UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Neisseria meningitidis lipopolysaccharide galactosyl transferase: mechanistic investigations and applications… Lougheed, Brenda 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1998-0689.pdf [ 8.47MB ]
JSON: 831-1.0061622.json
JSON-LD: 831-1.0061622-ld.json
RDF/XML (Pretty): 831-1.0061622-rdf.xml
RDF/JSON: 831-1.0061622-rdf.json
Turtle: 831-1.0061622-turtle.txt
N-Triples: 831-1.0061622-rdf-ntriples.txt
Original Record: 831-1.0061622-source.json
Full Text

Full Text

NEISSERIA MENINGITIDIS LIPOPOLYSACCHAMDE GALACTOSYL TRANSFERASE: MECHANISTIC INVESTIGATIONS AND APPLICATIONS FOR OLIGOSACCHARIDE SYNTHESIS BY BRENDALOUGHEED B.Sc, University of Winnipeg, 1994  A THESIS SUBMITTED ESf PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE F A C U L T Y OF GRADUATE STUDIES DEPARTMENT OF CHEMISTRY  We accept this thesis as conforming to the required standard:  THE UNIVERSITY OF BRITISH COLUMBIA February, 1998 ©BrendaLougheed, 1998  In presenting  this  degree at the  thesis  in  partial fulfilment  of  University of  British Columbia,  I agree  freely available for reference copying  of  department  this or  publication of  and study.  thesis for scholarly by  this  his  or  her  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  h e r e i n  purposes  a q / q f t  that the  may be It  thesis for financial gain shall not  C Vvg,vw V S t y " v |  requirements  I further agree  representatives.  permission.  Department of  the  that  advanced  Library shall make it  by the  understood be  an  permission for extensive  granted  is  for  that  allowed without  head  of  my  copying  or  my written  ABSTRACT  The objective of this study was to characterise the activity of the recombinant Neisseria meningitidis UDP-galactose: 4-a-galactosyl transferase, lgtC-19, which has been expressed in Escherichia coli. Advances in cloning and expression techniques over the past decade have enabled access to a variety of glycosyl transferases not previously available in sufficient quantity for detailed exploration. This recent availability has led to a resurgence of interest in the behaviour of these enzymes and their potential applications in the enzymic synthesis of oligosaccharides. Although there has been a significant expansion in our understanding of these biological catalysts, much of this work has focused on the study of inverting enzymes and there is still relatively little information available regarding the kinetic and mechanistic behaviour of retaining nucleotide diphosphate hexosyl transferases. Additionally, there has not yet been reported a crystal structure for this class of enzyme, so that the structure of the active site has yet to be elucidated. LgtC-19 is a retaining a-galactosyl transferase, catalysing the transfer of galactose from UDP-galactose to lipopolysaccharide acceptors. In Nature, this enzyme recognises the reducing end of a lactose moiety as a suitable acceptor for its transfer reaction. The ability of this enzyme to recognise and accept a variety of alternative, synthetic glycosides as acceptor substrates as well as donor substrates has been investigated. These studies have shown lgtC19 to exhibit significant flexibility with regard to acceptor substrate structure.  The  fluorescent-labelled sugars FITC-lactose, FCHASE-lactose, and even FCHASE-galactose  Ill  were capable of performing this function as were the simple disaccharide, lactose, and monosaccharide derivative, cc-galactosyl fluoride. Although there have been other glycosyl transferases reported to utilise synthetic glycosides as acceptors, lgtC-19 appears to be the first such enzyme capable of using a simple glycoside as a donor substrate. The transferase activity of lgtC-19 was seen when agalactosyl fluoride was provided as the glycosyl donor in the presence of catalytic amounts of UDP. This is the first report of any nucleotide diphosphate-dependent glycosyl transferase utilising a donor substrate which does not contain a nucleotide diphosphate functionality. With this flexibility, lgtC-19 shows excellent potential for use in the industrial-scale synthesis of oligosaccharides. The reaction conditions required by lgtC-19 were optimised by systematically altering the concentrations of substrates and co factors, investigating the effect of cofactors on enzyme activity, and varying the pH of enzyme storage solutions and of the assay itself. Once the conditions were optimised for maximum kinetic activity, the mechanism of enzyme action was investigated using Cleland's method. This analysis spanned a range of pH values and enabled the construction of pH curves for UDP-galactose, lactose, and oc-galactosyl fluoride. From the information obtained through a variety of kinetic and mechanistic studies, the mechanism of galactosyl transfer employed by lgtC-19 was determined to be bi bi sequential.  iv  TABLE OF CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS  iv  LIST OF TABLES  xiii  LIST OF FIGURES  xv  ABBREVIATIONS AND SYMBOLS  xx  ACKNOWLEDGEMENTS  xxiv  CHAPTER 1  1  INTRODUCTION  1  1.1 General Introduction  2  1.2 The Mechanism of Glycosyl Transfer  4  1.2.1  The Catalytic Mechanism of Glycosidases  5  1.2.2  The Proposed Mechanism for Glycosyl Transferases  9  V  1.2.3  Oxocarbenium Ion-like Transition States  11  Transition states of hydrolases  11  Transition states of transferases  15  Glycosyl-Enzyme Intermediate  17  Tests for a transferase intermediate  22  1.2.5  Acid/Base and Electrophilic Catalysis  24  1.2.6  Electrophilic Catalysis  26  1.2.7  Non-covalent Interactions  27  1.3 Enzymic Synthesis of Glycosides  29  1.2.4  1.3.1  Synthesis Using Glycosidases  29  1.3.2  Synthesis Using Glycosyl Transferases  31  1.4 Lipopolysaccharide Galactosyl Transferase 1.4.1  Applications of a(l->4)Galactose Oligosaccharides  32 34  1.5 Use of Cleland Kinetics to Probe Enzyme Mechanism  36  1.6 Aims of this Study  38  CHAPTER 2  CHARACTERISATION OF THE LGTC-19 MUTANT FROM N. MENINGITIDIS USING UDP-GALACTOSE AS THE DONOR SUGAR  40  vi  2.1 Introduction  41  2.2 Development of an Assay Method  43  2.2.1  Coupled Continuous Assay: UDP-galactose as glycosyl donor  2.3 lgtC-19 Kinetics using UDP-galactose as a Donor Substrate  45  49  2.3.1  Substrate Optimisation  49  2.3.2  Cofactor Optimisation  50  2.3.3  Metal Cofactor Specificity  55  2.4 Kinetic Mechanism  59  2.4.1  Cleland Kinetics at pH 7.5  59  2.4.2  Inhibition Using a Donor Substrate Analogue  61  2.5 pH Study  64  2.6 Probing the Catalytic Domain of lgtC-19  69  2.6.1  Inactivation and Labelling Studies  70  CHAPTER 3  74  a-GALACTOSYL FLUORIDE KINETICS AND CHARACTERISATION  74  vii 3.1 Introduction  75  3.2 Preliminary Studies: Assessment of a-Galactosyl Fluoride Suitability  as a Galactosyl Donor  3.3 Prep-scale Synthesis  76  79  3.4 Kinetic Characterisation of LgtC-19 using a-Galactosyl Fluoride as a Glycosyl Donor  82  3.4.1  Substrate Optimisation  82  3.4.2  Cofactor Optimisation  88  3.5 "Mini" pH Study of the IgtC-l 9-catalysed Galactosyl Transfer Reaction  88  3.6 Attempts to Identify a Galactosyl-enzyme Intermediate  89  3.7 Synthesis of UDP-Galactose: Indirect evidence of a galactosyl-enzyme intermediate  92  3.7.1  A Continuous Coupled Assay for UDP-Galactose Detection  92  3.7.2  Suitability of the UDP-Galactose 4-epimerase/UDP-Glucose Dehydrogenase Coupling System  3.7.3  3.7.4  93  End-product Inhibition of the UDP-Galactose 4-epimerase/UDP-Glucose Dehydrogenase Coupling System  95  Kinetic Parameters for LgtC-19  96  viii  CHAPTER 4  100  CONCLUSIONS AND FUTURE WORK  100  4.1 Conclusions  101  4.2 Future Work  106  4.2.1  Short-term Goals  106  4.2.2  Long-term Goals  107  CHAPTER 5  108  MATERIALS AND METHODS  108  5.1 Chemicals, Enzymes, and General Procedures  109  5.1.1  Reagent Sources  109  5.1.2  Analytical Methods  109  5.1.3  Sep Pack™ Purification  110  5.1.4  LgtC-19 Purification  111  5.2 Synthesis  112  5.2.1  112  General Compounds  ix 5.2.2  Prep-scale Enzymic Synthesis of Trisaccharidefroma-Galactosyl Fluoride and Lactose  114  5.3 Enzymology  115  5.3.1  General Procedures  115  5.3.2  Determination of Steady-state Kinetic Parameters  116  5.3.3  Assessment of Coupling System Suitability  118  5.3.4  Preincubation Studies  118  5.3.5  Metal Cofactor Specificity  119  5.3.6  Cleland Kinetics-GnV/ Experiments  119  5.3.7  pH Dependence Studies  120  5.3.8  Inactivation Experiment  121  5.3.9  Labelling Experiment with Dial-UDP  122  5.3.10 Labelling Experiment with UDP-galactose  123  5.3.11 Burst Experiment  123  5.3.12 "Mini" pH Study  124  5.3.13 Transglycosylation Experiment  124  5.3.14 Trouble-shooting Mn Interference with the F" Electrode  125  2+  5.3.15 Kinetic Measurements in the Presence of UDP-glucose  125  5.3.16 Enzymic Synthesis of UDP-galactose  126  REFERENCES  129  APPENDIX A  137  X  BASIC CONCEPTS OF ENZYME CATALYSIS  A.I  Basic Enzyme Kinetics  A.II Deriving Steady-State Velocity Equations  137  138  142  A.n.i  Random Bi Bi Sequential Reaction Mechanism  146  A.n.ii  Ordered Bi Bi Sequential Reaction Mechanism  150  A.n.iii Ping Pong BiBi Reaction Mechanism  A.III Enzyme Inhibition A.m.i  Competitive Inhibition  153  155 156  A.HI.ii Non-competitive Inhibition  157  A.III.iii Mixed Inhibition  157  A.HI.iv Uncompetitive Inhibition  158  A. IV Kinetics of Inactivation for lgtC-19  158  APPENDIX B  161  GRAPHICAL REPRESENTATION OF KINETIC DATA  161  B.I  Lineweaver-Burk Plots for the Transfer of Galactose from UDP-galactose to Lactose at 30°C, pH 7.5  162  xi  B.n Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 7.5 (in HEPES buffer) using Cleland's Method for a Bi Bi Sequential Reaction  163  B.DJ Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 6.0 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  165  B.IV Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 6.5 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  166  B.V Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 7.0 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  167  B.VI Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 7.0 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  168  B.VH Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 7.5 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  169  xii  B.VTfl Graphical Analysis of the Mechanism of Galactosyl Transfer from UDPGalactose to Lactose by LgtC-19 at 30°C, pH 7.7 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  170  B.LX Effect of MnCh Concentration of the Efficiency of Fluoride Ion Electrode Measurements of F Concentration at 30°C, pH 7.0  171  B.X Mass Spectra of Reaction Mixtures Containing FCHASE-lactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  172  B.XI Mass Spectra of Reaction Mixtures Containing FCHASE-galactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  173  B.XH Mass Spectra of Reaction Mixtures Containing FITC-lactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  174  LIST O F T A B L E S  Table 1 Partitioning of galactosyl P-galactosidase between water and methanol.  19  Table 2 Michaelis-Menten parameters for various donor substrates with galactosyl transferase using GlcNAc as acceptor.  28  Table 3 Reaction components for precipitation test.  51  Table 4 Results of clarity test.  52  Table 5 Relative activation of lgtC-19 by common divalent metal cations.  55  Table 6 Observed kinetic parameters for the metal ion binding sites of lgtC-19. Table 7 Kinetic constants for substrates utilised by lgtC-19. Table 8 Kinetic constants determinedfromreplots of l/K (a ) and W (app) as a function of 1/[lactose]. Table 9 T L C resultsfromthe lgtC-19-catalysed transfer of galactose to FITC-lactose (1 hour after initiation; run in 7:2:1:0.1 ethyl acetate/methanol/HzO/acetic acid). m  PP  max  58 61  68  77  Table 10 Molecular weights of fluorescent-labelled reaction products determined by electrospray mass spectroscopy.  78  Table 11 'H-NMR analysis of trisaccharide.  81  Table 12 NOE correlations between anomeric protons and the proton on the linkage carbon.  82  Table 13 TLC results of lgtC-19-catalysed transfer reactions. Determined in 4:2:1:0.1 (22 hours after initiation).  84  Table 14 Kinetic constants obtained using a-galactosyl fluoride as the donor substrate for lgtC-19-catalysed galactosyl transfer to lactose.  87  Table 15 Kinetic parameters for substrates and cofactors utilised by lgtC-19 as determined in 20 mM HEPES buffer at the indicated pH, 30°C.  XV LIST O F F I G U R E S  Figure 1-1  Reaction catalysed by a P-glucosidase.  Figure 1-2  Double displacement mechanism for a retaining  6  P-glucosidase.  8  Figure 1-3  One-step mechanism for an inverting p-glycosidase.  9  Figure 1-4  Two mechanisms proposed for glycosyl transfer: (A) inverting enzyme; (B) retaining enzyme.  10  Figure 1-5  Structure of a glycosyl cation.  13  Figure 1-6  Structural and electronic similarity of glucono-(l,5)lactone and glucosyl cation. Resemblance of nojirimycin to a glucosyl cation: (A) isoelectronic and (B) isosteric also.  13  Figure 1-7 Figure 1-8  14  Mannosidase inhibitors: (1) deoxymannojirimycin, (2) swainsonine, (3) kifunensine, (4) D-mannonolactam amidrazone.  15  Representative azasugar inhibitors of glycosyl transferases: (A) l,2-dideoxy-2-acetamido-nojirimycin derivative; (B) 5-membered azo-iV-acetylglucosamine, a transition state inhibitor of iV-acetylglucosaminyltransferase V.  16  Bisubstrate analogue inhibitor against GlcNAc: P-1,4galactosyl transferase from bovine milk.  16  Transition state structure for a glycosyltransferase reaction.  1^  Figure 1-12  Hydrolysis vs. methanolysis of aryl p-galactosides.  1^  Figure 1-13  Lifetimes of oxocarbenium ions in water at 25°C.  Figure 1-9  Figure 1-10  Figure 1-11  20  XVI Figure 1-14 Figure 1-15  Isotope exchange in the formation of an oxocarbenium ion intermediate.  22  Direct displacement mechanism in which no isotope exchange is observed.  23  Figure 1-16  P-D-Glucopyranosyl pyridinium salt.  24  Figure 1-17 Figure 1-18  Lipopolysaccharide-linkedtrisaccharide. Model for the domain structure and putative topology oflgtC.  33  Figure 2-1  34  Synthetic acceptor substrates. (A) FCHASE-lactose (B) FCHASE-galactose (C) FITC-lactose.  45  HPLCfromreaction mixtures containing 60 mM UDPgalactose, 62.5 mM lactose, 15 mM MnCfe, and 5 mM DTT in 100 mM HEPES buffer, pH 7.5 at 30°C. (a) Test reaction containing lgtC-19 (b) Control reaction without lgtC-19.  47  Figure 2-3  Correlation curve for the LDH/PK coupling system.  48  Figure 2-4  Rate observed as a function of preincubation time using standard assay conditions at pH 7.5,30°C (o) DTT added at time of reaction initiation (•) DTT added to preincubation mixture.  53  Figure 2-2  Figure 2-5  Figure 2-6  Figure 2-7  Effect of lgtC-19 preincubation time in the presence of 15 mM MnCfe on activity measured under standard assay conditions at pH 7.5,30°C. Rate of lgtC-19-catalysed galactosyl transfer from UDP-galactose to lactose under standard assay conditions at pH 7.5,30°C using selected metal salts as activators. Lineweaver-Burk plots for the rate dependence of the lgtC-19-catalysed galactosyl transfer reaction on metal concentration for Mg (left) and Mn (right). 2+  2+  5 4  56  XVII Figure 2-8  Figure 2-9 Figure 2-10  Figure 2-11  Double reciprocal plots for lgtC-19 activity as a function of UDP-galactose concentration over a range of lactose concentrations using standard assay conditions.  59  Replots of (A) 1/Vmax.app and (B) l/Km.uDPCapp) versus l/[lactose].  60  Double reciprocal plots of lgtC-19 activity as a function of UDP-galactose over a range of lactose concentrations in thepresence of 50 mM UDP-glucose (measured using standard assay conditions at pH 7.0).  63  Enzyme activity as a function of time when stored at a range of pH values.  65  Effect of storage pH on initial activities (v ) and inactivation rates (kob).  66  Effect of lgtC-19 concentration on vbs (measured as a change in absorbance at 340 nm.  67  Effect of pH on Vmax/Km for (A) UDP-galactose and (B) lactose in the lgtC-19-catalysed transfer of galactose using standard assay conditionsfrompH 6.07.7.  69  Figure 2-15  Synthetic route to dial-UDP.  71  Figure 2-16  Mechanism of dial-UDP inactivation.  71  Figure 2-17  LgtC-19 activity as a function of incubation time in dial-UDP (measured using standard conditions at pH 7.5).  72  Inactivation rate of lgtC-19 as a function of dial-UDP concentration. (A) kobs vs. [dial-UDP], (B) LineweaverBurkplot.  73  FfPLC tracings for reaction mixtures containing 15 mM MnCl , 5 mM DTT, 62.5 mM lactose, 150 mM Tris buffer (pH 7.5) and (A) 150 mM a-galactosyl fluoride and 7.5 mM UDP; (B) contents of A plus 0.03 mg lgtC-19; (C) 60 mM UDP-galactose (65 hours after initiation).  80  Figure 2-12  0  S  Figure 2-13 Figure 2-14  Figure 2-18  Figure 3-1  0  2  xvm Figure 3-2  Figure 3-3  'H-NMR of the trisaccharide product of the lgtC-19catalysed transfer reaction utilising a-galactosyl fluoride and lactose as substrates.  81  Rate of galactosyl transferfroma-galactosyl fluoride to lactose (100 mM) by lgtC-19 measured using standard assay conditions with the fluoride ion electrode where (A) the data fit to a polynomial; (B) a plot of v bs as a function of [a-gal-F] produces a straight line.  83  HPLC showing the formation of disaccharide by lgtC-19 using a-galactosyl fluoride as both acceptor and donor substrate.  85  HPLC evidence for the formation of disaccharide by lgtC-19 using a-galactosyl fluoride as both donor and acceptor substrate.  86  2  0  Figure 3-4  Figure 3-5  Figure 3-6  Figure 3-7  pH dependence of v b for the lgtC-19-catalysed transfer of galactosefroma-galactosylfluorideto lactose. 0  S  89  Graphical representation of the enzymatic burst anticipated for a reaction mixture containing 60 pM enzyme.  Figure 3-8  Fluoride ion concentration as a function of time for an assay mixture containing 47 uM lgtC-19.  Figure 3-9  Mechanism for the UDP-galactose 4-epimerase/ UDP-glucose dehydrogenase coupling reaction.  Figure 3-10  Linear correlation between AAA340 (v b ) and lgtC-19 concentration for the UDP-galactose 4-epimerase/ UDP-glucose dehydrogenase assay.  Figure 3-11  0  ^0 gj  Q  ^  S  Effect of UDP-glucuronic acid concentration of v b for the UDP-galactose 4-epimerase / UDP-glucose dehydrogenase coupling system in the presence of 0.04 mM UDP-galactose, (•) no UDP, (•) 0.5 mM UDP. 0  94  S  96  XIX Figure 3-12  Rate of UDP-galactose formation as a function of a-galactosyl fluoride concentration at 30°C using standard assay conditions for the UDP-galactose 4-epimerase/UDP-glucose dehydrogenase coupling system.  97  Effect of acceptor concentration on v b for the lgtC-19catalysed transfer of galactosefroma-galactosyl fluoride using the UDP-galactose 4-epimerase/UDPglucose dehydrogenase coupling system: (•) lactose acceptor; (o) cellobiose acceptor.  98  Velocity versus substrate concentration for a typical enzymatic reaction.  139  Figure A-2  A typical Lineweaver-Burk plot.  1^1  Figure A-3  Double-reciprocal plot for a random bi bi sequential reaction with a<l.  Figure A-4  Replot graphs: (A) slope rep lot (B) intercept replot.  Figure A-5  Basic King-Altman Figure for an ordered bi bi sequential reaction.  Figure A-6  Double-reciprocal plots for a ping pong reaction  Figure A-7  Replot graphs: (A) intercept replot (B) replot of apparent Michaelis constants.  Figure 3-13  Figure A-1  0  S  ^  1^  XX  ABBREVIATIONS AND S Y M B O L S  s  molar absorptivity  8  chemical shift  (a—>a)  retention of axial substitution in going from substrate to product  (e-»e)  retention of equatorial substitution in going from substrate to product  a-gal-F  a-galactosyl fluoride  v  initial velocity  0  A340  absorbance measured at 340 nm  AMPSO  3-[iV-Morpholino]ethanesulfonic acid  app  apparent  Asp  aspartate  ATP  adenosine triphosphate  au  absorbance units  AX  UDP-galactose (or any NDP-glycosyl donor)  B  lactose (or other glycosyl acceptor)  bi  a reaction step involving 2 reactants  BSA  bovine serum albumin  Cex  Cellulomonasfimi|3-l,4-exoglycanase  COSY  correlated spectroscopy  DCC  dicyclohexylcarbodiimide  dial-UDP  2'3'-uridine dialdehyde-5-diphosphate  DTT  dithiothreitol  XXI  EC  enzyme classification  EDTA  ethylenediamine tetra-acetic acid  EI  enzyme-inhibitor complex  ES  enzyme-substrate complex  FCHASE  6-(5-fluoresceincarboxamido)-hexanoic acid succimidyl ester  FLTC  fluorescein isothiocyanate  Fuc T V  a-1,3-fiicosyltransferase V  Gal  galactose  GDP  guanidine diphosphate  Glc  glucose  GlcNAc  2-acetamido-2-deoxyglucose  Glu  glutamate  HCA  hydrophobic cluster analysis  HEPES  A/-[2-hydroxyethyl]piperazine-iV-[2-ethanesulfonic acid]  HetCor  heteronuclear chemical shift correlation  FfPLC  high-performance liquid chromatography  I-mix  inhibition mixture  K  apparent binding constant  app  kcat/K  m  apparent second-order rate constant (also known as the specificity constant)  Ki  dissociation constant for inhibitor binding  ki  inactivation rate constant  kobs  pseudofirst-orderrate constant  LC  liquid chromatography  LDH  lactate dehydrogenase type DTfromrabbit muscle  lgt  lipopolysaccharide galactosyl transferase  lgtC  lgt obtained as a product of the 3 enzyme in the lgt gene cluster  LPS  lipopolysaccharide  MES  2-[iV-morpholino]ethanesulfonic acid]  mM  millimolar (1 millimole solute/ L solvent)  MS  mass spectrometry  NAD  rd  +  nicotinamide adenine dinucleotide (oxidized form)  NADH  nicotinamide adenine dinucleotide (reduced form)  NaOAc  sodium acetate  NDP  nucleotide diphosphate  nm  nanometers  NMR  nuclear magnetic resonance  NOE  nuclear Overhauser effect  PIPES  piperazine-iVAr-bis[2-ethanesulfonic acid]  PK  pyruvate kinase type IIIfromrabbit muscle  pK  a  Rf  r  -log K (where K is the ionization constant) a  a  the ratio of the distance a spot travelsfromthe point of origin to the distance the solvent travels along a TLC plate  SDS-PAGE  sodium dodecylsulfate polyacrylamide gel electrophoresis  TLC  thin-layer chromatography  Trien-HCl  tris[hydroxyethyl]-aminomethane hydrochloride  Tris  tris[hydroxymethyl]-aminomethane hydrochloride  TT  test tube  xxiii UDP  uridine-5'-diphosphate  UDPgal  uridine-5 '-diphosphate galactose  UGT  UDP-glucuronosyl transferase  UMP  uridine-5'-monophosphate  uni  reaction involving one reactant species  UTP  uridine-5'-triphosphate  uvms  ultraviolet/visible light  Vmax  maximum velocity of an enzyme-catalysed reaction  xxiv ACKNOWLEDGEMENTS  I would like to thank my supervisor, Dr. Steve Withers, for introducing me to the fascinating study of glycosyl transferases, for encouraging my independence throughout this study, and for his advice. I would also like to thank Dr. Warren Wakarchuk (NRC, Ottawa) for providing advice and enzyme through the course of this work. I am grateful to Dr. H. M. Holden (University of Wisconsin) and Mr. Rob Campbell (University of British Columbia) for their generous gifts of enzyme. Special thanks to my coworkers in the Withers' lab for their assistance and friendship. I am particularly indebted to Ms. Renee Mosi, who provided guidance when things seemed particularly unclear. I appreciate the help of Dr. Shouming He for performing mass spectrometric analysis, Mr. Cameron Mackereth and Ms. Marietta Austria for obtaining the NMR spectra. I thank Dr. Martin Tanner, Dr. Tommy Harvey, and Mr. Rob Campbell for their helpful discussions. I would like to thank Dr. Desiree Vanderwel for her encouragement and example. Finally, I am most thankful for my husband, Ayoub, whose love and support over these years apart meant so much.  CHAPTER 1  INTRODUCTION  2  1.1 G e n e r a l Introduction  Oligosaccharides are macromolecules composed of up to 20 monosaccharide units. Amongst the most important of biochemical building blocks, they constitute the basic components of complex carbohydrates.  These polysaccharides, or glycans,  comprise the bulk of the Earth's biomass. Unlike the biosynthesis of polymers such as nucleic acids and proteins which acquire structural order through template guided mechanisms, glycans are assembled by a special class of enzymes known as glycosyltransferases which operate without the strict control of a template (Kleene & Berger, 1993). Carbohydrates are important constituents of glycoproteins, proteoglycans, and glycolipids in which they are covalently bound to protein, glyceride, or hydrophobic entities (Collins & Ferrier, 1995). Collectively, these "glycoconjugates" show extensive variety in structure and function. They occur naturally in plants, animals, and microorganisms.  Far from the original belief that carbohydrates primarily function as  structural components, it has become increasingly evident that these polymers possess an unsurpassed ability to create unique structural arrays that are valuable for their content as well as their form. Glycoconjugates have particular importance in human biology and medicine where they are involved in many biological functions. These including blood clotting, lubrication, structural support, immunological protection, cell adhesion, hormone activity, and cell recognition events. Oligosaccharides are compounds with considerable potential both as therapeutic agents and as reagents for clinical assays. Most pathogens  3 use carbohydrate-binding proteins to attach to cells and initiate disease (Zopf & Roth, 1996). For mammals, thefirstline of defence against these infectious diseases consists of decoy oligosaccharides which are found in the mucous layer lining all exposed epithelial cells and can be detected in all body fluids. These attach to the carbohydrate-binding proteins so that these pathogens can be clearedfromthe organism. The ability to synthesise oligosaccharides is an important step towards the treatment of these diseases. Unfortunately, the chemical synthesis of many potentially interesting carbohydrates remains a daunting task because of the very nature of the saccharide units themselves. Many possibilities for the formation of positional isomers exist since there are numerous substituent groups on the sugars that may become involved in bond formation. To maintain control over these events, protection of the many functional moieties present on individual saccharide units must be carefully controlled, necessitating successive protection and deprotection steps with the addition of each sugar unit. In addition to difficulties associated with the control of regiochemistry in these natural products, there also exists potential for the formation of different anomeric forms. As a result of these problems and the poor yields of desired product which are obtained, the chemical synthesis of most oligosaccharides is not generally feasible on a commercial scale. To circumvent these problems, the help of enzymes involved in the biochemical synthesis and degradation of oligosaccharides and carbohydrates is increasingly being sought. There are two classes of enzymes that are used to transform carbohydrates in Nature:  glycosidases and glycosyl transferases.  Though their function within  biochemical processes is broad, the activity of glycosidases is largely catabolic.  4 Conversely, glycosyl transferases are more closely associated with anabolic processes. Despite the different functions of these two kinds of enzymes, the mechanism by which each assembles or disassembles oligosaccharides is believed to be quite similar. Enzymatic synthesis using glycosidases, glycosyl transferases or a combination thereof is now being considered as a possible solution to the difficulties associated with the chemical synthesis of oligosaccharides. Certainly, the combined use of chemical and enzymatic methods has enabled the synthesis of several novel oligosaccharides (Gross et al, 1987; Gross et al., 1989; Srivastava et al., 1993).  1.2 The Mechanism of Glycosyl Transfer To evaluate the importance of these enzymes in the enzymatic synthesis of oligosaccharides, an understanding of the mechanistic features of each class is required. Glycosidases have been extensively studied and much is known about the mechanistic paths they utilise and the substrates on which they act. They are readily available and relatively easy to work with. Glycosyl transferases have eluded significant investigation until recently. Because they are often membrane-associated, these enzymes tend to be difficult to isolate and purify, unstable in solution, and expensive to purchase. As a result, relatively few glycosyl transferases have yet been studied in detail with respect to mechanistic behaviour. Recent advances in techniques for cloning, expression, and purification of glycosyl transferases have led to an explosion of interest in this field. - Many glycosyl transferases that participate in the synthesis of glycoconjugates using nucleotide-activated glycosyl donors have been cloned in the last decade and mechanistic investigations are now getting underway. Despite the recent interest in thisfield,there  5 remains a vast difference in the amount of information which has been obtained for these two mechanistically-related enzyme systems. Thus, a discussion of glycosyl transferase mechanism and utility for oligosaccharide synthesis necessitates a parallel discussion of glycosidases.  1.2.1 The Catalytic Mechanism of Glycosidases Glycosidases are a broad class of enzymes active in the hydrolysis of oligo- and polysaccharides, which carry out, inter alia, the first step in the biochemical processing of nutrients. The ability of these enzymes to degrade biomass makes them extremely useful for industrial applications. For example, they are used in the processing of food and beverages, the processing of wood pulp into paper, and the manufacture of fuel. These enzymes have been widely studied and much is known about the mechanistic paths they utilise and the substrates on which they act (Legler, 1990; Hehre, 1989; Davies & Henrissat, 1995; Kempton & Withers, 1992; MacLeod et al., 1996; Gebler et al., 1992a; Jeong et al., 1996; Day & Withers, 1986; Kobayashi et al., 1996; Lawson et al., 1996). The simplest of these reactions is the cleavage of a disaccharide into two sugar molecules. The large family of glycosidases is divided into subgroups on the basis of several characteristics.  