Citation
Heavy Atom and Hydrogen Kinetic Isotope Effect Studies on Recombinant, Mammalian Sialyltransferases

Material Information

Title:
Heavy Atom and Hydrogen Kinetic Isotope Effect Studies on Recombinant, Mammalian Sialyltransferases
Copyright Date:
2008

Subjects

Subjects / Keywords:
Atoms ( jstor )
Enzyme substrates ( jstor )
Enzymes ( jstor )
Isotope effects ( jstor )
Isotopes ( jstor )
Kinetics ( jstor )
Oxygen ( jstor )
Phosphates ( jstor )
Purification ( jstor )
Rats ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Embargo Date:
7/30/2007

Downloads

This item has the following downloads:


Full Text












HEAVY ATOM AND HYDROGEN KINETIC ISOTOPE EFFECT STUDIES ON
RECOMBINANT, MAMMALIAN SIALYLTRANSFERASES













By

ERIN E. BURKE


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

ERIN E. BURKE
































This work is dedicated to my husband, my parents, and to all of my family and friends
who have supported me in this endeavor.















ACKNOWLEDGMENTS

My sincerest gratitude goes to my advisor, Dr. Nicole Horenstein, for her guidance,

support, and patience during my graduate studies and the course of this project. I would

also like to recognize and thank the members of my supervisory committee, Dr. Nigel

Richards, Dr. Jon Stewart, Dr. Ronald Castellano, and Dr. Art Edison for their

suggestions and support on this project over the last five years.

Special thanks go to the past and present Horenstein group members: Jen,

Jeremiah, Fedra, Andrews, Mirela, and Jingsong for their company and help throughout

the years. I am especially grateful to Jen, Jeremiah, and Fedra for their friendships and

steadfast support. Additionally, I wish to thank all of my colleagues in the Biochemistry

Division for their advice and for their generosity in allowing me to borrow equipment and

reagents when needed. I would also like to express my appreciation to JoAnne Jacobucci

and Romaine "Sugar Momma" Hughes for their administrative help and for keeping me

on an endless sugar high with their bottomless candy bowls.

A heartfelt thank you goes to my parents, Paul and Elizabeth Ringus, my brother,

Adrian, and my best girlfriends, Georgia, Felicia, Patricia, and Cerissa, for their support,

encouragement, and unwavering love throughout this endeavor. A sincere thanks also

goes to my relatives living in Melbourne, FL, for their love, encouragement, and for their

warm hospitality when I needed a break from lab work.

I am also deeply indebted to my wonderful husband, Andy, for his steadfast

patience, love, encouragement, support, and friendship at all times. I am so thankful to









have met him in graduate school and I look forward to our future together. Without his

support, none of my accomplishments in graduate school would have been possible.

Finally, I would like to express my eternal gratitude to my savior, Jesus Christ, and

to God. Thank you especially for answering all of my prayers and for bestowing me with

the talents to pursue a career in the field of Chemistry. Their unwavering support has

made this difficult journey worthwhile. Thus, ". .. with God all things are possible."(

Matthew 19:26).
















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 ............................... .... ...... ... ................. .x

ABSTRACT ........ .............. ............. ...... ...................... xiv

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

Sialic A cids ................. ...................... ....................... ............... .
Glycosyltransferases and Glycosidases ............................................. ............... 7
Fucosyltransferases .............................................. .... .. .... .............. .. 11
Sialyltransferases ............................................................... .. ... ......... 13
Sialyltransferase Inhibitors ............................................................ ............... 19

2 SYNTHESIS AND CHARACTERIZATION OF SUBSTRATES .........................24

In tro d u ctio n ................................................... .................. ................ 2 4
R results and D discussion ............................. ...... ............................ .... ...... ...... 24
Synthesis of CMP-NeuAc isotopomers............................................................24
Synthesis of [1-14C-N-acetyl, 2-10] CMP-NeuAc ..........................................27
Synthesis of [1-14C-N-acetyl, 1802] CMP-NeuAc ......................... ..........30
Synthesis of UMP-NeuAc Isotopomers ............................................................37
E x p e rim e n ta l ....................................................................................................4 1
M a te ria ls ....................................................................................................4 1
In stru m en ta l ................................................................................................... 4 2
Synthesis of [3H -N -acetyl] M anN A c.......................................... ....................... 43
Synthesis of [1-3H-N-acetyl] NeuAc and [1-3H-N-acetyl] CMP-NeuAc ............44
Cloning, Overexpression and Purification ofN-acetylneuraminic Acid
A ld olase [E C 4 .1.3 .3 ]............... .............................................. .... .. ........ .. 4 5
Overexpression of CMP-NeuAc Synthetase [EC 2.7.2.43] ..............................45
Cloning, Overexpression and Purification of Uridine Kinase [EC 2.7.1.48] ......46
Synthesis of 75 atom % [1-14C-N-acetyl, 2-180] CMP-NeuAc........................47
Synthesis of KH2P1804 ......... ................................................ 47
Synthesis of P180 3 CM P .................... ......... ........................ ............... 48









Synthesis of [1-14C-N-acetyl, P1O2] CMP-NeuAc...........................................49
Synthesis of U M P-N euA c ........................................................ ............. 49

3 PURIFICATION AND KINETIC CHARACTERIZATION OF
RECOMBINANT HUMAN ALPHA (2-3) SIALYLTRANSFERASE IV ............51

Intro du action ...................................... ................................................ 5 1
R results and D discussion .................... ........ ............................. .... ........ ......... 52
Overexpression and Purification of Recombinant Human a(2--3)
Sialyltransferase Isoform s........................................................... ............... 52
Kinetic Characterization of Recombinant h23 STGal IV Isoforms ..................64
C o n c lu sio n s........................................................................................................... 6 9
E x p e rim e n ta l ..................................................................................................6 9
M materials and M ethods ........................................ ...... .............................69
Preparation of pFastBacHTaInsulin/h23 STGal IV (Ins-h23 STGal IV)..............70
Preparation of pFastBacHTalnsulin/NtermHis6x-tag-h23 STGal IV Plasmid
(N term H is-h23STG al IV ) ................ ......................................... ..................... 71
Preparation of pFastBacHTalnsulin/CtermHis6x-tag-h23 STGal IV Plasmid
(CtermHis-h23 STGal IV) ............. ....... ............... ............... 72
Amplification of Recombinant Baculovirus Plasmids .............. .. ............ 73
Expression and Purification of Ins-h23STGal IV .........................................74
Expression and Purification ofNtermHis-h23STGal IV and CtermHis-
h23 ST G al IV Isoform s................................................................ ..................... 75
Sialyltransferase Enzyme Activity Assays........................ ......................76
Steady State Kinetics for Recombinant h23 STGal IV Isoforms.........................76

4 KINETIC ISOTOPE EFFECT STUDIES ON RECOMBINANT
SIALYLTRAN SFERA SES ......................................................... ............... 78

Introduction..................................... ... ........... ............... .......... 78
Kinetic Isotope Effect (KIE) Background ...................................... ............... 79
Isotope Effect Theory .......................................................... ............... 79
P rim ary Isotope E effects ............................................................ .....................84
Secondary Isotope Effects ....................... ............................. 85
Kinetic Isotope Effect Measurement Technique ...........................................87
The C om petitive M ethod.......................................................... ............... 87
The N oncom petitive M ethod.................................... .......................... .. ......... 90
Kinetic Isotope Effect M methodology ....................................... ............... 91
R results and D iscu ssion ............................. ........................ .. ...... .... ...... ...... 94
C o n c lu sio n s......................................................................................................... 1 0 4
E xperim mental ................................... ............................105
Enzyme Reaction General KIE Methodology ............................105

5 CONCLUSIONS AND FUTURE WORK........................................................109









L IST O F R E FE R E N C E S ......................................................................... ................... 112

BIOGRAPHICAL SKETCH ............................................................. ..................120















LIST OF TABLES


Table pge

2-1 CM P-N euA c isotopom er yields ........................................ .......................... 25

2-2 U M P-N euA c isotopom er yields .................................... ............................. ........ 38

3-1 Recombinant Ins-h23STGal IV purification table. ...............................................57

3-2 Recombinant NtermHis-h23STGal IV purification table ..................................61

3-3 Recombinant CtermHis-h23STGal IV purification table ..................................62

3-4 Kinetic paramters for sialyltransferase isoforms..................... ... ............... 65

4-1 KIEs measured for recombinant human a(2--3) sialyltransferase (Ins-h23STGal
IV ) ........................................................ .................................94

4-2 KIEs measured for recombinant rat a(2--3) sialyltransferase (r23 STGal IV).
The asterisk denotes the KIE previously measured by Michael Bruner.1 ...............94

4-3 KIEs measured for recombinant rat a(2--6) sialyltransferase (r26STGal I).
Asterisks denote KIEs previously measured by Michael Bruner.18,19...................94

4-4 Summary of predicted KIEs based on mechanism................................................100















LIST OF FIGURES


Figure pge

1-1 General structure of free sialic acid and N-acetyl neuraminic acid (NeuAc)
which is transferred by sialyltransferases .... ........... ....................................... 2

1-2 Structure of the viral sialidase inhibitor, DANA.....................................................4

1-3 Chemical structure of sialyltransferase inhibitor KI-8110............... ..................6

1-4 Proposed mechanism for inverting and retaining glycosyltransferases and
glycosidases from Lairson et al.30 ........................................................................ 8

1-5 3-D structural representations of the GT-A and GT-B fold groups of
glycosyltransferases from Coutinho et al.29 ......... ........................ ...............10

1-6 Proposed mechanism of a(--3) fucosyltransferase V from Murray et al.40...........12

1-7 Common topology of a type II membrane protein from Wang et al.43 The L, S,
and VS sialylmotifs of sialyltransferase are indicated in color.............................13

1-8 Reactions catalyzed by a(2--6) sialyltransferase and a(2--3) sialyltransferase. ....14

1-9 Proposed transition-state for sialyltransferase-catalyzed reaction from
Horenstein et al.4 ................................. ........................... .... ........ 16

1-10 Interaction of the ring oxygens of CMP-3FNeuAc with the phosphate oxygens
in CstIIA32 (left) and interactions of CMP and active site residues (right) from
Chiu et al.49........................ ..................................17

1-11 Structure of CMP-quinic acid and transition-state analogs ....................................22

1-12 Transition-state analogs of CMP-NeuAc synthesized by Horenstein and co-
w orkers.59 .......................................................................................... 23

2-1 Structure of labeled CMP-NeuAc. Asterisks denote sites of isotopic
substitution. ..................................................................... 25

2-2 Enzymatic synthesis of N-acetyl neuraminic acid and CMP-NeuAc. Asterisks
indicate sites of possible isotopic substitution. ....................................................... 26









2-3 Radioactive profile of HPLC fractions from a typical NeuAc reaction. The
composition of radioactive ManNAc and NeuAc in the reaction mixture was ~
20 % and 80 %, respectively after four days..................... ....... ........................ 27

2-4 Ring opening mechanism for NeuAc.................................................................... 28

2-5 31P-NMR of [1-14C-N-acetyl, 2- 0] CMP-NeuAc ............................................29

2-6 (-) ESI-MS of [2-180] CMP-NeuAc m/z 615 [M-H]-(top panel), zoom MS/MS
of [2-180] CMP-NeuAc [M-H]- (center panel), and MS/MS dissociation of m/z
6 15 [M -H ]- ion (bottom panel) ................................................................................30

2-6 Enzymatic synthesis of [P1803] CMP from KH2P1804. The enzymes used in this
synthesis were glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 3-
phosphoglycerate phosphokinase (3-PGK) and uridine kinase (UDK). .................31

2-7 Enzymatic synthesis of [1-14C-N-acetyl, P1802] CMP-NeuAc from [P1803] CMP. 31

2-8 (-) E SI-M S of K H 2P 1 0 4. ............................................................... ..................... 33

2-9 31P-NMR spectrum of KH2P104 (1M) in D20 with 4 mM EDTA. ........................34

2-10 10 % SDS-PAGE of purified UDK fractions from Red-A dye affinity column......35

2-11 (+) ESI-MS of P1803 CMP 2 (upper panel) and zoom-MS of the [M+H] ions
(low er panel). ...................................................... ................. 36

2-12 (+) ESI-MS spectrum of [P16021802] CMP-NeuAc......................... ...............37

2-13 Structure of labeled UMP-NeuAc. Asterisks denote sites of isotopic
substitution. .......................................... ............................ 38

2-15 Chemical deamination of CMP-NeuAc to UMP-NeuAc by sodium nitrite.............39

2-16 HPLC chromatogram of CMP-NeuAc deamination reaction after 48 hr................40

2-17 HPLC chromatogram of CMP-NeuAc deamination after 30 hrs incubation..........41

3-1 Diagram of the recombinant h23STGal IV constructs. The insulin signal peptide
sequence is represented in yellow, the truncated h23STGal IV sequence is
represented in blue and the His6x-tag sequence is represented in purple. ...............53

3-2 General scheme for the generation of recombinant baculoviruses and protein
expression with the BAC-TO-BAC expression system ............... .....................56

3-3 Structure of CDP-Hexanolamine affinity ligand synthesized for the purification
of recombinant Ins-h23 STGal IV enzyme. ........... ...............................................56









3-4 Typical elution chromatogram of recombinant Ins-h23 STGal IV from a
sepharose CDP-hexanolamine affinity column. Solid squares represent protein
concentration and open diamonds represent activity. ............................................59

3-5 10 % SDS-PAGE of purified recombinant Ins-h23STGal IV. Lane 1, MW
standard; Lane 2, Lane 3, and Lane 4 are 10 tL, 20 tL, and 30 gL loads of a
TCA precipitation of purified Ins-h23STGal IV, respectively(left gel). SDS-
PAGE of purified recombinant Ins-h23STGal IV digested with PNGase.18 Lane
1, MW standard; Lane 2, 30 tL load of a TCA precipitation of purified Ins-
h23STGal IV and Lane 3 is 30 tL load of PNGase digested Ins-h23STGal IV......60

3-6 Typical elution chromatogram of recombinant NtermHis-h23 STGal IV and
CtermHis-h23 STGal IV from a Ni2+-NTA affinity column. Solid squares
represent protein concentration and open diamonds represent activity. .................63

3-7 10 % SDS-PAGE gel of purified recombinant CtermHis-h23STGal IV
NtermHis-h23 STGal IV. Lane 1, TCA precipitation of purified CtermHis-
h23STGal IV; Lane 2, MW Standard; Lane 3, TCA precipitation of purified
N term H is-h23 ST G al IV ..........................................................................................63

3-8 Michaelis-Menten plots for recombinant Ins-h23 STGal IV with with varied
[CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-lactose]
and constant [CM P-NeuAc] (bottom panel). ................................... ..................... 66

3-9 Michaelis-Menten plots for recombinant NtermHis-h23 STGal with varied
[CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-lactose]
and constant [CM P-NeuAc] (bottom panel). ................................... .................67

3-10 Michealis-Menten plots for recombinant CtermHis-h23STGal IV with varied
[CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-lactose]
and constant [CMP-NeuAc] (top panel) and for with (lower panel)......................68

4-1 Expanded terms of the Bigeleisen equation. ................................. ...... ............ ...81

4-2 Free energy diagram depicting the looser potential energy wells in the transition-
state resulting in a normal (>1) isotope effect from Lowry et al.4 .........................83

4-3 Free energy diagram depicting the looser potential energy wells in the transition-
state resulting in an inverse (<1) isotope effect from Lowry et al.4.........................83

4-4 The symmetric stretching vibration mode in the transition-state of transfer
reactions. C represents the isotopically substituted atom that is transferred
betw een A and B ........................................................................... .. ...... 85

4-5 Typical to (top panel) and ti/2 (lower panel) UMP-NeuAc HPLC chromatograms
for KIE experiments on recombinant sialyltransferase........................................93









4-6 Positional isotope exchange (PIX) mechanism. PIX could not be dectected for
the sialyltransferase mechanism. If pixing is complete, the bridge 180 label
scrambles to give a 33 % 180 distribution at each oxygen .....................................98

4-7 Bond order analysis for protonation at the non-bridging phosphate oxygen of the
donor sub state. .................................................................... 100

4-8 Bond order analysis for protonation at the bridging phosphate oxygen of the
donor substrate (top panel) and no protonation of the phosphate oxygens of the
donor substrate (low er panel). .................................................................... ....... 101

4-9 Transition-state models proposed for recombinant r26STGal I (left) and
r23STG al IV (right) enzym es ........................................................ ....................103

4-10 Early and late transition state analogs for PNP from Schramm.31 .........................105

5-1 Proposed sialyltransferase transition-state inhibitor. ............................................111















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

HEAVY ATOM AND HYDROGEN KINETIC ISOTOPE EFFECT STUDIES ON
RECOMBINANT, MAMMALIAN SIALYLTRANSFERASES

By

Erin E. Burke

August, 2005

Chair: Nicole A. Horenstein
Major Department: Chemistry

Sialylated glycoproteins and glycolipids are key recognition molecules for a host of

biological processes such as cell-cell regulation, cell adhesion, and biological masking.

Sialyltransferases are glycosyltransferases that catalyze the biosynthesis of sialylated

oligosaccharides in the Golgi apparatus of many prokaryotic and eukaryotic cells. The

mechanism of sialyl transfer from activated donor substrate, 5'-cytidine monophosphate

N-acetylneuraminic acid (CMP-NeuAc), to terminal positions of carbohydrate groups on

glycoproteins and glycolipids is still not fully understood. Kinetic studies on

recombinant rat liver a(2--6) sialyltransferase propose a mechanism with a late

oxocarbenium ion-like transition-state and general acid catalysis to assist in glycosyl

transfer. This dissertation describes the mechanistic study and the transition-state

analysis of three recombinant sialyltransferases. The results from this study will provide

an increased understanding of the mechanism of glycosyl transfer which may be useful in

the future development of new sialyltransferase inhibitors.









The first part of this work describes the synthesis and purification of the

isotopically labeled CMP-NeuAc and UMP-NeuAc donor substrates required to conduct

the desired kinetic experiments. Details describing a novel enzymatic route for the

synthesis of non-bridging phosphate 180 labeled CMP-NeuAc are also presented in this

section. The characterization of these isotopically labeled substrates is shown here as

well.

Following the synthesis of the substrates, the next section describes the cloning,

overexpression, and purification of three recombinant human a(2--3) sialyltransferases,

two of which contain either a N-terminal His6x-tag or a C-terminal His6x-tag. The

purification yields, specific activities, and kinetic parameters of these recombinant human

a(2--3) sialyltransferases are also presented.

The dissertation concludes with the discussion of the kinetic isotope effect studies

on recombinant human a(2--3), rat liver a(2--3), and rat liver a(2--6) sialyltransferase

with the aforementioned isotopically labeled substrates. The kinetic isotope effects that

were measured on these enzymes include secondary 3-dideuterium, binding, control, and

primary and secondary 180 leaving group isotope effects. Comparisons were made

among isotope effects measured for the recombinant human and rat a(2--3)

sialyltransferases and for the recombinant rat a(2--6) sialyltransferase. The KIE data

provide new information regarding the nature of the transition-states for the a(2--3) and

a(2--6) sialyltransferase enzymes.














CHAPTER 1
INTRODUCTION

Sialyltransferases are glycosyltransferases that catalyze the transfer of sialic acid

(N-acetylneuraminic acid, NeuAc) from an activated CMP-NeuAc donor substrate to

non-reducing termini of glycoproteins and glycolipids with inversion of configuration at

the NeuAc glycon. Over the last two decades, research interest in sialyltransferases has

increased primarily because these enzymes play a critical role in the regulation of a host

of biological processes. The mechanism of sialyl transfer is still not fully understood.

Results from kinetic studies conducted previously by Bruner and Horenstein on

recombinant rat liver a(2--6) sialyltransferase (ST6Gal I) suggested that the mechanism

proceeds via a late oxocarbenium ion-like transition-state with general acid catalysis to

assist in glycosyl transfer.1'2 The work described in this dissertation represents an

investigation into the reaction catalyzed by recombinant a(2--3) sialyltransferase from

human placenta (h23STGal IV). The enzyme-substrate binding interactions at the

phosphate group of the donor substrate are of particular interest. Additionally,

comparisons can be made between enzymes in the same family using the information

obtained from work completed on recombinant rat liver a(2--3) sialyltransferase and

from the work previously reported on the recombinant ST6Gal I.

Sialic Acids

In order to understand the significance of studying this class of enzymes, one must

first appreciate the functional importance of sialic acids for a host of biological processes.

The purpose of this introduction is not to provide a comprehensive report on the subject










of sialic acids, but to merely highlight some of the key roles that sialic acid residues play

in these biological processes. This information will emphasize the importance of how

studying the enzymes involved in sialic acid regulation could lead to ways in which to

control and monitor a wide array of biological pathways.

In nature, sialic acids are linked in the terminal steps of the synthesis of cell surface

glycoproteins and glycolipids. The structure of sialic acid is unique in that it contains a

highly acidic carboxylate group on the anomeric carbon (Figure 1-1). The negative

charge on sialic acids is an important chemical feature of this molecule and it plays a

functional role in a variety of biological processes. For example, the negative charge on

sialic acids provides this molecule with the ability to attract and repel specific cells and

biomolecules.3 In the case of attraction, the negative charge allows sialic acids to bind to

positively charged molecules and assist in their transport.


OH OH
HO
H C2H

HO
Sialic Acid
(5-amino-3,5-dideoxy-D-glycero-D-galacto-nonulosonic acid)


OH OH
HO2H
AcHN OC02H
H O'""
HO
N-acetyl neuraminic Acid (NeuAc)
(5-amino-3,5-dideoxy-D-glycero-D-galacto-nonulsonic acid)


Figure 1-1. General structure of free sialic acid and N-acetyl neuraminic acid (NeuAc)
which is transferred by sialyltransferases.









On the other hand, the population of sialic acid residues on the periphery of cells

provides the cell with a net negative charge that is essential for the repulsion of other

cells or biomolecules. This is seen in erythrocytes and blood platelets where cell-surface

sialic acids can prevent the aggregation of these cells in the bloodstream.4 Furthermore,

the negative charge on these sugars has also been shown to contribute to the viscosity of

mucins lining intestinal endothelia cells.5 Thus, the electrochemical properties of sialic

acids appear to influence their unique function in a variety of biological phenomena.

One of the most important roles of sialic acids is their ability to function as

recognition elements for key processes. This function is facilitated by their chemical

properties and by their location on surface of cells. For example, sialic acids are

recognition molecules for bacterial and viral pathogens. The best known example of this

was observed over 50 years ago where sialic acid residues were identified as recognition

molecules for the binding of influenza A to human erythrocytes and respiratory tract

mucins.6 Since then, researchers have shown that influenza B virus can also bind to sialic

acid residues that are N-linked to cell-surface glycoproteins or glycolipids.7'8 The

binding of influenza to cell surface receptors bearing sialic acids is mediated by the viral

protein, Hemagglutinin (HA). HA works in conjunction with viral sialidase (or

neuraminidase) during the viral life cycle. Viral sialidase has been suggested to facilitate

the spread of influenza virus by cleaving sialic acid residues from the protecting mucin

layer of respiratory tract epithelia cells.9 This research lead to the development of viral

sialidase inhibitors such as 2-Deoxy-2,3-dehydro-N-acetylneuraminic acid (DANA)

which are being successfully used today for the treatment of influenza (Figure 1-2).













AcHN
HO-


DANA
(2-Deoxy-2,3-dehydro-N-acetylneuraminic acid)


Figure 1-2. Structure of the viral sialidase inhibitor, DANA.

Another example of sialic acids acting as essential recognition components is in

neural development. In this case, polysialic acid, a linear homopolymer of a(2--8)-

linked sialic acids, was discovered attached to the neural cell adhesion molecule

(NCAM).10 Two polysialyltransferases, ST8Sia II (STX) and ST8Sia IV (PST), regulate

the synthesis of polysialic acid on NCAM. Experiments involving NCAM-deficient mice

have suggested that polysialic acids on NCAM play critical roles in the regulation of

neural cell adhesion, cell migration, neurite outgrowth, and synapse formation.11'12

Furthermore, deficiencies and disorganization in polysialic production have been linked

to diseases such as schizophrenia and Alzheimer's disease.13'14

In contrast to their function as recognition molecules, sialic acids are also important

for the anti-recognition of certain biomolecules and cells. As the penultimate "capping"

molecule for cell-surface oligosaccharides, sialic acids serve as biological masking agents

by disguising and shielding their underlying sugars from receptor recognition. Ashwell

and Morell documented the first example of this function in 1974. Their experiments

investigated the role of sialic acids as masking agents by using sialidase to remove









terminally N-linked sialic acid residues from D-galactose molecules on radiolabeled

ceruloplasmin in the bloodstream.15 Upon removal of these sialic acids, exposed

galactosyl residues were quickly recognized by D-galactose receptors and the resulting

radiolabeled asialoceruloplasmin disappeared from the bloodstream within minutes.

Analysis of the liver showed increased radioactivity, thus signifying the degradation of

radiolabeled asialoceruloplasmin.

Another example of the sialic acid masking effect can be seen in the binding

recognition of siglecs in the immune system. Siglecs are sialic acid-binding

immunoglobulin-like lectins involved in cellular signalling functions and cell-cell

interactions in the nervous and immune systems.16 Siglecs use both cis and trans

interactions with sialic acid ligands when binding to the cell-surface. Researchers have

shown that the siglec receptor-binding site can be masked by cis interactions with sialic

acid ligands.16'17 These cis interactions with sialic acids are essential for the regulation of

siglec function by preventing or facilitating specific cell-cell interactions when necessary.

Sialic acids have also been implicated in the masking of tumor antigens.18 In this

case, the high sialic acid content on some tumor cells can mask antigen recognition sites.

As a result, tumor cells can elude immunological attack and, in some cases, continue to

grow uncontrollably. Clinical studies have observed that certain invasive cancer cell

lines have hypersialylated cell-surfaces and the patients with these cancers have increased

sialyltransferase activities in their blood serum.19 In response to these results, researchers

have been investigating the use of sialidases and sialyltransferase inhibitors as methods

for cancer treatment. Studies conducted on chemically induced malignant tumors in










small mammals showed marked tumor regression after treatment with sialidase.20

Additionally, sialyltransferase inhibitors such as KI-8110 have been shown to assist in

the reduction of tumor metastases by inhibiting the transfer of sialic acid onto cell-surface

oligosaccharides (Figure 1-3).21-24 Thus, these findings have opened the door to new

possibilities in the development of cancer treatments.

O

F
OAc NH
OCOOMe
AcOH2C N 0

AcHN
AcO
OAc

S O0

KI-8110
(5-fluoro-2',3'-O-isopropylidene-5'-O-(methyl
5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-a-D-galacto-2
-nonulopyranosylonate)-uridine


Figure 1-3. Chemical structure of sialyltransferase inhibitor KI-8110.

Since sialic acids participate in such a plethora of biological events, changes in

their chemical structure, concentration, or mutations to their biosynthetic pathways can

cause severe diseases. Some of the diseases include sialidosis, galactosialidosis, sialuria,

and sialic acid storage disorder (SASD).25 Sialidosis and galactosialidosis are genetically

inherited, lysosomal storage diseases that are characterized by the inability to degrade

sialylated glycoproteins due to a deficiency in the production of sialidase.26 In contrast,

sialuria is characterized by the overproduction of sialic acids in the cytoplasm resulting

from a lack of feedback inhibition of the rate-limiting enzyme, uridine diphosphate N-









acetylglucosamine 2-epimerase.27 Sialic acid storage disorder is the rarest of the

aforementioned diseases with less than 150 cases reported worldwide. Most of the

reported cases stem from a small region in northeastern Finland. SASD is typified by the

accumulation of unbound sialic acid within the lysosomes caused by a defect in sialin, a

transmembrane protein responsible for transport of sialic acid out of the lysosome. As a

result, patients with SASD have up to 100 times more unbound sialic acid in their urinary

excretions than normal. In general, patients suffering with these diseases display

neurodevelopmental delays, severe learning difficulties, coarse facial features, spinal

abnormalities, skin lesions, and visual impairment.2

The functional importance of sialic acids in biological systems is vast. Their

unique roles in a variety of cellular events make sialic acids indispensable to life.

Therefore, mechanistic studies on the enzymes involved in the biosynthesis and transfer

of sialic acids could provide new tools in which to investigate the complex nature of

these biomolecules.

Glycosyltransferases and Glycosidases

Glycosyltransferases are a class of enzymes that catalyze the transfer of

monosaccharide residues from mono- or diphosphate sugar nucleotides to the non-

reducing end of extending oligosaccharide chains, or, in general, different aglycon

acceptors. These membrane bound enzymes are highly specific for their donor and

acceptor substrates, and their function is critical for the proper glycosylation of a myriad

of oligosaccharide and polysaccharide chains. In mammalian cells, estimates indicate

that well over 100 different glycosyltransferases are required to biosynthesize all known

oligosaccharide structures.28 The resulting glycan chains regulate a diverse range of













cellular functions, including cell-cell interactions and signaling, host-pathogen


interactions, neuronal development, embyogenesis, and biological masking.29


Glycosyltransferases are classified by sequence similarity-based families and by the


type of mechanism they catalyze, which is retaining or inverting depending on the final


anomeric configuration of the product (Figure 1-4).30 The inverting glycosyltransferase


catalyzed mechanism is suggested to follow an SN2-like reaction whereby a general base


deprotonates the incoming nucleophile of the acceptor sugar, thus enabling the direct


displacement of the nucleoside diphosphate. In this mechanism, metal ions such as Mg2


or Mn2+ are believed to serve as acid catalysts for some glycosyltransferases.31


Inverting Mechanism

B
B\O H R' [HO
10- HO


rOP
AH


Retaining Mechanism

Nuc


HO _

OP
AH


S H


SHC


0O H- "-OR'
iB+
'v--^^ ~~ --ow


,OP
AH

Oxocarbenium Ion-Like
Transition State


Nuc
O /



'op
AH

Oxocarbenium Ion-Like
Transition State

S Nuc




OR
H
A

Oxocarbenium Ion-Like
Transition State


POH
t


B
-0
_\ -OR'


POH

Glycosyl-Enzyme
Intermediate


HO


Nuc
/


OR
H
2


P= Nucloside diphosphate or
monophosphate


Figure 1-4. Proposed mechanism for inverting and retaining glycosyltransferases and
glycosidases from Lairson et al.30









The retaining glycosyltransferase catalyzed mechanism is proposed to proceed via a

double-displacement reaction with a covalently bound glycosyl-enzyme intermediate. In

this mechanism, an aptly positioned amino acid within the active-site functions as a

nucleophile to catalyze the reaction.30'32 A divalent cation is believed to act as a Lewis

acid whereas the leaving diphosphate group has been suggested to serve as a general base

by deprotonating the incoming acceptor sugar hydroxyl to activate it for nucleophilic

attack. In comparison to inverting glycosyltransferases, retaining glycosyltransferase

reactions also proceed through oxocarbenium ion-like transition states. Despite this

similarity, the mechanism for retaining glycosyltransferases is still being explored since

some of the intermediates have yet to be conclusively identified.

Within the last decade, several crystal structures of glycosyltransferases have been

reported.28'32,33 Based on recent structural data, glycosyltransferases adopt one of two

general folds referred to as GT-A and GT-B (Figure 1-5).29 Glycosyltransferases

categorized under the GT-A (glycosyltransferase A) fold group typically contain a

conserved 'DXD' motif, a conical active site cleft formed by two closely associated

domains, and a/P proteins with a single Rossmann domain. The 'DXD' motif has been

shown to play a critical role in metal ion binding and catalysis. The divalent metal cation

coordinates the phosphate group oxygens of the sugar nucleotide donor in the enzyme

active-site.34'35 Additionally, binding of the nucleotide has been mainly observed on the

N-terminal domain of GT-A enzymes. Some of the glycosyltransferases that have been

classified under the GT-A fold group include phage T4 P-glucosyltransferase, glycogen

phosphorylase and a-1,3-galactosyltransferase.29'36









The GT-B (glycosyltransferase B) fold group consists of two Rossman-like P/a/P

domains that are separated by a deep substrate-binding cleft. Nucleotide binding for GT-

B enzymes takes place on the C-terminal domain while the acceptor substrates bind to the

N-terminal domain. The GT-B superfamily encompasses a diverse group of prokaryotic

and eukaryotic enzymes that are responsible for a variety of processes ranging from the

production of biologically active antibiotics to cell wall biosynthesis and gene

transcription.37












GTA GT'

Figure 1-5. 3-D structural representations of the GT-A and GT-B fold groups of
glycosyltransferases from Coutinho et al.29

In contrast to glycosyltransferases, glycosidases execute further modifications to

glycosylated biomolecules by catalyzing the cleavage of sugar residues. This class of

enzymes uses water as a nucleophile to trim carbohydrate residues from these

biomolecules in order to meet the requirements for a variety of biological processes.

Glycosidases follow similar mechanistic paths to those described above for

glycosyltransferases, but with a few exceptions. The characteristic mechanism for

retaining glycosidases involves a pair of aspartic or glutamic acid residues in enzyme

active-site, with one functioning as a nucleophile and the other acting as a general

acid/base catalyst. Unlike retaining glycosyltransferases, key intermediates in the









retaining glycosidase mechanism such as the covalently bound glycosyl-enzyme

intermediate have been conclusively identified and characterized using crystallographic

and spectroscopic methods. Recently, the Withers laboratory characterized this

covalently bound glycosyl-enzyme intermediate for the hen egg-white lysozyme (HEWL)

38
mechanism.3

Fucosyltransferases

Although glycosyltransferases have not been as well characterized as

glycosidases, one member of the glycosyltransferases superfamily that has been closely

studied is fucosyltransferase. Fucosyltransferases catalyze the transfer of 1-fucose from

an activated GDP-fucose donor substrate to oligosaccharide chains linked to proteins or

lipids. In recent years, the mechanism for a(1--3) fucosyltransferase V (FucTV) has

been investigated using kinetic isotope effect experiments and inhibitor studies.3941

FucTV catalyzes the final step in the biosynthesis of sialyl Lewis X and Lewis X

fucoglycoconjugates. These fucoglycoconjugates play an essential role in the regulation

of cell-cell interactions for a variety of immune system processes.

Kinetic isotope effect and pH-rate studies conducted on FucTV suggest that the

mechanism for fucosyltransfer is base catalyzed where the L-fucose is transferred to

acceptor sugars with inversion of configuration.39 Results from secondary isotope effect

studies using deuterated GDP-[ 1-2H]-Fucose as the donor substrate indicated that

cleavage of the glycosidic bond occurs prior to nucleophilic attack as illustrated in Figure

1-6.40 Furthermore, the transition-state structure is similar to glycosidases in that it is

proposed to have a flattened half-chair conformation with considerable oxocarbenium ion

character at the anomeric position.













Mn2+

OO"
OH 0
HO .zzk._O
0 P
OH OHHO 0 0 0 G
SGDP r-OH
Acceptor--OH boken first

OH
02C O
E Acceptor 'H, H
2C- E




OH /
HO

OH
OK 0 + GDP

O0 Acceptor



Figure 1-6. Proposed mechanism of (1--3) fucosyltransferase V from Murray et al.40

Product inhibition studies on FucT V have revealed that the mechanism is an

ordered, sequential, Bi-Bi mechanism in which the GDP-Fuc binds first followed by the

acceptor sugar. FucTV also requires a metal co-factor, typically Mn2, to achieve optimal

catalysis. Furthermore, FucTV can use both charged and uncharged sugar acceptor

substrates such as N-acetyllactosamine (LacNAc) and sialyl LacNAc, respectively. This

sugar acceptor substrate variability is similar for sialyltransferases however, the FucTV

donor substrate, GDP-Fuc, does not contain the highly acidic carboxylate group on its

anomeric carbon like the sialyltransferase donor substrate, CMP-NeuAc. Thus, this

group will undoubtedly alter the enzyme-donor substrate reactivity when compared to

GDP-Fuc. Nevertheless, the mechanism for FucTV may provide some insight into the

nature of the sialyltransferase and other glycosyltransferase catalyzed reactions.









Sialyltransferases

As a subfamily of glycosyltransferases, sialyltransferases are also localized in the

Golgi apparatus and their topology is characteristic of a type II membrane protein with a

short cytoplasmic domain, an N-terminal signal anchor and a large luminal catalytic

domain (Figure 1-7). There are presently 20 cloned cDNA's of sialyltransferases isolated

from bacteria, insects, and mammals. Their nomenclature and function are determined

by the different acceptor sugar substrates that sialyltransferases bind to in the transfer of

NeuAc. For example, the recombinant h23STGal IV catalyzes the transfer of NeuAc

from CMP-NeuAc donor substrate to the C3 terminal hydroxyl of Galpl,4GlcNAc or

Galpl,3GalNAc acceptor sugars while the recombinant ST6Gal I transfers NeuAc

residues to C6 terminal hydroxyl of Galpl,4GlcNAc substrates (Figure 1-8).42


Figure 1-7. Common topology of a type II membrane protein.













SNH nu HO' HO O 0
OH II 'i" NHAc OH
OH O--P-O O N SialylLacNAc
HO AeHN 0h CMP
HO OH OH

CMP-NeuAc '"
OH CO; OH HO

AcHNS HO
HO HO
NHAc OH
SialylLacNAc


Figure 1-8. Reactions catalyzed by a(2--6) sialyltransferase and a(2--3)
sialyltransferase.

Aside from common topological features, sialyltransfersases do not share any

sequence homology with other enzymes in the glycosyltransferase family. However,

sequence homology analysis of the sialyltransferase family revealed the existence of

several conserved protein motifs within the catalytic domain referred to as L, S, and VS

sialylmotifs. The L and S sialylmotifs are located at the center of the lumenal catalytic

domain of sialyltransferases and are composed of approximately 48 and 23 amino acid

residues, respectively. The VS sialylmotif is located in the C-terminus of

sialyltransferases and consists of two highly conserved glutamate and histidine residues

that are separated by four amino acid residues. The Paulson laboratory conducted a

series of site-directed mutagenesis studies on recombinant rat liver ST6Gal I to

investigate the roles of several conserved amino acids in the L & S sialylmotifs.43'44

Kinetic data from analysis of ST6Gal I mutant constructs suggested that the L sialylmotif

participated in the binding of CMP-NeuAc, while the S sialylmotif participated in the

binding of both donor and acceptor substrates. The Paulson group also found that the two

invariant cysteine residues present in each of the L and S sialylmotifs for an









intrachaindisulfide bond that is essential for retention of catalytic activity and proper

conformation of ST6Gal I.45

A closer examination of all known eukaryotic sialyltransferase sequences revealed

the presence of another highly conserved motif located between the S and VS

sialylmotifs referred to as the aromatic motif. 46 This motif is comprised of a stretch of

four highly conserved mostly aromatic residues. The functional role of these amino acid

residues was investigated using site-directed mutagenesis experiments on recombinant

human hST3Gal I. The results suggested that the highly conserved histidine (His299)

and tyrosine (Tyr300) residues of the aromatic motif are necessary for enzyme activity

since their mutation to alanine generated inactive enzymes.

Apart from research involving conserved residues in the sialylmotifs, there is still

limited information available concerning the catalytic mechanism and structure of the

sialyltransferase family. Previous work in the Horenstein laboratory used radiolabeled

CMP-NeuAc and UMP-NeuAc, a weak binding substrate analog, to conduct a series of

kinetic isotope effect and pH vs. rate experiments on recombinant rat liver ST6Gal I to

elucidate the mechanism of glycosyl transfer. A dissociative mechanism involving a late

oxocarbenium ion-like transition state was proposed in the model for sialyltransferase

catalysis based on KIE results (Figure 1-9).1,47 The pH-rate profile from experiments

using UMP-NeuAc and LacNAc as the donor-acceptor substrate pair fits a bell-shaped

curve for two ionizable groups with pKa values of 6.2 and 8.9. Further pH-rate

experiments and theoretical calculations suggested that glycosyl transfer proceeded via a

general acid catalyzed mechanism in which a non-bridging phosphate oxygen on CMP-

NeuAc may be protonated to facilitate the loss of CMP.1 The kinetic mechanism was









proposed to be steady-state random based on initial velocity, KIE, and isotope trapping

experiments.1,2


0 P 0


Op- O Cytosine
OH i 0
"H+"


AcHNO
HO O-




Figure 1-9. Proposed transition-state for sialyltransferase-catalyzed reaction from
Horenstein et al.47

A three dimensional crystal structure of sialyltransferase CstII from Campylobacter

jejuni was recently reported by Chiu et al.48 From the structure, sialyltransferase

CstIIA32 was categorized under the GT-A fold group because it contained a single

Rossmann domain. Aside from this feature, sialyltransferase CstII lacked the conserved

'DXD' motif and a bound metal in the active site which are common characteristics

among other glycosyltransferases in the GT-A group. Since sialytransferases do not

require a metal cofactor for catalysis,49-51 the lack of the conserved 'DXD' motif

responsible for the binding of a divalent metal cation was not surprising.

