptide, and N is the total number of amide protons available for exchange. D = N mt-m0%m100%-m0% [ ( )] D = N mt-m0%m100%-m0% [ ( )]D = N mt-m0%m100%-m0% [ ( )] Figure 3-2. Equation for Ca lculating the Number of Exchanged Amide Protons 131 One of the most important practical as pects of H/Dx HPLC-ESI-MS is backexchange of deuterium to hydrogen upon samp le analysis. Back-exchange is a pH, temperature, and time dependent process, thus the time needed for pepsin digestion and separation are optimized to minimize back-exchange.133 Furthermore, digestion and separation are performed at low pH and 0 oC.

PAGE 97

80 Global exchange and local exchange expe riments are designed to obtain maximum amounts of data (Figure 3-3).131 A global exchange experime nt refers to a non-digested protein, whereas a local exchange experime nt refers to a digested protein. H/D exchange of sample and quench with low pH buffer at different time intervals Pepsin digest for 30 to 45 1:1 enzyme:substrate HPLC ESI FTICR MS 4oC1 2 3 4 4oC H/D exchange of sample and quench with low pH buffer at different time intervals Pepsin digest for 30 to 45 1:1 enzyme:substrate HPLC ESI FTICR MS 4oC1 2 3 4 4oC Figure 3-3. Analytical Work Flow for H/Dx Experiment For a global exchange experi ment, three steps are needed. First, the protein is exchanged then quenched with a low pH buffer. Second, HPLC provides a quick desalt and allows fast exchanging protons found on ami no acid side chains to exchange back to protons. Lastly, proteins are analyzed w ith MS. The total mass increase between nonexchanged and exchanged protein yields the total number of solvent accessible amide protons. The local exchange protocol follo ws the same steps as the global exchange experiment; however, digestion with pepsin is added after the que nch step. Peptic peptides are mapped against the primary sequenc e of the selected protein to determine the percent coverage and possible local resolution. Hydrogen/Deuterium Exchange HPLC-ESI-M S Analysis of Selected Proteins It is the goal of structural proteomics to provide useful answers to biological questions based on well-designed experiments. Many laboratories have reported H/Dx

PAGE 98

81 HPLC-ESI-MS methodologies for describi ng protein-protein and protein-ligand interactions. Protein-protein interactions are not described in this dissertation; however, the experimental design is similar to experime nts used to solve protein-ligand interactions and dynamics. Often, H/Dx data are used to map regions of solvent accessibility on solved crystal structures. H/Dx data st rongly complements crys tal structure data; however, X-ray crystallography only presents a snap-shot of a highly dynamic protein structure. Also, the H/Dx data may desc ribe new motifs without a solved crystal structure. Finally, the requirement exists that the protein und er investigation is amendable to the experimental conditions. The following examples provided the basis for experiential design for the analysis of recombinant ST3Gal IV. First, di hydrodipicolinate reductase (28.7 kDa) was investigated in global and lo cal experiments to describe catalytically open and closed forms upon binding of NADH and inhibitors. Four peptic peptides were localized to a binding region and a hinge region needed fo r closing upon binding of inhibitor. The results clearly described the difference betw een catalytically inactive and active forms not described by the crystal structure.137 Second, diaminopimelate dehydrogenase exists as a homodimer (70.4 kDa) that binds to ei ther NADPH, diaminopimelate (DAP), or both. Using localization experiments, the exch ange of several different peptide regions under different conditions (bindi ng of NADPH, DAP, or both) resulted in different amide proton solvent accessibility. The final resu lts characterized a region that changes conformation based on which set of substrates were present.138 Third, purine nucleoside phosphorylase exists as a homotrimer (31.5 kDa) that has three separate active sites. Interestingly, when only one site was occupied by the inhibitor, Immucillin-H, the

PAGE 99

82 homotrimer was completely deactivated. Gl obular studies show that less amide protons are exchanged in the protein-inhibitor complex. Furthermore, localization studies with peptic peptides illustrated that 10 different peptides from all three subunits had different exchange rates when complexed to one equivalent of inhibitor. This study illustrated that binding of one molecule of inhibitor to one site in the homotrimer created allosteric effects in respect to the other subunits. Once again, this could not ha ve be described with X-ray crystallography, although the data wa s mapped onto the crystal structure for conformation of solvent accessible sites.139 Although there are many other systems described in the literature, they all foll ow the same H/Dx HPLC-MS protocols. Experimental Design for Anal ysis of Sialyltransferase ESI-FTICR-MS provides high resolution and the ability to trap and analyze large proteins, thus providing functional advantages over other MS inst ruments. Exact mass determination facilitates correct assignment of intact protein and peptides from pepsin digestion. Figure 3-4 provides the experi mental design for a complete study on recombinant hST3Gal IV secondary structure. First, hST3Gal IV is dena tured and exchanged, thus a llowing the percent back exchange to be calculated. Second, non-de natured ST3Gal IV is exchanged and quenched in thirty-second intervals. As opposed to the denatured ST3Gal IV case, only a certain number of amide protons are solven t accessible, thus the number of solvent accessible amide protons of non-substrate, nondenature ST3Gal IV is determined. Third, ST3Gal IV is digested with pepsin prio r to LC analysis. At this point, the peptide coverage, digestion efficienc y, and reproducibility of retent ion times are evaluated. Fourth, H/Dx data on the

PAGE 100

83 Recombinant ST3Gal IV H/Dx Pepsin Digest Substrate Introduction (1)(2)(3) (5) H/Dx Pepsin Digest Substrate Introduction (4) (6) H/Dx LC LC LC LC H/Dx Pepsin Digest LC ESI/FTICRMS H/Dx Introduction (1)(2)(3) (5) H/Dx Pepsin Digest Substrate Introduction (4) (6) H/Dx LC LC LC LC H/Dx Pepsin Digest LC Data Analysis-Bioinformatics Recombinant ST3Gal IV H/Dx Pepsin Digest Substrate Introduction (1)(2)(3) (5) H/Dx Pepsin Digest Substrate Introduction (4) (6) H/Dx LC LC LC LC H/Dx Pepsin Digest LC ESI/FTICRMS H/Dx Introduction (1)(2)(3) (5) H/Dx Pepsin Digest Substrate Introduction (4) (6) H/Dx LC LC LC LC H/Dx Pepsin Digest LC Data Analysis-Bioinformatics Figure 3-4. Six Experiments for Complete Ch aracterization of Reco mbinant ST3Gal IV constructs with H/Dx HPLC-MS substrate-ST3Gal IV complex is collected us ing the natural substr ates. The number of amide sites protected (or de-pro tected) from solvent is calculated, thus leading to the final two experiments. The last two experiments include diges tion of H/Dx protein, with and without substrate present. As stated before, these experiments identify peptides with changes of solvent accessibility. Data collected on important peptides are compared to known biological data on hST3Gal IV. Before anal yzing ST3Gal IV, conditions are optimized with standard proteins. Important stru ctural motifs and structural dynamics of recombinant hST3Gal IV may be determined by comparing H/Dx data sets of globular enzyme, global enzyme-substrate, pepsin digest of enzyme, and pepsin digest of enzymesubstrate complex.

PAGE 101

84 Results and Discussion Before performing experiments on recombinant human ST3Gal IV, several components of the analytical system had to be optimized. First, the ESI-FTICR-MS must be optimized for sensitivity and mass accuracy w ith large intact proteins before analysis of recombinant hST3Gal IV. Second, HPLC on-line with MS was optimized for quick separation of intact proteins Third, H/Dx optimization proceeds with hen egg white lysozyme (HEWL) as a standard. Fourth, HPLC-MS analysis of peptic peptides was optimized with digestion of bovine carbonic anhydrase II and bovine serum albumin. Finally, the investigation of in tact and pepsin digested recombinant ST3Gal IV proceeded with confidence. Optimization of Analytical System for Analysis of Intact Proteins Dynamic and gas assisted trapping and anal ysis of intact prot eins with ESI-FTICR MS The sensitivity, mass resolving power, and mass accuracy of the ESI-FTICR-MS was evaluated with protein standards of incr easing mass rage. The standards included bovine pancreas ribonuclease A (~13 .7 kDa), HEWL (~14.3 kDa), bovine chymotrypsinogen A (~25.6 kDa), human euka ryotic carbonic anhydrase I (28.7 kDa), bovine carbonic anhydrase II (~ 28.9 kDa), and bovine serum albumin (~66.4 kDa). The Apex II FTICR-MS, presented in Chapter 1, ha s three different modes for trapping ions. The first, SidekickTM, is normally associated with small molecule analysis. HPLC-MS analysis of peptides utilizes SidekickTM trapping. With SidekickTM trapping, as the molecule enters the Infinity cell, a voltage pulse knocks the ions off axis, thus allowing the molecules to be trapped between two endplates with ap plied DC voltages (0.60 to 1.0

PAGE 102

85 volts). For high resolving pow er, during excitation and detection, the DC voltages on the end plates must be kept below 1.0 volt. Since proteins have increased kinetic en ergy, they require an increase in cell pressure, termed Gas-assisted dynamic trapping (GADT), or an increase in the end plate voltages, termed dynamic trapping (DT). With GADT, a pulse of gas introduced into the cell to aid SidekickTM trapping of large molecules; however, the duty cycle greatly decreased because a 3 to 5 second pumping dela y that must be added in sequence before excitation and detection. Time delays of this length were not conducive for on-line HPLC-MS. Furthermore, the increased pre ssure may fragment hi gh-energy molecules. To improve the scan rate for on-line HPLC-M S, dynamic trapping (DT) was used. With DT, protein ions were trapped by first holding the end plates at increased voltages (1.7 to 3.0 volts), then dropped to 0.7 to 0.5 volts over a designated peri od in a stepwise manner. Figure 3-5 describes the DT experiment where the y-axis corresponds to the voltage applied to the end plates, the x-axis corres ponds to the number of steps towards the lower voltage (steps), and t corresponds to the am ount of time between each voltage step (step time). All three parameters were optimized for the high sensitivity and mass accuracy for large intact proteins. Experimentally, it was observed that the larger the trapping voltage, the total ion cooling time (steps multiplied by the step ti me) had to be increased. This observed phenomena is described by the adiabatic expa nsion of ions trapped within the cell.140 Ions trapped at an increased voltage contain a distribution of energies. During the ion cooling process, ions which are of high energy will be ejected from the ion packet before excitation and

PAGE 103

86 Table 3-1. Summar y of Different Trapping Methods SidekickGADTDTAs ions enter cell, voltage pulse knocks the ions off axis Relative low molecular weight species100 to 8000 Da (possible peptide analysis) High duty cycle (ms to s analysis time) Ions enter cell, pulse of gas added to help slowions All molecular weight species Low duty cycle (Need to pump gas away before excitation and detection) 3 to 5 s analysis time Ions enter cell, voltages are raised to trap ions, then lowered over ms time scale to allow ions to cool before excitation and detection. All molecular weight species Intermediate Duty cycle s analysis time SidekickGADTDTAs ions enter cell, voltage pulse knocks the ions off axis Relative low molecular weight species100 to 8000 Da (possible peptide analysis) High duty cycle (ms to s analysis time) Ions enter cell, pulse of gas added to help slowions All molecular weight species Low duty cycle (Need to pump gas away before excitation and detection) 3 to 5 s analysis time Ions enter cell, voltages are raised to trap ions, then lowered over ms time scale to allow ions to cool before excitation and detection. All molecular weight species Intermediate Duty cycle s analysis time t t t 0.6 2.0 Number of Steps = 3 123 0 Number of StepsDC Voltage on End Plates in Volts t t t 0.6 2.0 Number of Steps = 3 123 0 Number of StepsDC Voltage on End Plates in Volts Figure 3-5. Dynamic Trapping Experiment. detection. If the trapping voltages were too large and total io n cooling time was too short, poor resolution and mass accur acy was observed. The observed peaks had apparent shoulders from the high energy mol ecules still in the cell. The optimal end plate trapping voltages for protein standard s varied between 1.8 to 3.0 volts, with the excitation and detection end pl ate voltage at 0.7 volts. T ypically, increased trapping voltages were needed for increased molecular weights. For HPLC-MS analysis of intact

PAGE 104

87 proteins, trapping voltages of 2.40 to 2.65 volts were adopted for the entire distribution of protein standards. The number of steps for ion cooling had minimal effects on the resolution and mass accuracy as long as th e total ion cooling time was between 0.7 and 1.0 seconds. If the total ion cooling time wa s above 1.5 seconds, sensitivity suffered. Including the DT parameter considerations for sensitivity, as described in Chapter 1, ions generated by ESI were first collected with a hexapole over a defined time period (Hex Acc Time) before being pulsed to the ICR cell. The optimal time for hexapole accumulation was found to be between 0.01 and 2.00 seconds. No clear trends in the hexapole accumulation time versus sensitivit y were observed. With DT, several ion pulses from the hexapole were collected in the cell before excitation and detection (load cell). Typically, 3 to 10 load cell cycles were required for increased sensitivity. Also, the infusion rate of the sample must be considered. The ESI source was capable of providing stable electrospray at flow rates of 0.5 to 4.0 L/min. Lastly, slight changes in transfer optics were observed over several mont hs. If the optics needed to be tuned with drastic changes in voltages, this indicated th at the source optics were dirty or there was electrical issues related to the transfer opt ic voltage boards. Ta ble 3-2 represents a typical optimization experiment with HEWL as a standard (selecting the +10 charge state). All six parameters were tuned in a round robin fashion for optimal conditions. The infusion rate for all experiment s described in this section was 2 L/min. Three generalities may be extracted from the small data set presented in Table 3-2. First, the resolution decrease with increasing trapping volta ge. Second, the sensitivity decreased with increased total ion-cooling time Since more ions were ejected from the ion packet over time, fewer molecules were present for detection. Third, poor peak shape

PAGE 105

88 was observed if the ion-cooling time was too sh ort or if the step time was below 0.01 s. Lastly, the lower time limit for the DT hardware was 0.01 seconds per step. Table 3-2. Relationship Between the Six Impor tant Parameters and Signal Intensity/ Resolution of the Isotopically Resolved +10 Charge State of HEWL Initial Plate Voltage Final Plate Voltage Voltage Step Time / Steps Total Ion Cooling Time Hex Acc Time Load Cell Relative Intensity Resolution 2.50 V 0.50 V 0.02s / 50 Steps 1.00 s 1 5 2500000 43,000 2.50 V 0.50 V 0.01s / 50 Steps 0.50 s 1 5 poor peak shape 2.50 V 0.50 V 0.03s / 50 Steps 1.50 s 1 5 2000000 43,000 3.00 V 0.50 V 0.03s / 50 Steps 1.50 s 1 5 3000000 37,000 2.00 V 0.50 V 0.03s / 50 Steps 1.50 s 1 5 1500000 43,000 2.00 V 0.50 V 0.02s / 50 Steps 1.00 s 1 5 2000000 43,000 2.00 V 0.50 V 0.005s / 200 Steps 1.00 s 1 5 poor peak shape 2.00 V 0.50 V 0.02s / 200 Steps 4.00 s 1 5 1000000 40,000 2.00 V 0.50 V 0.005s / 200 Steps 1.00 s 1 5 poor peak shape 2.00 V 0.050 V 0.04 s / 25 Steps 1.00 s 1 5 1500000 43,000 Another practical consideration was the acquisition file size collected. The acquisition file size may be 8 k, 32 k, 128k, 256 k, 512 k, or 1 M data points. Theoretically, the larger the file size, th e longer the acquisition time, the better the resolution. Higher mass resolution was obtaine d with smaller molecules because the decay of the transient signal was slower than that of larger molecules. As the molecular weight of the molecule increases, the transi ent signal decays faster. Of the proteins

PAGE 106

89 analyzed, only HEWL was isotop ically resolved. For all th e other proteins, the charge states could be easily resolved with high mass accuracy and resolution. The radio frequency (rf) and the duration of attenuation for excitation was also tuned. If the rf power attenuation was too lo w and/or duration too short, poor signal was observed because the ions were not excited to the full radial orbit. If the rf power attenuation was too high and/or the duration too long, the ions were excited into an orbit that caused collisions with the ICR cell walls. Several different rf power attenuation and durations were investigated for optimal signal intensity; however, isotopic resolution of the large proteins was not obtained. To re iterate, even though isotopic resolution was not achieved, mass accuracy and sensitivity were not relinquished. The fast decay of the transient signal ma y be attributed to dampening due to collisions with residual gas, collisions with the end plates, and magnetron motion. Bovine serum albumin and carbonic anhydras e have large cross-sectional areas; therefore, they were more likely to collide with residual gas in the cell than smaller molecules. Also, large molecules with high er charge-states incr eased the likely hood of coalescence. For the larger proteins, 32 k files were collected with excellent mass accuracy after deconvoltion of the charge st ate spectra. The following figures represent typical spectra collected for protein standard s. Table 3-3 represents the mass accuracies of intact proteins analyzed with direct infusion ESI-FTICR-MS. All the spectra observed have charge state resolution and fall between 1000 to 2500 m/z. All proteins were diluted in 50% methanol, 50% water, 0.1 % acetic acid except carbonic anhydrase which was diluted in 20 % methanol, 80 % water, 0.1 % acetic acid. The spectrum presented in Figure 3-7 was collected with GADT, whereas a ll other spectra were collected with DT.

PAGE 107

90 The proteins analyzed with direct infu sion ESI FTICR MS were diluted to the given concentration with no further purifica tion. The mass errors associated with different protein standards was most likely due to impurities and salts left over from the purification. For example, carbonic anhydrase exhibited a 40 Da difference between the theoretical and experimental mass. Most likely, the mass error was due to potassium present in the lyophilized product. As w ill be seen with HPLC-MS, the mass error decreased when the sample was desalted. Table 3-3. Mass Accuracies for Different Protein Standards Standard Protein Concentration Hex Acc Load Cell Exp. m/z Theoretical m/z Error in ppm Bovine Pancreas Ribonuclease A 5 pmol/ L 0.50s10x 13,681.113,681.3 15 HEWL 10 pmol/ L 0.01s5x 14,301.914,298.3 250 Human Carbonic Anhydrase I 10 pmol/ L 0.50s10x 28776.7 28734.4 1300 Bovine Serum Albumin 20 pmol/ L 1.00s5x 66431.9 66,430.0 28 Broadband versus heterodyne excitation and detection The FTICR-MS may be utilized in broa dband or heterodyne excitation/detection modes. Up to this point, only broadband ex citation and detection was used for intact protein analysis. Broadband excitation and detection is the gene ral mode of operation which excites ions with a frequency sweep over a broad frequency range corresponding to 200 to 5000 m/z. To achieve increas ed resolution (from 30,000 ca to 100,000 ca for HEWL), heterodyne mode was utilized. W ith heterodyne mode, a small mass-to-charge window may be analyzed with a smaller sa mpling file (32 k versus 1 M). Simply,

PAGE 108

91 500 1000 1500 2000 2500 3000 3500 0 50000 100000 150000 200000 250000 300000 350000+8 +7 +9 +6 500100015002000250030003500 0 50000 100000 150000 200000 250000 300000 350000+8 +7 +9 +6 Figure 3-6. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of Ribonuclease A where: Hex Acc = 0.5 s; Steps = 35, Ion Cooling Time = 0.7 s; Infusion Rate = 2 L/min; Step Time = 0.02s; Load Cell = 10x; Initial Trap Voltage = 1.85 V; Detection Trap Voltage = 0.5 V. Broadband Mode; and the Deconvoluted Molecular Weight = 13, 681.1.

PAGE 109

92 0 500000 1000000 1500000 2000000 2500000 3000000 3500000 4000000 4500000 1200 1700+10 +9 +8 +11 0 500000 1000000 1500000 2000000 2500000 3000000 3500000 4000000 4500000 12001700+10 +9 +8 +11 Figure 3-7. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of HEWL where: Hex Acc = 0.01s; Infusion Rate = 2 L/min; Step Time = 0.02s; Load Cell = 5x; Gas Assisted Dynamic Tra pping; Broad Band Mode; and the Deconvoluted Molecular Weight = 14,301.9

PAGE 110

93 1430.5 1431.0 1431.5 1432.0 m/z 0 500000 1000000 1500000 2000000 2500000 3000000 3500000 4000000 4500000 Figure 3-8. Expanded +10 Charge State of the Spectra Presented in Figure 3-7. heterodyne detection reducing the sampling frequency of th e ICR signal with frequency mixing in regards to a reference frequency, t hus reducing the digitiza tion rate needed to satisfy the Nyquest sampling criterion. This means that there was a smaller sampling data set (32k versus 1 M) needed for a longer detection period. Figures 3-11 and 3-12 present the comp arison between broadband and heterodyne mode spectra and the resulting mass resolution of the +10 charge state of 10 pmol/ L HEWL. To initiate heterodyne mode, the APEX II FTICR-MS manual contains simple instructions for changing between the modes. Using heterodyne on la rger proteins than

PAGE 111

94 500 1000 1500 2000 2500 3000 0 50000 100000 150000 200000 250000 300000 350000+19 +20 +21 +22 +23 +17 +16 +15 +14 50010001500200025003000 0 50000 100000 150000 200000 250000 300000 350000+19 +20 +21 +22 +23 +17 +16 +15 +14 Figure 3-9. Direct Infusion of ESI FTICR-MS Analysis of 15 pmol/ L of Human Carbonic Anhydrase I where: Hex Acc = 0.5 s; Steps = 35, Ion Cooling Time = 0.7 s; Infusion Rate = 2 L/min; Step Time = 0.02s ; Load Cell = 10x; Initial Trap Voltage = 1.85 V; Detection Tr ap Voltage = 0.5 V. Broadband Mode; and the Deconvoluted Mol ecular Weight = 28, 776.7. HEWL did not result in higher resolution. This was because HEWL had a slower cyclotron decay rate in comparison to larger proteins. The benefits of heterodyne mode

PAGE 112

95 1400 1900 2400 2900 m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000+29 +28 +27 +30 +31 +32 +33 +34 +36 1400190024002900m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000+29 +28 +27 +30 +31 +32 +33 +34 +36 Figure 3-10. Direct Infusion ESI FTICR-MS Analysis of 20 pmol/ L of Bovine Serum Albumin where: Hex Acc = 1.0s; Step s = 50, Ion Cooling Time = 0.7 s; Infusion Rate = 2 L/min; Step Time = 0.02s; Load Cell = 5x; Initial Trap Voltage =2.85 V; Detection Trap Vo ltage = 0.5 V; Broadband Mode; and the Deconvoluted Molecular Weight = 66,431.9. were only observed if the transient was longlived. Heterodyne m ode was initiated to reduce the file size for on-line HPLC-MS and increase resolving powers for analysis of recombinant hST3Gal IV. As will be seen in the HPLC-MS section, one of the most important limitations

PAGE 113

96 1400 1600 1800 m/z 0 500000 1000000 1500000 2000000 2500000 3000000 3500000 1430.9 1431.4 1431.9 m/z 0 500000 1000000 1500000 2000000 2500000 +11 +10 +9 +8 140016001800m/z 0 500000 1000000 1500000 2000000 2500000 3000000 3500000 1430.91431.41431.9m/z 0 500000 1000000 1500000 2000000 2500000 +11 +10 +9 +8 Figure 3-11. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of HEWL with Expanded +10 Charge State. Broadba nd Mode Excitation a nd Detection with a 40,000 Mass Resolution for the Isotopica lly Resolved +10 Charge State. of our instrument was data storage due to the file size collected during an HPLC-ESIFTICR-MS run. Further explanation of the details corresponding to the scan rate and HPLC-MS will be covered in th e next section. The improved resolution was thought to benefit the analysis of reco mbinant ST3Gal IV; however, he terodyne mode analysis of

PAGE 114

97 1380 1400 1420 1440 1460 1480 m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000 1430.4 1430.9 1431.4 1431.9 m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000A B 138014001420144014601480m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000 1430.41430.91431.41431.9m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000A B Figure 3-12. Direct Infusion ESI FTICR-MS Analysis of 10 pmol/ L of (A) HEWL with (B) Expanded +10 Charge State. Hete rodyne Excitation and Detection with a 90,000 Mass Resolution for the Isotopical ly Resolved +10 Charge State.

PAGE 115

98 large proteins did not prove us eful or realistic. Thus, this method was abandoned for the purpose of high resolution pertaining to la rge proteins. Finally, heterodyne mode provided useful data for the analysis of H/ Dx HPLC-ESI-FTICR-MS of intact HEWL as presented later in this chapter. Optimization of HPLC-ESI-FTICR-MS After optimizing the instrument with di rect infusion ESI-FTICR-MS of standard proteins, the analysis of in tact proteins with quick se paration on-line HPLC-MS was performed. For chromatogra phy of intact proteins, a C4 column was purchased from Grace Vydac (Hesperin, Ca). The optimized gradient and flow rates are described in the Methods and Materials section. The following figures represent the spectra obtained on proteins after optimization of separation conditions and scan rates. Separation and MS analys is of intact proteins HPLC-MS was initiated by defining the num ber of frames (NF) collected. The total acquisition time for an HPLC-MS run was defined by the scan rate multiplied by the number of frames. The scan rate was de fined by the hexapole accumulation time, load cell cycle, and file size. In cluded within the frames collected, there was the possibility for signal averaging of multiple scans. For inta ct protein analysis, it was not necessary to signal average because of the load cell f unction. Signal averaging was utilized for HPLC-MS analysis of peptic peptides. Typically, the scan rate was 10 to 14 seconds/frame. This may appear to be a slow scan rate; however, the scan rate was the product of cell cycles and hexa pole accumulation time. Over the 10 to 14 second/frame, ions were collected in the cell prior to ex citation and detection. If GADT was used for HPLC-MS analysis, 3 to 4 seconds of informa tion would be lost. Typically, a 30 minute analysis included 150 frames. To illustrate th e utility of this experimental design, the

PAGE 116

99 following spectra were obtained on different protein standards. HEWL, under the conditions described in Figure 3-12, eluted at about 17 minut es with the apex at 19.15 minutes. In Table 3-4, note the improved mass error of the de-salted HEWL (60 ppm) in comparison to the directly infused HEWL (250 ppm). 0 2 4 6 8 10 12 14 16 18 20 22 24 min Figure 3-13. TIC of HEWL (75 pmol) Anal yzed with HPLC-ESI-FT-ICR-MS where: Hex = 1.25 s, Steps = 25; Ion Cooling Time = 1.0 s; Flow Rate = 2 L/min; Step Time = 0.04s; Load Cell = 5x; Initial Trap Voltage = 1.65 V; Final Trap Voltage = 0.5 V, NS = 1; Scan Rate = 10.44 s /scan; Acquisition File Size = 32 k, and with Broadband Excitation and Detection. Table 3-4. Mass Accuracies for Different Pr oteins with HPLC-ESI FTICR-MS Analysis Standard Protein Amount Analyzed Hex Acc Load Cell Experimental m/z Theoretical m/z Error in ppm Bovine Pancreas Ribonuclease A 10 pmol 1.0 s 5x 13, 680.32 13, 681.32 70 HEWL 10 pmol 1.0 s 5x 14, 299.78 14, 298.32 100 Chymotrypsinogen 10 pmol 0.5 s 5x 25, 645.50 25, 649.42 150 Carbonic Anhydrase I 10 pmol 0.5 s 5x 28, 731.50 28.734.22 110 To test the mass accuracy of the instrume nt, an in-house expressed protein, uridine kinase, was analyzed with the HPLC-ESI-FTICR-MS (Figur es 3-18 and 3-19). Based

PAGE 117

100 500 1000 1500 2000 2500 m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000+8 +9 +10 +11 5001000150020002500m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000+8 +9 +10 +11 Figure 3-14. Mass Spectrum at 19.15 mins from the TIC Presented in Figure 3-13. The Deconvoluted Spectra Molecular Weight = 14,299.9 the cDNA sequence, the theoretical ma ss was 24,353 Da, thus there was a mass difference of 138 Da between the deconvoluted spectra mass and theoretical mass. The mass difference may be due to mutations in the protein upon cloning, su ch as clipping at

PAGE 118

101 2 4 6 8 10 12 14 16 18 20 m 9 0 9 2 9 4 9 6 9 8 i. Figure 3-15. TIC of HEWL (50 pmol) and -Chymotrypsinogen A (50 pmol) Analyzed with HPLC-ESI-FT-ICR-MS the N-terminus methionine, or there was a shift in experimental m/z due to space charging from over loading the column and the ICR cell. Furthermore, when looking at the deconvoluted spectra, th ere was a mass shift of 98 Da corresponding to either phosphate adducts or possible phosphorylation. Next, hete rodyne mode was optimized with the +10 charge state of HEWL (Figure 3-15). This experiment preceded the H/Dx HPLC-MS experiment on HEWL H/Dx Direct Infusion and H/Dx HPLC -ESI-FTICR-MS of Intact HEWL Figures 3-22 and 3-23 represent the spectra collected (by direct infusion) after 20 M of HEWL was incubated with D2O and quenched at 4 oC with a buffer containing 20% methanol, 80 % H2O, and 2 % acetic acid at 1, 2, 5, 10, 15, and 20 minutes. All the spectra that were collected at the different quench times had the same m/z envelope. After deconvoltion, it was estimated that ther e were 110 exchanged protons. According to the literature, there shoul d only be 62 solvent assessable amide protons. Based on the

PAGE 119

102 500 1000 1500 2000 2500 3000 3500 4 0 20000 40000 60000 80000 100000 120000 140000+9 +10 +8 +7 Figure 3-16. HEWL Spectrum Observed at 12.30 Minutes (Figure 3-15). peak shape and the number of amide protons for exchange, the amino acid side chains retained deuterium upon ionization. When HPLC-ESI-MS was performed on exchanged HEWL, the +10 state was isotopically resolved.