First, they are classified according to the sugar, or glycone, moiety  toward which they are most reactive (e.g., a galactosidase is most reactive towards galactosides).  Specificity is often further characterised with regard to ring size (i.e.  pyranose, furanose, etc.).  Next, they are classed as "a" or "p" depending on the  stereochemistry of the glycosidic bond which they cleave. Finally, and most importantly from a mechanistic approach, the stereochemical outcome of the catalysed reaction is  6 considered. Hydrolysis may occur with either inversion or retention at the anomeric center of the reducing sugar product (Figure 1-1).  To reduce ambiguity in the  designation of pyranosyl enzymes, the enzyme can be classified in terms of the stereochemical configuration of the starting material and product. For example, a glycosidase that hydrolyses a P-glucoside to form p-glucose would be an (e—»e)glucosidase. If the product formed was a-glucose, the enzyme would be classified as an (e—>a)-glucosidase. Thus, the axial or equatorial positioning of the anomeric substituent in the preferred conformation is described.  Figure 1-1: Reaction catalysed by a /3-glucosidase  Glycosidases have been successfully classified into families based on sequence homology, with enzymes that are related by close sequence homology exhibiting similar catalytic behaviour.  In fact, this method has been so successful that sequence  comparisons between closely related enzymes has enabled the identification of catalytically-important residues. Although relaxed specificity for the glycone portion of  7 the substrate may be observed, these enzymes are specific with regards to the configuration (a or p) of the bond which they hydrolyse (Sinnott, 1987). Many retaining glycosidases transfer glycosyl residues to low molecular weight alcohols such as methanol, as well as to their natural acceptor, water. This activity is never found with inverting glycosidases. Were an inverting glycosidase to have nonspecific transferase activity, it would have to be anomerically ^discriminate (Sinnott, 1987). Enzymic glycosyl transfer therefore takes place with inversion of anomeric configuration only when the acceptor and donor have sufficient structural diversity to ensure that glycosyl transfer is accomplished with absolute anomeric selectivity while obeying the law of microscopic reversibility (Davies et al., 1997). The mechanism of glycosidases has been studied in detail. In 1953 Koshland first provided insight into the two fundamentally different mechanisms arising from the stereochemical outcome of these reactions. First, a double-displacement mechanism resulting in retention and second, a single displacement mechanism resulting in inversion (Figures 1-2 and 1-3, respectively).  His proposal accommodates all subsequent  mechanistic features uncovered for these enzymes. Important features of this mechanism include: (1)  A group located within the enzyme active site has an unshared pair of electrons enabling it to act as a nucleophile.  This functionality lies  adjacent to the anomeric center of the donor substrate, on the opposite side of the sugar ringfromthe aglycone. (2)  A covalent glycosyl-enzyme intermediate of inverted configuration forms between the nucleophile and C(l) of the sugar. This allows the second  8  substrate (water) to react, resulting in a second inversion, to form a final product having the same configuration as the initial substrate. (3)  The covalent intermediate which forms is reached, in both directions, via an oxocarbenium-ion-like transition state.  (4)  General acid catalysis may assist aglycone departure, but is not essential for all substrates. General base catalysis may assist attack of the acceptor (e.g. water) in the deglycosylation step.  (5)  Non-covalent interactions provide most of the rate acceleration.  Figure 1-2: Double displacement mechanism for a retaining fi-glucosidase  9  Figure 1-3: One-step mechanism for an inverting fi-glycosidase  For many of the glycosidases currently classified, these mechanisms have been confirmed and the residues acting as nucleophile and acid/base catalyst have been identified (MacLeod, 1996; Mcintosh et al., 1996; Lehmann & Schlesselmann, 1983). In all cases to date, these are carboxylate moieties found on glutamate or aspartate residues.  1.2.2 The Proposed Mechanism for Glycosyl Transferases All enzymes which transfer glycosyl groups to an acceptor substrate are considered glycosyl transferases. However, by convention, only those transferring glycosyl residues from nucleotide-activated sugar molecules to other carbohydrates or aglycons such as peptides or lipids, the so-called "Leloir donors", are called glycosyl transferases. These particular enzymes are important because many biologically active carbohydrate chains are constructed primarily by this type of transfer reaction. While the biological pathway of oligosaccharide formation is controlled by genes encoding the production of each glycosyl transferase, the individual oligosaccharide structures are determined by the donor and acceptor substrate specificities of the individual transferases involved (Ichikawa et al., 1992).  Each glycosidic linkage is  formed by the action of a unique enzyme with the sole function of synthesising that  10 particular bond. The mechanism of glycosyl transferases proposed by Saxena et al. (1995; Figure 1-4) closely resembles that of glycosidases. Although the action of this class of enzymes is believed to mirror that of the more extensively studied glycosidases, the validity of this comparison remains largely unexplored. ^  q  —  | i n v e r t i n g )  \  t  •cu  UDP  ^  o  —  H  UDP ,  UDP  Figure 1-4: Two mechanisms proposedfor glycosyl transfer: (A) inverting enzyme; (B) retaining enzyme (Saxena et al, 1995)  The sequence-based classification system for glycosidases has been of great utility in the study of glycosyl hydrolases. Until recently, no such system had been fully described for glycosyl transferases. However, hydrophobic cluster analysis and other sequence homology comparisons have now been used to develop a classification system  11 for this class of enzymes (Campbell et al., 1997a). The number of glycosyl transferases is staggering; already there are more than 500 entries of EC 2.4.1.x, the nucleotide diphosphate (NDP) sugar hexosyl transferases. Although there are close similarities between the reactions they catalyse, these enzymes differ substantially in the nature of the sugar donor they utilise. This variability may be responsible for the difficulties that have been encountered with this type of organisational scheme.  The recent success in  constructing a classification system for nucleotide-diphospho-sugar hexosyltransferases is based on sequence analysis of proteins listed in the SwissProt and EMBL/GenBank databanks. Campbell et al. (1997a) used representative sequencesfromeach EC family as templates for their BLAST similarity searches. Further analysis using hydrophobic cluster analysis (HCA) was used when direct comparisons showed low homology. Using this method, 552 sequences were classified into 25 families. Each family represented a grouping of two or more sequences of significant amino acid or HCA similarity over a length greater than 100 residues, with no similarity to other families. Less than 1% of the total sample they examined could not be assigned to any family using this method. Confirmation of these classifications will only be provided by x-ray crystal structures which, at this time, remain elusive. 1.2.3 Oxocarbenium Ion-like Transition States  Transition States of Hydrolases Based on Koshland's original discussion concerning the nature of glycosyl transfer reactions, much effort has been placed on elucidating the nature of the transition state and the ensuing glycosyl-enzyme intermediate formed by retaining glycosidases.  12 Speculation regarding the mechanism of this enzymic reaction stemsfromthe chemistry of acetals reactingfreein solution. In these systems, the primary effect of an oxygen atom adjacent to the reaction center is stabilisation of positive charge on the central atoms. This is accomplished when the lone pairs on oxygen overlap with the vacant porbital of a neighbouring carbenium ion center. The bond between C(l) and the leaving group, or aglycone, then may break to a much greater extent before there is significant bond formation with the attacking nucleophile. The limit of such processes in water is the complete dissociation of the leaving group to form a discrete oxocarbenium ion intermediate. This concept has been adopted in the discussion of enzymes where distinct oxocarbenium ions continue to be suggested as intermediates. Much of the evidence supporting the existence of oxocarbenium ion-like transition states is derived from studies of the interaction between glycosidases and transition state analogues (Legler, 1990; Jeong et al., 1996; Day & Withers, 1986). Additionally, the measurement of kinetic isotope effects can provide substantial information regarding the nature of the transition state in an enzyme-catalysed reaction (Kempton & Withers, 1992). Positive kinetic isotope effects (ko/kH>l) signify a change from sp to sp character upon the formation of the transition state, as would be expected 3  2  in the formation of an oxocarbenium ion-like transition state. Generally, transition state analogues act as competitive inhibitors. These are often stable monosaccharide derivatives which differfromtheir parent glycoside in one or both of the following characteristics: C(l) and 0(5) carry a full positive charge between them; C(l), C(2), C(5) and 0(5) are coplanar. If one accepts the paradigm that the catalytic strength that enzymes provide is founded on their ability to stabilise the  13 high-energy transition state more efficiently than the ground state (Fersht, 1985), it is clear that transition state mimics of such compounds would exhibit enhanced binding affinities toward the active sites of both glycosidases and glycosyltransferases relative to the normal substrates. The strong binding affinity observed for transition state analogues that resemble glycosyl cations supports the notion that an oxocarbenium ion-like transition state exists.  Figure 1-5: Structure of a glycosyl cation  For example, aldonolactones are bound tightly to (e-»e)-glycosidases with K, values two to three orders of magnitude lower than Kg values for substrates. These lactones resemble the corresponding glycosyl cations in that C(l), 0(5), C(2), and C(5) are coplanar. Also, the ester resonance places some positive charge on the ring oxygen, resulting in a charge distribution that closely resembles that of a glycosyl cation (Figure 1-6; Sinnott, 1987).  vs.  Figure 1-6: Structural and electronic similarity of glucono-(1,5) -lactone and glucosyl cation  14 Although the mechanistic data available for (a-»a)-glycosidases are much leaner than for (e-»e)-enzymes, these enzymes also appear to operate via a double-displacement mechanism with an enzyme carboxylate being P-glycosylated in the glycosyl-enzyme intermediate. However, there are no examples of (a->a)-glycosidase-catalysed reactions for which the glycosyl enzyme intermediate accumulates and their response towards tight-binding inhibitors is often significantly different from that of (e-»e)-enzymes, aldono lactones being bound -100 times less tightly to (a->a)- than to (e->e)-enzymes (Reese et al., 1971). Both Reese (1971) and Sinnott (1987) attribute this weaker binding to differences in the relative position of the acid catalyst in these two classes of enzymes. That is, the carbonyl oxygen atom of the lactone in the (e-»e) active site can form a hydrogen bond with the protonated acid catalyst but not in the active site of an (a-»a)enzyme. No crystallographic data has been provided for the support of this claim  A  B  Figure 1-7: Resemblance of nojirimycin to a glucosyl cation: (A) isoelectronic and (B) isosteric also  Alternatively, it has been found that compounds with nitrogen in the pyranose ring in place of oxygen, such as nojirimycin (Figure 1-7), are bound fairly tightly to (a-»a)glycosidases.  In fact, several azasugars are reported to be potent inhibitors of  glycosidases (Figure 1-8; Pan et al., 1992).  15  1  3  2  4  Figure 1-8: Mannosidase inhibitors: (1) deoxymannojirimycin, (2) swainsonine, (3) kifunensine, (4) D-manno-amidine.  Transition States of Transferases Like glycosidase reactions, glycosyltransferase reactions are believed to proceed through a transition state with a positively-charged, half-chair conformation. Retaining glycosyl transferases are also (a—»a)-enzymes. Thus, their response to inhibitors such as aldonolactones and azasugars might be expected to resemble that of the (a-»a)glycosidases. In fact, this is exactly what has been observed. Kim et. al. (1988a) have reported tight binding of glucono-l,5-lactone and deoxynojirimycin to glycogen synthetase from rabbit muscle. They found this lactone to be a strong non-competitive inhibitor versus UDP-glucose with K =0.091±0.004 mM and Kii=0.70±0.09 mM. is  Deoxynojirimycin had a more complicated inhibitory effect on this enzyme that probably stemsfromits ability to bind to both the glycogen and the UDP-glucose site. A large secondary isotope effect of ki^ko = 1.23 for V  max  , along with the inhibition studies, was  taken as evidence for the involvement of an oxocarbenium ion transition state or intermediate in the reaction mechanism. Azasugars function as effective inhibitors of glycosyltransferases (Figure 1-9; Kajimoto et al., 1991; Takaoka et al., 1993; Kim et al., 1988b), especially in the presence of nucleotide phosphates (Wong et al., 1992a).  16  A  B  Figure 1-9: Representative azasugar inhibitors of glycosyl transferases: (A) 1,2-dideoxy2-acetamido-nojirimycin derivative (Kajimoto et al., 1991); (B) a substituted pyrrolidine transition state analogue inhibitor of N-acetylglucosaminyltransferase V (Takaoka et al, 1993).  Recent inhibition studies of glycosyltransferases have incorporated both acceptor and donor substrate components to produce potent transition state analogue inhibitors (Figure 1-10; Takaoka et al., 1993; Hashimoto et al., 1997; Murray et al., 1997; Wong et al., 1992a).  O—UDP Figure 1-10: Bisubstrate analogue inhibitor against GlcNAc: f3-l,4-galactosyl transferasefrombovine milk (Hashimoto et al, 1997)  Further evidence for the proposed oxocarbenium-ion transition state is provided by the observation of positive cc-secondary kinetic isotope effects (k /k >1) when H  D  reactions catalysed by both glycosidases (Sinnott & Withers, 1974; Tull & Withers, 1994; Sinnott, 1978; Umezurike, 1988; Kempton & Withers, 1992; Legler & Julich, 1984) and glycosyltransferases (Kim et al., 1988a,b; Murray et al., 1997) are examined. These  17 effects were observed for both the glycosylation and deglycosylation steps of glycosidases. If the reaction intermediate, rather than the transition state, resembled an oxocarbenium ion, a reverse isotope effect (k /k <l) would be expected for the H  D  deglycosylation step. Collectively, these observations suggest that these reactions are bimolecular but involve exploded transition states in which the anomeric carbon has substantial sp character. This is consistent with a transition state in which there is very 2  little bonding with either the mcoming nucleophile or the departing leaving group (Figure 1-11). -1  t  Nu: i  Ott  UDP  Figure 1-11: Transition state structure for a glycosyltransferase reaction  1.2.4 Glycosyl-Enzyme Intermediate  Fersht (1985) lists three criteria to be considered as proof that an intermediate is formed along a reaction pathway: (1) The intermediate can be isolated and characterised. (2) The intermediate is formed sufficiently rapidly to be on the reaction pathway. (3) The intermediate reacts sufficiently rapidly to be on the reaction pathway.  18  He also warns that an intermediate which has been isolated might be the result of the rearrangement of another intermediate and might not itself be on the reaction pathway. This is why criteria (2) and (3) are required. An elegant experiment was performed by Stokes and Wilson (1972). They observed the partitioning ratios between hydrolysis and methanolysis products formed by P-galactosidase-catalysed solvolysis of a series of aryl p-galactosides in a mixture of buffer and methanol. They found that the ratio of methanolysis product to hydrolysis product was essentially constant for all of the substrates studied. The relative rates of the reactions involving these substrates were highly dependent on the nature of the aglycone portion of the substrate molecules (Table 1). These results suggest that both nucleophiles attack an enzymatic intermediate common to all substrates and that the rate at which that intermediate forms is rate-determining and governed by the aglycone leaving group ability.  These results do not, however, provide insight into the nature of that  intermediate. Obviously, the beautiful evidence gleanedfromthe study by Stokes and Wilson cannot be taken as "proof using Fersht's criteria since the enzyme-intermediate was never isolated. Gal OH  Gal—X  +  E  E»Gal—X  E » Gal -X GalOMe  Figure 1-12: Hydrolysis vs. methanolysis of aryl fi-galactosides  19 Table 1: Partitioning ofgalactosylfi-galactosidasebetween water and methanol (Stokes & Wilson, 1972).  Aglycone  Methanolysis/Hydro lysis  Vmax (relative)  2'-nitrophenyl  1.97  1.0  3-nitrophenyl  1.96  0.9  3'-chlorophenyl  2.08  0.5  4'-nitrophenyl  1.99  0.2  phenyl  1.94  0.1  4'-methoxyphenyl  2.14  0.1  4-chlorophenyl  2.13  0.02  4-bromophenyl  2.02  0.02  methyl  2.02  0.06  As shown by the kinetic isotope effects for glycosidases and glycosyltransferases (section 1.2.3), there is a significant change in hybridisation at the anomeric center of a glycoside when approaching the transition state. This effect is observed on both the glycosylation and deglycosylation steps for glycosidases and, therefore, points to an sp  3  hybridised intermediate species. Recall Koshland's (1953) proposed covalent glycosylenzyme intermediate. These observations support the existence of such a species. Those involved in the study of these enzymes are divided into two streams of thought: those who believe a covalent glycosyl-enzyme intermediate is involved in the reaction mechanism and those who believe an ion pair consisting of a discrete oxocarbenium ion and anionic nucleophile such as a carboxylate exists. The major argument against the existence of a discrete oxocarbenium ion intermediate arises from  20 the inherent instability of such species (Figure 1-13). The stability of these cationic species is highly dependent on the solvent environment. To prove that an intermediate is formed along the reaction pathway, it must be directly observed and demonstrated to have a sensible lifetime under the conditions of the reaction. +• +  o-  20 ns McClelland and Ahmad, 1978  700 ns  20 ns  Steenken and McClelland, 1989  Figure 1-13: Lifetimes of oxocarbenium ions in water at 25°C  Two criteria can be used to evaluate the "sensibility" of the lifetime of a reactive intermediate. First, the lifetime must be longer than a bond vibration, ~10 seconds. -13  Second, and less exactingly, the intermediate must live long enough for any other part of the precursor molecule to diffuse away. In water, at 25°C, this gives a lower limit of ~10 seconds. Compare the lifetimes of common cations (Figure 1-13) with those of _10  glycosyl-enzyme intermediates at ambient temperatures which range from 1-100 milliseconds (Weber & Fink, 1980): it seems unlikely that such an unstable species could exist as long as this unless a covalent bond had formed, even within the stabilising environment of an enzyme active site (Sinnott & Souchard, 1973). More recently, fluorinated sugars have been used to probe glycosidase mechanisms (Tull et al., 1996). The use of 2-deoxy-2-fluoro-p-D-glucosides has ,  21 addressed the criteria outlined by Fersht more comprehensively (Withers et al., 1987; Withers & Street, 1988). These mechanism-based inactivators function through the formation of a relatively stable glycosyl-enzyme intermediate. By utilising 2-deoxy-2fluoroglycosyl derivatives with good leaving groups, the glycosyl-enzyme intermediate can be formed at a significantly greater rate than it degrades. This effect is a function of the inductive destabilisation of the positively charged oxocarbenium transition state by the C(2) fluorine, slowing the rates of both glycosyl-enzyme formation and hydrolysis. The presence of a reactive leaving group increases the rate of intermediate formation relative to that of hydrolysis, resulting in accumulation of the intermediate. This method has enabled both the identification and isolation of a glycosyl-enzyme intermediate by Withers and Street (1988). Using F NMR, they were able to show that the glycosyl19  enzyme intermediate formed during the hydrolysis of 2-deoxy-2-fluoro-P-Dglucopyranosyl fluoride by a p-glucosidasefromAlcaligenes faecalis (now typed as Agrobacterium sp.)  contains  a  2-deoxy-2-fluoro-a-glycosyl-enzyme  linkage.  Additionally, the crystal structure of a trapped glycosyl-enzyme intermediate of another retaining p-glycosidase, the xylanase/glucanase (Cex)fromCellulomonasfimi,has been obtained and confirms the identity of the catalytic nucleophile as Glu 233 (White et al., 1996). These fluorinated glycosides have also been extensively employed in the identification of active site carboxylate nucleophiles in a number of glycosidases (Withers et al., 1990; Tull et al., 1991; Wang et al., 1993; Gebler et al. 1992b). The mechanistic picture that emerges for retaining glycosidases is now so uniform that Sinnott suggests that good evidence is required that a glycosylated carboxylate group is not involved before other mechanistic proposals can be seriously considered (1987).  22  Tests for a Transferase Intermediate A direct test for the existence of a discrete oxocarbenium intermediate on bovine milk galactosyl transferase was made by Kim et. al. (1988b) using 0-labelled UDP18  galactose and a-lactalbumin. Under these conditions, this enzyme catalyses the transfer of galactosefromUDP-galactose to glucose to form lactose. Thus, the reaction proceeds with net inversion at the anomeric carbon. By using [P- C>2, aP- 0]UDP-galactose, it 18  I8  was possible to determine whether or not the carbon-oxygen bond was cleaved during the formation of an intermediate. Positional isotope exchange of an oxygen-18 atom at the P-nonbridge position to the anomeric oxygen position can only occur if the carbonoxygen bond is completely broken and the transiently formed UDP is able to rotate about the P-phosphoryl group, suggesting the existence of a discrete oxocarbenium ion intermediate (Figure 1-14).  Figure 1-14: Isotope exchange in the formation of an oxocarbenium ion intermediate  23 Alternatively, isotope exchange would not be observed if the catalytic mechanism involved the direct displacement of UDP by the glucose acceptor since bond breakage and formation would occur at nearly the same time (Figure 1-15). Their failure to observe any significant isotope exchange was interpreted as evidence that this reaction occurs by means of a direct displacement mechanism rather than via the formation of a discrete oxocarbenium ion (Kim et al, 1988a).  Figure 1-15: Direct displacement mechanism in which no isotope exchange is observed  Identification of active-site residues for glycosyl transferases has had limited success. Although a variety of both retaining and inverting glycosidases have been studied in detail, almost all of the published data on glycosyl transferases has derived from the study of inverting enzymes. Thus, the identification of a glycosyl-enzyme intermediate or a catalytic nucleophile remains to be actualised for a member of this class. Additionally, only one glycosyl transferase three-dimensional structure has been reported, this being for an inverting glycosyl transferase, the DNA p-glucosyl transferase of Bacteriophage T4 (Vrielink et al., 1994). Unfortunately, the structure of the substratebound complex was only obtained at 2.2 A resolution and clear electron density was revealed only for the uridine diphosphate portion of the substrate. Perhaps crystal  24 structures of retaining glycosyl transferases will provide insight into the nature of the reaction intermediate. 1.2.5 Acid/Base and Electrophilic Catalysis  Acid catalysis of aglycone departure, as mentioned previously, is of varying importance and may not occur at all in some cases. For example, glycosidases such as A. wentii p-glucosidase A 3 hydrolyse P-D-glucopyranosyl pyridinium salts (Figure 1-16) in  which any acid catalysis of leaving group departure is structurally impossible (Legler et al., 1980). Despite the inability of this enzyme to protonate the leaving group of substrates of this type, rate accelerations of 10 -10 have been observed for a series of 8  13  such compounds when compared to rates for SNI departure of pyridine in water (Jones et al, 1977; Sinnott, 1979). Results such as these suggest that the action of the acid catalyst is of relatively low importance to the overall function of the enzyme.  x Figure 1-16:fi-D-Glucopyranosylpyridinium salt  The most convincing evidence for the existence of an acid/base catalyst, strategically placed within the active site of glycosidases, is derived from extensive structural studies. There are numerous x-ray crystal structures of glycosidases which clearly show the correct positioning of active site carboxylates to function as acid/base catalysts (Svensson & Segaard, 1993; White et al, 1996). In addition, the use of sitedirected mutagenesis to selectively replace residues identified from either sequence comparison or x-ray crystallography has enabled the acid/base function of these residues  25 to be determined (Birsan, 1996). These mutants exhibit kinetic behavior consistent with the removal of an acid/base catalyst and lend further support to the existence of acid/base catalysis within the mechanism of glycosyl hydrolysis. Difficulties in acquiring x-ray crystal structures for glycosyl transferases have eliminated this method as a source of information about active site residues. Evidence of catalytically important residues has been limited to studies utilising group-specific chemical reagents and affinity labels to map substrate binding sites and pH profiles to indicate the pKa of catalytically important residues. Much of the work that has been communicated in the literature centers around the fucosyl transferases. Several of these enzymes are involved in the synthesis of Sialyl Lewis X, an oligosaccharide which has been implicated in cancer cell recognition processes. The pH-rate profile of Human a1,3-fucosyltransferase V (FucT V) suggests that the reaction catalysed by this enzyme requires a catalytic base with a pKa of 4.1, presumably a carboxylate such as that found on aspartate (Asp) or glutamate (Glu). The same study showed a mechanistically relevant residue with a pKa of -6.9, suggestive of a histidine (Murray et al., 1996). Failure of this second residue to impact V  max  over the pH range studied indicates that it  may not be involved in the transition state. Other reports on this enzyme have identified an essential cysteine residue in the GDP-fucose binding site (Holmes et al., 1995). Human a-l,3-fucosyltransferase VI, which shows significant sequence homology to FucT V, has both cysteine and histidine residues of catalytic importance (Britten & Bird, 1997). Battaglia et al. (1994) used carboxyl-specific reagents to investigate three UDPglucuronosyl transferases (EC; UGT). These enzymes were representatives from the UGT1 family, of human origin, and the UGT2 family, of rat origin. They found that  26 dicyclohexylcarbodiimide (DCC), l-ethyl-3-(3-dimethylamkopropyl)-carrx)dim^ and A^ethyl-5-phenylisoxazolium-3'-sulfonate all caused rapid first-order inactivation of these enzymes.  Some protection was offered by UDP-glucuronic acid, UDP, and 4-  methylumbelliferone. DCC significantly decreased V  max  without affecting the apparent  K towards substrates. Thus, the data support the involvement of a carboxyl group in the m  catalytic process. They suggest that the reactive carboxyl group is probably conserved among the UGT superfamily. Additionally, they propose the potential involvement of a histidine in conjunction with either an Asp or Glu as the general acid/base catalyst, likely in the form of a charge-relay system Group-specific modification was also used to identify an essential lysine residue in the GDP-fucose binding site (Holmes, 1992). Affinity labelling of the UDP-galactose binding site of bovine galactosyl transferase suggests that an essential lysine residue may be a characteristic feature of this site (Powell & Brew, 1976). Despite the progress that has been made in this area, the identities of the catalytic residues responsible for abstraction of a proton from the acceptor OH in most glycosyltransfer reactions remain elusive. 1.2.6 Electrophilic Catalysis  The true substrate of an enzyme-catalysed reaction may actually be a substrateactivator complex. In these cases, the activator is usually a metal ion. This phenomenon predominates in reactions involving phosphorylated intermediates, especially nucleotides. In fact, almost all enzyme-catalysed reactions involving ATP require Mg . For these 2+  reactions, the actual substrate is MgATP \ In addition to forming an active complex with ATP, the metal ion may also combine with the enzyme at a specific activator site. This increased activation may be essential or nonessential (Segel, 1993).  