In order to observe enzyme-substrate binding interactions in the active site, Chiu et

al. crystallized sialyltransferase CstIIA32 in the presence of CMP-3FNeuAc, an

unreactive substrate analog of CMP-NeuAc (Figure 1-10). The structure of

sialyltransferase CstIIA32 completed with CMP-3FNeuAc offers some explanations

regarding the nature of sialyl transfer. In the crystal structure, the sialyl moiety of CMP-









3FNeuAc adopted a distorted skew boat conformation which favors formation of the

oxocarbenium ion. The leaving-group phosphate was oriented in a pseudo axial position

twisted above the plane of the sugar ring allowing the pro-R oxygen on the phosphate to

interact with the ring oxygen on NeuAc (Figure 1-10). Cleavage of the glycosidic bond

and departure of the CMP moiety was suggested to be facilitated both by the negative-

charge buildup on the pro-R phosphate oxygen, and by the hydrogen bonding interactions

with the non-bridging phosphate pro-S oxygen and active-site Tyrl56 and Tyrl62

residues. Mutagenesis experiments using Y156F and Y162F mutants of sialyltransferase

CstIIA32 resulted in a significant loss in catalytic activity with only one tyrosine residue

mutated and a total loss in catalytic activity with both tyrosine residues mutated.

Although it is unclear why acid catalysis would be necessary to assist in the departure of

a stable monophosphate leaving group, these results indicate that both residues are

critical for optimal catalytic efficiency of the CstIIA32 transferase mechanism.






OH I,
CCOAH

HO F T,''I'
o... N- : iq j .



Figure 1-10. Interaction of the ring oxygens of CMP-3FNeuAc with the phosphate
oxygens in CstIIA32 (left) and interactions of CMP and active site residues
(right) from Chiu et al.48

Deprotonation of the incoming hydroxyl group of the acceptor sugar was suggested

to be catalyzed by Hisl88. The close proximity ofHisl88 at 4.8k to the anomeric









carbon in the crystal structure made Hisl88 the only feasible candidate for the role of

general base catalyst; however, this identification is still ambiguous. Based on active-site

comparison studies, Hisl88 is located in a similar position to other catalytic bases

identified in inverting glycosidases.52 Additionally, the pH optimum of 8.0 for the

CstIIA32 catalyzed reaction favors deprotonation of the His188 imidazole, therefore,

allowing it to act as a general base catalyst. A complete loss of transferase activity was

also observed when His188 was mutated to alanine in CstIIA32. This same result was

also observed in histidine to lysine/alanine mutagenesis experiments on recombinant

human hST3Gal I and polysialyltransferases, ST8Sia II and IV.46'53 Thus, the results

from these experiments reinforce the hypothesis that a histidine residue plays an

important role in the sialyltransferase catalyzed reaction.

Despite the information obtained from this sialyltransferase crystal structure,

bacterial sialyltransferase CstIIA32 does not share sequence homology with any

mammalian sialyltransferases, which is where the primary research interest exists.54

Hence, one can not assume that the mammalian sialyltransferases will adopt the same

structural fold and active-site arrangement as sialyltransferase CstIIA32. To date, there

are no three dimensional crystal structures reported for a mammalian sialyltransferase.

This is primarily due to the fact that recombinant mammalian sialyltransferases are more

difficult to overexpress and purify. Sufficient quantities of pure enzyme are arduous to

obtain to conduct crystallization experiments. Furthermore, even if enough pure enzyme

was available to attain a crystal structure, these structures do not provide information

about the reaction's transition-state structure.









Therefore, other techniques may be used advantageously on mammalian

sialyltransferases in order to probe the mechanism of sialyl transfer. Methods such as

kinetic isotope effect (KIE) experiments can often provide detailed information on the

transition-state structure which will assist in acquiring mechanistic information for the

sialyltransferase catalyzed reaction. In this study, several CMP-NeuAc and UMP-NeuAc

radioisotopomers were synthesized to investigate the mechanism of sialyl transfer using

KIE experiments. The dual-label competitive method was used to measure the KIEs for

these radiolabelled substrates with recombinant human placental a(2--3)

sialyltransferase and recombinant rat liver a(2--3) sialyltransferase. The data from these

experiments will provide an increased understanding of the mechanism of glycosyl

transfer with regard to interactions at the phosphate leaving group via 180 isotopic

substitution at the glycosidic O and non-bridging phosphate oxygen atoms.

Sialyltransferase Inhibitors

Within the last decade, a burst in the design and synthesis of a variety of

sialyltransferase inhibitors occurred due, in part, to their interest as potential therapeutic

compounds for the treatment of tumor metastases and immunological diseases. Inhibition

studies on purified sialyltransferase also became more feasible in recent years because of

the increased commercial availability of recombinant sialyltransferases. Although there

have been numerous reported sialyltransferase inhibitors, only the more noteworthy

inhibitors will be discussed here.

The most common strategy used toward the design of sialyltransferase inhibitors

have been donor substrate based analogs of CMP-NeuAc. The idea behind the

development of these inhibitors was to alter functional groups on the sugar or nucleotide

portion of the donor substrate, but maintain the basic glycosidic linkage. In 1997,









Schauer and co-workers conducted a series of inhibition studies on recombinant a(2--6)

sialyltransferase from rat liver and a(2--3) sialyltransferases from porcine submandibular

gland using a variety of different nucleosides, nucleotides, sialic acid and sugar

nucleotide analogs as donor substrates.24 The goal of the study was to identify key

structural elements that were essential for inhibition of sialyltransferase. The inhibition

studies showed that donor substrate analogs containing a nucleotide monophosphate

moiety were the most effective sialyltransferase inhibitors, while the sialic acid analogs

displayed little to no inhibition. CMP, CDP, and CTP were natural competitive inhibitors

of sialyltransferase with Ki values of 90, 50, and 46 atM, respectively. These inhibition

constants are comparable to the Km value for the natural donor substrate CMP-NeuAc (46

atM).55 The enhanced inhibitory effect upon addition of one or more phosphate groups

was proposed to be caused by their ability to provide a negative charge similar to the

carboxylate group of CMP-NeuAc. Moreover, the results from this study suggest that the

nucleotide moiety, particularly cytidine monophosphate, is a fundamental structural

requirement for high binding affinity of the donor substrate to the enzyme active site.

This information led to the development of more sialyltransferase inhibitors that

incorporated the general cytidine or cytidine monophosphate scaffolding in the donor

substrate. Schmidt et al. synthesized a series of sialyltransferase inhibitors with cytidine

monophosphate linked to quinic acid analogs.56 These CMP-quinic acid based inhibitors

were advantageous to use because they not only included the CMP moiety for high

binding affinity, but they also blocked transferase activity by changing the glycosidic

bond to a more stable C-glycoside linkage. These compounds were also stable under

physiological conditions. Inhibition experiments using CMP-quinic acid as the donor









substrate with ST6Gal I gave a Ki value of 44 aM, which is approximately the same as

the Km value for the natural substrate CMP-NeuAc. With this information in hand,

Schmidt and co-workers modified their inhibitor design strategy by synthesizing

transition-state analogs that would mimic the oxocarbenium ion-like transition-state

proposed for CMP-NeuAc.5 A few of the transition state analog inhibitors synthesized

and tested by Schmidt et al. are shown in Figure 1-11.

These compounds contain a flattened ring with the anomeric carbon trigonal planar

to simulate the oxocarbenium ion coplanarity in the transition-state. A methylene group

was also added between the anomeric carbon and CMP to model their increased distance

in the proposed transition-state structure of CMP-NeuAc. Substitution of the methylene

hydrogen with a phosphonate group greatly increased the inhibitory activity of these

compounds with Ki values in the nanomolar range. The phosphonate group provided an

additional negative charge similar to the carboxylate of CMP-NeuAc. Furthermore,

replacing the glycerol side chain on the NeuAc ring with a phenyl group as seen in Figure

1-11, resulted in a 1,000-fold increase in binding affinity to ST6Gal I with a Ki of 29 nM.

These results demonstrate the enzyme's capability to tolerate bulky side chain

modifications made to the donor substrate without compromising binding affinity. To

date, the phenyl phosphonate compound is the most potent sialyltransferase inhibitor.

Horenstein and co-workers synthesized another unique set of transition state

analogs as sialyltransferase inhibitors.58 This new class of sialyltransferase inhibitor

employed an unsaturated bicyclic system with a conjugated carboxylate group to mimic

the conformation of the proposed transition-state (Figure 1-12). Additionally, the CMP

moiety attached to the bicyclic system was kept at an increased distance from the










anomeric carbon to imitate the late transition-state distance proposed for bond cleavage.

These compounds were highly efficient inhibitors of sialyltransferase with Ki values in

the low micromolar range. Furthermore, substitution of the NeuAc ring with a bicyclic

ring illustrates the enzyme's ability to accept a diverse range of structure changes in the

sugar portion of the donor substrate. Recently, Schmidt and co-workers demonstrated

that sialyltransferase also exhibit high binding affinities for transition-state analogs with

aryl and hetaryl ring systems substituted for the sugar portion.59 Thus, these studies

indicate that the neuraminyl ring only plays a minor role in binding since structure

variations to this part of the donor substrate are still tolerated by the enzyme.


NH2 NH2

o HO OH I
NI OH 0H-- N 0
o PN' O N ---N

OH OHOH
,OH
R=Hor Po.O
0
CMP-Quini Aid Phosphonate Compound
CMP-Quinic Acid Phosphonate Compound


Phenyl Phosphonate Compound


Figure 1-11. Structure of CMP-quinic acid and transition-state analogs.


















Ki = 20 tM Ki = 10 tM


Figure 1-12. Transition-state analogs of CMP-NeuAc synthesized by Horenstein and co-
workers.58

Moreover, these studies have enabled researchers to ascertain that the key

components required for sialyltransferase inhibitiors are: (i) a planar anomeric carbon;

(ii) an increased distance between the anomeric carbon and the leaving group CMP; (iii)

at least two negative charges near the cleavage site; and (iv) the cytidine moiety for

recognition.60 Transition-state analogs of CMP-NeuAc have been the most potent of all

sialyltransferase inhibitors reported to date. Therefore, information about the transition

state acquired from kinetic isotope effect studies may also prove useful toward the

development of new sialyltransferase inhibitors.














CHAPTER 2
SYNTHESIS AND CHARACTERIZATION OF SUBSTRATES

Introduction

Cytidine 5'-monophosphate neuraminic acid (CMP-NeuAc) is synthesized by

CMP-NeuAc synthetase in many prokaryotic and eukaryotic cells and serves as a key

intermediate in the sialyltransferase catalyzed biosynthesis of sialylated oligosaccharides.

During catalysis, sialic acid (N-acetylneuraminic acid, NeuAc) is transferred from an

activated CMP-NeuAc donor substrate to non-reducing termini of glycoproteins,

glycolipids, and oligosaccharide chains. Although sialyltransferases vary in acceptor

substrate specificity, all sialyltransferases use CMP-NeuAc as their donor substrate.

Thus, information obtained from experiments conducted on recombinant human and rat

a(2--3) sialyltransferases, may be applied other members in the sialyltransferase family.

Results and Discussion

Synthesis of CMP-NeuAc isotopomers

In order to probe the mechanism of the sialyltransferase catalyzed reaction, a series

of CMP-NeuAc isotopomers were synthesized to perform the desired experiments. These

CMP-NeuAc isotopomers either contain one radioactive trace label or a radioactive trace

label with several nonradioactive isotopic substitutions. The various sites of isotopic

substitution are illustrated in Figure 2-1. The isolated yields are shown in Table 2-1.









NH2

N
0
*,N 0



AcHN,-*Z 5 0 0coo
H O4 OH OH
HO IO 3

H*


Figure 2-1. Structure of labeled CMP-NeuAc. Asterisks denote sites of isotopic
substitution.

Table 2-1. CMP-NeuAc isotopomer yields.
CMP-NeuAc Isotopomer Isolated % Yield
[9-3H] 74
[1-3H-N-acetyl] 78
[1-14C-N-acetyl] 54
[1-14C-N-acetyl, P102] 35
[1-14C-N-acetyl, 2-180] 71


Chemical and enzymatic methods were employed to synthesize the various CMP-

NeuAc isotopomers (Figure 2-2). The first step of the synthesis required the use of N-

acetyl neuraminic acid (NANA) aldolase which was cloned, overexpressed in E. coli and

purified according to literature procedures.61-64 NANA adolase catalyzes the aldol

condensation reaction between pyruvate and N-acetyl mannosamine (ManNAc) to yield

NeuAc.65 Isotopomers of NeuAc were obtained by substituting nonradiolabeled

ManNAc with the appropriate radiolabeled ManNAc substrate. Several of these

radiolabeled ManNAc compounds were purchased commercially, such as, [6-3H] and [1-

14C-N-acetyl] ManNAc which give [9-3H] and [1-14C-N-acetyl] NeuAc, respectively.

The [1-3H-N-acetyl] NeuAc isotopomer was obtained by synthesizing the [3H-N-acetyl]

ManNAc starting material via acetylation of D-mannosamine with 3H acetic anhydride as











per the literature method.66 The progress of the radiolabeled NeuAc reactions was


monitored by making analytical injections on anion-exchange HPLC and collecting


fractions for LSC counting (Figure 2-3). Radiolabeled ManNAc and NeuAc eluted with


retention times of 3 min. and 6 min., respectively, on HPLC MonoQ (50-100 mM


NH4HCO3 gradient, 15% methanol, 2 mL/min, A271, 2 mL fractions collected).


ManNAc

*H *


HN NANA Aldolase
0 _O _
250C, 72 hrs
OH


NeuAc


HO OH

*


HH
IHO
0


OH


-0 CO2


CTP
CMP-NeuAc
Synthetase I
370C, 6 hrs
V: PiT


HO OH



H

HO
SHO
0


O


O-P/co2
0-
SD*CO2


CMP-NeuAc






Figure 2-2. Enzymatic synthesis of N-acetyl neuraminic acid and CMP-NeuAc.
Asterisks indicate sites of possible isotopic substitution.


Further isotopic modifications to the radiolabeled NeuAc were made once it was


isolated. Deuterium labels were incorporated into the NeuAc at the C3 position by


HO-


HO0
HO


-0 Na


N




OH OH









exchanging the protons with deuterium under alkaline conditions with D20. This

synthesis was performed by Mike Bruner of the Horenstein laboratory as previously

described in the literature.2 [3,3'-2H2] CMP-NeuAc was then synthesized with CMP-

NeuAc synthetase with the addition of CTP.



NeuAc Reaction Radioactive Profile

250000 NeuAc

200000

150000

100000
ManNAc
50000



0 5 10 15 20
Fraction #



Figure 2-3. Radioactive profile of HPLC fractions from a typical NeuAc reaction. The
composition of radioactive ManNAc and NeuAc in the reaction mixture was ~
20 % and 80 %, respectively after four days.

Synthesis of [1-14C-N-acetyl, 2-180] CMP-NeuAc

Synthesis of the [1-14C-N-acetyl, 2-180] CMP-NeuAc isotopomer was achieved by

first exchanging the C-2' hydroxyl oxygen on NeuAc via a ring opening mechanism with

H2180 (95% enrichment) under basic conditions (pH > 9.5) (Figure 2-4).67 The

enzymatic synthesis with [1-14C-N-acetyl, 2-180] NeuAc, CTP, and CMP-NeuAc

synthetase gave [1-14C-N-acetyl, 2-180] CMP-NeuAc in 71 % isolated yield after

purification by anion-exchange HPLC. The 31P-NMR spectra showed two peaks at -










4.243 ppm and -4.257 ppm representing the [1-14C-N-acetyl, 2-160] and [1-14C-N-acetyl,

2-180] CMP-NeuAc compounds, respectively (Figure 2-5). Integration of the peaks

indicated the relative abundance of the [1-14C-N-acetyl, 2-160] CMP-NeuAc compound

to be 25 % and the [1-14C-N-acetyl, 2- 0] CMP-NeuAc compound was 75 %. ESI-MS

spectral data was obtained from a parallel nonradiolabeled synthesis of [2-10] CMP-

NeuAc.

B:

HO OH HO OH
HO



H o3
HOH


dHO
HO HO


Figure 2-4. Ring opening mechanism for NeuAc.















[1-14C-N-acetyl, 2-180]
CMP-NeuAc




[1-14C-N-acetyl, 2-160]
CMP-NeuAc









-4.2 ppm -4.3

Figure 2-5. 31P-NMR of [1-14C-N-acetyl, 2-180] CMP-NeuAc.

The (-)ESI-MS from the [2-18O] CMP-NeuAc synthesis shows the most abundant

ions at m/z 615 [M-H]- for [2-1O] CMP-NeuAc and m/z 613 [M-H]- for the [2-160]

CMP-NeuAc compound (Figure 2-6). The [(M-H+Na)-H] adduct of m/z 615 was also

present at m/z 637. The m/z 615 ion underwent MS/MS to yield the labeled m/z 324

[CMP-H]- ion. Selected ion monitoring of the ions at m/z 615 ([2- 0] CMP-NeuAc) and

m/z 613 ([2-160] CMP-NeuAc) indicated a distribution of 75.6 % and 24.4 %,

respectively.















[2-1O] CMP-NeuAc [M-H]-
15 7







500 550 600 650 700 750 800 850 900 950 1000 105
m/z
[2-1O] CMP-NeuAc [M-H]


[2-180] CMP-NeuAc [M]


P801603 CMP [M]


Figure 2-6. (-) ESI-MS of [2- O] CMP-NeuAc m/z 615 [M-H]-(top panel), zoom
MS/MS of [2-10] CMP-NeuAc [M-H]- (center panel), and MS/MS
dissociation of m/z 615 [M-H]- ion (bottom panel).

Synthesis of [1-14C-N-acetyl, P1802] CMP-NeuAc

Synthesis of the [1-14C-N-acetyl, Pl02] CMP-NeuAc isotopomer was achieved


using a multi enzymatic synthesis route to selectively incorporate 180 labels into the non-


bridging phosphate oxygens of CMP-NeuAc (Figures 2-6 and 2-7).68


-1- -.- .....












KHP804 + 03PO" H NAD

NH2 OH
1[P8 N Glyceraldehyde-3-phosphate
[P'803 CMP N GAPDH

180 NO NADH
H18 0- I
18u \
O18H Cytidme
OH OH 0
UDK


03P18H + [7 -P1803]ATP PO 18OP1803
OH
OH 3-PGK OH
3-Phosphoglycerate 3-PGK OH
S1,3 Diphosphoglycerate


ADP

Figure 2-6. Enzymatic synthesis of [P 03] CMP from KH2P1804. The enzymes used in
this synthesis were glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 3-
phosphoglycerate phosphokinase (3-PGK) and uridine kinase (UDK).


PEP Pyruvate



Pyruvate Kinase

[a-P1803] CDP [a-P103] CTP

PEP ADP [l 14C-N-acetyl] NeuAc

Pyruvate Nucleoside CMP-NeuAc
Kinase Monophosphate Synthetase
Kinase
Pyruvat ATP PPi NH
PO83 CMP [-14C-N-acetyl, P 02] CMP-NeuAc N

HO OH 'II Ni O

SOH

r-0NC OH OH
OHO



Figure 2-7. Enzymatic synthesis of [1-14C-N-acetyl, P1802] CMP-NeuAc from [P1803]
CMP.









The first step of the synthesis involved the preparation of KH2Pl04 which was

synthesized in 74 % yield via hydration of PC15 with H2180 (95 % atom enrichment)

followed by the addition of 2 M KOH.69 This synthesis produced five different

phosphate species (P104, 160 1803, p16021802, p1603180, and P1604) due to the isotopic

distribution of the 1O label. HPLC/ (-) ESI-MS and 31P NMR analysis of KH2P1804

measured the relative isotopic abundance of the five phosphate species to be the

following: 80.4 % 04, 16.2 % 160 03, 1.6 % 16021802, 0.1 % 160310, and 1.7 % 1604

(Figures 2-8 & 2-9).70 These results are consistent with the statistical distribution of 180

for the synthesis of P1804 using 95 % atom enriched H2180. The KH2P1804 (80 % 1804)

compound was then used in an enzymatic synthesis with glyceraldehyde-3-phosphate,

(G3P), NAD+, ADP, and cytidine to give P 03 CMP in 64 % isolated yield after

purification on anion-exchange HPLC. The last step in the synthesis of Pls03 CMP

required the use of uridine kinase (UDK) which was cloned, overexpressed in E. coli, and

purified by dye affinity chromatography.71 Enzyme purification yielded 150 units and

uridine kinase (23 kDa) was 90-95 % pure base on SDS-PAGE analysis (Figure 2-10).


















1[H2P0]

1055


20

15

10

5 [H2P"04]
S9715 985

9I PA P7 mQ


-Il. .I *. ]-
1035

1067

[H2,P" [H2p'0261802] 1044
994 1014 1023 1076

99 100 101 102 103 104 105 106 107 10 109 110
mfz


Figure 2-8. (-) ESI-MS of KH2P1804.


100













































160 P160 180


2.5 2.4 2.3 2.Z
ppm


Figure 2-9. 31P-NMR spectrum ofKH2Ps04 (1M) in D20 with 4 mM EDTA.









MW F76 F78 F80

......... .... ..








29 LP ".,-
45

36 a



24

20 f

14 f



Figure 2-10. 10 % SDS-PAGE of purified UDK fractions from Red-A dye affinity
column. The sample load for each fraction was 20 kL and the gel was stained
with coomassie blue.

The HPLC/MS analysis of P1803 CMP showed four peaks at m/z 324, 326, 328 and

330 corresponding to the P1604 CMP, p1603180 CP, P16021802 CMP, P16003 CMP

compounds, respectively (Figure 2-11). An (+) ESI-MS scan measured the relative

abundances of the various CMP isotopomers to be: 60 % 1601803, 6.4 % 1602102, 1.1 %

1603180, 32.5 % 1604. The results show an approximate 20 % dilution of the 80 label

from the initial enrichment of P104 used in the synthesis. This may be explained by the

fact that glyceraldehyde-3-phosphate decomposes at neutral pH to release its phosphate,

thus resulting in a dilution of the [P1804 2-]. In response to this result, a shorter

incubation time and higher enzyme concentrations were used to help minimize the

decomposition of G3P in the reaction mixture. A higher [P104] was also used to reduce

unlabeled phosphate incorporation.















P180 CMP [M+H]
3301
100
90 P [M+Na]
80 -
70 -
60
60 3241
50 [cytidineH]46
112 1
40
30
20 3 3741
S2293 3682 3858 4001
90 I3220 5




P 803 CMP [M+H]+
329 9
100
90 3299
80 P1604 CMP [M+H]+






30 323 1 p1603180CMP [32H
33 P 0C[M+H]P [tme
324 1 250 3280 331 0
10 3238w 4 3249 3250 529 379 328 1 3290 329 3309
.3223 3227 3230 3236 ,,, 3 250, 3255 329 326 1 326 327 1 3278 3289 33-6 8
323 324 325 32627 7 328 329 330 331 332
mn/z



Figure 2-11. (+) ESI-MS of P803 CMP 2 (upper panel) and zoom-MS of the [M+H]
ions (lower panel).


The [l-14C-N-acetyl, P1802] CMP-NeuAc isotopomer was enzymatically


synthesized using P1803 CMP (60 % 1s03) and [1-14C-N-acetyl] NeuAc. The isolated


yield was 35 % after purification on anion exchange HPLC. Since ESI-MS spectral data


of a radiolabeled compound could not be obtained, a parallel synthesis was conducted to


determine approximate isotopic incorporation for the radiolabeled synthesis. The (+)


ESI-MS spectra from the nonradiolabeled synthesis of [P802] CMP-NeuAc shows the


most abundant ions at m/z 619 [M+H]+ for [P1602102] CMP-NeuAc and m/z 615


[M+H]+ for [p1604] CMP-NeuAc (Figure 2-12). The [M+Na]+ adducts of m/z 619 and


m/z 615 were also present at m/z 641 and m/z 637, respectively. The m/z 619 ion


underwent MS/MS to yield the labeled m/z 328 [CMP+H] ion which in a MS/MS/MS












scan produced the m/z 112 [Cytidine+H] ion. Other, less intense ions were detected at


m/z 310 and m/z 292 which correspond to fragments of the NeuAc moiety. Selected ion


monitoring of the ions at m/z 619 ([P1602102] CMP-NeuAc), m/z 617 ([P1603180] CMP-


NeuAc), and m/z 615 ([P1604] CMP-NeuAc) indicated a distribution of 56.6, 9.2 and 34.2


%, respectively.




P'602'02 CMP-NeuAc
[M+H]
61800 9
5 P'"04 CMP-NeuAc
[M+H]
0 614 9 p6002"02 CMP-NeuAc
[M+Na]
0640 9


70 8
65
60 6
55 P1604 CMP-NeuAc
-| [M+Na]
50 636 9



o5 \ 6169


25 633 0
S6469







Synthesis of UMP-NeuAc Isotopomers
6491













Previous work conducted in the Horenstein laboratory showed that there is a
/0 623 63 0 659 / 669O
6141 62 3 638 64 L 651
5 59118 5964 601... ,4 61 628 6] 4 64 I3 9 I I 66 1 I,
590 595 6OO 6O5 610 615 620 625 63O 635 640 645 65O 655 66O 665 670




Figure 2-12. (+) ESI-MS spectrum of [p160218021] CMP-NeuAc.


Synthesis of UMP-NeuAc Isotopomers


Previous work conducted in the Horenstein laboratory showed that there is a


commitment to catalysis for the CMP-NeuAc donor substrate when bound to the


enzyme.1 In other words, the sialyltransferase catalyzed reaction with CMP-NeuAc


donor substrate contains more than one kinetic barrier that is partially rate limiting.









Furthermore, non-chemistry rate limiting steps can mask the full expression of the kinetic

isotope effects. To circumvent this problem, UMP-NeuAc, an unnatural "slow" donor

substrate for sialyltransferase, was also synthesized to aid in the study of the

sialyltransferase catalyzed mechanism. UMP-NeuAc binds more weakly to

sialyltransferase and the chemistry step is slower than for CMP-NeuAc as indicated by its

higher Km and lower kcat values.1 Thus, these factors made UMP-NeuAc an ideal donor

substrate analog to use for the desired set of kinetic experiments.

UMP-NeuAc was synthesized with a variety of isotopic substitutions. The

locations of these isotopic labels were equivalent to those used in the synthesis of the

CMP-NeuAc isotopomers (Figure 2-13). The isolated yields of the UMP-NeuAc

isotopomers are shown in Table 2-2.

0


NH
0
N 0

*OH OH OH
HOHO 6 2 3
O 1
8 ac 5 I coo

H*


Figure 2-13. Structure of labeled UMP-NeuAc. Asterisks denote sites of isotopic
substitution.

Table 2-2. UMP-NeuAc isotopomer yields.
UMP-NeuAc % Isolated Yield
[9-3H] 47
[1-3H-N-acetyl] 33
[1-14C-N-acetyl] 42
[1-14C-N-acetyl, P102] 30
[1-14C-N-acetyl, 2- 0] 44










Synthesis of the UMP-NeuAc isotopomers was achieved through a chemical

deamination reaction using 1 M NaNO2, pH 3.8 and the appropriately labeled CMP-

NeuAc compounds Figure (2-15). This reaction was challenging due to the instability of

CMP-NeuAc under acidic conditions. Thus, deamination reactions were carried out at 4

C to help minimize the decomposition of CMP-NeuAc. The conversion of CMP-NeuAc

to UMP-NeuAc also varied depending on how the CMP-NeuAc compounds where

purified on anion-exchange HPLC. Deamination reactions using CMP-NeuAc

isotopomers purified on anion-exchange HPLC with an ammonium bicarbonate buffer

system proceeded slowly, and the conversion of CMP-NeuAc to UMP-NeuAc was 38

% with < 20 % decomposition after 48 hrs at 4 C (Figure 2-16). This result was

explained by considering the concentration of free ammonia present in the solution of

CMP-NeuAc after desalting the purified compound with Amberlite IR-120 H+.

Ammonia assays conducted on the purified and desalted CMP-NeuAc isotopomers

estimated the [NH4 ] in solution to be 50 mM. The excess NH4+ in the reaction

solution slows the progress of CMP-NeuAc deamination by reacting with NaNO2. This

was especially the case when the [CMP-NeuAc] in the reaction was in the micromolar

range.





0 Ny 0
SH HO OH

N N"


CMP-NeuAc UMP-NeuAc



Figure 2-15. Chemical deamination of CMP-NeuAc to UMP-NeuAc by sodium nitrite.









Several ideas were tested to remove more of the NH4 from the CMP-NeuAc

solution such as, filtering the solution through a minicolumn of zeolite, applying the

sample to a mini gel filtration column, and desalting the sample again with Amberlite IR-

120 H While these approaches were somewhat effective in removing the [NH4 ], the

additional purification methods led to a further loss and decomposition of the

radiolabeled CMP-NeuAc product. Additional NaNO2 was also added to the reaction

mixture to expedite the deamination of CMP-NeuAc, but this made the desired products

more difficult to isolate and purify on anion-exchange HPLC due to the increased

complexity of the chromatograms.



40-
CMP-NeuAc UMP-NeuAc

30- s w


20
CMP UMP










Figure 2-16. HPLC chromatogram of CMP-NeuAc deamination reaction after 48 hr.
The vertical lines represent the beginning and end of fraction collection.

To circumvent this problem, CMP-NeuAc isotopomers slated for deamination were

purified on anion-exchange HPLC using a sodium bicarbonate buffer system.

Deamination reactions using the sodium form of CMP-NeuAc resulted in a mixture

consisting of 11 % CMP-NeuAc, 66 % UMP-NeuAc, 3 % CMP, and 20 % UMP after 30










hours at 4 C (Figure 2-17). These yields were calculated from peak integration values

from an HPLC chromatogram. This reaction was advantageous because it allowed the

synthesis of UMP-NeuAc isotopomers in higher yield without significant decomposition

of the starting material and the final product. This aspect was essential in order to obtain

enough UMP-NeuAc isotopomer to carry out the desired set of experiments.

Furthermore, unreacted CMP-NeuAc could be recovered during purification of UMP-

NeuAc and recycled for use in experiments requiring a CMP-NeuAc isotopomer.



75 IMP-NeuAc




50 -



UMP
CMP-NeuAc


CMP




-12- 110 M0 .nutes 0



Figure 2-17. HPLC chromatogram of CMP-NeuAc deamination after 30 hrs incubation.
The vertical lines represent the beginning and end of fraction collection.

Experimental

Materials

Reagents and buffers were purchased from Sigma and Fisher and used without

further purification. Recombinant rat liver a(2--3) and a(2--6) sialyltransferase was

purchased from Calbiochem. The 10O water (95% atom enrichment) was purchased from









Isotec and Medical Isotopes, Inc. N-acetyl D-mannosamine isotopomers ([1-14C-N-

acetyl] and [6-3H]) were purchased from Moravek and American Radiolabelled

Chemicals. Glyceraldehyde-3-phosphate (G3P), glyceraldehyde-3-phosphate

dehydrogenase (GAPDH) [EC 1.2.1.12] and 3-phosphoglycerate phosphokinase (3-PGK)

[EC 2.7.2.3] were purchased from Sigma. Nucleoside monophosphate kinase (NMK)

[EC 2.7.4.4] was purchased from Roche Applied Science. Liquid scintillation fluid

(ScintiSafe 30 %) was purchased from Fisher. The E. coli expression plasmid pWV200B

haboring the E. coli CMP-NeuAc synthetase gene [EC 2.7.7.43] was a generous gift from

Dr. W. F. Vann at the National Institutes of Health.

Instrumental

A Rainin HPLC system consisting of a HPXL binary pump and a model UV-1

detector was used. HPLC separations were performed on a Mono Q HR 10/10 anion

exchange column (Amersham-Pharmacia) monitored at 271 nm. Data collection was

achieved on a personal computer using the Star Workstation Version 6.2 software

(Varian Inc.). A Rainin-Dynamax fraction collector (model FC-1) was used to collect

eluent samples from the HPLC. A Packard 1600 TR instrument was used for liquid

scintillation counting. Mass spectrometry (LC-MS) was performed on a ThermoFinnigan

(San Jose, CA) LCQ in electrospray ionization (ESI) mode. The system was interfaced

with an Agilent (Palo Alto, CA) 1100 binary pump HPLC system consisting of an

Applied Biosystems Model 785A programmable absorbance detector set at 254 nm.

HPLC separations for LC-MS were performed on a Phenomenex (Torrace, CA) Synergi

4u Hydro-RP 80A C18 column (mobile phases = 0.5 % HOAc in water/0.5 % HOAc in

methanol). 31P-NMR spectra were acquired on a 300 MHz Mercury NMR spectrometer.

E. coli cells were lysed using a French pressure cell with Carver hydraulic press.









Centrifugation was performed with a Sorvall RC 5B centrifuge. A Labconco CentriVap

Concentrator (speedvac) was used to concentrate small volume samples.

Synthesis of [3H-N-acetyl] ManNAc

[1-3H-N-acetyl] ManNAc was synthesized using a procedure adapted from that of

Roseman et. al.66 Freshly prepared Dowex 1 x 8-200 mesh (CO32-) (250 mg damp) and

D-mannosamine-HCl (12.3 mg, 56 [mol) was suspended in 250 gL of water and

carefully added to the ampoule containing [3H] acetic anhydride (5 mCi, s.a.-100

mCi/mmol, 50 [tmol). The reaction was stirred in an ice bath for 4 hours. The aqueous

solution was removed from the reaction solution and passed over an amberlite IR-120 H+

minicolumn. The eluate was collected in a 25 mL round bottomed flask. The reaction

ampule was washed three times with 1 mL of water and the solution was passed over the

amberlite column after each wash. Water (1 mL) was used to wash the amberlite

minicolumn twice more. The 25 mL flask containing the eluate was fitted to a short path

distillation apparatus with a 10 mL receiving flask. The solution was refluxed twice and

then cooled to room temperature. The mixture was then concentrated to dryness in

vacuo, keeping the bath temperature below 55 C. The pot residue was twice resuspended

in 2 mL of water and reconcentrated in vacuo. The residue was dissolved in 500 tL of

water and [1-3H-N-acetyl] ManNAc was purified from the reaction mixture via multiple

injections onto a HPLC C18, 1 x 30 column (5 % methanol (v/v), 95 % water, A220 nm, 1

mL/min). The [1-3H-N-acetyl] peak (r.t. 10 min) was collected after each injection. The

[1-3H-N-acetyl] ManNAc fractions were combined, the solution was concentrated to

dryness in vacuo, and the product was resuspended in 1 mL of water. The product

contained 150 tCi for an isolated yield of 6 %.









Synthesis of [1-3H-N-acetyl] NeuAc and [1-3H-N-acetyl] CMP-NeuAc

[1-3H-N-acetyl] D-mannosamine (30 [tCi) was concentrated to dryness in vacuo

using the speedvac. Phosphate buffer, pH 7 (40 mM, 100 pL) containing 1 mg/mL BSA

and 1 mg/mL NaN3 was added to the 1.5 mL reaction tube. Sodium pyruvate (10 mg, 1

mmol) and 2 units ofNANA aldolase were also added to the reaction tube.65 The

reaction mixture was placed at room temperature for four days and judged to be >80 %

complete by LSC counting of reaction aliquots analyzed by HPLC (2 mL HPLC Mono-Q

fractions, 15 % methanol (v/v), 500 mM NH4HCO3, pH 8.0, 10 20 % salt gradient, 2

mL/min). The retention times for [1-3H-N-acetyl] ManNAc and [1-3H-N-acetyl] NeuAc

were 3 min and 6.5 min, respectively. The reaction mixture was concentrated to dryness

in vacuo using the speedvac and resuspended in 120 uL of 10 mM HEPES, pH 7.5 buffer

containing NeuAc (10 mM, 0.4 mmol) and CTP (15 mM, 1 mmol). CMP-NeuAc

synthetase (3 units) and 4 pL of 2.5 M MnCl2 were also added to the reaction tube. The

mixture was incubated at 37 C for 8 hours and the judged to be >80 % complete by LSC

counting of reaction aliquots analyzed by HPLC (2 mL HPLC Mono-Q fractions, 15%

methanol (v/v), 75 mM NH4HCO3, pH 8.0, isocratic, 2 mL/min). The product was

purified by HPLC on a Mono-Q column. The [1-3H-N-acetyl] CMP-NeuAc fraction was

collected, desalted with Amberlite IR-120 H concentrated to dryness in vacuo, and the

material was resuspended in 300 [tL of diH20. The purified product contained 15 tCi for

a yield of 78 %. This procedure was also used to synthesize other isotopomers of NeuAc

and CMP-NeuAc by using the appropriately radiolabeled precursors.









Cloning, Overexpression and Purification of N-acetylneuraminic Acid Aldolase [EC
4.1.3.3]

The N-acetylneuraminic acid aldolase(NANA aldolase) gene62 was amplified from

E. coli K12 genomic DNA using PCR (upper primer-5'-

ATGGCAACGAATTTACGTGGCGTAA-3' and lower primer 5'-

TCACCCGCGCTCTTGCATCAACTGC-3'). The gel purified PCR product was ligated

into the pETBlue-1 vector and transformed into NovaBlue competent cells for plasmid

amplification. The plasmid was purified using the QIAprep Spin Miniprep Kit (Qiagen)

and subsequently used in a transformation reaction with TunerTM(DE3)pLacI competent

cells. IPTG induction of a 2 L culture of the TunerTM(DE3)pLacI cells harboring the

recombinant plasmid resulted in overexpression of the target enzyme. N-

acetylneuraminic acid lyase was purified following the published protocol and was

judged 90-95 % pure based on SDS-PAGE analysis.61,63,72

Overexpression of CMP-NeuAc Synthetase [EC 2.7.2.43]

The original pWV200B plasmid harboring the CMP-NeuAc synthetase gene was

given as a generous gift from Dr. W. F. Vann. The plasmid was transformed into E. coli

JM109 cells via electroporation and the cells were plated onto luria broth agar plates

containing 60 tg/mL ampicillin. Overnight incubation of the plates at 37 C yielded ~

200 colonies. Several colonies were selected from the plates with a sterile toothpick and

used to inoculate a culture tube containing 5 mL of luria broth supplemented with 100

[tg/mL ampicillin. The culture was grown in a 37 C shaking incubator (200 rpm) until an

O.D.600 nm = 0.8 1.0. The 5 mL culture was then used to inoculate a 2 L culture of luria

broth containing 100 [tg/mL ampicillin. The culture was grown in a 37 C shaking

incubator (200 rpm) until an O.D.600nm = 0.6 0.8. The culture was induced with IPTG









(1 mM final concentration) for 10 hours in a 37 C shaking incubator to overexpress

CMP-NeuAc synthetase. The cells were harvested via centrifugation at 5000 rpm, 4 C,

for 30 min. The cells were resuspended in 20 mL of purification buffer (10 mM HEPES,

pH 7.0 containing 1 mM EDTA, 10 mM MgCl2, and 1 mM PMSF) and lysed using a

French pressure cell and Carver hydraulic press (2 runs). The supernatant was separated

from cellular debris via centrifugation at 8000 rpm, 4 C for 30 min. CMP-NeuAc

synthetase was purified from the supernatant by Red-A (Millipore) dye affinity column

chromatography (2.5 x 7 in) using a linear salt gradient from 0-1 M KC1 in purification

buffer. Fractions were analyzed for protein using the Bradford assay method and for

CMP-NeuAc synthetase activity using the published assay.73'74 CMP-NeuAc synthetase

containing fractions were combined and concentrated in an Amicon concentrator to a

final protein concentration of>10 mg/mL. The protein solution was then saturated with

80 % (NH4)2S04 and the precipitate was stored at 4 C until further use. The yield was ~

60 units and the CMP-NeuAc synthetase was 90-95 % pure based on SDS-PAGE

analysis.

Cloning, Overexpression and Purification of Uridine Kinase [EC 2.7.1.48]

The uridine kinase (UDK) gene71 was amplified from E. coli K12 genomic DNA

using PCR (upper primer-5'-ATGACTGACCAGTCTCACCAGCAGTGCG-3' and

lower primer -5'-AAGCTTATTCAAAGAACTGACTTAT-3'). The PCR product was

gel purified using the QIAquick Gel Extraction Kit (Qiagen) and ligated into the

pETBlue-1 vector (Novagen). The recombinant plasmid was transformed into NovaBlue

(Novagen) competent cells and the plasmid was purified using QIAprep Spin Miniprep

Kit (Qiagen). Transformation of TunerTM(DE3)pLacI competent cells with the construct

and IPTG-induced overexpression of a 2 L culture yielded the target enzyme. Uridine









kinase was purified by Red-A (Millipore) dye affinity column (2.5 x 7 in) using a linear

salt gradient from 0-1 M KC1 (0.1 M Tris-HC1, pH 7.8, 4 mM EDTA). Fractions were

analyzed for protein using the Bradford assay method and for UDK activity using the

published assay.73'75 UDK containing fractions were combined and concentrated in an

Amicon concentrator to a final volume of 2 mL. The yield was 150 units and the UDK

was 90-95 % pure based on SDS-PAGE analysis.