PAGE 120

103 1000 1500 2000 2500 3000 m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000+11 +10 +12 +7 10001500200025003000m/z 0 20000 40000 60000 80000 100000 120000 140000 160000 180000+11 +10 +12 +7 Figure 3-17. -Chymotrypsinogen Spectrum Observed at 12.93 Minutes (Figure 3-15). Figure 3-25 represents the overlaid hete rodyne spectra of an H/Dx HPLC-ESIFTICR-MS (with the same reaction conditions described for Figure 3-22) and an HPLCMS experiment with HEWL. The chromat ogram shown in Figure 3-24 was reproducible for non-exchanged and exchanged HEWL. Spectrum 1 represents the non-H/Dx heterodyne spectrum of HEWL and spectrum 2 represents the H/Dx heterodyne spectrum

PAGE 121

104 2 4 6 8 10 12 14 16 18 20 22 m 80 85 90 95 Figure 3-18. TIC of Unknown Amount of Uridine Kinase Analyzed with HPLC-MS ESI-FTICR-MS for HEWL. According to the data, there are 44 amide protons exchanged to deuterium. According to the literature, investigat ors reported 62 exchanged amide protons. 141 According to our data, we have not yet completely optimized our H/Dx conditions; however, the results show that H/Dx of HEWL was observable with ou r analytical set-up. With this positive result, more attention was given to pepsin digestion of proteins and recombinant hST3Gal IV analysis. Separation and MS Analysis of Pepsin Digested Proteins Pepsin digests were analyzed with HPLC-MS on either a C4 or a C18 column purchased from Grace Vydac (Hesperin, Ca). Chromatography with the C4 column did not yield high separation efficiency; however, with the exact mass capability of the ESIFTICR MS, several peaks which co-eluted were resolved and identified using the protein mass fingerprinting algorithm described in Ch apter 1. Since quick separation was one of the requirements for the H/Dx HPLC experiment, many investigators use C4 columns as opposed to C18 columns because the C18 columns provide increased retention times. With the C4 column, most of the peptides eluted within the firs t minute of the gradient. When

PAGE 122

105 500 1000 1500 2000 2500 3000 3500 4000 4500 m/ z 0 20000 40000 60000 80000 100000 120000 140000+15 +14 +13 +11 +10 +18 +16 0 0243002450024700Deconvoluted Spectra 50010001500200025003000350040004500m/ z 0 20000 40000 60000 80000 100000 120000 140000+15 +14 +13 +11 +10 +18 +16 0 0243002450024700Deconvoluted Spectra Figure 3-19. Uridine Kinase Spectrum Observed at 12.00 Minutes (Figure 3-18). performing HPLC-MS analysis of a pepsin dige st, the file size and number of scans (NS) per frame were most important. The analysis of peptides used SidekickTM trapping, thus tuning was less difficult than DT. Simply, before each HPLC-MS run, signal intensity was maximized with Agilents HP standard mix. Unlike intact proteins, the peptic pept ides have masses between 600 and 4000 Da, thus they had a long transient decay times after excitation and we re easily monitored

PAGE 123

106 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 min 0.70 0.80 0.90 1.00 Figure 3-20. TIC of HEWL (25 pmol) An alyzed with HPLC-ESI-FT ICR-MS in Heterodyne Mode and Dynamic Trapping 1430.0 1431.0 1432.0 m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000 Figure 3-21. HEWL Spectrum Observed at 14.00 Minutes (Figure 3-20).

PAGE 124

107 1400 1500 1600 m/z 0 100000 200000 300000 400000 500000 600000+10 +9 140015001600m/z 0 100000 200000 300000 400000 500000 600000+10 +9 Figure 3-22. H/Dx Analysis of HEW EL with Direct Infusion ESI-FTICR-MS with a 1 M acquisition file. The typica l scan rate for HPLC-MS with SidekickTM trapping was between 0.5 and 2.5 seconds/scan, with th e time per frame defined by the hexapole accumulation time. If the scan rate was 2.5 seconds/scan, the total acquisition time was 40 minutes, and a 1 M file was used, the final total file size would be 960 M! This value was well outside the range of the memory of our data acquisition system. To decide which acquisition file size would provide the most realistic chromatic file size with minimal loss in mass accuracy, resolution, and sensitivity, HPLC-MS acquisition was performed with 32 k, 128k, and 256 k files. Th e number of scans, hexapole accumulation time, acquisition file size, and scan rate were optimized for sensitivity on pepsin digested HEWL, carbonic anhydrase, and BSA.

PAGE 125

108 1439.0 1440.0 1441.0 1442.0 1443.0 m/z 0 50000 100000 150000 200000 250000 300000 350000 400000 450000 500000 550000 600000 Figure 3-23. Expanded +10 Charge State Observed in Figure 3-22 After several different acquisitions, a 256 k file size provided the best trade-off for mass accuracy, resolution, and sensitivity. Figure 3-26 represents the sequence coverage for 5 pmol/ L of carbonic anhydrase digested with 3 pmol/ L of pepsin. A total of 3 L was loaded on to the C4 column, thus 15 picomoles were analyzed. Table 3-5 represents the peptides observed below a mass error of 50 ppm. There were several other peptides observed with mass errors between 50 and 200 ppm; however, they were of low signal intensity. Since the file size collected wa s 254 k, these low intensity peaks (above three times signal to noise) were not pr operly resolved. If a larger file size was collected, these

PAGE 126

109 2 4 6 8 10 12 14 16 18 20 22 24 26 28 min Figure 3-24. TIC of H/Dx-HEWL Sample ( 10 pmol) Analyzed with HPLC-ESI FTICR MS with Heterodyne Mode and Dynamic Trapping 1431 1433 1435 m/z 1.0e+05 2.0e+05 3.0e+05 4.0e+05 5.0e+05 6.0e+05 7.0e+05 8.0e+05 9.0e+05 1.0e+06 1.1e+06 +10 Charge StateMass Spectra of Non-H/D exchanged Lysozyme MWavg= 14,301 1Mass Spectrum of H/D exchanged Lysozyme MWavg= 14,345 2 4.4 m/z 44 Da 1431 1433 1435 m/z 1.0e+05 2.0e+05 3.0e+05 4.0e+05 5.0e+05 6.0e+05 7.0e+05 8.0e+05 9.0e+05 1.0e+06 1.1e+06 +10 Charge StateMass Spectra of Non-H/D exchanged Lysozyme MWavg= 14,301 1Mass Spectrum of H/D exchanged Lysozyme MWavg= 14,345 2 4.4 m/z 44 Da Figure 3-25. Spectra of (1) Non-H/Dx of 10 pmol of HEWL and (2) 15 min H/Dx of 10 pmol of HEW Analyzed with HPLC-E SI FTICR MS with Heterodyne Mode and Dynamic Trapping.

PAGE 127

110 low intensity peaks would have been better resolved. The sequence coverage presented in Figure 3-26 and Table 3-5 were manua lly identified with the theoretical in silico digestion of bovine carbonic anhydrase II wi th pepsin and the MS-Digest algorithm (http://prospector.ucsf.edu/ucsf html4.0/msdigest.htm). Intere stingly, digestion of HEWL with pepsin under the same conditions as bovine carbonic anhydrase II did not result in peptic peptides. Analysis with a C18 column provided separation of pe ptides; however, the realistic maximum acquisition file size was 128 k because the retention times of the peptides were greatly increased. A 45 minute acquisition w ith a 128 k acquisition f ile size created a 64 M file. After successful analysis and identification of 15 picomoles of carbonic anhydrase II digested with pepsin, attention was switched to the analysis of recombinant ST3Gal IV. Mass accuracies and percen t sequence coverage of bovine carbonic anhydrase II and bovine serum albumin could ha ve been improved; however, the analysis of recombinant hST3Gal IV proceed with conf idence at this level of sensitivity and mass accuracy. SHHWGYGKHNGPEHWHKDFPIANGERQSPVDIDTKAVVQDPALKPLALVYGEATSRRMVN NGHSFNVEYD DSQDKAVLKDGPLTGTYRLVQFHFHWGSSDDQGSEHTVDRKKYAAELHLVHWNTKYGDFGTAAQQPDGL AVVGVFLKVGDANPALQKVLDALDSIKTKGKSTDFPN FDPGSLLPNVLDYWTYPGSLTTPPLLESVTWIVLKE PISVSSQQMLKFRTLNFNAEGEPELLMLANWRPAQPLKNRQVRGFPK SHHWGYGKHNGPEHWHKDFPIANGERQSPVDIDTKAVVQDPALKPLALVYGEATSRRMVN NGHSFNVEYD DSQDKAVLKDGPLTGTYRLVQFHFHWGSSDDQGSEHTVDRKKYAAELHLVHWNTKYGDFGTAAQQPDGL AVVGVFLKVGDANPALQKVLDALDSIKTKGKSTDFPN FDPGSLLPNVLDYWTYPGSLTTPPLLESVTWIVLKE PISVSSQQMLKFRTLNFNAEGEPELLMLANWRPAQPLKNRQVRGFPK SHHWGYGKHNGPEHWHKDFPIANGERQSPVDIDTKAVVQDPALKPLALVYGEATSRRMVN NGHSFNVEYD DSQDKAVLKDGPLTGTYRLVQFHFHWGSSDDQGSEHTVDRKKYAAELHLVHWNTKYGDFGTAAQQPDGL AVVGVFLKVGDANPALQKVLDALDSIKTKGKSTDFPN FDPGSLLPNVLDYWTYPGSLTTPPLLESVTWIVLKE PISVSSQQMLKFRTLNFNAEGEPELLMLANWRPAQPLKNRQVRGFPK Figure 3-26. Sequence Coverage of Bovine Carbonic Anhydrase Digested with Pepsin MS Analysis of Recombinant Sialyltransferase Ins-ST, N-Tag-ST, C-Tag-ST, and comm ercially prepared recombinant rat (2 6)-(N)-sialyltransferase were analyzed with either direct infu sion ESI-FTICR-MS or

PAGE 128

111 Table 3-5. Peptides Observed from Bovine Carbonic Anhydrase Dige sted with Pepsin Presented in Figure 3-26 ExperimentalTheoretical Da Error in ppmPeptide 541.3724541.37080.00163LKPLA 591.3510591.35000.00102AVVGVF 641.3879641.37330.014623TTPPLL 650.363650.35370.009314RTLNF 658.4148658.41430.00051KVLDAL 666.3630666.3643-0.00132QMLKF 786.4794786.46520.014218QKVLDAL 786.4794786.46520.014218LQKVLDA 804.4582804.45150.00678LHLVHW 823.4004823.39850.00192WTYPGSL 827.4448827.44100.00385HKDFPIA 844.3992844.38350.015719NTKYGDF 846.4048846.39110.013716PNFDPGSL 858.3871858.36880.018321NAEGEPEL 957.4641957.46040.00374GTAAQQPDGL 971.4801971.46800.012112NAEGEPELL 997.5661997.55800.00818LKVGDANPAL 1019.55591019.51960.036336LPNVLDYW 1132.65291132.7152-0.062355KFRTLNFNA 1153.54561153.50310.042537YDDSQDKAVL 1174.74481174.68300.061853AVVGVFLKVGDA 1326.7321326.7632-0.031224DSIKTKGKSTDF 1405.68701405.67530.01178TSRRMVNNGHSF 1430.77681430.77500.00181HTVDRKKYAAEL 1488.57961488.57740.00221HFWHGSSDDQGSE 1516.75661516.73310.023515HLVHWNTKYGDF 1659.81321659.79600.017210VLKDGPLTGTYRLVQ 2154.08502154.07130.01376DSIKTKGKSTDFPNFDPGSL HPLC-ESI-FTICR-MS. The starting buffers and molecular weights associated with each construct are presented in Table 3-6. For ES I, the buffers presented in Table 3-6 were detrimental to ionization efficiency. Fu rthermore, for on-line HPLC-ESI-FTICR-MS with the C4 column, glycerol and triton were de trimental for column life-time and separation efficiency. Lastly, as presented in Chapter 3, there was a large decrease in activity after solvent exchange by dialysis or ultrafiltration. To facilitate direct infusion ESI-FTICR MS analysis of recombinant sialyltransferase, several different experime nts were performed to exchange the enzyme into solvent systems compatible with ES I. First, using Amicons Ultrafree-MC

PAGE 129

112 Table 3-6. Sialyltransferase Construct Concentration, Buffer, and Molecular Weights Construct Starting Buffer (Starting Concentration) Molecular Weight In-house prepared Ins-ST 350 mM KCl 50 mM MES, pH 6.8 20% glycerol 0.01% triton CF 54 (6 M) 34,004 Da without glycans Between 34,000 and 37,000 Da by SDSPAGE In-house prepared N-Tag-ST 100 mM KCl 50 mM MES, pH 6.8 135 mM imidazole 20% glycerol 0.01% triton CF 54 (6.5 M) 34,941 Da without glycans Between 35,000 and 38,000 Da by SDSPAGE In-house prepared C-Tag-ST 100 mM KCl 50 mM MES, pH 6.8 135 mM imidazole 20% glycerol 0.01% triton CF 54 (7.0 M) 34,941 Da with out glycans Between 35,000 and 38,000 Da by SDSPAGE Commercially prepared rat (2 6)-(N)-sialyltransferase 250 mM NaCl 50 mM MES, pH 6.2 0.5 mM 2-mercaptoethanol 50 % glycerol (5.5 M) ~41,000 Da Centrifugal Filter Units (Durapor e PVDF membrane;10,000 MWCO; 400 L capacity), Micron Centrifugal Filter Devices (MI CRON YM-10; regenerated cellulose; 10,000 MWCO; 450 L capacity), or Centriprep (YM -10; regenerated cellulose; 10,000 MWCO; 20 mL capacity), enzyme was diluted with either deionized water, 25 and 50 mM ammonium acetate at pH 7.0, 50 mM ME S pH 6.8, or 0.1 % acetic acid prior to ultrafiltration, Unfortunatel y, activity was lost upon exchange into these buffers. Furthermore, other organic acids, organic base s, or organic solvents could not be used with ultrafiltration membranes. For di rect infusion ESI-FT-ICR MS of solvent exchanged enzyme, triton and glycerol were the only components observed in the spectra.

PAGE 130

113 Second, 100 L of N-Tag-ST or C-Tag-ST was precipitated with trichloroacetic acid (TCA). The precipitated recombinant si alyltransferase was diluted in one of the following buffers: (a) 50 % trifluoroeth anol, 50 % deionized water; (b) 50 % trifluoroethanol, 49 % deionized water, 1.0 % acetic acid; (c ) 25 % trifluoroethanol 75 % deionized water; (d) 25 % methanol, 74 % deionized water, 1.0 % acetic acid; (e) 50 % methanol, 49 % deionized water, 1 % acetic acid; or (f) 99 % deionized water, 1.0 % acetic acid; or 99 % methanol, 1.0% acetic acid. Analysis of enzyme diluted with these solvent systems did not facilitate positive identif ication of sialyltransferase. As with the solvent exchange experiments, only surfactants were observed in the spectra after direct infusion ESI-FT-ICR-MS analysis. Lastly, sialyltransferase was analyzed with on-line HPLC-ESI-FTICR-MS with the C4 column. On separate occasions, 5 to 10 L of exchanged or crude sialyltransferase (25 to 70 picomoles) was injected onto the C4 column. Either water/methanol/acetic acid or water/acetonitrile/formic aci d mobile phase systems were used for elution of intact sialyltransferase constructs. Once again, only triton and glycerol were observed in the spectra. Furthermore, the analysis of cr ude sialyltransferase samples destroyed the column efficiency. After injection of crude sialyltransferase, detergent peaks were observed with subsequent HPLC-MS runs. The triton bleed off from the column inhibited the ionization of other large molecules. Besides problems with the solvents n eed for sialyltransferase stability, the recombinant enzyme was expressed as at leas t three glycoforms (SDS-PAGE analysis). Obviously, for MS analysis, the added hetero geneity with different molecular weights lowered the effective starting concentration (~5 M) of recombinant sialyltransferase to

PAGE 131

114 that of the relative concentrations of the gl ycoforms. To reduce th e heterogeneity of the Ins-ST construct, it was deglycosylat ed with either PNGase F or Endo Hf. PNGase F cleaves sugars at the core of N-li nked oligosaccharides, whereas Endo Hf cleaves one sugar up the glycan chain from the N-linkage si te. It was postulated that Ins-ST may be deglycosylated by Endo Hf, while retaining enzyme structure and activity. EndoHf and PNGase F digestion of intact Ins-ST Both, Endo Hf and PNGase F digests were in cubated with sialyltransferase (Chapter 2 Figure 2-3). Figur e 3-25 represents the Endo Hf deglycosylation reaction of Ins-ST while monitoring activity during the time course of the digestion. The activity reported was normalized to the original ac tivity before addition of buffers, Endo Hf, or heat. Figure 3-27 represents the activity of Ins-ST at different time points during digestion of glycans with Endo Hf where, Ins-ST and Endo Hf reaction corresponds to 88 L of Ins-ST diluted with 10 L of 10x Endo Hf reaction buffer plus 2 L of 20 units/ L Endo Hf. The buffer control corresponds to 88 L of Ins-ST diluted with 12 L of 10x of Endo Hf reaction buffer. The positive control was the stock solution of Ins-ST. The Endo Hf control corresponds to was 88 L of water diluted with 10 L of 10x Endo Hf reaction buffer plus 2 L of 20 units/ L of Endo Hf. All samples were incubated at 37 oC and assayed at 1 hour intervals. The data illustrates that the activity of In s-ST drops to 0% in the buffer control and the Ins-ST/ Endo Hf reaction. Also, the activity of the positive control drops to 50 % over three hours after being heated to 37 oC. Ins-ST was digested with Endo Hf at 25 oC and 4 oC; however, SDS-Page analysis did not re veal the loss of glyc osylation. Lastly, recombinant hST3Gal IV that was deglycosylated with Endo Hf or PNGase F and

PAGE 132

115 0 20 40 60 80 100 120 00.511.522.53Deglycosylation Time in HoursPercent Original Activity Ins-ST and Endo Hf Reaction Buffer Control Positive Control Endo Hf Control Figure 3-27. Activity of Ins-ST / Endo Hf Digestion analyzed with HPLC-ESI-FTICR-MS did not yield signal corresponding to sialyltransferase. Since intact sialyltransferase cannot be analy zed with direct infusion or HPLC-ESI-FTICR-MS, the digestio n of sialyltransferase with pepsin was performed. MS analysis of pepsin di gested sialyltransferase Before abandoning the H/Dx experiment s completely, pepsin digestion of recombinant sialyltransferase was investigat ed. Because the enzyme was sensitive to ultrafiltration, 5 M recombinant sialyltransferase was digested with 3.5 M pepsin at pH 2.2 (50 L final volume). ESI-FTICR-MS and LCQMS analysis did not reveal peptic peptides. The absence of peptides may be fo r several reasons. Fi rst, the glycans may inhibit digestion if the protei n has not been denatured. S econdly, several non-denatured proteins are known to be proteas e resistant, even without glyc osylation present. Lastly, recombinant sialyltransferase may precipitate at low pH. At this point, it was determined that H/Dx was not the correct experiment al choice for analysis of recombinant sialyltransferase because the threshold betw een H/Dx of recombinant sialyltransferase and ionization to the gas phase could not be met.

PAGE 133

116 Conclusion The ESI-FT-ICR MS was used to analyze standard intact proteins and pepsin digested standard proteins. The different im portant parameters have been optimized for proteins up to 66 kDa with a mass error of less than 100 ppm. The tuning of the instrument with large protein standard s has increased the confidence that unknown proteins may be identified. The analysis of in-house prepared Ins-ST, N-Tag-ST, and CTag-ST and commercially prepared rat (2 6)-(N)-sialyltransferase has not resulted in spectra. This was due to the difficulty of re moving sialyltransferase from glycerol/triton buffer systems. Unfortunately, ESI was not co mpatible with buffers that contain highsalt, glycerol, or detergents. Also, since th e protein has several glycoforms, this reduces the overall concentration of sial yltransferase to that of th e different glycoforms. To reduce heterogeneity, deglycosyla tion was performed with Endo Hf and PNGase F. ESIFTICR MS analysis of deglycosylated c onstructs also did no t provide signal. Furthermore, after analysis of the sialyl transferase constructs with HPLC-ESI-FTICRMS, the column lost separation efficiency due to the detergent being present. HPLC-ESI-FTICR MS of protein standards di gested with pepsin resulted in the observation of peptic peptid e. Digestion of carbonic anhydrase II resulted in 45% sequence coverage. Digestion of In s-ST and commercially prepared rat (2 6)-(N)sialyltransferase prior to HPLC -ESI-FTICR MS did not result in positive identification of peptides. Signal may not have been observed because the glycans ar e inhibiting digestion or sialyltransferase may need to be denatured prior to pepsin digestion. Since the activity of sialyltransferase can not be retained prior to MS analysis, H/Dx HPLC-MS was abandoned. Since structural prot eomic analysis of si alyltransferase is the overall goal of

PAGE 134

117 this project, attention is shifted to labeling with small molecule labeling agents (Chapter 4) and site-directed photoaffinity labeling agents (Chapter 5). Material and Methods Acetic acid (Ca # A507-500), formic acid (Ca # BP 1215-500), methanol (Ca # 230-4),acetonitrile (Ca # A998), and high-pur ity water (Ca # AH 365-4 or W 7-4) were purchased from Fisher-Scientific (Fair Lawn, NJ). D2O (99.9%) was purchased from Cambridge Isotope Laboratories, Inc. (Andove r, MA) Bovine pancreas ribonuclease A (Ca # R6513), HEWL (Ca # L6876), bovine ch ymotrypsinogen A (Ca # C4879), human eukaryotic carbonic anhydrase I (Ca # C4396), bovine carbonic anhydrase II (Ca # C3934), and bovine serum albumin (Ca # A9056 ) were purchased from Sigma-Aldrich (St Louis, MO). FT-ICR/MS Instrumentation Stable electrospray was achieved with Aglients (Waldbronn, Germany), off-axis ESI spray source equipped with nebulizi ng gas (Ca # G2427A). A Bruker Daltonics (Billerica, MA) APEX II 4.7 T FT-ICR MS was used to colle ct data in heterodyne and broadband modes with SideKickTM trapping, gas-assisted dynamic trapping (GADT), or dynamic Trapping (DT). Samples were dir ectly infused with a Harvard Apparatus (Holliston, MA) PHD 200 Infusion (Ca # 70 2000) at 0.5 to 2 L/min. Calibration was provided with Agilent Technologies (Palo Alto, CA) HP Tuning Mix (Ca #G2421A) with the four point (622.0290, 922.0098, 1521.9715, 1521.9715, 2121.9332) non-linear algorithm provided by the Bruker Xmass software. HPLC Instrumentation An Agilent Technologies (Palo Alto, CA ) HP1100 Series Binary Pump (Ca # G1312A) was used to deliver solvent at 50 L/min to Grace Vydac (Hesperia, CA)

PAGE 135

118 columns. A C4 RP, 300 (1.0 mm x 50 mm; Ca # 214T P5105) column was used for the analysis of intact-protein s at a flow rate of 50 L/min with the flow split post-column to 2 to 4 L/min prior to ESI-MS analysis. The gradient for the C4 column included a 4 minute hold at 5 % mobile phase B, followed by a 5 95 % mobile phase B ramp over 2 minutes. For peptide analysis with the C4 column, the same gradient was used. A C18 Polymer RP column 300 (0.150 mm x 50 mm ; Ca # 218MS5.1505) was also used for separation of digested protei ns at a flow rate of 2 L/min. The gradient for the C18 column included a 3 minute hold at 5 % mobile phase B, followed by a 5 95 % mobile phase B ramp over 15 minutes. Three separate buffer systems were used and each did not change the signal intensity observed. M obile phase A1 consisted of 5% acetonitrile, 94.5% H2O, and 0.5 % formic acid and mobile phase B1 consisted of 95 % acetonitrile, 4.5 % H2O, and 0.5 % formic acid. Mobile phase A2 consisted of 5 % methanol, 94.5% H2O, and 0.5 % acetic acid and mobile phase B2 consisted of 95 % methanol, 4.5 % H2O, and 0.5 % acetic acid. Mobile phase A3 consisted of 5% acetonitrile, 94.5% H2O, and 0.5 % acetic acid and mobile phase B3 c onsisted of 95 % acetonitrile, 9.5 % H2O, 0.5 % acetic acid. An Upchurch Scie ntific (Oak Harbor, WA) Micr o-injector Valve (M-435) or a VICI (Houston, TX) Cheminert (01S-0010H) inj ector valve were used to inject sample onto the column. The load loop was between 3 L and 10 L depending on the application. Intact Protein Analysis with Broadband Mode Direct infusion analysis of intact proteins For direct infusion (2 L/min) analysis, the concentrati ons of the standard proteins were kept between 5 and 50 pmol/ L. All proteins were d iluted to 50% methanol, 50%

PAGE 136

119 water, and 0.1 % acetic acid except human carbonic anhydrase I whic h was diluted to 20 % methanol, 80 % water, and 0.1 % acetic acid. After several iterati ons, it was found that the optimal conditions for bovine pancr eas ribonuclease A, HEWL, human carbonic anhydrase I and chymotrypsinogen A included an trapping voltage of 1.85 V, an ion cooling time of 0.7 s with 35 steps, a fina l detection voltage of 0.5 V, a hexapole accumulation time of 0.5 seconds, a load cell cy cle of 5 times, a p2 (t ime-of-flight effect) of 2000, and a 32 k acquisition file size. The optimal trapping voltage for bovine serum albumin was 2.85 V, an ion cooling time of 1.0 s with 50 steps, a fi nal detection voltage of 0.5 V, a hexapole accumulation time of 0.5 se conds, a load cell cycle of 5 times, and a p2 of 2200. The entire series of standard proteins could be analyzed with a trapping voltage of 2.65 V, an ion cooling time of 1.0 s with 25 steps, a final detection voltage of 0.5 V, a hexapole accumulation time of 1.0 seconds, a load cell cycle of 10 times, a p2 of 2200, and a and a 32 k acquisition file size. The direct infusion ESI-FTICR-MS analysis of recombinant sialyltransferase included para meters optimized for the different protein standards. HPLC-ESI-FT-ICR MS analysis of intact proteins Between 15 and 100 picomoles of intact protein was analyzed with the C4 column and chromatographic conditions describe above. After seve ral iterations, the optimal conditions for the analysis of all standard proteins were obtained. One chromatic frame acquisition was set to a trapping voltage of 2.65 V, an ion cooling time of 1.0 s with 25 steps, a final detection voltage of 0.5 V, a hexapole accumulation time of 1.0 seconds, and a load cell cycle of 10 times, a p2 of 2200, and a 32 k acquisition file size. This provided a scan rate of 12.72 s/scan. For reco mbinant sialyltransferase, 10 to 70 pmol

PAGE 137

120 was analyzed with the C4 column, chromatographic conditions, and instrumental conditions describe above. Pepsin Digestion and HPLC-ESI-FTICR-MS Analysis To initiate digestion of bovine carbonic anhydrase II (CA II), 7 L of CA II (70 M stock contrition; 4.6 M final concentration) was diluted with 93 L of 0.5 % formic acid and 5 L of pepsin (50 M stock solution in 0.5 % formic acid; 2.3 M final concentration) and incubated at 4oC for 30 to 60 minutes. Also, 65 L of CA II (70 M stock contrition; 30 M final concentration) was diluted with 75 L of 0.5 % formic acid and 10 L of pepsin (50 M stock solution in 0.5 % formic acid; 3.3 M final concentration) and incubated at 4oC for 30 to 60 minutes. To initiate digestion of bovine serum albumin (BSA), 57 L of BSA (83uM stock solution in 0.5 % formic acid; 30 M final concentration) was diluted with 83 L of 0.5 % formic acid and 10 L of pepsin (50 M stock solution in 0.5% formic acid; 3.3 M final concentration) and incubated at 4 oC for 30 to 60 minutes. To initiate digestion of HEWL 35 L of HEWL (120 M stock solution in 0.5 % formic acid; 28.5 M final concentration) was diluted with 102.5 L of 0.5 % formic acid and 10 L of pepsin (50 M stock solution in 0.5% formic acid; 3.4 M final concentration) and incubated at 4 oC for 30 to 60 minutes. To initiate digestion of Ins-ST or commercial rat recombinant -(2 6)-(N)-sialyltransferase (STcommercial), 50 L Ins-ST or STcommercial (5.5 to 6.0 M stock solution in buffers described in Table 3-6; 4.5 to 5.0 M final concentration) was diluted with 6.5 L of 5.0 % formic acid and 4 L of pepsin (50 M stock solution in 0.5% formic acid; 3.3 M final concentration) and incubated at 4 oC for 30 to 60 minutes. For HPLC-MS analysis, between 15 and 100

PAGE 138

121 picomoles of digested protei n was analyzed with the C4 or C18 column and the chromatographic conditions describe above. Th e instrument was tuned daily prior to the analysis of digested proteins. After several iterations, the optimal instrumental conditions for the analysis of digested proteins were obtained. One chromatic frame acquisition was defined with a hexapole accumulation time of 2.0 seconds, a p2 of 2000, the sum-average of 2 to 4 scans, and either a 128 k or 256 k acqui sition file size. This provided a scan rate of 5.7 to 11.4 s/scan because of the signal aver aging of 2 to 4 scans. While acquiring a 35 minute HPLC run, the 128 k acquisition file size produced a final file size of 24 M and the 256 k acquisition produced a final file size of 47 M. H/Dx Experiments To initiate H/Dx, 28.6 L of HWEL (700 M stock solution in water; 20 M final concentration) was diluted with 971.4 L of 99.9% D2O. At 1, 2, 5, 10, 15, and 20 minutes, 100 L aliquots were taken and quenched at 4 oC with 100 L of a buffer containing 20% methanol, 78 % water, 2% acetic acid. Direct infusi on analysis included the use of GADT in broadband mode with a hexapole accumulation time of 0.01 seconds, a load cell cycle of 5 times, a p2 of 2200, a 1 M acquisition file size, and signal averaging over 4 scans. For HPLC-ESI-FTICR-MS analysis of exchanged and non-exchanged HEWL, the C4 column and injector were submerged in an ice bath. The instrument was tuned in heterodyne mode (1326.52 to 1552.17 m/z window) with the HEWL +10 charge state prior to HPLC analysis. Optimal inst rument conditions incl uded a trapping voltage of 1.5 V, an ion cooling time of 1.0 s with 25 steps, a final detecti on voltage of 0.5 V, a hexapole accumulation time of 1.0 seconds, a load cell cycle of 5 times, a p2 (time-offlight effect) of 2200, and a 32 k acquisition file size.