27 Most glycosyl transferases require the presence of divalent metal cations for activity (Tsopanakis & Herries, 1978; Kuhn et al., 1991a; Permyakov et al., 1993; Navaratnam et al., 1988; Murray et al., 1996; Hendrickson and Imperiali, 1995; Kuhn et al., 1992). The association of many glycosyl-transferring enzymes with metal ions advances speculation that by enhancing leaving-group ability, these ions may constitute a possible source of enzymic rate-enhancement. Studies of bovine galactosyl transferase have indicated the existence of two metal binding sites, the first involved in mamtaining the structural integrity of the protein and the other associated with UDP-galactose binding (O'Keeffe et al., 1980). Navaratnam (1988), based on studies of galactosyl transferase from lactating rat Golgi membranes, proposes that thefirstbinding metal participates directly in the reaction mechanism, whereas the second site is a regulatory one that is subject to activation by proteins. Manganese is reported to depart from bovine galactosyl transferase as a UDP»Mn complex (Tsopanakis & Herries, 1978). Thus, for galactosyl 2+  transferase, leaving group departure may be facilitated by a metal cofactor involved in electrophilic catalysis. Obviously, given the conflicting nature of the data which have been published in the current literature, further examination of the role assumed by metal cofactors in glycosyl transferase reactions is required before any conclusions can be drawn. 1.2.7 Non-covalent Interactions  Beyond the effect of nucleophilic, acid/base, and electrophilic catalysis on the catalytic ability of glycosidases and glycosyl transferases, enormous rate enhancements remain unaccounted for. The catalytic strength of these enzymes is significantly determined by the existence of non-covalent binding interactions between substrates and  28 residues within the enzyme active site. The existence of non-covalent interactions, predominantly hydrogen-bond formation, between substrate hydroxyl groups and functionalities present within the binding site has been repeatedly confirmed by both xray crystal structures of glycosidases bound to substrate analogues, and by the systematic replacement of either key residues using site-directed mutagenesis or, alternatively, replacement of individual hydroxyl groups on the substrate itself (Namchuk & Withers, 1995; Roeser & Legler, 1981; McCarter et al., 1992). For these studies, deoxy-glycoside substrates have been extremely useful. Similarly, the utility of these glycosyl derivatives has been extended to the study of glycosyl transferases. Changes in Vmax and K for bovine galactosyl transferase result m  when non-covalent binding interactions are systematically altered (see Table 2). This effect shows that these interactions have a significant impact on the catalytic efficiency of this enzyme. Although the substrate specificity of this particular galactosyl transferase is broad enough to accept donors other than just UDP-galactose, it is clear that the glycosyl Table 2 : Michaelis-Menten parameters for various donor substrates with galactosyl transferase using GlcNAc as acceptor".  substrate  Vmax  UDP-galactose  (%) 100  (uM) 13.7"  5.5  26  lacks 4-OH (no axial substituent)  0.3  23  4'-OH is equatorial instead of axial  4.0  14  UDP-4-deoxygalactose  d  UDP-glucose  structural differences from UDP-galactose  c  e  lacks 6'-CH OH moiety but contains an axial 4-OH group "TakenfromBerliner and Robinson, 1982. "Bell et al (1976) . CalculatedfromBell et al (1976). "Andree UDP-arabinose  2  e  and Berliner (1978). from thiol protection experiments (Magee and Ebner, 1974) e  29 binding site is nonetheless quite specific.  Evidently, the 4-pyranosyl position can  tolerate an axial, equatorial, or complete lack of hydroxyl group. The removal of the primary -CH2OH group at the 6' position is also tolerated. However, both of these groups contribute to specific enzyme-substrate interactions (probably hydrogen bonds) which are important for optimal catalytic turnover rates (see Appendix A for a general review of enzyme kinetics). The effects of non-covalent interactions are seen in acceptor binding as well. The K values of acceptor sugars with a hydrophobic group at the reducing end were reduced m  by as much as 100-fold when compared to the parent glycoside for a series of related acceptor substrates utilised by Human FucT V (Murray et al., 1996). This significant increase in binding affinity was accompanied by only a 3-fold decrease in kc t. Thus, the a  hydrophobic interactions between the substrate and the enzyme binding site provide substantial binding, thereby increasing enzyme specificity (Sinnott, 1984).  1.3 Enzymic Synthesis of Glycosides 1.3.1 Synthesis using Glycosidases  For synthetic purposes, glycosidases must run in the direction opposite to that normally encountered in Nature. Thus, oligosaccharides may be synthesised by selectively increasing the likelihood of a transglycosylation reaction. There are two approaches to the synthesis of oligosaccharides by glycosidases: a thermodynamic approach and a kinetic approach (Sinnott, 1990; Davies et al., 1997). First, consider the equilibrium that exists between the starting glycoside and the hydrolysis product. If, for example, the substrate for the particular glycosyl hydrolase is  30  lactose, the hydrolysis products are glucose and galactose. A thermodynamic equilibrium exists between the lactose, galactose, and glucose such that the enzyme-catalysed reaction will favour the production of monosaccharide products until they are present at equilibrium amounts. Using the thermodynamic approach to disaccharide synthesis, the two monosaccharides are reacted with the glycosidase to get a small amount of disaccharide product. In some cases, the yield can be improved by moving to nonaqueous solvents, thereby lowering the water concentration and shifting the equilibrium towards disaccharide production. Alternatively, kinetic control of the activity of retaining glycosidases is possible. These enzymes operate via a double-displacement mechanism that involves the formation of a covalent enzyme-substrate intermediate and two transition states (Figure 1-2). When the concentration of substrate and acceptor species exceed their respective Km values, the transfer reactions of these glycosyl-enzyme intermediates are kinetically controlled. In this situation, many different transfer products may accumulate and decay before thermodynamic equilibrium is eventually achieved. The kinetic approach to synthesis involves the addition of an excess of the desired acceptor saccharide to the reaction mixture so that it can compete with water, reacting in its place to form the transglycosylation product. By manipulating the reaction conditions, the donor arid acceptor substrates used, and even the structural composition of the enzymes themselves, these enzymes are able to produce oligosaccharides simply, efficiently, and with fewer side products than currently available synthetic methods (Ichikawa et al., 1992; Baisch et al., 1996a,b,c; Palcic et al., 1989; Blanken & Van den Eijnden, 1985; Flitsch et al., 1991; Heidlas et al., 1992; Hokke et al., 1996; Monsan & Paul, 1995; Thiem, 1995).  31 Despite their general availability and ease of handling, glycosidases present difficulties involved in controlling their natural hydrolytic function, resulting in poor yields of product. Additionally, although stereochemical control is good and only one anomer is formed, it is often difficult to predict or control the regiochemistry of the glycosidic bond which forms (i.e., the formation of 1-2 versus 1-3 versus 1-4 versus 1-6). 1.3.2 Synthesis using Glycosyl Transferases  Unlike glycosidases, glycosyl transferases naturally function to synthesise oligosaccharides. The reactions they catalyse result in specific products with excellent stereochemical and regiochemical control. In addition, reaction proceeds with substantial yield because the reverse reaction does not occur. Despite these apparent advantages, the use of glycosyl transferases for the industrial scale synthesis of oligosaccharides has not, for the most part, proven economically feasible. The sugar nucleotide substrates required by these enzymes are very expensive. Furthermore, glycosyl transferases possessing the desired specificity to make many interesting oligosaccharides are not commercially available. Recent progress in cloning techniques, however, has made several glycosyl transferases available in sufficient quality and quantity, to make enzymatic oligosaccharide synthesis more practical (Paulson & Colley, 1989). Additionally, significant progress has been made in the use of multi-enzyme, substrate regenerating systems (Zervosen & Elling, 1996; Wong et al., 1992b; Hokke et al., 1996). These systems have substantially reduced the costs associated with large-scale oligosaccharide synthesis. Unfortunately, they involve complex enzyme interactions, usually utilising several enzymes as well as a plethora of cofactors and substrates. Complex systems such as these can be difficult to optimise and manage efficiently. Further exploration into the  32 mechanistic pathway of glycosyl transferases may help us to better understand this class of enzymes, enabling better utilisation of natural biochemical processes for the efficient enzymatic synthesis of oligosaccharides. Ultimately, the exploitation of these natural oligosaccharide factories may lead to the use of cheaper, synthetic substrates to serve as glycosyl donors in place of expensive sugar nucleotides.  1.4 Lipopolysaccharide Galactosyl Transferase Gram-positive bacteria have a polysaccharide coating which lies outside a lipid bilayer that surrounds the bacterial cell contents. Glycosyl transferases assemble this structure through the addition of individual sugar units. The enzyme featured in the present study, lipopolysaccharide galactosyl transferase (lgt) isolated from the lipopolysaccharide glycosyl transfer gene cluster of Neisseria meningitidis, is one such enzyme. The third enzyme located on this gene cluster, it is given the designation lgtC. LgtC functions to transfer galactosefromUDP-galactose onto the lipopolysaccharide of N. meningitidis to form an a(l,4) galactosyl linkage. This enzyme has been classified as a  member of Family 8 of the EC 2.4.1.x superfamily, which is composed of NDP-sugar hexosyl transferases. Specifically, it is designated by the acpession number U65788 (EMBL/GenBank databanks; Campbell et al, 1997a). This particular enzyme transfers a galactose unit to the growing lipid-linked polysaccharide, or glycoconjugate, to form a terminal Gala(l,4)Gaip(l,4)Glc linkage.  The enzyme is therefore UDP-galactose:  galactose 4-a-galactosyltransferase. The product formed is a lipopolysaccharide-linked trisaccharide:  33  Figure 1-17: Lipopolysaccharide-linked trisaccharide  The gene encoding lgtC has been cloned and expressed in Escherichia  coli  (Wakarchuk et al., 1997). The overall structure of the enzyme can be described as a short transmembranous domain sandwiched between two luminal domains.  The  transmembranous loop is extended on one side by a loosely folded 'stem' region that communicates with a tightly folded globular catalytic domain (Figure 1-18). This domain structure is typical of glycosyl transferasesfrombacterial species (Wakarchuk et al, 1997). Low interspecies homology of the stem region suggest its primary function may only be to serve as a spacer for the catalytic domain. The expression of glycosyl transferases is tissue and cell-specific. The recombinant enzyme, lgtC-19, differsfromthe enzyme expressed by N. meningitidis  in that 19 amino acid residues have been removedfromthe C-terminal end  of the protein. This modification removes the transmembranous portion of the enzyme, allowing for the expression of soluble protein. It has been shown not to affect enzyme function (Wakarchuk et al., 1997). LgtC-19 has a molecular weight of 33,177 kDa and exists as a monomer. The mutant enzyme, being expressed in E. coli, is not glycosylated.  34  Figure 1-18: Modelfor the domain structure and putative topology of lgtC  Very little is known about this enzyme. Only the most basic characterisation required for expression and purification has been done (personal communicationfromW. Wakarchuk, NRC). The interesting features of this galactosyl transferase include its ability to generate a(l,4) galactosyl linkages and the retaining mechanism by which it accomplishes its task. The mechanistic properties of this class of enzymes are not well understood, so it may be possible to improve catalytic properties through a combination of mechanistic and structural work. 1.4.1 Applications of a(1->4)Galactose Oligosaccharides  Interest in the particular carbohydrate product formed by lgtC-19 relates to the potential applications of the terminal oligosaccharide as a pharmacological agent. oc(l-»4) Galactosyl linkages are a particularly difficult synthetic challenge. Substantial savings of time and money would be provided by the one-step enzymatic synthesis of this linkage. As mentioned previously (section 1.1), many pathogens bind to cell surface carbohydrates. Oligosaccharides which mimic these receptors could find potential use as  35 anti-adhesive drugs. These therapeutic agents work by binding to the toxic proteins released by the infecting micro-organism, freeing them from the mucosal cells to which they bind. In this way, the toxins are dislodgedfromthe host organism's cells and can be swept along and expelledfromthe organism, thereby preventing (or treating) infection. Pathogen proteins, also called adhesins, lectins, or haemagglutinins, have strict requirements for their oligosaccharide ligands. Most adhesin binding sites accommodate oligosaccharides consisting of 3-5 monosaccharides at the terminal, non-reducing end of the oligosaccharide chain (Zopf & Roth, 1996). Adhesin specificity is one of the main factors that determines which species a pathogen can colonise and the site within the organism at which a successful colonisation can be initiated. Of  E. coli  strains isolated from the urinary tract of children with acute  pyelonephritis, 90% are coated with P-pili that bear an adhesin called PapG. PapG recognisesthe sugar sequence Gala(l,4)GaL characteristic of globoseries glyco lipids that are vastly enriched in cells lining the upper urinary tract and identical to the terminal, non-reducing end of the trisaccharide formed by lgtC-19. with  haemorrhagic  colitis,  hemolytic-uremic  E. coli  syndrome,  is closely associated and  thrombotic  thrombocytopenic purpura (Armstrong et al., 1991). The cytotoxic activity of this bacterial strain is attributed to the Shiga toxins that it produces. Additionally, Gala(l,4)Gaip(l,4)Glc forms the oligosaccharide part of the trihexosyl ceramide, or Pk, antigen of the P blood group glycolipids, providing a receptor site for Streptococcus suis, a bacterium known to cause sepsis and meningitis in humans (Haataja et al., 1993). Phase I clinical trials using Synsorb™-bound trisaccharides show that this carbohydrate  36 is a highly effective anti-adhesive agent in the treatment of bacterial infections resulting from these pathogens (Armstrong et al, 1991; Armstrong et al., 1995).  1.5 Use of Cleland Kinetics to Probe Enzyme Mechanism Until the 1950's, most studies of enzyme kinetic activity were based on the HenriMichaelis-Menten or Briggs-Haldane equations for unireactant enzymes (the basic concepts of enzyme kinetics are described in Appendix A). In the decade which followed this work, attempts were made to analyse the kinetic behaviour of bireactant and terreactant enzymes. Methodology for these studies evolvedfromthe equations formulated by Henri, Michaelis, and Menten and were based on those same assumptions. For many enzymes, however, the rapid equilibrium kinetics which form the backdrop for these theories did not apply. Most enzymes catalyse reactions between two or more substrates to yield two or more products so a system of kinetic analysis of these systems was desperately needed. The solution came in 1970 with a clear, uniform procedure for writing kinetic equations that describe multireactant, steady-state enzyme systems (Cleland, 1970). To understand the method Cleland devised for the kinetic analysis of multi-reactant enzymic reactions, an understanding of relevant nomenclature is imperative. First, the number of kinetically important reactants in a given direction is designated as the  reactancy  of the reaction and is identified by the prefix Uni, Bi, Ter, or Quad. Thus, a reaction with two substrates and one product is Bi Uni. This means that the reaction is bireactant in the forward direction and unireactant in the reverse direction. Second, kinetic mechanisms are classified into two major groups:  sequential  or ping pong. Sequential mechanisms are those  in which all reactants must combine with the enzyme before reaction can take place.  37 Alternatively, a ping pong mechanism is one in which one or more products is released before all substrates have bound to the enzyme. Further analysis of sequential reaction mechanisms shows that these reactions can occur in either an ordered or a random fashion. If the reactants and products must bind to or dissociate from the enzyme in an obligatory order, the reaction is said to be ordered whereas, if the order of combination or release is not obligatory the reaction is random Consider the following Bi Bi reaction: E  +  AX  +  B  **  E  +  A  +  BX  (where E = free enzyme, AX = NDP-sugar donor, and B = acceptor sugar)  There are many considerations which must be accounted for during the kinetic analysis of this enzyme-catalysed reaction: (1)  Do both AX and B bind to the enzyme, E, concurrently?  (2)  Is the order in which substrates bind random, or must binding occur in a specific sequence?  (3)  Does group X transfer directlyfromA to B while both substrates are bound to the enzyme or does the reaction proceed by the transfer of XfromA to a site on the enzyme followed by a transfer of Xfromthis site to B?  These questions raise the potential for at least three distinct reaction mechanisms: (1)  Random Bi Bi Sequential  (2)  Ordered Bi Bi Sequential  (3)  Ping Pong Bi Bi  38 Experimental determinations made during the initial steady-state period of the enzymic reaction can be used to differentiate between these mechanisms. During this part of the reaction, the concentration of all enzyme-intermediates is constant and little or no product has formed. See Appendix A for the experimental evidence that can be obtained for each of these three enzyme mechanisms along with a thorough description of the kinetic theory from which these classifications are made.  1.6 Aims of this Study N. meningitidis UDP-Galactose: galactose 4 oc-galactosyl transferase, or lgtC-19, is a  retaining a-galactosyl transferase which has been classified as a member of Family 8 of the NDP-sugar hexosyl transferases. To date, little information is available regarding the kinetic and chemical mechanisms of retaining nucleotide-diphosphate glycosyl transferases. Overall, the objective of this study is to characterise this enzyme as an example of a retaining glycosyl transferase and to determine its potential applications for industrial-scale synthetic use. First the kinetic mechanism will be investigated using experimental methods to enable analysis by Cleland's method. These studies will be performed using the natural donor substrate, UDP-galactose to transfer a galactose unit to lactose, a mimic of the reducing terminus of the natural acceptor substrate. In addition, several different synthetic substrates will be tested as potential acceptor sugars. These are the fluorescent-labelled sugars: FITC-lactose, FCHASE-lactose, and FCHASE-galactose. a-Galactosyl fluoride will be investigated as a potential synthetic donor. Exploration of the kinetic behaviour of lgtC-19 when presented with this alternate donor sugar may help  39 to determine the feasibility of using donor sugars other than UDP-galactose. It is hoped that this synthetic galactosyl donor will be useful for the industrial application of this enzymic process to large-scale oligosaccharide synthesis. Finally, the reaction conditions for the lgtC-19-catalysed galactosyl transfer using both natural and synthetic donor sugars with the synthetic acceptor, lactose will be optimised. The optimal conditions will be identified by determining the effect of pH on the reaction kinetics, ascertaining the effect of metal cofactor concentration on reaction rate, and defining the pH stability characteristics of this enzyme. Furthermore, it is hoped that pH-dependence studies of lgtC-19 activity and inhibition studies using reactive substrate analogues may provide insights into the identity of essential residues in the active site of this enzyme.  40  CHAPTER 2  CHARACTERISATION OF THE LGTC-19 MUTANT FROM N. MENINGITIDIS  USING UDP-GALACTOSE AS THE DONOR SUGAR  41  2.1 Introduction As with many enzymic reactions, the catalysis of glycosyl transfer is a complex, highly coordinated event. In the biological systems from which these catalysts have developed, control over enzyme activity is imperative for efficient function of the organism as a whole. As a result, complex regulatory systems have evolved that enable quick, efficient response to the rapidly changing demands of the organism. In addition to the enzyme responsible for a given reaction and its requisite substrates, there are countless other molecules, ions, and other enzymes present within a cell. These cellular components must interact in such a way that each enzyme is able to operate in an efficient, tightly controlled manner. In this environment, the temperature, ionic strength, and pH define the medium within which the enzyme must function. Additionally, specific cofactors may be present which help to either stabilise the protein or activate its catalytic ability, or both. Some enzymes, such as the glycosyl transferases, are bound to cellular membranes which anchor these proteins in a particular region of the cell. This enables the formation of highly structured synthetic pathways in which a substrate may pass directlyfromone membrane-bound enzyme to the next, facilitating the step-wise construction of complex carbohydrate oligomers (Kleene & Berger, 1993). Because of the complexity of the cellular domain, it is necessary to isolate the individual enzyme of interestfromall others found within the cellular milieu. This purification is one of the most important steps in the study of an enzyme. It is only after  the enzyme has been segregated from the other cellular components that the mechanistic and kinetic behaviour of the enzymic reaction can be studied.  42 Once an enzyme has been isolated from other cellular components, a variety of concerns must be addressed. Most importantly, the stability of the protein must be determined under a variety of conditions so that its activity can be best preserved during and after the purification process. This can be a complicated task. Considerations of ionic strength, pH, and the presence of stabilising cofactors or surfactants in the enzyme suspension have to be made. To ascertain the stability of the enzyme under a variety of these conditions, some method for measuring enzyme activity is essential. Thus, an assay must be developed to monitor the activity of the enzyme over time. In order to measure the activity of the enzyme, the conditions required by the enzyme for activity need to be met. Clearly, protein stability, assay procedure, and the conditions required for enzyme activity, are interdependent. Unravelling the complex requirements of a particular enzyme requires a systematic approach. The first step towards understanding an enzymic reaction is the development of an assay to monitor enzymic activity. Ideally, this assay should be highly specific for the reaction of interest, relatively fast, and simple enough to enable frequent monitoring of enzyme activity as required. Once an assay method is found, it is possible to begin isolating the protein of interest. With the purified protein in hand, the conditions required for optimal activity can be determined. This process is much the same as for any chemical reaction. The temperature and pH at which the enzyme is most active must be determined, cofactors required for activity identified and optimal concentrations resolved, ionic strength adjusted, and finally, the optimal substrate concentrations ascertained to enable the enzyme to function at peak efficiency.  43 2.2 Development of an Assay Method The concentration of substrates or products in a reaction mixture can be measured using a variety of methods. Typically, changes in IJV7VIS absorbance, radioactivity, or fluorescence are measured although NMR spectroscopy and pH changes can also provide quantitative measure of reaction progress (Kim et al, 1988b; Kuhn et al., 1991a; O'Keefe et al., 1980; Khatra et al., 1974; Hendrickson & Imperiali, 1995; Palcic et al., 1988). Continuous assays  are those in which the concentration of substrate or product is  continuously monitored over time. This is the method of choice. It allows the initial velocity of a reaction (v ) to be determined by extrapolation to the starting time and it is 0  often used for reactions that can be observed using a spectroscopic technique such as those in which a chromophoric leaving group is involved. Many examples of this assay technique can be found in the study of glycosidases. In these systems a phenolate ion is producedfromthe displacement of a substituted phenolic aglycone by water (Kempton & Withers, 1992; MacLeod et al., 1996; Wakarchuk et al., 1986; Sun & Budde, 1997). The rate of increase in absorbance is followed at the appropriate wavelength and is directly proportional to the turnover rate. When it is not possible to monitor the reaction directly, a stopped-assay  discontinuous-  or  may be used. This technique involves the removal of aliquotsfroma  reaction mixture at given time points. In these samples, the reaction progress is halted by changing the temperature, pH, or solvent so that no further catalysis is possible. The concentrations of the substrates or products are then determined at each time point using a variety of approaches such as NMR spectroscopy, chromatography, or scintillation counting to monitor radioactivity. The measured concentration can then be plotted as a  44 function of time to produce a kinetic curve from which v can be extrapolated. Stopped 0  assays are less convenient than continuous methods and are difficult to use when monitoring rapid reactions. A third alternative is the use of a coupled assay. When neither the substrate nor the product of an enzymic reaction has a distinguishable or measurable absorbance or fluorescence spectrum, it may be possible to couple the reaction with a second enzymic reaction for which such a change occurs. This is a process whereby the productfromthe reaction of interest serves as a substrate for the coupling enzyme.  By coupling the  reactions in this manner, a continuous spectroscopic procedure for monitoring the first reaction may be realised. For this method to provide an accurate measure of the rate of the enzyme reaction that is of interest, the second reaction must be fast and efficient such that the first reaction is rate-limiting and all of the product from the first reaction is immediately consumed by the second reaction. For example, the pyruvate kinase (PK) reaction can be monitored by coupling to lactate dehydrogenase (LDH):  H C=C—COO| OPO32  ADP  P  ATP  K  +  phosphoenolpyruvate (PEP)  "3°—9,— II O  C O  H C—C—COO|| o 3  +  pyruvate  °-  l  +  NADH  _  W  r  m H  -  H3C—C—COOI OH lactate  +  NAD  45 In this situation, a large excess of LDH ensures that all the pyruvate from the first reaction is immediately converted to lactate. Since NADH absorbs strongly at 340 nm (s=6.22 mNT'cm') but NAD does not, the rate of decrease in A 3 4 0 due to loss of NADH 1  +  is proportional to the rate of pyruvate formation by pyruvate kinase. 2.2.1  Coupled Continuous Assay: UDP-galactose as glycosyl donor  LgtC-19  catalyses  the  reaction:  acceptor  +  UDP-galactose —>  Gala(l-»4)acceptor + UDP, where the acceptor has a terminal galactose unit at the reducing end. Initial studies of this enzyme, conducted by collaborators at the National Research Center (NRC) in Ottawa, had provided insight into conditions conducive to enzyme stability and activity. Additionally, several synthetic glycosides were tested and shown to be effective acceptors for the lgtC-19-catalysed transfer of galactose from UDPgalactose (Figure 2-1 A, B, and C).  Figure 2-1: Synthetic acceptor substrates. (A) FCHASE-lactose (B) FCHASE-galactose (C) FITC-lactose  46 These fluorescent glycosides enabled the reaction progress to be monitored using capillary electrophoresis. Since both the acceptor and the product fluoresce, separation of reaction components enabled quantification of starting material and product over the course of the reaction using a discontinuous assay method. With this assay in place, they were able to develop a method for purifying the enzyme from the other components of the E. coli cells.  The facilities required for capillary electrophoresis were unavailable for our investigation of this enzyme reaction. Thus, an alternative assay was needed. For the reaction described above, direct measurement using a continuous method is not possible. In this situation, the fluorescence of product and substrate are indistinguishable, prohibiting direct measurement as a means of observing reaction progress. The two need to be separated and quantified. In addition, no chromophore is either produced or consumed during the reaction so that UV7VIS spectrometry could not be used to monitor the reaction. A coupled continuous spectrometric assay of glycosyl transferase activity has been described by Gosselin et al (1994). Of particular interest was its success in mapping the progress of the iV-acetylglucosaminyltransferase-catalysed transfer from UDP-iVacetylglucosamine. Their method involved a two-step coupling scheme in which the UDP produced by the transferase reaction is coupled to NADH oxidation via PK and LDH according to equations [l]-[3]: lgtC-19  UDP-sugar  +  acceptor  —>  sugar-acceptor +  UDP [1]  PK  UDP +  PEP - »  pyruvate  +  UTP  [2]  47  LDH  pyruvate  +  NADH  -»  lactate  +  NAD  +  [3]  Thus, the decrease in absorbance at 340 nm is proportional to the amount of transferase present and reflects the rate of UDP formation in the reaction of interest. Coupled with knowledge of the basic requirements for lgtC-19 activity received from collaborators at NRC, this assay procedure provided a starting point for the development of a continuous coupled assay to monitor the lgtC-19-catalysed reaction. Using enzyme provided by the NRC group, lactose was assessed for its viability as an acceptor in the lgtC-19-catalysed reaction with UDP-galactose as donor.  