Synthesis of 75 atom % [1-14C-N-acetyl, 2-180] CMP-NeuAc

A mixture ofN-acetyl neuraminic acid (2 mg, 6 [tmol) and [-14C-N-acteyl]

neuraminic acid (15 [tCi) was dissolved in 200 [tL of 1.9 mM glycine, 0.5 M MnC12, pH

9.5 buffer. The solution was concentrated to dryness in vacuo and the material was

resuspended in 300 [tL of H2180 (95 % enrichment). The reaction mixture was placed at

37 C for 16 hrs. Cytidine-5-triphosphate (15 mg, 28 [tmol) was then added to the

reaction tube along with CMP-NeuAc synthetase (2 units). The pH of the reaction

mixture was adjusted to pH 7.5 as necessary with 5 N NaOH. The reaction was

incubated at 37 C for 8 hrs and judged to be >75 % complete by LSC counting of 2 mL

HPLC Mono-Q fractions (15 % methanol (v/v), 75 mM NH4HCO3, pH 8.0, isocratic, 2

mL/min, A271). The product was purified by HPLC on a Mono-Q column. The [1-14C-N-

acetyl, 2-180] CMP-NeuAc fraction was collected, desalted with Amberlite IR-120 H+,

concentrated to dryness in vacuo, and the material was resuspended in 300 [tL of diH20.

The purified product contained 12 [tCi for a yield of 71 %.

Synthesis of KH2P1804

KH2P 04 was synthesized using a method similar to that ofRisley et al.69 The

80 water (95 % atom enrichment) (300 [tL, 15 [tmol) was added drop wise via syringe to

a two necked flask containing phosphorus pentachloride (428 mg, 2 mmol). The PC15









was weighed in a dry box. Once removed, it was immediately placed on a Schlenk line

under constant flow of N2 (g). This was done to reduce H2160 contamination. The

reaction was stirred at 0 C for 1 hr under constant flow of dry N2 (g). The flask was then

warmed to room temperature and heated in a water bath at 100 C for 30 min. The

remaining reaction solution was cooled to room temperature and 2 mL of deionized

water was added. The solution was titrated to pH 5 with 2 M KOH and KH2P1O4 was

precipitated from solution by addition of 95 % ethanol. The precipitate was collected by

concentration under reduced pressure.

Synthesis of P803 CMP

For the synthesis of PlO3 CMP a solution of KH2PsO4 (1M, pH 7.0) was made

from which 250-350 pL was mixed with glyceraldehyde-3-phosphate (60 tL, 2 mM), 3-

NAD (60 tL, 3 mM), ADP (40 pL, 1 mM), MgSO4 (50 tL, 2.8 mM), glycine (50 mM),

cytidine (40 tL, 2.5 mM), GAPDH (2 units), 3-PGK ( 1 unit), and uridine kinase (2

units) in a 1.5 mL microfuge tube and incubated at room temperature for 10 hrs.76 Special

care was taken to minimize contamination by unlabeled phosphate initially present in

some of the reagents used in the reaction. Thus, ADP was freshly prepared and 3-PGK

was dialyzed against 0.5 M Tris-HC1, pH 7.5 buffer to remove the orthopyrophosphate

storage buffer provided by the manufacturer. The reaction solution was filtered through a

microcon filtration unit (Millipore, MWCO 10 kDa) to remove enzymes and P103 CMP

2 was purified from the filtrate using isocratic, anion exchange HPLC (75 mM

NH4HCO3, pH 8.5, 15% methanol, 2 mL/min.). The PO103 CMP containing fractions

were pooled and desalted with Amberlite IR 120-H cation-exchange resin. The solution

was concentrated to dryness in vacuo and resuspended in 300 pL of deionized water.









Synthesis of [1-14C-N-acetyl, 1802] CMP-NeuAc

The [l-14C-N-acetyl, P O02] CMP-NeuAc was synthesized using a method adapted

from that of Ichikawa et al.77 The Pls03 CMP purified from above was concentrated to

dryness in vacuo and used in an enzymatic reaction with ATP (5 imol), PEP

monosodium salt (10 imol), MnCl2 (10 imol), MgCl2 (10 imol), NMK (2 units), PK

(500 units) and 700 iL of HEPES buffer (0.2 M, pH 7.5). The reaction mixture was

incubated for 24 h at 25 C and then filtered through a microcon filtration device

(Millipore, MWCO 10 kDa) to remove enzymes. This step is necessary to prevent the in-

situ recycling ofP1803 CMP and subsequent dilution of 180 labels resulting from the

decomposition of CMP-NeuAc in the following synthetic step. [1-14C-N-acetyl] NeuAc

(10-25 iCi) and CMP-NeuAc synthetase (3 units) were added to the filtrate and the

reaction was incubated at 37 C for 6 h. CMP-NeuAc isotopomers were purified using

isocratic, anion exchange HPLC (75 mM NH4HC03, 15 % methanol, pH 8.0, 2 mL/min).

CMP-NeuAc fractions were pooled, desalted with Amberlite IR 120-H cation-exchange

resin and concentrated in vacuo as previously described.

Synthesis of UMP-NeuAc

CMP-NeuAc was purified on HPLC Mono Q (15 % methanol (v/v), 75 mM

NaHCO3, pH 8.0, isocratic, 2 mL/min). The CMP-NeuAc fraction was collected,

desalted with Amberlite IR-120 H concentrated to dryness in vacuo, and the material

was resuspended in 300 [tL of diH20. The solution of CMP-NeuAc (2 mM, 370 tmols)

was made 1 N with NaNO2 and adjusted to pH 3.5 4.0 with 1 N HC1 while on ice. The

reaction mixture was placed at 4 C for 30 hours. The reaction proceeds to 65 70 %

completion with less than 20 % hydrolysis of starting material in this time. This method






50


was also used to synthesize the various UMP-NeuAc isotopomers by using the

appropriate CMP-NeuAc isotopomers.














CHAPTER 3
PURIFICATION AND KINETIC CHARACTERIZATION OF RECOMBINANT
HUMAN ALPHA (2-3) SIALYLTRANSFERASE IV

Introduction

Complex carbohydrates and polysaccharides are biosynthesized in living systems

via an intricate pathway of membrane-bound and secreted proteins. Many of these

carbohydrate moieties are added by glycosyltransferases to specific biomolecules as the

"finishing touches" during post-translational events in the Golgi apparatus of cells. The

biosynthesis of sialylated glycans are governed by a unique group of enzymes in the

glycosyltransferase family known as sialyltransferases. Sialyltransferase are membrane-

bound proteins that transfer sialic residues (NeuAc) from activated CMP-NeuAc to

specific acceptor oligosaccharides, glycolipids, and glycoproteins during post-

translational modification. The addition of these sialic acid residues assists in the

regulation of a myriad of cellular processes that are of significant biological importance.

This knowledge has prompted the need to study the sialyltransferase family of enzymes

in order to probe the functional roles of sialic acid 'capped' biomolecules required for

many key biological processes.

According to the most recent reviews on the sialyltransferase family, approximately

20 distinct cDNA sequences for sialyltransferases have been identified and cloned from

various mammalian and bacterial sources.29'42'55 The most widely studied

sialyltransferases are the a(2--6) and a(2--3) sialyltransferases from rat liver which are

commercially available in recombinant forms. In this study, three truncated forms of









recombinant human a(2-3) sialyltransferase (h23STGal IV, EC 2.4.99.4), which lacked

the first 61 amino acids coding for the NH2-signal anchor of the open reading frame, were

overexpressed in Spodopterafrugiperda (Sf-9) insect cells using a baculovirus expression

vector. The NH2-terminal signal anchor sequence was replaced in all three enzymes with

a cleavable canine insulin signal peptide to produce a soluble, catalytically active

secreted protein.78'79 The other two recombinant forms of h23 STGal IV contained a

His6x-tag sequence at the N- and C- termini of the sialyltransferase sequence. The three

enzymes were overexpressed, purified via affinity chromatography, and used to conduct

the desired sets of kinetic experiments.

Results and Discussion

Overexpression and Purification of Recombinant Human a(2--3) Sialyltransferase
Isoforms

The cDNA sequence of human a(2--3) sialyltransferase from placenta has been

reported79, which greatly facilitated the primer design for PCR amplification of the

recombinant sialyltransferase isoforms. The cDNA clones encoding the h23STGal IV

gene which lacked the transmembrane domain and the creation of the pFastBacHTa

vector harboring the recombinant h23STGal IV gene were prepared by Dr. Nicole

Horenstein. To produce the enzyme in a more readily purified form, the N-terminal His6x

tag sequence located downstream from the strong polyhedron (pPolh) promoter in the

pFastBacHTa vector was removed via a restriction digest with RsrII and BamHI

endonucleases and replaced with a 114 bp insert containing the sequence coding for the

cleavable canine pancreas insulin signal peptide.78 This allowed the h23STGal IV

enzyme to become a secreted protein. Two other recombinant h23 STGal IV constructs

containing N- and C-terminal His6x tags, additional to the insulin signal peptide, were









prepared in a similar manner by Bronson Anatao.80 These recombinant His6x tag

h23 STGal IV constructs were created to simplify the standard purification procedure.

One construct contained an N-terminal His6x tag located between the insulin signal

peptide and the catalytic domain of recombinant h23 STGal IV, while the other construct

contained a C-terminal His6x tag located at the end of the recombinant h23STGal IV

catalytic domain sequence. Figure 3-1 illustrates the order of the insulin peptide

sequence, recombinant h23 STGal IV gene sequence, and the His6x tag peptide sequence

for each of the three constructs described above. The recombinant pFastBacHTa

vectors harboring the three enzyme constructs were subsequently used to create the

recombinant baculoviruses following the protocols outlined in the BEVS manual supplied

by the manufacturer.81




Ins-h23STGal IV Construct




I -C
CtermHis-h23STGal IV
Construct


I -C

NtermHis-h23STGal IV Construct



Figure 3-1. Diagram of the recombinant h23STGal IV constructs. The insulin signal
peptide sequence is represented in yellow, the truncated h23STGal IV
sequence is represented in blue and the His6x-tag sequence is represented in
purple.









A baculovirus expression system using insect host cells was chosen to overexpress

recombinant h23 STGal IV because of its reported success in overexpressing other

catalytically active enzymes in the glycosyltransferase family.56'82-84 Baculoviruses are

one of the most prominent viral pathogens affecting the insect species. This system is

widely used because of its capability for expressing high levels of recombinant protein

and because of its ability to provide the eukaryotic post-translational modifications

required to produce active enzymes of this type. Several glycosyltransferases have also

been successfully overexpressed in active form with other eukaryotic host cells such as

Saccharomyces cerevisiae and methylotrophic yeast Pichiapastoris.83'8586 Attempts to

overexpress recombinant mammalian sialyltransferases in E. coli resulted in the

production of an insoluble, inactive form of the enzyme.87 This outcome is presumably

because E. coli lacks the capability to sufficiently glycosylate mammalian

sialyltransferases during post translational modification in the Golgi apparatus to allow

for proper folding of the enzymes into an active form. In our hands, attempts to express

the sialyltransferase gene in P. pastoris were also unsuccessful.88

Recombinant baculovirus was generated by first cloning the truncated h23 STGal

IV sialyltransferase constructs into the pFastBacHTa donor plasmids and transforming

the recombinant vectors into competent DH1OBac E. coli (Figure 3-2). Site-specific

transposition was used to insert the recombinant sialyltransferase constructs into bacmid

DNA provided in the E. coli host cells. Recombinant bacmid DNA was purified from E.

coli and used to transfect Sf-9 insect cells. Recombinant baculovirus particles were

generated after incubating the tranfection mixture with the insect cells for several days.

Baculovirus stocks were amplified several times by infecting fresh Sf-9 cultures grown to









a density of 2 x 106 cells/mL until a viral titer of 1 x 107 to 1 x 108 pfu/mL was achieved.

Large scale expression of the recombinant h23 STGal IV was achieved by infecting liter

cultures of Sf-9 insect cells with the amplified recombinant baculovirus stocks.

Recombinant h23 STGal IV was expressed and secreted into the culture medium during

the late phase of the viral life cycle. The average h23 STGal IV activity accumulated in

the cell supernatant of a 1 L Sf-9 cell culture was (2-3 U/L) after 72 hours post infection

with recombinant baculovirus. Purification of recombinant Ins-h23 STGal IV was

achieved using sepharose CDP-hexanolamine affinity column chromatography. Since

sepharose CDP-hexanolamine resin is not commercially available, the CDP-

hexanolamine ligand was synthesized and attached to a CNBr- activated sepharose 4B

resin as previously described in the literature (Figure 3-3).89,90 Table 3-1 summarizes the

data obtained from the purification of a 370 mL scale Sf-9 expression of Ins-h23 STGal

IV on sepharose CDP-hexanolamine. The results shown in Table 3-1, Figures 3-4 and 3-

5 were obtained in collaboration with Jeremiah D. Tipton of the Horenstein laboratory.91












pFAsriBA dcona plaamid
Cimr Genac inlEEI
\,


1 Tranaltnatln ,~Trans Bllon



Recombinant
Dcior Plarnid Competent DH10Bac E.hCela E cca (LacT-)
Containing Recombinant Bacmid

DAY 1 DAYS 2-3
MInlprep oh Ml
S** rntilarWeghitODNA
----------------------------- ----- -- _--______
_t DAYS 5-7 DAY 4
Determine Viral Titer RIctIrant
by Plaque Aaaay lBajculElIu
PacldEe/


"InLFEtClIN Reigent
j -CLLFECTIN Hneant


Inectlln o
InlEBt c;IIE

/


Recommbinant
Bacmid DNA


Recombinant Gene Expreeabn
orViral Amplificatbn




Figure 3-2. General scheme for the generation of recombinant baculoviruses and protein
expression with the BAC-TO-BAC expression system.81



NH2

N
0 0
H2NoI I IP- N 0OO

0- 0-

OH OH



Figure 3-3. Structure of CDP-Hexanolamine affinity ligand synthesized for the
purification of recombinant Ins-h23 STGal IV enzyme.


J^^^ 0uu^yy 0 0









Table 3-1. Recombinant Ins-h23STGal IV purification table.
Specific
Volume Activity Yield Concentration Total Scific
Activity
Step (mL) (mU) (%U) (mg/mL) (mg) (Umg
Step (U/mg)
Crude
370 823 100 0.080 29.6 0.029
Supernatant
Amicon
Amic 80 767 94 0.440 35.2 0.022
Ultrafiltration
CDP-
hexanolamine
28 227 28 0.050 1.40 0.162
pooled
fractions
Concentration 5 159 20 0.241 1.20 0.132


The crude supernatant was concentrated to a volume of 80 mL with an Amicon

Ultrafiltration unit fitted with a polyethersulfone membrane (MWCO 10 kDa) prior to

purification on the sepharose CDP-hexanolamine column. This step was necessary to

reduce the volume of supernatant applied to the affinity column in order to expedite the

purification process. Glycerol (20 % v/v) and Triton CF-54 (0.01 % v/v) were added to

the crude supernatant to help stabilize recombinant h23 STGal IV during concentration

and purification. The crude supernatant was loaded onto the sepharose CDP-

hexanolamine affinity column and the column was washed with three column volumes of

purification buffer containing 300 mM p-lactose prior to eluting the column with a KCl

step gradient. This step was necessary to elute gp64, a predominant baculovirus

membrane glycoprotein responsible for virus-cell fusion, that was found to co-elute with

recombinant h23STGal IV in earlier pilot purification trials.84'88'92 Gp64 is believed to

act as an acceptor substrate for recombinant h23STGal IV. Since p-lactose serves as an

acceptor substrate for recombinant h23 STGal IV, enzyme-gp64 binding interactions can

be disrupted by washing the column with high concentrations of p-lactose.









Purified recombinant h23 STGal IV was obtained in 28 % yield upon elution of the

affinity column with a KC1 step gradient. An activity assay and a protein assay were

performed on select fractions to determine the elution profile of the enzyme from the

affinity column (Figure 3-4). The final yield of purified recombinant h23STGal IV was

20 % with a specific activity of 0.132 U/mg after concentrating the pooled fractions with

an Amicon Ultrafiltration unit. Purified recombinant h23 STGal IV gave three bands in

the SDS-PAGE gel with sizes of- 40 kDa, 37 kDa, and 35 kDa, which represent three

different glycoforms of the enzyme (Figure 3-5). These results are consistent with the

results from a protein sequence analysis of recombinant h23STGal IV of MS experiments

where four potential asparagine N-linked glycosylation sites were identified following a

general protein sequence motif ofN/X/S (X represents any amino acid).91'93'94 Digestion

experiments performed by Jeremiah D. Tipton of the Horenstein laboratory on the

purified recombinant Ins-h23 STGal IV with PNGase, gave one protein band on an SDS-

PAGE gel at 33 kDa, which is the expect size for the deglycosylated protein.91'95'96

After several purification trials, it was found that recombinant Ins-h23 STGal IV

appeared to be sensitive to the concentration step following purification as shown in

Table 3-1 by the decrease in specific activity from 0.162 to 0.132 U/mg. The reason for

this result is unclear, but one explanation may be that the enzyme denatures during the

concentration step. Surfactants such as Triton CF-54 and Tween-80 have been shown to

help stabilize enzyme activity during the purification of other recombinant

sialyltransferases.84'86'97'98 Therefore, Triton CF-54 (0.01 % v/v) was added to the

enzyme buffer to stabilize the enzyme and help minimize the loss of activity during this

step. Additionally, a significant loss in enzyme activity was observed when a












concentrated sample of recombinant Ins-h23 STGal IV was diluted and reconcentrated. In


this case, the enzyme may denature upon dilution and appears to be incapable of


renaturing into active form when concentrated again. As a result, this step was avoided


during the purification of recombinant Ins-h23 STGal IV.


0.3



0.25
-J

E 0.2

o
E

0.15
U
0
C-)
o
- 0.1

a.
0.05


250 mM NaCI


350 mM NaCI


450 mM NaCI


0.045

0.04

0.035

0.03

0.025

0.02

0.015

0.01

0.005

0


2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74
Fraction #


Figure 3-4. Typical elution chromatogram of recombinant Ins-h23STGal IV from a
sepharose CDP-hexanolamine affinity column. Solid squares represent
protein concentration and open diamonds represent activity.






60
1 2 3 4 1 2 3



66 kDa


45
36

29
24 -


20


14



Figure 3-5. 10 % SDS-PAGE of purified recombinant Ins-h23STGal IV. Lane 1, MW
standard; Lane 2, Lane 3, and Lane 4 are 10 ptL, 20 ptL, and 30 pL loads of a
TCA precipitation of purified Ins-h23STGal IV, respectively(left gel). SDS-
PAGE of purified recombinant Ins-h23 STGal IV digested with PNGase.91
Lane 1, MW standard; Lane 2, 30 pL load of a TCA precipitation of purified
Ins-h23STGal IV and Lane 3 is 30 pL load of PNGase digested Ins-h23STGal
IV. The gels were stained with coomassie blue.

Ni2+-NTA affinity chromatography was employed for the purification of

NtermHis- and CtermHis-h23STGal IV enzymes. Glycerol (20 % v/v) and Triton CF-54

(0.01 % v/v) were added to the crude supernatant to stabilize the enzymes during the

purification. A 300 mM p-lactose wash was also performed as described previously to

remove gp64 prior to eluting the column with an imidazole step gradient. Purified

recombinant NtermHis-h23 STGal IV and CtermHis-h23 STGal IV were obtained in 47 %

and 32 % yield, respectively, after eluting Ni2+ the affinity column with an imidazole step

gradient. An activity assay and a protein assay were performed on select fractions to

determine the elution profile of the enzymes from the affinity column (Figure 3-6). The

final yield of purified recombinant NtermHis-h23 STGal IV and CtermHis-h23 STGal IV









was 12 % and 43 % with specific activities of 0.022 U/mg and 0.161 U/mg, respectively

after concentration and dialysis (Tables 3-2 & 3-3). The results shown in Tables 3-2 and

3-3 and in Figure 3-6 were obtained in collaboration with Jeremiah D. Tipton of the

Horenstein laboratory.91 Purified recombinant NtermHis-h23 STGal IV and purified

recombinant CtermHis-h23STGal IV gave three bands in the SDS-PAGE gel with sizes

of- 42 kDa, 41 kDa, and 40 kDa, which represent three different glycoforms of the

enzyme (Figure 3-7). The size of the enzyme glycoforms for the recombinant NtermHis-

and CtermHis-h23 STGal IV are larger than for the recombinant Ins-h23 STGal IV

enzyme glycoforms. Since the addition of the His6x tag peptide sequence in these

enzymes would only add an additional 1000 Da to the size of the recombinant Ins-

h23STGal IV glycoforms, the reason for the larger sizes is unclear. One explanation for

the size discrepancy may be that the presence of the His6x tag alters the addition of glycan

chains during post translational modification to produce recombinant sialyltransferase

with a higher degree of glycosylation.

Table 3-2. Recombinant NtermHis-h23STGal IV purification table.
Specific
Volume Activity Yield Concentration Total Activity
Step (mL) (mU) (%U) (mg/mL) (mg) (U/mg)
Crude
460 419 100 0.240 70.5 0.004
Supernatant
Ni2-NTA
Pooled 30 190 47 0.050 1.50 0.126
Fractions
Concentration
Concentration 4.5 50 12 0.500 2.25 0.022
& Dialysis









Table 3-3. Recombinant CtermHis-h23STGal IV purification table.
Specific
Volume Activity Yield Concentration Total Activity
Step (mL) (mU) (%U) (mg/mL) (mg) (U/mg)
Crude
470 560 100 0.150 71 0.008
Supernatant
Ni2+-NTA
Pooled 50 181 32 0.03 1.5 0.121
Fractions
Concentration
Concentration 5.5 243 43 0.273 1.5 0.161
& Dialysis


The recombinant NtermHis-h23STGal IV enzyme also appeared to be sensitive to

the concentration step after purification as shown in Table 3-2 by the decrease in specific

activity from 0.126 to 0.022 U/mg. Concentration and dialysis of this enzyme without

the presence of a surfactant such as Triton CF-54, resulted in a total loss of enzyme

activity. Therefore, as for the recombinant Ins-h23STGal IV, Triton CF-54 (0.01 % v/v)

was added to the enzyme buffer prior to concentration and dialysis. The recombinant

CtermHis-h23 STGal IV enzyme, however, did not lose activity during the concentration

and dialysis step as seen by the increase in specific activity from 0.121 to 0.161 U/mg.

This may be due to the overall stability of the recombinant CtermHis-h23 STGal IV

enzyme in comparison to the recombinant NtermHis-h23STGal IV. It is reasonable to

consider that the position of the His6x tag on the N- or C-terminus may change the protein

folding structure in a manner that would make recombinant NtermHis-h23 STGal less

stable than CtermHis-h23STGal IV during concentration. Additionally, the position of

the N-term His6x tag could also interfere with the addition of glycan chains during post

translational that are essential for enzyme stability and activity. This effect has been

observed for other His6x-tagged glycosyltransferases.83'99 The specific activity of the







63


recombinant CtermHis-h23 STGal IV enzyme was similar to that of recombinant Ins-

h23 STGal IV, but the yield was higher than for the recombinant Ins-h23 STGal IV.


0.4 0.016

0.35 -- 0.014

S0.3 0.012

0.25 \ 0.01

0.2- 0.008

0.15 I 0.006

0 0.1 -- 0.004

0.05 0.002


0 12 30 38 48 57 67 75 76 77 78 79 80 81 82 83 84 86
Fraction #


Figure 3-6. Typical elution chromatogram of recombinant NtermHis-h23STGal IV and
CtermHis-h23 STGal IV from a Ni2+-NTA affinity column. Solid squares represent
protein concentration and open diamonds represent activity.

1 2 3

66 kDa
O*Ift

45 v":
36 u

29
24


20

14

Figure 3-7. 10 % SDS-PAGE gel of purified recombinant CtermHis-h23STGal IV
NtermHis-h23STGal IV. Lane 1, TCA precipitation of purified CtermHis-
h23STGal IV; Lane 2, MW Standard; Lane 3, TCA precipitation of purified
NtermHis-h23STGal IV. The gel was stained with coomassie blue.









Kinetic Characterization of Recombinant h23STGal IV Isoforms

The kinetic parameters obtained for the three recombinant isoforms of human

a(2--3) sialyltransferase using CMP-NeuAc and a-lactose as the donor-acceptor

substrate pair were estimated by fitting the initial velocity kinetic data to the Michaelis-

Menten equation using a least squares analysis in Sigma Plot ver. 9.0 (Figures 3-8 3-

10). Recombinant Ins-h23STGal IV, NtermHis-h23STGal IV, and CtermHis-h23STGal

IV enzymes were concentrated a second time to specific activities of 0.05, 0.012, and

0.042 U/mg, respectively as determined by activity assays prior to their use in steady

state kinetic experiments in order to provide concentrated enough enzyme samples. The

measured kinetic parameters were similar to those obtained for a wild type a(2--3)

sialyltransferase purified from human placenta (Table 3-4).97 The CMP-NeuAc Km

values for NtermHis-h23STGal IV and Ins-h23STGal IV were 82 5 tM and 74 8 aM,

respectively. These Km values were on the same order of magnitude as the wild type

a(2--3) sialyltransferase from human placenta and were comparable to the literature Km

value of 74 aM obtained for recombinant rat liver a(2--3) sialyltransferases expressed

from insect cells.56 However, the CMP-NeuAc Km value obtained for CtermHis-

h23STGal IV was -3.5 fold higher than for the NtermHis-h23STGal IV and Ins-

h23STGal IV enzymes. This reason for this is unclear and literature kinetic data on other

C-terminal His6x tagged sialyltransferases have not been reported. However, a few

hypotheses for this result are that the addition of the CtermHis6x tag on recombinant

h23 STGal IV either interferes with the binding site for CMP-NeuAc or that it changes the

structural fold of the enzyme in a manner that alters CMP-NeuAc binding. The a-lactose

Km values for all three recombinant isoforms of human a(2--3) sialyltransferase were

similar to each other and for the wild type enzyme within experimental error.97









Table 3-4. Kinetic parameters for sialyltransferase isoforms.

CMP- NeuAc a-Lactose
Recombinant Km Vmax Km Vmax
Enzyme ([iM) (tmol-min-mg-1) (mM) (tmol-min-mg 1)
WTh23STGal IV7 63 -220 40 0.12 0.01
Ins-hST3Gal IV 82 5 0.072 + 0.002 171 + 18 0.072 + 0.004
NtermHis-hST3Gal
NtermHis-h74 + 8 0.008 + 0.0003 155 + 14 0.020+ 0.001
IV
CtermHis-hST3Gal
CtermHs-h267 20 0.041 + 0.002 158 + 11 0.041 0.002
IV


The Vmax value obtained for Ins-ST was 0.072 0.004 [tmol/(min-mg) which is

slightly lower than the 0.12 0.01 [tmol/(min-mg) Vmax reported for the wild type human

placenta a(2--3) sialyltransferase when CMP-NeuAc and a-lactose were used as the

donor-aceptor substrate pair.97 This difference in Vmax values was not large and reflects

the observed loss in enzyme specific activity after concentrating the enzyme. The Vmax

values obtained for CtermHis-h23STGal IV and NtermHis-h23STGal IV were 3-6 fold

lower than that of the wild type sialyltransferase. However, these values were similar to

the specific activities obtained for these enzymes after the second concentration step

described above. If the recombinant enzymes were not concentrated after purification,

than their Vmax values may have been more comparable to that reported for wild type

h23STGal IV enzyme.












0.06


0.05


0.04


0.03


3 0.02


0.01


0.00 Experimental
-- Predicted


0 50 100 150 200 250 300 350

[CMP-NeuAc], |aM



0.05



0.04 -



0.03



0.02



S0.01



0.00 Experimental
Predicted



0 50 100 150 200 250 300 350

[a-lactose], mM


Figure 3-8. Michaelis-Menten plots for recombinant Ins-h23STGal IV with with varied
[CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-lactose]
and constant [CMP-NeuAc] (bottom panel).












0.007 -


0.006 -


0.005


0.004 -


0.003 -


0.002 -


0.001


0.000 -


0 50 100 150 200 250 300 350


[CMP-NeuAc], |iM


0.016

0.014 -

0.012 -


0.010 -

0.008 -

0.006 -

0.004 -

0.002 -

0.000 -


0 50 100 150 200 250 300 350


[a-lactose], mM




Figure 3-9. Michaelis-Menten plots for recombinant NtermHis-h23STGal with varied
[CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-lactose]
and constant [CMP-NeuAc] (bottom panel).


Experimental
-- Predicted


Experimental
- Predicted












0.030


0.025


0.020


0.015


0.010


0.005


0.000


0 100 200 300 400


[CMP-NeuAc], iM


0.030


0.025


0.020


0.015


0.010


0.005


50 100 150 200 250 300 350


[a-lactose], mM


Figure 3-10. Michealis-Menten plots for recombinant CtermHis-h23STGal IV with
varied [CMP-NeuAc] and constant [a-lactose] (top panel) and with varied [a-
lactose] and constant [CMP-NeuAc] (top panel) and for with (lower panel).


Experimental
-- Predicted


Experimental
-- Predicted









Conclusions

Overall, the recombinant NtermHis-h23 STGal IV and CtermHis-h23 STGal IV

enzymes were significantly easier to purify than the recombinant Ins-h23 STGal IV

enzyme. Furthermore, the Ni2+-NTA resin can readily obtained from commercial sources

thus allowing for larger scale expressions of recombinant enzyme to be purified more

rapidly. This is not the case for the recombinant Ins-h23 STGal IV enzyme which requires

a tedious synthesis of a sepharose CDP-Hexanolamine affinity resin prior to purification.

The protein expression scale is also limited by the amount of affinity resin synthesized.

Of the three enzymes, recombinant CtermHis-h23STGal IV enzyme gave the best results

with the highest purification yield and specific activity. This enzyme also had

comparable kinetic parameters to the recombinant Ins-h23 STGal IV enzyme except for

the slightly larger Km value obtained for CMP-NeuAc.

Experimental

Materials and Methods

Reagents and buffers were purchased from Sigma and Fisher and used without

further purification. The restriction enzymes, E. coli strains JM109 & ER2925, Klenow,

Large Fragment (DNA Polymerase I), and T4 DNA ligase were purchased from New

England Biolabs. Shrimp Alkaline Phosphatase (SAP) was purchased from Roche

Molecular Biology. The Wizard Plus Minipreps Kit and dNTPs were purchased from

Promega. The QIAquick Nucleotide Removal Kit and QIAquick Gel Extraction Kit

were purchased from Qiagen. The BCA Protein Assay Kit was purchased from Pierce.

The BAC-TO-BAC Baculovirus Expression System (BEVS), Spodopterafrugiperda (Sf

9) insect cells, BACPACKTM Baculovirus Rapid Titer Kit, and DH10 BAC E. coli

competent cells were purchased from Invitrogen. Human placental cDNA was obtained









from Clontech. Primers for cloning and PCR analysis were obtained from Integrated

DNA Technologies. The protocol for recombinant virus preparation is found in

Invitrogen's instruction manual for BEVS version D April 6, 2004

(www.invitrogen.com). The sepharose CDP-hexanolamine affinity column was prepared

as per the literature.89' 100 N-acetyl neuraminic acid (NANA) aldolase [EC 4.1.3.3] used

in the synthesis of [9-3H] neuraminic acid (NeuAc) was cloned, overexpressed in E. coli

and purified according to literature procedures.61'63'72 The E. coli expression plasmid

pWV200B haboring 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 for sialyltransferase activity

determination and kinetic experiments were analyzed with a Packard 1600 TR liquid

scintillation analyzer. DNA sequencing was performed at the University of Florida ICBR

DNA sequencing core.

Preparation of pFastBacHTaInsulin/h23STGal IV (Ins-h23STGal IV)

MAX EFFICIENCY DH10BAC E. coli cells transformed with the pFASTBAC

plasmid containing the recombinant h23 STGal IV gene were previously prepared in our

lab by Dr. Nicole Horenstein. A canine insulin construct located upstream from the

h23STGal IV was also cloned into the recombinant plasmid to allow for secretion of the

enzyme into Sf-9 insect cell media. The construct, pFastBacHTaInsulin/h23 STGal IV,

was submitted for DNA sequencing to confirm the presence of the canine insulin

secretion peptide insert. The protocol for baculovirus preparation is found in the manual

BAC-To-BAC Baculovirus Expression Systems provided by Invitrogen.81 The

recombinant bacmid DNA was purified from E. coli using mini-prep proceduressl and









subsequently used in the transfection of Sf-9 insect cells to produce recombinant

baculovirus particles.

Preparation of pFastBacHTalnsulin/NtermHis6x-tag-h23STGal IV Plasmid
(NtermHis-h23STGal IV)

A fresh glycerol stock of E. coli JM109 cells harboring the

pFastBacHTaInsulin/h23STGal IV plasmid was used to inoculate a 5 mL luria broth

culture supplemented with 100 [tg/mL ampicillin. The culture was grown overnight at 37

C (200 rpm) and the plasmid was subsequently isolated using the Wizard Plus

Minipreps kit. The isolated 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 BamHI digested plasmid was digested for 6 hours at 37 C with Stul.

After thermally inactivating Stul at 65 C for 20 min., the doubly digested plasmid was

agarose gel purified using the QIAquick Gel Extraction Kit. The 5' ends generated by

digestion were dephosphorylated using SAP.

The entire NtermHis6x tag insert was created by allowing two complimentary

primers to anneal and then be extended by Klenow, Large Fragment. Primers were

mixed in equal volumes (2.5 pL) of NtermHISForUpper (5'-

GGTAGGCCCTGGCCATTAAGCGGATGCTGGAGATGGGAGCTATCAAGAACCT

CACGTCC-3'), (50 tiM, 0.80 tg/tiL) and NtermHisRev_Lower (5'-

AGCAGGCCTTGCTCTCTGCCTCACCCTGGAGGAGGCACGGCTCCTTCTTCTCG

CC-3'), (50 IM, 0.81 pg/iL) 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 11.16 [L deionized water, 420 iM of a dNTP mixture, 10 mM Tris-HCl pH

7.5, 5 mM MgC12, 7.5 mM dithiothreitol, and 5 U of Klenow, Large Fragment. This









reaction mixture was then allowed to incubate in a 25 C water bath for 80 minutes. After

inactivating the polymerase at 75 C for 20 minutes, the mixture was cooled to room

temperature and then doubly digested with BamHI and Stul at 37 C for 9 hours.

Following heat inactivation of Stul, a clean-up of the enzymatic reaction was performed

using the QIAquick Nucleotide Removal Kit. The NtermHis6x tag insert was then

ligated into the BamHI/StuI sites of the pFastBacHTaInsulin/h23 STGal IV vector. The

new construct, pFastBacHTalnsulin/NtermHis6x-tag-h23 STGal IV, was submitted for

DNA sequencing to confirm the presence of the NtermHis6x tag insert. Isolation of

bacmids and generation of baculovirus stocks followed the BEVS protocol.81

Preparation of pFastBacHTalnsulin/CtermHis6x-tag-h23STGal IV Plasmid
(CtermHis-h23STGal IV)

The pFastBacHTaInsulin/h23 STGal IV 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 thermally inactivating Eco0109I and Xhol at 65 C for 20 min., the doubly

digested plasmid was agarose gel purified using the QIAquick Gel Extraction Kit. The

5' ends generated by digestion were dephosphorylated using SAP.

The CtermHis6x tag insert was created following a similar procedure to that of the

NtermHis6x tag insert. Primers were mixed in equal volumes (2.3 tiL) of

CtermHISForUpper (5'-

GGTAGGCCCTGGCCATTAAGCGGATGCTGGAGATGGGAGCTATCAAGAACCT









CACGTCC-3'), (50 [iM, 0.882 [g/IL) and CtermHisRevLower (5'-

AGCCTCGAGTTAGTGATGGTGATGGTGATGACCGCCGAAGGACGTGAGGTTC

TTGATAGC-3'), (50 [iM, 0.917 [g/IL) 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 11.5 [L deionized water, 463 [M of a dNTP mixture, 10

mM Tris-HCl pH 7.5, 5 mM MgC12, 7.5 mM dithiothreitol, and 5 U of Klenow, Large

Fragment. This reaction mixture was then allowed to incubate in a 25 C water bath for 2

hrs. After inactivating the polymerase, the mixture was cooled to room temperature and

then digested with Eco0109I for 16 hours at 37 C. The enzymatic reaction was cleaned

up using the QIAquick Nucleotide Removal Kit. The clean mixture was digested with

Xhol at 37 C for 8 hours. Xhol was then thermally inactivated. The CtermHis6x tag

insert was then ligated into the Eco01091/XhoI sites of the

pFastBacHTaInsulin/h23STGal IV vector. The new construct,

pFastBacHTalnsulin/CtermHis6x-tag-h23 STGal IV, was submitted for DNA sequencing

to confirm the presence of the CtermHis6x tag insert. Isolation of bacmids and generation

of baculovirus stocks followed BEVS protocol.81

Amplification of Recombinant Baculovirus Plasmids

Recombinant baculovirus stocks, as prepared following the BEVS protocol, were

amplified four times until a titer of 1 x 107 3 x 108 pfu/mL was obtained.81 For

amplification and expression, cell cultures (50 mL) containing 2 x 106 cells/mL in 250

mL borosilicate shaker flasks were infected with 3 mL of 1.7 x 108 pfu / mL Ins-

h23STGal IV, 3 mL of 2.1 x 108 pfu / mL NtermHis-h23STGal IV, or 3 mL of 2.7 x 108

pfu / mL CtermHis-h23STGal IV.









Expression and Purification of Ins-h23STGal IV

A 1 L culture of Sf-9 insect cells (20, 250 mL shaker flasks containing 50 mL of

culture each) at concentrations of 2 x 106 cells/mL were infected with 3 mL of amplified

baculovirus stocks( -2 x 108 pfu/mL) harboring the recombinant sialyltransferase

expression constructs. The cultures were incubated for 70 hrs at 27 C after which the

cultures were combined and centrifuged at 14,000 rpm, 4 C, for 30 min to pellet the

cells. The supernatant was harvested and concentrated to 50 -100 mL using a 200 mL

Amicon Ultrafiltration unit equipped with a polyethersulfone membrane (MWCO 10

kDa). This concentration step was only done for the Ins-h23STGal IV expressions in

order to expedite the purification process on the CDP-hexanolamine affinity column.