PAGE 139

122 CHAPTER 4 MS BASED STRUCTURAL PROTEOMICS WITH SMALL MO LECULE LABELING AGENTS FOR IDENTIFYING SIALYLTR ANSFERASE ACTIVE SITE AMINO ACIDS Introduction to Bioconjugation Techniques The identification of the catalytically im portant amino acid residues for a protein can aid in rational drug design. Furtherm ore, an understanding of the allosteric movements of proteins, alrea dy described in Chapter 3, can aid in the study of proteinprotein interactions, dynamic movement of regions of the protein upon binding to ligand, and protein binding motifs. The second bi oconjugation technique described in this dissertation involves the covalent modificat ion of amino acid residues with labeling agents.142 The chemical derivatization of different amino acids has been used for several decades to allow identificati on of catalytic amino acids with many case specific studies on protein-protein interacti ons, protein-DNA interactions, protein-ligand interactions, protein topology, and membrane topology.143 With respect to this dissertation, chemi cal modification of proteins is a proven method to implicate the amino acid residue(s) needed for the function of protein-ligand interactions.144 Furthermore, there are two differe nt modification schemes: 1) small molecule labeling agents for derivatization of specific amino acids and 2) site-directed photoaffinity labeling. Chapter 5 presents th e study of site-directed photoaffinity labels applied to sialyltransferase. Prior to performing a labeling experiment on sialyltransferase, target amino acids were selected for bioconjugation, labeling agents were selected for the specific amino acids, a nd a protocol was devel oped. All three of

PAGE 140

123 these aspects are reviewed in this chapter, along w ith classical biochemistry experiments and MS based analysis. Practical Consideration and Types of Small Molecule Labeling Agents Small molecule labeling agents (SMLAs) ar e chosen based on the application being performed. Often, the use of SMLAs is case sp ecific for a particular protein. The amino acid (AA) side chain selected for deriva tization may be important for maintaining solubility, binding of substr ates, or catalyzing a reaction. Also, the AAs may be involved with H-bonding to maintain struct ure or hydrophobic interactions to protect regions from aqueous solvents; however, AAs of this type are not normally selected for small molecule labeling. Typically, SMLAs target the nucleophilic amino acids that contain primary amines, secondary amines, carboxylic acids, alcohols, or sulfhydryls.145 To efficiently use SMLAs with MS analysis the label must selectively bind to the target amino acid with a stable covalent bond. When a particular protein is folded into its particular three dimensional structur e, many residue may have different pKas based on the microenvironment, i.e. the surrounding AAs in the active-site. Furthermore, changes in pH, temperature, and detergent concentr ation may also change the reactivity of charged AAs towards the SMLAs. Generally, there are 4 requirements for id eal SMLA-proteomic MS experiments. First, the reaction must be fast and re producibility, with yields approaching 100%. Second, the covalent bond formed must be stable through sample cleanup, separation, and MS analysis. Third, the reagent must r eact at room temperature, physiological pH, and be water soluble. Fourth, the tertiary st ructure of the protein mu st be retained during the labeling experiment.146;147 Furthermore, a differential labeling scheme is used to

PAGE 141

124 identify amino acids which are protected when substrate is bound in the active-site and not protected when no substrate is present in the active site. Practical Considerations for MS Based Structural Proteomics with SMLAs Classically, bioconjugation of pr oteins used radioisotope (3H, 13C, and 32P) SMLAs. Currently, investigators still use this method for quantifying the number of labeled AAs on a certain protein; however, e xperiments are designed to avoid the use of radioactive material because of licensing, safety, and dispos al. Biochemical data from radioisotope and non-radioisotope SMLAs ha ve provided valuable information on the identity of specific AAs needed for binding, catalysis, and hinge movement for many protein systems. Furthermore, SMLAs are commercially available for bioconjugation experiments. The identification of deriva tized AAs may be performed with different biochemical experiments based on the SMLA us ed, with Edman degradation analysis, or with proteomic MS analysis. Of course, each method has their drawbacks; however, proteomic MS has the best capacity for high-throughput after reproducible bioconjugation protocols have been worked out. As presented in Chapter 3, bioconjugati on of proteins followed by polypeptide sequence analysis with Edman degradation or proteomic MS are considered a low resolution technique in compar ison to NMR spectroscopy or X -ray crystallography. Still, bioconjugation of proteins followed by proteo mic MS analysis is the best way to investigate secondary structure of low yield protein systems.148;149 Often, bioconjugation with proteomic MS analysis compleme nts already known crystal structures.148 Since crystal structures are static photographs of the protein, bioconjugation with SMLAs may provide a new view of a proteins dynami c movement upon binding of a substrate. The description of a proteins dynamic moveme nt provides a basis for the rational design

PAGE 142

125 of transition-state mimic inhibitors. Furtherm ore, if the crystal structure is un-solved, derivatized AAs may be compared to known pr otein structural motifs. Bioinformatics may then be used to generate a picture of the active-site based on the identity of AAs derived from differential labeling. Proteomic MS has provided the technology th at measures the mass shift associated with derivatization of amino acids. Several different small molecule labeling protocols have been developed or re-visited because of the improvements to MS instrumentation and sample work-up.150 In the complete workflow of a SMLA-proteomic MS experiment, sample workup is still the major bottleneck. Biotechnology has recently driven mass spectrometrist to develop instrumentation for the sole use of identifying posttranslational modifications on proteins. The same technology is applied to proteins which have been derivatized with SMLAs. For example, the Q-TOF and the more expensive LTQ-FT may be set to analyze a certain mass list of peptides which contain certain AA selectively derivatized with a known SMLA. There are many reports in the literature where a certain protein or cell-sta te has been investigated with bioconjugation up-stream from MS analysis (Table 4-1) Data may be collected for functional proteomics, i.e. disease state versus health st ate, or structural prot eomics in regards to protein-protein interactions, membrane t opology, receptor binding sites, and enzymeligand interactions.145;151 Literature Review on Ideal Cases One of the best strategies for SMLA-pro teomic MS is the use of isotope-coded affinity tags (ICAT) for the comparison between healthy and disease states. This method has applications in functional proteomics a nd is not related to the overall goal of the experiments presented in this dissertation. However, ICAT technology is the spawn of

PAGE 143

126 well designed bioconjugation techniques a nd was made commercially available by Applied Biosystems (Foster City, CA) in 2000. Figure 4-1 represents the ICAT reagent used for the workflow described in Figur e 4-2. The cysteines of polypeptides are targeted with ICAT type labels with iodo acetyl or vinyl functionalities. Cystines normally represent 1.1 % of the total AAs foun d in proteins. If there are no cysteines present in a target protein, lysine may be modified with amidination/guanidination or tryptophan may be modified with indole. Si nce lysines are much more prevalent, more peptides will be labeled, t hus providing more information; however, the data is more complex to analyze.142 Upon analysis with MS, there is an 8 Da mass shift between the differently labeled cell states (Figure 4-2). Differential comparison between the signal intensities of the light and h eavy labeled peptides (same pept ide but different cell states), provides relative quantitation on the level of e xpression of a particular protein in the two different cell states. Heavy Reagent: D8-ICAT Reagent (X=Deuterium) Light Reagent: D8-ICAT Reagent (X=Hydrogen) NH HN S O H NO ON H I O X X X X X X X X O Biotin Tag Linker Chain (Heavy or Light) Reactive Group Heavy Reagent: D8-ICAT Reagent (X=Deuterium) Light Reagent: D8-ICAT Reagent (X=Hydrogen) NH HN S O H NO ON H I O X X X X X X X X O Biotin Tag Linker Chain (Heavy or Light) Reactive Group Figure 4-1. ICAT Reagent Table 4-1 represents the selected refe rences that use SMLAs for structural proteomic studies. The examples in the follo wing table were chosen based on the time frame (published in 2004), the target amino acid for derivatization (tyrosine, lysine,

PAGE 144

127 Sample 1 Healthy Cell State Sample Label with Light ICAT Reagent Label with Heavy ICAT Reagent Mix and Digest Samples Avidin-Biotin Affinity Separation Relative Quantitation between Light and Heavy Isotopes with HPLC-MS Analysis Sample 1 Healthy Cell State Sample 2 Disease Cell State Label with Light ICAT Reagent Label with Heavy ICAT Reagent Mix and Digest Samples Avidin-Biotin Affinity Separation Relative Quantitation between Light and Heavy Isotopes with HPLC-MS Analysis Figure 4-2. ICAT Work-flow histidine, cysteine, or aspartic/glutamic acid) and/or the reaction conditions. The basis for the selecting the SMLAs for derivatization of sialyltransferase were extracted from the papers presented below. Table 4-1. Literature Review of Different Bioconjugation Experiments Protein System and Reference Labeling Agent (Target Amino Acid) Important Information 1) Protein-protein interactions of HIV binding viral envelope protein (gp120) to human receptor CD 4 on T-cells152 hydroxyphenolglyoxal (lysine) With a differential labeling strategy and MS analysis, the specific lysine involved with binding was identified. This paper also compared the new data to the crystal structure to confirm contact point. 2) Protein-protein interactions of IgG cross-linked to horseradish peroxidase153 In-house synthetic cross-linker: sulfosuccinimidyl(perfl uorobenzoamido)ethyl-1,3dithiopropionate (primary amines) Study to illustrate the increased cross-linking of one of the most common ELISA assay components. The fluorinated azide had higher insertion efficiency. This study is more relevant to Chapter 5 of this dissertation.

PAGE 145

128 Table 4-1. Continued. Protein System and Reference Labeling Agent (Target Amino Acid) Important Information 3) Protein-ligand interactions of IgG to test new labeling agent154 Acyl adenylate (Deoxycholyl adenylate) (lysine or arginine) Study to illustrate the direction of SMLAs for protein-protein and protein-ligand interactions for high-throughput in the postgenomic area. Specifically label lysines which may be in receptor binding sites. In-gel digestion with trypsin and MS analysis was used to identify specific site of binding. 4) Membrane topology (solvent accessibility) of glycine receptor 150 Tetranitromethane (tyrosine) This study revealed which tyrosines played an important role in partitioning properties due to hydrophobic ends of this type of receptor based on the solvent accessibility. In-gel digestion with trypsin and MS analysis was used to identify mass shifts. 5) Protein-nucleic acid interactions of human replication protein A (hRPA) and DNA155 No labeling agent: Irradiated protein-DNA complex for zerocrosslink with phenylalanine. Three different binding regions were identified with MS analysis. This solidified data on the binding region when compared to NMR, mutagenesis, and X-ray crystallography data. This study is more relevant to Chapter 5 of this dissertation. 6) Protein-ligand binding interaction of calmodulin to melitin149 Several different crosslinkers. 1-ethyl-(3dimethylaminopropopy l) carbodiimide hydrochloride, SulfoNHS, Sulfo-DST, and bis(sulfosuccinimidyl) suberate. (aspartic/glutamic acid, and lysine, asparagine) Study using three different crosslinkers of different length to map out the binding sites. In-gel digestion with trypsin with FTICRMS. From exact mass, took data one step further and generated low resolution crystal structures based on the difference between the cross-linker lengths and binding sites. 7) Structural characterization of fish egg vitelline envelope proteins156 No labeling agents Good paper which represents the use of different types of MS to identify proteins based on mass and also mapped proteins with different proteases and MS/MS analysis. Presented interpretation of non-standard MS/MS data.

PAGE 146

129 Table 4-1. Continued. Protein System and Reference Labeling Agent (Target Amino Acid) Important Information 8) Protein-ligand interactions of aminopeptidase A157 N-acetylimidazole (Tyrosine, serine, cystine, lysine, and histidine) This study is similar to sialyltransferase in that it contains a consensus sequence motif of HEXXH. Through a differential labeling scheme, Nacetylimidazole labeled a tyrosine which is necessary for stabilization of the transition state. Strictly a biochemical study. 10) Protein ligand interaction of leader peptidase158 Several SMLAs investigated N-Bromosuccinimide, 1-ethyl-(3dimethylaminopropopy l) carbodiimide hydrochloride, Nacetylimidazole, iodoacetic acid, 5,5dithiobis(2nitrobenzoic acid, succinic acid, and phenylglyoxal (all nucleophilic amino acids) Purely a biochemical study on the solvent accessible amino acids. NBromosuccinimide was found to react with tryptophan, 1-ethyl-(3dimethylaminopropopyl) carbodiimide hydrochloride was found to react with an aspartic or glutamic acid, and phenylglyoxal was found to react with arginine to deactivate the enzyme. All of the other SMLAs did not deactivate the enzyme. Used HPLC-UV and radioisotopically labeled SMLAs to identify derivatized peptides. 11) Protein-ligand of transglutaminase factor XIII (FXIII) in active and non-active state159 alkylmalemidides and acetic anhydride (cystines and lysines) Differential labeling of solvent accessible AAs followed by insolution digestion with trypsin and MS analysis. Obtained data on non-activated state and activated state which could not be determined from static X-ray crystal structure upon binding to different substrates and divalent cations. 12) Membrane binding topology of cytochrome P450144 acetic anhydride, glycinamide and 1ethyl-(3dimethylaminopropopy l) carbodiimide (lysine, serine, arginine, cysteine, glutamic acid and aspartic acid. Used differential labeling scheme for comparison of soluble and membrane bound states. The study found several lysines that were contact points for binding to the membrane. Used in-solution digestion with trypsin and MALDI-TOF MS for identification of conjugated AAs.

PAGE 147

130 Table 4-1. Continued. Protein System and Reference Labeling Agent (Target Amino Acid) Important Information 13) Solvent accessible amino acids on Neuroliginstransmembrane proteins160 iodoacetic acid and 4-vinyl-pyridine (cystine) This study mapped disulfide bonds and N/Oglycans sites. These proteins have similar topology to sialyltransferase. 14) Labeling of neuropeptides143 N-acetylimidazole (Tyrosine, serine, cystine, lysine, and histidine) Continuous-fast atom bombardment MS revealed a mass shift of 42 Da of a derivatized peptide. The modification was reversed by treatment with hydroxylamine. 15) Protein-ligand interactions of -glutamyl transpeptidase161 Iodoacetamide and N-acetylimidazole (tyrosine, serine, cystine, lysine, and histidine) N-acetylimidazole was found to react with a lysine which causes a cystine to be exposed for labeling with iodoacetamide. This is an example where the labeling agent has changed the structure of the enzyme. 16) Protein-ligand interaction of cinnamomin type II ribosome-inactivating protein and ribosomal RNA162 N-acetylimidazole (Tyrosine, serine, cystine, lysine, and histidine) This study revealed, through differential labeling, the identification of a tyrosine needed for binding to DNA. The reaction was terminated by the addition of tyrosine. Activity was regenerated with ethanolamine. 17) Protein-DNA interaction of capase-3 activated DNase CAD with DNA cleavage 163 2,4,6-trinitrobenzenesulphonic acid, and N-acetylimidazole (Tyrosine, serine, cystine, lysine, and histidine) This study provided identification of a lysines and tyrosine needed for binding and catalysis. Mutagenesis studies on different lysines and tyrosines allowed identification of individual amino acids important for catalysis. Limitations for MS Based Structural Proteomics with SMLAs The major limitations with SMLA-proteom ic MS includes peptide ionization and MS/MS fragmentation. The derivitation of basic amino acids may greatly reduce the ionization efficiency. Also, fragmentation patterns of non-derivatiz ed and derivatized proteolytic peptides do not always yield spectra exemplifying polypeptide backbone

PAGE 148

131 ladders. The intensities of fragmented pr oteolytic peptide ions are mediated by the charge state of the parent ion, the AAs presen t, and the size of the proteolytic digested.142 Interestingly, some investigators have chosen to synthesize SMLAs that enhance ionization by the addition of a basic tertiary amine.147 For example, cationic SMLAs were reported to increase signal intensities by a factor of forty for hydrophilic peptides; however, a decrease in MS/MS fragment ation coverage was observed. Secondly, coumarin coupled to the N-termini of proteolytic polypeptides with Nhydroxysuccinimide chemistry enhanced MADLI-TOF signal and improved MS/MS fragmentation.146 Enhancement of peptide ionization ma y be beneficial if the investigator is looking for a certain important polypept ide for functional proteomics. The main problems with these types of SMLAs are their solubility and difficult synthetic protocols.147 Practical Consideration for Labeling of Sialyltransferase with SMLAs Based on the primary structure motif data presented in Chapter 2, a general acid and a general base are the ta rgets for SMLAs. The targeted amino acids for a general acid and a general base include histidine, gl utamic acid, aspartic acid, lysine, serine, and/or tyrosine. After revi ewing the literature, the deci sion was made to label with iodoacetic acid (IOA), 1-ethyl-(3-dimet hylaminopropopal) carbodiimide hydrochloride (EDC) coupled with ethanolamine (ETAM), a nd N-acetylimidazole (NAI). The choice of these three SMLAs was also based on cost, commercial availability, reaction conditions, water solubility, and reaction quenc h conditions. Furthermore, experiments with radiolabel SMLAs were not performed because the main requirement for the labeling reaction was the deactivat ion of enzymatic activity.

PAGE 149

132 IOA normally reacts with cystine; howeve r, if the microenvironment shifts the pKa of the AAs found in the catalytic active-site, hi stidine or lysine may also be modified. EDC is normally associated with activat ion of carboxyl groups leading to amide formation by reaction with primary amines. EDC has been used as a cross-linker for several applications, and in this case, may create cross-linking betw een sialyltransferase molecules found in solution. To avoid prot ein-protein cross linki ng, ETAM is added in large excess to EDC. NAI may r eact with lysine, cystine, his tidine, serine, and tyrosine. Of particular interest, NAI is normally used to successfully identify tyrosine as the labeled AA through deacetylation with ETAM. The crystal structure of the bacterial sialyltransferase presented in Chapter 2, al ong with the structural motif information, make tyrosine a special target for this study. I OO+OO HN R N N OO R O O S R O O S+ R CH3Cys His Lys Met I OO+OO HN R N N OO R O O S R O O S+ R CH3Cys His Lys Met Figure 4-3. IOA Specificity Bioconjugation of Sialyltransferase with Small Affinity Labels The three different constructs presented in Chapter 2 were incubated with NAI, EDAC/ETAM, and IOA. First, Ins-ST was in cubated with the three labeling agents to determine which would deactivate the enzyme. Second, the identifica tion of active-site AAs with NAI acetylation was performed usi ng a differential protection scheme and the C-Tag-ST construct. Third, the reactivation of N-Tag-ST with ETA M after labeling with NAI was performed. Finally, all the derivatiz ed recombinant hST3Gal IV constructs

PAGE 150

133 H3CN C NNH+ CH3CH3Cl-+ROH O H3C N H N H+ N CH3CH3O R O+OH NH3+ H3C N H N H H+ N CH3CH3O+RN H O OH Figure 4-4. EDC Coupling with ETAM N N CH3O HO R+H N N O R+O Figure 4-5. Acetylation of Tyrosine with NAI were subsequently analyzed w ith SDS-PAGE, in-gel digested with trypsin, and analyzed with either MALDI-TOF MS or LC-Q MS. Differential Labeling Scheme Figure 4-6 illustrates the differential la beling scheme used. For simplicity and illustration, only CMP-NeuAc is shown in the ac tive-site of sialyltransferase and ETAM coupled to Glu or Asp with EDAC is shown as the labeling agent. Differential labeling is

PAGE 151

134 performed by first labeling sialyltransferas e without substrate present. All solvent accessible amino acids specific to the chemistry of the labeling agent will react. Second, sialyltransferase is labeled in the presence of donor substrate (CMP -NeuAc), of acceptor substrate ( -lactose), or with both present. The pr esence of substrate will block solvent accessible AAs in the active site; therefore, the SMLA will not be able to derivatize the catalytically important AAs. MS analysis of stable covalent bonds formed between the SMLA and the AA may aid in identifying the exact location of the general acid or the general base. Un-Protected Active SiteActive Site Peptide chain ON H NH3+ ON H NH3+ O N H +H3N HN O NH3+ O N H NH3+ O N H NH3+Active Site Peptide chain OOProtected Active Site O OO -O O N N O OH HO O P O O NH2OO CO2HO H N OH OH HO O HN O NH3+ O N H NH3+ ON H NH3+Un-Protected Active SiteActive Site Peptide chain ON H NH3+ ON H NH3+ O N H +H3N HN O NH3+ O N H NH3+ O N H NH3+Active Site Peptide chain OOProtected Active Site O OO -O O N N O OH HO O P O O NH2OO CO2HO H N OH OH HO O HN O NH3+ O N H NH3+ ON H NH3+ Figure 4-6. Differentia l Labeling Experiments Workflow The work flow for the labeling of sialyl transferase is described in Figure 4-7. There are three separate experiments that must be performed. The first data set includes

PAGE 152

135 the MS analysis of in-gel digested recombinant hST3Gal IV which has not been labeled with SMLAs. The main goal of this experi ment is to obtain 100 % coverage. This will ensure that an AA which has been derivati zed will be observed. The second data set includes MS analysis of in-gel digested un-protected, derivatized recombinant hST3Gal IV. This data set will identify AAs which are derivatized because they are solvent accessible. The third data set entails the MS analysis of in-gel digested active-site protected, derivatized recombin ant hST3STGal IV. When comparing this data set to the second data set, the identity of the AA which is no longer solvent accessible may be identified. The step including trichloroacetic acid (TCA) precipitation may be excluded if the final concentration of recombinant hST3Gal IV after derivatization and activity profiling is above the limitof-detection for SDS-PAGE a nd MS analysis. The step including Endo Hf or PNGase F glycan digestion is included to reduce the complexity of the peptides being analyzed because of the heterogeneity associated with the multiple glycoforms of expressed sialyl transferase. In-gel digesti on is most commonly performed with trypsin; however, to increase sequence c overage, other proteases such as Lys C, Glu C, Asp N, or -chymotrypsinogen A may be used indi vidually or in a double-digestion manner. MALDI-TOF analysis is used to determine the percent sequence coverage and to determine what peptides were derivatized. LC MS/MS is used to confirm MALDI-MS data on percent sequence coverage and to va lidate MALDI-TOF data on peptides which AAs have been derivatized. Several different software packages were used to analyze the peptide maps generated with MALDI-T OF and peptide fragmentation patterns generated with MS/MS analysis.

PAGE 153

136 Ins -His6x-ST Ins -ST -His6xIns -ST Covalent Modification: Un -protected Protected Mass Spectrometry LC MS/MS MALDI TOF MS Bioinformatics Data Analysis TCA precipitation Endo Hf or PNGaseF Deglycosylation Rxn SDS-Page Analysis In-Gel Digestion 1 3 2 Ins -His6x-ST Ins -ST -His6xIns -ST Ins -His6x-ST Ins -ST -His6xIns -ST Ins -His6x-ST Ins -ST -His6xIns -ST Covalent Modification: Un -protected Protected Mass Spectrometry LC MS/MS MALDI TOF MS Bioinformatics Data Analysis TCA precipitation Endo Hf or PNGaseF Deglycosylation Rxn SDS-Page Analysis In-Gel Digestion 1 3 2 Figure 4-7. Work Flow for Bioconjuga tion of Recombinant hST3Gal IV and Identification of Derivatized Amino Acids Results and Discussion Preliminary screening of NAI, EDAC/ETAM, and IOA The initial experiments to determine whic h SMLA deactivated Ins-ST involved the use of EDAC with ETAM, IOA, and NAI. A ll activity assays were performed in the same fashion as described in Chapter 2: Me thods and Materials: Assays. The Ins-ST concentration was held at 6.5 M and incubated in the presence of the particular SMLA at 37 oC for 20 minutes in a final reaction volume of 30 L. Table 4-2 presents the enzymatic activity for a 20 minute incubation. Prior to assaying, NAI was quenched with 17 mM tyrosine, IOA was quenched with 13 mM histidine, and EDAC/ETAM was quenched with 17 mM acetic acid. The pH during all steps was maintained at 6.8. NAI was the only SMLA to deactivate INS-ST with the reaction conditions described. After all of the reactions were performed, EndoHf was added to the que nched labeling reaction

PAGE 154

137 mixture. The results for SDS-PAGE analysis, in-gel digestion (IGD), and MS analysis of labeled INS-ST are presented later in this chapter. Table 4-2. SMLAs and Activit y after 20 Minute Incubation Sample Activity in CPM INS-ST Standard 1401 NAI [17 mM]f 0 IOA [13 mM]f 1306 EDAC [17 mM]f + Ethanolamine [17 mM]f 1420 NAI differential labeling experiments Several experiments were performed to optimize conditions (concentration of enzyme, reaction temperature, concentration of NAI, and concentration of substrate(s)) for the best differential results when CTag-ST was incubated with NAI, with and without CMP-NeuAc, -lactose, or both present. Fi gure 4-8 represents the steps involved for an experiment with substrate pr esent. First, it was found that the final concentration of C-Tag-ST had to be at least 5 M and the final reaction volume had to be at least 80 L for SDS-PAGE, In-gel digested (IGD) and MS analysis after incubation and activity profiling. Second, for protection of the active-site, the concentration of the substrate(s) had to be 5 times the Km for C-Tag-ST. Third, about 2500 times excess of labeling agent to C-Tag-ST was required for complete deactivation. This value was within a very narrow window. For example, if there was about 1800 times excess NAI, deactivation would not go to 100 %. Fu rthermore, about 3200 times excess NAI completely deactivated the protected sialyltran sferase. Fourth, accord ing to the literature, tyrosine may be used as a quenching r eagent to scavenge the un-reacted NAI.162 Tyrosine as a quenching agent had no affect on enzyme activity. Fifth, activity was determined in the same manner as described in Chapter 2, methods and materials, assays.

PAGE 155

138 Stock Solution Sialyltransferase 1 Added Substrate (5 x Km) 2 Add Labeling Agent (2500 x excess) 3 Activity At Time Points 5 Quench with free tyrosine 4 Stock Solution Sialyltransferase 1 Added Substrate (5 x Km) 2 Add Labeling Agent (2500 x excess) 3 Activity At Time Points 5 Quench with free tyrosine 4 Figure 4-8. Labeling Experiment The following graphs represent the data collected for positive results with differential labeling. Due to limited materi als and the stability of the recombinant sialyltransferase constructs, onl y the C-Tag-ST construct was fully investigated. Figure 4-9 presents the results fo r NAI derivatization of C-Tag-ST with CMP-NeuAc [525 M]final present and no substrate present. Fi gure 4-10 presents the results for NAI derivatization of C-Tag-ST with CMP-NeuAc [1.5 mM]final and saturated -lactose present, saturated -lactose present, and no substrate pr esent. The results indicated that both donor and acceptor substrates were need ed for differential protection. Final activities were corrected for maximal velocity based on the equation presented in Chapter 2, method and material, assays. 0 5000 10000 15000 20000 25000 30000 01234567891011 Reaction Time in MinutesActivity in fmol / min product --Active Site Protection with CMP-NeuAc --No Substrate Present 0 5000 10000 15000 20000 25000 30000 01234567891011 Reaction Time in MinutesActivity in fmol / min product --Active Site Protection with CMP-NeuAc --No Substrate Present Figure 4-9. NAI Derivatizati on of C-Tag-ST without Subs trate Present and with 525 M CMP-NeuAc Present

PAGE 156

139 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 024681012 Reaction Time in MinutesActivity in fmol / min product --Active site protected with 1.05 mM CMP -NeuAc / saturated -lactose --Active Site Protection with saturate -lactose --No Substrate Present 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 024681012 Reaction Time in MinutesActivity in fmol / min product --Active site protected with 1.05 mM CMP -NeuAc / saturated -lactose --Active Site Protection with saturate -lactose --No Substrate Present Figure 4-10. NAI Derivatizati on of C-Tag-ST without Subs trate Present, with 1.5 mM CMP-NeuAc and Saturated -Lactose Present, and with Saturated -Lactose Present. The reaction volumes, the concentration of labeling agent, the concentration of enzyme, and the temperature were held constant once the optimal conditions were determined. Based on the curves obtained, there was a change in conformation of the enzyme upon binding to both donor and acceptor substrates. This may include ordering of a floppy loop region to produce a protected active-site as described in the crystal structure of the bacterial sial yltransferase. The curve pr esented in Figure 4-9 suggests that CMP-NeuAc alone does not protect the enzyme from deactivation. Other experiments suggested that higher con centrations of CMP-NeuAc prolong the deactivation of C-Tag-ST. Furthermore, saturating -lactose will also protect the binding site from being labeled; however, act ivity is still lost over time. NAI acetylates protein residue s at rates proportional to their nucleophilicity and solvent accessibility. With NAI as a SMLA, the identity of the active-site amino acids may be a tyrosine, serine, histidine, cystei ne, or lysine. Before performing SDS-PAGE,

PAGE 157

140 in-gel digestion, and MS analysis, the beha vior of the labeled sialyltransferase is investigated with the addition of et hanolamine after deactivation with NAI Reactivation of NAI derivatized N-Tag-ST with ethanolamine To test the stability of the acetylated amino acid, N-tag-ST was treated with ethanolamine after derivatization with NAI. Fo ur separate activity pr ofiles were obtained as illustrated in Figure 4-11. As presente d in Chapter 2, the recombinant hST3Gal IV was not amendable to ultrafiltration or dialysis for removal of salts. Since this was the case, NAI was still present in the soluti on during this experiment. However, NAI hydrolyzes with a half-life of 30 mins in 10 mM Tris-HCl (pH 7.5) at 30o C.157 It was assumed that the half-life of NAI in 50 mM MES pH 6.8 will be similar. Under the conditions described, NAI may acetylate tyro sine, cystine, and lysine, but will not acetylate serine because of the low reactivity at neutral and mildly alkaline pH. Deacetylation of lysine normally requires alkaline conditions, leaving tyrosine and histidine as the main targets for labeled am ino acids. Figure 4-11 represents the the reactivation of N-Tag-ST versus different stan dards. The experiment was performed at 4o C over 5 days with activitys obtained daily. At the end of the 5 days, the pH was determined to be ~6.8 for all sample s. The top curve (x) represents 5 M of N-Tag-ST without the addition of NAI or ETAM. The ( ) curve represents 5 M of N-Tag-ST with addition of ETAM to a final concentration of 1 mM. The ( ) curve represents the addition of NAI [55mM]final to 5 M of N-Tag-ST without the addition of ETAM. The ( ) curve represents the N-Tag-ST/NAI solu tion with dilution of ETAM to the final concentration of 1 mM. Per cent activity represents the relative activity normalized to NTag-ST without addition of NAI of ETAM. The data reveals that ETAM does not affect

PAGE 158

141 the activity of N-Tag-ST and that the presence of ETAM increased the rate in which activity was recovered. Based on the reaction conditions descri bed, it is proposed that the identity of the derivatized AA is either a histidine or a tyrosine. 0 20 40 60 80 100 120 140 160 020406080100120 x --5 M C-Tag-ST (Standard) --5 M C-Tag-ST + 55 mM NAI --5 M C-Tag-ST + 1 mM ETAM --5 M C-Tag-ST + 55 mM NAI + 1 mM NAITime in HoursPercent Activity 0 20 40 60 80 100 120 140 160 020406080100120 x --5 M C-Tag-ST (Standard) --5 M C-Tag-ST + 55 mM NAI --5 M C-Tag-ST + 1 mM ETAM --5 M C-Tag-ST + 55 mM NAI + 1 mM NAITime in HoursPercent Activity Figure 4-11. Reactivation of N-Tag-ST with ETAM SDS-PAGE analysis and in-gel digest ion with trypsin or other proteases SDS-PAGE analysis separated the hST3Gal IV constructs from impurities, SMLAs, and endoglycosidases. As described in Chapter 1, if there was more than one protein present in the solution prior to trypsin digestion, ther e would be more peptides for MS analysis. If the digested sample was more complex, proteolytic peptides may coelute, thus increasing the probability that an import ant recombinant hST3Gal IV proteolytic peptide would not be observed. Fi gure 4-12 presents th e SDS-PAGE analysis TCA precipitated N-Tag-ST and C-Tag-ST. The boxed area represents the portion of the gel which was excised for in-gel digested (I GD) with trypsin prior to MS analysis. Figure 4-13 represents the SD S-PAGE analysis of TCA precipitated N-Tag-ST and CTag-ST following digestion with EndoHf. The EndoHf band was observed at 66 kDa.