Figure 2-2: HPLC from reaction mixtures containing 60 mM UDP-galactose, 62.5 mM lactose, 15 mMMnCh and 5 mMDTT in 100 mM HEPES buffer, pH 7.5 at 301. (a) Test reaction containing lgtC-19 (b) Control reaction without lgtC-19  The conditions used were those provided by our collaborators and are described in detail in section 5.2.2. Because it is inexpensive and readily available, it was hoped that lactose  48 would provide a more economical route to trisaccharide formation with this system Figure 2-2 shows the HPLC tracings obtainedfromthese reactions. It is clearfromthese results that lactose acts as an acceptor in this reaction. Using standard assay conditions described in 5.3.2, the ability of the PK/LDH coupling system to accurately monitor the lgtC-19-catalysed transfer of galactose from UDP-galactose to lactose was determined at pH 7.5. The initial velocity of the reaction, v , was measured at several different concentrations of lgtC-19. The best-fitting line for 0  the plot of v as a function of lgtC-19 concentration (Figure 2-3) was determined by the 0  least square method as calculated by GraFit 3.0 (Leatherbarrow, 1990). Any change in v  +  i  i  i  i  i  i  i  i  i  0  i  a > a > a > < D < j > a ) < D ( ! > a > a ) < D  [lgtC-19] (mg/ml) Figure 2-3: Correlation curve for the LDH/PK coupling system.  was directly proportional to the change in transferase concentration up to 1.97 x 10" mg/ml at which point a rate of 0.14 mM min" was observed. Thus, the LDH/PK 1  coupling system appears to provide an effective method for continuously monitoring the  49 reaction progress of the lgtC-19-catalysed reaction, over a range of enzyme concentrations and reaction velocities sufficiently broad to satisfy anticipated needs for the study of this reaction. Additionally, this can be accomplished using readily available instrumentation and reaction components  2.3 lgtC-19 Kinetics using UDP-galactose as a Donor Substrate With a convenient method for following the course of the transferase reaction in place, the next task was to optimise reaction conditions. While lactose has been shown to act as an acceptor for galactose in this reaction, no specific kinetic information for this transfer exists. A working K m of 40 pM had been determined for UDP-galactose when FCHASE-lactose was used as the acceptor substrate in this reaction (personal communication, W. Wakarchuk), but the affinity of the enzyme for this substrate may be affected by the change in acceptor. 2.3.1 Substrate Optimisation  The kinetic parameters for the lgtC-19-catalysed transfer of galactosefromUDPgalactose to lactose were determined using the continuous coupled assay described above and varying the concentration of lactose from 1 mM to 125 mM. Though the upper limit of this concentration range is merely 3 times that of the K m ultimately determined, the use of higher lactose concentrations may have led to effects on solvent structure. At a UDPgalactose concentration of 1 mM at pH 7.5, K m = 41 mM and kcat = 2.2 x 10 min" for 3  1  lactose. The kinetic parameters for UDP-galactose were determined using the same reaction conditions except that the concentration of lactose was held constant at 125 mM  50 while that of UDP-galactose was varied. Under these conditions, uM and  kcat =  Km,uDP-gaiactose  =  29.8  3.3 x 10 min" . 3  1  2.3.2 Cofactor Optimisation As mentioned previously (section 1.2.6), most glycosyl transferases require the assistance of divalent metal ions for catalysis.  Currently, the precise nature of this  dependence is unknown. Metal cofactors commonly utilised by glycosyl transferases include magnesium (Mg ) and manganese (Mn ). Of these, M n 2+  metal ion of preference (Kuhn et al., 1992).  2+  2+  is most often the  Our NRC collaborators had identified  manganese as the metal cofactor which supported lgtC-19 activity, although a detailed characterisation of the interaction between this cofactor and lgtC-19 had not been undertaken. Using standard reaction conditions, the concentration of MnCfe added to the reaction mixture was variedfrom0-35 mM in the presence of 1 mM UDP-galactose and 125 mM lactose at pH 7.5. Non-linear regression (GraFit 3.0; Leatherbarrow, 1990) of the plot, v versus MnCk concentration, gave the kinetic constants Km,Mn2+ = 4.9 mM and 0  kcat=  2.0x10 min" . Ideally, saturating concentrations of M n 3  1  2+  would be employed in  subsequent assays. However, since the interaction of MnCk with another essential assay component, dithiothreitol (DTT; see below), is a source of interference with the coupled assay, and all previous assays (2.4.1) have been performed in the presence of 15 mM MnCk, no changes were made to the composition of the standard assay mixture based on these results. Although the amount of metal salt used in the assay represents only 3 times its K m value, this concentration of MnCk should be sufficient for use in subsequent experiments.  51  Over time, cuvettes containing all the reaction components except lgtC-19 and UDP-galactose, became cloudy. There appeared to be an interaction between the components of the assay mixture or between the mixture and the cuvettes which not only interfered with the detector but may also have altered the reaction mixture in an unknown and uncontrolled manner. To determine the source of this problem, a crude experiment was performed in which eight glass test tubes were assembled with the suspect reaction components as described in Table 3.  Table 3: Reaction components for precipitation test. (+) component present; (-) component absent  Test Tube (TT):  # 1 # 2 # 3 # 4 # 5 _ # 6 # 7 # 8  13mM HEPES, pH7.5  +  +  +  +  +  +  +  +  0.1% BSA  + .  +  +  +  +  +  +  +  50mMKCl  .  +  .  +  +  -  +  15mMMnCl  -  -  +  +  +  5mM DTT  .  .  +  +  2  -  .  + +  -  +  Each component was added at a concentration typical of the assay mixture with TT#1 serving as a negative control and each of the remaining test-tubes containing various combinations of reaction components. The tubes were allowed to stand at room temperature and observed at timed intervals for any change in clarity. The appearance of a white precipitate (Table 4, #7 and #8), indicated that a reaction was occurring between the MnCh and DTT. Unfortunately, both components are required by lgtC-19 for  52 activity: Mn as described previously, and DTT as a reducing agent, which prevents 2+  disulphide bondsfromforming betweenfreecysteines in lgtC-19. Also, the presence of potassium chloride (KC1), required by PK in the coupling reaction, appeared to slow the formation of precipitate. To reduce this interference in subsequent experiments, DTT and MnCk were only combined in the reaction mixture within 2 hours of reaction initiation.  Table 4: Results of clarity test. (-) denotes a colourless, transparent reaction mixture and (+) denotes the appearance of cloudiness.  Reaction  1 hour  2 hours  3 hours  4 hours  -  -  5 hours  1 2  -  -  4  -  -  5  -  -  6  -  -  7  -  +  +  +  +  8  -  -  +  +  +*  3  * a white precipitate was observed  Next, the effect of lgtC-19 preincubation with DTT was examined. Two parallel sets of experiments were performed under standard assay conditions in the presence of 1 mM UDP-galactose and 125 mM lactose at pH 7.5: one in which the preincubation mixture contained all components except DTT and UDP-galactose until initiation, and another in which the preincubation mixture contained DTT. In this way, the effect of  53 dilution and of preincubation of the enzyme with other reaction components could be controlled. The data in Figure 2-4 suggests that preincubation of lgtC-19 in DTT is required for optimal activity and that changes in activity are small after 1 hour of preincubation under the conditions used.  1  60  1  1  T  1  i—"—i—i—r  h  O )  E o E  Z3  20 h  o o o J  0 0  20  i  I  o  40  i  I 60  i  I  i  80  I 100  i  o  L  120  incubation time (min) Figure 2-4: Rate observed as a function ofpreincubation time using standard assay conditions at pH 7.5, 30 °C (O) DTT added at time of reaction initiation (•) DTT added to preincubation mixture To test the possibility of storing the lgtC-19 in 5 mM DTT, preventing the need for a preincubation period, the effect of this method was compared to preincubation of the enzyme in the assay mixture. The following three reactions were assembled: (1) All assay components except DTT and lgtC-19; (2) All assay components except lgtC-19; and (3) All reaction components except UDP-galactose, including 5 mM DTT and lgtC-19. After 1.5 hours, reactions (1) and (2) were initiated using lgtC-19 that had been preincubated with 5 mM DTT and reaction (3) was initiated by adding UDP-galactose to a final concentration of 1 mM. The initial rates for each reaction were compared.  54 Maximal v was observed for the reaction in which lgtC-19 had been stored in 5 mM 0  DTT and which contained 5 mM DTT in the assay mixture (2). The relative rates of (1) and (3) were 88% and 67% respectively. Thus, storage of lgtC-19 in DTT should eliminate the need for preincubation of the enzyme in the assay mixture prior to assay as long as DTT is retained as a component of the assay mixture. 48  cn  E  I  1  I  1  I  1  I  1  I  I  1  I  1  1  46 h  o E 3  44 h  i  42  I  4 6 8 10 12 preincubation time (min)  i  I  14  16  Figure 2-5: Effect of lgtC-19 preincubation time in the presence of 15 mMMnCh on activity measured under standard assay conditions at pH 7.5, 30 f .  With the effect of DTT preincubation on lgtC-19 activity determined, the effect of preincubation with Mn was assessed. Using standard assay conditions, with 0.5 mM 2+  UDP-galactose and 125 mM lactose at pH 7.5, v was measured as a function of 0  preincubation time in an assay mixture containing 15 mM MnCk. The enzyme, stored in DTT to promote optimal activity, was added to the assay mixture and allowed to incubate for increasing time periods prior to initiation of the enzymic reaction with UDPgalactose. The results shown in Figure 2-5 indicate that a minimal preincubation time of  55 3 minutes is required for optimal enzyme activity. Enzyme activity appears relatively constant after the initial 3 minute lag period. 2.3.3 Metal Cofactor Specificity  Many glycosyl transferases are activated by millimolar concentrations of bivalent metal ions such as Mn . Although previous attempts at replacing Mn with Mg had 2+  2+  2+  proven unsuccessful in lgtC-19-catalysed reactions (personal communication from W. Wakarchuk), other metal species had not been evaluated for use with this enzyme. The relative velocity of lgtC-19 in the presence of various divalent metals (5 mM) was assessed at fixed concentrations of UDPgal (0.5 mM) and lactose (125 mM) using the standard assay at pH 7.5. The lgtC-19 was prepared in buffer containing 2 mM EDTA to prevent residual metal ions present in the enzyme suspensionfrominterfering with the experimental results (Table 5).  Table 5: Relative activation oflgtC-19 by common divalent metal cations  Metal Salt  Relative Activity  MnCl  100%  MgCl  53%  CaCl  undetectable  ZnCl  a  2  2  2  2  addition of ZnCl to the assay mixture caused a precipitate to form and decreased the pH of the mixture below pH 5.5 a  2  56  Figure 2-6: Rate of the lgtC-19-catalysed galactosyl transferfrom UDP-galactose to lactose under standard assay conditions at pH 7.5, 30 <C using selected metal salts as activators. (•) CaCh (•) MgCh (O) MnCh (complete curve obtained in a previous experiment (2.3.2); see Appendix B.I for results)  Next, the kinetic parameters for the metal activators were sought under these same reaction conditions. The effect of metal ion concentration on v is seen in Figure 0  2-6. A Lineweaver-Burk plot of l/v dependence on l/[MgCl2] indicates that a change in 0  the relative value of kinetic parameters is occurring as the concentration of magnesium increases (Figure 2-7, left). This can be compared with a similar plot for varying manganese concentration (Figure 2-7, right). Both of these double-reciprocal plots exhibit a break between two linear regions.  57  0  2  4  6  1/[Mg*] (mM) 2  0  2  4  6  l/JMn *] (mM)  1  2  1  Figure 2-7: Lineweaver-Burk plots for the rate dependence of the IgtC-l9-catalysed galactosyl transfer reaction on metal concentration for Mg (left) and Mn (right). 2+  2+  Navaratnam et al. (1988) have reported that manganese ion concentration has a similar effect on the activity of galactosyl transferasefromlactating rat golgi membranes. They interpreted the results as proof that two metal binding sites are present, one with a high specificity for Mn and the other with less rigid specificity for metal ion. They 2+  proposed that maximal activity is only possible when both sites are occupied. By extrapolationfromthe double-reciprocal plot, the apparent K m for Mn at those sites was 2+  determined. If this methodology is applied to the results obtainedfromthe current study of Mg activation, two binding sites with 2+  values of 0.37 mM and 11 mM exist on  lgtC-19. Furthermore, this suggests that binding to the first site provides a Vma with X  46% of the fully activated activity seen when both sites are occupied. Similarly, with Mn activation, the two binding sites have K values of 0.027 mM and 5.48 mM with 2+  m  just 24% of the fully activated activity present when only the first site is occupied (Table 6).  58 Table 6: Observed kinetic parameters for the metal ion binding sites of lgtC-19 (determined by the method of Navaratnam et al, 1988)  Metal ion (concentration)  K  m  (mM)  kcat (min )  Specificity constant  316±9 27±2 290+50 (i.23+6.bo7)xid  (mM^min ) 860+70 63±7 (l.l+0.4)xl0 (2.2+0.03)xl0  1  1  0.37±0.01 11±1 0.027+0.005 5.48±0.03  Mg (<10 mM) 2+  Mg "(>i6mM) Mn (<0.5 mM) Mri (>5 mM) 2+  2+  2+  3  4  2  If two metal binding sites are available on lgtC-19, it appears that thefirstbinding site has a higher specificity for Mn . For thefirstbinding site, the specificity constant 2+  (kcat/K ) m  for Mn was 13 times that of Mg . Only a 3 fold difference in specificity was 2+  2+  seen for Mn over Mg at the second binding site. 2+  2+  From this experiment alone it is impossible to draw conclusions regarding the actual nature of the interaction between lgtC-19 and metal ion activators. It is not clear why the presence of two binding sites might be implicated in this mechanism. The apparent change in the rate-determining step of this glycosyl transfer reaction as a function of metal ion concentration could impact the kinetic information gained in a number of studies in which the concentration of metal ion was held constant. Indeed, the function of divalent metal ions in the mechanism of glycosyl transferases remains a hotly debated topic (see section 1.2.6). Certainly, further investigation into these binding effects is required.  59 2.4 Kinetic Mechanism 2.4.1 Cleland Kinetics at pH 7.5 Having optimised the components required for the lgtC-19-catalysed transfer reaction, a more rigorous examination of the catalytic mechanism could be made. Using the method described by Cleland (1970), a kinetic experiment was devised in which the concentrations of UDP-galactose and lactose were varied independently. The method used for this grid experiment is outlined in 5.3.6. The resultant family of doublereciprocal plots (Figure 2-8) appears to be a series of parallel lines. This pattern of lines  Figure 2-8: Double reciprocal plots for lgtC-19 activity as a function of UDP-ga concentration over a range of lactose concentrations using standard assa conditions, (o) 145 mM lactose (•) 40 mM lactose (•) 20 mM lactose (•) lactose (A) 7.5 mM lactose ( A ) 5 mM lactose  60 is generally indicative of a ping pong mechanism however, it is impossible to determine the mechanism of the reactionfromthis data alone. There are cases in which both ordered bi bi sequential and random bi bi sequential reactions will produce apparently parallel lines in this type of plot (see Appendix A.II.iii for a more thorough discussion of these exceptions). 200 h 160  5  c | o E  E  1  120 V-  3  O  O  O  O  O  O  Q  O  O  O  1/[lactose] mM-  Figure 2-9: Replots of (A)  O  J  j  O  S  oo_^-^-k-»--kijro O > 0 0 N J * . O > 0 0  1/pactose] m M  1  1/V , pp  From the replot graphs of 1/V  —  max  a  and (B)  l/K uDP (a ) m>  PP  r  J  N >  1  versus 1/[lactose]  ( ) and l / K m ^ versus 1/[lactose], the true kinetic 1  max  app  constants were determined. Using this model,  = 52 pM, Klf  056  = 47 mM, and  k =3.4xl0 min" . 3  1  cat  The kinetic parameters for the lgtC-19-catalysed reaction, which have been determined from the studies using UDP-galactose as the donor substrate and lactose as the acceptor substrate, are summarised in Table 7. The values of Km and kcat determined using Cleland's method were calculated based on the ping pong model (described in 2.4.1).  61 Table 7: Kinetic constants for substrates utilised by lgtC-19  SUBSTRATE UDP-galactose  Km  kcat  29.8ll.7uM  (min"mM") 1  1  7.2 x 10  b  3.4 x 10  6.6 x 10  47mM  3.4 x 10  72  4  3  b  2+t  kcat/Km 3  4  3  54  41±7mM  (2.210.2) x 10  4.9 ± 0.5mM  (2.010.05) x 10  c  Mn  1  (3.3±0.05) x 10  3  52uM Lactose  (min")  3  410  3  All reactions were performed at 30°C at pH 7.5 in the presence of 5 mM DTT. apparent value: 125 mM lactose, 15 mM MnCl in 13 mM HEPES buffer. from replot analysis using Cleland's method; in 15 mM MnCl and 13 mM HEPES buffer. apparent value: 1 mM UDP-galactose, 15 mM MnCl in 13 mM MnCl . 125 mM lactose, 1 mM UDPgal, and 13 mM HEPES. 8  b  2  0  2  2  f  2  2.4.2 Inhibition Using a Donor Substrate Analogue  The identification of an enzyme-substrate intermediate is essential to prove that a ping pong mechanism is involved in the lgtC-19-catalysed transfer of galactose from UDP-galactose to lactose (see Chapter 1). In the absence of such information, other methods are needed to determine the catalytic mechanism of lgtC-19-catalysed galactosyl transfer. Typically, product inhibition studies can be useful for determining the order in which substrates bind and products leave. By deterrnining whether each substrate is competitive, non-competitive, or un-competitive with each product, the precise catalytic mechanism can be determined (Appendix A). Unfortunately, the PK/LDH coupling system utilises one of the reaction products, UDP. Thus, UDP can not be tested as an inhibitor. Only limited amounts of the trisaccharide product, Gala(l-»4)Gala(l-»4)Glc, are available, making inhibition experiments that require this product unpractical.  62 To understand the perils of assigning a kinetic mechanism to lgtC-19 using the information that has been gathered so far, one must examine the rate equations for each possible mechanistic pathway that might be used by this enzyme for the synthesis of trisaccharide. Random bi bi sequential, ordered bi bi sequential, and ping pong reactions are described by equations [1], [2], and [3] respectively (derivations in Appendix A). v  V^LAXHB] maxl/"Ml"J  _  ri-l  T  oK^tBJ ^K AX rDl + i „ofK AX17B  8  -rr  +oK , _r?Br[AX] + [ A•« X »-i ] i -[ r »B -i ] B  •-  A  L J  Vn.JAXjrB] K^K>K [AX] + K f [B] + [AX][B]  v =  m  V^tAXJtB] K*[AX] + K f [B] + [AX][B] Note that equations [2] and [3] differ by only one term, K K„. If the dissociation constant for AX, K^, is much less than  , this term will approach zero. Thus, the  velocity equations for an ordered bi bi system and a ping pong system would be indistinguishable under these conditions.  To test whether the results obtained are  consistent with a low, but non-zero value for  in equation [2], the dependence of rate  on UDP-galactose and lactose concentrations was examined in the presence of a fixed concentration of UDP-glucose (Ki«200 uM). There is precedence for UDP-glucose to act as a competitive inhibitor with respect to UDP-galactose in reactions catalysed by both human and bovine galactosyl transferases (Tsopanakis & Herries, 1978; Khatra et al., 1974). The addition of a competitive inhibitor serves to amplify  and K^* (in  [2]) by a factor of (1+[I]/Ki) and magnify any tendency of the lines in the reciprocal plot  V^JAXJfB]  to intersect: v = — — , £ [4] K **K (1 + [FJ/K,) + K [AX] + K f (1 + [Tj/K, )[B]+[AX] [B] m  m  63 From Figure 2-10 it can be seen that in the presence of UDP-glucose, the lines quite clearly intersect. Similar experimental results with bovine galactosyl transferase have been interpreted as evidence for an ordered bi bi sequential reaction mechanism (Tsopanakis & Herries, 1978; Khatra et al., 1974; Powell & Brew, 1974). Clearly, the  2.4 2 |  1.6  | o  1.2  1  0.8  CO  .>°  0.4 0 -0.4 -60 -40 -20  0 20 40 60 80 100 1/[UDPgal] (mM) Figure 2-10: Double reciprocal plots oflgtC-19 activity as afunction of UDP-galactos over a range of lactose concentrations in the presence of 50 mM UDP-glucose (measured using standard assay conditions at pH 7.0). (o) 125 mM lactose, (•) mM lactose, (•) 20 mM lactose, (•) 10 mM lactose, (A) 5 mM lactose, (A) 2.5 lactose 1  data are consistent with this previously described behaviour, indicating that the mechanism of lgtC-19 is described by equation [2] rather than [3] and is, therefore, an ordered bi bi sequential mechanism in which K « K^f. Thus, the reaction mechanism AX  may be described by: B ( k  k, E  +  AX  " k-i  2  E'AX  (E»AX»B k.2  kp k. p  BX k > -r^  k  3  E»A»BX)  k_3  4  E *A  ^ k_4  E  +  A  64 where E isfreeenzyme, AX is UDP-galactose, A is UDP, B is lactose, BX is the trisaccharide product, and K** = k i / k - i  «  =  k  2  k  3  k  4  k,k (k +k ) 2  4  3  Based on the results of this experiment, the values of K m and kcat reported in Table 7 were recalculated using the bi bi sequential model by plotting 1/[lactose] as a function of the slope of the double reciprocal plots of [UDP-galactose]/vbs vs. 1/[UDP0  galactose], and as a function of the intercepts of these same plots (as described in Appendix A). The kinetic constants determined in this manner were very similar to those determined using the previously described method: Km,i pM, and  kcat  ac  tose  =  52 mM,  Km,uDPgai  =  50  = 2.8 x 10 min", representing differences of 10%, 4%, and 18% 3  1  respectively. The replot graphs are shown in Appendix B (section II). The error in V  ma  x  determinedfromthis method is larger because of the scatter seen in the intercept replots. The poor fit of a regression line to this data also prevented determination of K**. As a result, the kinetic values have been derived using the replot graphs described for a ping pong mechanism since this approach enables the kinetic constants to be determined from just one replot graph and there is no significant difference in the results obtained using the two analytical methods.  2.5 pH Study Kinetic parameters are pH dependent. Thus, it is necessary to determine the value of these parameters as a function of pH. The methodology for this experiment is described in 5.3.7. Before this study was undertaken, the pH stability of lgtC-19 was assessed to ensure that loss of enzyme activity as a result of unfavourable conditions  65 would not affect the outcome of kinetic measurements. Using a standard assay at pH 7.5, v was measured as a function of enzyme incubation time in a series of buffersfrompH 0  3.0-9.0 and the data were plotted to indicate any trends.  time (min) Figure 2-11: Enzyme activity as a function of time when stored at a range ofpH values As shown in Figure 2-11, lgtC-19 appeared to undergo an initial period of activation at pH values from 6.0 to 8.0. The extent of this activation, and the speed with which it occurred, increased as the buffered enzyme storage solution became more alkaline. The effect disappeared at pH 9.0, at which time the initial activation may have been so rapid that the activity had begun to decline by the time of the first assay. The nature of this effect is undetermined. Activity enhancements at alkaline pH have been reported for other Mn -dependent enzymes and are believed to play a role in the cellular 2+  regulation of activity (Sherry et al, 1975; Kuhn et al., 1991b). Time-dependent loss of activity was fit to afirst-orderexponential decay (GraFit 3.0; Leatherbarrow, 1990) and  66 the decay constants  (ko ) b s  and initial activities (v„) were plotted as a function of pH  (Figure 2-12).  Figure 2-12: Effect of storage pH on initial activities (v ) and inactivation rates (k bj. 0  0  While the maximal activity of lgtC-19 is greatest when stored at higher pH, the rate of activity loss increases at pH<5 and pH >8. To maximise activity and minimise the timedependent inactivation of lgtC-19 over the course of an experiment, pH 7.0 buffer was chosen as the preferred storage solution for lgtC-19. Finally, a series of grid experiments were performed. Both the UDP-galactose and lactose concentrations were varied independently to obtain kinetic parameters at pH 6.0, 6.5, 7.0, 7.5, and 7.7.  The linear dependence of reaction rate on lgtC-19  concentration over this pH range ensured the adequacy of this coupling system for the pH study to be conducted (Figure 2-13). Using the method described by Cleland (1970), the double-reciprocal plots obtained at each pH were observed for their pattern of intersection. Each of these revealed apparently parallel lines (see Appendix B, sections  67 III-VIII)  as seen previously (Figure 2-8).  Closely positioned lines, representing  measurements at high concentrations of lactose, showed an increasing tendency to intersect as pH increased.  o o o  o o o  o o o o co  o b o  o b o a cn  2  o b o o a>  o o o  lgtC-19 (mg/ml)  Figure 2-13: Effect of lgtC-19 concentration on v b (measured as a change in 0  s  absorbance at 340 nm). (o) pH 7.7 HEPES, (•)pH 7.5 HEPES, (•)pH 7.0 HEPES, (•) pH 7.0 MES, (A)pH 6.5 MES, ( A ) p H 6.0 MES  This effect is slight and may simply resultfromerrors inherent in the measurement of v bs0  The kinetic constants were determined from the replot graphs and are summarised  in Table 8. The pH curves for UDP-galactose and lactose were constructedfromthe corrected data and fit using the pH algorithm in GraFit 3.0 (Leatherbarrow, 1990; Figure 2-14). While the data appear tofitclosely to the curve, there is only a 2-3 fold variation in  Vmax/Kn,  over the range of pH values studied. As a result, caution must be observed in  analysing these results.  68  Table 8: Kinetic constants determinedfrom replots of 1/K ( ) and l/V (ap ) function of 1/[lactose] m  PH  K m (lactose)  K m (UDPgal)  kcat  (mM)  (pM)  (min-)  MES 6.0  41.8  85.0  2.2  MES 6.5  28.6  26.0  MES 7.0  26.4  HEPES 7.0  app  max  P  as a  kcat/Km(lactose)  kcat/Km(UDPgal)  (min"mM)  (min^mM")  53  2.6 x 10  1.8 x 10  62  6.8 x 10  23.6  1.9 x 10  72  8.0 x 10  50.0  28.8  2.7 x 10  72*  8.0 x 10**  HEPES 7.5  41.0  44.3  2.9  91*  5.4 x 10**  HEPES 7.7  37.2  52.4  2.7 x 10  96*  4.4 x 10**  1  xlO  1  3  3  3  3  xlO  3  3  1  1  4  4  4  4  4  4  Reactions run in HEPES buffer have been normalised to account for buffer effects (*,**). *observed valuex(71.5/54.9) and """observed valuex(8.00/9.54)  The bell-shaped curve obtained for UDP-galactose (Figure 2-14, A) indicates the presence of two ionisable groups involved in the catalysis and/or release of this substrate, one with a p K a of 6.7 and the other with a p K a of 7.2. For lactose (Figure 2-14, B), only the second ionisable group appears to affect binding and catalysis. Cysteine, histidine, and lysine are amino acids which have functional side-chains that could be protonated or deprotonated, within the confines of the enzyme active site, over this pH range. Previous studies of glycosidases have shown that carboxylate moieties such as those of aspartate or glutamate can also ionise over this pH range when placed within the enzyme active site (Mcintosh et al., 1996; Kempton & Withers, 1992; Legler, 1990).  69  Figure 2-14: Effect ofpHon V JK for (A) UDP-galactose and (B) lactose in the IgtC19-catalysed transfer of galactose using standard assay conditions from pH6.0m  m  7.7.  Based on the information obtainedfromthe study of glycosidases, it is likely that the first ionisable group is the catalytic nucleophile. When this group deprotonates at pH 6.7, it can attack UDP-galactose to form the galactosyl-enzyme intermediate. Possibly the group which loses a proton at pH 7.2 acts as an acid catalyst, donating this proton to UDP to facilitate bond cleavage. However, it is also possible that this protonation could allow hydrogen bond formation between the ionised species and the ring oxygens of lactose.  2.6 Probing the Catalytic Domain of LgtC-19 Very little is known about the active site of lgtC-19. Without the benefit of a crystal structure, our ability to identify significant residues involved in substrate binding or catalysis is limited to more classical methods. The use of reactive molecules as affinity labels for reactive functionalities present within the catalytic domain is well documented (Gebler et al., 1995; Yadav & Brew, 1990; Withers & Aebersold, 1995;  70 Bazaes, 1987; Powell & Brew, 1976; Lee et al., 1983; Holmes, 1990; Ats et al., 1992; Hatanaka et al., 1996). Affinity labels are structurally similar to the substrate or ligand whose binding site they are meant to probe. The goal of this type of labelling is the quantitative covalent modification of a functional group belonging to a unique amino acid residue without affecting other functional groups or the conformation of the enzyme molecule. Affinity labelling involves site-specific binding followed by modification of an amino acid through the formation of a covalent bond. Thus, affinity labelling must show saturation kinetics (Bazaes, 1987). Identification of the residue implicated in such a reaction requires isolation of the peptide containing the chemically-modified site, followed by sequence analysis to identify the particular amino acid involved. 