The concentrated supernatant was brought to 20 % (v/v) glycerol and 0.01 % (v/v)

Triton CF-54 and applied to the sepharose CDP-hexanolamine affinity column (1.7 x 12

cm) previously equilibrated with purification buffer (50 mM MES, pH 6.8 buffer

containing 300 mM p-lactose, 20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF-54) at 4

C. The column was then washed with at least three column volumes of purification

buffer to remove gp64 glycoprotein from recombinant Ins-h23 STGal IV. Ins-h23 STGal

IV was purified from the supernatant by eluting the column with a KC1 step gradient of

50 mM, 250 mM, and 400 mM KC1 in 50 mM MES, pH 6.8 buffer with 20 % (v/v)

glycerol and 0.01 % (v/v) Triton CF-54. Fractions were analyzed for protein and activity

using the Bradford assay method the published activity assay.56'73 Fractions containing

activity were pooled, dialyzed, and concentrated. Bradford and bicinchoninic acid (BCA)

assays were used to estimate protein concentration.73'101









Expression and Purification of NtermHis-h23STGal IV and CtermHis-h23STGal IV
Isoforms

A 1 L culture of Sf-9 insect cells (20, 250 mL shaker flasks containing 50 mL of

culture each) at concentrations of 2 x 106 cells/mL were infected with 3 mL of amplified

baculovirus stocks( -2 x 108 pfu/mL) harboring the recombinant sialyltransferase

expression constructs. The cultures were incubated for 70 hrs at 27 C after which the

cultures were combined and centrifuged at 14,000 rpm, 4 C, for 30 min to pellet the

cells. The clarified supernatant was brought to 20 % (v/v) glycerol and 0.01 % (v/v)

Triton CF-54 and loaded onto a Ni2+-NTA column (2 x 10 cm) previously equilibrated

with 50 mM MES, pH 6.8 buffer containing 5 mM imidazole, 100 mM KC1, 20 % (v/v)

glycerol, and 0.01 % (v/v) Triton CF-54 at 4 C. After loading the supernatant onto the

column, the column was washed with at least three column equivalents of 50 mM MES,

pH 6.8 buffer containing 300 mM p-lactose, 5 mM imidazole, 100 mM KC1, 20 % (v/v)

glycerol, and 0.01 % (v/v) Triton CF-54, followed by a 5 column equivalent wash with

50 mM MES, pH 6.8 buffer containing 5 mM imidazole, 100 mM KC1, 20 % (v/v)

glycerol, and 0.01 % (v/v) Triton CF-54. NtermHis-h23STGal IV and CtermHis-

h23STGal IV enzymes were eluted from the column using imidazole step gradient of 50

mM, 75 mM, and 120 mM imidazole in 50 mM MES, pH 6.8 buffer with 100 mM KC1,

20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF-54. Fractions were analyzed for

protein using the Bradford assay method and for sialyltransferase activity using the

published assay.56'73 Fractions containing activity were pooled, dialyzed, and

concentrated. Bradford and BCA assays were used to estimate protein

concentration.73101









Sialyltransferase Enzyme Activity Assays

All activity assays reported for the recombinant h23 STGal IV isoforms were

performed using methods described by Paulson et al.56'90 The assay contained a mixture

of [9-3H] CMP-NeuAc (100-170 pM, 20,000 cpm, s.a.= 4-7 pCi/[mol) and 235 mM a-

lactose in 50 mM MES, pH 6.8 buffer containing 0.01 % (v/v) Triton CF-54, and 1

mg/mL BSA. A 10 [L aliquot from a selected sialyltransferase containing sample was

incubated with 10 [L of the [9-3H] CMP-NeuAc/a-lactose mixture for the appropriate

amount of time to limit the consumption of CMP-NeuAc to < 10 %. The reaction

mixture was quenched with 500 [L of 5 mM inorganic phosphate buffer, pH 6.8 and then

applied to 2.5 cm Dowex 1 x 8, 200 mesh (P042-) mini-columns equilibrated with 5 mM

Pi, pH 6.8.56,102 Reactions were eluted with 3.5 mL of 5 mM Pi buffer, pH 6.8 into liquid

scintillation vials. Liquid scintillation vials were counted for 5 min., and all tubes were

cycled through the counter 4-5 times to obtain an accurate measurement of the amount of

radioactive product. The definition of a unit of activity is the amount CMP-NeuAc

converted to sialyl-lactose per minute. The activity reported was obtained by correcting

the observed velocities with the obtained kinetic parameters and the concentrations of

substrate employed, as fit to the following bi-substrate equation:


v = (vmax x [A] x [B]) / (KmAx KmB + KmA x [B] + KmB[A] + [A] x [B]) eq. 3-1


Steady State Kinetics for Recombinant h23STGal IV Isoforms

The kinetic parameters for the recombinant human a(2-*3) sialyltransferase

isoforms with CMP-NeuAc and a-lactose as the donor-acceptor pair were estimated by

varying the CMP-NeuAc concentration while holding a-lactose at a near-saturating

concentration, and by vary the a-lactose concentration and holding CMP-NeuAc at a









near-saturating concentration. Reactions were conducted at 37 C for 12 min in 50 mM

MES, 0.2 mg/mL BSA, 0.05 % (v/v) Triton CF-54, pH 7.5 buffer with a final volume of

100 pL. Each reaction contained 100,000 cpm of [9-3H] CMP-NeuAc diluted to the

required specific activity. The apparent Km value for CMP-NeuAc was obtained by using

30-300 pM of CMP-NeuAc with 0.5 mM of a-lactose, and for a-lactose, using 50-500

mM of a-lactose with 300 pM CMP-NeuAc. The reactions were initiated by the addition

of 8-12 pg of recombinant enzyme. Aliquots of 20 pL were removed at 3, 6, 9, and 12

min and quenched in 500 pL of 5 mM phosphate, pH 6.8 buffer. The product was

quantified using the Dowex column methodology as described above.56'90 Samples were

counted in a liquid scintillation counter for 10 min., and all tubes were cycled through the

counter 6-10 times to obtain an accurate measurement of the amount of radioactive

product. The kinetic parameters obtained for the three recombinant isoforms of human

a(2--3) sialyltransferase using CMP-NeuAc and a-lactose as the donor-acceptor

substrate pair were estimated by fitting the kinetic data to the Michaelis-Menten equation

using a least squares analysis in Sigma Plot ver. 9.0.

Michaelis-Menten Equation:

v = Vmax [S]/Km+ [S] eq. 3-2














CHAPTER 4
KINETIC ISOTOPE EFFECT STUDIES ON RECOMBINANT
SIALYLTRANSFERASES

Introduction

Kinetic isotope effects (KIEs) serve as a valuable technique to elucidate the

transtition-state structure of organic and enzymatic reactions. Knowledge about the

transition-state structure of an enzyme catalyzed reaction is significant in that it offers

detailed information about the reaction mechanism in a way that enzyme crystal

structures are unable to provide. In this study, a series CMP-NeuAc and UMP-NeuAc

donor substrate radioisotopomers were synthesized to probe the mechanism of sialyl

transfer using KIE experiments. By measuring a variety of KIEs at different positions on

the donor substrate, one can gain valuable information about various aspects of the

transition-state structure which will assist in acquiring mechanistic information for the

sialyltransferase catalyzed reaction. The dual-label competitive method was used to

measure the KIEs for the various donor substrate radioisotopomers with recombinant

human placental a(2--3) sialyltransferase, recombinant rat liver a(2-*3)

sialyltransferase, and recombinant rat liver a(2--6) sialyltransferase. The data from these

experiments will provide an increased understanding of the mechanism of glycosyl

transfer with regard to interactions at the phosphate leaving group via 180 isotopic

substitution at the glycosidic O and non-bridging phosphate oxygen atoms for a family of

enzymes. Additionally, this data may prove useful toward the development of new









sialyltransferase inhibitors that are based on the transition-state structure of the donor

substrate.

Kinetic Isotope Effect (KIE) Background

Isotope Effect Theory

Isotope effects are simply explained as the perturbation of the reaction rate (kinetic

isotope effect, KIE) or of the reaction equilibrium constant (equilibrium isotope effect,

EIE) resulting from an isotopic substitution at one position in a reaction molecule. While

the description of isotope effects seems straight forward, the interpretation of isotope

effects can be quite complicated. In general, isotope effects are expressed as a ratio of

rate constants where the rate constant for the light molecule (kL) is divided by the rate

constant for of the heavy (kH).

The initial theoretical calculations for isotope effects and their use to investigate

chemical reaction mechanisms was published by Bigeleisen and Mayer in 1947.1 The

work presented by Bigeleisen and Mayer on the calculation of equilibrium isotope effects

established the foundation for the field of isotope effects. The Bigeleisen equation for

equilibrium isotope effects is shown in equation 4-1.



K1/K2 = MMI EXC- ZPE eq. 4-1



For this equation, K1 and K2 represent the equilibrium constants for the two isotopic

species being measured. The MMI term includes moments of inertia and the combined

molecular mass. The EXC term accounts for the isotope effect on the molecules if they

exist in excited vibrational states. Lastly, the ZPE term denotes the isotope effect

resulting from differences in vibrational zero-point energy.2











The Bigeleisen equation for equilibrium isotope effects was later extended to include

isotope effects on reaction rates (KIEs). Kinetic isotope effects are related to equilibrium

isotope effect theory via the transition-state theory. The basis of the theory is centered on

the supposition that the reactant and the transition-state are in equilibrium. The rate of

the reaction can then be derived from the transition-state theory as the difference in free

energy when going from the ground state of the reaction to the transition-state as

expressed in equation 4-2.



k = (kT/h)exp(-AGV/RT) eq. 4-2



In this equation k represents Boltzmann's constant, h denotes Planck's constant, T is the

temperature in Kelvin, -AG corresponds to the activation free energy, and R is the ideal

gas constant.

Under the assumptions provided in the transition-state theory, the Bigeleisen

equation can be applied to kinetic isotope effects with a slight modification to the normal

3N-6 vibrational modes for the ground-state EXC and ZPE terms. In the transition state,

one normal mode turns into a reaction coordinate mode with an imaginary frequency, vL.

The reaction coordinate mode accounts for the motion along the reaction coordinate since

the transition-state can convert either back to reactants or forward to products. Thus,

transition states have 3N-7 frequencies with one imaginary frequency.1 The equation for

KIE is expressed in eq. 4-3. The mathematical expansions of the individual terms in the






81

Bigeleisen equation for KIE are shown in Figure 4-1. In this equation u is equal to
hv/KT.


KIE = vYV]H MMI EXC ZPE eq. 4-3




MI M2 LAB C A2B/C2
MMI= IxM1 ABIC I A1B1CI
M M, ABC ABC
2 2 2 2

3-U(1)
EXC= 7 1 e'(2) 1- e
-ut 7 u"7(2)
S1-e I 1 -e

3N -7 (e(2)2 3N-6 (1)
ZPE = e(1/2)u(1) (2)
e i e

Figure 4-1. Expanded terms of the Bigeleisen equation.
Isotope effects, whether they are EIE or KIE, generally originate from the ZPE

term. Since molecules of biological interest are normally large, the contribution to the

isotope effect from the translational, rotational, and excited vibrational energies is usually

small and therefore insignificant. Consequently, the zero-point energy typically becomes

the dominating factor of the isotope effect. Most of the isotope effect stems from the

differences in zero-point energy between isotopomers of the initial and final states when

either going from reactant to product or in going from the ground-state to the transition-

state. When considering kinetic isotope effects, the contribution to the isotope effect

from the zero-point energy factor is determined by the change in force constants to the









isotopically substituted atom upon moving from the ground-state to the transition-state.

For example, if the bond to the isotopically substituted atom becomes looser in the

transition-state, then a normal (>1) KIE with be observed. In this case, the zero-point

energy term decreases because the force constant of this particular bond diminishes in the

transition-state. This change to the isotope effect is illustrated in Figure 4-2 where the

potential energy well becomes wider in going from the ground-state to the transition-state

on the reaction coordinate. Conversely, if the bond to the isotopically substituted atom

becomes tighter in the transition-state, then an inverse (<1) KIE will be observed. The

zero-point energy term in this situation increases because the force constant of this bond

becomes larger in the transition-state. This is depicted in Figure 4-3 where the potential

energy well becomes narrower in going from the ground-state to the transition-state on

the reaction coordinate. Lastly, if the bond to the isotopically substituted atom does not

change upon going from the ground-state to the transition-state, than a KIE of unity will

be observed. These general changes to the isotope effect may be summarized by the first

rule in isotope chemistry which states that the light isotopic molecule prefers a looser

bonding state where the restrictions to vibration are lower.3

Since isotope effects arise from changes in force constants and zero-point energies

on the bond attached to the isotopically substituted atom, they are therefore local effects

and can only extend a couple of bond distances. As a result, isotope effects are

categorized into different types depending on the location of the isotopic substitution to

the reaction center. The two major types of isotope effects are primary and secondary

effects which will be discussed in greater detail below.









Transition-state


Ground-state


Figure 4-2. Free energy diagram depicting the looser potential energy wells in the
transition-state resulting in a normal (>1) isotope effect from Lowry et al.4


Transition-state


Ground-state


Figure 4-3. Free energy diagram depicting the looser potential energy wells in the
transition-state resulting in an inverse (<1) isotope effect from Lowry et al.4









Primary Isotope Effects

Primary kinetic isotope effects occur when the isotopically substituted atom

experiences either bond formation or bond cleavage in the transition-state. In general,

primary isotope effects are larger than secondary isotope effects because the bond

changes taking place at the site of isotopic substitution translate into higher changes in

ZPE. The measurement of carbon heavy atom primary isotope effects for molecules of

biological interest is commonly achieved by isotopically substituting a carbon reaction

center with a 14C or 13C atom. The size of the primary isotope effect is contingent upon

the type of symmetry that occurs around the reaction center in the transition-state. If the

reaction center is symmetrical at the transition-state, than the symmetric stretching

vibration in the transition-state will be the major contributor to the primary isotope

effect.4 This is explained when one considers the different vibrational modes between the

ground-state and the transition-state. The vibrational modes that exist in the reactant are

the bending and stretching vibrations. In the transition-state, the vibrational modes

include the bending vibration, the symmetric stretching vibration, and the vibrational

mode that becomes the reaction coordinate. If the bending vibrations are similar between

the ground-state to the transition-state, these vibrations cancel to leave only the

symmetric stretching vibration. As a result, the symmetric stretching vibration becomes

the main contributor to the primary isotope effect. The symmetric stretching vibration is

depicted in Figure 4-4 where the isotopically substituted atom "C" is being moved from

"A" to "B". In a symmetrical transition-state, atom C is motionless and the symmetric

stretching vibration involves only A and B. The isotopically substituted atom, therefore,

does not significantly contribute to the vibrational frequency in the transition-state

because atom C will not have a zero-point energy difference. Symmetrical transition-









states occur in associative SN2 reactions. Associative SN2 reactions normally have larger

primary 14C isotope effects with values in the range of 1.08-1.15.

On the other hand, in an asymmetrical transition-state, the isotopically substituted

atom contributes more to the primary isotope effect because it keeps some of the

symmetrical stretching vibrational frequency. In this case, the symmetrical stretching

vibrational frequency attributed to atom C will partially cancel the zero-point energy

difference in the ground-state to give a decreased kinetic isotope effect.4 Asymetrical

transition-states are characteristic for dissociative SN1 reactions. Dissociate SNi-like

reactions typically have smaller primary 14C kinetic isotope effects with values in the

range of 1.02-1.05.5

A B C





Symmetric Transition-state







Asymmetric Transition-state


Figure 4-4. The symmetric stretching vibration mode in the transition-state of transfer
reactions. C represents the isotopically substituted atom that is transferred
between A and B.

Secondary Isotope Effects

Secondary isotope effects occur when a bond to the isotopically substituted atom is

neither broken or formed in the transition-state. Secondary isotope effects are typically

smaller than primary isotope effects and they do not directly report on bond breaking and




Full Text

PAGE 1

HEAVY ATOM AND HYDROGEN KINETI C ISOTOPE EFFECT STUDIES ON RECOMBINANT, MAMMALIAN SIALYLTRANSFERASES By ERIN E. BURKE 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 ERIN E. BURKE

PAGE 3

This work is dedicated to my husband, my parents, and to al l of my family and friends who have supported me in this endeavor.

PAGE 4

iv ACKNOWLEDGMENTS My sincerest gratitude goes to my advisor, Dr. Nicole Horenstein, for her guidance, support, and patience during my graduate studies and the course of this project. I would also like to recognize and thank the member s of my supervisory committee, Dr. Nigel Richards, Dr. Jon Stewart, Dr. Ronald Castellano, and Dr. Art Edison for their suggestions and support on this projec t over the last five years. Special thanks go to the past and present Horenstein group members: Jen, Jeremiah, Fedra, Andrews, Mirela, and Ji ngsong for their company and help throughout the years. I am especially grateful to Je n, Jeremiah, and Fedra for their friendships and steadfast support. Additionally, I wish to thank all of my co lleagues in the Biochemistry Division for their advice and for their genero sity in allowing me to borrow equipment and reagents when needed. I would also like to express my appreciation to JoAnne Jacobucci and Romaine “Sugar Momma” Hughes for their administrative help and for keeping me on an endless sugar high with their bottomless candy bowls. A heartfelt thank you goes to my parents, Paul and Elizabeth Ringus, my brother, Adrian, and my best girlfriends Georgia, Felicia, Patricia, and Cerissa, for their support, encouragement, and unwavering love throughout this endeavor. A sincere thanks also goes to my relatives living in Melbourne, FL, for their love, encouragement, and for their warm hospitality when I needed a break from lab work. I am also deeply indebted to my wonderful husband, Andy, for his steadfast patience, love, encouragement, support, and friend ship at all times. I am so thankful to

PAGE 5

v have met him in graduate school and I look fo rward to our future together. Without his support, none of my accomplishments in graduate school would have been possible. Finally, I would like to express my eternal gratitude to my savior, Jesus Christ, and to God. Thank you especially for answering al l of my prayers and fo r bestowing me with the talents to pursue a career in the fiel d of Chemistry. Their unwavering support has made this difficult journey worthwhile. Thus, “. . with God all th ings are possible.”( Matthew 19:26).

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES.............................................................................................................x ABSTRACT.....................................................................................................................xi v CHAPTER 1 INTRODUCTION........................................................................................................1 Sialic Acids................................................................................................................... 1 Glycosyltransferases and Glycosidases........................................................................7 Fucosyltransferases.....................................................................................................11 Sialyltransferases........................................................................................................13 Sialyltransferase Inhibitors.........................................................................................19 2 SYNTHESIS AND CHARACTER IZATION OF SUBSTRATES...........................24 Introduction.................................................................................................................24 Results and Discussion...............................................................................................24 Synthesis of CMP-NeuAc isotopomers...............................................................24 Synthesis of [1-14C -N-acetyl 2-18O] CMP-NeuAc.............................................27 Synthesis of [1-14CN-acetyl P18O2] CMP-NeuAc.............................................30 Synthesis of UMP-NeuAc Isotopomers..............................................................37 Experimental...............................................................................................................41 Materials..............................................................................................................41 Instrumental.........................................................................................................42 Synthesis of [3H-N-acetyl] ManNAc...................................................................43 Synthesis of [1-3HN-acetyl ] NeuAc and [1-3HN-acetyl ] CMP-NeuAc............44 Cloning, Overexpression and Purifica tion of N-acetylneuraminic Acid Aldolase [EC 4.1.3.3].......................................................................................45 Overexpression of CMP-NeuAc Synthetase [EC 2.7.2.43]................................45 Cloning, Overexpression and Purifica tion of Uridine Kinase [EC 2.7.1.48]......46 Synthesis of 75 atom % [1-14C-N-acetyl, 2-18O] CMP-NeuAc...........................47 Synthesis of KH2P18O4........................................................................................47 Synthesis of P18O3 CMP......................................................................................48

PAGE 7

vii Synthesis of [1-14C-N-acetyl, P18O2] CMP-NeuAc.............................................49 Synthesis of UMP-NeuAc...................................................................................49 3 PURIFICATION AND KINETI C CHARACTERIZATION OF RECOMBINANT HUMAN ALPHA (2 3) SIALYLTRANSFERASE IV............51 Introduction.................................................................................................................51 Results and Discussion...............................................................................................52 Overexpression and Purification of Recombinant Human (2 3) Sialyltransferase Isoforms................................................................................52 Kinetic Characterization of Reco mbinant h23STGal IV Isoforms.....................64 Conclusions.................................................................................................................69 Experimental...............................................................................................................69 Materials and Methods........................................................................................69 Preparation of pFastBacHTaInsulin /h23STGal IV (Ins-h23STGal IV)..............70 Preparation of pFastB acHTaInsulin/NtermHis6x-tag-h23STGal IV Plasmid (NtermHis-h23STGal IV)................................................................................71 Preparation of pFastB acHTaInsulin/CtermHis6x-tag-h23STGal IV Plasmid (CtermHis-h23STGal IV)................................................................................72 Amplification of Recombin ant Baculovirus Plasmids........................................73 Expression and Purificati on of Ins-h23STGal IV...............................................74 Expression and Purification of Nter mHis-h23STGal IV and CtermHish23STGal IV Isoforms.....................................................................................75 Sialyltransferase Enzyme Activity Assays..........................................................76 Steady State Kinetics for Recombinant h23STGal IV Isoforms.........................76 4 KINETIC ISOTOPE EFFECT STUDIES ON RECOMBINANT SIALYLTRANSFERASES........................................................................................78 Introduction.................................................................................................................78 Kinetic Isotope Effect (KIE) Background..................................................................79 Isotope Effect Theory..........................................................................................79 Primary Isotope Effects.......................................................................................84 Secondary Isotope Effects...................................................................................85 Kinetic Isotope Effect Measurement Technique.................................................87 The Competitive Method.....................................................................................87 The Noncompetitive Method...............................................................................90 Kinetic Isotope Effect Methodology...................................................................91 Results and Discussion...............................................................................................94 Conclusions...............................................................................................................104 Experimental.............................................................................................................105 Enzyme Reaction General KIE Methodology...................................................105 5 CONCLUSIONS AND FUTURE WORK...............................................................109

PAGE 8

viii LIST OF REFERENCES.................................................................................................112 BIOGRAPHICAL SKETCH...........................................................................................120

PAGE 9

ix LIST OF TABLES Table page 2-1 CMP-NeuAc isotopomer yields...............................................................................25 2-2 UMP-NeuAc isotopomer yields...............................................................................38 3-1 Recombinant Ins-h23STGal IV purification table...................................................57 3-2 Recombinant NtermHis-h23STG al IV purification table........................................61 3-3 Recombinant CtermHis-h23STG al IV purification table.........................................62 3-4 Kinetic paramters for si alyltransferase isoforms......................................................65 4-1 KIEs measured for recombinant human (2 3) sialyltransferase (Ins-h23STGal IV)............................................................................................................................ 94 4-2 KIEs measured for recombinant rat (2 3) sialyltransferase (r23STGal IV). The asterisk denotes the KIE prev iously measured by Michael Bruner.18...............94 4-3 KIEs measured for recombinant rat (2 6) sialyltransferase (r26STGal I). Asterisks denote KIEs previous ly measured by Michael Bruner.18,19......................94 4-4 Summary of predicted KI Es based on mechanism.................................................100

PAGE 10

x LIST OF FIGURES Figure page 1-1 General structure of free sialic ac id and N-acetyl neuraminic acid (NeuAc) which is transferred by sialyltransferases...................................................................2 1-2 Structure of the viral sialidase inhibitor, DANA........................................................4 1-3 Chemical structure of sial yltransferase inhibitor KI-8110.........................................6 1-4 Proposed mechanism for inverting a nd retaining glycosyltransferases and glycosidases from Lairson et al.30..............................................................................8 1-5 3-D structural representations of the GT-A and GT-B fold groups of glycosyltransferases from Coutinho et al.29.............................................................10 1-6 Proposed mechanism of (1 3) fucosyltransferase V from Murray et al.40...........12 1-7 Common topology of a type II memb rane protein from Wang et al.43 The L, S, and VS sialylmotifs of sialyltran sferase are indicated in color................................13 1-8 Reactions catalyzed by (2 6) sialyltransferase and (2 3) sialyltransferase.....14 1-9 Proposed transition-state for sialy ltransferase-catalyzed reaction from Horenstein et al.48.....................................................................................................16 1-10 Interaction of the ring oxygens of CMP-3FNeuAc with the phosphate oxygens in CstII 32 (left) and interactions of CMP a nd active site residues (right) from Chiu et al.49...............................................................................................................17 1-11 Structure of CMP-quinic acid and transition-state analogs......................................22 1-12 Transition-state analogs of CMP-Ne uAc synthesized by Horenstein and coworkers.59.................................................................................................................23 2-1 Structure of labeled CMP-NeuAc. Asterisks denote sites of isotopic substitution...............................................................................................................25 2-2 Enzymatic synthesis of N-acetyl neur aminic acid and CMP-NeuAc. Asterisks indicate sites of possible isotopic substitution.........................................................26

PAGE 11

xi 2-3 Radioactive profile of HPLC fracti ons from a typical NeuAc reaction. The composition of radioactive ManNAc and NeuAc in the reaction mixture was ~ 20 % and 80 %, respectively after four days............................................................27 2-4 Ring opening mechanism for NeuAc.......................................................................28 2-5 31P-NMR of [1-14CN-acetyl 2-18O] CMP-NeuAc..................................................29 2-6 (-) ESI-MS of [2-18O] CMP-NeuAc m/z 615 [M-H]-(top panel), zoom MS/MS of [2-18O] CMP-NeuAc [M-H](center panel), and MS /MS dissociation of m/z 615 [M-H]ion (bottom panel).................................................................................30 2-6 Enzymatic synthesis of [P18O3] CMP from KH2P18O4. The enzymes used in this synthesis were glyceraldehyde-3-pho sphate dehydrogenase (GAPDH), 3phosphoglycerate phosphokinase (3-PGK) and uridine kinase (UDK)...................31 2-7 Enzymatic synthesis of [1-14CN-acetyl P18O2] CMP-NeuAc from [P18O3] CMP.31 2-8 (-) ESI-MS of KH2P18O4..........................................................................................33 2-9 31P-NMR spectrum of KH2P18O4 (1M) in D2O with 4 mM EDTA.........................34 2-10 10 % SDS-PAGE of purified UDK fractions from Red-A dye affinity column......35 2-11 (+) ESI-MS of P18O3 CMP 2 (upper panel) and zoom-MS of the [M+H]+ ions (lower panel)............................................................................................................36 2-12 (+) ESI-MS spectrum of [P16O2 18O2] CMP-NeuAc.................................................37 2-13 Structure of labeled UMP-NeuAc. Asterisks denote sites of isotopic substitution...............................................................................................................38 2-15 Chemical deamination of CMP-NeuA c to UMP-NeuAc by sodium nitrite.............39 2-16 HPLC chromatogram of CMP-NeuA c deamination reaction after 48 hr.................40 2-17 HPLC chromatogram of CMP-NeuAc deamination after 30 hrs incubation...........41 3-1 Diagram of the recombinant h23STGal IV constructs. The insulin signal peptide sequence is represented in yellow, th e truncated h23STGal IV sequence is represented in blue and the His6x-tag sequence is represented in purple.................53 3-2 General scheme for the generation of recombinant baculoviruses and protein expression with the BAC-TO-BAC expression system.7.........................................56 3-3 Structure of CDP-Hexanolamine affin ity ligand synthesized for the purification of recombinant Ins-h23STGal IV enzyme...............................................................56

PAGE 12

xii 3-4 Typical elution chromatogram of recombinant Ins-h23STGal IV from a sepharose CDP-hexanolamine affinity column. Solid squares represent protein concentration and open diam onds represent activity...............................................59 3-5 10 % SDS-PAGE of purified recomb inant Ins-h23STGal IV. Lane 1, MW standard; Lane 2, Lane 3, and Lane 4 are 10 L, 20 L, and 30 L loads of a TCA precipitation of purified Ins-h23STG al IV, respectively(left gel). SDSPAGE of purified recombinant Ins-h 23STGal IV digested with PNGase.18 Lane 1, MW standard; Lane 2, 30 L load of a TCA precipitation of purified Insh23STGal IV and Lane 3 is 30 L load of PNGase digested Ins-h23STGal IV......60 3-6 Typical elution chromatogram of recombinant NtermHis-h23STGal IV and CtermHis-h23STGal IV from a Ni2+-NTA affinity column. Solid squares represent protein concen tration and open diamonds represent activity...................63 3-7 10 % SDS-PAGE gel of purified recombinant CtermHis-h23STGal IV NtermHis-h23STGal IV. Lane 1, TCA precipitation of purified CtermHish23STGal IV; Lane 2, MW Standard; La ne 3, TCA precipitation of purified NtermHis-h23STGal IV...........................................................................................63 3-8 Michaelis-Menten plots for recombinan t Ins-h23STGal IV with with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ -lactose] and constant [CMP-NeuAc] (bottom panel)............................................................66 3-9 Michaelis-Menten plots for recombin ant NtermHis-h23STGal with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ -lactose] and constant [CMP-NeuAc] (bottom panel)............................................................67 3-10 Michealis-Menten plots for recombinan t CtermHis-h23STGal IV with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ -lactose] and constant [CMP-NeuAc] (top panel) and for with (lower panel)........................68 4-1 Expanded terms of the Bigeleisen equation.............................................................81 4-2 Free energy diagram depicting the looser potential energy wells in the transitionstate resulting in a normal (>1) is otope effect from Lowry et al.4...........................83 4-3 Free energy diagram depicting the looser potential energy wells in the transitionstate resulting in an inverse (<1) isotope effect from Lowry et al.4.........................83 4-4 The symmetric stretching vibration m ode in the transition-state of transfer reactions. C represents th e isotopically substituted atom that is transferred between A and B......................................................................................................85 4-5 Typical t0 (top panel) and t1/2 (lower panel) UMP-NeuAc HPLC chromatograms for KIE experiments on recombinant sialyltransferase............................................93

PAGE 13

xiii 4-6 Positional isotope exchange (PIX) mech anism. PIX could not be dectected for the sialyltransferase mechanism. If pixing is complete, the bridge 18O label scrambles to give a 33 % 18O distribution at each oxygen.......................................98 4-7 Bond order analysis for protonation at the non-bridging phosphate oxygen of the donor substrate.......................................................................................................100 4-8 Bond order analysis for protonation at the bridging phosphate oxygen of the donor substrate (top panel) and no prot onation of the phosphate oxygens of the donor substrate (lower panel).................................................................................101 4-9 Transition-state models proposed fo r recombinant r26STGal I (left) and r23STGal IV (right) enzymes.................................................................................103 4-10 Early and late transition stat e analogs for PNP from Schramm.31.........................105 5-1 Proposed sialyltransferase transition-state inhibitor..............................................111

PAGE 14

xiv 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 HEAVY ATOM AND HYDROGEN KINETI C ISOTOPE EFFECT STUDIES ON RECOMBINANT, MAMMALIAN SIALYLTRANSFERASES By Erin E. Burke August, 2005 Chair: Nicole A. Horenstein Major Department: Chemistry Sialylated glycoproteins a nd glycolipids are key recogni tion molecules for a host of biological processes such as cell-cell regulation, cell adhesion, and biological masking. Sialyltransferases are glycosyltransferases th at catalyze the biosynt hesis of sialylated oligosaccharides in the Golgi apparatus of many prokaryotic and eukaryotic cells. The mechanism of sialyl transfer from activated donor substr ate, 5Â’-cytidine monophosphate N-acetylneuraminic acid (CMP-NeuAc), to terminal positions of carbohydrate groups on glycoproteins and glycolipids is still not fully understood. Kinetic studies on recombinant rat liver (2 6) sialyltransferase propose a mechanism with a late oxocarbenium ion-like transitionstate and general acid catalysi s to assist in glycosyl transfer. This dissertation describes th e mechanistic study and the transition-state analysis of three recombinant sialyltransferases. The results from this study will provide an increased understanding of th e mechanism of glycosyl transfer which may be useful in the future development of new sialyltransferase inhibitors.

PAGE 15

xv The first part of this work describe s the synthesis and purification of the isotopically labeled CMP-NeuAc and UMP-NeuA c donor substrates required to conduct the desired kinetic experiments. Details describing a novel enzymatic route for the synthesis of non-bridging phosphate 18O labeled CMP-NeuAc are al so presented in this section. The characterization of these isotopically labeled su bstrates is shown here as well. Following the synthesis of the substrates the next section describes the cloning, overexpression, and purification of three recombinant human (2 3) sialyltransferases, two of which contain either a N-terminal His6x-tag or a C-terminal His6x-tag. The purification yields, specific activ ities, and kinetic parameters of these recombinant human (2 3) sialyltransferases are also presented. The dissertation concludes with the discussi on of the kinetic isot ope effect studies on recombinant human (2 3), rat liver (2 3), and rat liver (2 6) sialyltransferase with the aforementioned isotopical ly labeled substrates. The kinetic isotope effects that were measured on these enzymes include secondary -dideuterium, binding, control, and primary and secondary 18O leaving group isotope effects. Comparisons were made among isotope effects measured for the recombinant human and rat (2 3) sialyltransferases and for the recombinant rat (2 6) sialyltransferase. The KIE data provide new information regarding the na ture of the transition-states for the (2 3) and (2 6) sialyltransferase enzymes.

PAGE 16

1 CHAPTER 1 INTRODUCTION Sialyltransferases are glycosyltransferases th at catalyze the transfer of sialic acid (N-acetylneuraminic acid, NeuAc) from an activated CMP-NeuAc donor substrate to non-reducing termini of glycoproteins and glycol ipids with inversion of configuration at the NeuAc glycon. Over the last two decades, research interest in sialyltransferases has increased primarily because these enzymes play a critical role in th e regulation of a host of biological processes. The mechanism of si alyl transfer is stil l not fully understood. Results from kinetic studies conducted previously by Bruner and Horenstein on recombinant rat liver (2 6) sialyltransferase (ST6Gal I) suggested that the mechanism proceeds via a late oxocarbenium ion-like tran sition-state with gene ral acid catalysis to assist in glycosyl transfer.1,2 The work described in this dissertation represents an investigation into the reac tion catalyzed by recombinant (2 3) sialyltransferase from human placenta (h23STGal IV). The enzyme -substrate binding in teractions at the phosphate group of the donor substrate are of particular interest. Additionally, comparisons can be made between enzymes in the same family using the information obtained from work completed on recombinant rat liver (2 3) sialyltransferase and from the work previously reported on the recombinant ST6Gal I. Sialic Acids In order to understand the si gnificance of studying this cl ass of enzymes, one must first appreciate the functional impor tance of sialic acids for a hos t of biological processes. The purpose of this introduction is not to pr ovide a comprehensive report on the subject

PAGE 17

2 of sialic acids, but to merely highlight some of the key roles that sial ic acid residues play in these biological processes. This info rmation will emphasize the importance of how studying the enzymes involved in sialic acid re gulation could lead to ways in which to control and monitor a wide ar ray of biological pathways. In nature, sialic acids are linked in the term inal steps of the synt hesis of cell surface glycoproteins and glycolipids. The structure of sialic acid is unique in that it contains a highly acidic carboxylate group on the anomeric carbon (Figure 1-1) The negative charge on sialic acids is an important chemi cal feature of this molecule and it plays a functional role in a variety of biological processes. For example, the negative charge on sialic acids provides this molecule with the ab ility to attract and repel specific cells and biomolecules.3 In the case of attraction, the negative charge allows sialic acids to bind to positively charged molecules and assist in their transport. Figure 1-1. General structure of free sialic acid and N-acetyl neuraminic acid (NeuAc) which is transferred by sialyltransferases. O OH CO2H H2N HO OH HO HOSialic Acid (5-amino-3,5-dideoxy-Dglycero -Dgalacto -nonulosonic acid)O OH CO2H AcHN HO OH HO HON-acetyl neuraminic Acid (NeuAc) (5-amino-3,5-dideoxy-Dglycero -Dgalacto -nonulsonic acid)

PAGE 18

3 On the other hand, the population of sialic acid residues on the periphery of cells provides the cell with a net negative charge th at is essential for the repulsion of other cells or biomolecules. This is seen in eryt hrocytes and blood platelets where cell-surface sialic acids can prevent the aggregat ion of these cells in the bloodstream.4 Furthermore, the negative charge on these sugars has also be en shown to contribute to the viscosity of mucins lining intestin al endothelia cells.5 Thus, the electrochemical properties of sialic acids appear to influence their unique func tion in a variety of biological phenomena. One of the most important roles of sialic acids is their ability to function as recognition elements for key processes. This function is facilita ted by their chemical properties and by their locati on on surface of cells. For ex ample, sialic acids are recognition molecules for bacterial and viral pathogens. The best known example of this was observed over 50 years ago where sialic acid residues were identi fied as recognition molecules for the binding of influenza A to human erythrocytes and respiratory tract mucins.6 Since then, researchers have shown that influenza B virus can also bind to sialic acid residues that are N-linked to cellsurface glycoproteins or glycolipids.7,8 The binding of influenza to cell surface receptors be aring sialic acids is mediated by the viral protein, Hemagglutinin (HA). HA works in conjunction with viral sialidase (or neuraminidase) during the viral life cycle. Vi ral sialidase has been suggested to facilitate the spread of influenza virus by cleaving si alic acid residues from the protecting mucin layer of respiratory tract epithelia cells.9 This research lead to the development of viral sialidase inhibitors such as 2-Deoxy2,3-dehydro-N-acetylneuraminic acid (DANA) which are being successfully used today for the treatment of influenza (Figure 1-2).