PAGE 159

142 (1) (2) (3) (4)12 345666 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa 66 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa (1) (2) (3) (4)12 3456 (1) (2) (3) (4)12 345666 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa 66 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa Figure 4-12. SDS-PAGE Analysis of C-tagST and N-tag-ST where: Lane 1 = TCA Precipitated C-tag-ST, IGD with Tryps in (1); Lane 2 = Molecular Weight Markers; Lane 3 = 20 uL of Purified N-tag-ST; Lane 4 = TCA Precipitated N-Tag-ST, IGD with Trypsin (3); La ne 5 = TCA Precipitated C-Tag-ST, IGD with Trypsin (4); Lane 6 = Molecular Weight Markers In noting the purity of the sialyltransferas e constructs, the lanes that do not contain EndoHf have bands that correspond to GP 64 as in Chapter 2. The purpose of these experiments was to determine the percent se quence coverage after IGD and MS analysis. Figure 4-14 represents the SDS-PAGE anal ysis of N-Tag-ST and C-Tag-ST TCA precipitated prior to re-suspension in 1x endoglycosidas e reaction buffer and addition either EndoHf or PNGaseF. Unfortunately, for this batch of samples, the deglycosylation reaction did not work. The boxed area repr esents the portion of the gel which was excised for IGD with a variety of differen t proteases for the purpose of enhanced sequence coverage with MS analysis. Since the proteases have di fferent AA specifities for cleavage, different peptide coverage ma ps may be created which may allow for 100% sequence coverage.

PAGE 160

143 (1) (2) (3) (4) (5) (6) (7) (8)12 3456789 10 11 (1) (2) (3) (4) (5) (6) (7) (8)12 3456789 10 11 Figure 4-13. SDS-Page Analysis of C-Tag-ST and N-Tag-ST After Deglycosylation with EndoHf where: Lane 1 = 20 L of Purified N-Tag-ST; Lane 2 = 20 L of Purified C-Tag-ST; Lane 3 = Molecu lar Weight Markers; Lane 4 = TCA Precipitated C-Tag-ST, IGD with Trypsin (1); Lane 5 = TCA Precipitated CTag-ST, IGD with -Chymotrypsinogen A (2); La ne 6 = TCA Precipitated EndoHf Digested C-Tag-ST, IGD with Tryps in (3); Lane 7 = TCA Precipitated EndoHf Digested C-Tag-ST, IGD with -Chymotrypsinogen A (4); Lane 8 = TCA Precipitated N-Tag-ST, IGD w ith Trypsin (5); Lane 9 = TCA Precipitated N-Tag-ST, IGD with Chym otrypsinogen (6) ; Lane 10 = TCA Precipitated N-Tag-ST, IGD with Trypsin (7); Lane 11 = TCA Precipitated EndoHf Digested N-Tag-ST, IGD with -Chymotrypsinogen A (8). Figure 4-15 represents the SDS-PAGE anal ysis of Ins-ST incubated with EDC, IOA, and NAI (Table 4-2). Obviously, the purification preceded this batch of Ins-ST sample was not adequate. The difficulty asso ciated with the purification of the Ins-ST construct was one of the main reasons for th e incorporation of the polyhistidine tag into the recombinant hST3Gal IV construct. Fu rthermore, the electrophoretic mobility was hampered due to the increased concentrati on of labeling and quenching reagents To improve the resolution of SDS-PAGE, TCA preci pitated sample provided the best results (Figures 4-12 and 4 -13). Figure 4-16 represents the DS -PAGE analysis of differentially labeled C-Tag-ST with NAI (Figures 4-9 and 4-10). The boxed ar ea represents the portion of the gel which was excised for IGD with trypsin prior to MS analysis. The electrophoretic mobility was

PAGE 161

144 12 3456789188 kDa 98 kDa 62 kDa 49 kDa 38 kDa 28 kDa 17 kDa 16 kDa 7 kDa 4 kDa (1) (2) (3) (4) (5) (6) (7) (8) 12 3456789188 kDa 98 kDa 62 kDa 49 kDa 38 kDa 28 kDa 17 kDa 16 kDa 7 kDa 4 kDa (1) (2) (3) (4) (5) (6) (7) (8) Figure 4-14. SDS-PAGE Analysis of TCA Precipitated C-Tag-ST and N-Tag-ST with Subsequent IGD with Multiple Proteases where: Lane 1 = Molecular Weight Markers; Lane 2 = TCA Pr ecipitated C-Tag-ST, IGD with Trypsin (1); Lane 3 = TCA Precipitated C-Tag-ST Pr ior to Digestion with EndoHf, IGD with Lys C (2); Lane 4 = TCA Precipitated C-Ta g-ST Prior to Digestion with EndoHf, IGD with trypsin and Glu C (3); Lane 5 = TCA Precipitated C-Tag-ST Prior to Digestion with PNGase F, IGD with Glu C and Arg C (4); Lane 6 = TCA Precipitated N-Tag-ST, IGD with As p N and Lys C (5); Lane 7 = TCA Precipitated N-Tag-ST Prior to Digestion with EndoHf, IGD with Asp N and Lys C (6); Lane 8 = TCA Precipitated N-Tag-ST Prior to Digestion with EndoHf, IGD with Glu C and Arg C (7); La ne 9 = TCA Precipitated N-Tag-ST Prior to Digestion with PNGase F, IGD with Trypsin and Glu C (8). hampered due to the increased labeling and quenching agent concentrations. In comparison to the other gels presented, the re solution was poor for this particular SDSPAGE run. Furthermore, the bands that were excised provided positive identification of tryptic peptides. However, MS analys is did not reveal acetylated AAs. MS analysis of IGD sample MALDI-TOF MS was used for analysis of digestion efficiency with two seperate

PAGE 162

145 1 2345678 (1) (2) (3) (4) (5) (6) (7)66 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa 1 2345678 (1) (2) (3) (4) (5) (6) (7)66 kDa 45 kDa 36 kDa 29 kDa 20 kDa 14 kDa Figure 4-15. SDS-Page Analysis of Initia l Screen of Ins-ST with EDAC/ETAM, NAI, and IOA where: Lane 1 = 20 uL of Puri fied Ins-ST, IGD with Trypsin (1) Lane 2 = Un-protected Ins-ST, Labele d with 17 mM NAI, IGD with Trypsin (2); Lane 3 = Un-protected Ins-ST, Labeled with 13 mM IOA, IGD with Trypsin (3); Lane 4 = CMP-NeuAc Pr otected Ins-ST, Labeled with 13 mM IOA, IGD with Trypsin (4); Lane 5 = Protected Ins-ST, Labeled with 17 mM NAI, IGD with Trypsin (5); Lane 6 = Un-protected Ins-ST, Labeled with 17 mM EDC/ETAM, IGD with Trypsin (6 ); Lane 7 = Molecular Weight Markers; Lane 8 = Un-protected In s-ST, Labeled with 17 mM EDC/ETAM, IGD with Trypsin (7). sample preparation techniques. It was f ound that a on-spot wate r wash method did not remove all the salts; however, a pep tide map corresponding to 45% coverage was obtained. Using the Zip-Tip method, only 35% coverage was obs erved; however, the peaks obtained were of much higher signalto-noise. Figure 4-17 represents a normal spectrum acquired with MALDI-TOF after IGD with trypsin and the on-spot water wash method. Figure 4-18 represents a normal spect rum after IGD with trypsin and Zip-Tip clean up. The sequence coverage from IGD with trypsin and on-spot water wash prior to MALDI-TOF is presented in Fi gure 4-19. The sequence coverage from IGD with trypsin and Zip-Tip clean up prior to MALDI-TOF is presented in Figure 4-20. The grey corresponds to identified peptid es, whereas the red corresponds to un-identified peptides.

PAGE 163

146 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)12 3456789 10 11 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)12 3456789 10 11 Figure 4-16. SDS-Page Analysis of C-Ta g-ST Labeled with NAI, with and without Substrate Present where: Lane 1 = Un -Protected C-Tag-ST Labeled with 80 mM NAI, IGD with Trypsin (1) and (2 ); Lane 2 = Protected C-Tag-ST Labeled with 80 mM NAI, IGD with Tr ypsin (3); Lane 3 = Un-protected CTag-ST Labeled with 55 mM NAI IGD with Trypsin (4); Lane 4 = Protected C-tag-ST Labeled with 55 mM NAI, I GD with Trypsin (5); Lane 5 = Unprotected C-Tag-ST Labeled with 55 mM NAI, IGD with Trypsin (6); Lane 6 = Protected C-Tag-ST Labeled with 55 mM NAI, IGD with Trypsin (7); Lane 7 = Protected C-Tag-ST Labeled with 55 mM NAI IGD with Trypsin (8); Lane 8 = Protected N-Tag-ST Labeled with 50 mM NAI IGD with Trypsin (9); Lane 9 = Un-protected C-TagST Labeled with 55 mM NAI, IGD with Trypsin (10); Lane 10 = Protected C-Ta g-ST Labeled with 55 mM NAI IGD with Trypsin (11); Lane 11 = Protecte d C-Tag-ST Labeled with 55 mM NAI, IGD with Trypsin (12). 799.01339.41879.82420.22960.63501.0 Mass ( m/z ) 0 5010.8 0 10 20 30 40 50 60 70 80 90 100% Intensity Spec-2_01_05_1201042153.4801 1868.5606 888.2341 1746.4627 2244.5381 1233.2539 1457.2671 1324.2335 908.2353 2259.5314 1728.4235 1450.2857 2526.3938 1257.2912 1871.6131 2318.5462 2156.4457 956.2924 1440.1966 1620.3317 2807.6238 1080.2822 2089.6532 1910.9954 1761.4306 1460.3692 2276.6076 3311.5825 2669.7276 1226.4638 902.3268 2844.4259 1750.2204 1045.5450 2459.6975 1987.0154 1252.2405 914.2100 1494.8537 799.01339.41879.82420.22960.63501.0 Mass ( m/z ) 0 5010.8 0 10 20 30 40 50 60 70 80 90 100% Intensity Spec-2_01_05_1201042153.4801 1868.5606 888.2341 1746.4627 2244.5381 1233.2539 1457.2671 1324.2335 908.2353 2259.5314 1728.4235 1450.2857 2526.3938 1257.2912 1871.6131 2318.5462 2156.4457 956.2924 1440.1966 1620.3317 2807.6238 1080.2822 2089.6532 1910.9954 1761.4306 1460.3692 2276.6076 3311.5825 2669.7276 1226.4638 902.3268 2844.4259 1750.2204 1045.5450 2459.6975 1987.0154 1252.2405 914.2100 1494.8537 Figure 4-17. MALDI-TOF MS Spectrum with On-Spot Water Wash Sample Preparation

PAGE 164

147 9001000110012001300140015001600170018001900200021002200 m/z, amu 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 1182.6455 1198.6489 888.4651 902.4913 1134.6585 1450.7006 1016.4725 842.5059 1472.6799 1328.6869 1205.6325 1894.8027 910.4738 1721.7738 2246.0781 DQPIFLR NSSLGDAINK LLSLPMQQPR LFYPESAHFDPK LLSLPM*QQPR (Oxidized Methionine) 9001000110012001300140015001600170018001900200021002200 m/z, amu 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 1182.6455 1198.6489 888.4651 902.4913 1134.6585 1450.7006 1016.4725 842.5059 1472.6799 1328.6869 1205.6325 1894.8027 910.4738 1721.7738 2246.0781 DQPIFLR NSSLGDAINK LLSLPMQQPR LFYPESAHFDPK LLSLPM*QQPR (Oxidized Methionine) Figure 4-18. MALDI-TOF MS Spectrum after ZipTip Sample Preparation FVDSRGGSEKKEPCLQGEAESKASK LFGNYSRDQPIFLRLEDYFWVKTPSAY ELPYGTKGSEDLLLR VLAITSSSIPKNIQSLRCRRCVVVGNGHRLRNSSLGDAI N KYDVVIR LNNAPVAGYEGDVGSKTTMRLFYPESAHFDPK VENNPDTLLVL VAFKAMDFHWIETILSDKKRVRKGFWK QPPLIWDVNPK QIR ILNPFFMEIAA DKLLSLPMQQ PR KIKQKPTTGLLAITLALH LCDLVHIAGF GYPDAYNKKQ TIHYYEQITLKSMAGSGHNVSQEALAIK RMLEMGAIK NLTSFGGHHHHHH Figure 4-19. Sequence Coverage of C-TagST IGD with Trypsin, On-Spot Water Wash Cleanup, and MALDI-TOF MS Analysis FVDSRGGSEK KEPCLQGEAESKASK LFGNYSR DQPIFLR LEDYFWVKTPSAY ELPYGTKGSEDLLLRVLAITSSSIPK NIQSLRCRRCVVVGNGHRLRNSSLGDAI NK YDVVIRLNNAP VAGYEGDVGSKTTMRLFYPESAHFDPK VENNPDTLLVL VAFK AMDFHWIE TILSDKK RVR KGFWK QPPLIWDVNPK QIRILNPFFMEIAA DK LLSLPMQQPR KIKQKPTTGLLAITLALHLCDLVHIAGFGYPDAYNKKQTI HYYEQITLK SMAGSGHNV SQEALAIKR MLEMGAIKNLTSFGGHHHHHH Figure 4-20. Sequence Coverage of C-TagST IGD with Trypsin, Zip-Tip Cleanup, and MALDI-TOF MS analysis In comparison, LC-Q MS/(MS) analysis provided higher percent coverage in comparison to MALDI-TOF data. Figure 4-21 represents a typical total ion chromatogram of a LCQ MS run perfor med on a LCQ Deca (Thermo Electron, San

PAGE 165

148 Antonio, TX). Figure 4-22 represents the MS spectra +1 parent ion corresponding to the tryptic peptide LEDYFWVK and Figure 4-23 represents the MS/MS spectrum of that peptide. Figures 4 -24 through 4-26 repres ent the percent sequen ce coverage obtained with LCQ MS for each of the three recombinant ST3Gal IV constructs. One reason for the increased percent cove rage of LCQ MS over MALDI-TOF MS analysis was the quality of data acquired. Si nce there were salts present after the on-spot water wash, the spectrum observed in Fi gure 4-17 has many salt adducts. When performing peak analysis with bioinformatic s software, several peaks are assigned as false positives. Using LCQ MS/MS provides two extra dimensions for positive identification of tryptic peptides. This incl udes the separation of tryptic peptides with chromatography and MS/MS of the pr ecursor tryptic ions. 0 5 10 15 20 25 30 35 40 Time (min) 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Relative Abundance 16.85 842.6 21.70 1051.9 23.77 1100.0 26.47 1290.3 19.16 1114.2 26.67 1289.9 27.00 914.4 16.24 947.7 28.58 755.8 42.21 431.7 13.13 730.1 29.31 861.7 12.21 609.5 32.32 854.1 41.85 432.1 40.12 432.4 10.50 892.0 9.49 1012.9 3.81 1357.3 1.83 445.1 Figure 4-21. Typical Total I on Chromatogram of Trypsin IGD of N-Tag-ST with an LCQ-DECA. The Retention Times and Base Peaks are Labeled.

PAGE 166

149 [] 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 m/z 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Relative Abundance 1099.4 1100.6 1105.7 1256.4 1077.6 1106.4 1028.8 550.9 926.0 1382.1 1279.7 898.4 1249.9 1451.2 844.1 761.2 699.4 628.9 407.0 495.3[M+H]+ [] 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 m/z 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Relative Abundance 1099.4 1100.6 1105.7 1256.4 1077.6 1106.4 1028.8 550.9 926.0 1382.1 1279.7 898.4 1249.9 1451.2 844.1 761.2 699.4 628.9 407.0 495.3[M+H]+ Figure 4-22. MS Spectra at 23.70 Minut es of the Tryptic Peptide LEDYFWVK @[] 300 400 500 600 700 800 900 1000 1100 m/z 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Relative Abundance 854.0 742.2 856.8 741.4 431.4 953.1 497.1 808.5 579.2 723.9 835.1 498.4 609.9 722.4 952.2 857.8 1064.5 567.6 790.3 662.0 1061.7 459.3 414.7 562.0 1019.3 402.0 356.4 872.9 LEDYFW b2YFWVK y3LEDYFWV b1WVK y5FWVK y4 @[] 300 400 500 600 700 800 900 1000 1100 m/z 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Relative Abundance 854.0 742.2 856.8 741.4 431.4 953.1 497.1 808.5 579.2 723.9 835.1 498.4 609.9 722.4 952.2 857.8 1064.5 567.6 790.3 662.0 1061.7 459.3 414.7 562.0 1019.3 402.0 356.4 872.9 LEDYFW b2YFWVK y3LEDYFWV b1WVK y5FWVK y4 Figure 4-23. MS/MS Spectra at 23.80 Mi nutes of the Tryptic Peptide LEDYFWVK FVDS R SGGE KK EPCLQGEAES K AS K LFGNYSRDQPIFLRLEDYFWVKTPSAY ELPYGTKGSEDLLLRVLAITSSSIPKNIQSLR C RR CVVVGNGH R L R NSSLGDAI NKYDVVIRLNNAPVAGYEGDVGSK TTM R LFYPESAHFDPKVENNPDTLLVL VAFKAMDFHWIETILSDK KR V RK GFW K QPPLIWDVNPK QI R ILNPFFMEIAA DKLLSLPMQQPR K I K QKPTTGLLAITLALHLCDLVHIAGFGYPDAYN KK QTI HYYEQITLKSMAGSGHNVSQEALAIK R MLEMGAIK NLTSF Figure 4 -24. Sequence Coverage of InsST Digested with Trypsin / LC MS/MS

PAGE 167

150 FVDS R GGSHHHHHHGGE KK EPCLQGEAES K AS K LFGNYSRDQPIFLRLEDYF WVKTPSAYELPYGTKGSEDLLLRVLAITSSSIPKNIQSLR C RR CVVVGNGH R L R NSSLGDAINKYDVVIRLNNAPVAGYEGDVGSK TTM R LFYPESAHFDPKVEN NPDTLLVLVAFKAMDFHWIETILSDK KR V RK GFW K QPPLIWDVNPK QI R ILN PFFMEIAADKLLSLPMQQPR K I K QKPTTGLLAITLALHLCDLVHIAGFGYPDA YN KK QTIHYYEQITLKSMAGSGHNVSQEALAIK R MLEMGAIK NLTSF Figure 4-25. Sequence Coverage of N-TagST Digested with Trypsin / LC MS/MS FVDS R SGGE KK EPCLQGEAES K AS K LFGNYSRDQPIFLRLEDYFWVKTPSAY ELPYGTKGSEDLLLRVLAITSSSIPKNIQSLR C RR CVVVGNGH R L R NSSLGDAI NKYDVVIRLNNAPVAGYEGDVGSK TTM R LFYPESAHFDPKVENNPDTLLVL VAFKAMDFHWIETILSDK KR V RK GFW K QPPLIWDVNPK QI R ILNPFFMEIAA DKLLSLPMQQPR K I K QKPTTGLLAITLALHLCDLVHIAGFGYPDAYN KK QTI HYYEQITLKSMAGSGHNVSQEALAIK R MLEMGAIK NLTSFGGHHHHH Figure 4-26. Sequence Coverage of C-TagST Digested with Trypsin / LC MS/MS For the LC MS/MS sequence coverage presen ted in Figures 4-24 to 4-26, note that most of the N-Terminal region was not obs erved (Red), a small region which has two cysteines (Blue) was not observed, some sma ll peptides (yellow) were not observed, a large peptide which contains sialylmotifs L (Green) was not observe d, and the C-terminal region which contains a poten tial N-glycosylation site (Orange) was not observed. Arginine and lysine (trypsin cleaves on the C-terminal side of R and K when proline is not the next amino acid) have been highlighted for regions that have not been observed. The percent coverage for all three recombin ant hST3Gal IV isoforms was similar. Digestion with proteases other than trypsin did not increase percent sequence coverage. Unfortunately, acetylated peptides were not identified with MS analysis. Most likely, the AA that was labeled was a labile species and was easily de-acetylated in the present of reducing agents needed for SDS-PA GE analysis. Experiments were performed that do not include reduction with DTT or -mercaptoethanol. When reducing agent was not present prior to SDS-PAGE analysis, IGD with trypsin did not yield tryptic peptides and the mobility of recombinant hST3GalIV was different than that of the reduced form.

PAGE 168

151 This result was similar to pepsin digeste d, non-denatured recombinant sialyltransferase presented in Chapter 3. Therefore, reduc tion of recombinant hST3Gal IV prior to proteolytic digestion was a requirement for positive identification of peptides. To identify peptides from MADLI-TOF, MS-Fit (http://prospector.ucsf.edu/ ucsfhtml4.0/msfit.htm), FindMod (http:// www.expasy.org/tools/findmod/), FindPept (http://www.expasy.org/tools/fi ndpept.html), Mascot (http ://www.matrixscience.com/), and PeptideMass (http://www.expasy.org/tools/pe ptide-mass.html) were utilized. Of the programs mentioned, Mascot and MS-Fit provid ed the best outputs for illustration of sequence coverage of non-la beled recombinant sialyltran sferase. The FindMod and FindPept software packages were used to search for possible acetylated amino acids. After submitting the mass lists extracted from the MS data files, several acetylated peptides were identified by the programs. After exhaustive manual interpretation and verification, none of the peptide matches were positive hits because the observed m/z corresponded to a signal below signal-to-noi se or the error was grater than 2.0 Da. Therefore, based on MALDI-TOF data, no acetyla ted tryptic peptides were identified. Three different programs were used to analyze the LCQ MS/MS data. PEAKS software, a de novo algorithm, was not useful becaus e it provide many false positives. Also, after searching through the positive results, many AAs in the ladder output did not correspond to the proper sequence of tryptic peptides. The ot her two programs used to interpret LCQ-MS data included Sequest a nd Mascot. On average, both programs produce the same number of correct peptide h its. However, when visually verifying the data, Sequest proved to be superior with th e extraction of MS/MS spectra and mass lists. Furthermore, it was absolutely necessary to have an in-house licensed version of both

PAGE 169

152 software packages to provide the full utility of each individual softwa re package. Sequest (Bioworks Browser) comes as a licensed softwa re package and was absolutely necessary for creating dta files needed for data analys is on the Mascot server. When trying to analyze data on the public freeware Mascot serv er, it would only interp ret 300 dta files. On average, there were over 2000 dta files (scans) obtained that may be analyzed with Mascot. Also, to simplify data analysis, a sear chable fastfa data base file was created in-house which only contained the recombinan t hST3GalIV sequence. Finally, many of the impurity peptides identified with Masc ot corresponded to human keratins. Most likely, these impurities were introduced during sample handling. Manual interpretation of MS data Exhaustive manual interpretation of MS data did not facilitate the positive identification of acetylated amino acids. Manua l interpretation of LCQ MS/(MS) data of both standard and labeled recombinant sial yltransferase was cons istent with data analyzed with bioinformatics software. Conclusion Based on the data presented in this chapter, the identity of one of the amino acids in the active-site of human ST3Gal IV was either a tyrosine or histidine. Furthermore, the data suggests that there was a change in c onformation of the enzyme based on the fact that both donor substrate (CMP -NeuAc) and acceptor substrate ( -lactose) were needed for differential protection. It may be useful to investigate other labeling agents or re-visit EDC/ETAM and IOA under different reaction cond itions. Unfortunately, MS analysis of derivatized recombinant hST3GalIV has not yiel ded the identity of the particular amino acid needed for catalysis. This was due to the fact that an acetylated tyrosine or histidine would not be stable to the r eagents used for reduction prior to SDS-PAGE analysis.

PAGE 170

153 Pepsin (Chapter 3) and trypsin digestion of non-denatured si alyltransferase did not yield proteolytic peptides. In-gel di gestion with trypsin prior to analysis has been reproducible between different batches of recombinant hST3 Gal IV. In keeping with the overall goal of identifying AAs needed for catalysis with structural proteomic based MS, Chapter 5 investigates the use of site-d irected photoaffinity labeling. Methods and Materials Iodoacetic acid (Ca # I4386), 1-ethyl -(3-dimethylaminopropopal) carbodiimide hydrochloride (Ca # 03449), ethanolamine (Ca # 411000), iodoacetamide (Ca # 57670), N-acetylimidazole (Ca # 01194) were purchased from Sigma-Aldrich (St. Louis, MO). Acetic acid (Ca # A507-500), formic acid (Ca # BP 1215-500), methanol (Ca # 230-4), acetonitrile (Ca # A998-1), and high-purit y water (Ca # AH 365-4 or W 7-4) were purchased from Fisher-Scientific (Fair Lawn, NJ). Reaction conditions for preliminary screening of SMLAs Standard (Table 4-2) Ins-ST (20 L, 4.6 ug, 6.4 M) was diluted with 15 L of 50mM MES, pH 6.8 and incubated at 30 oC for 10 minutes, after which the activity was determined. To perform a deglycosylation reaction, 3 L of EndoHf buffer and 2 L of EndoHf was added and incubated at 37 oC overnight. NAI (Table 4-2) Ins-ST (20 L, 4.6 g, 6.4 M) was diluted with 5 L of 50 mM MES, pH 6.8 then 10 L of 50 mM NAI in 50 mM MES, pH 6.8 and incubated at 30 oC for 10 minutes, after which the activity wa s determined. Following the assay, the reactions were quenched with 5 L of 100 mM tyrosine. To perform a deglycosylation reaction, 3 L of EndoHf buffer and 2 L of EndoHf was added and incubated at 37oC overnight.

PAGE 171

154 IOA (Table 4-2) Ins-ST (20 L, 4.6 g, 6.4 M) was diluted with 5 L of 50 mM MES, pH 6.8 then 10 L of 40 mM IOAA in 50 mM MES, pH 6.8 and incubated at 30oC for 10 minutes, after which the activity wa s determined. Following the assay, the reactions were quenched with 5 L of 80 mM histidine. To perform a deglycosylation reaction, 3 L of EndoHf buffer and 2 L of EndoHf was added and incubated at 37oC overnight. EDAC and ETAM (Table 4-2) Ins-ST (20 L, 4.6 g, 6.4 M) was diluted with 5 L of 50mM MES, pH 6.8 then 5 L 50 mM EDAC and 5 L 50 mM ETAM in 50 mM MES, pH 6.8 and incubated at 30 oC for 10 minutes, after which the activity was determined. Following the assay, the reactions were quenched with 5 L of 50 mM acetic acid. To perform a deglycosylation reaction, 3 L of EndoHf buffer and 2 L of EndoHf was added and incubated at 37oC overnight. NAI labeling experiments No substrate present (Figure 4-9) C-Tag-ST (40 L of 833 g/mL; 22.5 M; [ST]f =11.25 M) was diluted with 30 L of 50 mM MES, pH 6.8 and allowed to equilibrate for 10 minutes at 30 oC. Following equilibration, 10 L of 440 mM NAI ([NAI]f = 55 mM) was added to initiate labe ling. At different time points, 8 L of the reaction mixture was added to 4 L of 3.9 M tyrosine ([Tyr] f = 1.3 mM) and incubated for 5 minutes at 30 oC. To keep the concentration of the assays normalized between protected and un-protected reactions, th e pre-made assay was spiked with 1.8 L of 1.4 mM CMP-NeuAc, pH 6.8 (11.8 L total; [CMP-NeuAc] f = 163 M; specific activity =

PAGE 172

155 8.1 Ci/ mol). Activity was determined by the addition of 10 L of the quench mixture to the spiked assay and following the a ssay protocol presented in Chapter 2. Active site protection with CMP-NeuAc (Figure 4-9) C-Tag-ST (40 L of 833 ug/mL; 22.5 M; [ST]f =11.25 M) was diluted with 30 L of 1.4 mM CMP-NeuAc ([CMP-NeuAc]f = 525 M) and allowed to equilibr ate for 10 minutes at 30 oC. Following equilibration, 10 L of 440 mM NAI ([NAI]f = 55 mM) was added to initiate labeling. At different time points, 8 L of the reaction mixture was added to 4 L of 3.9 M tyrosine ([Tyr] f = 1.3 mM) and incubated for 5 minutes at 30 oC. To keep the concentration of the assays normalized betw een protected and un-pr otected reactions, the pre-made assays were spiked with 1.8 L of 50 mM MES, pH 6.8 (11.8 L total volume ([CMP-NeuAc]f = 206 M; specific activity = 6.4 Ci/ mol). Activity was determined by the addition of 10 L of the quench mixture to the spiked assay and following the assay protocol presented in Chapter 2. No Substrate Present (Figure 4-10) C-Tag-ST (40 L of 833 g/mL; 22.5 M; [ST]f =11.25 M) was diluted with 30 L of 50 mM MES, pH 6.8 and allowed to equilibrate for 10 minutes at 30 oC. Following equilibration, 10 L of 440 mM NAI ([NAI]f = 55 mM) was added to initiate labe ling. At different time points, 8 L of the reaction mixture was added to 4 L of 3.9 M tyrosine ([Tyr] f = 1.3 mM) and incubated for 5 mins at 30 oC. To keep the concentration of the assays normalized between protected and un-protected reactions, the pre-made assays were spiked with 2 L of 2.8 mM CMP-NeuAc, pH 6.8 (12.0 L total volume, [CMP-NeuAc]f = 300 M; specific

PAGE 173

156 activity = 4.33). Activity was de termined by the addition of 10 L of the quench mixture to the spiked assay and following the a ssay protocol presented in Chapter 2 Active site protected with 1.05 mM CMP-NeuAc/saturated -lactose (Figure 410) C-Tag-ST (40 L of 833 ug/mL; 22.5 M; [ST]f =11.25 M) was diluted with 30 L of 2.8 mM CMP NeuAc ([CMP-NeuAc]f =1.05 mM) and saturated -lactose and allowed to equilibrate for 10 minutes at 30oC. Following equilibration, 10 L of 440 mM NAI ([NAI]f = 55 mM) was added to initiate labeling (30 oC). At different time points, 8 L of the reaction mixture was added to 4 L of 3.9 M tyrosine ([Tyr] f = 1.3 mM) and incubated for 5 minutes at 30 oC. To keep the concentra tion of the assays normalized between protected and un-prote cted reactions, the pre-made assay was spiked with 2 L of 50 mM MES, pH 6.8 (12.0 L total volume [CMP-NeuAc]f = 365 M before consumption during labeling expe riment; specific activity = 3.6 Ci/ mol). Activity was determined by the addition of 10 L of the quench mixture to the spiked assay and following the assay protocol presented in Chapter 2. Protected with saturated -lactose (Figure 4-10) C-Tag-ST (40 L of 833 g/mL; 22.5 M; [ST]f =11.25 M) was diluted with 30 L of saturated -lactose and allowed to equilibrate for 10 minutes 30oC. Following equilibration, 10 L of 440 mM NAI ([NAI]f = 55 mM) was added to initiate labe ling. At different time points, 8 L of the reaction mixture was added to 4 L of 3.9 M tyrosine ([Tyr] f = 1.3 mM) and incubated for 5 minutes at 30 oC. To keep the concentra tion of the assays normalized between protected and un-prote cted reactions, the pre-made assay was spiked with 2 L of 2.8 mM CMP-NeuAc, pH 6.8 (12.0 L total volume, [CMP-NeuAc]f = 300 M;