2.6.1 Inactivation and Labelling Studies  A lysine residue common to the UDP binding site of two glycosyl transferases, bovine colostrum galactosyl transferase and an ccl-»3fucosyl transferasefromhuman small cell carcinoma NC1-H69 cells, has been identified using affinity labelling (Powell & Brew, 1976; Yadav & Brew, 1990; Holmes, 1992). A glycosyl nucleotide diphosphate analogue, in which the nucleotide moiety had been modified by treatment with periodate (IOY) to form a dialdehyde, was used as an affinity label for the UDP binding site (Figure 2-15). In the study upon which the current experiment is based (Powell & Brew, 1976), a Schiff base formed between one of the aldehyde functionalities and the e-NH2 group of a lysine located within the UDP binding site. Although the enzyme-inhibitor complex (Figure 2-16, A) was stable to extensive dialysis, inactivation was rendered fully irreversible by borohydride reduction of the enzyme following treatment with dial-UDP.  71  HN  HN  X  0  N  X  I0 -  II H C—O—P—O—P—OH  4  2  K OH  N  o  o  II  II  H,C-0—P-O—P-OH  o-  0-  J  O-  CH HC II II O O  OH  UDP  dial-UDP  Figure 2-15: Synthetic route to dial-UDP  V  R—CH  + 0H  OH  I  +  HN 2  R.  R—CH  I  R,  NH  H  R  H  +  R—CH R,  NH  H  V  2  II  ]  x  +  (A)  _ l NaBH4  R  H  ?  — I (B)  Figure 2-16: Mechanism of dial-UDP inactivation. R=dial-UDP, R\=lysine residue, (A) Schiffbase, (B) inactivation complex  Incubation of lgtC-19 with several different concentrations of dial-UDP, in the presence of MnCk and DTT, resulted in concentration-dependent inactivation over time. At each concentration of dial-UDP, the data were fit to afirst-orderexponential decay (Figure 2-17). The rate of inactivation was slowed by the addition of UDP-galactose to the reaction mixture, suggesting that both substrate and inhibitor compete for the same binding site (INSET, Figure 2-17).  72  0  200  400  600 800 time (min)  1000 1200  Figure 2-17: LgtC-19 activity as a function of incubation time in dial-UDP (measured using standard conditions at pFL 7.5). (o) 0.5 mM dial-UDP, (A) 0.7 mM dialUDP, (•) 1.0 mM dial-UDP, (•) 10 mM dial-UDP, (•) 20 mM dial-UDP; INSET: 10 mM dial-UDP (•) no UDP-galactose (A) 0.3 mM UDP-galactose  Plots of kobs as a function of inhibitor concentration failed to provide reliable values of the inactivation rate constant, ki, and the equilibrium constant for inhibitor binding, Ki (Figure 2-18 A, B; these parameters are defined in Appendix A). Large errors in these measurements as a result of data scatter preclude the assignment of any experimentally derived values to these parameters. Additionally, mass spectra of the lgtC-19 after treatment with the dialdehyde did not indicate any labelling of protein either with or without sodium borohydride reduction. Oxidised nucleotide diphosphates have been shown to be unstable, decomposing intofreephosphate and polymeric material (Bazaes, 1987). Although the breakdown of the dial-UDP can be minimised by low pH and low temperature, decomposition may have occurred under the conditions necessary  73  1/[dial-UDP] mM  [dial-UDP]  1  Figure 2-18: Inactivation rate oflgtC-19 as a function of dial-UDP concentration. (A) k vs. [dial-UDP] (Ki= 5 ±3 mMandki = 2.0 ±0.4 min' ), (B) LineweaverBurkplot (l/Kbs-45 ±31, k /[dial-UDP]=199 ±26 mMmin' ) 1  obs  1  ob  for lgtC-19 activity. In retrospect, the addition of a-galactosyl fluoride or another galactose analogue to the inhibition mixtures may have promoted binding of the dialUDP to this enzyme.  CHAPTER 3  a-GALACTOSYL FLUORIDE KINETICS AND CHARACTERISATION  75  3.1 Introduction Enzymic synthesis of oligosaccharides provides regio- and stereoselectivity and does not require the multiple protection and deprotection steps that complicate chemical synthetic methods. However, one drawback to the use of glycosyl transferases for the industrial-scale synthesis of oligosaccharides is that the corresponding nucleotide sugars are often difficult to synthesise or isolate and are, therefore, relatively expensive (Wong et al., 1992b; Hokke et al., 1996). The development of synthetic methodology which utilises less expensive donor substrates will improve the utility of transferases for largescale manufacture of oligosaccharides. Glycosyl hydrolases have exhibited significant flexibility with regard to both donor and acceptor substrates (Kempton & Withers, 1992; MacLeod et al., 1996; Day & Withers, 1986). Glycosyl fluorides, among other synthetic glycosides, have proved to be effective glycosyl donors for glycosidases (Okada et al., 1979; Liu et al., 1991) and are easily prepared at low cost (see 5.2.1 for synthesis). These qualities make simple glycosyl fluorides ideal substrates for large-scale syntheses. Because a strong similarity exists between glycosidases and glycosyl transferases (Chapter 1), the use of these synthetic substrates may also be possible for reactions catalysed by transferases. Glycosyl transferases have exhibited the ability to utilise both synthetic donor and acceptor substrates (Paquet et al., 1984; Berliner et al., 1984; Wong et al., 1992b; Srivastava et al., 1990; Baisch et al., 1996a,b; Blanken & Van den Eijnden, 1985; Flitsch et al., 1991). To date, however, these enzymes have not been shown to utilise monosaccharide donors that lack the nucleotide phosphate functionality.  76 3.2 Preliminary Studies: Assessment of a-G alactosyl Fluoride Suitability as a Galactosyl Donor Based on the successful use of glycosyl fluorides as synthetic glycosyl donors for glycosidases, cc-galactosyl fluoride was investigated as a potential donor substrate for the galactosyl transfer reaction catalysed by lgtC-19. Because lgtC-19 operates via a retaining mechanism, it should form a galactosyl-enzyme intermediate during the course of the transfer reaction (Koshland, 1953). Given that successful donor analogues have incorporated the nucleotide portion of the substrate, the potential requirement for binding energy provided by interaction of the UDP moiety of UDP-galactose with the enzyme binding site was recognised. Studies of various recombinant enzymes, engineered to remove catalytically-important residues, have shown that the activity of these crippled enzymes can be substantially restored in the presence of rescue substrates. These are substrates which incorporate the missing functionality, providing the interactions necessary for enzyme activity (Kim et al., 1992; Dhalla et al., 1994; Carlow et al, 1995; Zor et al., 1997). By analogy, the addition of UDP, along with a-galactosyl fluoride, might provide the binding energy required for lgtC-19 activity. In this situation, the anomeric fluoride should serve as an excellent leaving group, encouraging the formation of a glycosyl-enzyme intermediate without preventing UDP binding to the enzyme active site. In this situation, UDP would serve as an activator and would not be consumed over the course of the reaction, minimising the cost of this synthetic approach. To determine whether lgtC-19 would accept a-galactosyl fluoride as a galactosyl donor, the following experiment was performed. Four reactions were assembled, each containing 100 mM HEPES (pH 7.5), 15 mM MnCl , 5 mM DTT, 5 mM FITC-lactose 2  77 (acceptor), and 2.5 U lgtC-19. The glycosyl donors added to each reaction are listed in 1  Table 9. From the TLC analysis of each reaction mixture, made 1 hour after initiation of the reaction with lgtC-19 (Table 9), it appears that UDP is required by lgtC-19 for the utilisation of a-galactosyl fluoride as a glycosyl donor. Table 9: TLC resultsfromthe lgtC-19-catalysed transfer of galactose to FITC-lactose (1 hour after initiation; run in 7:2:1:0.1 ethyl acetate/methanol/H20/acetic acid).  Reaction Donor substrate  Rf of transfer product  1  a-galactosyl fluoride (29mM)  -  2  a-galactosylfluoride(29mM) + UDP(17mM)  0.29  3  Galactose (29 mM) + UDP(17mM)  4  UDP-galactose (0.3 mM)  a  -  a  a  0.29  No reaction occurred; the TLC remains unchanged after 1 hour  To ensure that the product formed from a-galactosyl fluoride and UDP was the same as that formed using the natural donor sugar, UDP-galactose, the products recoveredfromboth donor systems were compared. One reaction series utilised the natural donor and the other, a-galactosyl fluoride and UDP. Negative controls of each were obtainedfromparallel reactions run in the absence of lgtC-19. The transfer reaction was monitored by TLC using three different fluorescent-labelled acceptor sugars:  1 U of enzyme activity represents that amount of enzyme required to catalyse the transfer of 1 pmol/min of galactosefromUDP-galactose to FCHASE-lactose under standard reaction conditions at pH 7.5.  78 FCHASE-lactose, FCHASE-galactose, and FITC-lactose. Each reaction mixture was analysed by mass spectrometry when it reached > 80% completion to identify the products formed. Table 10 shows that the product formed by lgtC-19, using either donor sugar, appears to be the same for each of the three acceptors tested. Additionally, the mass expected for the product of galactosyl transfer was observed in each case (see Appendix B.X-XTI for the mass spectra). Table 10: Molecular weights of fluorescent-labelled reaction products determined by electrospray mass spectroscopy. (+) control contains lgtC-19 and UDP-galactose as the donor sugar; (-) control contains either donor sugar but no lgtC-19; test reaction contains lgtC-19, with a-galactosyl fluoride and UDP as donor substrates.  ACCEPTOR  DONOR  (+)CONTROL  (-)CONTROL  m/z  m/z  TEST REACTION m/z  FCHASE-gal FCHASE-lac FITC-lac  UDPgal a-Gal-F + UDP UDPgal a-Gal-F + UDP UDPgal a-Gal-F + UDP  905.4 -  1067.6 -  985.2 -  small M" peak (an unidentified peak at m/z 477.2 is the largest)  743.2 743.2 905.4 905.4 823.2 823.2  -  905.4 -  1067.4  3  -  985.2  1  In addition, UTP and UMP were investigated as potential replacements for UDP in the galactosyl transfer reaction. Using FCHASE-lactose as an acceptor, UMP did not promote the formation of the galactosyl transfer product from a-galactosyl fluoride. A very slight amount of product was observed with the reaction mixture containing UTP but it was impossible to determine the true source of activation, since contamination with UDP is unavoidable in most commercially available preparations of UTP. In this situation, the UTP was certified to be 99% pure.  79 3.3 Prep-scale Synthesis To unambiguously identify the product formed from the lgtC-19-catalysed transfer of galactose from a-galactosyl fluoride to acceptor, NMR spectra are required. Thus, milligram quantities of product are needed. The fluorescein-derivatised acceptors are costly, time-consuming to prepare, and have limited solubility. Because the nonreducing terminus of the natural acceptor comprises a lactose moiety, lactose itself would seem to be a logical alternative as an acceptor substrate. Indeed, as shown in Figure 3-1, lactose acts as an acceptor in this enzymatic synthesis and was used to obtain milligram quantities of trisaccharide product for NMR analysis. The 'H-NMR spectrum clearly indicates a trisaccharide, since the protons located at the anomeric position are easily assigned (see spectrum in Figure 3-2; assignments in Table 11). The chemical shifts of these protons compare favourably with those reported in the literature for the trisaccharide, aGal(l-*4)pGal(l-»4)Glc (Table 3-l;Urashima et al., 1997; Edebrink et al., 1996).  80 1 0.8 0.6 0.4 Negative Control  0.2 0  J  8  i  L  10 12 14  i  r  16 18 20  B  Test Reaction  8  10 12 14 time(min)  16 18 20  Figure 3-1: HPLC tracings for reaction mixtures containing 15 mMMnCh, 5 mMDTT, 62.5 mM lactose, 150 mM Tris buffer (pH 7.5) and (A) 150 mM a -galactosyl fluoride and 7.5 mM UDP; (B) contents of A plus 0.03 mg lgtC-19; (C) 60 mM UDP-galactose (65 hours after initiation).  81  aGal  BGlc BGal  aGlc  1 |  5.?  5.0  ~8 o r  _  J 4.9 PP-  6~!~o  7T0  4.6  4.4  5T0  4^0  FTo  2T0  TTo  ppm  Figure 3-2: 'H-NMR of the trisaccharide product of the lgtC-19-catalysed transfer reaction utilising a-galactosyl fluoride and lactose as substrates  Table 11: H-NMR analysis of trisaccharide 1  SOURCE  aGal (l->4)  PGal(l->4)  aGlc  pGlc  8'H (J, Hz)  8'H (J, Hz)  8 U (J, Hz)  8 H (J, Hz)  observed  4.95 (3.6)  4.52 (~8)  5.22 (3.6)  4.66(~8)  UrasWma et al., 1997  a  4.52 (7.7)  5.22 (3.7)  4.67 (8.1)  Edebrink et al., 1996  4.96 (4.3)  4.54 (7.6)  b  !  l  5.44 (4.0)  —  "in this reference, the structure was aGal(l-»3)BGal(l-»4)Glc; in this structure, the chemical shift of the anomeric proton of aGal(l-»3) was 5.14 ppm. in this reference, the structure for which these shift values were determined was aGal(l->4)pGal(l-»4)aGlc(l->2)->... b  82 Additional information was gainedfromthe NOE correlation between anomeric protons and those located on the linkage carbons (Table 12). The crowding of ring proton resonances between 3.6-4.0 ppm precludes a more detailed analysis. However, it is the character of the newly formed Gal-Gal linkage that is most important to determine, and this can be accomplishedfromthe information available.  Table 12: NOE correlations between anomeric protons and the proton on the linkage carbon  Source  aGal(l->4) pGal  pGal(l->4)Glc  Edebrink et al. (1996)  5 4.96-4.06  5 4.54-3.70  experimentally observed  5 4.95-4.0  8 4.52-3.6  3  (JGal(l->4)aGlc  3.4 Kinetic Characterisation of LgtC-19 u s i n g a-Galactosyl Fluoride as a Glycosyl Donor  3.4.1 Substrate Optimisation  The kinetic constants for a-galactosyl fluoride and UDP were determined using a fluoride electrode to monitor the release of F" in a continuous assay. Some difficulties were experienced when trying to obtain kinetic information for a-galactosyl fluoride in this manner. Most importantly, the reaction fails to obey Michaelis-Menten kinetics over the range of a-galactosyl fluoride concentrations used, as the enzyme does not become  83 saturated with substrate. In fact, as the concentration of a-galactosyl fluoride surpassed 200 mM, the observed rate of F" release became much greater than predicted. When v b 0  S  is plotted as a function of a-galactosyl fluoride concentration (Figure 3-3), a parabolic  [a -Galactosyl Fluoride] (mM)  [a-Galactosyl Fluoride] (mM) 2  2  Figure 3-3: Rate ofgalactosyl transferfroma-galactosyl fluoride to lactose(100 lgtC-19 measured using standard assay conditions with the fluoride ion e where (A) the datafitto a polynomial: v = 0.007[a-gal F] + 0.00006[a FJ + 6.0x10' and the slope of the line tangent to the initial curve gives k 6.0x10" mM min ; (B) a plot of v as a function of [a-gal-F] produces a straight line. 2  obs  1  obs  plot results. This behaviour suggests that a-galactosyl fluoride may itself be acting as an acceptor, competing with lactose at high concentrations. Thus, saturation is not detected and the observed rate escalates at an increasingly significant pace as the amount of agalactosyl fluoride in the reaction mixture increases. To determine whether the enzyme is capable of using galactosyl fluoride in the capacity of an acceptor, the following experiment was performed. Into three vials were placed 400 mM a-galactosyl fluoride, 3 mM UDP, 5 mM DTT, 15 mM MnCl , and 100 2  mM Tris buffer (pH 7.5). The first vial served as a negative control and 50 mM lactose was added. LgtC-19 (0.8 mg) was added to the second vial. The third vial served as a  84 positive control containing both lgtC-19 and lactose. If a-galactosyl fluoride were to act as an acceptor, the transfer product should be the disaccharide, Gakx(l-»4)Gal-F. After 22 hours, the reaction mixtures were analysed by TLC and evidence for the formation of disaccharidefroma-galactosyl fluoride was obtained (Table 13). Table 13: TLC results of lgtC-19-catalysed transfer reactions. Determined in 4:2:1:0.1 (ethyl acetate/methanol/water/acetic acid) 22 hours after initiation. (Vindicates the component is present; X indicates the component is absent)  Rf (reaction component)  (+)Control  (-)ControI  Test Reaction  0.57 (a-galactosyl fluoride)  •  •  •  0.48 (Gal-Gal-F)  X  X  •  0.45 (Galactose)  •  •  •  0.38 (Disaccharide)  •  •  •  0.32 (Trisaccharide)  •  X  •  Next, the reaction mixtures were subjected to HPLC analysis to isolate components. When the HPLC tracings of the negative control reaction and the test reaction were compared (Figure 3-4; black and blue lines, respectively), no peak could be assigned to the fluorinated disaccharide, Gal-Gal-F, although a small peak with a retention time of ~8 minutes appeared to indicate the presence of the disaccharide, Gal-Gal.  To determine whether a-Gal-Gal-F, the anticipated product of the  transglycosylation reaction, was co-eluting with Tris buffer, the test reaction was heated in a boiling water bath for 5 minutes to promote the hydrolysis of any glycosyl fluorides.  85  time (min) Figure 3-4: HPLC showing the formation of disaccharide by lgtC-19 using a-galactosyl fluoride as both acceptor and donor substrate (black = control containing all components of the test reaction except lgtC-19, and lactose; blue = test reaction with a-galactosyl fluoride but no lactose; red = test reaction after heating in boiling water bath for 5 minutes).  The HPLC tracing of the reaction mixture after heat treatment (Figure 3-4, red line) shows a substantially diminished a-galactosyl fluoride peak (TR « 4.8 minutes) and greatly enhanced galactose peak (T « 6.5 minutes) suggestive of a-galactosyl fluoride R  hydrolysis. In addition, the Tris buffer peak (T « 5.9 minutes) has been reduced with a R  concomitant increase in the disaccharide peak (TR « 8 minutes). To confirm the changes in the composition of the reaction upon heating, a new reaction mixture was assembled.  After 11 minutes, the reaction was stopped by  immersion in a boiling water bath. This promoted hydrolysis of any glycosyl fluorides.  86  4  0 0  2  4  6  10  8  12  14  16  18  20  time (min)  Figure 3-5: HPLC evidence for the formation of disaccharide by lgtC-19 using agalactosyl fluoride as both donor and acceptor substrate (blue = lactose standard; black = standard reaction mixture containing 400 mM a-galactosyl fluoride in the absence of lactose, 11 minutes after initiation;, red = same reaction mixture after heating in a boiling water bath for 5 minutes).  When HPLC tracingsfrombefore (Figure 3-5, black line) and after (Figure 3-5, red line) heat treatment are compared, a new disaccharide peak (TR«9.5min) appears, suggesting that Gala(l-»4)Gal-F is the initial disaccharide product formed. Attempts to obtain mass spectra of these reaction products were unsuccessful due to the leaching of column packing material into samples elutedfromthe HPLC column. Having determined that a-galactosyl fluoride was capable of acting as both a donor and acceptor in the galactosyl transfer catalysed by lgtC-19,  kcat/Km  was  determinedfromthe slope of the line tangent to the initial part of the curve described by v versus a-galactosyl fluoride concentration (dashed line in Figure 3-3, A). This is 0  and at low  possible since v = K  m  + [ S ]  [S],  keJEL,  v= V  [ S ] . Thus, the second-order J  87 rate constant ( k c a t / K ) can be determined since [E] is known. The kinetic constants for 0  m  a-galactosyl fluoride and UDP are shown in Table 14.  Table 14: Kinetic constants obtained using a-galactosylfluorideas the donor substrate for lgtC-19-catalysed galactosyl transfer to lactose  Substrate  K  m > a p  p  (mM)  kcat,app  (min") 1  kcat/Km  (min"mM") 1  a-galactosyl fluoride  220  c  1.3  6.0 x 10"  UDP  0.41  1.3  3.2  a  b  1  3  measured at pH 7.5 with 100 mM lactose and 15 mM MnCl , 30°C. "determined at 2 mM UDP. determined at 107 mM a-galactosylfluoride. calculated based on the k-j, determined for UDP. 2  b  c  The dependence of reaction rate upon UDP concentrations followed MichaelisMenten kinetics, allowing apparent values of both k c a and K m for UDP to be determined t  directly using Grafit 3.0 (Leatherbarrow, 1990; Table 14). An estimate of K m can be made by assuming that  kcat  =  1-3 min" (the apparent value observed under these 1  conditions when the concentration of a-galactosyl fluoride remained fixed at 107 mM). Using this value and the measured k c a t / K m , a K  m  ,a p P  of 220 mM was calculated for a-  galactosyl fluoride. Although this estimate of the K m ,  app  for a-galactosyl fluoride may be  slightly lower than the actual value of this kinetic constant, it should provide a reasonable estimate of the true value (within a factor of two). This determination can not be made directlyfromthe experimental data due to the large error associated with V  ma  x  in those  results. Additionally, it is important to notice that UDP need only be available to the reaction in small amounts and that its function is catalytic in nature. UDP is not consumed by the reaction.  88 3.4.2 Cofactor Optimisation  An attempt to determine optimal Mn concentrations for the lgtC-19-catalysed 2+  transfer reaction using a-galactosylfluoridein the presence of UDP was unsuccessful (see 5.3.14). The concentration of Mn present in reaction mixtures had a profound 2+  effect on the efficiency of thefluorideelectrode. Electrode accuracy diminished and response time increased as the concentration of MnCfe added to reaction mixtures increased (see Appendix B.IX). While the exact cause of this disturbance can not be explained, its effect made comparisons of reaction mixtures containing variable concentrations of MnCfe impossible.  3.5 "Mini" pH Study of the LgtC-19-cataly sed Galactosyl Transfer Reaction LgtC-19 has been reported to exhibit optimal activity at pH 7.5 (W. Wakarchuk, personal communication). To determine if this pH optimum holds for the transfer of galactose from a-galactosylfluoride,rather than from UDP-galactose, a series of assays were performed from pH 6.0 through 8.0. The observed rates offluoriderelease were plotted as a function of pH (Figure 3-6). The pH "curve" clearly shows that the measured rate of this reaction is greatest at pH 7.0. Only 24% of the activity observed at pH 7.0 can be seen at pH 7.5. A significant increase in activity was seen at pH 8.0 but the formation of a brown precipitate, likely a metal oxide, indicates that undetermined changes occurred in the reaction mixture. Because the effect of these changes upon the reaction mixture and the assay itself are unknown, it is not possible to draw conclusionsfromthe observed rate change.  89 0.001 0.0008 •at 0.0006 V  0.0004 0.0002 0 I I I I I I I I I I I I I I I I L-O -! I I 7  6  6.5  7  7.5  8  PH  Figure 3-6: pH dependence of v bsfor the lgtC-19-catalysed transfer of galactose from a-galactosyl fluoride to lactose where (O) denotes experimentally observed rates (the connecting line is to guide the eye only). NOTE: no activity was seen at pH 7.7 when duplicate measurements were made 0  Unfortunately, a satisfactory determination of the pK* values of residues involved in the catalytic mechanism could not be madefromthe data  3.6 Attempts to Identify a Galactosyl-enzyme Intermediate One method for indirectly observing the formation of an enzyme-substrate intermediate is to measure the size of any burst resultingfromthe rapid, presteady-state reaction of a-galactosyl fluoride with free lgtC-19 in which F is released as the galactosyl-enzyme intermediate is formed. This would be followed by a slower steadystate of F" release as galactose is transferred to acceptor. Consider the general reaction between enzyme and substrate: Fast  E + Gal-F  ->  Slow  E-Gal + F"  (+ acceptor) -» E + Gal-Acceptor  90 In a situation such as this, the plot of F" concentration as a function of reaction time will illustrate this phenomenon (Figure 3-7). 120 100 80 ~  60  LL  40 20 0 0  20  40  60  time (s)  Figure 3-7: Graphical representation of the enzymatic burst anticipatedfor a reactio mixture containing 60 pM enzyme  Observation of a burst requires that the first (glycosylation) step be faster than the second (glycosyl transfer) step. If this situation exists, a measurable burst should be seen if high concentrations of lgtC-19 are used. No such burst was seen (Figure 3-8) when 100 mM a-galactosyl fluoride was incubated with 47 pM lgtC-19 under standard fluoride ion assay conditions (as described in 5.3.2). Because the rate of the spontaneous hydrolysis of a-galactosyl fluoride was 70% of that observed upon addition of enzyme, the reaction was monitored until more than 157 pM of fluoride ion was present in the reaction mixture to ensure that any burst would be observed. In light of the probable ordered bi bi sequential mechanism of lgtC-19, lactose was added to the assay mixture to a final concentration of 83 pM, since it is possible that the binding energy supplied by the  91 acceptor is required for the formation of the galactosyl-enzyme intermediate. This failed to affect the rate of fluoride ion release. A slight upward curvature is observed which may resultfrompH changes since the experiment was performed above the pH optimum for lgtC-19. It is possible that a burst was not seen because the rate of glycosyl transfer is faster than that of glycosylation. Normally, adding donor substrate in the absence of acceptor might permit observation of a burst in this type of situation. However, previous experiments have shown (3.4.1) that a-galactosyl fluoride can itself act as an acceptor, pulling the reaction forward and preventing the accumulation of a galactosyl-enzyme intermediate. This hypothesis could be tested by using an incompetent acceptor such as  3.65 3.6 3.55 3.5 0.35 mM (dilution effect when lgtC-19 added) 3.45 E, •i• 3.4 LL 3.35 0 157 uM CD 3.3 d 3.25 i o j I I L 3.2 100 200 400 500 300 time(s) Figure 3-8: Fluoride ion concentration as a function of time for an assay mixture containing 47 JJM lgtC-19  4-deoxylactose which could potentially provide the required binding energy while failing to act as an acceptor in the reaction. The addition of low concentrations of a-galactosyl  92 fluoride to such a system might enable the covalent donor-enzyme intermediate to accumulate.  3.7 Enzymatic Synthesis of UDP-galactos e: Indirect evidence of a galactosyl-enzyme intermediate The requirement for UDP to be present in order for a-galactosyl fluoride to function as a glycosyl donor to a suitable acceptor meant that it was possible that, in the absence of added acceptor, this donor might transfer its sugar residue to UDP itself, forming UDP-galactose. Detection of this NDP-sugar would provide indirect proof of an enzyme-substrate intermediate. 3.7.1 A Continuous Coupled Assay for UDP-Galactose Detection  Since the amount of F" released does not provide evidence of UDP-galactose formation, some method for measuring the formation of this product was required. The PK/LDH coupling system could not be used since it measures UDP release. A new continuous coupled assay was developed to measure any UDP-galactose formed from a-galactosyl fluoride and UDP in the presence of lgtC-19. This system involved two coupling enzymes: UDP-galactose 4-epimerase and UDP-glucose dehydrogenase (see Figure 3-9). Consider the equilibrium expected for the formation of UDP-galactose by lgtC-19: a-galactosyl fluoride + UDP +E  <-»  UDP-galactose + E + F" (where E =  lgtC-19). As the natural substrate for lgtC-19, UDP-galactose binds tightly to the enzyme. Thus, strong product (UDP-galactose) inhibition would be expected to stop this  93 reaction from proceeding very far. The new assay system alleviates this problem by consuming UDP-galactose as it is released from lgtC-19. As this glycoside is converted to UDP-glucuronic acid, a change in absorbance is measured at 340 nm that corresponds  0  H  UDP  Figure 3-9: Mechanism for the UDP-galactose 4-epimerase/UDP-glucose dehydrogenase coupling reaction (UDPgal 4-epimerase/UDPglc dehydrogenase)  to the reaction rate, with 2 moles of NAD converted to NADH for every mole of UDP+  galactose that forms. In this way, the rate of UDP-galactose production can be measured. 3.7.2 Suitability  of the UDP-galactose  4-epimerase  /  UDP-glucose  dehydrogenase Coupling System  a-Galactosyl fluoride and UDP together serve as a UDP-galactose analogue for the galactosyl transfer reaction catalysed by lgtC-19. Thus, it is possible that the  94 coupling enzymes, UDP-galactose 4-epimerase and UDP-glucose dehydrogenase, might also utilise these substrates. To ensure this was not possible, a control reaction was assembled into which all assay components except lgtC-19 were added. This reaction mixture showed no appreciable absorbance change (340 nm) over a 10 minute period. Addition of UDP-galactose (0.04 mM) to the assay mix produced a sharp increase in absorbance  (AA340  «0.5) that was directly attributed to the added substrate. Thus, the  coupling system was functioning but did not appear to utilise a-galactosyl fluoride and UDP as substrates. 0.14 0.12 0.1  -  0.08 — 3 »  >°  0.06 0.04 0.02 0  I l l  0  I  1  I  1  0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 [lgtC-19] (mg/ml)  Figure 3-10: Linear correlation between AAA340 (Vobs) and lgtC-19 concentration for the UDP-galactose 4-epimerase / UDP-glucose dehydrogenase assay.  Next, the effect of varying concentrations of added lgtC-19 was determined (experimental details in 5.3.16). As shown in Figure 3-10, a linear correlation was seen between the rate of absorbance change at 340 nm and the amount of lgtC-19 added to the assay mixture. This indicated that UDP-galactose was produced by the lgtC-19-catalysed transfer of galactose from a-galactosyl fluoride to UDP and that the UDP-galactose 4-  95 epimerase / UDP-glucose dehydrogenase coupling system was able to accurately measure the rate of the reaction. 