PAGE 19

4 Figure 1-2. Structure of the vi ral sialidase inhibitor, DANA. Another example of sialic acids acting as esse ntial recognition components is in neural development. In this case, polysialic acid, a linear homopolymer of (2 8)linked sialic acids, was disc overed attached to the neur al cell adhesion molecule (NCAM).10 Two polysialyltransferases, ST8Sia II (STX) and ST8Sia IV (PST), regulate the synthesis of polysialic acid on NCAM. Experiments involving NCAM-deficient mice have suggested that polysialic acids on NCAM play critical roles in the regulation of neural cell adhesion, cell migration, neurite outgrowt h, and synapse formation.11,12 Furthermore, deficiencies and disorganizati on in polysialic production have been linked to diseases such as schizophr enia and Alzheimer’s disease.13,14 In contrast to their function as recognition molecules, sialic acids are also important for the anti-recognition of certain biomolecules and cells. As the penultimate “capping” molecule for cell-surface oligosaccharides, sia lic acids serve as biological masking agents by disguising and shielding their underlying sugars from receptor recognition. Ashwell and Morell documented the first example of this function in 1974. Their experiments investigated the role of sia lic acids as masking agents by using sialidase to remove O COOH HO AcHN OH HO HODANA (2-Deoxy-2,3-dehydro-N-acetylneuraminic acid)

PAGE 20

5 terminally N-linked sialic acid residues fr om D-galactose molecules on radiolabeled ceruloplasmin in the bloodstream.15 Upon removal of thes e sialic acids, exposed galactosyl residues were quick ly recognized by D-galactose receptors and the resulting radiolabeled asialoceruloplasmin disappeared from the bloodstream within minutes. Analysis of the liver showed increased radi oactivity, thus signifyi ng the degradation of radiolabeled asialoceruloplasmin. Another example of the sialic acid mask ing effect can be seen in the binding recognition of siglecs in the immune sy stem. Siglecs are sialic acid-binding immunoglobulin-like lectins involved in cel lular signalling functions and cell-cell interactions in the nervous and immune systems.16 Siglecs use both cis and trans interactions with sialic acid ligands when binding to the ce ll-surface. Researchers have shown that the siglec receptor-b inding site can be masked by cis interactions with sialic acid ligands.16,17 These cis interactions with sialic acids are essential for the regulation of siglec function by preventing or facilitating spec ific cell-cell interactions when necessary. Sialic acids have also been implicat ed in the masking of tumor antigens.18 In this case, the high sialic acid c ontent on some tumor cells can ma sk antigen recognition sites. As a result, tumor cells can el ude immunological attack and, in some cases, continue to grow uncontrollably. Clinical studies have observed that certain invasive cancer cell lines have hypersialylated cellsurfaces and the patients with these cancers have increased sialyltransferase activities in their blood serum.19 In response to these results, researchers have been investigating the us e of sialidases and sialyltran sferase inhibitors as methods for cancer treatment. Studies conducted on ch emically induced malignant tumors in

PAGE 21

6 small mammals showed marked tumor regr ession after treatment with sialidase.20 Additionally, sialyltransferase i nhibitors such as KI-8110 have been shown to assist in the reduction of tumor metastases by inhibiti ng the transfer of sia lic acid onto cell-surface oligosaccharides (Figure 1-3).21-24 Thus, these findings have opened the door to new possibilities in the developm ent of cancer treatments. Figure 1-3. Chemical structure of sialyltransferase inhibitor KI-8110. Since sialic acids participat e in such a plethora of bi ological events, changes in their chemical structure, concentration, or mutations to their biosynthetic pathways can cause severe diseases. Some of the diseases include sialidosis, gala ctosialidosis, sialuria, and sialic acid storage disorder (SASD).25 Sialidosis and galactosialidosis are genetically inherited, lysosomal storage dise ases that are characterized by the inability to degrade sialylated glycoproteins due to a deficiency in the production of sialidase.26 In contrast, sialuria is characterized by the overproduction of sialic acids in the cytoplasm resulting from a lack of feedback inhibition of th e rate-limiting enzyme, uridine diphosphate NO AcOH2C OAc OAc AcHN AcO COOMe O O O O N NH F O OKI-8110 (5-fluoro-2',3'-O-isopropylidene-5'-O-(methyl 5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-D-galacto-2 -nonulopyranosylonate)-uridine

PAGE 22

7 acetylglucosamine 2-epimerase.27 Sialic acid storage disord er is the rarest of the aforementioned diseases with less than 150 cases reported worldwide. Most of the reported cases stem from a small region in north eastern Finland. SASD is typified by the accumulation of unbound sialic acid within the lysosomes cause d by a defect in sialin, a transmembrane protein responsible for transport of sialic acid out of the lysosome. As a result, patients with SASD have up to 100 tim es more unbound sialic acid in their urinary excretions than normal. In general, pa tients suffering with these diseases display neurodevelopmental delays, severe learning diffi culties, coarse facial features, spinal abnormalities, skin lesions, and visual impairment.26 The functional importance of sialic acids in biological systems is vast. Their unique roles in a variety of cellular events make sialic acids indispensable to life. Therefore, mechanistic studies on the enzyme s involved in the bios ynthesis and transfer of sialic acids could provide new tools in which to investigate the complex nature of these biomolecules. Glycosyltransferases and Glycosidases Glycosyltransferases are a class of enzymes that catalyze the transfer of monosaccharide residues from monoor diphosphate sugar nucleotides to the nonreducing end of extending oligosaccharide ch ains, or, in general, different aglycon acceptors. These membrane bound enzymes are highly specific for their donor and acceptor substrates, and their function is criti cal for the proper glycosylation of a myriad of oligosaccharide and polysaccharide chains In mammalian cells, estimates indicate that well over 100 different glycosyltransferases are requi red to biosynthesize all known oligosaccharide structures.28 The resulting glycan chains regulate a diverse range of

PAGE 23

8 cellular functions, in cluding cell-cell interacti ons and signalin g, host-pathogen interactions, neuronal development, embyogenesis, and biological masking.29 Glycosyltransferases are classified by sequence similarity-based families and by the type of mechanism they catalyze, which is retaining or inverting depending on the final anomeric configuration of the product (Figure 1-4).30 The inverting glycosyltransferase catalyzed mechanism is suggested to follow an SN2-like reaction whereby a general base deprotonates the incoming nucleophile of the acce ptor sugar, thus enabling the direct displacement of the nucleoside diphosphate. In this mechanism, metal ions such as Mg2+ or Mn2+ are believed to serve as acid catalys ts for some glycosyltransferases.31 Figure 1-4. Proposed mechanism for inverti ng and retaining glycos yltransferases and glycosidases from Lairson et al.30 O B OPHOR' HO AH O B OPHOR' HO AH Oxocarbenium Ion-Like Transition State POHO HO OR' B A O Nuc OP HO AH O Nuc OP HO AH O Nuc OP HO AH O Nuc H HO A POH Oxocarbenium Ion-Like Transition State Oxocarbenium Ion-Like Transition StateO Nuc HO A-Glycosyl-Enzyme Intermediate HORInverting Mechanism Retaining Mechanism ORP= Nucloside diphosphate or monophosphate

PAGE 24

9 The retaining glycosyltransf erase catalyzed mechanism is proposed to proceed via a double-displacement reaction with a covalently bound glycosyl-enzyme intermediate. In this mechanism, an aptly positioned amino acid within the active-site functions as a nucleophile to catalyze the reaction.30,32 A divalent cation is believed to act as a Lewis acid whereas the leaving diphosphate group has b een suggested to serv e as a general base by deprotonating the incoming acceptor sugar hydroxyl to activate it for nucleophilic attack. In comparison to inverting glycosyl transferases, retaining glycosyltransferase reactions also proceed through oxocarbenium ion-like transition states. Despite this similarity, the mechanism for retaining glycosy ltransferases is still being explored since some of the intermediates have yet to be conclusively identified. Within the last decade, several crystal stru ctures of glycosyltran sferases have been reported.28,32,33 Based on recent structural data, gl ycosyltransferases adopt one of two general folds referred to as GT-A and GT-B (Figure 1-5).29 Glycosyltransferases categorized under the GT-A (glycosyltransf erase A) fold group typically contain a conserved ‘DXD’ motif, a conica l active site cleft formed by two closely associated domains, and / proteins with a single Rossmann domain. The ‘DXD’ motif has been shown to play a critical role in metal ion binding and catalysis. The divalent metal cation coordinates the phosphate group oxygens of the sugar nucle otide donor in the enzyme active-site.34,35 Additionally, binding of the nucleotide has been mainly observed on the N-terminal domain of GT-A enzymes. Some of the glycosyltransferases that have been classified under the GT-A fold group include phage T4 -glucosyltransferase, glycogen phosphorylase and -1,3-galactosyltransferase.29,36

PAGE 25

10 The GT-B (glycosyltransferase B) fo ld group consists of two Rossman-like / / domains that are separated by a deep substr ate-binding cleft. Nu cleotide binding for GTB enzymes takes place on the C-terminal domai n while the acceptor substrates bind to the N-terminal domain. The GT-B superfamily encompasses a diverse group of prokaryotic and eukaryotic enzymes that ar e responsible for a variety of processes ranging from the production of biologically active antibiotics to cell wall biosynthesis and gene transcription.37 Figure 1-5. 3-D structural representations of the GT-A and GT-B fold groups of glycosyltransferases from Coutinho et al.29 In contrast to glycosyltransferases, glyc osidases execute further modifications to glycosylated biomolecules by catalyzing the clea vage of sugar residues. This class of enzymes uses water as a nucleophile to trim carbohydrate residues from these biomolecules in order to meet the requirem ents for a variety of biological processes. Glycosidases follow similar mechanistic paths to those described above for glycosyltransferases, but with a few excep tions. The characteristic mechanism for retaining glycosidases involves a pair of aspa rtic or glutamic acid residues in enzyme active-site, with one functioning as a nucle ophile and the other acting as a general acid/base catalyst. Unlike retaining glyc osyltransferases, key intermediates in the

PAGE 26

11 retaining glycosidase mechanism such as the covalently bound glycosyl-enzyme intermediate have been conclusively identi fied and characterized using crystallographic and spectroscopic methods. Recently, the W ithers laboratory characterized this covalently bound glycosyl-enzyme intermediate for the hen egg-white lysozyme (HEWL) mechanism.38 Fucosyltransferases Although glycosyltransferases have not been as well characterized as glycosidases, one member of the glycosyltran sferases superfamily that has been closely studied is fucosyltransferase. Fucosyltransfer ases catalyze the transfer of 1-fucose from an activated GDP-fucose donor substrate to ol igosaccharide chains linked to proteins or lipids. In recent years, the mechanism for (1 3) fucosyltransferase V (FucTV) has been investigated using ki netic isotope effect experi ments and inhibitor studies.39-41 FucTV catalyzes the final step in the bi osynthesis of sialyl Lewis X and Lewis X fucoglycoconjugates. These fucoglycoconjugate s play an essential role in the regulation of cell-cell interactions for a variet y of immune system processes. Kinetic isotope effect and pH-rate studi es conducted on FucTV suggest that the mechanism for fucosyltransfer is base cataly zed where the L-fucose is transferred to acceptor sugars with inversion of configuration.39 Results from secondary isotope effect studies using deuterated GDP-[1-2H]-Fucose as the donor substrate indicated that cleavage of the glycosidic bond o ccurs prior to nucleophilic atta ck as illustrated in Figure 1-6.40 Furthermore, the transition-state structure is similar to glycosidases in that it is proposed to have a flattened half-chair conf ormation with considerable oxocarbenium ion character at the anomeric position.

PAGE 27

12 Figure 1-6. Proposed mechanism of (1 3) fucosyltransferase V from Murray et al.40 Product inhibition studies on FucT V have revealed that the mechanism is an ordered, sequential, Bi-Bi mechanism in wh ich the GDP-Fuc binds first followed by the acceptor sugar. FucTV also requir es a metal co-factor, typically Mn+2, to achieve optimal catalysis. Furthermore, FucTV can use both charged and uncharged sugar acceptor substrates such as N-acetyllactosamine (Lac NAc) and sialyl LacNAc, respectively. This sugar acceptor substrate variability is simila r for sialyltransferases however, the FucTV donor substrate, GDP-Fuc, does not contai n the highly acidic carboxylate group on its anomeric carbon like the sialyltransferase dono r substrate, CMP-NeuAc. Thus, this group will undoubtedly alter the enzyme-donor substrate reactivity when compared to GDP-Fuc. Nevertheless, the mechanism for FucTV may provide some insight into the nature of the sialyltransferase a nd other glycosyltransferase catalyzed reactions. O OH HO OH O GDP OH Acceptor O2CEO O AcceptorHEO2C P O O O P O OObroken firstO OH HO OH O Acceptor+ GDPO OH HO OH O HO OH G Mn2+

PAGE 28

13 Sialyltransferases As a subfamily of glycosyltransferases, sial yltransferases are also localized in the Golgi apparatus and their topology is characteristic of a type II membrane protein with a short cytoplasmic domain, an N-terminal si gnal anchor and a large luminal catalytic domain (Figure 1-7). There are presently 20 cl oned cDNAÂ’s of sialyltr ansferases isolated from bacteria, insects, and mammals. Th eir nomenclature and f unction are determined by the different acceptor sugar substrates that si alyltransferases bind to in the transfer of NeuAc. For example, the recombinant h23STG al IV catalyzes the transfer of NeuAc from CMP-NeuAc donor substrate to the C3 terminal hydroxyl of Gal 1,4GlcNAc or Gal 1,3GalNAc acceptor sugars while the reco mbinant ST6Gal I transfers NeuAc residues to C6 terminal hydroxyl of Gal 1,4GlcNAc substrates (Figure 1-8).42 Figure 1-7. Common topology of a type II membrane protein.

PAGE 29

14 Figure 1-8. Reactions catalyzed by (2 6) sialyltransferase and (2 3) sialyltransferase. Aside from common topological features, sialyltransfersases do not share any sequence homology with other enzymes in the glycosyltransferase family. However, sequence homology analysis of the sialyltran sferase family revealed the existence of several conserved protein motifs within the catalytic domain referred to as L, S, and VS sialylmotifs. The L and S sialylmotifs are lo cated at the center of the lumenal catalytic domain of sialyltransferases and are com posed of approximately 48 and 23 amino acid residues, respectively. The VS sialylmo tif is located in the C-terminus of sialyltransferases and consists of two highly conserved glut amate and histidine residues that are separated by four amino acid resi dues. The Paulson laboratory conducted a series of site-directed mutagenesis studies on recombinant rat liver ST6Gal I to investigate the roles of several conserve d amino acids in the L & S sialylmotifs.43,44 Kinetic data from analysis of ST6Gal I mutant constructs suggested that the L sialylmotif participated in the binding of CMP-NeuAc, while the S sialylmotif participated in the binding of both donor and acceptor substrates. The Paulson group also found that the two invariant cysteine residues present in each of the L and S sialylmotifs for an O HO AcHN HO P O O O O OH OH N N NH2OCMP-NeuAcO HO OH AcHN HO O CO2HO HO O HO OH HO O O HO HO NHAca ( 2 6 ) S i a l y l t r a n s f e r a s ea ( 2 3 ) S i a l y l t r a n s f e r a s eSialylLacNAc SialylLacNAcO CO2OH O O HO O O HO HO NHAc OH HOH2C O HO OH AcHN HO CO2HO OH OHL a c N A cCMPL a c N A cCMP

PAGE 30

15 intrachaindisulfide bond that is essential for retention of catalytic activity and proper conformation of ST6Gal I.45 A closer examination of all known eukaryot ic sialyltransferase sequences revealed the presence of another highly conserve d motif located between the S and VS sialylmotifs referred to as the aromatic motif. 46 This motif is comprised of a stretch of four highly conserved mostly aromatic residues. The functional role of these amino acid residues was investigated us ing site-directed mutagenesi s experiments on recombinant human hST3Gal I. The results suggested th at the highly conserve d histidine (His299) and tyrosine (Tyr300) residues of the aroma tic motif are necessary for enzyme activity since their mutation to alanine gene rated inactive enzymes. Apart from research involving conserved residues in the sialylmotifs, there is still limited information available concerning the catalytic mechanism and structure of the sialyltransferase family. Previous work in the Horenstein laboratory used radiolabeled CMP-NeuAc and UMP-NeuAc, a weak binding s ubstrate analog, to co nduct a series of kinetic isotope effect and pH vs. rate expe riments on recombinant rat liver ST6Gal I to elucidate the mechanism of glycosyl transfer A dissociative mechanism involving a late oxocarbenium ion-like transition state was pr oposed in the model fo r sialyltransferase catalysis based on KIE results (Figure 1-9).1,47 The pH-rate profile from experiments using UMP-NeuAc and LacNAc as the donor-ac ceptor substrate pair fits a bell-shaped curve for two ionizable groups with pKa values of 6.2 and 8.9. Further pH-rate experiments and theoretical calculations suggest ed that glycosyl transfer proceeded via a general acid catalyzed mechanism in wh ich a non-bridging phosphate oxygen on CMPNeuAc may be protonated to facilitate the loss of CMP.1 The kinetic mechanism was

PAGE 31

16 proposed to be steady-state random based on in itial velocity, KIE, and isotope trapping experiments.1,2 Figure 1-9. Proposed transition-state for si alyltransferase-catalyzed reaction from Horenstein et al.47 A three dimensional crystal structur e of sialyltransferase CstII from Campylobacter jejuni was recently reported by Chiu et al.48 From the structure, sialyltransferase CstII 32 was categorized under the GT-A fold group because it contained a single Rossmann domain. Aside from this feature, sialyltransferase CstII lacked the conserved ‘DXD’ motif and a bound metal in the active site which are common characteristics among other glycosyltransferases in the GT -A group. Since sialytransferases do not require a metal cofactor for catalysis,49-51 the lack of the conserved ‘DXD’ motif responsible for the binding of a diva lent metal cation was not surprising. In order to observe enzyme-substrate binding interactions in the active site, Chiu et al. crystallized sialyltransferase CstII 32 in the presence of CMP-3FNeuAc, an unreactive substrate analog of CMP-NeuA c (Figure 1-10). The structure of sialyltransferase CstII 32 complexed with CMP-3FNeuAc offers some explanations regarding the nature of sialyl transfer. In the crystal struct ure, the sialyl moiety of CMPO HO AcHN OH OH HO O O O P O O O Cytosine"H+"

PAGE 32

17 3FNeuAc adopted a distorted skew boat c onformation which favors formation of the oxocarbenium ion. The leaving-group phosphate was oriented in a pseudo axial position twisted above the plane of the sugar ring allowing the pr o-R oxygen on the phosphate to interact with the ring oxygen on NeuAc (Figur e 1-10). Cleavage of the glycosidic bond and departure of the CMP moiety was suggest ed to be facilitated both by the negativecharge buildup on the pro-R phosphate oxygen, and by the hydrogen bonding interactions with the non-bridging phosphate pro-S oxygen and active-site Tyr156 and Tyr162 residues. Mutagenesis experiments using Y 156F and Y162F mutants of sialyltransferase CstII 32 resulted in a significant loss in catalytic activity with only one tyrosine residue mutated and a total loss in catalytic activity with both tyrosine residues mutated. Although it is unclear why acid catalysis would be necessary to assist in the departure of a stable monophosphate leaving group, these results indicate that both residues are critical for optimal catalytic efficiency of the CstII 32 transferase mechanism. Figure 1-10. Interaction of the ring o xygens of CMP-3FNeuAc with the phosphate oxygens in CstII 32 (left) and interactions of CMP and active site residues (right) from Chiu et al.48 Deprotonation of the incoming hydroxyl gr oup of the acceptor sugar was suggested to be catalyzed by His188. The close proximity of His188 at 4.8 to the anomeric

PAGE 33

18 carbon in the crystal structure made His188 the only feasible candidate for the role of general base catalyst; however, this identification is still ambiguous. Based on active-site comparison studies, His188 is located in a similar position to other catalytic bases identified in inverting glycosidases.52 Additionally, the pH optimum of 8.0 for the CstII 32 catalyzed reaction favors deprotonation of the His188 imidazole, therefore, allowing it to act as a general base catalyst. A complete loss of transferase activity was also observed when His188 was mutated to alanine in CstII 32. This same result was also observed in histidine to lysine/alani ne mutagenesis experiments on recombinant human hST3Gal I and polysialyltr ansferases, ST8Sia II and IV.46,53 Thus, the results from these experiments reinforce the hypothe sis that a histidin e residue plays an important role in the sialy ltransferase catalyzed reaction. Despite the information obtained from this sialyltransferase crystal structure, bacterial sialyltransferase CstII 32 does not share sequence homology with any mammalian sialyltransferases, which is where the primary research interest exists.54 Hence, one can not assume that the mammalian sialyltransferases will adopt the same structural fold and active-site arra ngement as sialyltransferase CstII 32. To date, there are no three dimensional crystal structures re ported for a mammalian sialyltransferase. This is primarily due to the fact that reco mbinant mammalian sialyltransferases are more difficult to overexpress and purify. Sufficien t quantities of pure enzyme are arduous to obtain to conduct crysta llization experiments. Furthermor e, even if enough pure enzyme was available to attain a crys tal structure, these structur es do not provide information about the reactionÂ’s transition-state structure.

PAGE 34

19 Therefore, other techniques may be used advantageously on mammalian sialyltransferases in order to probe the mechanism of sialyl transfer. Methods such as kinetic isotope effect (KIE) experiments can often provide detailed information on the transition-state structure which will assist in acquiring mechanistic information for the sialyltransferase catalyzed reaction. In th is study, several CMP-NeuAc and UMP-NeuAc radioisotopomers were synthesized to investig ate the mechanism of sialyl transfer using KIE experiments. The dual-label competitive method was used to measure the KIEs for these radiolabelled substrates with recombinant human placental (2 3) sialyltransferase and recombinant rat liver (2 3) sialyltransferase. The data from these experiments will provide an increased unders tanding of the mechanism of glycosyl transfer with regard to interactions at the phosphate leaving group via 18O isotopic substitution at the glycosidic O and non-bridging phosphate oxygen atoms. Sialyltransferase Inhibitors Within the last decade, a burst in th e design and synthesis of a variety of sialyltransferase inhibitors occurred due, in pa rt, to their interest as potential therapeutic compounds for the treatment of tumor metastas es and immunological diseases. Inhibition studies on purified sialyltransferase also becam e more feasible in recent years because of the increased commercial availability of r ecombinant sialyltransf erases. Although there have been numerous reported sialyltransf erase inhibitors, only the more noteworthy inhibitors will be discussed here. The most common strategy used toward the design of sialyltransferase inhibitors have been donor substrate based analogs of CMP-NeuAc. The idea behind the development of these inhibito rs was to alter functional groups on the sugar or nucleotide portion of the donor substrate, but maintain the basic glycosidic linkage. In 1997,

PAGE 35

20 Schauer and co-workers conducted a series of inhibition studies on recombinant (2 6) sialyltransferase from rat liver and (2 3) sialyltransferases fr om porcine submandibular gland using a variety of diffe rent nucleosides, nucleotides, sialic acid and sugar nucleotide analogs as donor substrates.24 The goal of the st udy was to identify key structural elements that were essential for i nhibition of sialyltransf erase. The inhibition studies showed that donor substrate an alogs containing a nu cleotide monophosphate moiety were the most effective sialyltransfer ase inhibitors, while the sialic acid analogs displayed little to no inhibition. CMP, CDP, and CTP were natural competitive inhibitors of sialyltransferase with Ki values of 90, 50, and 46 M, respectively. These inhibition constants are comparable to the Km value for the natural donor substrate CMP-NeuAc (46 M).55 The enhanced inhibitory effect upon addition of one or more phosphate groups was proposed to be caused by their ability to provide a negative charge similar to the carboxylate group of CMP-NeuAc. Moreover, the results from this study suggest that the nucleotide moiety, particularly cytidine monophosphate, is a funda mental structural requirement for high binding affinity of th e donor substrate to the enzyme active site. This information led to the development of more sialyltransferase inhibitors that incorporated the general cy tidine or cytidine monophosphate scaffolding in the donor substrate. Schmidt et al synthesized a series of sialyltr ansferase inhibitors with cytidine monophosphate linked to quinic acid analogs.56 These CMP-quinic acid based inhibitors were advantageous to use because they not only included the CMP moiety for high binding affinity, but they also blocked transferase activity by changing the glycosidic bond to a more stable C-glycoside linkage. These compounds were also stable under physiological conditions. Inhibition experi ments using CMP-quinic acid as the donor

PAGE 36

21 substrate with ST6Gal I gave a Ki value of 44 M, which is approximately the same as the Km value for the natural substrate CMP-Ne uAc. With this information in hand, Schmidt and co-workers modified their inhibitor design strategy by synthesizing transition-state analogs that would mimic the oxocarbenium ion-like transition-state proposed for CMP-NeuAc.57 A few of the transition stat e analog inhibitors synthesized and tested by Schmidt et al are shown in Figure 1-11. These compounds contain a flattened ring w ith the anomeric car bon trigonal planar to simulate the oxocarbenium ion coplanarity in the transition-state. A methylene group was also added between the anomeric carbon a nd CMP to model their increased distance in the proposed transition-state structure of CMP-NeuAc. Substitution of the methylene hydrogen with a phosphonate group greatly incr eased the inhibitory activity of these compounds with Ki values in the nanomolar range. The phosphonate group provided an additional negative charge similar to the ca rboxylate of CMP-NeuAc. Furthermore, replacing the glycerol side ch ain on the NeuAc ring with a phenyl group as seen in Figure 1-11, resulted in a 1,000-fold increase in bi nding affinity to ST6Gal I with a Ki of 29 nM. These results demonstrate the enzymeÂ’s capability to tolerate bulky side chain modifications made to the donor substrate without compromising binding affinity. To date, the phenyl phosphonate compound is the mo st potent sialyltran sferase inhibitor. Horenstein and co-workers synthesized another unique set of transition state analogs as sialyltransferase inhibitors.58 This new class of sialyltransferase inhibitor employed an unsaturated bicyclic system with a conjugated carboxylate group to mimic the conformation of the proposed transition-st ate (Figure 1-12). Additionally, the CMP moiety attached to the bicyclic sytem was kept at an increased distance from the

PAGE 37

22 anomeric carbon to imitate the late transitio n-state distance proposed for bond cleavage. These compounds were highly efficient inhi bitors of sialyltransferase with Ki values in the low micromolar range. Furthermore, s ubstitution of the NeuAc ring with a bicyclic ring illustrates the enzymeÂ’s ability to accept a diverse range of structure changes in the sugar portion of the donor substrate. Recen tly, Schmidt and co-workers demonstrated that sialyltransferase also exhibit high binding affinities for transition-state analogs with aryl and hetaryl ring systems substituted for the sugar portion.59 Thus, these studies indicate that the neuraminyl ring only plays a minor role in binding since structure variations to this part of the donor su bstrate are still tolerated by the enzyme. Figure 1-11. Structure of CMP-quini c acid and transition-state analogs. O P O O O OH HO HO CO2O OH OH N N NH2O O HO AcHN O P O O O H R HO O OH OH N N NH2O OH OH P OH O OR = H orO O HO AcHN O P O O O O OH OH N N NH2OCMP-Quinic AcidPhosphonate Compound Phenyl Phosphonate CompoundH P O OH O

PAGE 38

23 Figure 1-12. Transition-state analogs of CMP-NeuAc synthe sized by Horenstein and coworkers.58 Moreover, these studies have enabled re searchers to ascert ain that the key components required for sialyltr ansferase inhibitiors are: (i ) a planar anomeric carbon; (ii) an increased distance be tween the anomeric carbon and th e leaving group CMP; (iii) at least two negative charges near the cleav age site; and (iv) the cytidine moiety for recognition.60 Transition-state analogs of CMP-NeuAc have been the most potent of all sialyltransferase inhibitors re ported to date. Therefore, information about the transition state acquired from kinetic isotope effect studies may also prove useful toward the development of new sialyltransferase inhibitors. H O CO2P O O O O OH OH N N NH2O H O CO2P O O O O OH OH N N NH2O2Na+2Na+Ki = 20 M Ki = 10 M

PAGE 39

24 CHAPTER 2 SYNTHESIS AND CHARACTERIZATION OF SUBSTRATES Introduction Cytidine 5Â’-monophosphate neuraminic acid (CMP-NeuAc) is synthesized by CMP-NeuAc synthetase in many prokaryotic an d eukaryotic cells and serves as a key intermediate in the sialyltransferase catalyzed biosynthesis of sialylat ed oligosaccharides. During catalysis, sialic acid (N-acetylneuram inic acid, NeuAc) is transferred from an activated CMP-NeuAc donor substrate to non-reducing termini of glycoproteins, glycolipids, and oligosacchar ide chains. Although sialyltr ansferases vary in acceptor substrate specificity, all sialyltransferases use CMP-NeuAc as their donor substrate. Thus, information obtained from experiment s conducted on recombinant human and rat (2 3) sialyltransferases, may be applied other members in the sialyltransferase family. Results and Discussion Synthesis of CMP-NeuAc isotopomers In order to probe the mechanism of the si alyltransferase catalyzed reaction, a series of CMP-NeuAc isotopomers were synthesized to perform the desired experiments. These CMP-NeuAc isotopomers either co ntain one radioactive trace la bel or a radioactive trace label with several nonradioactive isotopic s ubstitutions. The various sites of isotopic substitution are illustrated in Figure 2-1. The isolated yi elds are shown in Table 2-1.

PAGE 40

25 Figure 2-1. Structure of la beled CMP-NeuAc. Asterisk s denote sites of isotopic substitution. Table 2-1. CMP-NeuA c isotopomer yields. CMP-NeuAc Isotopomer Isolated % Yield [9-3H] 74 [1-3HN-acetyl ] 78 [1-14CN-acetyl ] 54 [1-14CN-acetyl P18O2] 35 [1-14CN-acetyl 2-18O] 71 Chemical and enzymatic methods were em ployed to synthesize the various CMPNeuAc isotopomers (Figure 2-2). The first st ep of the synthesis required the use of Nacetyl neuraminic acid (NANA) aldolas e which was cloned, overexpressed in E. coli and purified according to l iterature procedures.61-64 NANA adolase catalyzes the aldol condensation reaction between pyruvate and N-acetyl mannosamine (ManNAc) to yield NeuAc.65 Isotopomers of NeuAc were obtai ned by substituting nonradiolabeled ManNAc with the appropriate radiolabeled ManNAc substrate. Several of these radiolabeled ManNAc compounds were pur chased commercially, such as, [6-3H] and [1-14CN-acetyl ] ManNAc which give [9-3H] and [1-14CN-acetyl ] NeuAc, respectively. The [1-3HN-acetyl ] NeuAc isotopomer was obtained by synthesizing the [3HN-acetyl ] ManNAc starting material via acet ylation of D-mannosamine with 3H acetic anhydride as O HO OH HO AcHN HO COO O N NH2O N O OH OH O P O O * * H H * 1 2 3 4 5 6 7 8 9

PAGE 41

26 per the literature method.66 The progress of the radiolabeled NeuAc reactions was monitored by making analytical injections on anion-exchange HPLC and collecting fractions for LSC counting (Figure 2-3). Ra diolabeled ManNAc and NeuAc eluted with retention times of 3 min. and 6 min., respectively, on HPLC MonoQ (50-100 mM NH4HCO3 gradient, 15% methanol, 2 mL/min, A271, 2 mL fractions collected). Figure 2-2. Enzymatic synthesis of N-acet yl neuraminic acid and CMP-NeuAc. Asterisks indicate sites of possible isotopic substitution. Further isotopic modifications to the ra diolabeled NeuAc were made once it was isolated. Deuterium labels were incorpor ated into the NeuAc at the C3 position by O HO HO HO OH HN O O O NANA Aldolase 25oC, 72 hrs Na O OH CO2HO O H N + CTP PPi CMP-NeuAc Synthetase 37oC, 6 hrs O N NH2O N O OH OH O P OO O *H * * HO HO OH OCO2HO O H N HO HO OH D* D* * *ManNAc NeuAc CMP-NeuAc

PAGE 42

27 exchanging the protons with deuter ium under alkaline conditions with D2O. This synthesis was performed by Mike Bruner of the Horenstein laboratory as previously described in the literature.2 [3,3Â’-2H2] CMP-NeuAc was then synthesized with CMPNeuAc synthetase with the addition of CTP. Figure 2-3. Radioactive profile of HPLC fractions from a typical NeuAc reaction. The composition of radioactive ManNAc and NeuAc in the reaction mixture was ~ 20 % and 80 %, respectively after four days. Synthesis of [1-14C -N-acetyl 2-18O] CMP-NeuAc Synthesis of the [1-14C -N-acetyl 2-18O] CMP-NeuAc isotopomer was achieved by first exchanging the C-2Â’ hydroxyl oxygen on NeuAc via a ring opening mechanism with H2 18O (95% enrichment) under basic c onditions (pH > 9.5) (Figure 2-4).67 The enzymatic synthesis with [1-14CN-acetyl 2-18O] NeuAc, CTP, and CMP-NeuAc synthetase gave [1-14CN-acetyl 2-18O] CMP-NeuAc in 71 % isolated yield after purification by anion-exchange HPLC. The 31P-NMR spectra showed two peaks at NeuAc Reaction Radioactive Profile0 50000 100000 150000 200000 250000 05101520 Fraction #CPM NeuAc ManNAc

PAGE 43

28 4.243 ppm and -4.257 ppm representing the [1-14CN-acetyl 2-16O] and [1-14CN-acetyl 2-18O] CMP-NeuAc compounds, respectively (Fi gure 2-5). Integration of the peaks indicated the relative abundance of the [1-14CN-acetyl 2-16O] CMP-NeuAc compound to be 25 % and the [1-14CN-acetyl 2-18O] CMP-NeuAc compound was 75 %. ESI-MS spectral data was obtained from a para llel nonradiolabeled synthesis of [2-18O] CMPNeuAc. Figure 2-4. Ring opening mechanism for NeuAc. O OH CO2HO O H N HO HO OH B: O HO O H N HO HO OH OH CO2

PAGE 44

29 Figure 2-5. 31P-NMR of [1-14CN-acetyl 2-18O] CMP-NeuAc. The (-)ESI-MS from the [2-18O] CMP-NeuAc synthesis shows the most abundant ions at m/z 615 [M-H]for [2-18O] CMP-NeuAc and m/z 613 [M-H]for the [2-16O] CMP-NeuAc compound (Figure 2-6). The [(M-H+Na)-H]adduct of m/z 615 was also present at m/z 637. The m/z 615 ion underwent MS/MS to yield the labeled m/z 324 [CMP-H]ion. Selected ion monitoring of the ions at m/z 615 ([2-18O] CMP-NeuAc) and m/z 613 ([2-16O] CMP-NeuAc) indicated a di stribution of 75.6 % and 24.4 %, respectively. [114 CN-acetyl 218 O] CMP-NeuAc [114 CN-acetyl 216 O] CMP-NeuAc ppm

PAGE 45

30 Figure 2-6. (-) ESI-MS of [2-18O] CMP-NeuAc m/z 615 [M-H]-(top panel), zoom MS/MS of [2-18O] CMP-NeuAc [M-H](center panel), and MS/MS dissociation of m/z 615 [M-H]ion (bottom panel). Synthesis of [1-14CN-acetyl P18O2] CMP-NeuAc Synthesis of the [1-14C -N-acetyl P18O2] CMP-NeuAc isotopomer was achieved using a multi enzymatic synthesis route to selectively incorporate 18O labels into the nonbridging phosphate oxygens of CM P-NeuAc (Figures 2-6 and 2-7).68 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.7616.6 614.4 673.5 637.6 324.7 695.5 671.5 581.9 328.6 610.5 611.0 611.5 612.0 612.5 613.0 613.5 614.0 614.5 615.0 615.5 616.0 616.5 617.0 617.5 618.0 618.5 619.0 619.5 620.0m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.5 616.3 617.2615.0 615.7 613.4 613.5 613.7 613.3 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 324.2322.5[2-18O] CMP-NeuAc[M-H]-[2-18O] CMP-NeuAc[M-H]-[2-18O] CMP-NeuAc[M] P18O16O3CMP [M] 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.7616.6 614.4 673.5 637.6 324.7 695.5 671.5 581.9 328.6 610.5 611.0 611.5 612.0 612.5 613.0 613.5 614.0 614.5 615.0 615.5 616.0 616.5 617.0 617.5 618.0 618.5 619.0 619.5 620.0m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.5 616.3 617.2615.0 615.7 613.4 613.5 613.7 613.3 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 324.2322.5 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500m/z 0 10 20 30 40 50 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.7616.6 614.4 673.5 637.6 324.7 695.5 671.5 581.9 328.6 610.5 611.0 611.5 612.0 612.5 613.0 613.5 614.0 614.5 615.0 615.5 616.0 616.5 617.0 617.5 618.0 618.5 619.0 619.5 620.0m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 615.5 616.3 617.2615.0 615.7 613.4 613.5 613.7 613.3 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 324.2322.5[2-18O] CMP-NeuAc[M-H]-[2-18O] CMP-NeuAc[M-H]-[2-18O] CMP-NeuAc[M] P18O16O3CMP [M]

PAGE 46

31 Figure 2-6. Enzymatic synthesis of [P18O3] CMP from KH2P18O4. The enzymes used in this synthesis were glyceraldehyd e-3-phosphate dehydrogenase (GAPDH), 3phosphoglycerate phosphokinase (3-PGK) and uridine kinase (UDK). Figure 2-7 Enzymatic synthesis of [1-14CN-acetyl P18O2] CMP-NeuAc from [P18O3] CMP. P18O3 CMP [ -P18O3] CDP Nucleoside Monophosphate Kinase Pyruvate Kinase [ -P18O3] CTP CMP-NeuAc Synthetase ADP ATP PEP Pyruvate [1-14CN-acetyl P18O2] CMP-NeuAc PEPPy r uva t e Pyruvate Kinase O N NH2O N O OH OH O P O18O18O CO2HO O H N HO OH OH*[1-14CN-acetyl ] NeuAc PPi O3POH O OHGlyceraldehyde-3-phosphateNAD+GAPDHNADH O3PO18OP18O3O OH1,3 DiphosphoglycerateADP3-PGKO3PO18OH O OH -P18O3] ATPUDKCytidine[P18O3] CMP KH2P18O4+ 3-Phosphoglycerate +N NH2O N O OH OH O P H18O O18H18O

PAGE 47

32 The first step of the synthesis involved the preparation of KH2P18O4 which was synthesized in 74 % yield via hydration of PCl5 with H2 18O (95 % atom enrichment) followed by the addition of 2 M KOH.69 This synthesis produced five different phosphate species (P18O4, P16O18O3, P16O2 18O2, P16O3 18O, and P16O4) due to the isotopic distribution of the 18O label. HPLC/ (-) ESI-MS and 31P NMR analysis of KH2P18O4 measured the relative isotopic abundance of the five phosphate species to be the following: 80.4 % 18O4, 16.2 % 16O18O3, 1.6 % 16O2 18O2, 0.1 % 16O3 18O, and 1.7 % 16O4 (Figures 2-8 & 2-9).70 These results are consistent w ith the statistical distribution of 18O for the synthesis of P18O4 using 95 % atom enriched H2 18O. The KH2P18O4 (8O % 18O4) compound was then used in an enzymatic synthesis with glyceraldehyde-3-phosphate, (G3P), NAD+, ADP, and cytidine to give P18O3 CMP in 64 % isolated yield after purification on anion-exchange HPLC. The last step in the synthesis of P18O3 CMP required the use of uridine kinase ( UDK) which was cloned, overexpressed in E. coli and purified by dye affinity chromatography.71 Enzyme purificati on yielded 150 units and uridine kinase (23 kDa) was 90-95 % pure base on SDS-PAGE analysis (Figure 2-10).