PAGE 174

157 specific activity = 4.3 Ci/ mol). Activity was determined by the addition of 10 L of the quench mixture to the spiked assay and following the assay protocol presented in Chapter 2. SDS-PAGE analysis and in-gel dige stion of recombinant hST3Gal IV Recombinant hST3Gal IV samples were analyzed with SDS-PAGE, followed by in-gel digestion with Trypsin Gold (P romega), Asp N, Glu C, Lys C, or chymotrypsinogen A (Sigma-Aldrich). St andard Ins-ST, N-Tag-ST, or C-Tag-ST samples were precipitated with trichloro acetic acid (TCA) before dilution with 20 L of a buffer containing 100mM Tris, 1% SDS, 7 M -mercaptoethanol, 20% glycerol, and 0.2% w/v bromophenol blue (SDS-PAGE load buffer). To perform non-reducing SDSPAGE analysis, the load buffe r described did not contain -mercaptoethanol. For the initial screen labeling experiments, 20 L of the quenched reactions was diluted to 40 L with SDS-PAGE load buffer prior to SDS-P AGE analysis. For the SDS-PAGE analysis of the labeling experiment pres ented in Figures 4-9 and 4-10, 20 L of the quenched reactions was diluted to 40 L with the SDS-PAGE load buffer. Gels were stained with a buffer contai ning 0.2 % (w/v) coomassie blue R-250, 40% MeOH, and 5 % acetic acid followed by de-s taining with a buffer containing 5 % MeOH and 10% acetic acid. Bands corresponding to recombinant hST3Gal IV were cut into 1mm cubes and washed 2 times with water in 500 L microfuge tubes. The cubes were then de-stained with 100 to 150 L of 25 mM ammonium bicar bonate/50 % acetonitrile at 27 oC for 4 to 8 hours with removal and reappl ication 3 times. After all the coomassie blue stain was extracted from the gel slices, they once again were washed with water two times. To digest recombinant hST3Gal IV, 50 L of 20 ng/ L of trypsin in 20 mM

PAGE 175

158 ammonium bicarbonate buffer was added and allo wed to absorb into the gel cubes for 30 minutes at 27 oC. Next, the appropriate volume of 20 mM ammonium bicarbonate was added to cover the gel cubes and allowed to incubate at 27 oC overnight. After overnight incubation, the supernatant was collected and saved. The peptides were then extracted from the gel slices with 100 L of 5% v/v acetic acid, 30 % acetonitrile three times. The extracted peptide solution was added to the other saved supernatan t. This solution was then concentrated to dryness with a La bconco CentriVap Concen trator (Model number 7810000) at 37 oC. Mass spectrometry analysis For MALDI-TOF MS analysis, peptide samples were re-suspended in 10 to 20 L of 30% acetonitrile, 0.1 % TFA for the on-spot cold water wash or 0.1 % TFA for ZipTip cleanup. For LCQ MS analysis, samples were re-suspended in 10 to 20 L of 5 % acetonitrile, 0.1 % TFA or 5 % ac etonitrile, 0.5 % acetic acid. All three constructs were analyzed with MALDI-TOF MS and LCQ MS after enzymatic di gestion. MALDI-TOF analysis was performed on a Br uker Daltonics Reflex II (Biller ica, MA) retrofitted with delayed extraction, an Applied Biosystems Voyager-DE Pro (Foster City, CA), or an Applied Biosystems QStar XL (Foster Cit y, CA). All the systems described were equipped with a 337 nm UV laser. Analysis settings for the Reflex II were: a source voltage of 20 kV, extraction voltage of 14.7 kV, focus voltage of 6 kV, and reflectron voltage of 21.5 kV. Analysis settings for the Voyager DE Pro were: a source voltage of 20 kV and an extraction voltage of 14.4 kV. Analysis settings for the QStar XL were: source voltage of 19 kV, extraction voltage of 13.2 kV, focus voltage of 6 kV, and reflectron voltage of 20.5 kV. Spectra for all instruments were acquired in reflectron

PAGE 176

159 mode by averaging a period of 50 to 100 scans (4 -5 Hz) of highest sign al to lowest laser power attenuation with a mass range of 500 to 5500 m/z. All three instruments yielded similar peptide identification; however, the Re flex II was more difficult to calibrate as compared to the other two instruments. The matrix, -cyano-4-hydroxycinnaminic acid, was re-crystallized from ethanol prior to dilution to 10 mg mL in 50% acetonitrile, 0.1% trifluoroacetic acid. For the on-spot wash method, 3 L of dilute peptides was added to 5 L of 10 mg/mL matrix, after which, 1 L was spotted onto a stainles s steel plate and allowed to dry. Upon drying, 1 L of ice cold water was applied a nd allowed to inc ubate at room temperature for 5 minutes, when upon, a Chem Wipe was used to remove the water by capillary action. After wash ing the spot three times, a solution containing 30 % acetonitrile, 0.1 % TFA was applied to the sp ot and allowed to dry. ZipTip cleanup was performed with Milipor es (Billerica, MA) ZipTip -C18 Pipette Tip with 0.2 L C18 resin (Catalogue # ZTC18M960). All actions were performed with the ZipTip on a 10 L Eppendorf pipette without allowing the resin to dry. First, the tips were activated with a buffer containing 50% acetonitrile, 50 % water, and 0.1 % TFA. Second, the tips were washed with a buffer containing 100 % water, and 0.1 % TFA to remove organic solvent. Third, 5 to 10 L of sample was loaded onto the column bed. Fourth, the sample was washed three times with a buffer cont aining 100 % water, and 0.1% TFA to remove salt impurities. Lastly, 0.5 L of a buffer containing 50% acetonitrile, and 0.1% TFA was used to elute peptides directly to th e MALDI-TOF support plate. Once eluted onto the plate, 0.5 L of 10 mg/mL -cyano-4-hydroxycinnaminic aci d was added to the spot

PAGE 177

160 before drying. Both the on-spot water wash and ZipTip cleanup produced symmetric matrix crystals. LCQ-MS was performed on a Thermo El ectron LCQ Deca (San Jose, CA) or a Thermo Electron LCQ Classic operated in po sitive ion mode. For the LCQ Deca, a spray voltage of 1.8 kV +/0.2 was applied at the liquid junction before a capillary column. An Eldex MicroPro pump delivered solvent at 6 L/min to a LC Packings C18 PepMap Nano-Prepcolumn (0.3 x 1 mm). Five microliters of sample was loaded onto the nano-prepcolumn and washed for 5 mins at 0 % mobile phase B. After washing, the direction of flow was revers ed with a new flow rate of 200 nL/min after pre-column splitting from 8 L/min. Elution of peptides fr om a New Objective (0.75mm x 5 mm) Proteopep capillary column included a gr adient of 3-60% mobile phase B over 30 minutes and 60-90% mobile phase B over 1 minute with mobile phase A as 0.1% acetic acid, 3% acetonitrile and mob ile phase B as 0.1% acetic ac id, 95% acetonitrile. For the LCQ Classic, a spray voltage of 3.3 kV +/0.2 was applied at the ESI tip and 15.0 kV was applied at the capillary colu mn inlet. An Agilent (HP) 1100 series pump delivered solvent at 200 L/min to a Phenomenex (Torrance, CA) Luna C18 guard column (2 x 4 mm) in-line with a Waters (Milford, MA) Symmetry Shield RP18 column (2.1 x 150 mm) or a Phenomenex Synergi Hydro-RP 80 column (2 x 150 mm; 4 m). Using a Manual Rhodyne 7125 valve, 20 L of sample was injected. Elution of peptides included a gradient of 0-15% mobile phase B over 5 minutes and 15-95% mobile phase B over 75 minutes with solvent A as 0.5% acetic acid, 5% methanol and solvent B as 0.5% acetic acid, 95% methanol.

PAGE 178

161 CHAPTER 5 MS BASED STRUCTURAL PROTEO MICS WITH SITE-DIRECTED PHOTOAFFINITY LABELING FO R THE INVESTIGATION OF SIALYLTRANSFERASE ACTI VE SITE AMINO ACIDS Introduction to Site Directed Photo-affinity Labeling The identification of active-site amino acids may be investigated using site-directed photoaffinity labels. In a very similar ma nner to the small molecule labeling agents presented in Chapter 4, amino acids may be specifically labeled with site-directed photoaffinity labels to aid in proteomics for studying protein function, mechanism, and interaction networks.164 A photoreactive group is inert until irradiation unmasks a highly-reactive intermediate for reaction with amino acid si de chains. Site-directed photoaffinity labels differ than the small affinity labels because they are designed to localize in the active-site region of the protein before the unmasking the reactive species. Photo-reactive probes are designed to have high affinity for the target site, be stable under storage and reaction conditions, generate an excitation state faster than the Kd of the protein-ligand complex to allow covalent linkage, form a single covalent adduct, and yield an active form that will insert into C-H and X-H bonds.165 To expand on the requirements, the molecules should closely re semble the natural substrate and should be somewhat simple to synthesize. Furthe rmore, the activation wavelength of the photoaffinity label should not be less than 254 nm, should have a high molar absorbtivity, and should have a high quantum yield. There are several types of photoaffinity labels, many of which are commercially available as cross-linking agents. For this application, a CMP-NeuAc mimic with an aryl

PAGE 179

162 azide was synthesized by Dr. Nicole Horenste in. The derivative was chosen because it closely resembles the structure of known sialyl transferase inhibitors that contain aromatic rings. Aryl azides as covalent labeling agents have had success with identifying receptors in a mixture of proteins, identifying multi-s ubunit binding regions, functional proteomics, photoaffinity time dependent studies, and iden tifying protein/substr ate binding sites. Photoreactive Groups Photoaffinity labels ar e commercially available (u sually for cross linking experiments) or synthesized in-house for a case specific purpose (cross-linking or sitedirected affinity label). The most co mmonly used photoreactive moieties include benzophenones, diazirines, aryl azides, and perfluorinated aryl azides. Benzophenone O N3N3N N F3C FF F FBenzophenone Aryl Azide Perfluorinated Aryl Azide Diazirine O N3N3N N F3C FF F FBenzophenone Aryl Azide Perfluorinated Aryl Azide Diazirine Figure 5-1. Different P hotoreactive Moieties preferentially reacts with normally un-reactiv e C-H bonds and will not react with water. The main limitations with benzophenone are th e geometry required for H abstraction and steric hindrance. Diazirine will form a carbene upon photolysis that will react with all 20 amino acids. The main limitations with diazir ines are its difficult synthesis and general

PAGE 180

163 stability. Aryl azides (phenylazides) and perf luorinated aryl azides form nitrenes upon photolysis and may proceed through a variety of different reaction pathways (Figure 5-2). The main limitation of aryl azides is the rearrangement of the singlet nitrene to dehydroazepine (Figure 5-2). The rearrangem ent limits the reactivity to nucleophilic N N N N H N NH N R N R R N UV Light+ -:R NH2 R H R Ring Expansion Phenylazide NitreneFormation Activated Hydrogen (C-H) Insertion Activated Hydrogen (N-H) Insertion Addition Reactions Nucleophile N N N N H N NH N R N R R N UV Light+ -:R NH2 R NH2 R H R H R Ring Expansion Phenylazide NitreneFormation Activated Hydrogen (C-H) Insertion Activated Hydrogen (N-H) Insertion Addition Reactions Nucleophile Nucleophile N N N N H N NH N R N R R N UV Light+ -:R NH2 R H R Ring Expansion Phenylazide NitreneFormation Activated Hydrogen (C-H) Insertion Activated Hydrogen (N-H) Insertion Addition Reactions Nucleophile N N N N H N NH N R N R R N UV Light+ -:R NH2 R NH2 R H R H R Ring Expansion Phenylazide NitreneFormation Activated Hydrogen (C-H) Insertion Activated Hydrogen (N-H) Insertion Addition Reactions Nucleophile Nucleophile Figure 5-2. Reaction Pathways of Activated Phenylazide amino acids. To reduce the likelihood of rearrangement, the aryl azide may be fluorinated.166 The bottle neck in using photoreac tive molecules for proteomics is the design and synthesis of the photoaffinity la bel. Although these phot oreactive moieties have been utilized for over 50, the mechanis m of insertion is s till poorly understood. However, not understanding the mechanis m does not limit the usefulness of these molecules for solving structural proteomic questions.

PAGE 181

164 Practical Considerations for Site-Directed Photoaffinity Labeling When performing a site-directed photoaffi nity labeling experiment, the ratio of substrate to protein should be close to unity. If the concentration of the labeling agent is much greater than the protein concentra tion, non-specific binding may occur. Also, before irradiation, the reaction mixture shoul d be equilibrated for 10 to 30 minutes. Next, short irradiation times (1 to 6 minutes ) are preferred because long irradiation times will damage the protein. Lastly, aryl azides will react with thiols (reducing environments) or buffers such as Tris-HCl.167 Traditionally, identification of photoaffin ity labeled amino acids was performed with radiolabel molecules followed by prot ein digestion, SDS-PAGE analysis, HPLC-UV detection, and/or Edman degrad ation. As stated before, MS analysis allows for highthroughput analysis of photolabeled proteins with a t op-down or bottom-up approach. Furthermore, MS/MS analysis may allow iden tification of the ami no acid derivatized by the photoaffinity labeling agent, a feat that in not usually reported by radiolabel incorporation.168 Examples of Photoaffinity Labeling Over the last 50 years, many laboratories ha ve used photoaffinity labeling to solve interesting biological questions. The following table presents just a few successful case specific photoaffinity labeling experiments. To keep the list short, only aryl azides and the important information pertaining to e xperimental conditions and analysis are summarized. Practical Consideration for Site Directed Ph otoaffinity Labeling of Sialyltransferase The design of the aryl azide inhibitor prep ared may allow for and investigation of all members of the sialyltransferase family. The literature on sialyltransferases donor

PAGE 182

165 Table 5-1. Literature Review on Ca se Specific Photoaffinity Labeling Protein System and Reference Number Site-Directed Photoaffinity Labeling Molecule Important Information 1) RNA polymerase subunit crosslinking167 aryl azide cross-linkers: N-[(5-azido-2nitrobenzyl)oxy) succinimide and Nhydroxysuccinimidyl)-4azidosalicyclic acid Tested experimental factors influencing cross-linking. The azide was activated by 5 minutes exposure to 366 nm light at a distance of 3.4 cm. DTT and 2mercaptoethanol were necessary for maintaining enzyme activity. Found that DTT at a pH lower than 8.0 was less effective at reducing azide and increased cross-linking efficiency. 2) Active-site labeling of glucosidase I169 Derivatized natural inhibitor 1deoxynojirimycin (DNM) to 4-(pazidosalicylamido)butyl5-amido-pentyl-1-DNM Identified 24 k Da peptide that was radiolabeled with 125I. The protein is normally associated with low level expression and purification, thus photoaffinity labeling is used for secondary structure information. Inhibitor was preincubated for 10 minutes. The azide was activated by 1 minute exposure to 254 nm light at a distance of 7.5 cm. 3) Active-Site labeling of sialidase170 Derivatized NeuAc to5N-acetyl-9-(4azidosaicoylamido)-2deoxy-2,3didehydroneuraminic acid Activity was lost upon irradiation of protein in the presence of derivatized substrate. At 4oC, the substrate was pre-equilibrated for 5 minutes before irradiation for 1 minute at 254 nm at a distance of 1 cm. Proposed the use of photoreactive NeuAc on sialyltransferases. specificity illistrates that the enzyme can accep t several substrates with modifications to sialic acid. Interestingly, ma ny investigators have chosen to derivatize the 4, 5, or 9 positions of sialic acid with photoaffinity or fluorescent labels for transfer to oligosaccharides with sialyltransferase. Th ese derivatized sialic acids were useful for

PAGE 183

166 Table 5-1. Continued Protein System and Reference Number Site-Directed Photoaffinity labeling molecule Important Information 4) Active-site labeling of UDPGlucose Dehydrogenase171 Derivatized UDPGlucose to [32P]5N3UDPGlucose Used differential labeling scheme to protect active site from labeling when natural substrate UDPglucose present with azide derivative. Cystine identified as labeled amino acid by tryptic digestion, HPLC, and Edman degradation. Labeling was achieved with a hand-held 254-nm UV lamp for 6 minutes at 4 oC. After 10 minutes pre-incubation of protein with 5N3UDP-glucose, 30 % of the activity was lost, and after 5 minutes of irradiation, 90 % of the activity was lost 5) Active-site labeling of RNA polymerase III transcription complexes166 Four different photoreactive groups were attached to dUMP: aryl azide, benzophenone, perfluoronated aryl azide, or a diazirine was attached to 5-aminoallyl deoxy uridine Found that diazirine was more reactive towards different DNAbinding proteins than aryl azide, benzophenone, or perfluorinated aryl azide. 6) Membranetopology probe for uridine diphosphatesugar binding proteins172 Derivatized UTPP3-(4azidoanilido)uridine 5triphosphate[ -P32] The design of the experiment allowed answers to functional proteomics questions on membrane bound glycosyltransferases. The substrate was allowed to equilibrate for 15 minutes on ice prior to a 1 to 2 minute irradiation with a hand-held UV lamp (UVG11) at 254 nm at a distance of 7.6 cm. Four glycosyltransferases and 3 sugar transporters were identified. studying disease states and si alic acid binding proteins.173-176 In these studies, the large bulky groups did not affect the transfer of the derivatized sialic acid to the target

PAGE 184

167 oligosaccharide. In the scope of this di ssertation, the NeuAc moiety is completely removed and replace with 5-azido mande lic acid for photoaffinity labeling of sialyltransferase itself. The most important interactions fo r binding of the donor substrate to sialyltransferase prior to catalysis are related to the CMP moiety.177 CMP and CDP alone are sialyltransferase inhibitors with Kis of 50 M and 19 M respectively. A study on the series of nucleosides cytidin e, deoxycytidine, and cytosine, illustrate that cytidine is the best inhibitor, whereas cytosine is the least effective inhibitor. This information suggests that the ribose of th e nucleotide is also needed for strong enzyme-substrate binding properties. The importance of phosphate binding to the active si te is still unclear, and the complete NeuAc sugar is not essential for increased affinities.177;178 Recently, Schmidt and co-workers synthe sized molecules which contain mandelic acid instead of the NeuAc moiety.179 These inhibitors have Kis between 0.2 and 10 M based on the stereochemistry of a negatively charged phosphate or ca rboxylate (Figure 53).177-179 Since sialyltransferase are strongly inhibited by the mandelic acid containing substrate analogue molecules, this provides st rong basis for the use of aryl acids as a photoaffinity labeling agent for identific ation of active-site amino acids. Results and Discussion CMP-o-azido Mandelate Synthesis and Characterization The first generation of donor substrat e analogues with an aryl azide was synthesized by Dr. Nicole Horenstein (Figur e 5-4). UV analysis provided two absorption

PAGE 185

168 O N N NH2O O H H O O P O O O Y X ZZ = Ph, X = H, Y = P(O)(OH)(ONa) Ki = 0.20 uM Z = Ph, X = P(O)(OH)(ONa), Y = H Ki = 0.35 uM Z = Ph, X = H, Y = CO2NaKi = 10 uM Z = Ph, X = CO2Na, Y = HKi = 7 uM O N N NH2O O H H O O P O O O Y X ZO N N NH2O O H H O O P O O O Y X ZZ = Ph, X = H, Y = P(O)(OH)(ONa) Ki = 0.20 uM Z = Ph, X = P(O)(OH)(ONa), Y = H Ki = 0.35 uM Z = Ph, X = H, Y = CO2NaKi = 10 uM Z = Ph, X = CO2Na, Y = HKi = 7 uM Figure 5-3. Aromatic Ring Substituted for NeuAc and Ki Values179 bands at 254 nm and 280 nm, with the mo lar absorbtivity determined to be 11800 cm-1 and 11400 cm-1 M-1 respectively. CMP-o-azido mande late was purified with strong cation exchange after synthesis. The collected fractions were exhaustively desalted with Amberlite resin in the H+ form, rotovaped to dryness, and re-suspended in deionized water. The purification and desalting procedure produced pure CMP-o-azido mandelate as observed with NMR and MS analysis. N O OOH OH O OHO P NNH2ON N N O O Figure 5-4. CMP-o-azido Mandelate

PAGE 186

169 MS analysis on a Bruker Reflex II FT-I CR MS operated in negative ion mode yielded a clean spectrum (Figure 53). To validate the utility of the molecule, i.e. the aryl azide decomposes, the photolysis of the aryl azide was monitored with MS in negative ion mode ESI-FTICR-MS. CMP-o-azido mandelate was diluted to 200 M and irradiated at 254 nm for 6 minutes with 50 L aliquots removed at 1 minute intervals. The aliquots were then diluted to 100 L with 100 % methanol prior to MS analysis. Figure 5-5 plots the change in signal intensity of the stari ng material and products. The starting CMP-o-azido mandelate is 498.0895 Da. The major product af ter irradiation has a molecular weight of 488.0850 Da, thus corresp onding to the loss of nitrogen and the addition of water. The other product s formed, m/z of 1316.88 and 1150.90, are most likely CMP-o-azido mandelate self insertion a dducts. The purpose of this experiment was to determine if the CMP-o-azido mandela te would be useful as a site-directed labeling agent. The results indicate that CMP-o-azido mandelate will generate a reactive species; therefore, no further characteriza tion of the molecule was performed and irradiation of CMP-o-mandelate with reco mbinant hST3Gal IV was performed. Site-Directed Photoaffinity Labeling of Sialyltransferase with CMP-o-Azido Mandelate The Ki of CMP-o-azido mandelate was not determined; however, Figure 5-7 suggest that a mode of inhibition ma y be occurring. The N-Tag-ST [6.5 M]final construct was incubated at 37 oC without and with inhibitor [180 M]final over two hours with the activity dassayed at 5, 55, and 130 minutes. Over the time period, N-Tag-ST without inhibitor showed a loss of 8 % of th e original activity. Next, it was postulated that if a cysteine was in the active site n ear the azide moiety, th e inhibitor would have been a suicide inhibitor. Since activity does not drop to 0 at the high substrate to

PAGE 187

170 491496501506m/z 0 1000000 2000000 3000000 4000000 5000000 6000000 7000000 8000000 9000000 10000000 11000000 Figure 5-5. Negative Ion Mode ESI-FTICR MS Spectrum of CMP-o-azido Mandelate 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 3.00E+07 3.50E+07 4.00E+07 4.50E+07 0123456Time in MinutesIntensit y Series 1 ( ): 498.0895 C17H19N6O10P1Series 2 ( ): 488.0850 C17H21N6O11P1Series 3 ( ): 1316.8821 Unknown Series 4 (x): 1150.9030 Unknown 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 3.00E+07 3.50E+07 4.00E+07 4.50E+07 0123456Time in MinutesIntensit y Series 1 ( ): 498.0895 C17H19N6O10P1Series 2 ( ): 488.0850 C17H21N6O11P1Series 3 ( ): 1316.8821 Unknown Series 4 (x): 1150.9030 Unknown Figure 5-6. Negative Ion Mode ESI-FTICR MS Analysis of CMP-o-azido Mandelate Decomposition Products

PAGE 188

171 protein ratio (28:1), this s ubstrate may not be a suicide inhibitor; however, further inhibition studies are required to solidify th is postulation. Since there was a drop in activity, it was likely that the inhibitor was binding in the active site of the N-Tag-ST. Table 5-2 represents the percent original ac tivity at the different time points plotted in Figure 5-7. Note that there was a loss of 23 % of the initial activity after the first 5 minutes. The loss of activity presented in Table 5-2 will be compared to the loss of activity observed when N-Tag-ST/inhi bitor complexes are irradiated. -5.00 5.00 15.00 25.00 35.00 45.00 55.00 65.00 75.00 85.00 95.00 105.00 115.00 125.00 020406080100120140Incubation Time in Minutes @ 37oCActivity in nmol/min Product N-Tag-ST Without Inhibitor Present N-Tag-ST With 180 uM Inhibitor Present -5.00 5.00 15.00 25.00 35.00 45.00 55.00 65.00 75.00 85.00 95.00 105.00 115.00 125.00 020406080100120140Incubation Time in Minutes @ 37oCActivity in nmol/min Product N-Tag-ST Without Inhibitor Present N-Tag-ST With 180 uM Inhibitor Present Figure 5-7. Activity Profile of N-Tag-ST Incubated With and Without Inhibitor Present. Table 5-2. Percent Activity Remaining afte r Incubation of N-Tag-ST With and Without Inhibitor Present. Time Point N-Tag-ST no Inhibitor N-Tag-ST with 180 M inhibitor 5 min 100 % 77 % 55 min 96 % 49 % 130 min 92 % 40 % Irradiation of N-Tag-ST from 5 to 12 minut es without inhibitor at 254 nm with a Hamamatsu Corporation (Middlesex, NJ) 0.65 amp lamp, at 254 nm with a short wave

PAGE 189

172 Ultra-violet products incor porated (Upland, CA) 0.16 amp lamp, or at 365 nm with a Ultra-violet products incor porated (Upland, CA) 0.16 amp lamp reduced the activity by 4 to 8 %. Table 5-3 presents the da ta collected upon irradiation of 6.5 M N-Tag-ST with 50, 98, or 180 M inhibitor present. The top row corresponds to the concentration of CMP-o-azido mandelate. The percent activity reported was normalized to 100 % of the activity prior to irradiation or addition of inhibitor. The left column corresponds to the time irradiated at 254 nm with the 0.65 amp lamp at 5 cm from the top of an open 500 L microfuge tube. The 50 M, 98 M, and (1)-180 M samples were pre-incubated for 30 minutes, whereas the (2)-180 M sample was pre-incubated for 90 mins. Time 0 corresponds to the activity after pre-equilibration. Table 5-3. Irradiation of N-Tag-ST at 254 nm (0.65 Amps) with Different Concentrations of Inhibitor Irradiation Final Concentration of Inhibitor and Percent Activity Time in mins 50 M 98 M 180 M (1) 180 M (2) 0 62 % 76 % 46% 21 % 1 54 % 35 % 2 59 % 38 % 46 % 3 38 % 45 % 4 56 % 43 % 45 % 5 6 16 % 12 13 % These experiments suggest that labeling was not occurring, however inhibition was. The 98 M data set appears to demonstrate la beling, however this data set was not consistent with the other findings. According to the literate presented, deactivation of the enzyme should proceed over a 1 to 6 minute irradiation period. Interestingly, the preequilibration values (0 minutes in Table 5-3) are similar to the tr end observed in Table 52. Overall, there was a decrease in activity after irradiation; howe ver, the change in

PAGE 190

173 activity was within th e 4 to 8% observed with irradiat ed N-Tag-ST without inhibitor present. To determine if the N-Tag-ST c onstruct was labeled with the CMP-o-mandelate, samples were analyzed with SDS-PAGE, in-gel digested with trypsin, and analyzed with LC-MS. At this point, commercially available sialyltransferase was purchased to see if they could be labeled with the inhibitor. Recombinant rat 2,3-(N)-sialyltransferase and recombinant rat 2,6-(N)-sialyltransferase, purchased from Calbiochem (San Diego, CA), were investigated under similar conditions as the N-Tag-ST construct. In Table 5-4, the top row corresponds to the different UV lamp wave lengths and powers used for irradiation. For all the experiments presented, the lamps were held at a distance of 5 cm. Prior to irradiation, the final concentration of i nhibitor was brought to 130 M in the presence of recombinant rat 2,3-(N)-sialyltransferas e (Table 5-4) to a final volume of 50 L and allowed to pre-equilibrate for 30 minutes. The percent activity reported was based on 100 % activity prior to irradiation or addition of inhi bitor with the same dilution factor as the solution co ntaining inhibitor. In a similar manner, 2,6-(N)-sialyltransferase (Table 5-5) was pre-incubated with 130 M of inhibitor with a final volume of 50 L for 30 minutes prior to irradiation with the lamps already described. In Ta ble 5-5, w/o inhibitor corresponds to 2,6-(N)sialyltransferase without inhibitor present and w/ inhibitor corresponds to 2,6-(N)sialyltransferase in the presence of 130 M inhibitor. Percent activities were normalized to the initial activity observed prior to addition of inhi bitor or irradiation. Also, the 2,6(N)-sialyltransferases acceptor substrat e was N-acetyllactosamine as opposed to lactose, thus the activity assay cont ained N-acetyllactosamine instead of -lactose.