3.7.3 End-product Inhibition of the UDP-galactose 4-epimerase / UDPglucose dehydrogenase Coupling System  To ensure that the coupling enzymes remained active over the time period during which linear response to lgtC-19 concentration was assessed, sequential additions of UDP-galactose (4 ul; 2 mM) were made to the previously mentioned (3.7.2) control reaction. Each addition resulted in a rapid increase in absorbance at 340 nm followed by a plateau, indicating the reaction had reached completion. With each addition, the reaction rate decreased. The cause of the apparent decline in coupling system activity could be either time-dependent loss of activity or inhibition by the reaction end-product, UDP-glucuronic acid. To determine whether UDP-glucuronic acid inhibited the coupling system, the reaction rate was measured at constant UDP-galactose concentration (0.04 mM) in the presence of increasing concentrations of UDP-glucuronic acid (Figure 3-11). Manipulation of the data as described in section 5.3.16 (Figure 3-11; INSET) was used to estimate the "working" Ki for UDP-glucuronic acid with the UDP-galactose 4epimerase/UDP-glucose dehydrogenase system to be 270 ± 65 pM. In fact, UDPglucuronic acid has been identified as a competitive inhibitor of UDP-glucose in the reaction catalysed by UDP-glucose dehydrogenase, with a reported Ki of 200 pM (Campbell et al., 1997b).  96  0  200  400  600  800  [UDP-glucuronic acid] (uM)  Figure 3-11: Effect of UDP-glucuronic acid concentration on v bsfar the UDP-galactose 4-epimerase/UDP-glucose dehydrogenase coupling system in the presence of 0.04 mM UDP-galactose, (•) no UDP, (•) 0.5 mM UDP. (INSET: inverse curve used to determine K = 270 ±56 juM) 0  m  The effect of UDP on the inhibition of the coupling system by UDP-glucuronic acid was examined. As shown in Figure 3-11 (•), the addition of 0.5 mM UDP to an assay mixture containing 100 uM UDP-glucuronic acid resulted in a substantially greater decrease in v than when no UDP was present (•). This suggests that UDP also inhibits ODS  the turnover of UDP-galactose by this coupling system.  The influence of UDP  concentration on the coupling system could have implications for kinetic measurements, since UDP is consumed in the formation of UDP-galactosefroma-galactosyl fluoride and UDP. 3.7.4 Kinetic Parameters for LgtC-19  The adequacy of the UDP-galactose 4-epimerase / UDP-glucose dehydrogenase coupling system had been determined for assay mixtures containing 0.5 mM UDP. Knowing that UDP inhibits this coupling system, subsequent assays were performed with  97 assay mixtures containing no more than 0.5 mM UDP. In addition, v b was determined 0  S  during thefirst30 seconds of the reaction to ensure the concentration of product (UDPglucuronic acid) in the assay mixture remained less than 200 u.M. Under these conditions, a standard assay (5.3.16) was used to determine the value of MichaelisMenten parameters for a-galactosyl fluoride and lactose. A plot of v„bs as a function of a-galactosyl fluoride concentration (0-200 mM; Figure 3-12) produced a straight line.  ?  0.02  |  o E  e o.oi CO XI O  >  0 0 20 40 60 80 100 120 140 160 180 200 [a-Galactosyl Fluoride] (mM) Figure 3-12: Rate of UDP-galactose formation as a function of a-galactosyl fluoride concentration at 30 °C using standard assay conditions for the UDP-galactose 4epimerase / UDP-glucose dehydrogenase coupling system.  No saturation was observed over this range of concentrations however,  kcat/Km  was  determined to be 3.9 + 0.1 (x 10') min". The Lineweaver-Burk plot provided K m = 270 3  1  ± 30 mM. These values compare favourably with those obtained by direct assay using the fluoride electrode (kcat/ K m = 6.0x10" min" and K m , p p 220 mM). 3  1  =  a  In the presence of lactose, the rate of UDP-glucuronic acid production should decrease. First, lactose should compete with UDP as an acceptor for glycosyl transfer  98 and second, any UDP-galactose formed may then be consumed by lgtC-19 to form trisaccharide (see Figure 3-9). By adding varying concentrations of lactose (0-120 mM) to a standard assay mixture containing 50 mM a-galactosyl fluoride, its effect on the rate of UDP-galactose release could be measured. In this way, v bs was observed to decrease 0  in the presence of increasing lactose concentrations (Figure 3-13). Utilising the method  0.06 * -  o 83  c E  3  0.02  0.04 0  20  AO  60 80 [lactose] nM  100  120  0.02 J  0  I  20  I  J  l__l_  40  60  I  80  I  I  100 120  [lactose] mM Figure 3-13 : Effect of acceptor concentration on v bsfor the lgtC-19-catalysed transfer of galactosefroma-galactosylfluoride using the UDP-galactose 4-epimerase/ 0  UDP-glucose dehydrogenase coupling system: (+) lactose acceptor; (O) cellobiose acceptor. (INSET: inverse curve used to determine Km=7.0±0.8  mM)  described above to calculate the kinetic constants for UDP-glucuronic acid, Km for lactose was determined to be 7.0 ± 0.8 mM under these conditions.  Cellobiose  (GlcP(l-»4)Glc) differsfromlactose (Gaip(l-»4)Glc) at the 4-position of the nonreducing sugar. For lactose, the hydroxyl group on this carbon is axial while it is positioned equatorially for cellobiose. Because lgtC-19 transfers galactose to an acceptor to form an a(l->4) linkage with the non-reducing terminus, the stereochemistry at this position is critical for a substrate to act as an acceptor. Thus, cellobiose would not be  99 expected to accept galactose in the lgtC-19-catalysed galactosyl transfer. However, when cellobiose (40 mM) was added to the assay mixture described above, in place of lactose, an 18% reduction in the rate of UDP-glucuronic acid production was observed (Figure 2  3-13). At the same concentration, lactose effected a 72% decrease in v bs- How then, 0  does cellobiose effect this observed reduction in rate if it is not competing with the coupling system for UDP-galactose? Previous experiments (2.4.2) have suggested that the galactosyl transfer reaction catalysed by lgtC-19 occurs via an ordered bi bi sequential mechanism. It is possible that the acceptor, as the second-binding substrate, induces a conformational change in lgtC-19 that strains the sugar-UDP bond or otherwise increases its susceptibility to cleavage. If such a change in conformation were to occur, it is possible that the UDP binding site might be malpositioned for the formation of a bond between UDP and galactose. Thus, UDP would be a less efficient acceptor in the formation of UDP-galactose. This effect would manifest as a decrease in v b for UDP-glucuronic acid. In this way, cellobiose 0  S  might be able to affect UDP-galactose synthesis by inducing the conformational change. From the data gathered here, it is impossible to substantiate these claims. A crystal structure of lgtC-19 showing bound substrates is needed to support any assertions of this nature.  2  It has subsequently been determined that UDP-galactose acts as a substrate for  UDP-glucose dehydrogenase and is converted to UDP-galacturonic acid. Thus, a mixture of UDP-galacturonic acid and UDP-glucuronic acid is produced from the coupling reaction.  CHAPTER 4  CONCLUSIONS AND FUTURE WORK  101  4.1 Conclusions The recent availability of many uncharacterised glycosyl transferases presents a unique opportunity to explore the behaviour of a class of enzymes which may prove to have the most significant impact on the commercial production of oligosaccharides this century. So little is known about these biological catalysts that the process of defining their activities and optimising the reactions they catalyse presents a serious challenge. Of these enzymes, lgtC-19 is one of the few retaining glycosyl transferases to be studied. The experimental results presented in this thesis contribute to an expanded understanding of the basic requirements, kinetic properties, and catalytic mechanism of lgtC-19. Conditions required for maintaining enzyme activity during storage as well as those necessary for activity were explored. Reaction conditions were determined for the enzymatic synthesis of a commercially promising trisaccharide using a novel synthetic glycosyl donor and the kinetic constants for a variety of substrates and cofactors were determined. Finally, insight into the mechanism of the galactosyl transfer reaction was provided. The basic requirements for lgtC-19 storage and activity have been determined. This enzyme is most stable between pH 6.0-7.0. Even under these optimal storage conditions however, 80-83% of enzyme activity is lost after only 24 hours at 30°C (based on enzyme concentrations of 0.048 mg/ml). On the other hand, negligible changes in activity are seen over much greater periods (up to 30 days) when concentrated enzyme (>9 mg/ml) is kept at 4°C. This inherent instability means that enzyme activity must be  102 determined with a standard assay performed at frequent intervals during experiments to ensure that changes in enzyme activity do not affect the results. Requirements for cofactors and other reaction components were investigated. As evidenced by an absolute requirement for the reducing agent, DTT, lgtC-19 requires free cysteines for activity. While enzyme activity can be seen when Mg is available to 2+  activate lgtC-19, optimal activation occurs with Mn . The kinetic parameters for this 2+  cofactor have been determined (Table 15) and the optimal concentration of Mn for 2+  catalysis was found to be 15 mM. Suprising flexibility in both the donor and acceptor substrates utilised by lgtC-19 was observed. In Nature, this enzyme catalyses the transfer of galactose from UDPgalactose, the natural donor substrate, to a lipopolysaccharide element in which a lactose moiety defines the non-reducing end. A number of synthetic acceptor molecules have been admitted by this enzyme. From complex glycosides such as FCHASE-lactose, FCHASE-galactose, and FITC-lactose to a simple disaccharide such as lactose, and even a monosaccharide derivative (a-galactosyl fluoride), lgtC-19 will transfer galactose to each of these alternative substrates. Even more significant is the finding that lgtC-19 can use alternate donor substrates.  The ability of a nucleotide diphosphate dependent  glycosyl transferase to transfer a glycosyl unit from a simple synthetic substrate has not previously been demonstrated. The combination of UDP and a-galactosyl fluoride required to make this reaction work provides evidence that simple, inexpensive substrates can be used, enabling progress toward industrial application of this technology to be made. The kinetic parameters for UDP-galactose, lactose, Mn , UDP, and a-galactosyl 2+  fluoride are summarised in Table 15.  103  Table 15 : Kinetic parameters for substrates and cofactors utilised by lgtC-19 as determined in 20 mM HEPES buffer at the indicated pH, 30 °C.  SUBSTRATE  k^t (min")  K (pH)  1  m  kcat/Km  (min^mM") 1  UDP-galactose  b  52  uM (7.5)  b  3.4 x 10  b  29  uM (7.0)  b  2.7 x 10  b  47  mM (7.5)  b  3.4 x 10  41 ± 7 mM (7.5)  a  mM (7.0)  b  2.7 x 10  a  ( 2 . 0 ± 0 . 0 5 ) x 10  b  a  Mn  2+  a-galactosyl fluoride  b  50  a  4 . 9 ± 0.5 mM (7.5)  °*220 mM (7.5) e  UDP  7.2 x 10  3 0 ± 2 uM (7.5)  b  Lactose  (3.3 ± 0.05) x 10  a  270 ± 30 mM (7.5) 0.41  c  mM (7.5)  a  3  3  3  a  6.6 x 10  (2.2 ± 0.2) x 10  a  3  4  72  54  b  3  4  9.5 x 10  b  3  4  55  a  410  ' 1.3  c  6.0 x  e  1.0  e  3.9 x 10"  c  1.3  c  3.2  c  d  3  IO"  3  3  " apparent values determined using standard reaction conditions with either ImM UDP-galactose or 125mM lactose. determinedfromreplots using Cleland's method. performed in lOOmM HEPES under standard conditions for thefluorideion assay. calculated based on experimentally observed kcat for UDP and kcat/Km for a-galactosylfluoride. obtained using standard reactions conditions using the UDPgalactose 4-epimerase/UDP-glucose dehydrogenase coupling system. b  c  d  e  Results of pH studies indicate that two residues may be involved in the catalytic mechanism of this galactosyl transferase, a nucleophile with pKa=6.7 and a base catalyst with pKa=7.2. Kinetic behaviour was analysed using the method described by Cleland (1970).  The double-reciprocal plots of rate dependence on the independently varying  concentrations of lactose and UDP-galactose appear to be parallel lines.  Some  104 intersection of these lines is seen at high lactose concentrations, increasingly so as pH increases. This may result from error inherent in the measurements, amplified by the close proximity of the lines in this region of the reciprocal plot, or it may represent a fundamental component of the mechanistic course of the reaction. Additionally, the K m of UDP-galactose increases as the concentration of lactose increases and vice versa. The increasing K m values may indicate that the two substrates compete for the same enzyme form The out-of-order binding of one substrate inhibits binding of the other through the formation of a dead-end enzyme-inhibitor complex (Segel, 1993). There are four situations in which parallel lines are observed in this type of plot. The most common is for a ping pong reaction mechanism Secondly, this may occur in an ordered bi bi system in which the dissociation constant for the first-binding substrate is much smaller than its K  m  value. In this situation (see Appendix A), the family of  reciprocal plots intersect far to the left of the 1/v-axis and far below the 1/[AX]-axis. The same effect is observed when [B] is varied. The "best fitting" lines may appear parallel, allowing for small errors in determining v . The product inhibition patterns for this 0  system would not correspond to those for a ping pong reaction. Unfortunately, these inhibition studies could not be performed because of the type of assays used for measuring activity. The amplification of the difference between a ping pong and bi bi ordered system can be made, however, by adding a competitive inhibitor to the reaction mixture. Third, parallel initial velocity patterns occur in rapid equilibrium bi bi random systems in which the binding of one substrate strongly inhibits the binding of the other (i.e. a » 1). This mechanism can be determined by product inhibition studies. If a competitive inhibitor is added to this reaction mixture, the appearance of the velocity  105 pattern should remain unchanged. Finally, a non-rapid equilibrium random system will also yield seemingly parallel plots if the rate constants for the release of UDP-galactose and lactose are lower than V  max  for the forward reaction. Again, this system should  remain unaffected by the addition of a competitive inhibitor of one substrate. The fact that intersecting double-reciprocal plots were observed when UDP-glucose was added as a competitive inhibitor of UDP-galactose (both UDP-galactose and lactose concentrations were varied) appears to indicate that the galactosyl transfer reaction catalysed by lgtC-19 operates via an ordered bi bi reaction mechanism (Tsopanakis & Herries, 1978; Powell & Brew, 1974; Khatra et al., 1974). Recall that this is described by:  E  +  A X  k, ^ " k_i  B ( k E«AX  B X k  2  k  _2  kj  p  (E«AX»B  k .p  j  E«A»BX)  E »A k  .3  k ^ k  4  *~  E  +  A  .4  It is possible that lgtC-19 produces UDP-galactose from a-galactosyl fluoride and UDP which then acts as a substrate for galactosyl transfer to lactose. The true nature of this reaction mechanism is unclear. It suggests, however, that UDP-galactose may bind to lgtC-19 first, followed by lactose in an ordered bi bi sequential mechanism Additionally, nucleophilic attack of the sugar-UDP bond must occur in the absence of lactose. Thus, the binding of lactose to the enzyme may change the affinity of the enzyme towards UDP, expediting its expulsion from the active site. Determination of the precise mechanism and binding order requires detailed product inhibition studies that are out of reach using currently available assay methods for lgtC-19.  106 4.2 Future Work 4.2.1 Short-term Goals  The work presented in this thesis is a starting point for more detailed mechanistic studies of lgtC-19. While the evidence appears to suggest that the transfer of galactose from donor to acceptor occurs in an ordered bi bi sequential mechanism, product inhibition studies could provide additional support for this theory as well as information about the order of product releasefromlgtC-19. A new assay, able to operate in the presence of varying concentrations of both UDP and UDP-galactose, will need to be developed to monitor these effects. Until such a system is devised, a study of the inhibition patterns of UDP-glucose with UDP-galactose and lactose may provide some insight into the mechanism of this enzyme. Kinetic information for lactose derivatives in which the anomeric stereochemistry is fixed should be gathered to assess its importance for reaction rates. It is possible that the enzyme may prefer the P-anomer (which mimics the natural substrate), providing a more efficient route to trisaccharide synthesis.  Evidence for the formation of a  galactosyl-enzyme intermediate may be obtained using 2-deoxy-2-fluoro-a-D-galactosyl fluoride or 5-deoxy-5-fiuoro-a-D-galactosyl fluoride as donor analogues in the presence of UDP and a non-competent acceptor.  These compounds are currently being  synthesised by members of the Withers lab. Isolation of the labelled peptide and subsequent sequencing would provide the identity of the catalytic nucleophile.  107 4.2.2 Long-term Goals There remains a vast amount of information to uncover regarding the behaviour of glycosyl transferases in general and lgtC-19 in particular. The crystal structure of lgtC-19 will provide valuable information about the interaction between enzyme and substrate. Selective mutagenesis of key residues in the enzyme active site, along with detailed kinetic information about the role these residues play in the catalysis of glycosyl transfer, may ultimately provide the key to understanding the complexities of a vast collection of enzymes capable of efficient oligosaccharide synthesis. Future studies will generate greater insight into the character of these incredible biological catalysts.  CHAPTER 5  MATERIALS AND METHODS  109  5.1 Chemicals, Enzymes, and General Procedures  5.1.1 Reagent Sources  HF-pyridine was purchased from Aldrich. All other solvents and reagents were of analytical, reagent, or HPLC grade and were obtained from Sigma and used without further purification. Acceptor substrates obtained from Dr. W. Wakarchuk (N.R.C., Ottawa) include FCHASE-lactose, FCHASE-galactose, and FITC-lactose. LgtC-19 mutant galactosyl transferase that had been isolated from E. coli cells and lyophilised was a generous gift from Dr. Wakarchuk. Coupling enzymes; type H rabbit muscle PK and LDH, were obtainedfromSigma as suspensions in 3.2 M ammonium sulphate, pH 6.0. Coupling enzymes; UDP-galactose 4-epimerase from E. coli and UDP-glucose dehydrogenase from group A Streptococcus, were gifts from Dr. H. M. Holden (University of Wisconsin) and R. Campbell (University of British Columbia) respectively.  The UDP-galactose 4-epimerase was received in 10 mM potassium  phosphate buffer (pH 7.0) and required extensive dialysis in 10 mM HEPES buffer (pH 7.0, 1 mM EDTA) at 4°C prior to use. This was accomplished using 10,000 kDa molecular weight cut-off Spectra/Por® Molecular porous dialysis tubing. The UDPglucose dehydrogenase was stored in 50 mM Trien-HCl buffered at pH 8.7 and was used without further purification. 5.1.2 Analytical Methods  Thin-layer chromatography (TLC) separations were performed using Merck Kieselgel 60 F254 analytical plates. Compounds were detected visually (when possible)  110 under long-wave U.V. light, or by charring with either 5% sulphuric acid in methanol or 10% ammonium molybdate with 2 M sulphuric acid. Mass spectrometry was performed using a Sciex API 300 triple quadrupole LC/MS/MS electrospray mass spectrometer. High pressure liquid chromatography (HPLC) was performed using a Waters HPLC system equipped with a Waters 410 differential refractometer on a 4.6 mm x 150 mm Rainen Dynamax™ column equipped with a guard column (60 A, 8 urn NH-linked 2  packing). A mixture of acetonitrile and water suitable for the separation of reaction components was used for all HPLC separations. A Pharmacia Phast System separation and control unit was used to separate protein bands of column fractions for analysis by SDS-PAGE. NOE experiments were performed on a Briiker 300 MHz spectrometer. 'H NMR, COSY and HetCor experiments were performed using a Varian Unity 500 MHz spectrometer with referencingfromthe water signal at 4.72 ppm and water suppression techniques employed. The C NMR experiment was performed on a Varian 300 MHz 13  spectrometer. All NMR measurements were made at 25°C. 5.1.3 Sep Pack™ Purification  Products from enzymic syntheses using fluorescent-labelled acceptors were isolatedfromreaction mixtures by reverse-phase chromatography using 1 ml Sep Pack™ C-18 cartridges. The cartridges were pre-conditioned by a 10 ml acetonitrile wash followed by 10 ml of deionised water. Reaction mixtures were diluted to 1 ml using deionised water then applied to the Sep Pack™ column. Reaction components other than the fluorescent-labelled compounds were elutedfromthe column by a 10 ml water wash.  Ill  The labelled substances eluted in 2 ml of 50% acetonitrile and were evaporated in vacuo to a final volume of 0.2-0.5 ml prior to analysis. 5.1.4 LgtC-19 Purification Crude, lyophilised enzyme was dissolved into 10 mM sodium acetate buffer (pH 5.7, 2 mM EDTA) at a concentration of approximately 1 mg of protein per millilitre of buffer. The lyophilate was difficult to dissolve and required gentle shaking for 4-6 hours at 4°C. Particulate matter was removed by first centrifuging the enzyme solution then filtering the supernatant through a 0.45 pm Millipore™ syringe filter. Approximately 0.4 L of enzyme solution was loaded onto a 20 ml Source S-15™ ion exchange column using a peristaltic P-50 pump (Pharmacia) to deliver 2 ml/minute. Flow-through eluent was collected and reloaded onto the column before washing to baseline with 10 mM NaOAc buffer (pH 5.7, 2mM EDTA). Protein was eluted from the column using a salt gradient of 0 - 40 mM over 10 column volumes. Three protein peaks were detected but only one exhibited lgtC-19 activity. Fractionsfromthis peak were examined for purity by SDSPAGE and those containing > 80% lgtC-19 (as determinedfromcomparison to a purified standard provided by Dr. Wakarchuk) were pooled, concentrated to a final volume of 1 ml using Amicon Centriprep™ 10,000 molecular weight cut-off concentrators. The concentrated eluent was then applied to a Hi Prep® 16/60 Sephacryl S-100 high resolution pre-packed column (Pharmacia Biotechnology). 50 mM Ammonium acetate buffer (pH 7.0, 2 mM EDTA) was used to elute the protein in one peak. Fractions containing > 95% lgtC-19 were pooled and concentrated then stored at 4°C.  112 Approximately 15 mg of pure lgtC-19 were obtained from each 400 mg of crude lyophilate.  5.2 Synthesis  5.2.1 General Compounds  q-D-Galactosyl Fluoride Preparation of 1,2,3,4,6-penta-O-acetyl a-D-galactopyranose (1)  Galactose (60g) was dissolved in dry pyridine (420 ml) then cooled to 0°C. Acetic anhydride (285 ml) was added slowly with stirring then allowed to stir for 5 days at room temperature. The reaction was quenched by the addition of ice water (1.5 L) and extracted into ethyl acetate (500 ml). The organic extract was washed with 10% HC1 until pH<6, washed with 5% sodium bicarbonate solution until basic to litmus, then with saturated NaCl. Solvent was removed in vacuo to give a white solid in 88% yield (114 g). The product, recrystallised from 95% ethanol, had a mp. of 109-110°C (lit. mp. 112113°C; Wolfrom& Thompson, 1963) Preparation of 2,3,4,6-tetra-O-acetyl a-D-galactosyl fluoride (2)  Compound (2) was prepared by the method of Hayashi et al. (1984). Compound (1) (5 g) was dissolved in 70% HF-pyridine (4 ml/mg glycoside). After 24 hours at 0°C, the reaction was complete by TLC. The reaction was quenched in ice water (200 ml) and CHCI3  (200 ml). The organic layer was removed and the aqueous solution extracted  against  CHCI3  (5 x 50 ml). The pooled organic extract was washed with ice water (200  ml), followed by sodium bicarbonate solution (200 ml) after which the aqueous layer  113 remained basic, then with water (50 ml). The organic layer was dried over MgSO*4, filtered, and the solvent removed in vacuo to give a colorless gum. Toluene (3 x 100 ml) was added during the removal of solvent to pull out any remaining pyridine. The product was a rnixture of peracetylated a- and P-galactosyl fluorides which were separated by column chromatography (120 g silica gel, 5 cm column diameter, 30% ethyl acetate/ 70 % hexanes). Column fractions containing only product (2) were pooled and solvent removed in vacuo to yield 4.3 g of a clear oil (96% yield). 'H NMR (CDC1): 8 5.82 (dd, 3  1 H, JJ,F53.0 HZ, J 2.8 Hz, H-l), 5.55 (dd, 1 H, J , 1.8 Hz, J , 2.3 Hz, H-4), 5.50 (dd, 1 u  4 5  3 4  H, J 2.3 Hz, J 10.0 Hz, H-3), 5.27 (ddd, 1 H, J 21.8 Hz, J 10.0 Hz, J 2.8 Hz, 3)4  2s3  2 J  H-2), 4.34 (td, 1 H, J 1.8 Hz, J 5.1 Hz, J 4>5  5>6  I. 97, 2.25, 2.40, 2.50, (4s, 12 H, OAc).  5fi  19  2;3  u  7.3 Hz, H-5), 4.10 (m, 2 H, H-6, H-6'),  F NMR (CDC1, decoupled): 8 -74.74 (s, 3  a-fluoride). Preparation of a-D-galactosylfluoride  Compound (2) (0.428 g) was deacetylated with sodium methoxide/methanol then neutralised over a silica plug in dry methanol. The solvent was removed in vacuo to give a clear oil. This wasfreeze-driedto a white solid (0.221 g, 95% yield) which was pure by TLC. Enzymic Synthesis of FCHASE- and FITC-linked Oligosaccharides: Into an eppendorf vial were added 3 JJ1 . of 115 mM a-galactosyl fluoride (0.35 umoles), 3 ul of 100 mM UDP (0.3 umoles), 5 ul of 5 mM acceptor (0.025 umoles), and 4 ul of 500 mM HEPES buffer (pH 7.5) containing 50 mM MnCl and 25 mM DTT. The reaction 2  was initiated by adding 5 ul of stock lgtC-19 (6 mg/ml). Reactions were run concurrently as both positive (containing 6 ul of UDP-galactose instead of a-galactosyl fluoride and  114 UDP) and negative (containing either UDP-galactose or a-galactosyl fluoride and UDP but no enzyme) controls. The reaction progress was monitored by TLC (solvent system, 7:2:1:0.1 ethyl acetate, methanol, water, acetic acid).  After 17.5 hours at room  temperature, the reactions were complete. The Rf values for each reaction component were as follows: FCHASE-lactose acceptor Rf = 0.5, FCHASE-trisaccharide product Rf = 0.3, FCHASE-galactose acceptor Rf = 0.6, FCHASE-lactose product Rf = 0.5, FITClactose acceptor Rf = 0.4, FITC-trisaccharide product Rf = 0.25, a-galactosyl fluoride donor Rf = 0.37. After Sep Pack™ purification (see 5.1.3), the fluorescent components were analysed by electrospray mass spectroscopy.  5.2.2 Prep-scale Enzymic Synthesis of Trisaccha ride from a-Galactosyl Fluoride and Lactose  Into an eppendorf vial were added 15 ul of 1 M a-galactosyl fluoride (15 umoles), 5 ul of 1 M UDP (5 umoles), 100 ul of 500 mM Tris buffer (pH 7.5) containing 50 mM MnCfe and 25 mM DTT. Approximately 2.6 mg of purified lgtC-19 was added, followed by 40 ul of 250 mM lactose (10 pinoles) to initiate the lgtC-19-catalysed transfer of galactosefroma-galactosyl fluoride to lactose. The reaction was monitored hourly at room temperature. After 9 hours, the pH had dropped to 5.0 and a white precipitate had formed. NaOH was added to a pH of 7.5. After 23 hours, the reaction had reached >90% completion (by TLC). White precipitate was visible but the pH remained at 7.5. The enzyme was removedfromthe reaction mixture by centrifuging the reaction contents in a 10,000 kDa molecular weight cut-off Amicon™ microconcentrator. The retained protein  115 was recovered and the remaining reaction components were separated using HPLC. The reaction was repeated at half this scale using the recovered enzyme. A visible loss of product occurred when the column eluent was filtered through a 0.22 pm Millipore™ syringe filter. However this filtration was necessary to remove leached column packing materialfromthe sample prior to NMR analysis. A total of 5 mg (9.9 umoles) of purified trisaccharide product was obtained, representing an overall yield of 66%. H NMR, C J  13  NMR, HetCor, and NOE analyses of the product were made. H NMR (D2O): (where X  A^x-Gal(l,4), B=|3-Gal(l,4), and C^Glc) 8 5.22 (d, 0.4 H, J = 3.6 Hz, H-lC ), 4.95 (d, u  1 H, Ji.2 = 3.6 Hz, H-1A), 4.66 (d, 0.6 H, J  u  a  = 8 Hz, H-lCp), 4.52 (d, 1 H, J  u  = 8 Hz,  H-1B), 4.39 (t, 2 H, J , = 8 Hz, H-5A,B), 4.08 (t, 2 H, J = J 6 = 3.6 Hz, H-4A,B), 3.65 6  4)5  5>  4.0 (m, 12 H, H-2A,B,C; H-3A,B,C; H-4C; H-6A,B,C), 3.35 (t, 1 H, J = 8 Hz, H-5C). 5>6  5.3 Enzymology  5.3.1 General Procedures  All absorbance measurements were made on a UV/VIS Pye-Unicam 8700 spectrophotometer equipped with a temperature-controlled circulating water bath. Generally, acryl cuvettes with a 1 cm path length were used for measurements taken at 340 nm When reaction volumes less than 800 pi were required, quartz microcuvettes were used and reaction volumes decreased to 300 pi. LgtC-19 concentrations were determined based on £ 2 8 o l -74 ml mg^cm" (provided by Dr. W. Wakarchuk, personal =  1  communication) using matched quartz cells.  116 Fluoride ion concentrations were determined using an Orion® fluoride ion electrode equipped with a temperature-controlled water bath. Reaction volumes were maintained at 300 ul or greater to optimise electrode stability. Prior to use, a three-point calibration of the electrode was made using standard solutions of NaF (at concentrations bracketing those measured during experimental determinations; typically, 0.1 M, 0.001 M, and 0.0001 M) in buffer, MnCfe and DTT concentrations reflecting those used for the particular experiment. 5.3.2 Determination of Steady-state Kinetic Parameters  Michaelis-Menten parameters for the lgtC-19-catalysed transfer of galactose from UDP-galactose to lactose were determined using a continuous, coupled assay. Standard assay mixtures had a total volume of 800 ul and contained 20 mM HEPES buffer at either pH 7.5 or 7.0, 1 mg/ml BSA, 50 mM KCl, 15 mM MnCl , 5 mM DTT, 95 U PK , 3  2  215 U LDH ,125 mM lactose, and 0.5 mM UDP-galactose. Solutions were preincubated 2  within the spectrophotometer at 30°C until thermally equilibrated and reactions were initiated by the addition of enzyme unless otherwise specified. The release of UDP upon transfer of galactosefromUDP-galactose to lactose was measured by the concomitant decrease in absorbance at 340 nm as the coupling system converted NADH to NAD . +  Reaction rates were determined based on an 8340=6.22 mM'cm" for NADH (Palcic et al., 1  1988). To ensure linear kinetics and a sufficient absorbance change for accurate initial rate calculations, the concentration of the enzyme added and the reaction time were  3  1U of enzyme activity is defined as that amount of enzyme which catalyses the  turnover of 1 umole of substrate per minute at 30°C under saturating conditions.  