PAGE 48

33 Figure 2-8. (-) ESI-MS of KH2P18O4. 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 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 105.5 103.5 106.7 104.4 97.5 107.6 102.3 101.4 98.5 [H2P18O4][H2P16O18O3][H2P16O4] [H2P16O2 18O2] 99.4 [H2P16O3 18O]

PAGE 49

34 Figure 2-9. 31P-NMR spectrum of KH2P18O4 (1M) in D2O with 4 mM EDTA.

PAGE 50

35 Figure 2-10. 10 % SDS-PAGE of purified UDK fractions from Red-A dye affinity column. The sample load for each fraction was 20 L and the gel was stained with coomassie blue. The HPLC/MS analysis of P18O3 CMP showed four peaks at m/z 324, 326, 328 and 330 corresponding to the P16O4 CMP, P16O3 18O CMP, P16O2 18O2 CMP, P16O18O3 CMP compounds, respectively (Figure 2-11). An (+) ESI-MS scan measured the relative abundances of the various CMP isotopomers to be: 60 % 16O18O3, 6.4 % 16O2 18O2, 1.1 % 16O3 18O, 32.5 % 16O4. The results show an approximate 20 % dilution of the 18O label from the initial enrichment of P18O4 used in the synthesis. Th is may be explained by the fact that glyceraldehyde-3-phosphate decomposes at neutral pH to release its phosphate, thus resulting in a dilution of the [P18O4 2-]. In response to this result, a shorter incubation time and higher enzyme concentr ations were used to help minimize the decomposition of G3P in the reaction mixture. A higher [P18O4] was also used to reduce unlabeled phosphate incorporation. F76 F78 F80 66 45 36 29 24 20 14 MW

PAGE 51

36 Figure 2-11. (+) ESI-MS of P18O3 CMP 2 (upper panel) and zoom-MS of the [M+H]+ ions (lower panel). The [1-14CN-acetyl P18O2] CMP-NeuAc isotopomer was enzymatically synthesized using P18O3 CMP (60 % 18O3) and [1-14CN-acetyl ] NeuAc. The isolated yield was 35 % after purification on anion ex change HPLC. Since ESI-MS spectral data of a radiolabeled compound could not be obt ained, a parallel synthesis was conducted to determine approximate isotopic incorporation for the radiolabeled synthesis. The (+) ESI-MS spectra from the nonradiolabeled synthesis of [P18O2] CMP-NeuAc shows the most abundant ions at m/z 619 [M+H]+ for [P16O2 18O2] CMP-NeuAc and m/z 615 [M+H]+ for [P16O4] CMP-NeuAc (Figure 2-12). The [M+Na]+ adducts of m/z 619 and m/z 615 were also present at m/z 641 and m/z 637, respectively. The m/z 619 ion underwent MS/MS to yield th e labeled m/z 328 [CMP+H]+ ion which in a MS/MS/MS 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 m/z 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance 330.1 352.2 324.1 346.2 112.1 374.1 229.3 368.2 385.8 400.1 205.2 249.3 211.0 267.6 379.8 361.9 151.0 184.9 309.2 298.9 115.3 175.2 242.5 291.0 394.5 406.1 341.4 195.0 416.0 134.8 168.0 323 324 325 326 327 328 329 330 331 332 m/z 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance 329.9 329.9 323.9 324.0 324.1 323.9 329.8 328.0 331.0 325.0 324.1 331.1 323.8 327.9 325.0 329.0 330.9 324.9 329.7 325.9 328.1 326.1 323.6 331.2 327.1 330.6 327.8 322.7 325.5 328.9 322.3 326.8 323.0 331.8P18O3CMP [ M+H ] + P18O3 CMP [M+Na] + [cytidine+H] + P16O4 CMP [M+H] + P16O2 18O2 CMP [M+H]+ P16O3 18O CMP [M+H]+ P18O3 CMP [M+H] +

PAGE 52

37 scan produced the m/z 112 [Cytidine+H]+ ion. Other, less intense ions were detected at m/z 310 and m/z 292 which correspond to fragme nts of the NeuAc moiety. Selected ion monitoring of the ions at m/z 619 ([P16O2 18O2] CMP-NeuAc), m/z 617 ([P16O3 18O] CMPNeuAc), and m/z 615 ([P16O4] CMP-NeuAc) indicated a distribution of 56.6, 9.2 and 34.2 %, respectively. Figure 2-12. (+) ESI-MS spectrum of [P16O2 18O2] CMP-NeuAc. Synthesis of UMP-NeuAc Isotopomers Previous work conducted in the Horenstein laboratory showed that there is a commitment to catalysis for the CMPNeuAc donor substrate when bound to the enzyme.1 In other words, the sialyltransfer ase catalyzed reaction with CMP-NeuAc donor substrate contains more than one kineti c barrier that is pa rtially rate limiting. 590 595 600 605 610 615 620 625 630 635 640 645 650 655 660 665 670 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 618.9 614.9 640.9 636.9 651.1 655.0 616.9 633.0 646.9 653.0 619.9 638.0 642.0 657.1 649.1 623.0 639.0 659.1 669.0 614.1 660.1 623.9 634.8 642.9 630.4 667.1 664.0 604.5 591.8 643.9 610.0 605.4 601.6 628.7 596.4 P16O2 18O2 CMP-NeuAc [M+H]+ P16O4 CMP-NeuAc [M+H]+ P16O2 18O2 CMP-NeuAc [M+Na]+ P16O4 CMP-NeuAc [M+Na]+

PAGE 53

38 Furthermore, non-chemistry rate limiting steps ca n mask the full expression of the kinetic isotope effects. To circumvent this pr oblem, UMP-NeuAc, an unnatural “slow” donor substrate for sialyltransferase, was also synthesized to aid in the study of the sialyltransferase catalyzed mechanism. UMP-NeuAc binds more weakly to sialyltransferase and the chemistry step is sl ower than for CMP-NeuAc as indicated by its higher Km and lower kcat values.1 Thus, these factors made UMP-NeuAc an ideal donor substrate analog to use for the desi red set of kinetic experiments. UMP-NeuAc was synthesized with a vari ety of isotopic substitutions. The locations of these isotopic labels were equiva lent to those used in the synthesis of the CMP-NeuAc isotopomers (Figure 2-13). Th e isolated yields of the UMP-NeuAc isotopomers are shown in Table 2-2. Figure 2-13. Structure of labeled UMP-Ne uAc. Asterisks denote sites of isotopic substitution. Table 2-2. UMP-NeuA c isotopomer yields. UMP-NeuAc % Isolated Yield [9-3H] 47 [1-3HN-acetyl ] 33 [1-14CN-acetyl ] 42 [1-14CN-acetyl P18O2] 30 [1-14CN-acetyl 2-18O] 44 O HO OH HO AcHN HO COO O NH O N O OH OH O P O O * * H H * O 1 2 3 4 5 6 7 8 9

PAGE 54

39 Synthesis of the UMP-NeuAc isotopom ers was achieved through a chemical deamination reaction using 1 M NaNO2, pH 3.8 and the appropriately labeled CMPNeuAc compounds Figure (2-15). This reaction was challenging due to the instability of CMP-NeuAc under acidic conditions. Thus, deam ination reactions were carried out at 4 C to help minimize the decomposition of CM P-NeuAc. The conversion of CMP-NeuAc to UMP-NeuAc also varied depending on how the CMP-NeuAc compounds where purified on anion-exchange HPLC. De amination reactions using CMP-NeuAc isotopomers purified on anion-exchange HPLC with an ammonium bicarbonate buffer system proceeded slowly, and the convers ion of CMP-NeuAc to UMP-NeuAc was ~ 38 % with < 20 % decomposition after 48 hrs at 4 C (Figure 2-16). This result was explained by considering the concentration of free ammonia present in the solution of CMP-NeuAc after desalting the purifi ed compound with Amberlite IR-120 H+. Ammonia assays conducted on the purified and desalted CMP-NeuAc isotopomers estimated the [NH4 +] in solution to be ~ 50 mM. The excess NH4 + in the reaction solution slows the progress of CMP-Ne uAc deamination by reacting with NaNO2. This was especially the case when the [CMP-Neu Ac] in the reaction was in the micromolar range. Figure 2-15. Chemical deamination of CM P-NeuAc to UMP-NeuAc by sodium nitrite. O N NH2O N O OH OH O P O O OCO2HO O H N HO HO OH NaNO2N2O NH O N O O H OH O P O O OCO2HO O H N HO HO OH O CMPN euAc UMPN euAc

PAGE 55

40 Several ideas were tested to remove more of the NH4 + from the CMP-NeuAc solution such as, filtering the solution th rough a minicolumn of zeolite, applying the sample to a mini gel filtration column, and desalting the sample again with Amberlite IR120 H+. While these approaches were some what effective in removing the [NH4 +], the additional purification methods led to a further loss and decomposition of the radiolabeled CMP-NeuAc product. Additional NaNO2 was also added to the reaction mixture to expedite the deamination of CMPNeuAc, but this made the desired products more difficult to isolate and purify on ani on-exchange HPLC due to the increased complexity of the chromatograms. Figure 2-16. HPLC chromatogram of CMPNeuAc deamination reaction after 48 hr. The vertical lines represent the begi nning and end of fraction collection. To circumvent this problem, CMP-NeuAc is otopomers slated for deamination were purified on anion-exchange HPLC using a sodium bicarbonate buffer system. Deamination reactions using the sodium fo rm of CMP-NeuAc resulted in a mixture consisting of 11 % CMP-NeuAc, 66 % UMP-NeuAc, 3 % CMP, and 20 % UMP after 30 CMP-NeuAc CMP UMP-NeuAc UMP

PAGE 56

41 hours at 4 oC (Figure 2-17). These yields were cal culated from peak integration values from an HPLC chromatogram. This reacti on was advantageous because it allowed the synthesis of UMP-NeuAc isotopomers in high er yield without significant decomposition of the starting material and the final product. This aspect was essential in order to obtain enough UMP-NeuAc isotopomer to carry out the desired set of experiments. Furthermore, unreacted CMP-NeuAc could be recovered during pur ification of UMPNeuAc and recycled for use in experime nts requiring a CMP-NeuAc isotopomer. Figure 2-17. HPLC chromatogram of CMPNeuAc deamination afte r 30 hrs incubation. The vertical lines represent the begi nning and end of fraction collection. Experimental Materials Reagents and buffers were purchased fr om Sigma and Fisher and used without further purification. Recombinant rat liver (2 3) and (2 6) sialyltransferase was purchased from Calbiochem. The 18O water (95% atom enrichment) was purchased from CMP NeuAc CMP UMP NeuAc UMP

PAGE 57

42 Isotec and Medical Isotopes, Inc. N-acetyl D-mannosamine isotopomers ([1-14CNacetyl ] and [6-3H]) were purchased from Morave k and American Radiolabelled Chemicals. Glyceraldehyde-3-phosphat e (G3P), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) [EC 1.2.1.12] and 3-phosphoglycerate phosphokinase (3-PGK) [EC 2.7.2.3] were purchased from Sigma. Nucleoside monophosphate kinase (NMK) [EC 2.7.4.4] was purchased from Roche Applie d Science. Liquid scintillation fluid (ScintiSafe 30 %) was purchased from Fisher. The E. coli expression plasmid pWV200B haboring the E. coli CMP-NeuAc synthetase gene [EC 2.7.7.43] was a generous gift from Dr. W. F. Vann at the Nati onal Institutes of Health. Instrumental A Rainin HPLC system consisting of a HPXL binary pump and a model UV-1 detector was used. HPLC separations were performed on a Mono Q HR 10/10 anion exchange column (Amersham-Pharmacia) monitored at 271 nm. Data collection was achieved on a personal computer using the Star Workstation Ve rsion 6.2 software (Varian Inc.). A Rainin-Dynamax fraction co llector (model FC-1) was used to collect eluent samples from the HPLC. A Packar d 1600 TR instrument was used for liquid scintillation counting. Mass spectrometry (LC-MS) was performed on a ThermoFinnigan (San Jose, CA) LCQ in electrospray ionizati on (ESI) mode. The system was interfaced with an Agilent (Palo Alto, CA) 1100 bina ry pump HPLC system consisting of an Applied Biosystems Model 785A programmabl e absorbance detector set at 254 nm. HPLC separations for LC-MS were performe d on a Phenomenex (Torrace, CA) Synergi 4u Hydro-RP 80A C18 column (mobile pha ses = 0.5 % HOAc in water/0.5 % HOAc in methanol). 31P-NMR spectra were acquired on a 300 MHz Mercury NMR spectrometer. E. coli cells were lysed using a French pressu re cell with Carver hydraulic press.

PAGE 58

43 Centrifugation was performed with a Sorval l RC 5B centrifuge. A Labconco CentriVap Concentrator (speedvac) was used to concentrate small volume samples. Synthesis of [3H-N-acetyl] ManNAc [1-3HN-acetyl ] ManNAc was synthesized using a procedure adapted from that of Roseman et. al.66 Freshly prepared Dowex 1 x 8-200 mesh (CO3 2-) (250 mg damp) and D-mannosamine-HCl (12.3 mg, 56 mol) was suspended in 250 L of water and carefully added to the ampoule containing [3H] acetic anhydride (5 mCi, s.a.-100 mCi/mmol, 50 mol). The reaction was stirred in an ice bath for 4 hours. The aqueous solution was removed from the reaction solu tion and passed over an amberlite IR-120 H+ minicolumn. The eluate was collected in a 25 mL round bottomed flask. The reaction ampule was washed three times with 1 mL of water and the solution was passed over the amberlite column after each wash. Water (1 mL) was used to wash the amberlite minicolumn twice more. The 25 mL flask contai ning the eluate was fitted to a short path distillation apparatus with a 10 mL receiving flask. The solution was refluxed twice and then cooled to room temperature. The mixture was then concentrated to dryness in vacuo keeping the bath temperature below 55 C. The pot residue was twice resuspended in 2 mL of water and reconcentrated in vacuo The residue was dissolved in 500 L of water and [1-3HN-acetyl ] ManNAc was purified from the reaction mixture via multiple injections onto a HPLC C18, 1 x 30 column (5 % methanol (v/v), 95 % water, A220 nm, 1 mL/min). The [1-3HN-acetyl ] peak (r.t. 10 min) was collect ed after each injection. The [1-3HN-acetyl ] ManNAc fractions were combined, the solution was concentrated to dryness in vacuo, and the product was resusp ended in 1 mL of water. The product contained 150 Ci for an isolated yield of 6 %.

PAGE 59

44 Synthesis of [1-3HN-acetyl ] NeuAc and [1-3HN-acetyl ] CMP-NeuAc [1-3HN-acetyl ] D-mannosamine (30 Ci) wa s concentrated to dryness in vacuo using the speedvac. Phosphate buffer, pH 7 (40 mM, 100 L) c ontaining 1 mg/mL BSA and 1 mg/mL NaN3 was added to the 1.5 mL reaction tube. Sodium pyruvate (10 mg, 1 mmol) and 2 units of NANA aldolase were also added to the reaction tube.65 The reaction mixture was placed at room temperat ure for four days and judged to be >80 % complete by LSC counting of reaction aliquots analyzed by HPLC (2 mL HPLC Mono-Q fractions, 15 % methanol (v/v), 500 mM NH4HCO3, pH 8.0 10 – 20 % salt gradient, 2 mL/min). The retention times for [1-3HN-acetyl ] ManNAc and [1-3HN-acetyl ] NeuAc were 3 min and 6.5 min, respectively. The r eaction mixture was concentrated to dryness in vacuo using the speedvac a nd resuspended in 120 uL of 10 mM HEPES, pH 7.5 buffer containing NeuAc (10 mM, 0.4 mmol) and CTP (15 mM, 1 mmol). CMP-NeuAc synthetase (3 units) and 4 L of 2.5 M MnCl2 were also added to the reaction tube. The mixture was incubated at 37 C for 8 hours and the judged to be >80 % complete by LSC counting of reaction aliquots analyzed by HPLC (2 mL HPLC Mono-Q fractions, 15% methanol (v/v), 75 mM NH4HCO3, pH 8.0, isocratic, 2 mL/min). The product was purified by HPLC on a Mono-Q column. The [1-3HN-acetyl ] CMP-NeuAc fraction was collected, desalted w ith Amberlite IR-120 H+, concentrated to dryness in vacuo and the material was resuspended in 300 L of diH2O. The purified product contained 15 Ci for a yield of 78 %. This procedure was also used to synthesize othe r isotopomers of NeuAc and CMP-NeuAc by using the appropriat ely radiolabeled precursors.

PAGE 60

45 Cloning, Overexpression and Purification of N-acetylneuraminic Acid Aldolase [EC 4.1.3.3] The N-acetylneuraminic acid aldolase(NANA aldolase) gene62 was amplified from E. coli K12 genomic DNA using PCR (upper primer-5’ATGGCAACGAATTTACGTGGCGTAA3’ and lower primer 5’TCACCCGCGCTCTTGCATCAACTGC-3’). Th e gel purified PCR product was ligated into the pETBlue-1 vector and transformed into NovaBlue competent cells for plasmid amplification. The plasmid wa s purified using the QIAprep Spin Miniprep Kit (Qiagen) and subsequently used in a tr ansformation reac tion with TunerTM(DE3)pLacI competent cells. IPTG induction of a 2 L culture of the TunerTM(DE3)pLacI cells harboring the recombinant plasmid resulted in overexpression of the target enzyme. Nacetylneuraminic acid lyase was purified following the published protocol and was judged 90-95 % pure based on SDS-PAGE analysis.61,63,72 Overexpression of CMP-NeuAc Synthetase [EC 2.7.2.43] The original pWV200B plasmid harbori ng the CMP-NeuAc synthetase gene was given as a generous gift fr om Dr. W. F. Vann. The plasmid was transformed into E. coli JM109 cells via electroporation and the cells we re plated onto luria broth agar plates containing 60 g/mL ampicillin. Overni ght incubation of the plates at 37 C yielded ~ 200 colonies. Several colonies were selected from the plates with a sterile toothpick and used to inoculate a culture tube containing 5 mL of luri a broth supplemented with 100 g/mL ampicillin. The culture was grown in a 37 C shaking incubator (200 rpm) until an O.D.600 nm = 0.8 – 1.0. The 5 mL culture was then us ed to inoculate a 2 L culture of luria broth containing 100 g/mL ampicillin. The culture was grown in a 37 C shaking incubator (200 rpm) until an O.D.600nm = 0.6 – 0.8. The culture was induced with IPTG

PAGE 61

46 (1 mM final concentration) for 10 hours in a 37 C shaking incubator to overexpress CMP-NeuAc synthetase. The cells were ha rvested via centrifugation at 5000 rpm, 4 C, for 30 min. The cells were resuspended in 20 mL of purification buffer (10 mM HEPES, pH 7.0 containing 1 mM EDTA, 10 mM MgCl2, and 1 mM PMSF) and lysed using a French pressure cell and Carver hydraulic pre ss (2 runs). The supernatant was separated from cellular debris via centrifugation at 8000 rpm, 4 C for 30 min. CMP-NeuAc synthetase was purified from the supernatant by Red-A (Millipore) dye affinity column chromatography (2.5 x 7 in) using a linear salt gradient from 0-1 M KCl in purification buffer. Fractions were analyzed for prot ein using the Bradford assay method and for CMP-NeuAc synthetase activity using the published assay.73,74 CMP-NeuAc synthetase containing fractions were combined and con centrated in an Amicon concentrator to a final protein concentration of >10 mg/mL. The protein solution was then saturated with 80 % (NH4)2SO4 and the precipitate was stored at 4 C until further use. The yield was ~ 60 units and the CMP-NeuAc synthetase was 90-95 % pure based on SDS-PAGE analysis. Cloning, Overexpression and Purifica tion of Uridine Kinase [EC 2.7.1.48] The uridine kinase (UDK) gene71 was amplified from E. coli K12 genomic DNA using PCR (upper primer-5’-ATGACT GACCAGTCTCACCAGCAGTGCG-3’ and lower primer –5’-AAGCTTATTCAAAGAACT GACTTAT-3’). The PCR product was gel purified using the QIAquick Gel Extrac tion Kit (Qiagen) and ligated into the pETBlue-1 vector (Novagen). The recombinan t plasmid was transformed into NovaBlue (Novagen) competent cells and the plasmid was purified using QIAprep Spin Miniprep Kit (Qiagen). Transformation of TunerTM(DE3)pLacI competent cells with the construct and IPTG-induced overexpression of a 2 L cultu re yielded the target enzyme. Uridine

PAGE 62

47 kinase was purified by Red-A (Millipore) dye affinity column (2.5 x 7 in) using a linear salt gradient from 0-1 M KCl (0.1 M Tris-H Cl, pH 7.8, 4 mM EDTA). Fractions were analyzed for protein using the Bradford assay method and for UDK activity using the published assay.73,75 UDK containing fractions were combined and concentrated in an Amicon concentrator to a final volume of 2 mL. The yield was ~ 150 units and the UDK was 90-95 % pure based on SDS-PAGE analysis. Synthesis of 75 atom % [1-14C-N-acetyl, 2-18O] CMP-NeuAc A mixture of N-acetyl neuraminic acid (2 mg, 6 mol) and [1-14CN-acteyl ] neuraminic acid (15 Ci) was dissolved in 200 L of 1.9 mM glycine, 0.5 M MnCl2, pH 9.5 buffer. The solution was concentrated to dryness in vacuo and the material was resuspended in 300 L of H2 18O (95 % enrichment). The reaction mixture was placed at 37 C for 16 hrs. Cytidine-5-triphosphate (15 mg, 28 mol) was then added to the reaction tube along with CMP-NeuAc syntheta se (2 units). The pH of the reaction mixture was adjusted to pH 7.5 as nece ssary with 5 N NaOH. The reaction was incubated at 37 C for 8 hrs and judged to be >75 % complete by LSC counting of 2 mL HPLC Mono-Q fractions (15 % methanol (v/v), 75 mM NH4HCO3, pH 8.0, isocratic, 2 mL/min, A271). The product was purified by HPLC on a Mono-Q column. The [1-14CNacetyl 2-18O] CMP-NeuAc fraction was collected, desalted with Amberlite IR-120 H+, concentrated to dryness in vacuo and the material was resuspended in 300 L of diH20. The purified product contained 12 Ci for a yield of 71 %. Synthesis of KH2P18O4 KH2P18O4 was synthesized using a method si milar to that of Risley et al.69 The 18O water (95 % atom enrichment) (300 L, 15 mol) was added drop wise via syringe to a two necked flask containing phosphorus pe ntachloride (428 mg, 2 mmol). The PCl5

PAGE 63

48 was weighed in a dry box. Once removed, it was immediately placed on a Schlenk line under constant flow of N2 (g). This was done to reduce H2 16O contamination. The reaction was stirred at 0 C for 1 hr under constant flow of dry N2 (g). The flask was then warmed to room temperature and heated in a water bath at 100 C for 30 min. The remaining reaction solution was cooled to r oom temperature and ~ 2 mL of deionized water was added. The solution was titrated to pH 5 with 2 M KOH and KH2P18O4 was precipitated from solution by a ddition of 95 % ethanol. The precipitate was collected by concentration under re duced pressure. Synthesis of P18O3 CMP For the synthesis of P18O3 CMP a solution of KH2P18O4 (1M, pH 7.0) was made from which 250-350 L was mixed with gl yceraldehyde-3-phosphate (60 L, 2 mM), NAD (60 L, 3 mM), ADP (40 L, 1 mM), MgSO4 (50 L, 2.8 mM), glycine (50 mM), cytidine (40 L, 2.5 mM), GAPDH (2 units), 3-PGK ( 1 unit), and uridine kinase (2 units) in a 1.5 mL microfuge tube and in cubated at room temperature for 10 hrs.76 Special care was taken to minimize contamination by unlabeled phosphate in itially present in some of the reagents used in the reaction. Thus, ADP wa s freshly prepared and 3-PGK was dialyzed against 0.5 M Tris-HCl, pH 7.5 buffer to remove the orthopyrophosphate storage buffer provided by the manufacturer. The reaction solution was filtered through a microcon filtration unit (Millipore, MWCO 10 kDa) to remove enzymes and P18O3 CMP 2 was purified from the filtrate using isocratic, anion exchange HPLC (75 mM NH4HCO3, pH 8.5, 15% methanol, 2 mL/min.). The P18O3 CMP containing fractions were pooled and desalted with Amberlite IR 120-H+ cation-exchange resin. The solution was concentrated to dryness in vacuo and resuspended in 300 L of deionized water.

PAGE 64

49 Synthesis of [1-14C-N-acetyl, P18O2] CMP-NeuAc The [1-14CN-acetyl P18O2] CMP-NeuAc was synthesized using a method adapted from that of Ichikawa et al .77 The P18O3 CMP purified from above was concentrated to dryness in vacuo and used in an enzymatic reaction with ATP (5 mol), PEP monosodium salt (10 mol), MnCl2 (10 mol), MgCl2 (10 mol), NMK (2 units), PK (500 units) and 700 L of HEPES buffer (0.2 M, pH 7.5). The reaction mixture was incubated for 24 h at 25 C and then filtered through a microcon filtration device (Millipore, MWCO 10 kDa) to remove enzymes. This step is necessary to prevent the insitu recycling of P18O3 CMP and subsequent dilution of 18O labels resulting from the decomposition of CMP-NeuAc in the following synthetic step. [1-14CN-acetyl ] NeuAc (10-25 Ci) and CMP-NeuAc synthetase (3 units) were added to the filtrate and the reaction was incubated at 37 C for 6 h. CMP-NeuAc isotopomers were purified using isocratic, anion exch ange HPLC (75 mM NH4HCO3, 15 % methanol, pH 8.0, 2 mL/min). CMP-NeuAc fractions were pooled, de salted with Amberlite IR 120-H+ cation-exchange resin and concentrated in vacuo as previously described. Synthesis of UMP-NeuAc CMP-NeuAc was purified on HPLC Mono Q (15 % methanol (v/v), 75 mM NaHCO3, pH 8.0, isocratic, 2 mL/min). The CMP-NeuAc fraction was collected, desalted with Amberlite IR-120 H+, concentrated to dryness in vacuo and the material was resuspended in 300 L of diH2O. The solution of CMP-NeuAc (2 mM, 370 mols) was made 1 N with NaNO2 and adjusted to pH 3.5 – 4.0 with 1 N HCl while on ice. The reaction mixture was placed at 4 C for 30 hours. The reaction proceeds to 65 – 70 % completion with less than 20 % hydrolysis of st arting material in this time. This method

PAGE 65

50 was also used to synthesize the various UMP-NeuAc isotopomers by using the appropriate CMP-NeuAc isotopomers.

PAGE 66

51 CHAPTER 3 PURIFICATION AN D KINETIC CHARACTERIZA TION OF RECOMBINANT HUMAN ALPHA (2 3) SIALYLTRANSFERASE IV Introduction Complex carbohydrates and polysaccharides are biosynthesized in living systems via an intricate pathway of membrane-bound and secreted proteins. Many of these carbohydrate moieties are added by glycosyltransf erases to specific biomolecules as the “finishing touches” during post-tr anslational events in the Golgi apparatus of cells. The biosynthesis of sialylated glycans are gove rned by a unique group of enzymes in the glycosyltransferase family known as sialyltran sferases. Sialyltransferase are membranebound proteins that transfer sialic residues (NeuAc) from activated CMP-NeuAc to specific acceptor oligosaccharides, glycol ipids, and glycoproteins during posttranslational modification. The addition of th ese sialic acid residues assists in the regulation of a myriad of cellu lar processes that are of signi ficant biological importance. This knowledge has prompted the need to st udy the sialyltransferas e family of enzymes in order to probe the functional roles of si alic acid ‘capped’ bi omolecules required for many key biological processes. According to the most recent reviews on th e sialyltransferase family, approximately 20 distinct cDNA sequences for sialyltransferas es have been identified and cloned from various mammalian and bacterial sources.29,42,55 The most widely studied sialyltransferases are the (2 6) and (2 3) sialyltransferases from rat liver which are commercially available in recombinant forms. In this study, three truncated forms of

PAGE 67

52 recombinant human (2 3) sialyltransferase (h23STGal IV, EC 2.4.99.4), which lacked the first 61 amino acids coding for the NH2-signal anchor of the open reading frame, were overexpressed in Spodoptera frugiperda ( Sf-9) insect cells using a baculovirus expression vector. The NH2-terminal signal anchor sequence was replaced in all th ree enzymes with a cleavable canine insulin signal peptide to produce a soluble, catalytically active secreted protein.78,79 The other two recombinant form s of h23STGal IV contained a His6x-tag sequence at the Nand Ctermini of the sialyltransferase sequence. The three enzymes were overexpressed, purified via affi nity chromatography, and used to conduct the desired sets of kinetic experiments. Results and Discussion Overexpression and Purification of Recombinant Human (2 3) Sialyltransferase Isoforms The cDNA sequence of human (2 3) sialyltransferase fr om placenta has been reported79, which greatly facilitated the primer design for PCR amplification of the recombinant sialyltransferase isoforms. The cDNA clones encoding the h23STGal IV gene which lacked the transmembrane doma in and the creation of the pFastBacHTa vector harboring the recombinant h23STGal IV gene were prepared by Dr. Nicole Horenstein. To produce the enzyme in a more readily purified form, the N-terminal His6x tag sequence located downstream from the st rong polyhedron (pPolh) promoter in the pFastBacHTa vector was removed via a restriction digest with RsrII and BamHI endonucleases and replaced with a 114 bp in sert containing the se quence coding for the cleavable canine pancreas insulin signal peptide.78 This allowed the h23STGal IV enzyme to become a secreted protein. Two other recombinant h23STGal IV constructs containing Nand C-terminal His6x tags, additional to the insulin signal peptide, were

PAGE 68

53 prepared in a similar manner by Bronson Anatao.80 These recombinant His6x tag h23STGal IV constructs were created to simp lify the standard purif ication procedure. One construct contained an N-terminal His6x tag located between the insulin signal peptide and the catalytic doma in of recombinant h23STGal IV, while the other construct contained a C-terminal His6x tag located at the end of the recombinant h23STGal IV catalytic domain sequence. Figure 3-1 illustrates the order of the insulin peptide sequence, recombinant h23STGal IV gene sequence, and the His6x tag peptide sequence for each of the three constructs describe d above. The recombinant pFastBacHTa vectors harboring the three enzyme constructs were subsequently used to create the recombinant baculoviruses following the prot ocols outlined in the BEVS manual supplied by the manufacturer.81 Figure 3-1. Diagram of the recombinant h23STGal IV constructs. The insulin signal peptide sequence is represented in yellow, the truncated h23STGal IV sequence is represented in blue and the His6x-tag sequence is represented in purple. Ins h23STGalIVConstruct CtermHis-h23STGal IV Construct NtermHis-h23STGal IV Construct

PAGE 69

54 A baculovirus expression system using inse ct host cells was c hosen to overexpress recombinant h23STGal IV because of its reported success in overexpressing other catalytically active enzymes in the glycosyltransferase family.56,82-84 Baculoviruses are one of the most prominent viral pathogens aff ecting the insect species. This system is widely used because of its capability for e xpressing high levels of recombinant protein and because of its ability to provide the eukaryotic post-translational modifications required to produce active enzymes of this type Several glycosyltran sferases have also been successfully overexpressed in active form with other eukaryotic ho st cells such as Saccharomyces cerevisiae and methylotrophic yeast Pichia pastoris .83,85,86 Attempts to overexpress recombinant mammalian sialyltransferases in E. coli resulted in the production of an insoluble, inactive form of the enzyme.87 This outcome is presumably because E. coli lacks the capability to suffi ciently glycosylate mammalian sialyltransferases during post translational m odification in the Golgi apparatus to allow for proper folding of the enzymes into an activ e form. In our hands, attempts to express the sialyltransferase gene in P. pastoris were also unsuccessful.88 Recombinant baculovirus was generated by first cloning the truncated h23STGal IV sialyltransferase construc ts into the pFastBacHTa donor plasmids and transforming the recombinant vectors into competent DH10Bac E. coli (Figure 3-2). Site-specific transposition was used to insert the recombinant sialyltransferase constructs into bacmid DNA provided in the E. coli host cells. Recombinant bacmid DNA was purified from E. coli and used to transfect Sf-9 insect cells. Recombinant baculovirus particles were generated after incubating the tranfection mixt ure with the insect ce lls for several days. Baculovirus stocks were amplifie d several times by infecting fresh Sf-9 cultures grown to

PAGE 70

55 a density of 2 x 106 cells/mL until a viral titer of 1 x 107 to 1 x 108 pfu/mL was achieved. Large scale expression of the recombinant h23STGal IV was achieved by infecting liter cultures of Sf-9 insect cells with the amplified recombinant baculovirus stocks. Recombinant h23STGal IV was expressed and secreted into the culture medium during the late phase of the viral lif e cycle. The average h23STGal IV activity accumulated in the cell supernatant of a 1 L Sf-9 cell culture was (2-3 U/L) after 72 hours post infection with recombinant baculovirus. Purifica tion of recombinant Ins-h23STGal IV was achieved using sepharose CDP-hexanolamine affinity column chromatography. Since sepharose CDP-hexanolamine resin is not commercially available, the CDPhexanolamine ligand was synthesized and attached to a CNBractivated sepharose 4B resin as previously described in the literature (Figure 3-3).89,90 Table 3-1 summarizes the data obtained from the purif ication of a 370 mL scale Sf-9 expression of Ins-h23STGal IV on sepharose CDP-hexanolamine. The results shown in Table 3-1, Figures 3-4 and 35 were obtained in collaboration with Jeremi ah D. Tipton of the Horenstein laboratory.91

PAGE 71

56 Figure 3-2. General scheme for the generati on of recombinant baculoviruses and protein expression with the BAC-TO-BAC expression system.81 Figure 3-3. Structure of CDP-Hexanolami ne affinity ligand synthesized for the purification of recombinan t Ins-h23STGal IV enzyme. H2N O P O O O P O O O O OH OH N N NH2O

PAGE 72

57 Table 3-1. Recombinant Ins-h23STGal IV purification table. Step Volume (mL) Activity (mU) Yield (%U) Concentration (mg/mL) Total (mg) Specific Activity (U/mg) Crude Supernatant 370 823 100 0.080 29.6 0.029 Amicon Ultrafiltration 80 767 94 0.440 35.2 0.022 CDPhexanolamine pooled fractions 28 227 28 0.050 1.40 0.162 Concentration 5 159 20 0.241 1.20 0.132 The crude supernatant was concentrated to a volume of 80 mL with an Amicon Ultrafiltration unit fitted with a polyether sulfone membrane (MWCO 10 kDa) prior to purification on the sepharose CDP-hexanolamin e column. This step was necessary to reduce the volume of supernatant applied to th e affinity column in order to expedite the purification process. Glycer ol (20 % v/v) and Triton CF-5 4 (0.01 % v/v) were added to the crude supernatant to help stabilize recombinant h23STG al IV during concentration and purification. The crude supernatant was loaded onto the sepharose CDPhexanolamine affinity column and the column was washed with three column volumes of purification buffer containing 300 mM -lactose prior to eluti ng the column with a KCl step gradient. This step was necessary to elute gp64, a predominant baculovirus membrane glycoprotein responsib le for virus-cell fusion, that was found to co-elute with recombinant h23STGal IV in earli er pilot purification trials.84,88,92 Gp64 is believed to act as an acceptor substrate for recombinant h23STGal IV. Since -lactose serves as an acceptor substrate for recombinant h23STGal IV, enzyme-gp64 binding interactions can be disrupted by washing the column with high concentrations of -lactose.

PAGE 73

58 Purified recombinant h23STGal IV was obtained in 28 % yield upon elution of the affinity column with a KCl step gradient. An activity assay and a protein assay were performed on select fractions to determine the elution profile of the enzyme from the affinity column (Figure 3-4). The final yi eld of purified recombinant h23STGal IV was 20 % with a specific activity of 0.132 U/mg af ter concentrating the pooled fractions with an Amicon Ultrafiltratio n unit. Purified recombinant h23STGal IV gave three bands in the SDS-PAGE gel with sizes of ~ 40 kDa, 37 kDa, and 35 kDa, which represent three different glycoforms of the enzyme (Figure 35). These results are consistent with the results from a protein sequence analysis of recombinant h23STGal IV of MS experiments where four potential asparagine N-linked gl ycosylation sites were identified following a general protein sequence motif of N/ X/S (X represents any amino acid).91,93,94 Digestion experiments performed by Jeremiah D. Tipt on of the Horenstein laboratory on the purified recombinant Ins-h23STGal IV with PNGase, gave one pr otein band on an SDSPAGE gel at 33 kDa, which is the exp ect size for the deglycosylated protein.91,95,96 After several purif ication trials, it was found that recombinant Ins-h23STGal IV appeared to be sensitive to the concentra tion step following purification as shown in Table 3-1 by the decrease in specific activity from 0.162 to 0.132 U/mg. The reason for this result is unclear, but one explanation may be that th e enzyme denatures during the concentration step. Surfactants such as Tr iton CF-54 and Tween-80 have been shown to help stabilize enzyme activity during the purification of other recombinant sialyltransferases.84,86,97,98 Therefore, Triton CF-54 ( 0.01 % v/v) was added to the enzyme buffer to stabilize the enzyme and he lp minimize the loss of activity during this step. Additionally, a significant loss in enzyme activity was observed when a

PAGE 74

59 concentrated sample of recombinant Ins-h23STG al IV was diluted and reconcentrated. In this case, the enzyme may denature upon d ilution and appears to be incapable of renaturing into active form when concentrated again. As a result, this step was avoided during the purification of recombinant Ins-h23STGal IV. Figure 3-4. Typical elution chromatogram of recombinant Ins-h23STGal IV from a sepharose CDP-hexanolamine affinity column. Solid squares represent protein concentration and ope n diamonds represent activity. 0 0.05 0.1 0.15 0.2 0.25 0.3 261014182226303438424650545862667074 Fraction #Protein Concentration (mg/mL)0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045Activity (U) 250 mM NaCl350 mM NaCl450 mM NaCl

PAGE 75

60 Figure 3-5. 10 % SDS-PAGE of purified r ecombinant Ins-h23STGal IV. Lane 1, MW standard; Lane 2, Lane 3, and Lane 4 are 10 L, 20 L, and 30 L loads of a TCA precipitation of purified Ins-h23STG al IV, respectively(left gel). SDSPAGE of purified recombinant Ins-h 23STGal IV digested with PNGase.91 Lane 1, MW standard; Lane 2, 30 L lo ad of a TCA precipitation of purified Ins-h23STGal IV and Lane 3 is 30 L load of PNGase digested Ins-h23STGal IV. The gels were stained with coomassie blue. Ni2+-NTA affinity chromatography was employed for the purification of NtermHisand CtermHis-h23STGal IV enzymes. Glycerol (20 % v/v) and Triton CF-54 (0.01 % v/v) were added to the crude supern atant to stabilize th e enzymes during the purification. A 300 mM -lactose wash was also performe d as described previously to remove gp64 prior to eluting the column with an imidazole step gradient. Purified recombinant NtermHis-h23STGal IV and CtermH is-h23STGal IV were obtained in 47 % and 32 % yield, respectively, after eluting Ni2+ the affinity column with an imidazole step gradient. An activity assay and a protein a ssay were performed on select fractions to determine the elution profile of the enzymes from the affinity column (Figure 3-6). The final yield of purified recombinant NtermH is-h23STGal IV and CtermHis-h23STGal IV 1 1 2 3 4 2 3 66 kDa 45 36 29 24 20 14

PAGE 76

61 was 12 % and 43 % with specific activities of 0.022 U/mg and 0.161 U/mg, respectively after concentration and dialysis (Tables 3-2 & 3-3). The resu lts shown in Tables 3-2 and 3-3 and in Figure 3-6 were obtained in colla boration with Jeremiah D. Tipton of the Horenstein laboratory.91 Purified recombinant Nter mHis-h23STGal IV and purified recombinant CtermHis-h23STGal IV gave thr ee bands in the SDS-PA GE gel with sizes of ~ 42 kDa, 41 kDa, and 40 kDa, which repr esent three different glycoforms of the enzyme (Figure 3-7). The size of the enzyme glycoforms for the recombinant NtermHisand CtermHis-h23STGal IV are larger th an for the recombinant Ins-h23STGal IV enzyme glycoforms. Since the addition of the His6x tag peptide sequence in these enzymes would only add an additional 1000 Da to the size of the recombinant Insh23STGal IV glycoforms, the reason for the larg er sizes is unclear. One explanation for the size discrepancy may be that the presence of the His6x tag alters the addition of glycan chains during post translati onal modification to produce r ecombinant sialyltransferase with a higher degree of glycosylation. Table 3-2. Recombinant NtermHis-h23STGal IV purification table. Step Volume (mL) Activity (mU) Yield (%U) Concentration (mg/mL) Total (mg) Specific Activity (U/mg) Crude Supernatant 460 419 100 0.240 70.5 0.004 Ni2+-NTA Pooled Fractions 30 190 47 0.050 1.50 0.126 Concentration & Dialysis 4.5 50 12 0.500 2.25 0.022

PAGE 77

62 Table 3-3. Recombinant CtermHish23STGal IV purification table. Step Volume (mL) Activity (mU) Yield (%U) Concentration (mg/mL) Total (mg) Specific Activity (U/mg) Crude Supernatant 470 560 100 0.150 71 0.008 Ni2+-NTA Pooled Fractions 50 181 32 0.03 1.5 0.121 Concentration & Dialysis 5.5 243 43 0.273 1.5 0.161 The recombinant NtermHis-h23STGal IV enzyme also appeared to be sensitive to the concentration step after purification as s hown in Table 3-2 by the decrease in specific activity from 0.126 to 0.022 U/mg. Concentrati on and dialysis of this enzyme without the presence of a surfactant such as Triton CF-54, resulted in a total loss of enzyme activity. Therefore, as for the recombinan t Ins-h23STGal IV, Tr iton CF-54 (0.01 % v/v) was added to the enzyme buffer prior to c oncentration and dialys is. The recombinant CtermHis-h23STGal IV enzyme, however, did not lose activity duri ng the concentration and dialysis step as seen by the increase in specific activity from 0.121 to 0.161 U/mg. This may be due to the overall stability of the recombinant CtermHis-h23STGal IV enzyme in comparison to the recombinant Nter mHis-h23STGal IV. It is reasonable to consider that the position of the His6x tag on the Nor C-terminus may change the protein folding structure in a manner that would make recombinant NtermHis-h23STGal less stable than CtermHis-h23STGal IV during c oncentration. Additiona lly, the position of the N-term His6x tag could also interfer e with the addition of gl ycan chains during post translational that are essential for enzyme stability and activity. This effect has been observed for other His6x-tagged glycosyltransferases.83,99 The specific activity of the

PAGE 78

63 recombinant CtermHis-h23STGal IV enzyme was similar to that of recombinant Insh23STGal IV, but the yield was higher th an for the recombinant Ins-h23STGal IV. Figure 3-6. Typical elution chromatogram of recombinant NtermHis-h23STGal IV and CtermHis-h23STGal IV from a Ni2+-NTA affinity column. Solid squares represent protein concentration and open diamonds represent activity. Figure 3-7. 10 % SDS-PAGE gel of purif ied recombinant CtermHis-h23STGal IV NtermHis-h23STGal IV. Lane 1, TCA precipitation of purified CtermHish23STGal IV; Lane 2, MW Standard; La ne 3, TCA precipitation of purified NtermHis-h23STGal IV. The gel was stained with coomassie blue. 1 2 66 kDa 45 36 29 24 20 14 3 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 01230384857677576777879808182838486 Fraction #Protein Concentration (mg/mL)0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016Activity (U)

PAGE 79

64 Kinetic Characterization of Reco mbinant h23STGal IV Isoforms The kinetic parameters obtained for the three recombinant isoforms of human (2 3) sialyltransferase using CMP-NeuAc and -lactose as the donor-acceptor substrate pair were estimated by fitting the in itial velocity kinetic data to the MichaelisMenten equation using a least squares analysis in Sigma Plot ver. 9.0 (Figures 3-8 – 310). Recombinant Ins-h23STGal IV, Nter mHis-h23STGal IV, and CtermHis-h23STGal IV enzymes were concentr ated a second time to speci fic activities of 0.05, 0.012, and 0.042 U/mg, respectively as determined by activ ity assays prior to their use in steady state kinetic experiments in order to provide concentrated enough enzyme samples. The measured kinetic parameters were sim ilar to those obtained for a wild type (2 3) sialyltransferase purified from human placenta (Table 3-4).97 The CMP-NeuAc Km values for NtermHis-h23STGal IV and Ins-h23STGal IV were 82 5 M and 74 8 M, repectively. These Km values were on the same order of magnitude as the wild type (2 3) sialyltransferase from human placenta a nd were comparable to the literature Km value of 74 M obtained fo r recombinant rat liver (2 3) sialyltransferases expressed from insect cells.56 However, the CMP-NeuAc Km value obtained for CtermHish23STGal IV was ~3.5 fold higher than for the NtermHis-h23STGal IV and Insh23STGal IV enzymes. This reason for this is unclear and literature kinetic data on other C-terminal His6x tagged sialyltransferases have not been reported. However, a few hypotheses for this result are th at the addition of the CtermHis6x tag on recombinant h23STGal IV either interferes with the binding site for CMP-NeuAc or that it changes the structural fold of the enzyme in a ma nner that alters CMP-NeuAc binding. The -lactose Km values for all three reco mbinant isoforms of human (2 3) sialyltransferase were similar to each other and for the wild type enzyme within experimental error.97

PAGE 80

65 Table 3-4. Kinetic parameters for sialyltransferase isoforms. CMPNeuAc -Lactose Recombinant Enzyme Km (M) Vmax (molmin-1mg-1) Km (mM) Vmax (molmin-1mg-1) WT h23STGal IV97 63 220 40 0.12 0.01 Ins-hST3Gal IV 82 5 0.072 0.002 171 18 0.072 0.004 NtermHis-hST3Gal IV 74 8 0.008 0.0003 155 14 0.020 0.001 CtermHis-hST3Gal IV 267 200.041 0.002 158 11 0.041 0.002 The Vmax value obtained for Ins-ST was 0.072 0.004 mol/(min mg) which is slightly lower than the 0.12 0.01 mol/(min mg) Vmax reported for the wild type human placenta (2 3) sialyltransferase when CMP-NeuAc and -lactose were used as the donor-aceptor substrate pair.97 This difference in Vmax values was not large and reflects the observed loss in enzyme specific activity after concentrating the enzyme. The Vmax values obtained for CtermHis-h23STGal IV and NtermHis-h23STGal IV were 3-6 fold lower than that of the wild type sialyltransfer ase. However, these values were similar to the specific activitie s obtained for these enzymes after the second concentration step described above. If the recombinant enzymes were not concentrated after purification, than their Vmax values may have been more comparable to that reported for wild type h23STGal IV enzyme.