PAGE 191

174 Table 5-4. Irradiation of Recombinant Rat 2,3-(N)-sialyltransferase with 130 M Inhibitor Present Irradiation Percent Activity with Different UV Lamps Time in Minutes 254 nm (0.65 amps) 365 nm (0.15 amps) 254 nm (0.15 amps) 365 nm (0.15 amps) 0 27 % 27 % 14 % 29 % 5 24 % 39 % 21 % 43 % Table 5-5. Irradiation of Recombinant Rat 2,6-(N)-sialyltransferase without Inhibitor Present and with 130 M Inhibitor Present Irradiation Percent Activity with Different UV Lamps Time in minutes 254 nm (0.65 amps) 254 nm (0.65 amps) 365 nm (0.15 amps) 365 nm (0.15 amps) w/o inhibitor w/ inhibitor w/o inhibitor w/ inhibitor 0 100 % 52 % 100 % 29 % 5 166 % 153 % 203 % 129 % The results presented in Table 5-4 for th e higher power 254 nm lamp have the same results observed with the irradiated N-Tag-ST/inhibitor complex; the final activity was within the observed 4 to 8 % loss of activity for N-Tag-ST without inhibitor present. Interestingly, irradiation of recombinant rat 2,3-(N)-sialyltransferase in the presence of inhibitor with the lower powered 254 nm a nd 365 nm lamps resulted in an increased activity in comparison to the pre-equilibrated activity value. The results for the 2,6-(N)-sialyltransferase (Table 55) were much different than the N-Tag-ST or the recombinant rat 2,3-(N)-sialyltransferase. After 5 minutes of irradiation of 2,6-(N)-sialyltransfer ase with and without inhibitor present, activity increased. The data suggests that irradiation of recombinant rat 2,6-(N)sialyltransferase produces a different species which has higher catalytic activity. This may be due to activation, cross-linking, or de rivatization of aromatic amino acids. More studies are needed to dete rmine why the activity of 2,6-(N)-sialyltransferase increases upon irradiation. SDS-PAGE analysis of commercially prepared recombinant rat 2,3-

PAGE 192

175 (N)-sialyltransferase or commercially prepared recombinant rat 2,6-(N)sialyltransferase did not yield observable ba nds with coomassie blue stain because the amount of protein in the solution was below the limit-of-detection. MS Analysis of In-Gel Digested Ir radiated N-Tag-ST/Inhibitor Samples Even though the activity profiles did not provide positive evidence for labeling, irradiated N-Tag-ST/inhibitor samples were analyzed with SDS-PAGE, in-gel digested with trypsin, and analyzed with LCQ-MS. Be fore analyzing the MS data, the theoretical mass shift associated with addition of th e CMP-o-azido mandelate was determined. Theoretically, there are at lease 3 different mass shifts possible based on the lability of the phosphate and cytosine bonds. Figure 5-8 represents the different mass shifts that may be observed with a derivatized tryptic peptide. During sample work up, there are three possible decomposition sites for the attach ed label. At positi on 1 (Figure 5-8), the cytosine may depyrimidinylate through acid catalyzed hydrolysis. At position 2 and 3 (Figure 5-8), hydrolysis may occur. Table 5-6 represents the theoretical +1 charge state of tryptic peptides with zero-missed cleavages and the theoretical a ddition of the four possible molecular weights described in Figur e 5-6. Missed cleavag e corresponds to the number lysines or arginines in a proteins se quence that were not cl eaved by trypsin. Ingel digestion of recombinant hST3GalIV w ith trypsin followed by MS analysis only provided peptides which did not have missed cleavages (zero-missed cleavages). Table 5-7 represents the same theoretical data as Table 5-6; however, this represents the +2 charge state series of theoretically labeled peptides. The bold amino acids in Tables 5-7 and 5-8 correspond to possible glycosylati on sites and the underlined amino acids correspond to sialylmotifs presented in Chapter 2.

PAGE 193

176 O OH HO N N O NH2O P O OH O O OHN R 1 2 3 Cleavage 2 = 244.118 (C8 H7 N1 O6 P1) Cleavage 1 = 360.233 (C13 H15 N1 O9 P1) Cleavage 3 = 165.146 (C8 H6 N1 O3) Total Molecule Molecular Weight = 471.335 (C13 H15 N1 O9 P1) O OH HO N N O NH2O P O OH O O OHN R 1 2 3 Cleavage 2 = 244.118 (C8 H7 N1 O6 P1) Cleavage 1 = 360.233 (C13 H15 N1 O9 P1) Cleavage 3 = 165.146 (C8 H6 N1 O3) Total Molecule Molecular Weight = 471.335 (C13 H15 N1 O9 P1) O OH HO N N O NH2O P O OH O O OHN R 1 2 3 O OH HO N N O NH2O P O OH O O OHN R 1 2 3 Cleavage 2 = 244.118 (C8 H7 N1 O6 P1) Cleavage 1 = 360.233 (C13 H15 N1 O9 P1) Cleavage 3 = 165.146 (C8 H6 N1 O3) Total Molecule Molecular Weight = 471.335 (C13 H15 N1 O9 P1) Figure 5-8. Possible Labile Bonds of CMP-oAzido Mandelate and Possible Mass Shifts Associated with Derivatized Tryptic Peptide Before manual interpretation of each LCQ-MS data set, Mascot analysis of the LCQ-MS data revealed peptides which were not labeled. The percent sequence coverage for the non-labeled N-Tag-ST was compar ed to the irradiated N-Tag-ST/98 M inhibitor and N-Tag-ST/180 M inhibitor samples (Table 5-3). Figures 5-9 and 5-10 represent the overlaid spectra for the N-Tag-ST sequence coverage presented in Chapter 4 and the sequence coverage for the LCQ-MS analysis of N-Tag-ST/98 or 180 M inhibitor data sets (Table 5-3) respectively. The red letter s correspond to tryptic peptides observed with both the non-labeled N-Tag-ST and irradiated N-Tag-ST/inhibitor samples. The blue letters correspond to peptides missing from LCQ-MS data from irradiated N-TagST/inhibitor samples. Typically, in-gel digestion of N-Tag-ST with trypsin provided peptides with zeromissed cleavages after analysis with Mascot ; therefore, the assumption was made that only zero-missed cleaved peptides need to be considered. Figure 5-9 illustrated that only two peptides were not observed with LCQMS. The m/z of the theoretical labeled

PAGE 194

177 Table 5-6. Mass-to-Charge of +1 Charge State of Theoretical Tryptic Peptides and Four Possible Mass Shifts Row # Pe p tide Se q uenceTheoretical m/z 471.335360.233244.118165.146 1 FVDSR623.315 1094.65983.548867.433788.461 2 GGSEK477.23 948.565837.463721.348642.376 3 EPCLQGEAESK1190.536 1661.871550.771434.651355.68 4 LFG NYS R856.431 1327.771216.661100.551021.58 5 DQPIFLR888.494 1359.831248.731132.611053.64 6 LEDYFWVK1099.546 1570.881459.781343.661264.69 7 TPSAYELPYGTK1326.658 1797.991686.891570.781491.8 8 GSEDLLLR902.494 1373.831262.731146.611067.64 9 VLAITSSSIPK1115.667 15871475.91359.791280.81 10 NIQSLR730.421 1201.761090.65974.539895.567 11 CVVVGNGHR 940.478 1411.811300.711184.61105.62 12 NSS LGDAINK 1018.516 1489.851378.751262.631183.66 13 YDVVIR 764.43 1235.771124.661008.55929.576 14 LNNAPVAGYEGDVGSK 1590.776 2062.111951.011834.891755.92 15 TT MR508.255 979.59868.488752.373673.401 16 LFYPESAHFDPK1450.7 1922.041810.931694.821615.85 17 VENNPDTLLVLVAFK1671.932 2143.272032.171916.051837.08 18 AMDFHWIETILSDK1705.825 2177.162066.061949.941870.97 19 QPPLIWDVNPK1306.715 1778.051666.951550.831471.86 20 ILNPFFMEIAADK1508.782 1980.121869.021752.91673.93 21 LLSLPMQQPR1182.666 16541542.91426.781347.81 22 QKPTTGLLAITLALHLCDL V HIAGF GYPDAYNK3553.887 4025.223914.123798.013719.03 23 QTIHYYEQITLK1536.806 2008.141897.041780.921701.95 24 SMAGSGH NVS QEALAIK1699.843 2171.182060.081943.961864.99 25 MLEMGAIK892.463 1363.81252.71136.581057.61 26 NLT SF581.293 1052.63941.526825.411746.439 FVDSRGGSHHHHHHGGEKKE PCLQGEAESKASK LFGNYSR DQPIFLRLEDYFWVKTPSAY ELPYGTKGSED LLLRVLAI TSSSIPKNIQSLR CRRCVVVGNGHRLR NSSLGDAINKYDV VIRLNNAPVAGYEGDVGSK TTMR LFYPESAHFDPK VENN PDTLLVLVAFK AMDFHWIETILSDK KRVRKGFWK QPPLI WDVNPK QIR ILNPFFMEIAADKLLSLPMQQPR KIKQKPTT GLLAITLALHLCDL VHIAGFGYPDAYNKK QTIHYYEQITL K SMAGSGHNVSQEALAIK R MLEMGAIK N LTSF Figure 5-9. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/180 M Inhibitor Experiment where Red Corres ponds to Tryptic Peptides Common to Both LCQ MS/MS Analysis, Black Corresponds to Tryptic Peptides not Observed with either Sample, and the Blue Corresponds to Tryptic Peptides not Observed from the N-Tag-ST:180 M Sample Analysis but Observed from the Standard N-Tag-ST Sample Analysis.

PAGE 195

178 Table 5-7. Mass-to-Charge of +2 Charge State of Theoretical Tryptic Peptides and Four Possible Mass Shifts Row # Peptide SequenceTheoretical m/z 471.335360.233244.118165.146 1 FVDSR623.315 547.829492.278434.221394.735 2 GGSEK477.23 474.787419.236361.178321.692 3 EPCLQGEAESK1190.536 831.44775.889717.831678.345 4 LFG NYS R856.431 664.387608.836550.779511.293 5 DQPIFLR888.494 680.419624.868566.81527.324 6 LEDYFWVK1099.546 785.945730.394672.336632.85 7 TPSAYELPYGTK1326.658 899.501843.95785.892746.406 8 GSEDLLLR902.494 687.419631.868573.81534.324 9 VLAITSSSIPK1115.667 794.005738.454680.397640.911 10 NIQSLR730.421 601.382545.831487.774448.288 11 CVVVGNGHR 940.478 706.411650.86592.802553.316 12 NSS LGDAINK 1018.516 745.43689.879631.821592.335 13 YDVVIR 764.43 618.387562.836504.778465.292 14 LNNAPVAGYEGDVGSK 1590.776 1031.56976.009917.951878.465 15 TT MR508.255 490.299434.748376.691337.205 16 LFYPESAHFDPK1450.7 961.522905.971847.913808.427 17 VENNPDTLLVLVAFK1671.932 1072.141016.59958.529919.043 18 AMDFHWIETILSDK1705.825 1089.081033.53975.476935.99 19 QPPLIWDVNPK1306.715 889.529833.978775.921736.435 20 ILNPFFMEIAADK1508.782 990.563935.012876.954837.468 21 LLSLPMQQPR1182.666 827.505771.954713.896674.41 22 QKPTTGLLAITLALHLCDL V HIAGF GYPDAYNK3553.887 2013.121957.561899.511860.02 23 QTIHYYEQITLK1536.806 1004.57949.024890.966851.48 24 SMAGSGH NVS QEALAIK1699.843 1086.091030.54972.485932.999 25 MLEMGAIK892.463 682.403626.852568.795529.309 26 NLT SF581.293 526.818471.267413.21373.724 FVDSRGGSHHHHHHGGEKKEPCLQGEAESKASK LFGNYSR DQPIFLRLEDYFWVKTPSAYE LPYGTKGSEDLLLRVLAI TSSSIPKNIQSLR CRRCVVVGNGHRLR NSSLGDAINKYDV VIRLNNAPVAGYEGDVGSK TTMR LFYPESAHFDPKVENN PDTLLVLVAFKAMDFHWIETILSDK KRVRKGFWK QPPLI WDVNPK QIR ILNPFFMEIAADKLLSLPMQQPR KIKQKPTT GLLAITLALHLCDLVHIAGFGYPDAYNKK QTIHYYEQITL K SMAGSGHNVSQEALAIK R MLEMGAI K N LTSF Figure 5-10. Overlaid Sequence Coverage of Standard N-Tag-ST and N-Tag-ST/98 M Inhibitor Experiment where Red Corres ponds to Tryptic Peptides Common to Both LCQ MS/MS Analysis, and Black Corresponds to Tryptic Peptides not Observed with either Sample.

PAGE 196

179 peptides for these two candidates were inve stigated manually; however, no matches were observed. Figure 5-10 illustrated that no pe ptides were missing between the non-labeled N-Tag-ST and the irradiated N-Tag-ST/98 M inhibitor. Conclusion CMP-o-azido mandelate may be consider ed a first generation site-directed photoaffinity label and there were several pos sible reasons why this molecule did not react with sialyltransferase. First, CMP-oazido mandelate may not effectively label the enzyme because the nitrene species was s hort lived, thus creating a dehydroazepine intermediate (7 member ring with the nitr ogen). If this was the case, then only nucleophilic amino acids would be labeled, and if labeled, may be labile. Although there was no direct evidence for the formation of the dehydroazepine, the generation of this molecular species should not be excluded. To improve the design, a perfluorinated aryl azide may be beneficial for pr olonging the life time of the nitrene intermediate. Second, if there was water in the activ e-site, the nitrene or dehydroaze pine would react with this molecule first. Third, the azide may not be w ithin reach of amino acids in the active site pocket. Improvements in molecule may involve incr easing the distance of the aryl azide from the phosphate bond, synthesizing CMP-pazido mandelate, or synthesizing CMP-mazido mandelate. The direction for improvement in molecule design was not clear at this time and a library of many different aryl azid e donor substrate mimics may need to be synthesized and irradiated in the presence of sialyltransferase. Currently, a method for synthesis of these types of molecules is in hand.

PAGE 197

180 Methods and Materials Synthesis and Characterization of CMP-o-azido Mandelate Following synthesis and purification by Dr. Nicole Horenstein, CMP-o-azido mandelate was desalted with Amberlite resin in the H+ form. After desalting, the sample was concentrated to dryness followed by th e recovery of 4.0 mg in 1.5 mL of water yielding a final concentration of 5.35 mM. After re-suspens ion in water, the pH was slightly acidic. The pH was adjusted to ne utrality with a minimal amount of ammonium hydroxide. To determine th e spectral properties, 5 L of sample was diluted to 200 L (135 M) for UV analysis. UV analysis provid ed absorption bands at 254 nm (1.51 Abs) and 280 nm (1.57 Abs). The molar absorb tivity was calculated to be 11,400 cm-1 M-1 at 280 nm and 11,800 cm-1 M-1 at 254 nm. A Bruker APEX II ESI FTICR MS (Billerica, MA) operated in negative ion mode was us ed to determine purity and the photolysis kinetics of CMP-o-azido mandelate. Photoaffinity Labeling Experiments Incubation of N-Tag-ST wi th and without inhibitor For the standard activity curve, 90 L of N-Tag-ST (250 g/mL; 7.2 M) in the purification buffer described in Chapter 2, was diluted with 10 L of a buffer containing 20 % glycerol, 0.01 % triton CF 54, and 50 mM MES, pH 6.8 ([ST]final = 6.5 M). To initiate inhibition, 10 L of 1.8 mM inhibitor in a buffe r containing 20 % glycerol, 0.01 % triton CF 54, and 50 mM MES, pH 6.8 was added to 90 L of N-Tag-ST ([ST]final =6.5 M). Both samples were incubated at 37oC with 10 L aliquots assayed at 5, 55, and 120 minutes.

PAGE 198

181 Photoactivation of inhibitor in the presence of N-Tag-ST For the standard activity value, 90 L of N-Tag-ST (250 g/mL; 7.2 M) in the purification buffer reported in Chapter 2, was diluted with 10 L of a buffer containing 20 % glycerol, 0.01 % triton CF 54, 50 mM MES pH 6.8 ([ST]final =6.5 M). Stock solutions of inhibitor were prepared at 500 M, 980 M and 1.8 mM in a buffer containing 20 % glycerol, 0.01 % triton CF 54, and 50 mM MES, pH 6.8. Prior to irradiation, 10 L of the particular stock solution was added to 90 L of N-Tag-ST ([ST]final =6.5 M) and allowed to pre-equilibrate for 30 minutes prior to irradiation, except the 180 M (2) sample. The 180 M (2) sample was allowed to pre-equilibrate for 90 mins prior to irradiation. The sample s were kept on ice during all steps of the experiment. A 254 nm Hamamatsu Corpor ation (Middlesex, NJ) 0.65 amp lamp was used for continuous UV irradiation for 6 to 12 minutes at 4.5 cm from the top of an open 500 L microfuge tubes. At 1, 2, 3, 4, 5, 6, or 12 minute intervals, a 10 L aliquot was assayed under the assay conditions described later in this se ction. All activities were presented as the percent activity of the orig inal standard without inhibitor present and prior to irradiation. Photoactivation of inhibitor in th e presence of commercially available sialyltransferase For the standard activity value, 5 L of recombinant rat 2,3-(N)-sialyltransferase from Calbiochem (San Diego, CA) in a bu ffer containing 75 mM NaCl, 75 mM MES, 2.5 mM MgCl2, 500 M -mercaptoethanol, and 50 % glycer ol, pH 6.0 was diluted with 40 L of a buffer containing 50 mM MES, 0.1 mg/ml BSA, and 0.25% Triton X-100, pH 7.4 and 5 L of water. The stock solution for inhib itor was prepared to 1.33 mM in water.

PAGE 199

182 Prior to irradiation, 5 L the inhibitor stock solution ([133 M]final) was added to 5 L of recombinant rat 2,3-(N)-sialyltransferase in a buffer containing 75 mM NaCl, 75 mM MES, 2.5 mM MgCl2, 500 M -mercaptoethanol, and 50 % glycerol, pH 6.0 followed by dilution with 40 L of a buffer containing 50 mM MES, 0.1 mg/ml BSA, and 0.25% Triton X-100, pH 7.4. After a 30 minute pre-equilibration period, a 254 nm Hamamatsu Corporation (Middlesex, NJ) 0.65 amp lamp, a 254 nm short wave UltraViolet Products Inc. (Upland, CA) 0.16 amp lamp, or a 365 nm long wave Ultra-Violet Products Inc. (Upland, CA) 0.16 amp lamp was used for continuous UV irradiation for 5 minutes at 4.5 cm from the top of open 500 L microfuge tubes. At time 0 and 5 minutes, a 10 L aliquot was assayed under the conditions described later in this chapter. All activities were presented as the percent activity of th e original standard without inhibitor present and prior to irradiation. For the standard activity value, 5 L of recombinant rat 2,6-(N)-sialyltransferase from Calbiochem (San Diego, CA) in a bu ffer containing 250 mM NaCl, 50 mM MES, 2.5 mM MgCl2,500 M -mercaptoethanol, and 50 % glycerol, pH 6.2 was diluted with 40 L of 50 mM MES, 0.1% Tween 20, pH 7.4 and 5 L water. The stock solution for inhibitor was prepared to 1.33 mM in water. Prior to irradiation, 5 L the inhibitor stock solution [133 M]final was added to 5 L of recombinant rat 2,6-(N)-sialyltransferase from Calbiochem (San Diego, CA) in a bu ffer containing 250 mM NaCl, 50 mM MES, 2.5 mM MgCl2, 500 M -mercaptoethanol, and 50 % glycer ol, pH 6.2 was diluted with 40 L of a buffer containing 50 mM MES and 0.1% Tween 20, pH 7.4 and 5 L water. After equilibation for 30 minutes, a 254 nm Hamamatsu Corporation (Middlesex, NJ) 0.65 amp lamp, a 254 nm short wave Ultra-Vi olet Products Inc. (Upland, CA) 0.16 amp

PAGE 200

183 lamp, or 365 nm long wave Ultra-Violet Pr oducts Inc. (Upland, CA) 0.16 amp lamp was used for continuous UV irradiation for 5 mi ns at 4.5 cm from the top of open 500 L microfuge tubes. At time 0 and 5 minutes, a 10 L aliquot was assayed under the assay conditions described later in this chapter. All activities were pres ented as the percent activity of the origin al standard without inhibitor pr esent and prior to irradiation. Activity assays for recombin ant sialyltransferase All activity assays reported on recombinant sialyltransfer ase were performed in the following way: a 10 L aliquot from a selected sa mple was incubated with a 10 L mixture of 100-170 M [9 3H] CMP-NeuAc (10,000 20,000 counts-per-minute; specific activity = 10 20 Ci/ mol) and 235 mM -lactose (rat recombinant 2,3 sialyltransferase a nd N-Tag-ST) or 75 M N-acetyl-lactosamine (rat recombinant 2,6 sialyltransferase) for the a ppropriate amount of time to limit the consumption of CMPNeuAc to less than 10%. The reaction mi xture was quenched with 500 mL of 5 mM inorganic phosphate buffer, pH 6.8 and th en applied to Dowex 1 X 8, 200 mesh (PO4 2-) mini-columns, pH 6.8. Reactions were eluted with 3.5 mL of 5 mM Pi buffer pH 6.8 into liquid scintillation vials. Th e counts-per-minute from liquid scintillation were used to calculate relative percent activities based on de scribed standards. The definition of a unit of activity is the amount CMP-NeuAc convert ed to sialyl-lactose per minute. In-Gel Digestion with Trypsin Recombinant sialyltransferase samples we re analyzed with SDS-PAGE, followed by in-gel digestion with Trypsin Gold (Prome ga). The protocol described in Chapter 4 for in-gel digestion with trypsin was u tilized for N-Tag-ST/inhibitor samples.

PAGE 201

184 MS Analysis LCQ MS was performed on a Thermo Electron LCQ Deca (San Jose, CA), operated in positive ion mode with a spray volta ge of 1.8 kV +/0.2 applied at the liquid junction before a capillary column. An Eldex MicroPro pump delivered solvent at 6 L/min to a LC Packings C18 PepMap Nano-Prepcolumn (0.3 x 1 mm). Five microliters of sample was loaded onto the nano-prepcol umn and washed for 5 minutes at 0 % mobile phase B. After washing, the direction of flow was reversed with a new flow rate of 200 nL/min after pre-column splitting from 8 L/min. Elution of peptides from a New Objective (0.75mm x 5 mm) Proteopep capilla ry column included a gradient of 3-60% mobile phase B over 30 minutes and 60-90% mobile phase B over 1 minute with mobile phase A as 0.1% acetic acid, 3% acetonitrile and mobile phase B as 0.1% acetic acid, 95% acetonitrile.

PAGE 202

185 CHAPTER 6 MS ANALYSIS OF DNA HARVESTED WITH NANOHARVESTING AGENTS Introduction to Oligonucleotide Analysis with Mass Spectrometry One of the overall goals for proteomics is to identify protei ns which are involved with disease states. Also, as shown in Chapters 2 through 5, studying the threedimensional structure of a partic ular class of enzymes, i.e. si alyltransferase, aids in the rational design for inhibitors related to that particular class of en zymes. In a totally different mindset, human disease may also be studied at the genomic level by analyzing oligonucleotides with MS. As of 2000, at le ast 50 tumor types may be defined at the genomic level. For example, different type s of breast and colon cancers, Huntingtons disease, ataxia telangiectasia, cystic frib rosis, and myotonic dystrophy were identified through analysis of oligonucleot ides. Often, the cancer states are due to mutations in the tumor suppressor p53 codon.180 At the genetic level, there is a high fr equency (1 in 800 62,000 base pairs) of single nucleotide polymorphisms (SNPs). Id entification of the SNPs is useful for studying simple and complex tra its related to genotype-pheno types. Furthermore, SNP identification is useful for associat ion reconstruction of human evolution.181 Also, oligonucleotide analysis finds applicati on in microbial and forensic studies.182 Lastly, MS may be used as a tool for quality cont rol of synthetic nucleotides and polymerase chain reaction (PCR) products. The advent of matrix assisted lase r desorption (MALDI) and electrospray ionization (ESI) allowed for larg e oligonucleotides to be anal yzed with MS in a similar

PAGE 203

186 manner to intact proteins and proteolytic peptid es. MS analysis of oligonucleotides is not considered to be main stream because several other methods have been established and widely accepted. Capillary electrop horesis (CE), high performance liquid chromatography (HPLC), polyacrylamide gel electrophoresis (PAGE), UV spectrophotometry, and thin layer chromatogra phy are all proven techniques for analysis of oligonucleotides.180 With these methods, detection involves the use of fluorescent labels or radioactive 32P. To obtain sequence information, Sanger sequencing is analyzed with PAGE. The advent of 96-capillary electrophoresis has improved throughput dramatically and was the method used to map the human and other genomes. The main advantage MS has over Sanger sequencing is the ability to provide information on base methylation or th e frequency of mutated DNA based on mass differences. Furthermore, oligonucleotides ma y be fragmented in a similar fashion to proteins or peptides. As with polypeptides the length, charge state, and the selected collisional energy will provi de different fragmentation patterns of oligonucleotides.181 Also, in comparison to the other method pr esented, MS has the capability of being accurate and fast for verification of genome sequencing. For verification and diagnosis, isolated re gions of oligonucleotides analyzed with MS is more sensitive and fa ster than PAGE analysis.180;181 The accuracy of oligonucleotide identification depends largely on the accuracy provided by the instrument and the size of the oligonucleot ide. As with positive mode ESI of proteins and peptide, negative mode ESI of oligonucleotides provides multiply-charged ions that fall within the typical mass range observed with modern MS.183 Because the cost of automated

PAGE 204

187 synthesis of oligonucleotides has been reduced dramatically, high-thr ough put analysis is a requirement for modern biological studies, i.e. microarray technologys.184 For example, four 96-well plates for P CR analysis may generate 386 PCR products in parallel, thus creating highthrough put needs. Currentl y, ESI FTICR-MS was cited to have analyzed over 300 samples per day with flow injection, and 1700 samples per day with automation.185 Realistically, with MS analysis the genotyping of polymorphism is possible for femtomolar amounts of PCR produc ts of less than 150 base pairs (bp). Currently, a 500 bp (310,000 Da) limit with ESI ha s been reported from the enzymatic digests of DNA analyzed with ESI-FTICR-MS.182;183 There are several difficult issues with MS analysis of oligonucleotides. First, oligosaccharides are analyzed in negative i on mode instead of positive ion mode. Often, in negative ion mode, arching will occur betw een the ESI tip and the orifice if the solvent is greater than 50 % aquous. Electrons may collect at the ESI tip, thus creating a corona discharge that will destr oy the electrospray Taylor c one. Secondly, the negatively charged phosphate backbone may form str ong charge interactions with sodium, potassium, and manganese. PCR reactions contain DNA polymerase as well excess primers, dNTPs, and manganese which must be removed before ESI-MS is performed.185 Sample clean-up has been the main bo ttleneck for high-purity oligonucleotide sample analysis. Several different off-line techniques are used to isolate oligonucleotides, including: immobilization by streptavidin/biotin, phenol/chloroform extractions, size-exclusion filters, silica re sin, cation-exchange resin, anion-exchange HPLC, purification pipette-tips, spin-columns, mico-dialysis, gel-filtration, and ethanol precipitation.185 As illustrated in Chapter 1, salts will inhibit electrospray or lower the

PAGE 205

188 sensitivity due to salt adduct charge-state distributions.184;186 Furthermore, the chargestate distribution has been shown to be dependent on base pair composition and the solvent system. Third, high efficiency separa tion of oligonucleotides normally utilizes strong-cation exchange chromatography (SCE). The salt buffers used with SCE are not conducive for ESI-MS analysis. Separation on C18 material is possible; however, all the salt may not be removed from the oligosaccharides and substantial peak broadening may occur, thus lowering sensitivity. To provide on-line separation with MS, volatile ionparing agents must be used with a re verse phase stationary phase (IP-RP).187 In this application, chromatography w ill be used for pre-concentration of oligosaccharides after selec tion with nanoharvesting agents (NHAs). Instead of using chromatography to separate and desalt complex oligosnucleotides samples, nanoharvesting agents were designed to select a particular oligonucleotides of interest. Normally, oligonucleotides sele cted with NHAs are detected with fluorescent probes. Compared to using fluorescent probes, ESI-FTICR-MS analysis of nanoharvested oligonucleotides facilitates ex act mass identification. Nanoharvesting Agents In collaboration with Josh Smith from the Tan Group, oligonucleotides may be analyzed with MS after selection NHAs. The focu s of this chapter is the MS analysis of a synthetic 5 mer and five synthetic 19 me rs on the way to analyzing NHA selected oligonucleotides. NHAs were used to select and isolate trace amounts of oligonucleotides from complex samples. NHAs are small nanoparticles which have a magnetic material at the core a surface silica coating which may be modified, an avidin cross linking layer, and a bound streptavidin derivatized oligonuc leotide of varying length acting as a target r ecognition element.

PAGE 206

189 Selection of target oligonucleotides wa s performed by first mixing the particles with complex mixture containing the targ et DNA. Secondly, after equilibration, a magnetic field was applied to collect the NHAs Third, supernatant was removed and the magnetic particles were washed extensively. Fourth, the NHAs were heated to denature the bound target DNA. To finish the purification of the target DNA, the NHAs were again extracted by intr oduction of a magnetic field, and once again the supernatant was removed (containing the purified DNA). Fl uorescent studies performed by the Tan Group have shown the capture and releas e of target oligonucleotides. All NHA preparation and selection of DNA were performed by the Tan Group. The particles for this experiment were designed to select for 19 mers of complimentary, single miss match, and doubl e miss match sequences. One of the questions that fluorescence de tection cant answer is the absolute sequence of the oligonucleotide harvested with the nanoparticle. Five oligon ucleotides were designed to test the binding affinity of 1 and 2 miss-matched base pairs in comparison to the complementary 19 mer target (Table 6-1). Th e overall goal was to develop a protocol for the on-line analysis of oligonuc leotides selected with NAHs to take advantage of the exact mass capabilities of negative io n mode ESI-FTICR-MS. The different oligonucleotides were designed to have di fferent molecular weights for intact mass analysis, thus the requirement for MS/MS were not needed for initial studies. The red highlighted base pairs correspond to miss-matche s base pairs in reference to the trapping DNA found on the NHA. Finally, compared to the complementary 19 mer sequence, the modified oligonucleotides will have a lower dissociation temperature. The 5 mer is designed as an internal standard.