117 selected so that less than 10% of the total substrate was converted to product during the assay. Kinetic parameters for the transfer of galactose from a-galactosyl fluoride to lactose by lgtC-19 were determined by a continuous assay in which F release was monitored. Standard assays, 300 pi in volume, contained 100 mM HEPES (pH 7.5 or 7.0), 15 mM MnCk, and 5 mM DTT. Lactose and a-galactosyl fluoride concentrations were adjusted to appropriate levels for individual experiments. Reaction cells were preincubated in a 30°C water bath until thermally equilibrated then the background hydrolysis of a-galactosyl fluoride was measured for 5 minutes before lgtC-19 was added to initiate the transfer reaction. After initiation, the reactions were monitored for 10 minutes and the reaction rate corrected for background hydrolysis. The concentration of enzyme used was adjusted to ensure the reaction rate was at least twice the rate of background hydrolysis and yet consumed less than 10% of the substrate over the course of the assay. For each substrate, K^app and Vma^app were determined by measuring initial rates as a function of substrate concentration with all other variables held constant in accord with the standard reaction conditions. In general, concentrations of the substrate under study variedfrom0.2 to 7 times the K value. The data are presented as double m  reciprocal plots (Appendix B.I) to allow rapid visual evaluation of the data. Kinetic parameters were determined directly by nonlinear regression analysis (Leatherbarrow, 1990).  118 5.3.3 Assessment of Coupling System Suitability  Using standard reaction conditions (5.3.2), the rate of galactosyl transfer from UDP-galactose to lactose was measured as a function of lgtC-19 concentration. Enzyme concentration was varied from 0 - 0.001 mg/ml, producing initial rates bracketing those anticipated for all kinetic determinations. The coupling system was deemed adequate if the measured rate exhibited an immediate linear response to changes in lgtC-19 concentration. 5.3.4 Preincubation Studies  To assess the effect of preincubation time with various reaction components, lgtC19 activity was measured as a function of preincubation time.  The effect of  preincubation time in the presence of DTT was determined by running two parallel reaction series. All reaction mixtures were identical except that series #1 contained all standard assay components except UDP-galactose and DTT and series #2 contained all components except UDP-galactose. For series #1 reactions, DTT was added immediately before the reaction was initiated with UDP-galactose. In this way, lgtC-19 was permitted to preincubate in the presence of all other reaction components for a time interval equal to that of the series #2 reactions in which DTT was present in the preincubation mixture. Thus, the only variable between the two reaction series was the amount of time lgtC-19 spent in the presence of DTT. To determine the optimal time required for Mn to 2+  activate lgtC-19, the enzyme was diluted and stored in buffer containing 5 mM DTT and left for one hour prior to assay. All reaction components were assembled as before except for MnCk and UDP-galactose. The MnCk was added to the preincubation mixture then the reaction initiated with UDP-galactose after measured time intervals. For  119 each of these experiments, initial rates were plotted as a function of this preincubation time to determine the time required for activation. 5.3.5 Metal Cofactor Specificity  The dependence of Vmax on metal ion concentration was determined as follows. A final concentration of 5 mM MnCfe, MgCk, CaCfe, or ZnCfe was added to a standard assay mixture containing all components except MnCk and lgtC-19. After thermal equilibration at 30°C, the reaction was initiated by the addition of lgtC-19 and the change in A340 measured for initial rate detenninations. After the reaction was complete, the pH of each assay mixture was tested to ensure that it had not fluctuated over the course of the reaction. The enzyme activity afforded by each of these metals was then compared with that of MnCk. Kinetic parameters were sought for MgCk and CaCfe using the methods outlined in 5.3.2. 5.3.6 Cleland Kinetics-Grid Experiments  The effect of independently varying UDP-galactose and lactose concentrations on kinetic parameters was determined as follows. A range of 5-7 different concentrations of both UDP-galactose (0.01022-1.022 mM) and lactose (5-145 mM) were chosen with limitsfromapproximately 0.2 to 20 times K and 0.1 to 3 times Km, respectively. At m  each concentration of UDP-galactose, the concentration of lactose was varied. In this way, experiments in which 5 different concentrations of each substrate were used resulted in a total of 25 reactions. All reaction components except the variable substrates and lgtC-19 were prepared in multi-component assay mixtures. Cuvettes were prepared within two hours of the assay to minimise interaction between DTT and MnCfe. Standard  120 cells containing 0.5 mM UDP-galactose and 125 mM lactose were used to measure enzyme activity prior to each set of reactions. A set is defined as a group of reactions in which the concentration of lactose is constant. In this way, observed rates could be corrected to account for loss of enzyme activity. This was necessary since, although the lgtC-19 was maintained at 4°C until added to each reaction mixture (10 pi of 0.0367 mg/ml; 0.00037 mg), these experiments spanned 10-35 hours and lgtC-19 loses activity over that time period at the low concentrations required for the assay. 5.3.7 pH Dependence Studies The stability of lgtC-19 as a function of pH was determined by storing the diluted enzyme in solutions buffered from pH 3.0 through 9.0. 20 mM AMPSO buffer (pH 9.0), HEPES buffer (pH 7.0, 8.0), and citrate buffer (pH 3.0, 4.0, 5.0, 6.0) containing 5 mM DTT were used. None of the buffers contained BSA to ensure that variations in the protective properties of BSA with pH would not influence the results.  After a  preincubation time of 1 hour at 4°C, the activity of each enzyme mixture was measured under standard assay conditions at pH 7.5 to provide a baseline measurement at t=0. After this initial assay, the enzyme mixtures were immersed in a 30°C water bath to mimic standard reaction conditions.  The activity of each enzyme mixture was  determined by removing aliquots for assay at timed intervals. The observed activity was then plotted as a function of time for each mixture. The change in activity with respect to time was fitted to a first order exponential decay (GraFit 3.0; Leatherbarrow, 1990) which yielded rate constants for each pH value. LDH has a pH optimum of 6.0 but retains 84% and 50% of maximal activity at pH 7.0 and 8.0 respectively. It is stable for 180 minutes at 25°C from pH 5.0-10.0 but  121 unstable at pH<5 (Keesey, 1987). PK has a pH optimum of 7.0-7.5 and is highly unstable at pH<6.0. It is stable for 1 hour at pH 6.5-8.0 and after 3 hours at pH 9, 94% of activity remains (Keesey, 1987). At pH>7.7, the manganese added to the assay mixture precipitated as a metal oxide. Based on the requirements of the coupling system and the stability of reaction components, a pH range of 6.0-7.7 was chosen for study. Next, the adequacy of the PK/LDH coupling system was determined at pH 6.0, 6.5, 7.0, 7.5, and 7.7. Two different buffers were required to cover this pH range: HEPES and MES. A solution of each buffer was prepared at pH 7.0 and duplicate measurements made, enabling correction for buffer effects. Finally, a grid experiment was performed (see 5.3.6) at each pH using five lactose concentrations and 5-7 different UDP-galactose concentrations.  Km pp >a  and Vmax.app were determined at each concentration  of lactose and the data reploted as described in Appendix A to permit the determination of the true values of  Km  and k ^ -  Plots of  k^/Km  (the apparent second-order rate  constant) as a function of pH for each substrate yielded pH curvesfromwhich the p K a values of relevant functional groups were determined using GraFit 3.0 (Leatherbarrow, 1990). 5.3.8 Inactivation Experiment The pseudo-first order rate constant (kobs) for inactivation was calculated for lgtC19 using the inactivator dial-UDP. Inhibition mixtures (I-mixes) contained 20 mM HEPES (pH 7.5), 5 mM DTT, 15 mM MnCl and varying concentrations of dial-UDP 2  (from 0.5 to 20 mM). A control mixture was assembled which contained all components except dial-UDP.  Additionally, two I-mixes, one with no dial-UDP and another  containing 10 mM dial-UDP, contained 0.3 mM UDP-galactose as a competitive  122 inhibitor of the inactivation. All of the I-mixes were incubated at 23°C and inactivation times were measuredfromthe timed addition of dial-UDP (or UDP-galactose in the case of controls). Aliquots (lOul) were removedfromthe I-mixes at timed intervals and diluted into reactions cells containing standard assay components at pH 7.5, 30°C. Inactivation was arrested at the time the assay was initiated by both the dilution of inactivator in the assay cell and substrate competition for the enzyme since UDPgalactose was present at saturating concentration (0.3 mM). Residual enzyme activity was determinedfromthe rate of galactosyl transfer and is directly proportional to the amount of active enzyme. Inactivation was monitored until >90% of enzyme activity was lost,  kobs  for inactivation was calculated by direct fit of the  v„b  S  versus time curve for  each concentration of dial-UDP to afirst-orderfunction using GraFit 3.0 (Leatherbarrow, 1990). Plots of kobs as a function of dial-UDP concentration were constructed and kinetic parameters for k i and K i determined by nonlinear regression analysis (GraFit 3.0; Leatherbarrow, 1990). 5.3.9 Labelling Experiment with Dial-UDP Residues involved in the formation of covalent bonds with inactivators can be identified using mass spectroscopy techniques (Miao et al., 1994). If a protein forms a covalent bond with an inactivator, the mass of the protein will increase by the mass of the label. To see if an increase in mass occurred when lgtC-19 was incubated with dial-UDP, the following experiment was performed. Four samples were assembled. Each contained approximately 4 mg/ml of lgtC-19 in 50 mM ammonium sulphate buffer, pH 7.0. Samples 1,2, and 3 were diluted 2-fold by the addition of 13 mM HEPES (pH 7.5, 5 mM DTT) and dial-UDP to a final concentration of 0 mM, 5 mM and 10 mM for samples 1, 2  123 and 3, respectively. After 38 hours, a 25 mM NaBFLt solution was added to sample 3 to give a final concentration of 5 mM dial-UDP, 12.5 mM NaBHU, 13 mM HEPES, 15 mM MnCl , 5 mM DTT, and 1 mg/ml lgtC-19. Sample 4 was similarly diluted with NaBFL; 2  solution to provide a controlfreeof dial-UDP but subject to the reducing effects of NaBHt. Samples 1 and 2 were diluted with buffer to give protein concentrations of 1 mg/ml then all of the samples were analysed by electrospray mass spectroscopy. 5.3.10 Labelling Experiment with UDP-galactose  Into an eppendorf vial were placed lgtC-19 (0.08 mg), MnCl (15 mM), HEPES 2  buffer (100 mM, pH 7.5) and DTT (5 mM) to yield a 40 ul mixture. From this, three 10 ul aliquots were removed and placed in separate vials. One vial served as a control and was diluted 2-fold with deionised water, the second vial was similarly diluted with a solution of UDP-galactose to yield a final concentration of 10 mM UDP-galactose, and the third vial was treated the same as the second except it also contained 125 mM lactose. The mixtures were incubated at 30°C for 20 minutes then analysed by electrospray mass spectroscopy to determine if lgtC-19 had formed a covalent galactosyl-enzyme intermediate as evidenced by a mass increase of 163 g/mol. 5.3.11 Burst Experiment  The burst experiment for lgtC-19 involving the addition of a-galactosyl fluoride in the presence of UDP was performed by following the release of F. To 100 mM agalactosyl fluoride in 100 mM HEPES, pH 7.5 with 15 mM MnCl , 2 mM UDP and 5 2  mM DTT at 30°C was added 30 ul of enzyme for a final concentration of 1.56 mg/ml. Based on the amount of enzyme added, a burst of 47 uM was expected. The  124 concentration of F" was measured for 1 minute prior to reaction initiation with lgtC-19 then every 5 seconds after initiation for a period of 10 minutes. After this time, more than 160 pM of F" had been released. With no evidence of a burst, lactose was added to a final concentration of 83 mM. Still, no burst was observed. 5.3.12 "Mini" pH Study  The pH dependence of Vmax for the lgtC-19-catalysed transfer of galactose from a-galactosyl fluoride to lactose was determined in two separate experiments. A series of 200 mM HEPES buffers were preparedfrompH 7.0 to 8.0. Standard assays were made in duplicate for each pH using 80 mM a-galactosylfluoride,100 mM lactose, and 2.054 mg/ml lgtC-19. The pH of each reaction mixture was checked immediately following the assay to ensure the pH remained unchanged. Next, a series of buffers was prepared from MES (pH 6.0, 6.5), PIPES (pH 6.5, 7.0), and HEPES (pH 7.0, 7.5, 8.0). The standard assay described above was repeated using each of these buffered solutions. V b from 0  S  both experiments were plotted as a function of pH but p K a values for the resulting curve could not be determined satisfactorily using the pH algorithm of GraFit 3.0 (Leatherbarrow, 1990). 5.3.13 Transglycosylation Experiment  The ability of a-galactosylfluorideto act as both acceptor and donor substrate in an lgtC-19-catalysed transglycosylation reaction was determined by the following method. Three reaction vials were assembled containing 100 mM Tris buffer (pH 7.5), 15 mM MnCl , 400 mM a-galactosylfluoride,3 mM UDP, and 5 mM DTT. Vial #1 2  served as a negative control into which 50 mM lactose were placed but no transferase, the  125 reaction in vial #2 was initiated with 10 ul lgtC-19 to give afinalconcentration of 4 mg/ml, and 50 mM lactose was added to vial #3 prior to initiation with lgtC-19. Reaction progress was monitored by TLC using a 4:2:1:0.1 (ethyl acetate/methanol/water/acetic acid) solvent system The reactions were left at room temperature for 2 days then the enzyme was removed by microconcentration. Components of the reaction mixtures were separated by HPLC (60:40 acetonitrile/water) and identified by comparison of retention times to standards. The individual components were collected, concentrated in vacuo, then submitted for mass spectral analysis. 5.3.14 Trouble-shooting M n  2+  Interference with the F electrode  The F" electrode was calibrated using buffered standards of 20 mM, 2 mM, and 0.2 mM F. Next, a series of 1 mM F" standards were prepared in 50 mM ammonium acetate buffer, pH 7.0. Each standard contained variable concentrations of MnCL. (0-30 mM). The concentration of F" was measured continuously for 5 minutes for each standard then plotted as a function of time. Next, the average concentration measured for each standard was plotted as a function of MnCL. concentration to show the concentration-dependent interference of MnCfe on electrode performance (see Appendix B.LX for results). 5.3.15 Kinetic Measurements in the Presence of UDP-glucose  An estimate of Ki was determined for UDP-glucose by the following method. Standard assay cells were prepared for assay with the PK/LDH coupling system. Each reaction was performed at pH 7.0 with 125 mM lactose and 0.04 mM UDP-galactose. The concentration of UDP-glucose in each reaction was variedfrom0 to 1.75 mM and  126 1/vobs  plotted as a function of UDP-glucose concentration.  competitive inhibitor, v =  V^tS]  In the presence of a  therefore, a plot of l/v bs versus [I] will 0  [S]+K a+rrj/K ) B  I  give a straight line along which lies a point with coordinates of l/Vmax and -Ki. By this method, an estimate of Ki can be made if V  max  is known.  To determine the effect of UDP-glucose, a potential competitive inhibitor of UDP-galactose, on the kinetic parameters for UDP-galactose and lactose with lgtC-19, the following experiment was performed. Using standard assay conditions at pH 7.0, a grid experiment was performed, varying UDP-galactosefrom0.01022 - 1.022 mM and lactosefrom5-125 mM. Each cuvette also contained 0.25 mM UDP-glucose. The data were plotted as double reciprocal plots to determine the pattern of intersection. 5.3.16 Enzymic Synthesis of UDP-galactose  For experiments in which a-galactosyl fluoride and UDP served as substrates for the lgtC-19-catalysed synthesis of UDP-galactose, a continuous coupled assay was used to monitor product formation. The coupling enzymes were UDP-galactose 4-epimerase (1) and UDP-glucose dehydrogenase (2). UDP-galactose serves as a substrate for (1) which converts it to UDP-glucose, which is then converted to UDP-glucuronic acid by (2) with the simultaneous reduction of two equivalents of NAD to NADH for each +  equivalent of UDP-galactose. Thus, the rate of the lgtC-19 reaction could be determined from the rate of increased absorbance at 340 nm Reaction cells contained 100 mM HEPES buffer (pH 7.5), 15 mM MnCl , 50 mM 2  a-galactosyl fluoride, 0.5 mM UDP, 5 mM DTT, 0.5 mM NAD , 4.3 mg/ml UDP+  galactose 4-epimerase, and 1.03 mg/ml UDP-glucose dehydrogenase. A control reaction  127 was monitored (in the absence of lgtC-19) to observe any interaction between agalactosyl fluoride and UDP with the coupling enzymes. After 10 minutes with no change in absorbance, 4 ul of 2 mM UDP-galactose (0.04 mM final concentration) were added and an immediate increase in absorbance was observed. Next, suitability of the coupling system was determined by the method described previously (5.3.3). To determine the effect of increasing concentration of product, UDP-glucuronic acid, on the observed rate of the coupling system, the following experiments were performed. First, UDP-galactose was added to the control reaction mentioned above in 4 ul (2 mM) serial additions for which the initial rate was determinedfromthe rate of change in A340. After all UDP-galactose had been converted into product, the next addition was made and monitored in the same manner. Second, into standard assay mixtures containing 100 mM HEPES buffer (pH 7.5), 15 mM MnCl , 5 mM DTT, 0.5 2  mM NAD and 0.04 mM UDP-galactose were placed variable concentrations of UDP+  glucuronic acid (0-400 uM). The reactions were initiated after thermal equilibration of the assay mixture to 30°C using 10 ul of the coupling enzyme mixture to give a final concentration of 0.0358 mg/ml and 0.0103 mg/ml for UDP-galactose 4-epimerase and UDP-glucose dehydrogenase, respectively.  Two cells were prepared with 200 uM  UDP-glucuronic acid, one as described above and the other containing 0.5 mM UDP. The observed reaction rates were plotted as a function of UDP-glucuronic acid concentration. By inverting the plot, a second-order curve was obtained which was fit to a nonlinear regression by GraFit 3.0 (Leatherbarrow, 1990) to obtain a 'Vorking" Ki for UDP-glucuronic acid with this coupling system  128 Michaelis-Menten parameters for a-galactosyl fluoride and lactose were determined using the continuous coupled assay described above. Standard assay cells contained 100 mM HEPES (pH 7.5), 15 mM MnCl , 0.5 mM UDP, 5 mM DTT, 0.5 mM 2  NAD , 4.3 mg/ml UDP-galactose 4-epimerase, and 1.03 mg/ml UDP-glucose +  dehydrogenase.  Reactions were initiated with 10 pi of 10.2 mg/ml lgtC-19 (final  concentration, 0.51 mg/ml). The concentration of a-galactosyl fluoride was varied from 0-200 mM. Data were analysed as described previously (5.3.2). Five cells were assembled with 50 mM of a-galactosyl fluoride and varying concentrations of lactose (0120 mM) or cellobiose (40 mM). The effect of acceptor concentration on the rate of UDP-glucuronic acid production was determined by plotting  v bs versus 0  lactose  concentration, the data were treated as described above (for the effect of UDP-glucuronic acid concentration) so that Km, pp for lactose could be determined under these conditions. a  129  REFERENCES  130 Andree, P. J., & Berliner, L. J. (1978) Biochim. etBiophys.  Acta  Armstrong, G. D., Fodor, E., & Vanmaele, R. (1991) J. Infec.  544,489-495.  Dis. 164,1160-1167.  Armstrong, G. D., Rowe, P. C., Goodyer, P., Orrbine, E., Klassen, T. P., Wells, G., MacKenzie, A., Lior, H., Blanchard, C., Auclair, F., Thompson, B., Rafter, D. J., & McLaine, P. N. (1995)7. Infec. Dis. 171,1042-1045. Ats, S. C., Lehmann, J., & Perry, S. (1992) Carb. Res. 233, 125-139. Baisch, G., Ohrlein, R., & Ernst, B. (1996a) Bioorg. Med. Chem.  Lett. 6,  749-754.  Baisch, G., Ohrlein, R., Katopodis, A , & Ernst, B. (1996b) Bioorg. Med. Chem. 759-762. Baisch, G., Ohrlein, R., Streiff, M., & Ernst, B. (1996c) Bioorg. Med. Chem. 758.  Lett. 6,  Lett. 6,  755-  Battaglia, E., Senay, C., Fournel-Gigleux, S., Herber, R., Siest, G., & Magdalou, J. (1994) FEBS Lett. 346,146-150. Bazaes, S. (1987) in Chemical Modification ofEnzymes (Eyzaguirre, J., Ed.) pp 35-45, Ellis Horwood Limited, Chichester, West Sussex. Bell, J. E., Beyer, T. A., & Hill, R. L. (1976) JBC 251,3003-3013. Berliner, L. J., Davis, M. E., Ebner, K. E., Beyer, T. A., & Bell, J. E. (1984) Mol. Cell Biochem. 62, 37-42. Berliner, L. J., & Robinson, R. D. (1982) Biochem. 21,6340-6343. Birsan, C. (1996) MSc. Thesis in Chemistry, University of British Columbia, Vancouver, pp 118. Blanken, W. M., & Van den Eijnden, D. H. (1985) JBC 260,12927-12934. Britten, C. J., & Bird, M. I. (1997) Biochim.  Biophys. Acta 1334,  57-64.  Campbell, J. A., Davies, G. J., Bulone, V., & Henrissat, B. (1997a) Biochem. 939.  J. 326,  929-  Campbell, R. E., Sala, R. F., van de Rijn, I., & Tannner, M. E. (1997b) JBC 272, 34163422. Carlow, D. C, Smith, A. A., Yang, C. C, Short, S. A., & Wolfenden, R. (1995) Biochem. 34, 4220-4224.  131  Cleland, W. W. (1970) in The Enzymes pp 1-65. Collins, P., & Ferrier, R. (1995) Monosaccharides: Their Chemistry and Their Roles in Natural Products, John Wiley & Sons Ltd., New York. Davies, G., & Henrissat, B. (1995) Structure 3, 853-859. Davies, G., Withers, S. G., & Sinnott, M. L. (1997) in Comprehensive Biological Catalysis (Sinnott, M. L., Ed.), Academic Press, London. Day, A. G., & Withers, S. G. (1986) Biochem.  Cell Biol.  64,914-922.  Dhalla, A. M., Li, B., Alibhai, M. F., Yost, K. J., Hemmingsen, J. M., Atkins, W. M., Schineller, J., & Villafranca, J. J. (1994) Prot. Science 3,476-481. Edebrink, P., Jansson, P. E., Widmalm, G., Holme, T., & Rahman, M. (1996) Carb. 295,127-146.  Res.  Fersht, A. (1985) Enzyme Structure and Mechanism, 2 ed., W. H. Freeman and Company, New York. nd  Flitsch, S. L., Taylor, J. P., & Turner, N. J. (1991) J.  Chem. Soc, Chem. Com.,  382-384.  Gebler, J., Gilkes, N. R„ Claeyssens, M., Wilson, D. B., Beguin, P., Wakarchuk, W. W., Kilburn, D. G., Miller, R. C, Jr., Warren, R. A. J., & Withers, S. G.(1992a) JBC 267,12559-12561. Gebler, J. C , Aebersold, R., & Withers, S. G. (1992b) JBC 267, 11126-11130. Gebler, J. C, Trimbur, D. E., Warren, R. A. J., Aebersold, R., Namchuk, M., & Withers, S. G. (1995) Biochem. 34, 14547-14553. Gosselin, S., Alhussaini, M., Streiff, M. B., Takabayashi, K., & Palcic, M. M. (1994) Anal. Biochem. 220,92-97. Gross, H. J., Bunsch, A., Paulson, J. C, & Brossmer, R. (1987) Eur. J. Biochem. 168, 595-602. Gross, H. J., Rose, U., Drause, J. M., Paulson, J. C, Schmid, K., Feeney, R. E., & Brossmer, R. (1989) Biochem. 28, 7386-7392. Haataja, S., Tikkanen, K., Liukkonen, J., Francois-Gerard, C, & Finne, J. (1993) JBC 268,4311-4317.  132 Hashimoto, H., Endo, T., & Kajihara, Y. (1997) J. of Org. Chem. 62,1914-1915. Hatanaka, Y., Hashimoto, M., Nishihara, S., Narimatsu, H., & Kanaoka, Y. (1996) Carb. Res. 294,95-108. Hayashi, M., Hashimoto, S. I., & Noyori, R. (1984) Chem. Lett., 1747-1750. Hehre, E. J. (1989) Denpun Kagaku 36,197-205. Heidlas, J. E., Lees, W. J., & Whitesides, G. M. (1992) J. Org. Chem. 57, 152-157. Hendrickson, T. L., & Imperiali, B. (1995) Biochem. 34,9444-9450. Hokke, C. H., Zervosen, A., Elling, L., Joziasse, D. H., & Van den Eijnden, D. H.(1996) Glycoconj. J. 13, 687-692. Holmes, E. H. (1990) JBC265,13150-13156. Holmes, E. H. (1992) Arch. Biochem. Biophys. 296, 562-568. Holmes, E. H., Xu, Z., Sherwood, A. L., & Macher, B. A. (1995) JBC 270, 8145-8151. Ichikawa, Y., Look, G. C, & Wong, C. H. (1992) Anal. Biochem. 202,215-238. Jeong, J. H., Murray, B. W., Takayama, S., & Wong, C. H. (1996) JACS118, 4227-4234. Jones, C. C, Sinnott, M. L., & Souchard, I. J. L. (1977) J. Chem. Soc., Perkin Trans. 2, 1191. Kajimoto, T., Liu, K.K.C., Pederson, R.L., Zhong, Z., Ichikawa, Y., Porco, J.A., Wong, C.H. (1991) •/yfCS'113, 6187-6196. Keesey, J. (1987), Beohringer Mannheim Biochemicals, Indianapolis. Kempton, J. B., & Withers, S. G. (1992) Biochem. 31,9961-9969. Khatra, B. S., Herries, D. G., & Brew, K. (1974) Eur. J. Biochem. 44, 537-560. Kim, S. C, Singh, A. N., & Raushel, F. M. (1988a) Arch. Biochem. Biophys. 267, 54-59. Kim, S. C., Singh, A. N., & Raushel, F. M. (1988b) JBC 263, 10151-10154. Kim, Y., Mlsna, D., Monzingo, A. F., Ready, M. P., Frankel, A, & Robertus, J. D. (1992) Biochem. 31, 3294-3296. Kleene, R., & Berger, E. G. (1993) Biochem. Biophys. Acta 1154,283-325.  133  Kobayashi, S., Kiyosada, T., & Shoda, S. (1996) JACS118,13113-13114. Koshland, D. E., Jr. (1953) Biol. Rev. 28, 416-436. Kuhn, N. J., Stankiewicz, M., & Ward, S. (1992) Biochem. Soc. Trans. 20, 714-716. Kuhn, N. J., Talbot, J., & Ward, S. (1991a) Arch. Biochem. Biophys. 286,217-221. Kuhn, N. J., Ward, S., & Leong, W. S. (1991b) Eur. J. Biochem. 195,243-250. Lawson, S. L., Wakarchuk, W. W., & Withers, S. G. (1996) Biochem. 35,10110-10118. Leatherbarrow, R. J. (1990) GraFit, 3.0 ed., Erithacus Software Limited, Staines, UK. Lee, T. K., Wong, L. J. C, & Wong, S. S. (1983) JBC 258,13166-13171. Legler, G. (1990) Adv. Carbohyd. Chem. Biochem. 48,319-383. Legler, G., & Jtilich, E. (1984) Carb. Res. 128,61-72. Legler, G., Sinnott, M. L., & Withers, S. G. (1980) J Chem. Soc, Perkin Trans. 2,1376. Lehmann, J., & Schlesselmann, P. (1983) Carb. Res. 113,93-99. Liu, W., Madsen, N. B., Braun, C., & Withers, S. G. (1991) Biochem. 30,1419-1424. MacLeod, A. M., Tull, D., Rupitz, K., Warren, R. A. J., & Withers, S. G. (1996) Biochem. 35,13165-13172.  Magee, S. C, & Ebner, K. E. (1974) JBC 249, 6992-6998. McCarter, J. D., Adam, M. J., & Withers, S. G. (1992) Biochem. J286, 721-727. McClelland, R. A., & Ahmad, M. (1978) JACS 100, 7031-7036. Mcintosh, L. P., hand, G., Johnson, P. E., Joshi, M. D., Korner, M., Plesniak, L. A., Ziser, L., Wakarchuk, W. W., & Withers, S. G. (1996) Biochem. 35,9958-9966. Miao, S., McCarter, J. D., Grace, M. E., Grabowski, G. A., Aebersold, R., & Withers, S. G. (1994) JBC 269,10975-10978. Monsan, P., & Paul, F. (1995) FEMS Micro. Rev. 16,187-192. Murray, B. W., Takayama, S., Schultz, J., & Wong, C. H. (1996) Biochem. 35,1118311195.  134  Murray, B. W., Wittmann, V., Burkart, M. D., Hung, S. C, & Wong, C. H. (1997) Biochem. 36, 823-831. Namchuk, M. N., & Withers, S. G. (1995) Biochem. 34,16194-16202. Navaratnam, N., Ward, S., Fisher, C, Kuhn, N. J., Keen, J. N., & Findlay, J. B. C. (1988) Eur. J. Biochem. 171, 623-629. Okada, G., Genghof, D. S., & Hehre, E. J. (1979) Carb. Res. 71, 287-298. O'Keeffe, E. T., Hill, R. L., & Bell J. E. (1980) Biochem. 19,4954-4962. Palcic, M. M., Heerze, L. D., Pierce, M., & Hindsgaul O. (1988) Glycoconj. J. 5,49-63. Palcic, M. M., Venot, A. P., Ratcliffe, R. M., & Hindsgaul, O. (1989) Carb. Res. 190,111. Pan, Y. T., KaushaL G. P., Papandreou, G., Ganem, B., & Elbein, A. D. (1992) JBC 267, 8313-8318. Paquet, M. R., Narasimhan, S., Schachter, H., & Moscarello, M. A. (1984) JBC 259, 4716-4721. Paulson, J. C., & Colley, K. J. (1989) JBC 264, 17615-17618. Permyakov, E. A., Reyzer, I. L., & Berliner, L. J. (1993) J Prot. Chem. 12, 633-638. Powell, J. T., & Brew, K. (1974) Eur. J. Biochem. 48,217-228. Powell J. T., & Brew, K. (1976) Biochem. 15, 3499-3505. Reese, E. T., Parrish, R. W., & Ettlinger, M. (1971) Carb. Res. 18,381-388. Roeser, K. R., & Legler, G. (1981) Biochim. Biophys. Acta 657, 321-333. Saxena, I. M , Brown, R. M., Fevre, M., Geremia, R. A, & Henrissat, B. (1995) J. Bacteriology 177,1419-1424.  Segel I. H. (1993) Enzyme Kinetics: Behavior and Analysis ofRapid Equilibrium and Steady-State Enzyme Systems, Wiley Classics Library Edition ed., John Wiley & Sons, Inc., New York. Sherry, A. D., Newman, A. D., & Gutz, C. G. (1975) Biochem. 14,2191-2196. Sinnott, M. L. (1978) FEBSLett. 94,1-9.  135  Sinnott, M. L. (1979) Chem. Br. 15,293. Sinnott, M. L. (1984) in The Chemistry of Enzyme Action (Page, M. I., Ed.) pp 389-431, Elsevier Science Publishers B.V., Amsterdam, The Netherlands. Sinnott, M. L. (1987) in Enzyme Mechanisms (Page, M. I., & Williams, A., Eds.) pp 259297, Royal Society of Chemistry, London. Sinnott, M. L., & Souchard, I. J. L. (1973) Biochem. J. 133, 89-98. Sinnott, M. L., & Withers, S. G. (1974) Biochem. J. 143,751-762. Srivastava, G., Alton, G., & Hindsgaul, O. (1990) Carb. Res. 207,259-276. Srivastava, G., Hindsgaul, O., & Palcic, M. M. (1993) Carb. Res. 245, 137-144. Steenken, K., & McClelland, R. A. (1989) JACS 111, 4967-4973. Stokes, T. M., & Wilson, I. B. (1972) Biochem. 11,1061-1064. Sun, G., & Budde, R. J. (1997) Biochem. 36, 2139-2146. Svensson, B., & Segaard, M. (1993) J. Biotech. 29, 1-37. Takaoka, Y., Kajimoto, T., & Wong, C. H. (1993) J. Org. Chem. 58,4809-4812. Thiem, J. (1995) FEMS Micro. Rev.16,193-211. Tsopanakis, A. D., & Herries, D. G. (1978) Eur. J. Biochem. 83,179-188. Tull, D., Burgoyne, D. L., Chow, D. T., Withers, S. G., & Aebersold, R. (1996) Anal. Biochem. 234,119-125.  Tull, D., & Withers, S. G. (1994) Biochem. 33, 6363-6370. Tull, D., Withers, S. G., Gilkes, N. R., Kilburn, D. G., Antony, R., Warren, J., & Aebersold, R. (1991) JBC 266,15621-15625. Umezurike, G. M. (1988) Biochem. J. 254, 73-76. Urashima, T., Kusaka, Y., Nakamura, T., Saito, T., Maeda, N., & Messer, M. (1997) Biochim. Biophys. Acta 1334, 247-255.  136 Vrielink, A., Ruger, W., Driessen, H. P. C, & Freemont, P. S. (1994) EMBOJ. 