PAGE 81

66 Figure 3-8. Michaelis-Menten plots for recomb inant Ins-h23STGal IV with with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ -lactose] and constant [CMP-NeuAc] (bottom panel). [ -lactose], mM 050100150200250300350 /Et ( mol/min/mg) 0.00 0.01 0.02 0.03 0.04 0.05 Experimental Predicted [CMP-NeuAc], M 050100150200250300350 /E t ( mol/min/mg) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Experimental Predicted

PAGE 82

67 Figure 3-9. Michaelis-Menten plots for reco mbinant NtermHis-h23STGal with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ -lactose] and constant [CMP-NeuAc] (bottom panel). [CMP-NeuAc], M 050100150200250300350 Et ( mol/min/mg) 0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 Experimental Predicted [ -lactose], mM 050100150200250300350 /Et ( mol/min/mg) 0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 Experimental Predicted

PAGE 83

68 Figure 3-10. Michealis-Menten plots for recombinant CtermHis-h23STGal IV with varied [CMP-NeuAc] and constant [ -lactose] (top panel) and with varied [ lactose] and constant [CMP-NeuAc] (top panel) and for with (lower panel). [CMP-NeuAc], M 0100200300400500 Et ( mol/min/mg) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 Experimental Predicted [ -lactose], mM 050100150200250300350 Et ( mol/min/mg) 0.000 0.005 0.010 0.015 0.020 0.025 0.030 Experimental Predicted

PAGE 84

69 Conclusions Overall, the recombinant NtermHis-h23S TGal IV and CtermHis-h23STGal IV enzymes were significantly easier to purify than the recombinant Ins-h23STGal IV enzyme. Furthermore, the Ni2+-NTA resin can readily obtai ned from commercial sources thus allowing for larger scal e expressions of recombinant enzyme to be purified more rapidly. This is not the case for the recomb inant Ins-h23STGal IV enzyme which requires a tedious synthesis of a sepharose CDP-Hexanol amine affinity resin pr ior to purification. The protein expression scale is also limited by the amount of affinity resin synthesized. Of the three enzymes, recombinant CtermHish23STGal IV enzyme gave the best results with the highest purification yield and sp ecific activity. This enzyme also had comparable kinetic parameters to the reco mbinant Ins-h23STGal IV enzyme except for the slightly larger Km value obtained for CMP-NeuAc. Experimental Materials and Methods Reagents and buffers were purchased fr om Sigma and Fisher and used without further purification. The restriction enzymes, E. coli strains JM109 & ER2925, Klenow, Large Fragment (DNA Polymerase I), and T4 DNA ligase were purchased from New England Biolabs. Shrimp Alkaline Phos phatase (SAP) was purchased from Roche Molecular Biology. The Wizard Plus Minipreps Kit and dNTPs were purchased from Promega. The QIAquick Nucleotide Removal Kit and QIAquick Gel Extraction Kit were purchased from Qiagen. The BCA Protei n Assay Kit was purchased from Pierce. The BAC-TO-BAC Baculovirus Expression System (BEVS), Spodoptera frugiperda ( Sf 9 ) insect cells, BACPACKTM Baculovirus Rapid T iter Kit, and DH10 BAC E. coli competent cells were purchased from Invitrogen. Human placental cDNA was obtained

PAGE 85

70 from Clontech. Primers for cloning and PCR analysis were obtaine d from Integrated DNA Technologies. The protocol for reco mbinant virus preparation is found in InvitrogenÂ’s instruction manual for BEVS version D April 6, 2004 (www.invitrogen.com). The sepharose CDP-he xanolamine affinity column was prepared as per the literature.89, 100 N-acetyl neuraminic acid (NANA) aldolase [EC 4.1.3.3] used in the synthesis of [9-3H] neuraminic acid (NeuAc) was cloned, overexpressed in E. coli and purified according to literature procedures.61,63,72 The E. coli expression plasmid pWV200B haboring the E. coli CMP-NeuAc synthetase ge ne [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 for sialyltransferase activity determination and kinetic experiments were analyzed with a P ackard 1600 TR liquid scintillation analyzer. DNA se quencing was performed at the University of Florida ICBR DNA sequencing core. Preparation of pFastBacHTaInsulin /h23STGal IV (Ins-h23STGal IV) MAX EFFICIENCY DH10BAC E. coli cells transformed with the pFASTBAC plasmid containing the recombinant h23STGal IV gene were previously prepared in our lab by Dr. Nicole Horenstein. A canine in sulin construct located upstream from the h23STGal IV was also cloned into the recombin ant plasmid to allow for secretion of the enzyme into Sf-9 insect cell media. The construc t, pFastBacHTaIns ulin/h23STGal IV, was submitted for DNA sequencing to confir m the presence of the canine insulin secretion peptide insert. The protocol for baculovirus preparation is found in the manual BAC-To-BAC Baculovirus Expression Syst ems provided by Invitrogen.81 The recombinant bacmid DNA was purified from E. coli using mini-prep procedures81 and

PAGE 86

71 subsequently used in the transfection of Sf-9 insect cells to produce recombinant baculovirus particles. Preparation of pFastB acHTaInsulin/NtermHis6x-tag-h23STGal IV Plasmid (NtermHis-h23STGal IV) A fresh glycerol stock of E. coli JM109 cells harboring the pFastBacHTaInsulin/h23STGal IV plasmid was used to inoculate a 5 mL luria broth culture supplemented with 100 g/mL ampicillin. The culture was grown overnight at 37 C (200 rpm) and the plasmid was subsequently isolated using the Wizard Plus Minipreps kit. The isolated plas mid was digested for 12 hours at 37 C with BamHI Clean-up of the enzymatic digesti on was performed using the QIAquick Nucleotide Removal Kit. The BamHI digested plasmid was digested for 6 hours at 37 C with StuI After thermally inactivating StuI at 65 C for 20 min the doubly digested plasmid was agarose gel purified using the QIAquick Gel Extraction Kit. Th e 5Â’ ends generated by digestion were dephosphorylated using SAP. The entire NtermHis6x tag insert was created by allowing two complimentary primers to anneal and then be extended by Klenow, Large Fragment. Primers were mixed in equal volumes (2.5 L) of NtermHISFor_Upper (5Â’GGTAGGCCCTGGCCATTAAGCGGATGCT GGAGATGGGAGCTATCAAGAACCT CACGTCC-3Â’), (50 M, 0.80 g/ L) and NtermHisRev_Lower (5Â’AGCAGGCCTTGCTCTCTG CCTCACCCTGGAGGAGGCA CGGCTCCTTCTTCTCG CC-3Â’), (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 11.16 L deionized water, 420 M of a dNTP mixture, 10 mM Tris-HCl pH 7.5, 5 mM MgCl2, 7.5 mM dithiothreitol, and 5 U of Klenow, Large Fragment. This

PAGE 87

72 reaction mixture was then allowed to incubate in a 25 C water bath for 80 minutes. After inactivating the polymerase at 75 C for 20 minutes, the mixture was cooled to room temperature and then doubly digested with BamHI and StuI at 37 C for 9 hours. Following heat inactivation of StuI a clean-up of the enzymatic reaction was performed using the QIAquick Nucleotide Removal Kit. The NtermHis6x tag insert was then ligated into the BamHI/StuI sites of the pFastBacHTaInsu lin/h23STGal IV vector. The new construct, pFastBacHTaInsulin/NtermHis6x-tag-h23STGal IV, was submitted for DNA sequencing to confirm the presence of the NtermHis6x tag insert. Isolation of bacmids and generation of baculovirus stocks followed the BEVS protocol.81 Preparation of pFastB acHTaInsulin/CtermHis6x-tag-h23STGal IV Plasmid (CtermHis-h23STGal IV) The pFastBacHTaInsulin/h23STGal IV 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 Dcmsensitive restriction enzymes such as Eco01091 The isolated plasmid was digested with Eco0109I for 10 hours at 37 C. Clean-up of the enzymatic diges tion was performed using the QIAquick Nucleotide Removal Kit. The Eco0109I digested plasmid was digested for 8 hours at 37 C with XhoI After thermally inactivating Eco0109I and XhoI at 65 C for 20 min., the doubly digested plasmid was agarose ge l purified using the QIAquick Gel Extraction Kit. The 5Â’ ends generated by digestion were dephosphorylated using SAP. The CtermHis6x tag insert was created following a si milar procedure to that of the NtermHis6x tag insert. Primers were mixed in equal volumes (2.3 L) of CtermHISFor_Upper (5Â’GGTAGGCCCTGGCCATTAAGCGGATGCT GGAGATGGGAGCTATCAAGAACCT

PAGE 88

73 CACGTCC-3Â’), (50 M, 0.882 g/ L) and CtermHisRev_Lower (5Â’AGCCTCGAGTTAGTGATGGTGATGGTG ATGACCGCCGAAGGACGTGAGGTTC TTGATAGC-3Â’), (50 M, 0.917 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 11.5 L deionized water, 463 M of a dNTP mixture, 10 mM Tris-HCl pH 7.5, 5 mM MgCl2, 7.5 mM dithiothreitol, and 5 U of Klenow, Large Fragment. This reaction mixture was then allowed to incubate in a 25 C water bath for 2 hrs. After inactivating the pol ymerase, the mixture was cooled to room temperature and then digested with Eco0109I for 16 hours at 37 C. The enzymatic reaction was cleaned up using the QIAquick Nucleotide Removal Kit. The cl ean mixture was digested with XhoI at 37 C for 8 hours. XhoI was then thermally inactivated. The CtermHis6x tag insert was then ligated into the Eco01091/XhoI sites of the pFastBacHTaInsulin/h23STGal IV vector. The new construct, pFastBacHTaInsulin/CtermHis6x-tag-h23STGal IV, was submitted for DNA sequencing to confirm the presence of the CtermHis6x tag insert. Isolation of bacmids and generation of baculovirus stocks followed BEVS protocol.81 Amplification of Recombinant Baculovirus Plasmids Recombinant baculovirus stocks, as prep ared following the BEVS protocol, were amplified four times until a titer of 1 x 107 3 x 108 pfu/mL was obtained.81 For amplification and expression, cell cu ltures (50 mL) containing 2 x 106 cells/mL in 250 mL borosilicate shaker flasks were infected with 3 mL of 1.7 x 108 pfu / mL Insh23STGal IV, 3 mL of 2.1 x 108 pfu / mL NtermHis-h23STGal IV, or 3 mL of 2.7 x 108 pfu / mL CtermHis-h23STGal IV.

PAGE 89

74 Expression and Purification of Ins-h23STGal IV A 1 L culture of Sf-9 insect cells (20, 250 mL shak er flasks containing 50 mL of culture each) at concentrations of 2 x 106 cells/mL were infected with 3 mL of amplified baculovirus stocks( ~2 x 108 pfu/mL) harboring the recombinant sialyltransferase expression constructs. The culture s were incubated for 70 hrs at 27 C after which the cultures were combined and centrifuged at 14,000 rpm, 4 C, for 30 min to pellet the cells. The supernatant was harvested and c oncentrated to 50 -100 mL using a 200 mL Amicon Ultrafiltration unit equipped with a polyethersulfone membrane (MWCO 10 kDa). This concentration step was only done for the Ins-h23STGal IV expressions in order to expedite the purif ication process on the CDP-he xanolamine affinity column. The concentrated supernatant was brought to 20 % (v/v) glycer ol and 0.01 % (v/v) Triton CF-54 and applied to the sepharose CDP-hexanolamine affinity column (1.7 x 12 cm) previously equilibrated with purification buffer (50 mM MES, pH 6.8 buffer containing 300 mM -lactose, 20 % (v/v) glycerol, a nd 0.01 % (v/v) Triton CF-54) at 4 C. The column was then washed with at least three column vol umes of purification buffer to remove gp64 glycoprotein from reco mbinant Ins-h23STGal IV. Ins-h23STGal IV was purified from the supernatant by eluti ng the column with a KCl step gradient of 50 mM, 250 mM, and 400 mM KCl in 50 mM MES, pH 6.8 buffer with 20 % (v/v) glycerol and 0.01 % (v/v) Triton CF-54. Fractio ns were analyzed for protein and activity using the Bradford assay met hod the published activity assay.56,73 Fractions containing activity were pooled, dialyzed, and concentrated Bradford and bicinchoninic acid (BCA) assays were used to esti mate protein concentration.73,101

PAGE 90

75 Expression and Purification of NtermHis -h23STGal IV and CtermHis-h23STGal IV Isoforms A 1 L culture of Sf-9 insect cells (20, 250 mL shak er flasks containing 50 mL of culture each) at concentrations of 2 x 106 cells/mL were infected with 3 mL of amplified baculovirus stocks( ~2 x 108 pfu/mL) harboring the recombinant sialyltransferase expression constructs. The culture s were incubated for 70 hrs at 27 C after which the cultures were combined and centrifuged at 14,000 rpm, 4 C, for 30 min to pellet the cells. The clarified supernatant was brought to 20 % (v/v) glycer ol and 0.01 % (v/v) Triton CF-54 and loaded onto a Ni2+-NTA column (2 x 10 cm) previously equilibrated with 50 mM MES, pH 6.8 buffer containing 5 mM imidazole, 100 mM KCl, 20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF-54 at 4 C. After loading the supernatant onto the column, the column was washed with at leas t three column equivalents of 50 mM MES, pH 6.8 buffer containing 300 mM -lactose, 5 mM imidazole, 100 mM KCl, 20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF-54, followe d by a 5 column equivalent wash with 50 mM MES, pH 6.8 buffer containing 5 mM imidazole, 100 mM KCl, 20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF-54. NtermHis-h23STGal IV and CtermHish23STGal IV enzymes were eluted from the co lumn using imidazole step gradient of 50 mM, 75 mM, and 120 mM imidazole in 50 mM MES, pH 6.8 buffer with 100 mM KCl, 20 % (v/v) glycerol, and 0.01 % (v/v) Triton CF -54. Fractions were analyzed for protein using the Bradford assay method a nd for sialyltransferase activity using the published assay.56,73 Fractions containing activit y were pooled, dialyzed, and concentrated. Bradford and BCA assa ys were used to estimate protein concentration.73,101

PAGE 91

76 Sialyltransferase Enzyme Activity Assays All activity assays reported for the r ecombinant h23STGal IV isoforms were performed using methods described by Paulson et al .56,90 The assay contained a mixture of [9-3H] CMP-NeuAc (100-170 M, 20,000 cpm, s.a.= 4-7 Ci/mol) and 235 mM lactose in 50 mM MES, pH 6.8 buffer containing 0.01 % (v/v) Triton CF-54, and 1 mg/mL BSA. A 10 L aliquot from a selected sialyltransferase containing sample was incubated with 10 L of the [9-3H] CMP-NeuAc/ -lactose mixture for the appropriate amount of time to limit the consumption of CMP-NeuAc to < 10 %. The reaction mixture was quenched with 500 L of 5 mM inorganic phosphate buffer, pH 6.8 and then applied to 2.5 cm Dowex 1 x 8, 200 mesh (PO4 2-) mini-columns equilibrated with 5 mM Pi, pH 6.8.56,102 Reactions were eluted with 3.5 mL of 5 mM Pi buffer, pH 6.8 into liquid scintillation vials. Liquid sc intillation vials were counted for 5 min., and all tubes were cycled through the counter 4-5 times to obtai n an accurate measurement of the amount of radioactive product. The de finition of a unit of activity is the amount CMP-NeuAc converted to sialyl-lactose per minute. Th e activity reported was obtained by correcting the observed velocities with the obtained kine tic parameters and the concentrations of substrate employed, as fit to the following bi-subst rate equation: = (vmax x [A] x [B]) / (KmA x KmB + KmA x [B] + KmB[A] + [A] x [B]) eq. 3-1 Steady State Kinetics for Reco mbinant h23STGal IV Isoforms The kinetic parameters for the recombinant human (2 3) sialyltransferase isoforms with CMP-NeuAc and -lactose as the donor-accepto r pair were estimated by varying the CMP-NeuAc con centration while holding -lactose at a near-saturating concentration, and by vary the -lactose concentration and holding CMP-NeuAc at a

PAGE 92

77 near-saturating concentration. Reactions were conducted at 37 oC for 12 min in 50 mM MES, 0.2 mg/mL BSA, 0.05 % (v/v) Triton CF-5 4, pH 7.5 buffer with a final volume of 100 L. Each reaction contained 100,000 cpm of [9-3H] CMP-NeuAc diluted to the required specific activity. The apparent Km value for CMP-NeuAc was obtained by using 30-300 M of CMP-NeuAc with 0.5 mM of -lactose, and for -lactose, using 50-500 mM of -lactose with 300 M CMP-NeuAc. The reactions were initiated by the addition of 8-12 g of recombinant enzyme. Aliquots of 20 L were removed at 3, 6, 9, and 12 min and quenched in 500 L of 5 mM phosphate, pH 6.8 buffer. The product was quantified using the Dowex column methodology as described above.56,90 Samples were counted in a liquid scintillati on counter for 10 min., and all tu bes were cycled through the counter 6-10 times to obtain an accurate measurement of the amount of radioactive product. The kinetic parameters obtained fo r the three recombinan t isoforms of human (2 3) sialyltransferase using CMP-NeuAc and -lactose as the donor-acceptor substrate pair were estimated by fitting the kinetic data to the Michaelis-Menten equation using a least squares analys is in Sigma Plot ver. 9.0. Michaelis-Menten Equation: = Vmax [S]/ Km + [S] eq. 3-2

PAGE 93

78 CHAPTER 4 KINETIC ISOTOPE EFFECT STUDIES ON RECOMBINANT SIALYLTRANSFERASES Introduction Kinetic isotope effects (KIEs) serve as a valuable technique to elucidate the transtition-state structure of organic and enzymatic reactions. Knowledge about the transition-state structure of an enzyme catalyz ed reaction is significant in that it offers detailed information about the reaction m echanism in a way that enzyme crystal structures are unable to pr ovide. In this study, a seri es CMP-NeuAc and UMP-NeuAc donor substrate radioisotopomers were synthe sized to probe the mechanism of sialyl transfer using KIE experiments. By measuri ng a variety of KIEs at different positions on the donor substrate, one can gain valuable information about various aspects of the transition-state structure which will assist in acquiring mechanistic information for the sialyltransferase catalyzed reaction. The dua l-label competitive method was used to measure the KIEs for the various donor subs trate radioisotopomers with recombinant human placental (2 3) sialyltransferase, recombinant rat liver (2 3) sialyltransferase, and recombinant rat liver (2 6) sialyltransferase. The data from these experiments will provide an increased unders tanding of the mechanism of glycosyl transfer with regard to interactions at the phosphate leaving group via 18O isotopic substitution at the glycosidic O and non-br idging phosphate oxygen atoms for a family of enzymes. Additionally, this data may prove useful toward the development of new

PAGE 94

79 sialyltransferase inhibitors that are based on the transition-state structure of the donor substrate. Kinetic Isotope Effect (KIE) Background Isotope Effect Theory Isotope effects are simply explained as the perturbation of the reaction rate (kinetic isotope effect, KIE) or of th e reaction equilibrium constant (equilibrium isotope effect, EIE) resulting from an isotopic substitution at one position in a reaction molecule. While the description of isotope effects seems stra ight forward, the interpretation of isotope effects can be quite complicated. In general, isotope effects are expressed as a ratio of rate constants where the rate constant for the light molecule (kL) is divided by the rate constant for of the heavy (kH). The initial theoretical calculations for is otope effects and their use to investigate chemical reaction mechanisms was pub lished by Bigeleisen and Mayer in 1947.1 The work presented by Bigeleisen and Mayer on th e calculation of equilibrium isotope effects established the foundation for th e field of isotope effects. The Bigeleisen equation for equilibrium isotope effects is shown in equation 4-1. K1/K2 = MMI EXC ZPE eq. 4-1 For this equation, K1 and K2 represent the equilibrium constants for the two isotopic species being measured. The MMI term incl udes moments of inertia and the combined molecular mass. The EXC term accounts for th e isotope effect on the molecules if they exist in excited vibrational states. Lastly, the ZPE term denotes the isotope effect resulting from differences in vibrational zero-point energy.2

PAGE 95

80 The Bigeleisen equation for equilibrium isotope effects was later extended to include isotope effects on reaction rates (KIEs). Kinetic isotope effect s are related to equilibrium isotope effect theory via the transition-state th eory. The basis of the theory is centered on the supposition that the reactant and the transiti on-state are in equilibrium. The rate of the reaction can then be derived from the tran sition-state theory as the difference in free energy when going from the ground state of the reaction to the transition-state as expressed in equation 4-2. k = (kT/h)exp(G‡/RT) eq. 4-2 In this equation k represents Boltzmann’s cons tant, h denotes Planck’s constant, T is the temperature in Kelvin, G‡ corresponds to the activation fr ee energy, and R is the ideal gas constant. Under the assumptions provided in the transition-state theory, the Bigeleisen equation can be applied to kine tic isotope effects with a sli ght modification to the normal 3N-6 vibrational modes for the ground-state EX C and ZPE terms. In the transition state, one normal mode turns into a reaction coordi nate mode with an imaginary frequency, vL. The reaction coordinate mode accounts for th e motion along the reaction coordinate since the transition-state can convert either back to reactants or forward to products. Thus, transition states have 3N-7 freque ncies with one imaginary frequency.1 The equation for KIE is expressed in eq. 4-3. The mathematical expansions of the individual terms in the

PAGE 96

81 Bigeleisen equation for KIE are show n in Figure 4-1. In this equation u is equal to h /KT. KIE = ‡ L/ ‡ H MMI EXC ZPE eq. 4-3 2 3 1 2‡ 2 ‡ 1 MMI 2 1 1 1 1 2 2 2C B A C B A C B A C B A ‡ 2 ‡ 2 ‡ 2 ‡ 1 ‡ 1 ‡ 1 7 3N‡ ‡ ) 1 ( ‡ ) 2 (e 1 e 1 EXCi u ui i 6 N 3) 2 ( ) 1 (1 1i u ui ie e 7 N 3 2 1 2 1‡ ‡ ) 1 ( ‡ ) 2 ( ZPEi u ui ie e 6 N 3 2 1 2 1) 2 ( ) 1 (i u ui ie e Figure 4-1. Expanded terms of the Bigeleisen equation. Isotope effects, whether they are EIE or KIE, generally originate from the ZPE term. Since molecules of biological interest are normally large, the contribution to the isotope effect from the translational, rotationa l, and excited vibrationa l energies is usually small and therefore insignificant. Conseque ntly, the zero-point en ergy typically becomes the dominating factor of the isotope effect. Most of the isotope effect stems from the differences in zero-point energy between isotopo mers of the initial and final states when either going from reactant to product or in going from the ground-state to the transitionstate. When considering kinetic isotope eff ects, the contribution to the isotope effect from the zero-point energy factor is determin ed by the change in force constants to the

PAGE 97

82 isotopically substituted atom upon moving from the ground-state to the transition-state. For example, if the bond to th e isotopically subs tituted atom becomes looser in the transition-state, then a normal (>1) KIE with be observed. In this case, the zero-point energy term decreases because the force constant of this particular bond diminishes in the transition-state. This change to the isotope effect is illu strated in Figure 4-2 where the potential energy well becomes wider in going fr om the ground-state to the transition-state on the reaction coordinate. Conversely, if th e bond to the isotopically substituted atom becomes tighter in the transition-state, then an inverse (<1) KIE will be observed. The zero-point energy term in this situation increa ses because the force constant of this bond becomes larger in the transition-state. This is depicted in Figure 4-3 where the potential energy well becomes narrower in going from the ground-state to the transition-state on the reaction coordinate. Las tly, if the bond to the isotopical ly substituted atom does not change upon going from the ground-state to the transition-state, than a KIE of unity will be observed. These general changes to the is otope effect may be summarized by the first rule in isotope chemistry which states that the light isotopic mol ecule prefers a looser bonding state where the restricti ons to vibration are lower.3 Since isotope effects arise from changes in force constants and zero-point energies on the bond attached to the isotopically subst ituted atom, they are th erefore local effects and can only extend a couple of bond distances As a result, isotope effects are categorized into different types depending on the location of the isotopic substitution to the reaction center. The two major types of isotope effects are primary and secondary effects which will be discussed in greater detail below.

PAGE 98

83 Ground-state Transition-state GH GDC-H C-D Figure 4-2. Free energy diagra m depicting the looser poten tial energy wells in the transition-state resulting in a normal (> 1) isotope effect from Lowry et al.4 Figure 4-3. Free energy diagra m depicting the looser poten tial energy wells in the transition-state resulting in an inverse (<1) isotope effect from Lowry et al.4 Ground-state Transition-state GH GDC-H C-D

PAGE 99

84 Primary Isotope Effects Primary kinetic isotope effects occur when the isotopically substituted atom experiences either bond formation or bond cleavag e in the transition-state. In general, primary isotope effects are larger than secondary isotope effects because the bond changes taking place at the site of isotopic s ubstitution translate into higher changes in ZPE. The measurement of carbon heavy atom primary isotope effects for molecules of biological interest is commonly achieved by isotopically substituting a carbon reaction center with a 14C or 13C atom. The size of the primary isotope effect is contingent upon the type of symmetry that occurs around the re action center in the tran sition-state. If the reaction center is symmetrical at the transi tion-state, than the symmetric stretching vibration in the transition-st ate will be the major contri butor to the primary isotope effect.4 This is explained when one considers the different vibrati onal modes between the ground-state and the transition-state. The vibrat ional modes that exist in the reactant are the bending and stretching vibrations. In th e transition-state, the vibrational modes include the bending vibration, the symmetric stretching vibration, and the vibrational mode that becomes the reaction coordinate. If the bending vibrations are similar between the ground-state to the transition-state, th ese vibrations cancel to leave only the symmetric stretching vibration. As a result the symmetric stretc hing vibration becomes the main contributor to the primary isotope e ffect. The symmetric stretching vibration is depicted in Figure 4-4 where the isotopically substituted atom “C” is being moved from “A” to “B”. In a symmetrical transition-state, atom C is motionless and the symmetric stretching vibration involves onl y A and B. The isotopically substituted atom, therefore, does not significantly contribute to the vi brational frequency in the transition-state because atom C will not have a zero-point energy difference. Symmetrical transition-

PAGE 100

85 states occur in associative SN2 reactions. Associative SN2 reactions normally have larger primary 14C isotope effects with valu es in the range of 1.08-1.15. On the other hand, in an asymmetrical tr ansition-state, the isotopically substituted atom contributes more to the primary isotope effect because it keeps some of the symmetrical stretching vibrational frequency. In this case, the symmetrical stretching vibrational frequency attributed to atom C will partially cancel the zero-point energy difference in the ground-state to give a decreased kinetic isotope effect.4 Asymetrical transition-states are characteristic for dissociative SN1 reactions. Dissociate SN1-like reactions typically have smaller primary 14C kinetic isotope effects with values in the range of 1.02-1.05.5 Figure 4-4. The symmetric stretc hing vibration mode in the transition-state of transfer reactions. C represents th e isotopically substituted atom that is transferred between A and B. Secondary Isotope Effects Secondary isotope effects occur when a bond to the isotopically substituted atom is neither broken or formed in the transition-st ate. Secondary isotope effects are typically smaller than primary isotope effects and th ey do not directly re port on bond breaking and Symmetric Transition-state Asymmetric Transition-state AB C

PAGE 101

86 bond formation in going from the ground-state to the transition-state. The two common types of secondary isotope effects are and -secondary isotope effects. In -secondary isotope effects, the site of is otopic substitution is at the reaction center. These effects generally occur when the reaction center undergoes a change in hybridization or a change in nonbonding inte ractions. A normal isotope effect is observed if the hybridization st ate changes from sp3 to sp2 in the transition-state as shown in Figure 4-2. Conversely, if the hybrid ization state changes from sp2 to sp3 in the transition-state, an inverse isotope effect will be observed as de picted in Figure 4-3. Unlike primary carbon isotope effects, -secondary isotope effects are unable to differentiate SN1 and SN2 type reaction mechanisms. In -secondary isotope effects, the site of is otopic substitution is adjacent to the reaction center. -secondary isotope effects result when hyperconjugation occurs between the isotopically s ubstituted atom, such as -deuterium (C-H/D), and the electron deficient p-orbital at the adjacent carbon in a dissociative transition-state.6 A normal isotope effect is generally observed for -secondary isotope effects because the C-H/D bond loosens in the transition-state as shown in Figure 4-2. The magnitude of the secondary isotope effect relies on the degree of positive charge build-up in the transitionstate and the dihedral angle ( ) between the empty p-orbital on the reaction center and the bond to the isotopically substituted atom.7 Maximal effects are seen for dihedral angles of 0 and 180. When = 90, induction from the ad joining electron deficient carbon tightens the C-D bond, producing a small inverse isotope effect. The inductive effect associated with deuterium isotope substitutions also contributes to the isotope effect, but is seldom included in the analysis of the -secondary isotope effects due to its small size.8

PAGE 102

87 Thus, -secondary isotope effects can not only pr ovide key information about the amount of positive charge formation on the reaction cen ter at the transition-state, but they can also give pertinent information about the c onformation of the reaction center due to the angular dependence of the isotope effect.9 The -secondary deuterium isotope effects for traditional SN2 and SN1 reactions fall in the ra nge of 1.00-1.02 and 1.08-1.15 per deuterium substitution, respectively.5 Kinetic Isotope Effect Measurement Technique Given that the majority of isotope effects are generally small with the exception of hydrogen primary isotope effects, the establis hment of reliable methods that accurately and precisely measure the isotope effect is qui te important in order to acquire the most information about the reaction mechanism. Depending on the type of method used to measure the isotope effect, one can determ ine whether the effect is on V/K or Vmax. The two commonly used methods to measure kine tic isotope effects on enzyme catalyzed reactions are the competitive and non-competitive methods. These methods will be discussed in greater detail below. The Competitive Method The competitive method uses a mixture of tw o isotopically labeled substrates in a reaction with an enzyme. The reaction rate s are then simultaneously measured for the two isotopically labeled subs trates based on changes in th e light/heavy isotope ratio over the time course of the reaction. In the competitive method, only V/K effects can be measured because the two isotopically labeled substrates are considered to be competitive inhibitors of each other and, theref ore, both can never be saturating.10,11 The two most commonly used labels for the substrate isotopomer pair are 3H and 14C radiolabels. In this case, one substrate contai ns a radiolabel at the isotop ically sensitive position, while

PAGE 103

88 the other substrate contains a radiolabel at a remote position. In a ddition to these labels, stable isotopes such as 18O, 15N, and 2H, can also be used to labe l the substrate. In this situation, a dual-labeled substrat e is synthesized that contai ns a stable isotope in the isotopically sensitive position and a radioa ctive trace label in a remote position. The competitive method is advantageous to use becau se the experimental errors are generally small, <1 %, even when unstable substrates are used.11 Gas-ratio mass spectroscopy and dual-channe l liquid scintillati on counting are the two common techniques used to measure competitive method KIEs. The gas-ratio mass spectroscopy method can be advantageous to use because the experimental error is generally around 0.001%.11 The major caveat to using this method is that it requires the isotopically labeled substrate to be converted into a gas. Additionally, this conversion must take place without an isotope effect occurring.11 In this method, substrates containing stable isotop e labels are generally preferred over those containing radiolabels because these substrates are more difficult to synthesize in larg e quantities and the radioactive precursors are often more expensive to purchase. Dual-channel liquid scintillation counti ng is often chosen over gas-ratio mass spectroscopy to measure KIEs because it easier to use for substrates containing 3H and 14C radiolabels. The advantages of this tec hnique are that it can clearly differentiate between the two radioisotopomers, it is highl y sensitive, and it can easily be used for substrate radioisotopomers that have different specific ac tivities without affecting the KIE results. One of the drawbacks to using th is technique is a lowe r degree of precision than for gas-ratio mass spectroscopy methods, but the attainable value of 0.25 % is often sufficient for the problem to which it is applied. Furthermore, the synthesis of the

PAGE 104

89 various substrate isotopomers still remains an obstacle since it is often very difficult to synthesize substrates with a radioisotope or a stable isotope label in the desired positions. The synthesis and purification of the desire d substrates is usua lly “rate limiting step” toward obtaining KIE measurements. Synt hetic routes that employ enzymes to synthesize many of the desired substrate isot opomers has grown in popularity in recent years because enzymes are highly stereospec ific and selective. Additionally, enzymes are able to rapidly produce hi gh yields of the desired produ ct often in a convenient onepot reaction mixture. The pitfalls of using enzymatic synthesis over traditional chemical synthesis are that the synthe sis is limited by the types of reactions that enzymes can perform, and that many of the enzymes re quired are not commercia lly available. When using the dual-channel liquid scintill ation technique to measure KIE values, sample vials should be counted for at least 10 minutes each for a minimum of 6 cycles in order to reduce the relative standard deviati on. The standard deviation for this technique approximately equals the square root of the number of counts, s = cpm(1/2), for a single determination of counts. For example, th e percent relative deviation for a sample containing at least 350,000 cpm is 0.3 %.11 KIE experiments should have a percent relative standard deviation of <0.25 % for this method.11 The calculation of the KIEs using the dual-channel liquid scintillation met hod will be discussed in greater detail in the experimental section. As with any experimental method, the establis hment of several cont rols that test the validity of the method is paramount. When using dual-labeled substrates, one control that should be performed is the one that wi ll demonstrate that the site of the remote radioactive trace label is in an isotopically insensitive positio n. This control is usually

PAGE 105

90 done by measuring the KIE for the substrate is otopomers that only c ontain the radiolabel in the remote position. If a KIE is observed at this remote position, than this value can be used to correct future KIEs m easured on dual-labeled substrates with these remote labels. The measurement of KIEs using the dua l-labeled competitive method typically employs the use of column chromatography (FPLC or HPLC) to separate reaction products from the remaining substrate. Therefor e, another type of c ontrol that should be performed is the one that will test whether there is any isotopic fractionation from the column chromatography method. This contro l is typically accomplished by measuring the 3H/14C ratios of a substrate mixture prior to and after chromatograp hy. Fractions are collected after the chromatography step into liquid scintillation vial s for counting using a dual-channel liquid scintillation counter. Specia l care should be taken to collect the entire radioactive peak in order to ensu re that there is an identical 3H/14C composition in all fraction vials for multiple trials. Identical 3H/14C ratios indicate that there is no isotopic fractionation from the column chromatography methods. The Noncompetitive Method In the noncompetitive method or “direct ra te method”, the reaction rates of two isotopically labeled substrates are measured individually using separate enzyme reaction mixtures. Unlike the competitive method, this method can measure both V/K and Vmax values for the overall reaction or for a single turnover. The use of the noncompetitive method to measure small KIEs values is often undesirable because of the large experimental error (2-10 %) a ssociated with this method.11 Thus, primary deuterium KIEs are typically the only KIEs measured with this method because their large KIE values are generally less affected by the magnit ude of the error. However, errors of less than 1 % have been obtained for small heavy atom 15N and 18O kinetic isotope effects

PAGE 106

91 when special spectrophotometric techniques were used.12-14 Another caveat of this method is that it is difficult to exactly re produce the experimental conditions for every reaction. Changes in substrate and enzyme concentrations, sp ecific activities, and reaction temperature may give variable KIE va lues. Time-point assays are not a suitable method to use for the measurement of noncom petitive KIEs because of they can have errors greater than 10 %.11 Therefore, noncompetitive KIEs are usually measured on a UV-vis spectrophotomer using th e continuous assay method. Kinetic Isotope Effect Methodology The KIE methodology used in the follo wing experiments was the dual-label competitive method. KIE measurements were made by incubating recombinant sialyltransferase with a mixture of dual-labeled CMP-NeuAc or UMP-NeuAc radioisotopomers and acceptor substrate until a substrate to product conversion of 40-60 % was achieved.15 Unreacted substrate from a non-enzyme initiated reaction (t0) and an enzyme initiated reaction (t1/2) were isolated by anion-ex change HPLC and directly collected into liquid scintilla tion vials for dual-channel liqui d scintillation counting. The typical t0 and t1/2 UMP-NeuAc HPLC chromatograms fr om a KIE experiment are shown in Figure 4-5. UMP-NeuAc isotopomers were primarily used as the donor substrate for KIE experiments with the recombinant sialyltransfer ases. This was done in order to avoid the commitment factor associated with the na tural donor substrate, CMP-NeuAc. As discussed previously in Chapter 2, previous work conducted in our laboratory showed that there is a commitment to catalysis (commitment factor, Cf) for the CMP-NeuAc donor substrate when bound to sialyltransferase.16 This means that the sialyltransferase catalyzed reaction with CMPNeuAc donor substrate contains more than one kinetic

PAGE 107

92 barrier that is partially rate limiting.17 When a commitment factor exists, the nonchemistry rate limiting steps can mask the fu ll expression of the KIEs. As a result, the observed KIE values are typically smaller than the actual (intrinsic) KIEs. The observed V/K isotope effects can be corrected for the commitment factors using equation 4-4. KIEintrinsic = KIEobserved + (KIEobserved x Cf) – Cf eq. 4-4 One method to eliminate the commitment factor is to alter the reaction conditions in a way that will reduce the rate of the isotopi cally sensitive step. In our case, UMP-NeuAc removes the commitment factor in two ways. First, by raising the ki netic barrier for the chemistry step as indicated by its lower kcat value (kcat = 1.2 s-1 for UMP-NeuAc and kcat = 3 s-1 for CMP-NeuAc).16 Secondly, the Km for UMP-NeuAc is higher (Km UMPNeuAc = 1.2 mM vs. Km CMP-NeuAc = 16 M).16 UMP-NeuAc is also an ideal substrate for these isotope effect studies beca use the structure is similar to that of the natural donor substrate, CMP-NeuAc. This substrate only differs from CMP-NeuAc by one substitution, an amino group substituted for a keto group at the C4 position of the pyrimidine ring, which is distant from the anomeric carbon reaction center of NeuAc.