PAGE 207

190 Table 6-1. Synthetic Oligonuc leotides Prepared for NAH Se lection and On-line HPLCMS Analysis Name Synthetic Oligonucle otide Theoretical m/z Complimentary GCG ACC ATA GCG ATT TAG A 5834.015 5 modification GC C ACC ATA GCG ATT TAG A 5796.848 mid-modification GCG ACC ATA T CG ATT TAG A 5809.009 3 modification GCG ACC ATA GCG ATT G AG A 5859.002 5x2 modification G G G AC T ATA GCG ATT TAG A 5889.021 5 mer standard ATC GA 1486.302 On-line HPLC MS of Oligonucleotides As stated before, HPLC-MS of oligonuc leotides was the main bottleneck to overcome because of the difficulty with sepa rating negatively charge oligonucleotides. Typically, oligonucleotides are first purified with strong ca tion-exchange resin (ECE), and secondly purified off-line with ethanol pr ecipitation, solid-phase extraction, affinity purification, or IP-RP-HPLC.182 To facilitate on-line HPLC-MS analysis of oligonucleotides, the best results were obtained w ith octadecylated poly(styrene/divinylbenzene) particels187 or monolithic capillary columns that were prepared by copolymerization of styren e and divinylbenzene in a 200 m i.d. fused silica with ionpairing agents in the mobile phase. The monolithic columns have been reported to be more efficient than packed-columns because the mass transfer was faster.186 For example, using typical C18 material columns for the analysis of oligonucleotides provided decreased resolution and reten tion time because of low ma ss transfer. Furthermore, simple un-derivatized Poly(styrene-divi nyl benzene) was not able to resolve oligonucleotides. Lastly, highly resolved peak s required the heating of the columns to 50 oC.188 To increase retention times and desalt the oligonucleotide on-line with MS, triethanolamine as ion pairing agent is know to ion-pair with the phosphate backbone,

PAGE 208

191 thus improving resolution and out-com peting sodium and potassium binding.184 Including triethanolamine, several other ionpairing agents have been investigated in mobile phases for separation on the column s presented. These include, 25 mM butyldimethylamine bicarbonate pH 8.5, 25 mM butyldimethylamine hydrocarbonate, 25 mM aqueous triethylammonium hydroge n carbonate, 0.5 to 1.0 % triethylamine adjusted to pH 8.5, or 1.5 % spermidine. 181;182;184;189 Typically, oligonucleotides eluted at 5 to 15 % organic. Direct infusion analysis of off-line purified oligonucleotides in cluded the use of the buffers described above, or the addition of imidizole, ammonium acetate, and/or piperidine to acetonitrile and/or 2-propanol. As stated before, at low percent organic, there was a high likelihood that arching will o ccur during the ESI process. To reduce the possibility of arching, a post-column makeup flow was normally added to enhance ionization.189 Post-column solvents normally cons ists of 100 % organic content and improved signal intensity with the follo wing series: isopropanol < methanol < acetonitrile.187 Once the complete solvent system is worked out, there are several instrumentation parameters that must be optimized. MS Instrumentation Considerations There are three important instrumental considerations for MS analysis of oligonucleotide analysis in the negative i on mode with a Bruker Apex II ESI-FT-ICR MS. First, the potential applied to the ESI tip or the capillary entrance must be lower for higher charge states.189 Compared to proteins and pe ptides, the potential applied for oligonucleotide analysis was usually 1000 V less with opposite polarity. Second, a higher potential difference betw een the first skimmer and the heated capillary in the source was reported to produce si ngly charged oligonucleotides;189 therefore, the

PAGE 209

192 potential difference will be lower between these two elements because higher chargestates are desired. Lastly, it was reported th at there was a dramatic change between the charge state distributions of 20 mers and 80 mers based on the voltages applied on the transfer optics. The instrument was tuned da ily with a 19 mer standard to ensure that optimal signal was observed prio r to IP-RP-HPLC-ESI-FTICR-MS. To sequence oligonucleotides, there are several techniques available, each producing similar fragmentation patters. The techniques include nozzle-skimmer dissociation (NS), collisiona lly activation dissociation (C AD), and infrared multiphoton dissociation (IRMPD). The complete seque nce coverage for 42 to 108-mer DNAs were first reported by McClafferty in 1996 with the three different fr agmentation methods.180 Furthermore, MS/MS/MS fragmentation may al so provide unique structural information that is not observable with the othe r analysis methods presented earlier. As with all aspects of MS analysis of oligonucleotides, automated bioinformatic analysis of MS/MS spectra has lagged behind. Currently, three of the main bioinformatics programs for computer-aided data interpretation of MS and MS/MS spectra include a MS comparative sequence program from Stanford University (http://insertion.stanf ord.edu/sequencing)182 is suitable for oli gonucleotides less than 30 nucleotides; the comparative sequencing of long-chain oli gonucleotides (COMPAS) from Saarland University (http ://www.uni-saarland.de/fak8/huber/co mpas.htm) is useful for sequencing up to 80 nucleotides; and the simp le oligonucleotide sequencer (SOS) from Rega Institute for Medical Resear ch, Katholieke Universities Leuven (http://rna.rega.kuleuven.ac.be/sos/)181 is useful for sequencing oligonucleotides less than 50 nucleotides.

PAGE 210

193 Experiment Design The overall goal of this project was to load the crude solution containing NHAs, after selection of target DNA, into tubing at tached to an inject or (with an applied magnetic field), wash impurities away from th e matrix solution, apply a heat gradient to denature the target-NHA DNA complex, coll ect DNA on a short column, and finally analyze the DNA with on-line IP-RP-HPLC-MS. First, solvent systems for make-up flow and chromatography were investigate d. The resolving power of FT-ICR-MS was characterized for the synthetic 5 and 19 me rs synthesized. Second, TEA or TEAA was investigated as an ion-pairing agents with a PS-DVB monolithic column for synthetic 5 and 19 mers. Third, solutions of NHA selected 19 mers were injected onto the IP-RPHPLC-ESI-FTICR-MS isystem. Finally, a solution of target-NHA/DNA complex was loaded into tubing attached to an inject or for complete on-line analysis. The next goal of this project was to provi de answers to the se lectivity of NHAs for short oligonucleotides (19mers) with si ngle and double mismatched base pairs (in comparison to the complementary oligonucleotid e). Mixtures of equimolar solutions of complementary 19 mer and one of the mismatched 19 mers were mixed with NHAs. Since the mismatched 19 mer will dissociated at a lower temperature, an added degree of selectivity was provided. Also, to provide relative quantificat ion and serve as an internal standard, the 5 mer was evaluated as a com ponent in the post-column make up flow. Finally, our Bruker Apex II may be operated in NS dissociation, CAD, and IRPD for obtaining sequence information; however, this wa s not a goal of this pa rticular project. Results and Discussion The following section presents preliminary data on solvent system optimization for directly infused oligonucleotides synt hesized and purified in the Tan Group.

PAGE 211

194 Furthermore, the mass accuracy and signal intensity of each synthetic 19 mer was evaluated. To improve mass accuracy and provide possible quan tification of on-line separation, a 5 mer was evaluated as an additi ve for post column make-up flow. Lastly, initial studies on the separation of oligonuc leotides are presente d. NHA collection and IR-RP-HPLC MS analysis of synthetic 19 mers has not yet provided retention of oligonucleotide; therefore, the fi nal goal of complete on-line analysis has yet to be met. All samples were analyzed on a Bruker Apex II 4.7 T ESI FT-ICR MS operated in negative ion mode with SideKickTM trapping. Typically, the instrument had to be tuned daily for maximum signal intensities, with th e capillary held at 2470 V. After external calibration, all the mass errors were lower than 10 ppm (0.0005 Da). Solvent System Studies Direct infusion (0.5 to 2 L /min) analysis was performed with several different buffer systems. Table 6-2 presents the fina l solvent systems whic h provided the highest signal intensities. Typical spectra observe d for the different synthetic 19mers are presented in Figures 6-1 and 6-2. Because the purity of the synthetic nucleotides was not constant from batch-to-batch, spectra with consistent charge stat es were not observed between the different 19 mers. The 3modifi ed 19 mer was determined to be of the highest quality (least amount of sodium adducts observed); ther efore, it was used as the model oligonucleotide for this study. In the different spectra, note the difference between the signal intensities and the change in charge state between the different buffer systems. For buffer 1, ammonium acetate was present, whereas for buffer 2 ammonium acetate was not. The data illustrated th at ammonium acetate suppressed the higher charge-states and signal intensities. Figure 6-3 represents the si gnal intensity of the

PAGE 212

195 Table 6-2. Solvent Systems for Direct Infusion of Synthetic Oligonucleotides Buffer System 1 Buffer System 2 Buffer System 3 Buffer System 4 60 % acetonitrile 60 % acetonitrile 60 % acetonitrile 60 % acetonitrile 20 % isopropanol 20 % isopropanol 20 % isopropanol 20 % isopropanol 20 mM piperidine 20 mM imidazole 20 mM piperidine 20 mM piperidine 5 mM ammonium acetate 20 mM piperidine 5 mM ammonium acetate 25 mM triethylamine 20 mM imidazole 20 mM im idazole 20 mM imidazole 25 mM triethylamine 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 m/z 0.0e+00 2.0e+06 4.0e+06 6.0e+06 8.0e+06 1.0e+0710 pmol/uL Buffer 1[M -3H]-3[M -4H]-4[M -6H]-6 2004006008001000120014001600180020002200m/z 0.0e+00 2.0e+06 4.0e+06 6.0e+06 8.0e+06 1.0e+0710 pmol/uL Buffer 1[M -3H]-3[M -4H]-4[M -6H]-6 Figure 6-1. Spectrum of 10 pmol/ L of 3Modified 19 mer Diluted in Buffer 1 different charge-states versus the different bu ffer systems presented in Table 6-2. Buffer system 4 offers the highest intensity for the high-charge states. Ba sed this data, buffer 2 was used as a make-up flow so lvent for post-column addition. In a similar manner, the intensities of the 5 mer were investigated at 1 and 10 pmol/ L diluted in buffers 1 and 2. Note that Figure 6-4 follows the same trend observed in Figure 6-3 for the ammonium acetate buffer. To determine the linear dynamic range of

PAGE 213

196 200 400 600 800 1000 1200 1400 1600 1800 2 0.0e+00 2.0e+06 4.0e+06 6.0e+06 8.0e+06 1.0e+07 1.2e+0710 pmol/uL Buffer 2[M -7H]-7[M -4H]-4[M -6H]-6[M -5H]-5 20040060080010001200140016001800 2 0.0e+00 2.0e+06 4.0e+06 6.0e+06 8.0e+06 1.0e+07 1.2e+0710 pmol/uL Buffer 2[M -7H]-7[M -4H]-4[M -6H]-6[M -5H]-5 Figure 6-2. Spectrum of 10 pmol/ L of 3Modified 19 mer diluted in Buffer 2 0.00E+00 2.00E+06 4.00E+06 6.00E+06 8.00E+06 1.00E+07 1.20E+07 10 pmol/uL in Buffer 1 10 pmol/uL in Buffer 2 10 pmol/uL in Buffer 3 10 pmol/uL in Buffer 4 Charge-state[ M 8 H ]8[ M 7 H ]7[ M 6 H ]6[ M 5 H ]5[ M 4 H ]4[ M + 1 N a 5 H ]4[ M + 1 N a 4 H ]3[ M 3 H ]3[ M + 2 N a 5 H ]3Intensity 0.00E+00 2.00E+06 4.00E+06 6.00E+06 8.00E+06 1.00E+07 1.20E+07 10 pmol/uL in Buffer 1 10 pmol/uL in Buffer 2 10 pmol/uL in Buffer 3 10 pmol/uL in Buffer 4 Charge-state[ M 8 H ]8[ M 7 H ]7[ M 6 H ]6[ M 5 H ]5[ M 4 H ]4[ M + 1 N a 5 H ]4[ M + 1 N a 4 H ]3[ M 3 H ]3[ M + 2 N a 5 H ]3Intensity Figure 6-3. Charge State Distributio n and Signal Intensity of 10 pmol/ L of 3Modifed 19 mer Diluted with the Different Buffer Systems

PAGE 214

197 the method, 5mer was diluted to 1, 10, 50, and 100 pmol/ L with buffer 2 prior to direct infusion at 2 L/min (Figure 6-5). The 742.1148 m/z [M 2H]-2 was selected because it had the highest signal intensity. The LOD for the system was determined to be 500 fmol/ L with direct infusion at 0.5 L / min. 0.00E+00 2.00E+06 4.00E+06 6.00E+06 8.00E+06 1.00E+07 1.20E+07 1.40E+07 1 pmol/uL in Buffer 1 1 pmol/uL in Buffer 2 10 pmol/uL in Buffer 1 10 pmol/uL in Buffer 2[ M 3 H ]3[ M 2 H ]2[ M + N a 3 H ]2[ M + 2 N a 4 H ]2[ M 1 H ]1[ M + N a 2 H ]1[ M + 2 N a 3 H ]1[ M + 3 N a 4 H ] 1 Charge-stateIntensity 0.00E+00 2.00E+06 4.00E+06 6.00E+06 8.00E+06 1.00E+07 1.20E+07 1.40E+07 1 pmol/uL in Buffer 1 1 pmol/uL in Buffer 2 10 pmol/uL in Buffer 1 10 pmol/uL in Buffer 2[ M 3 H ]3[ M 2 H ]2[ M + N a 3 H ]2[ M + 2 N a 4 H ]2[ M 1 H ]1[ M + N a 2 H ]1[ M + 2 N a 3 H ]1[ M + 3 N a 4 H ] 1 Charge-stateIntensity Figure 6-4 Charge State Di stribution and Signal Inte nsity of 1 or 10 pmol/ L of 5 mer Diluted in Buffer System 1 or 2 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 0102030405060708090100110Concentration of 5 mer in uMIntensit y 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 02.557.51012.51517.520 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 0102030405060708090100110Concentration of 5 mer in uMIntensit y 0.00E+00 5.00E+06 1.00E+07 1.50E+07 2.00E+07 2.50E+07 02.557.51012.51517.520 Figure 6-5. Linear Dynamic Range of the -2 Charge State of 5 mer Diluted with Buffer 2

PAGE 215

198 Unfortunately, there was not a linear response to an increased concentration of the 5 mer. Furthermore, when the 5 mer and 19 mer were diluted to10 pmol/ L in the same solution, the 19 mer suppressed ionization of the 5 mer. The 5 mer can not act as an internal standard because of the non-linear calibrati on curve and ion-suppre ssion in the presence of the 19 mer. IP-RP-HPLC ESI-FTICR-MS Analysis of Synthetic Oligonucleotides A PS-DVB monolithic column (200 m x 50 mm) was obtained from LC Packings (Sunnyvale, CA). TEA or TEAA was added to mobile phase A (100 % water) and mobile phase B (30 % acetonitrile) in 25 and 50 mM concentrations. Figure 6-6 represents the typical elution profile for a 2 L injection of 20 picomoles of 5 mer obtained with the current column and either buffer system. Analysis of synthetic 19 mers also provided the same elution profiles. Acco rding to the data, the column did not retain the synthetic oligonucleotides and that they elut e at the solvent front with a flow rate of 2.5 L/min 336.8 703.7 1179.4 1475.4 1735.0 2382.7 3100.9 3557.7 4267.4 4787.7 +MS, 2.5min #22 0.0 0.2 0.4 0.6 0.8 5 x10 Intens. 500 1000 1500 2000 2500 3000 3500 4000 4500 m/z 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 Time [min] 9.6 9.7 9.8 9.9 7 x10 Intens. [M 1H]-1 Elution at dead volume 336.8 703.7 1179.4 1475.4 1735.0 2382.7 3100.9 3557.7 4267.4 4787.7 +MS, 2.5min #22 0.0 0.2 0.4 0.6 0.8 5 x10 Intens. 500 1000 1500 2000 2500 3000 3500 4000 4500 m/z 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 Time [min] 9.6 9.7 9.8 9.9 7 x10 Intens. [M 1H]-1 Elution at dead volume Figure 6-6. IP-RP HPLC ESI FTICR MS Anal ysis of 15 picomoles of 5mer with a PSDVB Monolithic Column and 25 mM TEAA as an ion-pairing agent.

PAGE 216

199 With monolithic columns, there was the po ssibility that differe nt batches may not provide the same efficiency. Furthermore, ot her column materials should be investigated with our system. Other options include a packed octadecyla ted poly-(styrene / divinylbenzene) column (200 m id x 50 mm), a packed C18 resin columns, or an inhouse copolymerization of styren e and divinylbenzene in 200m i.d. fused silica. Lastly, the temperature of the column needs to be heated to at least 50 oC for efficient separation. Conclusion The NHAs have demonstrated the ability to separate oligonucleot ides that differ by a single base pair and are of the same le ngth (19 mer). Solvent systems have been optimized for maximum signal intensity and high charge-state character. Results on different concentrations of oligonucleotide versus signal intensity produced a non-linear response. Unfortunately, retention of oligonucleotides on the PS DVB monolithic column was problematic. To solve this issu e, an in-house packed octadecylated poly(styrene/divinylbenzene) column (200 um id x 50 mm), purchase of a Phenomenex Monolithic Onyx C18 (CHO-7646), or purchase of Waters XTerra MS C18 column (186002471) is recommended for proper ion-pair reversed phase HPLC MS-ESI-FTICRMS analysi. Once chromatography is functi onal, release of oligonucleotide from NHAs will be performed in-line via a modified inje ctor port. Also, experiments should proceed which facilitate fragmentation for sequencing a nd the analysis of larg er oligonucleotides. The continuation of this project will be beneficial to trace analysis of selected oligonucleotides for verificati on of sequence and selectivity.

PAGE 217

200 Methods and Materials Acetic acid (Ca # A507-500), formic acid (Ca # BP 1215-500), methanol (Ca # 230-4), 2-propanol (Ca # A414500), acetonitrile (Ca # A998-1), triethanolamine (Ca # 04885-1) and high-purity water (Ca # AH 365-4 or W 7-4) were purchased from FisherScientific (Fair Lawn, NJ). Piperidine (Ca # 104094), ammonium acetate (Ca # 431311), and imidazole (Ca # I5513) were purchased fr om Sigma-Aldrich (St. Louis, MO). All synthetic oligonucleotides were provided by the Tan Group in lyophilized form and diluted with a minimal amount of water to provide a final concentration between 50 and 500 pmol/ L. NHA synthesis and selection expe riments were performed in the Tan Group. Samples were diluted in one of the f our buffers presented earlier to the working concentration described in the result s and discussion section. FT-ICR/MS Instrumentation Stable electrospray was achieved with Aglients (Waldbronn, Germany), off-axis ESI spray source equipped with nebulizi ng gas (Ca # G2427A). A Bruker Daltonics (Billerica, MA) APEX II 4.7 T FT-ICR MS was used to colle ct data in broadband mode with SideKickTM trapping. For direct infusion acqui sition, 1 M files were collected with signal averaging of 4 spectra and a 1 to 3 second hexapole accumulation time. Samples were directly infused with a Harvard Apparatus (Holliston, MA) PHD 200 Infusion (Ca # 70-2000) at 0.5 to 2 L/min. For HPLC-MS acquisition, 32 k files were collected with signal averaging of 4 spectra and a 1 to 3 second hexapole accumulation time. The signal averaging of 4 spectra and 3 second hexapole accumulation time provide d a scan rate of 14.7 seconds per scan. (3.69 seconds per one acquisition multiplied by 4 scans). Calibration was provided Agile nt Technologies (Palo Alto, CA) HP Tuning Mix (Ca #

PAGE 218

201 G2421A) with the four point (601.9784, 1033.9875, 1633.9491, and 2233.9107 m/z) nonlinear algorithm provided by the Bruker Xmass software. HPLC Instrumentation An Aligent Technologies (Palo Alto, CA) HP 1100 Series (Ca # G1312A) was used to deliver solvent at 2.5 to 3.0 L/min. to a LC Packings (Sunnyvale, CA) monolithic PS-DVB (200 m x 5mm) column. Two separate so lvent systems were used. Mobile Phase A consited of 25 mM triethylamine (TEA) at pH 11.5 or 25 mM triethylamine acetate (TEAA) at pH 8.5 and mobile phase B consisted of 25 mM TEA at pH 11.5 or 25 mM TEAA at pH 8.5 in 30% acet onitrile. The gradient was held at 0 % mobile phase for 5 minutes, then ramped to 95% mobile phase B over 5 minutes. The load loop was made from 75 m i.d. fused silica, with a final volume of 2.0 L. A Harvard Apparatus (Holliston, MA (PHD 200 Infusion (Ca # 70-20 00) syringe pump delivered solvent at 2mL/min to a t-union for post column additi on. The make up flow buffer consited of either acetonitrile, Buffer 2 (Table 61), or Buffer 2 spiked with 5pmol/ L 5 mer.

PAGE 219

202 CHAPTER 7 CONCLUSIONS AND FUTURE WORK This dissertation presents the investigation of peptides, proteins, and oligonucleotides with common biochemi cal, novel bioconjugation/selection, and common analytical chemistry techniques. As presented in Chapter 1, systems biology means to look at all classes of bimolecules as an integrated system. Investigation of the interactions between the genome, transcript ome, proteome, and metabolome, require a basic understanding of the tools n eeded for analysis of each particular system. Presented in this dissertation are separation and ma ss spectrometry (MS) techniques needed for proteomic analysis in regards to the proteo me and detection of DNA in regards to the genome. Normally, MS is associated with functi onal proteomics anal ysis; however, this dissertation described the use of MS for stru ctural proteomic analysis of recombinant sialyltransferase and analysis of nanoparticle selected oligonu cleotides. Before using MS to solve structural proteomics, several bioc hemical experiments were performed. First, sialyltransferase was expressed, purified, and characterized. Second, three separate bioconjugation techniques were investigated for derivitation of si alyltransferase for studying secondary structure. The major bottleneck of this work was finding a bioconjugation technique that e fficiently derivatized sialyltr ansferase in an active form while creating a stable bond pr ior to HPLC-MS analysis. MS analysis of DNA afforded exact mass identification of known 19 base pair oligonucleotides; however, on-line

PAGE 220

203 HPLC-MS proved to be the major bottle neck for analysis of nano particles selected DNA. The Sialyltransferase Project The overall goal of this project was to id entify the catalytically important amino acids of human placental -(2 3) sialyltransferase (hST3Gal IV). To identify amino acids, MS based structural proteomic techniqu es were utilized. The evaluation of amino acids needed for catalysis proceeded with three separate bioconjugation techniques. These techniques included hydrogen/deuterium exchange high-performance liquid chromatography electrospray ionization F ourier transform ion cyclotron MS (H/Dx HPLC ESI FT-ICR MS), derivatization with small molecule labeling agents followed by matrix-assisted laser desorption ionizati on time-of-flight MS (MALDI-TOF MS) or liquid chromatography quadruple ion trap MS (LCQ-MS) analysis, and modification with a synthetic site-directed photoaffinity la beling agent followed by MALDI-TOF MS or LCQ-MS analysis. Before derivati zation with bioconjugation techniques, sialyltransferase had to be e xpressed and purified. During the course of this study, three sepa rate forms of recombinant hST3Gal IV were expressed in Spodoptera frugiperda ( Sf 9) insect cells, pur ified with affinity chromatography, and characterized. The firs t form included the catalytic domain of hST3Gal IV with a cleavable canine insulin se cretion peptide cloned at the end of the Nterminal region (Ins-ST). The second form included a His6x fusion tag cloned between the canine insulin secretion pe ptide and the catalytic domain of hST3Gal IV (N-Tag-ST). The third form included a His6x fusion tag cloned at the C-terminal end of the Ins-ST form (C-Tag-ST). All three forms had sim ilar expression yields (1 2.5 U/L) after Sf 9

PAGE 221

204 insect cells were infected with recombinant baculovirus. The Ins-ST construct was purified with an in-house synthesized se pharose cytidine diphosphate-hexanolamine affinity column (sCDP-Hex), whereas the NTag-ST and C-Tag-ST were purified with Ni+2-NTA affinity resin purchased commercially The final purification values for each construct were: Ins-ST (1 60 mU, 20 % yield, 280 ug, 0.132 U/ mg), N-Tag-ST (155 mU, 27 % yield, 160 ug, 0.216 U/mg) and C-TagST (240 mU, 43 % yield, 1.5 mg, 0.162 U/mg). These were somewhat similar a nd comparable to literature values. The N-Tag-ST and C-Tag-ST constructs we re much easier to purify than Ins-ST because the sCDP-Hex resin was mu ch easier to compress than the Ni+2-NTA resin. When comparing the specific activities betw een the three purified constructs, N-Tag-ST purified with Ni+2-NTA affinity resin provided the highest purity. When comparing yields (based on activity), the C-Tag-ST provided about 15 to 20 % higher yield. Although the final yield for the C-Tag-ST was higher than the other two constructs, the addition of the His6x fusion tag to the C-terminal e nd of recombinant hST3 Gal IV changed the observed kinetic parameters in comparison to the N-Tag-ST and Ins-ST constructs. Lastly, all three forms of recombinant hST3Gal IV exhibited multiple glycoforms upon SDS-PAGE analysis. To confirm that the multiple bands were glycoforms, recombinant hST3Gal IV was digested with Endo Hf and PNGase F. Upon SDS-PAGE analysis of the digested construc t, the multiple bands collapsed to one band. Kinetic parameters were investigated in collaboration with Dr. Erin Burke. All three recombinant hST3Gal IV constructs had similar Km values (Ins-ST: 171 + 18 mM; N-Tag-ST: 155 + 14 mM; C-Tag-ST: 158 + 11mM) for -lactose. The Ins-ST and NTag-ST constructs had similar Km values for CMP-NeuAc (Ins-ST: 82 + 5 M; N-Tag-

PAGE 222

205 ST 75 + 8 M), whereas the C-Tag-ST had a much large Km value (267 + 20 M). Normally, the N-terminal end of the cataly tic domain is extended through a stem region to an anchor which ligates the protein to the Golgi apparatus; therefore, changes in the in the N-terminal region of the catalytic doma in of ST3Gal IV would not necessarily perturbate structure. Since the C-Tag-ST Km was larger for the donor substrate, it was proposed that the polyhistid ine tag may slightly change the secondary structure. The major obstacle encountered during the preparation of Ins-ST, N-Tag-ST, and C-Tag-ST was enzyme stability during sample desalting after collection and concentration of purified fractions. Dialysis and dilution/concentration with different membranes and ultrafiltration units destroyed 90 % of the initial activity. To perform bioconjugation experiments, the c onstructs were used in the final purified form without desalting the solution. H/Dx HPLC ESI FT-ICR MS Prior to analysis of recombinant hS T3Gal IV with H/Dx ESI-FTICR-MS, our Bruker Apex II 4.7 T ESI-FTICR-MS was optimized with commercially prepared intact proteins and pepsin digests of commercially prepared proteins. Di rect infusion or HPLC ESI-FTICR-MS were used to provide sp ectra on 20 to 100 picomoles of bovine ribonuclease A, (~13.7 kDa), hen egg wh ite lysozyme (~14.3 kDa), bovine chymotrypsinogen A (~25.6 kDa), human euka ryotic carbonic anhydrase I (28.7 kDa), bovine carbonic anhydrase II (~ 28.9 kDa), and bovine serum albumin (~66.4 kDa) with mass error of less than 100 ppm. Unfortuna tely, direct infusion and HPLC ESI-FT-ICR MS analysis of in-house prepared Ins-ST N-Tag-ST, C-Tag-ST, and commercially prepared rat (2 6)-(N)-sialyltransferase did not resulted in positive identification.

PAGE 223

206 Recombinant hST3Gal IV constructs could not be exchanged into ESI compatible solvent systems. Furthermore, analysis of in tact recombinant hST3Gal IV from crude purification samples with HPLC ESI-FTICR-MS on a C4 resin destroyed column efficiency and provided detergent carryover p eaks upon subsequent analysis. Also, since the expressed recombinant hST3Gal IV pr esented several glycoforms, the overall concentration of sialyltransferase was reduced to that of the different glycoforms. Furthermore, literature on the ESI-FTICR-MS analysis of intact membrane bound glycoproteins is sparse because of the difficulty associated with ionizing normally membrane bound proteins and proteins which contain glycosylation. To reduce sample heterogeneity, deglycosylati on was performed with Endo Hf and PNGaseF. Direct infusion or HPLC ESI-FTICR-MS analysis of deglycosylated constr ucts did not provide positive identification. Furthermore, Endo Hf digestion of intact non-denatured recombinant hST3Gal IV did not f acilitate retention of activity. HPLC-ESI-FTICR-MS analysis of 5 M bovine carbonic anhydrase digested with 3 M of pepsin resulted in 45 % coverage of 15 picomoles. Digestion of 5 M Ins-ST or 5 M commercially prepared rat (2 6)-(N)-sialyltransferase with 3 M of pepsin prior to HPLC-ESI-FTICR-MS did not result in positive identification of peptides. Signal may not have been observed because the glycans inhibited digestion or sialyltransferase may need to be denatured prior to pepsin dige stion. Since it was a requirement to maintain the activity and tertiary structure during H/Dx HPLC-MS analysis of intact recombinant hST3Ga l IV, H/Dx HPLC-MS was abandoned as an analytical method to describe 3-dimenstional structure.

PAGE 224

207 Since structural proteomic analysis of sial yltransferase was the overall goal of this project, attention was shifted to derivatization with small molecule labeling agents and site-directed photoaffinity labeling agen ts. Although this case specific study on sialyltransferase did not result in positive data, our Bruker Apex II ESI-FT-ICR MS has been optimized for analysis of large intact proteins. Already, samples submitted to the Mass Spectrometry Core Facility have result ed in positive identification of modified intact proteins. Small Molecule Labeling Agents Based on the background presented in Ch apter 2, the target amino acids for derivatization with small molecule labeling ag ents included a genera l acid or a general base. Iodoacetic acid (IOA), 1-ethyl -(3-dimethylaminopropopal) carbodiimide hydrochloride (EDAC) coupled with ethanol amine (ETAM), and N-acetylimidazole (NAI) were initially screened for deactiva tion of Ins-ST, N-Tag-ST, and C-Tag-ST. Under the conditions describe d in Chapter 4, deactivation proceeded with NAI. Next, several experiments were performed to achieve differential labeling with NAI. Solutions containing recombinant hST3Gal IV; recombinant hST3Gal IV and CMP-NeuAc; recombinant hST3Gal IV and -lactose; and recombinant hST3Gal IV, CMP-NeuAc, and -lactose were incubated with NAI. The solu tion that contained sialyltransferase, CMPNeuAc, and -lactose retained 100 % of the initia l activity over a 6 minute incubation, whereas the solutions containing recomb inant hST3Gal IV and CMP-NeuAc or -lactose resulted in decreased activ ity, 90% and 50 % respectively. Furthermore, activity was regenerated when ethanolamine was added to NAI deactivated recombinant hST3Gal IV.

PAGE 225

208 Based on this data, the identity of one of the amino acids in the active-site of recombinant hST3Gal IV was either a tyrosine or histidine. Furthermore, the data suggested that there was a change in c onformation of the enzyme during catalysis because both donor substrate (CMP-NeuAc) and acceptor substrate ( -lactose) were needed for protection of the ac tive-site from NAI deactivation. In future work, it would be useful to investigate other labeling agents, EDC/ETAM, or IOA under different reaction conditions. Unfortunately, MS analysis of NAI deri vatized recombinant hST3Gal IV did not yield the identity of the particular amino aci d differentially protected. This was likely due to labile bond formation. Finally, in-gel di gestion with trypsin prior to MS analysis was reproducible between different batches of recombinant hST3Gal IV; therefore, the reproducibility of digestion was not in questi on. Digestion with tr ypsin, Asp N, Glu C, Lys C, or -chymotrypsinogen A resulted in 72% per cent sequence coverage. In keeping with the overall goal of identifying amino aci ds needed for catalysis with structural proteomic based MS, the next strategy util ized site-directed photoaffinity labeling. Site Directed Photoaffinity Labeling The first generation of CMP-NeuAc mimic substrates containi ng an aryl azide, CMP-o-azido mandelate, was synthesized by Dr. Nicole Horenstein. Unfortunately, irradiation of CMP-o-azido mandelate in the presence of recombinant hST3Gal IV, commercially prepared rat recombinant 2,6-(N)-sialyltransferase, or commercially prepared recombinant rat 2,3-(N)-sialyltransferase did not provide deactivation. CMPo-azido mandelate may not effectively label the enzymes because the nitrene may be short lived, thus creating a dehydroazepine intermediate (7 member ring with the

PAGE 226

209 nitrogen). If this was the case, only nucl eophilic amino acids woul d be labeled, and if labeled, may be labile. To improve molecula r design, a perfluorinated aryl azide may be beneficial for prolonging the life time of the nitrene intermediate. Furthermore, if water was in the active-site, the nitrene or dehydroazep ine would react with th is molecule first. Spatially, within the active-site of sialyl transferase, the azide may not be within reach of the catalytic amino acid. Improvements in the molecule may involve increasing the distance of the aryl azide from the phos phate bond or synthesizing the para or meta forms of the inhibitor. The direction for improvement was not clear at this time and a library of many different aryl azides may need to be synthesized and irradiated in the presence of sialyltransferase. Currentl y, a method for synthesis of these types of molecules is in hand and may be used to ge nerate a library of compounds for MS based structural proteomics on sialyltransferase. MS Analysis of Oligonucleotides Selected with Nanoharvesting Agents With fluorescent studies, NHAs have dem onstrated the ability to separate oligonucleotides that differ by a single base pair and were the same length as reported by collaborators in the Tan group. With MS an alysis of oligonucleotides, solvent systems have been optimized for maximum signal inte nsity and high charge-s tate character by direct infusion ESI-FTICR-MS analysis of 5 and 19 mers. Furthermore, results on the concentration of oligonucleotid e versus signal intensity reve aled a non-linear response by direct infusion ESI-FTICR-MS. Unfortunatel y, retention of oli gonucleotides on a PSDVB monolithic column on-line with ESI-FTI CR-MS was problematic. Oligonucleotide retention time corresponded to elution at the de ad volume time. To solve this issue, an in-house packed octadecylated poly-(styrene/divinylbenzene) column (200 um id x 50 mm), purchase of Phenomenexs Monolithic Onyx C18 (CHO 7646), or purchase of

PAGE 227

210 Waters XTerra MS C18 column (186002471) is recommended for proper ion-pair reversed phase HPLC MS. Once chro matography is functional, release of oligonucleotide from NHAs will be performed in-line via a modified injector port. During future work, experiments should pr oceed which facilitate fragmentation for sequencing and the analysis of larger oligonu cleotides. The continuation of this project will be beneficial to trace an alysis of selected oligonucleotides for verification of the DNA sequence selected and selectivity of the NHAs. Final Conclusions To conclude, while it is clear that MS is a powerful bioanalytical technique, it is quite clear that every bioconjugation system studied presented unique challenges to be over come on a case by case basis. First, H/Dx HPLC-ESI-FTICR-MS analysis of sialyltransferase was not possible because the analyte could not be transferred into the gas phase for MS analysis. This was larg ely due to the stability of recombinant sialyltransferase and the expression of mu ltiple glycoforms. Second, small molecule labeling with NAI and a differen tial labeling scheme revealed th at the identity of one of the amino acids in the active site was either a tyrosine or histidine; however, MS analysis could not identify the deriva tized amino acid. This was mainly due to labile bond formation. Lastly, CMP-o-azido mandelate irradiat ed in the presence of recombinant sialyltransferase did not provi de deactivation. This may be due to the life-time of the nitrene or the need for an increased dist ance between the aryl azide and the phosphate bond or synthesizing the para or meta form s of the inhibitor. Based on all the experiments presented in this dissertation, site -directed photoaffinity labeling is the best route for describing active-site amino acids. A synthetic route is in hand for synthesizing

PAGE 228

211 these types of CMP-NeuAc mimic substrat es. Once our model, hST3Gal IV, is deactivated and the covale ntly labeled amino acid is identified by MS, other sialyltransferase family members may be irradi ated in the presence of the aryl azide to evaluate secondary structure homology.

PAGE 229

212 LIST OF REFERENCES 1. Yates, J. R. Journal of Mass Spectrometry 1998, 33 1-19. 2. Simpson, R. J., Ed.; Introduction to Proteomics Cold Springs Harbor Press, Cold Springs Harbor, NY 2005 pgs 1 -37. 3. Tyers, M.; Mann, M. Nature 2003, 422 193-97. 4. Wu, Y.; Engen, J. R. Analyst 2004, 129 290-96. 5. Liu, H. L.; Hsu, J. P. Proteomics 2005, 5 2056-68. 6. Lane, C. S. Cellular and Molecular Life Science 2005, 62 848-69. 7. Blueggel, M.; Chamrad, D.; Meyer, H. E. Current Pharmaceutical Biotechnology 2004, 5 79-88. 8. Ashcroft, A. E. Nat.Prod.Rep. 2003, 20 202-15. 9. Zimmermann, J. G.; Brown, L. R. Mass Spectrometry Reviews 2001, 20 1-57. 10. Mosely, M. A. Trens in Biotechnology 2001, 19 S10-S16. 11. Peng, J.; Eilas, J. E.; Thoreen, C. C.; Licklider, L. J.; Gygi, S. P. Journal of Proteome Research 2003, 2 43-50. 12. Washburn, M. P.; Wolters, D.; Yates, J. R., III Nature Biotechnology 2001, 19 242-47. 13. Zenzmaier, C.; Gesslbauer, B.; Grobusche k, N.; Jandrositz, A.; Preisegger, K. H.; Kungl, A. J. Biochemical and Biophysical Research Communications 2005, 328 968-72. 14. Shen, Y.; Zhang, R.; Moore, R. J.; Kin, J.; Metz, T. O.; Hixson, K. K.; Zhao, R.; Livesay, E. A.; Udseth, H. R.; Smith, R. D. Analytical Chemistry 2005, 77 3090100. 15. Sali, A.; Glaser, R.; Earnest, T.; Baumeister, W. Nature 2003, 422 216-25. 16. Yamashita, M.; Fenn, J. B. Journal of Physical Chemistry 1984, 88 4451-59.

PAGE 230

213 17. Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Communications in Mass Spectrometry 1988, 2 151-53. 18. Aebersold, R.; Mann, M. Nature 2003, 422 198-207. 19. Anderson, J. S.; Svensson, B.; Roepstroff, P. Nature Biotechnology 1996, 14 44957. 20. Paizs, B.; Suhai, S. Mass Spectrometry Reviews 2005, 24 508-48. 21. Bogdanov, B.; Smith, R. D. Mass Spectrometry Reviews 2005, 24 168-200. 22. Novak, P.; Kruppa, G. H.; Young, M. M.; Schoeniger, J. Journal of Mass Spectrometry 2004, 39 322-28. 23. Sze, S. K.; Ge, Y.; Oh, H. B.; McLafferty, F. W. PNAS 2002, 99 1774-79. 24. Barret, J.; Brophy, P. M.; Hamilton, J. V. International Journal for Parasitology 2005, 35 543-53. 25. Lanne, B.; Panfilov, O. Journal of Proteome Research 2005, 175-79. 26. Reinders, J.; Lewandrowski, U.; Moebius, J.; Wagner, Y.; Sickmann, A. Proteomics 2004, 4 3686-704. 27. Shi, Y.; Xiang, R.; Horvath, C.; Wilkins, J. A. Journal of Chromatography A 2004, 1053 27-36. 28. Wilm, M. S.; Mann, M. International Journal of Mass Spectrometry and Ion Processes 1994, 136 167-80. 29. Kennedy, R. T.; Jorgenson, J. W. Analytical Chemistry 1989, 61 1128-35. 30. Ishihama, Y.; Rappsilber, J. ; Anders, J. S.; Mann, M. Journal of Chromatography A 2002, 979 233-39. 31. Ishihama, Y. Journal of Chromatography A 2005, 1067 73-83. 32. Le, J. C.; Hui, J.; Katta, V.; Rohde, M. F. Journal of the Amer ican Society for Mass Spectrometry 1997, 8 703-12. 33. Liu, S.; Griffiths, W. J.; Sjovall, J. Analytical Chemistry 2003, 75 791-97. 34. Wolters, D. A.; Washburn, M. P.; Yates, J. R., III Analytical Chemistry 2001, 73 5683-90. 35. Davis, M. T.; Lee, T. D. Journal of the American Soc iety for Mass Spectrometry 1997, 8 1059-69.

PAGE 231

214 36. Vissers, J. P. C.; Blackburn, R. K.; Moseley, M. A. Journal of the American Society for Mass Spectrometry 2002, 13 760-71. 37. Pan, P.; McLuckey, S. A. Analytical Chemistry 2003, 75 1491-99. 38. Schmidt, A.; Karas, M.; Dulcks, T. Journal of the American Society for Mass Spectrometry 2003, 14 492-500. 39. Hendrickson, C. L.; Emmett, M. R. Annu.Rev.Phys.Chem 1999, 50 517-36. 40. Cech, N. B.; Enke, C. G. Mass Spectrometry Reviews 2001, 20 362-87. 41. Loo, J. A.; Loo, R. O. Electrospray Ionization Mass Spectrometry Cole, R. B., Ed.; John Wiley and Sons, Incorporated, NY 1997 Chapter 11. 42. Henery, K. D.; Williams, E. R; Wang, B. H; McLafferty, F. W.; Shabanowitz, J; Hunt, D. F. Proceedings of the National Academy of Sciences USA 1989, 86 907578. 43. Garcia, M. C.; Hogenboom, A. C.; Zappey, H.; Irth, H. Journal of Chromatography A 2002, 957 187-99. 44. Knochenmuss, R.; Zenobi, R. Chemical Reviews 2003, 103 441-52. 45. Dreisewerd, K. Chemical Reviews 2003, 103 395-425. 46. Karas, M.; Kruger, R. Chemical Reviews 2003, 103 427-39. 47. Siuzdak, G.; Bothner, B.; Yeager, M.; Brugidou, C.; Fauguet, C. M.; Hoey, K.; Chang, C. M. Chemistry and Biology 1996, 3 45-48. 48. McLuckey, S. A.; Wells, J. M. Chemical Reviews 2001, 101 571-606. 49. Watson, J.T., Ed.; Introduction to Mass Spectrometry, 3rd ed.; Lippincott-Raven Publishers: Philadelphia, PA 1997 50. Kim, Y. J.; Freas, A.; Fenselau, C. Analytical Chemistry 2001, 73 1544-48. 51. Luxembourg, S. L.; Mize, T. H.; McDonnell, L. A.; Heeren, R. M. A. Analytical Chemistry 2004, 76 5339-44. 52. Mills, K.; Johnson, A. W.; Diettrich, O. ; Clayton, P. T.; Winchester, B. G. Tetrahedron: Asymmetry 2000, 11 75-93. 53. Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrometry Reviews 1998, 17 1-35. 54. Amster, I. J. Journal of Mass Spectrometry 1996, 31 1325-37.

PAGE 232

215 55. Mirgorodskaya, E.; Roepstorff, P.; Zubarev, R. A. Analytical Chemistry 1999, 71 4431-36. 56. Hofstadler, S. A.; Sannes-Lowe ry, K. A.; Griffey, R. H. Analytical Chemistry 1999, 71 2067-70. 57. Belov, M. E.; Nikolaev, E. N.; Anderson, G. A.; Udseth, H. R.; Conrads, T. P.; Veenstra, T. D.; Masselon, C. D. ; Gorshkov, M. V.; Smith, R. D. Analytical Chemistry 2001, 73 253-61. 58. Bossio, R. E.; Marshall, A. G. Analytical Chemistry 2002, 74 1674-79. 59. Luscombe, N. M.; Greenbaum, D.; Gerstein, M. Methods of Information in Medicine 2001, 4 346-58. 60. Johnson, R. S; Davis, M. T.; Taylor, A.; Patterson, S. D. Methods 2005, 35 223-36. 61. Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proceedings of the National Academy of Sciences 2004, 101 9528-33. 62. Engen, J. K.; McCormack, A. L.; Yates III, I. J. R. Journal of the American Society for Mass Spectrometry 1994, 5 976-89. 63. Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Electrophoresis 1999, 20 3551-67. 64. Khersonsky, S. M.; Ho, C. M.; Garcia, M .F.; Chang Y.T. Current Topics in Medicinal Chemistry 2003, 3 617-43. 65. Lairson, L. L.; Withers, S. G. Chemical Communication 2004, 2243-48. 66. Dwek, R. A. Chemical Reviews 1996, 96 683-720. 67. Unligil, U. M.; Rini, J. M. Current Opinion in Structural Biology 2000, 10 510-17. 68. Dube, D. H.; Bertozzi, C. R. Nature 2005, 4 477-89. 69. Angata, T.; Varki, A. Chemical Reviews 2002, 102 239-469. 70. Aebi, M.; Hennet, T. Trends in Cell Biology 2001, 11 136-41. 71. Coutinho, P. M.; Deleury, E.; Davies, G. J.; Henrissat, B. Journal of Molecular Biology 2003, 328 307-17. 72. Harduin-Leppers, A.; Vallejo-Ruiz, V.; Krzewinski-Recchi, M. A.; Samyn-Petit, B.; Julien, S.; Delannoy, P. Biochimie 2001, 727-37. 73. Kiefel, M. J.; von Itzstein, M. Chemical Reviews 2002, 102 471-90.

PAGE 233

216 74. Pilatte, Y.; Bignon, J.; Lambre, C. R. Glycobiology 1993, 3 201-17. 75. Dall'Olio, F.; Chiricolo, M. Glycoconjugate Journal 2001, 18 841-50. 76. Hennet, T.; Chui, D.; Paulson, J. C.; Marth, J. D. Proceedings of the National Academy of Sciences of th e United States of America 1998, 4504-09. 77. Harduin-Lepers, A.; Recchi, M. A.; Delannoy, P. Glycobiology 1995, 5 741-58. 78. Datta, A. K.; Paulson, J. C. The Journal of Biological Chemistry 1995, 270 1497500. 79. Jeanneau, C.; Chazalet, V.; Auges, C.; So umpasis, D. M.; Harduin-Leppers, A; Delannoy, P.; Imberty, A.; Breton, C. The Journal of Biological Chemistry 2004, 279 13461-68. 80. Koles, K.; Irvine, K. D.; Panin, I. M. The Journal of Biological Chemistry 2004, 279 4346-57. 81. Harduin-Lepers, A.; Mollicone, R.; Delannoy, P.; Oriol, R. Glycobiology 2005, 15 805-17. 82. Alviano, C. S.; Travassos, L. R.; Schauer, R. Glycoconjugate Journal 1999, 16 545-54. 83. Marchal, I.; Jarvis, D. L.; Cacan, R.; Verbet, A. Biological Chemistry 2001, 382 151-59. 84. Jackson, R. J.; Hall, D. F.; Kerr, P. J. Journal of Virology 1999, 73 2376-84. 85. Paterlini-Brechot, P.; Saigo, K.; Muraka mi, Y.; Chami, M.; Gozuacik, D.; Mugnier, C.; Lagorce, D.; Brechot, C. Oncogene 2003, 22 3911-16. 86. Gilbert, M.; Cunningham, A.; Watson, D. C.; Martin, D. C.; Richards, J. C.; Wakarchuk, W. W. Europe Journal of Biochemistry 1997, 249 187-94. 87. Yamamoto, T.; Nakashizuka, M.; Terada, I. The Japanese Biochemical Society 1998, 123 94-100. 88. Chiu, C. P. C.; Watts, A. P.; Lairson, L. L.; Gilbert, M.; Lim, D.; Wakarchuk, W. W.; Withers, S. G.; Strynadka, N. C. J. Nature Structural and Molecular Biology 2004, 11 163-70. 89. Sasaki, K.; Watanabe, E.; Kawashima, K. ; Sekine, S.; Dohi, T.; Oshima, M.; Haai; Nishi, T.; Hasegawa, M. The Journal of Biological Chemistry 1993, 268 22782-87. 90. Bianco, A.; Brufani, M.; Ciabatti R.; Melchioni, C.; Pasquali, V. Molecules Online 1998, 2 129-36.

PAGE 234

217 91. Grob, H. J.; Brossmer, R. Glycoconjugate Journal 1995, 12 739-46. 92. Wen, D. X.; Livingston, B. D.; Medzihradszky, K. F.; Kelm, S.; Burlingame, A. L.; Paulson, J. C. The Journal of Biological Chemistry 1992, 267 21011-19. 93. Datta, A. K.; Sinha, A.; Paulson, J. C. The Journal of Biological Chemistry 1998, 273 9608-14. 94. Kawaguchi, S. K.; Kabata, S.; Arita, M. The Journal of Biological Chemistry 2001, 276 15696-703. 95. Datta, A. K.; Chammas, R.; Paulson, J. C. Journal of Biological Chemistry 2001, 276 15200-07. 96. Bruner, M.; Horenstein, B. A. Biochemistry 1998, 37 289-97. 97. Bruner, M.; Horenstein, B. A. Biochemistry 2000, 39 2261-68. 98. Horenstein, B. A.; Bruner, M. Journal of the American Chemical Society 1996, 118 10371-79. 99. Horenstein, B. A. Journal of the American Chemical Society 1997, 119 1101-07. 100. Horenstein, B. A.; Bruner, M. Journal of the American Chemical Society 1998, 120 1357-62. 101. Borman, S. Chemical and Engineering News 2004, 82 10. 102. Wang, X.; Niu, Y.; Cao, X.; Zhang, L.; Zhang, L.H.; Ye, X. S. Bioorganic & Medicinal Chemistry 2003, 11 4217-24. 103. Skropeta, D.; Schworer, R.; Haag, T.; Schmidt, R. R. Glycoconjugate Journal 2004, 21 205-19. 104. Wang, X.; Zhang, L. H.; Ye, X. S. Medicinal Research and Reviews 2002, 23 3247. 105. Sun, H. B.; Yang, J. S. ; Amaral, K. E.; Horenstein, B. A. Tetrahedron Letters 2001, 42 2451-53. 106. Hamamoto, T.; Young-Choon, L.; Kurosawa, N.; Nakaoka, T.; Kojima, N.; Tsuji, S. Bioorganic & Medicinal Chemistry 1994, 2, 79-84. 107. Borsig, L.; Ivanov, S. X.; Herrman, G. F.; Kragl, U.; Wandrey, C.; Berger, E. G. Biochemical and Biophysical Research Communications 1995, 210 14-20. 108. Borsig, L.; Berger, E. G.; Malissard, M. Biochemical and Biophysical Research Communications 1997, 240 586-89.

PAGE 235

218 109. Malisard, M.; Zeng, S.; Berger, E. G. Biochemical and Biophysical Research Communications 2000, 267 169-73. 110. Kitagawa, H.; Paulson, J. C. Journal of Biological Chemistry 1994, 269 1394-401. 111. Kim, H. G.; Yang, S. M.; Lee, Y. C.; Do, S. I.; Chung, I. S.; Yang, J. M. Biochemical and Biophysical Research Communications 2003, 305 488-93. 112. Seo, N. S.; Hollister, J. R.; Jarvis, D. L. Protein Expression and Purification 2001, 22 234-41. 113. Ivannikova, T.; Bintein, F.; Malleron, A.; Juliant, S.; Cerutti, M.; Harduin-Leppers, A.; Delannoy, P.; ge, C.; bineau, A. Biochemistry 2003, 338 1153-61. 114. Williams, M. A.; Hiroshi, K.; Datta, A.; Paulson, J. C.; Jamieson, J. C Glycoconjugate Journal 1995, 12 755-61. 115. Colley, K. J.; Lee, E. U.; Adler, B.; Browne, J. K.; Paulson, J. C. Journal of Biological Chemistry 1989, 264 17619-22. 116. Gillespie, W.; Kelm, S.; Paulson, J. C. Journal of Biological Chemistry 1992, 267 21004-10. 117. Ford, C. F.; Suominen, I.; Glatz, C. E. Protein Expression and Purification 1991, 95-107. 118. Nilsson, J.; Stahl, S.; Lundeber g, J.; Uhlen, M.; Nygren, P. A. Protein Expression and Purification 1997, 11 1-16. 119. Joziasse, D. H.; Bergh, M. L. E.; ter Hart Hennie, G. J.; Koppen, P. L.; Hooghwinkel, G. J. M.; Van Den Eijnden, D. H. Journal of Biological Chemistry 1985, 260 4941-61. 120. Bac-to-Bac Baculovirus Expression System. www.invitrogen.com. 6-15-2004. 121. Sadler, J. E.; Rearick, J. I.; Paulson, J. C.; Hill, R. L. Journal of Biological Chemistry 1979, 254 4434-43. 122. Weinstein, J.; Souza-e-Silvia, U.; Paulson, J. C. Journal of Biological Chemistry 1982, 257 13845-53. 123. Nemansky, M.; Van Den Eijnden, D. H. Glycoconjugate Journal 1993, 10 99-108. 124. Barker, R.; Olsen, K. W.; Sharper, J. H.; Hill, R. L. Journal of Biological Chemistry 1972, 25 7135-47. 125. Aisaka, K.; Igarashi, A.; Yamaguchi, K.; Uwajima, T. Biochemical Journal 1991, 276 514-46.

PAGE 236

219 126. Aisaka, K.T.; Uwajima, T. Applied and Environmental Microbiology 1986, 51 562-65. 127. Engen, J. R.; Smith, D. L. Analytical Chemistry 2001, 73 256A-65A. 128. Busenlehner, L. S.; Armstrong, R. N. Archives of Biocemistry and Biophysics 2005, 433 34-46. 129. Engen, J. R.; Smithgall, T. E.; Gmeiner, W. H.; Smith, D. L. Journal of Molecular Biology 1999, 287 645-56. 130. King, D.; Lumpkin, M.; Bergmann, C.; Orlando, R. Rapid Communications in Mass Spectrometry 2002, 16 1569-74. 131. Zhang, Z.; Post, C. B.; Smith, D. L. Biochemistry 1996, 35 779-91. 132. Chung, E. W.; Nettleton, E. J.; Morgan, C. E.; Gross, M.; Mirankers, A.; Radford, S. E.; Dobson, C. M.; Robinson, C. V. Protein Science 1997, 6 1316-24. 133. Lam, T. T.; Lanman, J. K.; Emmett, M. R.; Hendrickson, C. L.; Marshall, A. G.; Prevelige, P. E. Journal of Chromatography A 2002, 982 85-95. 134. Englander, J. J.; Mar, C. D.; Li, W.; Engl ander, S. W.; Kim, J. S.; Stranz, D. D.; Hamuro, Y.; Woods Jr., V. L. Proceedings of the National Academy of Sciences of the United States of America 2003, 100 7057-62. 135. Kaltashov, I. A.; Eyles, S. J. Mass Spectrometry Reviews 2002, 21 37-71. 136. Jarrold, M. F. Annual Reviews in Physical Chemistry 2000, 51 179-207. 137. Wang, F.; Blanchard, J. S.; Tang, X. Biochemistry 1997, 36 3755-59. 138. Wang, F.; Scapn, G.; Blanchard, J. S.; Angeletti, R. H. Protein Science 1998, 7 293-99. 139. Wang, F.; Miles, R. W.; Kicska, G.; Nieves E.; Schramm, V. L.; Angeletti, R. H. Protein Science 2000, 9 1660-68. 140. Gorshkov, M. V.; Masselon, C. D.; Anderson, G. A.; Udseth, H. R.; Harkewicz, R.; Smith, R. D. Journal of the American Soc iety for Mass Spectrometry 2001, 12 1169-73. 141. Katta, V.; Chait, B. T Journal of the American Chemical Society 1993, 115 631721. 142. Leitner A.; Linder W. Journal of Chromatography B 2004, 813 1-26. 143. Silberring J.; Nyberg, F. Journal of Chromatography 1991, 562 459-67.

PAGE 237

220 144. Izumi, S.; Kaneko, H.; Yamazaki, T.; Toshifumi, H.; Kominami, S. Biochemistry 2003, 42 14663-69. 145. Wong, S. S.; Wong, L. J. C. Enzyme and Microbial Technology 1992, 14 866-74. 146. Pashkova, A.; Moskovets, E.; Karger, B. L. Analytical Chemistry 2004, 76 455057. 147. Suzuki, Y.; Tanji, N.; Ikeda, C.; Honda, A.; Ookubo, E.; itterio, D.; uzuki, K. Analytical Sciences 2004, 20 475-82. 148. Henry, C. M. Chemical Engineering and News 2000, 22-26. 149. Schulz, D. M.; Ihling, C.; Clore, G. M.; Sinz, A. Bichemistry 2004, 43 4703-15. 150. Leite, J. F.; Cascio, M. Biochemistry 2002, 6140-48. 151. Trester-Zedliltz, M.; Kamada K.; Burley, S. K.; Fenyo, D.; Chait, B. T.; Muir, T. W. Journal of the American Chemical Society 2003, 125 2416-25. 152. Hager-Braun, C.; Tomer, K. B. Biochemistry 2002, 41 1759-66. 153. Pandurangi, R. S.; Lusiak, P. ; Desai, S.; Kuntz, R. R. Bioorganic Chemistry 1998, 26 201-12. 154. Mano, N.; Nagaya, Y.; Saito, S.; Kobayashi, N.; Goto, J. Biochemistry 2004, 43 2041-48. 155. Doneau, C. E.; Gafken, P. R.; Bennett, S. E.; Barofsky, D. F. Analytical Chemistry 2004, 76 5667-76. 156. Darie, C. C.; Biniossek, M. L.; Jovine, L.; Litscher, E. S.; Wassarman, P. M. Biochemistry 2004, 43 7459-78. 157. Vazeux, G.; Iturrioz, X.; Cor vol, P.; Llorens-Cortes, C. Biochemistry journal 1997, 327 883-89. 158. Kim, Y. T.; Muramatsu, T.; Takahashi, K. Journal of Biochemistry 1995, 117 53544. 159. Turner, B. T.; Sabo.T, M; Wilding.D; Maurer.M, C Biochemistry 2004, 43 975565. 160. Hoffman, R. C.; Jennings, L. L.; Tsigelny, I.; Comoletti, D.; Flynn, R. E.; Sudhof, T. C.; Taylor, P. Biochemistry 2005, 1496-506. 161. Smith, T. K.; Meister, A. The Journal of Biological Chemistry 1995, 270 1247680.

PAGE 238

221 162. Hu, X.; Liu, W. Y. Protein and Peptide Letters 2003, 10 503-09. 163. Korn, C.; Scholz, S. R.; Gimadutdinow, O.; Pingoud, A.; Meiss, G. Nucleic Acid Research 2002, 30 1325-32. 164. Yin J; Marnett LJ; Ziaohua L; Walsh C T Journal of the American Chemical Society 2004, 126 7754-55. 165. Tyagi, N. K.; Kinne, R. K. H. Analytical Chemistry 2003, 323 74-83. 166. Tate, J. J.; Persinger, J.; Bartholomew, B. Nucleic Acids Research 1998, 26 142126. 167. McMahan, S. A.; Burgess, R. R. Biochemistry 1994, 12092-99. 168. Jahn, O; Tezval, H; Spiess, J; Eckart, K International Journal of Mass Spectrometry 2003, 228 527-40. 169. Romaniouk, A. V.; Silva, A.; Feng, J.; Vijay, I. K. Glycobiology 2004, 14 301-10. 170. van der Horst, G. T. J.; Mancini, G. M. S.; Brossmer, R.; Rose, U.; Verheijen, F. W. The Journal of Biological Chemistry 1990, 265 10801-04. 171. Huh, J. W.; Lee, H. J.; Choi, M. M.; Yang, S. J.; Yoon, S. Y.; Kim, D. W.; Kim, S. Y.; Choi, S. Y.; Cho, S. W. Bioconjugate Chemistry 2005, 16 710-16. 172. Rancour, D. M.; Menon, A. K. Biochemistry journal 1998, 333 619-69. 173. Contradt, H. S; Bunsch, A; Brossmer, R Federation of European Biochemical Studies 1984, 170 295-99. 174. Yarema, J. A.; Sun, Z. Nature Chemical Biology 2005, 1 69-70. 175. Saxon, E.; Bertozzi, C. R. Science 2000, 287 2007-10. 176. Gross, H. J.; Brossmer, R. Europeran Journal of Biochemistry 1988, 177 583-89. 177. Wang, X.; Zhang, L.; Ye, S. Medicinal Research and Reviews 2002, 23 32-47. 178. Wang, X.; Niu, Y.; Cao, X.; Zhang, L.; Zhang, L.; Ye, X. Bioorganic & Medicinal Chemistry 2003, 11 4217-24. 179. Skropeta, D.; Schworer, R.; Haag, T.; Schmidt, R. R. Glycoconjugate Journal 2004, 21 205-19. 180. Little, D. P.; Aaserud, D. J.; Valaskovic, G. A.; McLafferty, F. W. Journal of the American Society for Mass Spectrometry 1996, 118 9852-359.

PAGE 239

222 181. Oberacher, H.; Parson, W.; Oefner, P. J.; Mayr, B. M.; Huber, C. G. Journal of the American Society for Mass Spectrometry 2004, 15 510-22. 182. Oberacher, H.; Oefner, P. J.; Holzl, G.; Pr emstaller, A.; Davis, K.; Huber, C. G. Nucleic Acids Research 2002, 30 e67-e71. 183. Oefner P.J.; Huber C.G. Journal of Chromatography B 2002, 782 27-55. 184. Bellis, G. D.; Salani, G.; Battagila, C.; Pietta, P.; Rosti, E.; Mauri, P. Rapid Communications in Mass Spectrometry 2000, 14 243-49. 185. Null, A. P.; George, L. T.; Muddiman, D. C. Journal of the American Society for Mass Spectrometry 2002, 13 338-44. 186. Premstaller, A.; Oberacher, H.; Huber, C. G. Analytical Chemistry 2000, 72 438693. 187. Huber, C. G.; Krajete, A. Journal of Chromatography A 2000, 870 413-24. 188. Huber, C. G.; Oefner, P. J.; Bonn, G. K. Analytical Chemistry 1993, 212 351-58. 189. Oberacher H; Walcher W; Huber C G Journal of Mass Spectrometry 2003, 38 10816.

PAGE 240

223 BIOGRAPHICAL SKETCH Jeremiah Dwayne Tipton, son of Mr. a nd Mrs. James D. Tipton, was born in Inverness, Florida, on February 9, 1977. Jeremi ah lived in Inverness, Florida, until the age of 1, when upon his family moved to New Port Richey, Florida. There, he attended Gulf High School and graduated in 1995 with highest honors. He then attended the Florida Institute of Technology, Florida, where he graduated cum honorae in May, 2000 with a Bachelor of Science degree in research chemistry. Ne xt, he was accepted to the University of Florida, Department of Chemis try (Gainesville, Florida) to pursue his Doctor of Philosophy degree in bio-analytical chem istry. Jeremiah received his Ph.D. in chemistry in December, 2005 under the guidance of Dr. David H. Powell and Dr. Nicole A. Horenstein. After graduation, Jeremiah relo cated to Jupiter, Flor ida, to start his postdoctorial studies at the newly founded Scripps Research Institute of Florida.