PAGE 108

93 Figure 4-5. Typical t0 (top panel) and t1/2 (lower panel) UMP-NeuAc HPLC chromatograms for KIE experiments on r ecombinant sialyltransferase. The vertical lines represent the beginni ng and end of fraction collection. UMP-NeuAc U MPNeu A c U MP

PAGE 109

94 Results and Discussion The kinetic isotope effects measured for recombinant human (2 3) sialyltransferase (Ins-h23STGal IV), rat (2 3) sialyltransferase (r23STGal IV), and rat (2 6) sialyltransferase (ST6Gal I) are listed in Tables 4-1 – 4-3, respectively. Table 4-1. KIEs measured for recombinant human (2 3) sialyltransferase (Insh23STGal IV). Ins-h23STGal IV Isotopomeric Pair Type of KIE KIE Value [1-3H-NAc], [1-14C-NAc] CMP-NeuAc Control 1.000 0.005 [1-3H-NAc; 3,3'-2H2], [1-14C-NAc] UMP-NeuAc -Secondary 1.161 0.010 [9-3H], [1-14C-NAc] CMP-NeuAc Binding 0.963 0.006 [9-3H], [1-14C-NAc] UMP-NeuAc Binding 0.922 0.010 Table 4-2. KIEs measured for recombinant rat (2 3) sialyltransferase (r23STGal IV). The asterisk denotes the KIE prev iously measured by Michael Bruner.18 r23STGal IV Isotopomeric Pair Type of KIE KIE Value [1-3H-NAc], [1-14C-NAc] UMP-NeuAc Control 0.988 0.008 [9-3H], [1-14C-NAc] UMP-NeuAc Binding 0.910 0.006 [1-3H-NAc; 3,3'-2H2], [1-14C-NAc] UMP-NeuAc -Secondary 1.160 0.010* [1-3H-NAc], [1-14C-NAc, P18O2] UMP-NeuAc Secondary 18O 0.981 0.011 [1-3H-NAc], [1-14C-NAc, 2-18O] UMP-NeuAc Primary 18O 1.018 0.005* [9-3H], [1-14C-NAc, 2-18O] UMP-NeuAc Primary 18O 1.020 0.003 Table 4-3. KIEs measured for recombinant rat (2 6) sialyltransferase (ST6Gal I). Asterisks denote KIEs previous ly measured by Michael Bruner.18,19 r26STGal I Isotopomeric Pair Type of KIE KIE Value [1-3H-NAc], [1-14C-NAc, P18O2] UMP-NeuAc Secondary 18O 0.998 0.008 [1-3H-NAc], [1-14C-NAc, 2-18O] UMP-NeuAc Primary 18O 1.034 0.007* A KIE of 1.161 0.010 was measured for recombinant human (2 3) sialyltransferase (Ins-h23STGal IV) using the isotopomeric pair of [1-3H-NAc; 3,3'-2H2]

PAGE 110

95 and [1-14C-NAc] UMP-NeuAc in the presence of 300 mM -lactose. This KIE value is characteristic of an SN1-like reaction mechanism because it is dissociative at the transition state without nuc leophilic participation (SN1-like). This value was also identical within experimental error to the value previously meas ured by Michael Bruner for rat (2 3) sialyltransferase (r23STGal IV).19 These KIE values were obtained after correcting for the small inverse KIE (0.988 0.0 08) that was observed at the remote [1-3H-NAc] label. The -secondary dideuterium KIE result s obtained for recombinant Insh23STGal IV and r23STGal IV ar e interesting because they ar e significantly smaller than the 1.218 0.010 -secondary KIE reported for recombin ant r26STGal I using the same donor substrate isotopomeric pair.16 The smaller -secondary dideuterium KIE values suggests that there is less positive charge at the reaction center on the NeuAc ring for the recombinant Ins-h23STGal IV and r23STGal IV transition-states. Furthermore, because there is less hyperconjugation occurring between the C3 NeuAc ring deuterons and the oxocarbenium ion in the transi tion-state, this suggests th at the conformation of the NeuAc ring may be different for the recombinant (2 3) sialyltransferase enzymes in the transition-state, compared to the ST6Gal I enzyme. In addition to the -secondary dideuterium KIEs, sim ilar KIEs were also observed for the recombinant Ins-h23STGal IV and r 23STGal IV enzymes at the C9 position on the NeuAc glycerol tail. Binding KIEs of 0.922 0.010 and 0.910 0.006 were measured for recombinant Ins-h23STGal IV and r23STGal IV, respectively, using the isotopomeric pair of [9-3H] and [1-14C-NAc] UMP-NeuAc in the presence of 300 mM lactose. The large inverse isotope eff ect observed at the C9 position on the donor substrate indicates that the bond to the tritium label becomes tighter in the transition-state

PAGE 111

96 when enzyme bound. This result would occur if binding interactions existed between the NeuAc glycerol tail hydroxy moieties and ami no acids within the en zyme active-site. The binding KIEs with UMP-NeuAc are larg er than the effect observed at this position for recombinant Ins-h23STGal IV w ith the natural subs trate CMP-NeuAc (0.963 0.006). This result indicates that the majo rity of the commitmen ts associated with CMP-NeuAc in the sialyltransferase cataly zed reaction have been removed and the binding KIEs measured for recombinant Ins-h23STGal IV and r23STGal IV with UMPNeuAc are truly intrinsic. Furthermore, the binding KIE values for recombinant Insh23STGal IV and r23STGal IV were slig htly larger than the 0.944 0.010 binding KIE reported for recombinant ST6Gal I using th e same donor substrate isotopomeric pair.16 This result suggests that the binding interactions between the NeuAc glycerol tail and enzyme active-site amino acid residues may be different for the recombinant ST6Gal I enzyme in the transition-state. One way this would occur would be if recombinant ST6Gal I has different active-site ami no acid residues intera cting with the hydroxy moiety of the glycerol tail. Despite this difference, the binding KIE re sults for the recombinant Ins-h23STGal IV, r23STGal IV and ST6Gal I enzymes may be useful in the area of sialyltransferase inhibitor design. The binding KIE results sugge st that this segment of the donor substrate experiences a binding interaction with the enzy me. Thus, the glycerol tail may be an important component to consider when desi gning new sialyltransf erase inhibitors. Many of the reported sialyltransferase inhibitors maintain this portion of the donor substrate.2022 However, some of the most potent sialyl transferase inhibitors contain a phenyl group instead of the glycerol tail on the NeuAc ring.21,23 These results suggest that

PAGE 112

97 sialyltransferases are capable of tolerating a va riety of modifications ma de to this part of the donor substrate by utilizing compensatory binding interactions. Since one of the primary goals of this proj ect is to probe the mechanism of glycosyl transfer with respect to in teractions at the phosphate l eaving group, several primary and secondary leaving group 18O KIEs were measured for recombinant r23STGal IV and r26STGal I enzymes. As mentioned in Chapter 1, the profile from the pH-rate experiments on recombinant r26STGal I us ing UMP-NeuAc and LacNAc as the donoracceptor substrate pair fits a bell-shaped curv e for two ionizable groups with pKa values of 6.2 and 8.9.16 These data suggest that glycosyl transfer proceeds via a general acid catalyzed mechanism in which one of the phosphate oxygens on the donor substrate may be protonated to facilitate the loss of the nucleotide monophosphate moiety. Thus, the results from the leaving group 18O KIE experiments may be us eful in pinpointing the location of the proton at either the bridgi ng or non-bridging phos phate oxygens of the donor substrate. Before discussing the resu lts of the leaving group 18O KIE experiments, it is important to mention that results from pos itional isotope exchange (PIX) experiments with CMP-NeuAc indicated that the phosphate group of the leaving CMP moiety did not rotate and re-establish the bond with the NeuAc oxocarbenium ion prior to it departing as shown in Figure 4-6.19 This aspect of the mechanism is important to consider since the following leaving group 18O KIEs were measured usi ng donor substrates containing specific phosphate 18O labels. Rotation of the phospha te group would have adversely affected the measurement and in terpretation of the leaving group 18O KIEs because the isotopically labeled positions on the phosphate group of the donor substrates would no

PAGE 113

98 longer be static. Indeed, the results obtained in this work (Tables 4-2, 4-3) for primary vs. secondary 18O KIEs argue against any significant PIX, since one wold have anticipated identical KIEs if pixing had occu rred to any significant extent. Figure 4-6. Positional isotope exchange (PIX ) mechanism. PIX could not be dectected for the sialyltransferase mechanism. If pixing is complete, the bridge 18O label scrambles to give a 33 % 18O distribution at each oxygen. Secondary 18O KIEs of 0.981 0.011 and 0.998 0.008 were measured for recombinant r23STGal IV and r26STGal I enzymes, respectively, using [1-3H-NAc] and [1-14C-NAc, P18O2] UMP-NeuAc as the donor substrate isotopomeric pair. Additionally, primary 18O KIEs of 1.018 0.005 and 1.034 0.007 were observed for recombinant O HO OH HO *AcHN HO CO2 18O P OR O O O HO OH HO *AcHN HO CO2 18O P OR O ORotateO HO OH HO *AcHN HO CO2R = Cytidine H18O P OR O18OHO HO OH HO *AcHN HO CO2O P OR O18OP o s i t i o n a l I s o t o p e E X c h a n g e ( P I X ) (1/3) (1/3) (1/3) (1/3) (1/3) (1/3)

PAGE 114

99 r23STGal IV and r26STGal I en zymes, respectively, using [1-3H-NAc] and [1-14C-NAc, 2-18O] UMP-NeuAc as the donor substrate isotopomeric pair.18 The primary 18O KIE experiment for recombinant r23STGal IV was repeated using [9-3H] and [1-14C-NAc, 2-18O] UMP-NeuAc as the donor substrate isotopome ric pair in order to confirm the results obtained by Bruner. This experiment resulte d in an observed isot ope effect of 1.020 0.003 after correcting the KIE for the isotope effect at the remote 9-3H label. This result was virtually identical to the primary 18O KIE measured by Bruner with this enzyme. In order to determine the m eaning of these leaving group 18O KIEs, a series of theoretical calculations were done to predic t the KIEs using the change in bond order for each protonation case (non-bridging, bridgi ng, and no protonation) in going from the ground-state to the transiti on-state based on a maximum 18O EIE of 1.06 for a loss of a complete bond.24 Protonation of one of the non-br idging phosphate oxygens resulted in an increase in the bond order of 0.25 for the [P18O2] CMP-NeuAc model and a decrease in the bond order of 0.5 for the [2-18O] CMP-NeuAc model as shown in Figure 4-7. This protonation case resulted in pr edicted primary and secondary 18O KIEs of ~1.02 and 0.98 – 1.00 for the [2-18O] CMP-NeuAc and [P18O2] CMP-NeuAc models, respectively (Table 4-4). Furthermore, protonation of the bri dging phosphate oxygen resulted in predicted primary and secondary 18O KIEs of unity since the ch ange in bond order was zero for both of the bridgi ng and nonbridging 18O labeled CMP-NeuAc models (Figure 4-7 and Table 4-4). Finally, the no protonation case resulted in a decrease in bond order of 0.67 for the [2-18O] CMP-NeuAc model and a decrease in bond order of 0.17 for the [P18O2] CMP-NeuAc model (Figure 4-8). These resu lts gave predicted primary and secondary

PAGE 115

100 18O KIEs of >1.02 and ~1.01 for the bridging and non-bridging 18O labeled substrate models, respectively. Table 4-4. Summary of predicte d KIEs based on mechanism. Mechanism Bridge 2-18O KIE Non-bridge P18O2 KIE Non-bridge Protonation ~1.02 0.98 – 1.00 Bridge Protonation ~1.00 ~1.00 No Protonation >1.02 ~1.01 Figure 4-7. Bond order analysis for prot onation at the non-bri dging phosphate oxygen of the donor substrate. O HO OH HO *AcHN HO CO2 18O P O O O O OH OH N N NH2O2 Bonds to 18O LabelO HO OH HO *AcHN HO CO2O P O O18 18O O OH OH N N NH2O1.5 Bonds to 18O LabelO HO OH HO *AcHN HO CO2 18O P O O O O OH OH N N NH2O1.5 Bonds to 18O LabelO P O O18O18O OH OH N N NH2O1.75 Bonds to 18O Label Bonding = 0.5 Bonding = + 0.25Non-brid g e ProtonationH H O HO OH HO *AcHN HO CO2

PAGE 116

101 No Protonation18O P O O O O OH OH N N NH2O O HO OH HO *AcHN HO CO2 18O P O O O O OH OH N N NH2O O HO OH HO *AcHN HO CO2O P O O18 18O O OH OH N N NH2O2 Bonds to 18O Label1.33 Bonds to 18O LabelO P O O18 18O O OH OH N N NH2O1.33 Bonds to 18O Label 1.5 Bonds to 18O Label Bonding = 0.67 Bonding = 0.17O HO OH HO *AcHN HO CO2O HO OH HO *AcHN HO CO2 Figure 4-8. Bond order analysis for protona tion at the bridging phosphate oxygen of the donor substrate (top panel) and no pr otonation of the phosphate oxygens of the donor substrate (lower panel). O HO OH HO *AcHN HO CO2 18O P O O O O OH OH N N NH2O2 Bonds to 18O LabelO HO OH HO *AcHN HO CO2O P O O18 18O O OH OH N N NH2O1.5 Bonds to 18O LabelO18P O O O O OH OH N N NH2O2 Bonds to 18O LabelO P O O18O18O OH OH N N NH2O1.5 Bonds to 18O Label Bonding Bonding = 0Brid g e ProtonationH H O HO OH HO *AcHN HO CO2O HO OH HO *AcHN HO CO2

PAGE 117

102 When compared to the mode ls, the observed leaving group 18O KIEs agreed with the predicted 18O KIE values for a mechanism invol ving proton transfer at the nonbridging phosphate oxygen. The small inverse secondary 18O isotope effects measured on recombinant r23STGal IV and r26STGal I support a dissociative type mechanism in which the P-O bond order increases in the tran sition-state as shown in the model for nonbridge protonation (Figure 4-7).25,26 Although it is not obvious why acid catalysis would be necessary to assist in the departure of a stable leaving group, the anionic non-bridging phosphate oxygen would be a reasonable lo cation for a proton since this oxygen should be more basic than the bridging phosphate oxyge n at the transition-state. On the other hand, the bridging phosphate oxygen would also a ppear to be a reasonable location for a proton since protonation here would weaken the glyc osidic bond by making the phosphoryl group more electrophilic and possibly e xpedite the rate of glycosyl transfer.16 Protonation of the bridging phosphate oxygen in this case is extremely thermodynamically unfavorable given the lo w basicity of this phosphate oxygen. Glycosides, such as lysozyme, that use aci d catalysis must prot onate at the bridging oxygen because this is the onl y suitable location for prot onation to occur on the donor substrate.27,28 However, glycosyltransferases use donor substrates that contain at least two non-bridging phosphate oxygens which are si gnificantly more basi c than the bridging oxygen. Protonation of these non-bridging phosphate oxygens would still render the nucleotide monophosphate a better l eaving group. Thus, the results suggest that glycosyl transfer proceeds via a general acid cata lyzed mechanism in which a non-bridging phosphate oxygen is protonated to f acilitate the loss of CMP.

PAGE 118

103 In addition to the secondary 18O KIE results, the magnitudes of the primary 18O KIEs indicate that there is significant bond cleavage to the leaving group in the transitionstate.25,29,30 The observation that the primary 18O KIEs measured for recombinant r23STGal IV are slightly smaller than the primary 18O KIE measured for recombinant r26STGal I suggests that th ere is less bond cleavage to the leaving group for the r23STGal IV enzyme in the transition-state. This could be due to the reaction proceeding with an earlier transition-state. The -secondary dideuterium KIEs previously measured for the two recombinant enzymes also suggest that there is less glycosidic bond cleavage in transition-state for recombinant r2 3STGal IV, so the results of both 2H and 18O KIEs are in agreement.16,19 Thus, two different transitionstates can be proposed for the recombinant r26STGal I and r23STGal IV en zymes based on the KIE results as shown in Figure 4-9. Furthermore, it is also likely that the transition-state for recombinant Insh23STGal IV will be similar to the transiti on-state proposed for recombinant r23STGal IV given that the secondary -dideuterium and the 9-3H binding KIEs were similar for these two enzymes. Figure 4-9. Transition-state models proposed for recombinant r26STGal I (left) and r23STGal IV (right) enzymes. O P O O O O OH OH N N NH2O O HO OH HO AcHN HO CO2H O P O O O O OH OH N N NH2O O HO OH HO AcHN HO CO2H

PAGE 119

104 Conclusions The library of isotope eff ects measured for the recombinant human (2 3), rat (2 3), and rat (2 6) sialyltransferases s upport a dissociative, SN1-like transition-state with substantial oxocarbenium ion characte r. The results from the leaving group 18O KIE studies on recombinant rat (2 3), and rat (2 6) sialyltransferases suggest that the sialyltransferase catalyzed mechanism pro ceeds via acid catalysis on the non-bridging phosphate oxygen. Furthermore, the secondary -dideuterium and primary 18O leaving group effects for the recombinant rat (2 3)enzyme indicate that there is less charge build-up and C-O bond cleavage for this enzyme than the rat (2 6) sialyltransferase enzyme in the transition-state. These effect s suggest that donor s ubstrate structures for the recombinant rat (2 3) and rat (2 6) sialyltransferase catalyzed mechanisms are slightly different in the transition-state. This knowledge about the behavior of the transition-states for these two enzymes is im portant in the area of sialyltransferase inhibitor design because it may lead to the creation of new sialyltransferase transitionstate inhibitors that are specific to each en zyme. Transition-state analogs that mimic an early transition-state may be more specific to the recombinant (2 3) sialyltransferases, while those that mimic a late transition-state may better suited for recombinant (2 6) sialyltransferase inhibition. Evidence to s upport this inhibitor design strategy can been seen in the work conducted by Vern L. Schr amm where early and late transition-state analogs were used to target the inhibiti on of purine nucleoside phosphorylase (PNP) from M. tuberculosis (Figure 4-10).

PAGE 120

105 Figure 4-10. Early and late transitio n state analogs for PNP from Schramm.31 Experimental Enzyme Reaction General KIE Methodology The competitive method was employed to measure the kinetic isotope effects (KIEs) for recombinant human (2 3), rat (2 3), and rat (2 6) sialyltransferases. For each different KIE experiment, a master mixture of the appropriate 3Hand 14Clabeled donor substrate was prepared from wh ich aliquots were withdrawn to make the desired 3H/14C reference mixture at time zero and th e individual reaction mixtures. The reaction mixtures contained about 100,000 cpm of the appropriate 3H/14C-labeled isotopomeric pair of CMP-NeuAc or UMP-NeuAc donor substrate and 300 mM -lactose acceptor substrate in 50 mM MES, pH 7.2 containing 0.2 mg/mL BSA and 0.05 % (v/v) Triton CF-54 for the recombinant human and rat (2 3) sialyltransferase KIEs. KIEs measured on recombinant rat (2 6) sialyltransferase used 10 mM N-acetyl

PAGE 121

106 lactosamine (LacNAc) as the acceptor subs trate in 50 mM MES, pH 7.0 containing 1 mg/mL BSA, and 0.05 % (v/v) Triton CF-54. Reaction mixtures (20–100 L) were initiated by the addition of the appropriate am ount of enzyme to give 40-60 % conversion in < 30 minutes at 37 C.2 Unreacted substrate was then isolated from the reaction mixture using anion-exchange HPLC (MonoQ column, 75 mM NH4HCO3, 15 % methanol, pH 8.0, 2 mL/min, A271) and 2 mL fractions were directly collected into scintillation vials. Special care was take n to collect the entire CMP-NeuAc or UMPNeuAc peak. The reaction conversion was dete rmined by a peak integration of the CMPNeuAc or UMP-NeuAc and CMP or UMP peaks in the HPLC chromatogram. The reference 3H/14C ratio at time zero was obtained fo r each enzyme reaction mixture by injecting aliquots of the appropriate 3H/14C labeled donor substrate mixture with acceptor substrate in buffer onto the anion-excha nge column and coll ecting the entire donor substrate peak into liquid scintillation vials. Dual channel liquid scintillation counting was used to determine the 3H/14C ratios for the donor substrates (channel A, 0-15 keV; channel B, 15-90 keV). Each tube was counted for 8-10 minutes and all tubes were cycled through the counter 8-10 times. Triplicate samples of [14C] CMP-NeuAc were used to determine the ratio of 14C counts in channels A and B (A:B14). Since 3H is only detected in channel A, the 3H/14C ratio in a given sample tube was calculated with the following equation 32: 3H/14C = (cpm A-cpm B x A:B14) / (cpm B + cpm B x A:B14) eq. 4-5

PAGE 122

107 The reported value and error of a KIE represents the mean and standard deviation of 5-6 individual KIE experiments taken over 8-10 cycles through the liquid scintillation counter. The following equation wa s used to calculate the KIEs: 3H KIEobserved = (3H/14C)t1/2 / (3H/14C) t0 eq. 4-6 14C KIEobserved = (14C/3H)t1/2 / (14C/3H)t0 eq. 4-7 The KIEs were then corrected for per cent conversion using the following equation 33: KIEcorrected = ln(1-f) / ln((1-f) x KIEobserved) eq. 4-8 f = fraction of conversion For KIEs involving stable isotopes where co mplete incorporation was not achieved (18O KIEs), the following equation was used to correct the observed KIEs 34: KIEcorrected = (KIEobserved – 1 + f) / f eq. 4-9 f = fraction of stable isotope incorporation Additionally, the following equations were used to correct the KIE if a KIE was observed at the remote position on the dual-labeled substrates11: Corrected KIE for 14C KIE = KIEobserved x control KIE eq. 4-10

PAGE 123

108 Corrected KIE for 3H KIE = KIEobserved/control KIE eq. 4-11

PAGE 124

109 CHAPTER 5 CONCLUSIONS AND FUTURE WORK This study has provided an increased unders tanding of the transition-state and the mechanism of sialyl transfer for the r eactions catalyzed by the recombinant human (2 3), rat (2 3), and rat (2 6) sialyltransferases. In addition to the mechanistic information, valuable knowledge has been gained pertaining to the purification and kinetic properties of three recombinant human (2 3) sialyltransferase isoforms expressed from insect cells. Kinetic isotope effect met hods were used to obtain information pertaining to the transition-state structure of th e sialyl donor. The series of isotope effects measured for the recombinant Ins-h23STGal IV, r23STGal IV and ST6Gal I sialyltransferases support a dissociative, SN1-like transition-state with substan tial oxocarbenium ion character. Kinetic isotope effect methods were also used to probe the mechanism of glycosyl transfer with regard to enzy me interactions at the phosphate leaving group. Primary and secondary 18O leaving group isotope effects were measured on recombinant r23STGal I and ST6Gal I using a set of UM P-NeuAc isotopomers containing 18O labels either at the bridging or non-bridging phosphate oxygens. The results from the leaving group 18O KIE experiments on recombinant r23STGal IV and ST6Gal I, as well as the results from theoretical calculations, sugge st that glycosyl transfer proceeds via a general acid catalyzed mechanism in which a non-bridging phosphate oxygen is protonated to facilitate the loss of CMP.

PAGE 125

110 Additional KIE studies are now needed to as certain whether acid catalysis occurs at either the proR or proS non-bridging phos phate oxygen of the donor substrate. These studies would require the synthe sis of proR and proS phosphate 18O donor substrate isotopomers. Future work may involve the design and use of a new enzymatic synthesis route to selectively incorporate an 18O label into the proR or proS phosphate oxygen of the donor substrate. The -secondary dideuterium isotope effects observed for recombinant r23STGal I and h23STGal IV were smaller than those observed for recombinant ST6Gal I, suggesting that the degree of glycosidic bond cl eavage is slightly different between the (2 3) and (2 6) sialyltransferases at the tran sition-state. These results in conjunction with the primary 18O leaving group isotope effect results suggest that the recombinant r23STGal I and h23STGal IV cata lyzed reactions proceed with a slightly earlier transition-state than the reco mbinant ST6Gal I catalyzed reaction. Thus, the information gleaned from this study may be useful in the area of sialyltransferase inhibitor design. Our re sults point towards a transition-state for (2 3) sialyltransferases that contains less positiv e charge and glycosidic bond cleavage than the transition-state for (2 6) sialyltransferase. Therefore, transition-state inhibitors that mimic the early and late transition-states of the (2 3) and (2 6) sialyltransferases would presumably make the inhibito rs more specific to each enzyme.128 Furthermore, the most potent sialyltransferase inhibitors are th e transition-state inhibitors that include a planar anomeric carbon, charge mimicry, the cy tidine moiety, and an increased distance between the anomeric carbon and the leaving group CMP.43 These criteria are included in the proposed transition-state inhibitor s hown in Figure 5-1 which is currently being

PAGE 126

111 synthesized in our laboratory for fu ture inhibition studies on both (2 3) and (2 6) sialyltransferases. Figure 5-1. Proposed sialyltransfer ase transition-state inhibitor. In addition to the sialyltran sferase inhibition studies, future investigations of the sialyltransferase catalyzed mechanism may involve the measurement of KIEs on the bacterial sialyltransferase, CstII 32. Although bacterial sialyl transferases do not share sequence similarity with any mammalian sial yltransferases, KIE studies may allow for mechanistic comparisons to be made between these two seemingly unrelated groups of sialyltransferases. The cr ystal structure of CstII 32 could then be used to model the transition-state structure of th e sialyl donor in the enzyme active site. This information may provide new insight into the nature of the mammalian sialyltransferase catalyzed reactions. N H N O COO P O O O O OH OH N N NH2O

PAGE 127

112 LIST OF REFERENCES (1) Bruner, M. and Horenstein, B. A. Biochemistry 2000, 39 2261-2268. (2) Bruner, M. and Horenstein, B. A. Biochemistry 1998, 37 289-297. (3) Schauer, R. Archives of Bioche mistry and Biophysics 2004, 426 132-141. (4) Jovtchev, S., Djenev, I., Stoeff, S., and Stoylov, S. Colloids and Surfaces 2000, 164 95-104. (5) Robbe, C., Capon, C., Maes, E., Rousse t, M., Zweibaum, A., Zanetta, J. P., and Michalski, J. C. J.Biol.Chem. 2003, 278 46337-46348. (6) Hirst, G. K. J.Exp.Med. 1942, 76 195-209. (7) Skehel, J. J. and Wiley, D. C. Annu.Rev.Biochem. 2000, 69 531-569. (8) Chu, V. C. and Whittaker, G. R. Proc.Natl.Acad.Sci. 2004, 101 18153-18158. (9) Schauer, R. Adv.Carbohydr.Chem.Biochem. 1982, 40 131-234. (10) Finne, J. J.Biol.Chem. 1982, 257 11966-11970. (11) Angata, K. and Fukuda, M. Biochimie 2003, 85 195-206. (12) Rutishauser, U. and Landmesser L. Trends in Neuroscience 1996, 19 422-427. (13) Barbeau, D., Liang, J. J., Robitalill e, Y., Quirion, R., and Srivastava, L. K. Proc.Natl.Acad.Sci. 1995, 92 2785-2789. (14) Mikkonen, M., Soininen, H., Alafuzof, I., and Miettinen, R. Rev.Neurosci. 2001, 12 311-325. (15) Ashwell, G. and Morell, A. G. Adv.Enzymol. 1974, 41 99-128. (16) Crocker, P. R. Current Opinion in Structural Biology 2002, 12 609-615. (17) Crocker, P. R. and Varki, A. Immunology 2001, 103 137-145. (18) Codinton, J. F. Masking of cell-surface antigens on cancer cells ; Williams and Wilkins Co.: Baltimore, MD, 1975; p 399.

PAGE 128

113 (19) Dall'Olio, F. and Chiricolo, M. Glycoconjugate Journal 2001, 18 841-850. (20) Powell, L. D., Whiteh eart, S. W., and Hart, G. W. The Journal of Immunology 1987, 139 262-270. (21) Kijima-Suda, I., Miyazawa, T., Itoh, M., Toyoshima, S., and Osawa, T. Cancer Res. 1986, 48 858-862. (22) Kijima-Suda, I., Miyamoto, Y., Toyoshima, S., Itoh, M., and Osawa, T. Cancer Res. 1988, 48 3728-3732. (23) Wagner, H. E., Thomas, P., Wolf B. C., Rapoza, A., and Steele, G. Jr. Arch.Surg. 1990, 125 351-354. (24) Kleineidam, R. G., Schmelter, T., Schwarz, R. T., and Schauer, R. Glycoconjugate Journal 1997, 14 57-66. (25) Strehle, E. M. Genetic Testing 2003, 7 113-121. (26) Gahl, W. A.; Schne ider, J. A.; Aula, P. P. The Metabolic and Molecular Bases of Inherited Diseas ; McGraw-Hill: New York, 1995; pp 3763-3797. (27) Weiss, P., Tiet ze, F., Gahl, W. A., Seppala, R., and Ashwell, G. J.Biol.Chem. 1989, 264 17635-17636. (28) Bourne, Y. and Henrissat, B. Current Opinion in Structural Biology 2001, 11 593-600. (29) Coutinho, P. M., Deleury, E ., Davies, G. J., and Henrissat, B. J.Mol.Biol. 2003, 328 307-317. (30) Lairson, L. L. and Withers, S. G. Chem.Commun. 2004, 2243-2248. (31) Pederson, L. C. Tsuchi da K. Kitagawa H. Sugahara K. Darden T. A. and Negishi M. J.Biol.Chem. 2000. (32) Davies, C. and Henrissat, B. Structure 1995, 3 853-859. (33) Unligil, U. M. and Rini, J. M. Current Opinion in Structural Biology 2000, 10 510-517. (34) Wiggins, C. A. R. and Munro, S. Proc.Natl.Acad.Sci. 1998, 95 7945-7950. (35) Hagen, F. K., Hazes, B., Ra ffo, R., deSa, D., and Tabak, L. A. J.Biol.Chem. 1999, 274 6797-6803. (36) Breton, C., Mucha, J., and Jeanneau, C. Biochimie 2001, 83 713-718. (37) Hu, Y. and Walker, S. Chemistry and Biology 2002, 9 1287-1296.

PAGE 129

114 (38) Vocadlo, D. J., Davies, G. J., Laine, R., and Withers, S. G. Nature 2001, 412 835-838. (39) Murray, B. W., Wittmann, V., Burkar t, M. D., Hung, S. C., and Wong, C. H. Biochemistry 1996, 35 11183-11195. (40) Murray, B. W., Wittmann, V., Burk art, M. D., Hung S.C., and Wong, C. H. Biochemistry 1997, 36 823-831. (41) Javaud, C., Dupuy, F., Maftah, A., Raymond, J., and Petit, J. M. Genetica 2003, 118 157-170. (42) Harduin-Lepers, A., Vallejo-Ruiz, V., Krzeqinski-Recchi, M., Samyn-Petit, B., Julien, S., and Delannoy, P. Biochimie 2001, 83 727-737. (43) Wang, X., Zhang, L., and Ye, X. Medicinal Research Reviews 2003, 23 32-47. (44) Datta, A. K. and Paulson, J. C. J.Biol.Chem. 1995, 270 1497-1500. (45) Datta, A. K., Sinha A., and Paulson, J. C. J.Biol.Chem. 1998, 273 9608-9614. (46) Datta, A. K., Chammas, R., and Paulson, J. C. J.Biol.Chem. 2001, 276 1520015207. (47) Jeanneau, C., Chazalet, V., Aug, C. Soumpasis, D. M., Harduin-Lepers, A., Delannoy, P., Imberty, A., and Breton, C. J.Biol.Chem. 2004, 13461-13468. (48) Horenstein, B. A. and Bruner, M. J.Am.Chem.Soc. 1998, 120 1357-1362. (49) Chiu, C. P. C., Watts, A. G., Lairso n, L. L., Gilbert, M., Lim, D., Wakarchuk, W. W., Withers, S. G., and Strynadka, N. C. J. Nature Structural and Molecular Biology 2004, 11 163-170. (50) Weinstein, J., Souza-eSilva, U., and Paulson, J. C. J.Biol.Chem. 1982, 257 13845-13853. (51) Scudder, P. R. and Chantler, E. N. Biochemical and Biophysical Acta 1981, 660 136-141. (52) Bartholomew, B. A., Jo urdian, G. W., and Roseman, S. J.Biol.Chem. 1973, 248 5751-5762. (53) Pederson, L. C., Tsuchida, K., Kita gawa, H., Sugahara, K., Darden, T. A., and Negishi, M. J.Biol.Chem. 2000, 275 34580-34585. (54) Kitazume-Kawaguchi, S. and Arita, M. J.Biol.Chem. 2001, 276 15696-15703. (55) Harduin-Lepers, A., Mollic one, R., Delannoy, P., and Oriol, R. Glycobiology 2005, 15 805-817.

PAGE 130

115 (56) Williams, M. A., Hiroshi, K., Datta, A. K., Paulson, J. C., and Jamieson, J. C. Glycoconjugate Journal 1995, 12 755-761. (57) Schaub, C., Mller, B., and Schmidt, R. R. Eur.J.Org.Chem. 2000, 1745-1758. (58) Schaub, C., Mller, B., and Shmidt, R. R. Glycoconjugate Journal 1998, 15 345354. (59) Sun H., Yang, J., Amaral K. E., and Horenstein, B. A. Tet.Lett. 2001, 42 24512453. (60) Skropeta, D., Schwrer, R ., Haag, T., and Schmidt, R. R. Glycoconjugate Journal 2004, 21 205-219. (61) Ohta, Y., Shimosaka, M., Murata, K., Tsukada, Y., and Kimura, A. Appl.Microbiol.Biotechnol. 1986, 24 386-391. (62) Ohta, Y., Watana be, K., and Kimura, A. Nucleic Acids Research 1985, 13 88438852. (63) Aisaka, K. and Uwajima, T. Applied and Environmental Microbiology 1986, 51 562-565. (64) Uchida, Y., Tsuka da, Y., and Sugimori, T. J.Biochem. 1984, 96 507-522. (65) Kim, M. J., Hennen, W. J., Sweers, M., and Wong, C. H. J.Am.Chem.Soc. 1988, 110 6481-6486. (66) Roseman, S. and Ludowieg, J. J.Am.Chem.Soc. 1954, 76 301-302. (67) Ambrose, M. G., Freese, S. J., Re inhold, M. S., Warner, T. G., and Vann, W. F. Biochemistry 1992, 31 775-780. (68) Burke, E. E. and Horenstein, N. A. Journal of Labelled Compounds and Radiopharmaceuticals 2004, 47 1007-1017. (69) Risley, J. M. and Van Etten R. L. Journal of Labelled Compounds and Radiopharmaceuticals 1978, 533-538. (70) Cohn, M. and Hu A. Proc.Natl.Acad.Sci. 1978, 75 200-203. (71) Neuhard, J. and Tarp, L. Journal of Bacteriology 1993, 175 5742-5743. (72) Aisaka, K., Igarashi, A ., Yamaguchi, K., and Uwajima, T. Biochem.J. 1991, 276 541-546. (73) Bradford, M. Anal.Biochem. 1976, 72 248-254.

PAGE 131

116 (74) Baykov, A. A., Evtushenko, O. A., and Avaeva, S. M. Anal.Biochem. 1988, 171 266-270. (75) Valentin-Hansen, P. Methods in Enzymol. 1978, 51 308-314. (76) Hackney, D. D. Stempel K. E. and Boyer P. D. Methods in Enzymol. 1980, 60-83. (77) Ichikawa, Y. Shen G. J. Wong C. H. J.Am.Chem.Soc. 1991, 4698-4700. (78) Colley, K. J., Lee, E. U., Adler, B., Browne, J. K., and Paulson, J. C. J.Biol.Chem. 1989, 264 17619-17622. (79) Gillespie, W., Kelm, S., and Paulson, J. C. J.Biol.Chem. 1992, 267 21004-21010. (80) Anatao, B. and Burke E.E., Horenstein N. A. Unpublished work 2005. (81) BEVS Instruction Manual. www.invitrogen.com 4-6-2004. (82) Kim, G. H., Yang, S. M., Lee, Y. C., Do, S. I., Chung, I. S., and Yang, J. M. Biochemical and Biophysical Research Communications 2003, 305 488-493. (83) Borsig, L., Berger, E. G., and Malissard, M. Biochemical and Biophysical Research Communications 1997, 240 586-589. (84) Seo, N. S., Hollister, J. R., and Jarvis, D. L. Protein Expression and Purification 2001, 22 234-241. (85) Malissard, M., Zeng, S., and Berger, E. G. Biochemical and Biophysical Research Communications 2000, 267 169-173. (86) Borsig, L., Ivanov, S. X., Herrmann, G. F., Kragl, U., Wandrev, C., and Berger, E. G. Biochemical and Biophysical Research Communications 1995, 210 14-20. (87) Hamamoto, T., Y oung-Choon, L., Kurosawa, N., Nakaoka, T., Kojima, N., and Tsuji, S. Bioorganic & Medicinal Chemistry 1994, 2 79-84. (88) Horenstein, N. A. Unpublished work 2005. (89) Barker, R., Olsen, K. W., Sharper, J. H., and Hill, R. L. J.Biol.Chem. 1972, 25 7135-7147. (90) Paulson, J. C., Bera nek, W. E., and Hill, R. L. J.Biol.Chem. 1977, 252 23562362. (91) Tipton, J. D. and Horenstein, N. A. Unpublished work 2005. (92) Lu, W., Chapple, S. D ., Lissini, O., and Jones, I. M. Protein Expression and Purification 2002, 24 196-201.

PAGE 132

117 (93) Krokhin, O., Ens, W., Standing, K. G., Wilkins, J., and Perrault, H. Rapid Communications in Mass Spectrometry 2004, 18 2020-2030. (94) Mills, K., Johnson, A. W., Diettrich, O., Clayton, P. T., and Winchester, B. G. Tetrahedron: Asymmetry 2000, 11 75-93. (95) Kitagawa, H. and Paulson, J. C. J.Biol.Chem. 1994, 269 1394-1401. (96) Kitagawa, H. and Paulson, J. C. Biochemical and Biophysical Research Communications 1993, 194 375-382. (97) Joziasse, D. H., Berg, M. L. E., Te r Hart, H. G. J., Koppen, P. L., Hooghwinkel, G. J. M., and Van den Eijnden, D. H. J.Biol.Chem. 1985, 260 4941-4951. (98) Sadler, J. E., Rearick, J. I., Paulson, J. C., and Hill, R. L. J.Biol.Chem. 1979, 254 4434-4443. (99) Bencrov, M., Rendic, D., Fabini, G., Kopecky, E. M., Altmann, F., and Wilson, I. B. H. Biochimie 2003, 85 413-422. (100) Sadler, J. E., R earick, J. I., Paulson, J. C., and Hill, R. L. Journal of Biological Chemistry 1979, 254 4434-4443. (101) Smith, P. K. Anal.Biochem. 1985, 150 76-85. (102) Weinstein J, Souza-eSilvia U, and Paulson JC Journal of Biological Chemistry 1982, 257 13845-13853. (103) Bigeleisen J. and Mayer, M. G. The Journal of Chemical Physics 1947, 15 261267. (104) Cook, P. F. Enzyme Mechanisms from Isotope Effects ; CRC Press, Inc.: Boca Raton, 1991. (105) Cleland, W. W.; O'Leary, M.; Northrop, D. B. Isotope Effects on Enzyme Catalyzed Reactions ; University Park Pre ss: New York, 1976; pp 233-240. (106) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic Chemistry ; Harper Collins Publishers: New York, 1987; pp 233-240. (107) Melander, L.; Saunders, W. H. Carbon Isotope Effects; In Reaction Rates of Isotopic Molecules ; Robert E. Krieger Publishing Company: Malabar, 1987; pp 225-257. (108) Goitein, R. K., Chel sky, D., and Parsons, S. M. J.Biol.Chem. 1978, 253 29632971. (109) Zhang, Y., Bommuswam y, J., and Sinnott, M. L. J.Am.Chem.Soc. 1994, 116 7557-7563.

PAGE 133

118 (110) Bennett, A. J. and Sinnott, M. L. J.Am.Chem.Soc. 1986, 108 7287-7294. (111) Zhang, Y., Bommuswam y, J., and Sinnott, M. L. J.Am.Chem.Soc. 1994, 116 7557-7563. (112) Cleland, W. W. Isotopes in Organic Chemistry ; Elsevier: New York, 1987. (113) Parkin, D. W. Methods for the De termination of Competitive and Noncompetitive Kinetic Isotope Effcts; In Enzyme Mechanisms from Isotope Effects ; CRC Press, Inc.: Boca Raton, 1991; pp 269-290. (114) Rosenberg, S. and Kirsh, J. F. Biochemistry 1981, 20 3189. (115) Rosenberg, S. and Kirsch, J. F. Anal.Chem. 1979, 51 1379-1383. (116) Burgner, J. W., II, Oppenhe imer, N. J., and Ray, W. J. Jr Biochemistry 1987, 26 91-96. (117) Duggleby, R. G. and Northrop, D. B. Bioorg.Chem. 1989, 17 177-193. (118) Cleland, W. W. Methods in Enzymol. 1982, 87 625-641. (119) Bruner, M. and Horenstein, N. A. Unpublished work 2005. (120) Bruner, M. J. 1999. University of Florida. (121) Mller, B. Martin T. J. Schaub C. and Schmidt R. R. Tet.Lett. 1998, 39 509-512. (122) Sawyer, C. B. and Kirsch, J. F. J.Am.Chem.Soc. 1973, 95 7375-7382. (123) Weiss, P. M. and Cleland, W. W. J.Am.Chem.Soc. 1989, 111 1928-1929. (124) Weiss, P. M., Knight W. B., and Cleland, W. W. J.Am.Chem.Soc. 1986, 108 2761-2762. (125) Rosenberg, S. and Kirsch, J. F. Biochemistry 1981, 20 3196-3204. (126) Bennett, A. J. and Sinnott, M. L. J.Am.Chem.Soc. 1986, 108 7287-7294. (127) Gerratana, B., Frey, P. A., and Cleland, W. W. Biochemistry 2001, 40 29722977. (128) Schramm, V. L. Acc.Chem.Res. 2003, 36 588-596. (129) Parkin, D. L. Enzyme Mechanisms from Isotope Effects ; CRC Press: Boca Raton, 1991.

PAGE 134

119 (130) Northrop, D. B. Methods in Enzymol. 1982, 87 607-625. (131) Rosenberg, S. and Kirsch, J. F. Biochemistry 1981, 20 3204.

PAGE 135

120 BIOGRAPHICAL SKETCH Erin Elizabeth Burke, daughter of Mr. and Mrs. Paul M. Ringus, was born in Greenfield, Massachusetts, on October 14, 1978. Erin lived in Athol, Massachusetts, until the age of six and then her family re located to Moosup, Connecticut, in 1985 where she lived until the age of twelve. In the Fall of 1990, Erin Â’s family relocated to Barnwell, South Carolina, where she completed her sec ondary education culmin ating in graduation from Barnwell High School in 1996. She attend ed Columbia College in Columbia, South Carolina, where she graduated summa cum laude and cum honorae in 2000 with a Bachelor of Science degree in chemistry. Erin relocated to Gainesville, Florida, to pursue her Doctor of Philosophy degree in chemistry fr om the University of Florida. While in graduate school, Erin met and married her husband, Andrew Paul Burke, on December 7, 2002. Erin received her Ph.D. in chemistry in the summer of 2005 under the guidance of Dr. Nicole A. Horenstein. After graduation, Erin and her husband will be relocating to Charleston, South Carolina, to begin their new jobs.