13, 34133422. Wakarchuk, W. W., Cunningham, A., Watson, D. C, & Young, N. M. (1998) Prot. Eng. (in press). Wakarchuk, W. W., Kilburn, D. G., R C Miller Jr., & Warren, R. A. J. (1986) Mol. Gen. Genetics 205,146-152.  Wang, Q., Tull, D., Meinke, M., Gilkes, N. R., Antony, R., Warren, J., Aebersold, R.,& Withers, S. G. (1993) JBC 268,14096-14102. Weber, J. P., & Fink, A. L. (1980) JBC 255, 9030-9032. White, A., Tull, D., Johns, K., Withers, S. G., & Rose, D. R. (1996) Nat. Struct. Biol. 3, 149-154. Withers, S. G., & Aebersold, R. (1995) Prot. Science 4, 361-372. Withers, S. G., Antony, R., Warren, J., Street, I. P., Rupitz, K., Kempton, J. B., & Aebersold, R. (1990) JACS112, 5887-5889. Withers, S. G., & Street, I. P. (1988) JACS 110, 8551-8553. Withers, S. G., Street, I. P., Bird, P., & Dolphin, D. H. (1987) JACS 109, 7530-7531. Wolfrom, M. L., & Thompson, A. (1963) Meth. Carbohyd. Chem., Vol. 2. Wong, C. H., Dumas, D. P., Ichikawa, Y., Koseki, K., Danishefsky, S. J., Weston, B. W., & Lowe, J. B. (1992a) JACS 114, 7322-7324. Wong, C. H., Wang, R., & Ichikawa, Y. (1992b) J Org. Chem. 57,4343-4344. Yadav, S., & Brew, K. (1990) JBC 265,14163-14169. Zervosen, A, & Elling, L. (1996) JACS 118,1836-1840. Zopf, D., & Roth, S. (1996) The Lancet 347,1017-1021. Zor, T., Bar-Yaacov, M., Elgavish, S., Shaanan, B., & Selinger, Z. (1997) Eur. J. Biochem. 249, 330-336.  APPENDIX A  BASIC CONCEPTS OF ENZYME CATALYSIS  138 A.I: Basic Enzyme Kinetics  The concentration of a metabolite within a cell is at a steady state level when it is being produced at the same rate as that at which it is being degraded. This steady state concept can be extended to the concentration of enzyme-bound intermediates in enzyme kinetics. When an enzyme is mixed with a large excess of substrate, there is an initial time period during which the concentration of intermediates builds up to their steady state levels. This initial period is known as the pre-steady state and is characterised by a rapidly increasing reaction rate. After the steady state has been reached, the reaction rate changes little with time. Normally, the rates of enzymic reactions are measured during the steady state. In deriving kinetic equations to describe the steady state, two assumptions are made: (1)  The concentration of enzyme is negligible compared with that of the substrate.  (2)  The initial rate (v ) of product formation is being measured. That is, there have not 0  been significant accumulations of product nor have the substrates been appreciably depleted. Under these conditions, changes in the concentration of reagents are generally linear with time. Consider the Michaelis-Menten mechanism for enzyme-catalysed, single-substrate reactions: E+S  ES  •  E+P  This describes a typical enzyme-catalysed reaction in which the enzyme and substrate first combine to give an enzyme-substrate complex, ES, held together by physical forces alone. This step is assumed to be rapid and reversible. No chemical changes have taken place. It is  139  this concept of an ES complex that forms the foundation of enzyme kinetics. In honour of its introducer, this non-covalently bound complex is often labelled the Michaelis In the second step of this mechanism, a chemical process occurs with a first-order  complex.  rate constant, kcat (also known as the turnover number). This enzymic reaction is described by the basic equation of enzyme kinetics, the Michaelis-Menten equation:  K +[S] m  where [E] represents the total enzyme concentration, [S] is the substrate concentration, kcat is 0  the catalytic rate constant, v is the velocity of the reaction (measured as the rate of substrate utilisation or product formation), and K is called the Michaelis constant. Note that v will m  exhibit saturation kinetics with respect to [S]; when  [S]«Km  v will increase linearly with [S],  but at high substrate concentration v approaches a limiting value of V = kcat[E] . max  f /  v  =  V  m  a  [  S  K  x  m  1  0  v = Vmax = kcattE] /  V  Vmsw  2  [S]  Figure A-1: Velocity versus substrate concentration for a typical enzymatic reaction K  m  is defined as that substrate concentration for which v = '/Wmax- The value of K m  can be treated as an apparent dissociation constant for all enzyme bound species. That is,  140 rFirsi  = K . Thus, the value of Km may be used to indicate the stability of the m  S[ES] bound enzyme-substrate complex. Using this measure, a substrate with a low Km is said to bind tightly to the enzyme. The catalytic efficiency of an enzyme can be measured at low substrate concentrations since the Michaelis-Menten equation reduces to v = ——[E][S] 0  when [S]«Km. At very  low [S], most of the enzyme is unbound so that [E] «[E], thus: 0  v = ^L[E][S]  In this form, the term kcat/Km is the apparent second-order rate constant that relates to the properties and reactions of thefreeenzyme and free substrate. The importance of k^t/Km is that it relates the reaction rate to the concentration offree,rather than total, enzyme. This result has been shown to hold at any substrate concentration. Historically, the Michaelis-Menten equation has found greatest application to experimental analysis in its reciprocal form. By inverting both sides of the equation, a linear relationship between 1/v and 1/[S] is described: 1, v  1  ,  K  m  V^JS]  Graphing 1/v versus 1/[S] results in a straight line with y-intercept equal to 1/V x, xma  intercept equal to -1/Km, and slope of K /V ax. This Lineweaver-Burk plot is useful for m  m  analysing data graphically and detecting deviationsfromideal behaviour. A disadvantage  141  1/v  /  -1/Km  /  *  Slope = Km/Vma>  1 A/max  0  1/[S]  Figure A-2: A typical Lineweaver-Burkplot  of this type of analysis is that errors in measurements taken at lower substrate concentrations are amplified relative to those at higher substrate concentration. The overall effect is to place greatest weight on those values obtainedfromreactions using low substrate concentrations. With the advent of powerful computer-based analysis packages, the data can now be fit directly to determine the values of the kinetic parameters. The utility of the LineweaverBurk plot is now more qualitative than quantitative. Many of the principles developed for single-substrate systems can be extended to multi-substrate systems. In general, the solutions to these equations are complicated but can be dramatically simplified if the concentration of one substrate is held constant while the other is varied. Under these conditions, most reactions obey Michaelis-Menten kinetics. It is prudent to remember, however, that the Km for a particular substrate at one fixed set of cosubstrate concentrations may not be the actual Km. In reality, this is an apparent value that changes as co-substrate concentrations are modified. Similarly, observed values of V x for a ma  preparation with a saturating concentration of one substrate may not be the same as V observed when another substrate is saturating. The true  Km is  max  measured when all other  142 substrates are saturating and the true V  max  when all substrates are saturating. It is  necessary to distinguish between the equation and mechanism proposed by MichaelisMenten. Their general equation holds for many mechanisms, but their mechanism is often an over-simplification of the overall catalytic event. A.ll: Deriving Steady-State Velocity Equations  The basic premise of the King-Altman Method for deriving steady-state velocity equations can be illustrated using a simple Uni Uni reaction: E+A  ^  kj  EA  k  2  EP ^  k . ^ 3  E+P  The velocity equation is obtained as follows: (1)  The different enzyme species are arranged into a simple geometric figure with each species at one corner of the figure. For the Uni Uni reaction shown above, there are three enzyme species, E, EA, and EP. Thus, the basic figure is a triangle:  The steps by which two species are interconverted are represented by lines connecting the two species. Each line is labelled with the appropriate rate constant or rate constant x ligand concentration for each direction. All the  143 ways in which any one species can be formed are, in this way, represented by the lines leading to that species. (2)  All possible patterns containing one less line than the number of enzymes species are drawn. In a basic figure with three corners, the possible two-line patterns are:  0  (5)  «  These patterns represent all the ways any one enzyme species can be formedfromthe others by paths that do not involve closed loops. The total number of interconversion patterns is given by: m! # of patterns with (n -1) lines = (n-l)!(m-n + l)!  where n  = the number of enzyme species (and the number of corners in the basic figure).  m  = the number of lines in the basic figure.  (3)  The relative proportion of the total enzyme, [E] , represented by any species, T  [EX], is given by:  [EX] I z terms [E] a denominator T  where z = the number of acceptable (n-1) lined patterns as defined above.  144 The denominator is the same for all enzyme species and represents the sum of all numerator terms for all enzyme species. The numerator term for any one interconversion pattern is the product of all rate constants and associated ligand concentrations read along the lines leading to the species in question. This is illustrated below for the Uni Uni reaction:  A < a  [E]  +k.ik +k k 3  2  =  3  denominator  T  [EA] k. k [A] + k k [A]+k. [P]k = [E] denominator 2  ,  [E]  <C ^  A 1  3  1  3  2  , and  T  [EP] _ kJA]k +k, k, [P]+k k, [P] [E] denominator 2  1  3  2  3  T  The denominator is the sum of all numerator terms. Thus, grouping similar terms: denominator = [E]T= (k . k .1 + k .ik 3 + k 2k 3) + k 1 (k .2 + k + k 3)[A] 2  2  + k_3(k., + k.2 + k )[P] 2  (4)  The steady-state net velocity is given by the difference between the forward and reverse velocities of any step. For example: v = k [EA]-k_ [EP] 2  2  Thus, using the expressions we have developed for the concentration of enzyme species, we get:  v  _k (k, k [A] + k k [A]+k, k [P])[E] -k, (k kJA] + k k, [P]+k k [P])^ denominator 2  2  1  3  1  3  2  T  2  2  1  3  3  2  145 or v _ [E]  T  k k k [A]-k. k. k. [P] 1  2  3  1  2  3  (k ,k. +k.,k +k k ) + k (k. +k +k )[A]+k ( k , +k. +k )[P] 2  3  2  3  1  2  2  3  3  2  I  2  Equation I can be written in the coefficient form of Cleland (as described by Segal, 1993): v=  num [A]-num [P] ]  2  const + Coef [A]+Coef [P] A  II  P  where numi =the coefficient of the positive numerator term; the product of [ E ] and T  the rate constants that multiply [A] in the numerator nuni2  =the coefficient of the negative numerator term; the product of [E]T and the rate constants that multiply by [P] in the numerator  const =the constant term in the denominator; the group of rate constants in the denominator which is not associated with any ligand concentration term CoefA^the coefficient of [A] in the denominator CoefB=the coefficient of [P] in the denominator (5)  The final velocity equation is obtained by redefining the coefficients of equation II as kinetic constants that can be experimentally determined. This is accomplished  using  the  following  definitions:  V,  num Coef ' A  const Coef ' A  V.max, r  num Coefp ' 2  const Coefp  Also,  146 The numerator and denominator of II are multiplied by numi/numa and the above substitutions are made to yield: Pi'  (  V  ,V m ma ax x,, r1 ma ax x,, fI m  III  V =  K * V _ + V _max,rL^J-f ,[A] + r  ~  When [P] = 0, equation III reduces to the usual Michaelis-Menten equation for the forward velocity. When [A] = 0, we expect equation III to reduce to the Michaelis-Menten equation for the velocity in the reverse direction with P as the substrate: - V V =•  [PI max, r I J  K +[P] m  However, it actually becomes: - V  v= tn max,r V  [P]  ,V max,! * max,r K  —V TPI rl J T^T = —:—• m^ax, '-"  eq  v  Vmax,fP3 v K  eq  >  m  r  r  J  K^V^jKgq — r L"J  K V K A  r-  therefore,  i v  m m max,r^eq v  „ p ^ =K  It is evident from the complexity of the equations involved that even a simple Uni Uni reaction yields a complex rate equation. A.ll.i Random Bi Bi Sequential Reaction Mechanism  If two substrates, AX and B, bind randomly to an enzyme and the binding of one substrate changes the dissociation constant of the other substrate by a factor a, the system can be described by the equilibria shown below:  147 K  E + AX  AX  ^  w  BX  B  E«AX i  aK  )  B  E» A - -»  E«AX«B =S=S= E»A»BX(k ) cat  E  + B  — *-  E«B  aK  A  ^E«BX  AX  E+ A  -»  E + BX  In this reaction scheme, the complexes E»AX, E»B, and E«AX»B are called transitory complexes. That is, they undergo unimolecular reactions in which either a substrate or product is released (or are capable of isomerising into an intermediate capable of doing this). Additionally, E«AX»B is also known as a central complex. This means that, although it is a transitory complex, it cannot participate in a bimolecular reaction with another ligand because all of its binding sites are occupied. Enzyme forms which cannot undergo a unimolecular dissociation reaction (or isomerise to one which can) are called stable enzyme forms. In the reactions described above, free enzyme (E) is the only stable enzyme form In general, sequential reactions only have one stable enzyme form whereas ping pong reactions will have two or more. In the case of a random sequential mechanism such as this one, the overall equilibrium between E«AX»B and E must be path independent. The rate equations for this reaction mechanism can be derived as follows: [E.AX]  [E«B] v [E]  T  [E«AX«B]  [E»AX«B]  k^tE.AX.B] [E] + [ E * A X ] +  [E«B]+[E«AX«B]  Where kcat is the rate determining step and occurs after all substrates have bound to the enzyme. Thus, V  max  = k c a t [ E ] i and  148 [E»AX«B] [E]+[E.AX] + [E.B] + [E.AX.B] which can be expressed in terms of substrate concentration and dissociation constants as, [B][AX] v V_  t  , [AX] [B] [B][AX] K a K^K |  i  B  8  By fixing the concentration of one substrate (B, for example) and varying the other (AX), this equation is arranged to reflect the effect of substrate B on the kinetic constants of the rate equation as they relate to AX. V^JAXJtB] v= • aK^[B] + aK^K +aK [AX] + [AX][B] B  V^tAX] ^(  B  a K  <  aK^  IV  + [AX| 1 + 1+ [B] [B]  This equation now resembles the familiar form of the Michaelis-Menten equation except that in this situation the Michaelis constants can be described as follows:  V (app) = nHX  aK  1+  and K (app) = oKA X m  [B]  KB A 1+ [B] ct K 1+ [B]  The resulting equations are symmetrical to those in which [AX] is fixed. In reciprocal form, the equation becomes: 1 a KA X  K 1 1 1+ -+[B] [AX] V n  aK  1+  [B]  If a<l, the binding of one ligand increases the affinity of the enzyme towards the other ligand and K a  P P  for the varied ligand decreases as the concentration of the fixed ligand  increases. Graphically, this outcome is represented in Figure A-3 where the family of  149 reciprocal plots for varied [AX] intersect at a point above the l/[AX]-axis and pivot clockwise about this point. 4  ,[B] = 0.4K^, [B] = 0.6K  i-  [B] = 1.2K  8  I  [B] = oo  h  2  Figure A-3: Double-reciprocal plotfor a random bi bi sequential reaction with a<l  From 1/v versus 1/[AX], Vmax and K** can be determined: Slope  y  /AX  =——  AX  1  K 1+ [B]  V  •+-  max  J  [B]  Similarly, the plot of 1/v versus 1/[B] provides values for V K 1+  Slope = y  7B  Vmax  AX \  The slope replots (Figure A-4, A) also give values for 1/Vmax  versus 1/[B] will give V  relationship:  m  a  x  ,  x  V  n  and K since, B  1 aK -+ [AX]  [AX]J  \  m a  aK**  M  and K . In addition, replots of B  K**, and aK (Figure A-4, B) based on the following B  A.ll.ii Ordered Bi Bi Sequential Reaction Mechanism  An ordered reaction that is bireactant in both directions has the general form shown below: B  k, E  + AX  •» -  BX  ( k E»AX  k.i  k  2  i.  k.  "  p  (E«AX»B  2  k  "  k.  E'A'BX)  j  J  - ^  k.  p  3  k E »A  4  •* *  E + A  k.4  In this type of reaction, one substrate (AX in this example) must bind to the enzyme before the second substrate (in this case, B) can bind to form the ternary intermediate. If the equilibrium between the central complexes of E»AX«B and E»A«BX is the rate limiting step in catalysis, then E, AX, B and E«AX«B are all in equilibrium and the velocity of the reaction is given by: y  =  VrcaJAXltB]  V  K^K* + Q A X ] + K f [B] + [AX][B]  J  151 where K** = dissociation constant of A X and K^f = the Michaelis constant for A X  In the situation described above, the  (Vi V m a x ) .  equation  [B]  term is absentfromthe rate  since k3 is very small compared to the other rate constants and  Coef  k2k k 3  n  4  k,k (k +k )  CoefAB  2  4  reduces to zero. But, if the equilibrium between E«AX«B  3  and E«A«BX is not rate limiting, steady state assumptions must be used to derive the rate equation. To determine the correct rate equation in this way, a King-Altman figure with four corners can be constructed:  0-  ki[AX]  E*AX  k^[BX]  k [B] 2  k. [A] 3  E»BX  =^ (E*AX»B =^=^ E»A«Bx)  Figure A-5: Basic King-Altman Figure for an ordered bi bi sequential reaction  The distribution of enzyme species is obtainedfromfour three-lined conversion patterns such that:  [E] = k k.ik.2 + k. [A]k.ik.2 + k2[B]k k + k.!k k , 4  3  3  4  3  4  [E.AXl^k^itAXJk.z + kjtAXlkotAJk.z + k [BX]k .3 [A]k . +k k ki[AX], A  [E«AX«B + E«A«BX] =  2  3  4  k k 1 [AX]k [B] + k 1 [AX]k [B]k .3 [A] + 4  2  2  k [BX]k .3 [A]k 2 [B] + k k ^[BX] k .3 [A] and, . 4  . 1  [E«BX]=k.2k. k^[BX]+k [AX]k [B]k + k [BX]k [B]k + k.ik^[BX]k . 1  1  2  3  A  2  3  3  152  By assuming a steady-state exists, it is evident that v = k i[AX][E] - k _i[E»AX] (or any other difference between the forward and reverse velocities of a given step). Thus, we obtain the following complex rate expression after grouping the denominator terms: v [E]  =  T  k k k k [AX][B]-k, k, k. k, [A][BX] k ,k (k. + k ) + k,k (k. +k )[AX] + k k k [B] + k. k k. [A] + k ,k (k +k )[BX] + k k (k +k )[AX][B] + k k k [AX][A] + k k^(k. +k. )[A][BX] + k k k [B][BX]+k k k. [AX][B][A] + k k. k. [B][A][BX] 1  4  2  4  2  3  2  3  3  4  1  4  3  1  3  2  2  ]  2  2  3  4  3  2  2  3  2  3  4  1  4  3  1  4  1  2  2  2  3  3  3  4  In the absence of products, those terms derivingfromthe reverse reaction can be discounted and all product terms (those containing [A] and [BX]) removed. This is the situation when initial velocities are measured. With some manipulation, the above expression can then be simplified to the following:  y-i^qw  o r  K " K ° + K » [ A X ] + K^.[B]+[AX][B]  1 KAX 1 1 1+ v V„ K f [B] [AX] V„ m  where  vn  1 + —2[B]  is the dissociation constant for the E»AX complex and K^f is the concentration  of AX that results in an observed velocity of V 2 V  max  .  Notice that equations V and VII are of  the same general form. Because of this similarity, the graphical representation of datafroma random bi bi sequential reaction and an ordered bi bi sequential reaction are indistinguishable (see Figures A-4A,B). Hence, it is impossible to distinguish between these two mechanisms on the basis of reciprocal plots alone. Isotope incorporation and product inhibition studies are required to differentiate between them. However, if K** « K  x m  for an ordered bi bi  153  reaction, the slopes of the plots become insensitive to changes in the concentration of the co-substrate (B) and the family of plots will be essentially parallel. A.H.iii Ping Pong Bi Bi Reaction Mechanism  If a product is released between the addition of two substrates, the reaction operates by a ping pong bi bi mechanism: A  E + A X ^5=  =(E»AX=  B  E X - A W ==== E X ^ =  ;  (EX«B=^==^ B»BX)^^  4  "  E + BX  With this type of mechanism, two stable enzyme species exist, E and EX. Ping pong mechanisms are common in group transfer reactions. In this situation, group X is transferred from the first-binding substrate (AX) to the enzyme. After the newly modified substrate dissociates as the first product of the reaction, the second substrate (B) binds and the enzyme transfers X to this substrate, thereby forming the final product, BX. Using steady-state assumptions to derive the rate equation in the same manner that was used for the ordered bi bi sequential reaction in the previous example, the following velocity equation for a ping pong reaction is obtained: V.tAXlIB]  v =K• ^ [ A X ] + K f [ B M A X P ] VIII and at a fixed [B], v=•  V^JAX]  K Kf+tAXj 1 +  and  f J^AX "\ 1 1 —= •+ v V ^max j [AX] V  1+  n  m  [B]  [B]  Thus, the slope of the double-reciprocal plot will be independent of [B] but V  max  will not.  The consequence of these effects is illustrated in the graphical representation shown below:  154  t[B]  5 1h 0 1/[AX]  Figure A-6: Double-reciprocal plots for a ping pong reaction  The family of reciprocal plots shows a series of parallel lines. Again, a replot of the data can be madefromwhich the kinetic constants may be extracted since 1  V^Capp)  [B] + : V^max J  1  and  1  KfCapp)  m Tf AX  V m K  J  1 1 AX [B]• +K m  Thus, K for AX and B , along with Vmax, can be determinedfromthe replot of m  1/Vmax  (app) vs. 1/[B] and }/  / A vAXf X  1/[B]  (app)  vs. 1/[B] (Figure A-7A,B)  ^  1/IB]  Figure A-7: Replot graphs: (A) intercept replot (B) replot of apparent Michaelis constants  155 Although it would appear to be a simple task to differentiate between a sequential and a ping pong mechanism, there are several cases in which an ordered bi bi sequential reaction might produce a series of parallel lines when data are expressed as a double-reciprocal plot: (1)  K * * «  -  K £ *  As mentioned previously, this situation results in slope insensitivity to changes in substrate B. The enzyme has a very high affinity for AX but because of the relative values of the other rate constants,  (2)  » K**.  Rapid Random Equilibrium -  systems where a » 1 so that the binding of one substrate strongly inhibits binding of the other.  (3)  Non-rapid Random Equilibrium -  systems in which the rate constants for the release of [AX] and [B] are much lower than Vmax for the forward reaction.  -  parallel lines are obtainedfromreciprocal plots if [AX] and [B] are in the region of K m for both substrates.  Thus, a full kinetic analysis of a complex, multi-substrate system requires additional study to support any proposed mechanism The use of product inhibition patterns, isotope exchange information, or the identification of an enzyme-substrate intermediate (in the case of a ping pong mechanism) is required to substantiate these kinetic results. A.lll: Enzyme Inhibition  A decrease in enzyme activity, inhibition, can resultfromthe binding of an inhibitor molecule to the enzyme. Irreversible inhibition resultsfromthe formation of a covalent bond  156  between the enzyme and the inhibitor. When a non-covalent bond forms between the inhibitor and the enzyme, the inhibition is reversible and can be classified based on the nature of the interaction. A.lll.i: Competitive Inhibition  When an inhibitor (I) binds reversibly to the active site of thefreeenzyme (E), it prevents substrate (S)frombinding and vice versa. Because I and S must compete for the active site, I is said to be a competitive inhibitor. Thus, in the simple Michaelis-Menten mechanism (see A.I, where K = K ), an additional equilibrium must be considered: m  E + S  s  ^  w  ES  E + P  + I  EI  Solving the equiUbrium and rate equations using [E] = [ES] + [EI] + [E] gives, 0  [EUS]!^  v= D5] + K (l + % ) ' m  K  m  is apparently increased by a factor of (1+fTJ/Ki) yet Vmax is unaffected. This equation  holds for all mechanisms obeying the Michaelis-Menten equation (Fersht, 1 9 8 5 ) .  157 A.lll.ii: Non-competitive Inhibition  An inhibitor (I) may bind to enzyme (E) at the same time as substrate (S) binds to the enzyme. In this situation, the substrate and inhibitor are not competing for the same site: Km  E+ S  v  ES  +  + '  I  I  cat  E+ P  K,  K'  EI + S  where  K i  = K'I, K  m  = K'  m  ,  ESI  k',cat  E + P+ I  and k'cat = 0. Because S and I bind independently, the affinity of  each for either free enzyme or the ES complex is unchanged. Thus, [E] [S] V 0  k  [S] + K  and the inhibitor is said to be non-competitive. the enzyme is unchanged. V  max  m  Km  is unaffected since substrate affinity for  decreases by a factor of (1+P]/Kj).  A.lll.iii: Mixed Inhibition  The most complex form of inhibition, mixed inhibition, occurs when the inhibitor (I) binds to both free enzyme (E) and the enzyme-substrate complex (ES) with different affinities. In addition, ES or ESI may be active but with different kcat values. This situation is more common than that seen with non-competitive inhibition. The mechanistic scheme is  158 the same except that K j *  K\ K  m  * K'  m  ,  and k'cat * 0 or k^t. In this case,  Vmax,  a p P  .  and Km, . are complex functions of [I] and the related constants. app  A.lll.iv: Uncompetitive Inhibition  If the inhibitor (I) only binds to the E S complex, not tofreeenzyme ( E ) , it is an uncompetitive inhibitor. In this situation, inhibition can only occur in the presence of the  substrate. The mechanism is illustrated as: K  E +S  kgat  m  -  ES  •  E+ P  + I Ki  ESI where K , = 4  ™ and v =  [rEnSnI n ]  [S] + K  m  /  /(l+m/K^  Both K and Vmax are decreased by a factor (1+[I]/Ki). m  V  m  a  x  decreases because the substrate  is unable to compete to saturate the active site of the enzyme. The decrease in K m occurs because the inactivation of the E S complex as the E S I complex by inhibitor gives the appearance of better apparent binding of the substrate and subsequently, a lower K m . A.IV : Kinetics of Inactivation for lgtC-19  The inactivation of a glycosyl transferase such as lgtC-19 is a two-step process that can be described as follows:  159 E +I where ki=  ^  Ki  kj  E«I  w  •  E-I  the rate constant of inactivation  Ki =  the dissociation constant of enzyme-bound inactivator  E = lgtC-19 I =  irreversible inhibitor  First, inactivator binds to the enzyme in a reversible manner. Next, there is an irreversible bond formed between the enzyme and inactivator to produce an inactivated enzyme intermediate (E-I) in the rate-determining step. The concentration of inactivator remains essentially constant during the reaction if the initial concentration is much greater than that of the enzyme (I»[E] ). The kinetics, therefore, are pseudo-first-order with respect to enzyme 0  concentration. The equation for the inactivation can be written as:  v  _k [E] lTJ 1  0  Kt+rT] where v = the rate of inactivation. ki = the rate constant of inactivation. Ki = the apparent dissociation constant for all species of enzyme-bound inactivator, given by K = [E]P]£[EI] t  Since [I] is assumed to be constant over the course of the inactivation reaction, v = kobs[E] where, T  k^.jya.vi ohs  Kj+[I]  160 The value of k o b (the pseudo first-order rate constant of inactivation) at each S  inactivator concentration can be determined by fitting the velocity data to a standard exponential decay function (VII) using GraFit 3.0 (Leatherbarrow, 1990). Since each molecule of A released corresponds to one molecule of enzyme inactivated, v =  _aTTJ _aTE dt dt  rg -(kobs)t ]oe  =  (  v  n  k =  k  o  t  e  [  E  ]  T  k ^ Z ^ E k [E]  ^  T  t h e r e f o r e 5  l n ^ L = -k  [E]  o b s  t and [E]T =  0  )  The inactivation parameters  ki  and  K i  can be determined by substituting kobs values into VI.  161  APPENDIX B  GRAPHICAL REPRESENTATION OF KINETIC DATA  162 A.V : Lineweaver-Burk Plots for the Transfer of Galactose from UDPgalactose to Lactose at 30°C, pH 7.5.  £  §  ro  1/[lactose] (mM)-'  [Enzyme]=0.00197 mg/ml, 1=340 nm, Ae=6.22 mM~ cm~ 1  -40  -20  0 20 1/[UDP^at] (mM)-  40  1  60  1  -1 [Enzyme]=0.000229 mg/ml, X=340 nm, As=6.22 mNT cm 1  0.04  -0.2  -0.1  0 0.1 1/[MnCIJ (mM)"  0.2  1  [Enzyme]=0.000213 mg/ml, A=340 nm, As=6.22 ml\/r cm' 1  1  163 B.I: Graphical Analysis of the Mechanism of Galactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 7.5 (in HEPES buffer) using Cleland's Method for a Bi Bi Sequential Reaction  2500  a /  0/  2000  /S  1500  ° slop =  1000  e  max  500 0  — v ^ t a ^ m , lactose  / I  -500 3 3  O O  CD  O  p  ro  03  1/[lactose] (mM)-  1  o  I  I  o  o  i  i  i  1/[lactose] (mM)-  1  Replots of Data from [UDP-galactose]/v vs. [UDP-galactose]  164  10  S  8 h  o  1.022 mM UDP-gal  h  •  0.1022 mM UDP-gal  •  0.0511 mM UDP-gal  •  0.0383 mM UDP-gal  A  0.0256 mM UDP-gal  A  0.01022 mM UDP-gal  6  4 h 2  h  -25  25  50  75  100  125 150  [lactose] (mM)  [Enzyme]=0.000459 mg/ml, 1=340 nm, As=6.22 mM~ cm~ 1  1/[UDP-gal] (mM)  1  1  1/[UDPgal] (mM)  Replots of Data from [lactose]/v vs. [lactose]  1  165 B.ll: Graphical Analysis of the Mechanism of Galactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 6.0 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  Replots from Lineweaver-Burk Plots  166 B.lll: Graphical Analysis of the Mechanism of Galactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 6.5 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  -200  0  200  400  1/[UDPgal] (mM)"  1  [Enzyme]=0.000595 mg/ml, A=340 nm, As=6.22 mM~ cm 1  J  o 8  °  i o 2  i  i i i i i i i o o o § i5 S ^ 1/Ilactose] (mM)-  I  i  i  i  1  i  i  I  o  0  £  °  o  o o 2 .8 fS S 1/[lactose] (mM)  Replots from Lineweaver-Burk Plots  1  .  1  I  167 B.IV : Graphical Analysis of the Mechanism of G alactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 7.0 (in MES buffer) using Cleland's Method for a Ping Pong Reaction  Replots from Lineweaver-Burk Plots  168  B.V : Graphical Analysis of the Mechanism of G alactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 7.0 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  Replots from Lineweaver-Burk Plots  169 B.VI: Graphical Analysis of the Mechanism of Galactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 7.5 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  Replots from Lineweaver-Burk Plots  170  B.VII: Graphical Analysis of the Mechanism of Galactosyl Transfer from UDP-galactose to Lactose by LgtC-19 at 30°C, pH 7.7 (in HEPES buffer) using Cleland's Method for a Ping Pong Reaction  -100  0  O  125 mM lactose  •  50 mM lactose  •  20 mM lactose  •  10 mM lactose  A  5 mM lactose  100  1/[UDP-gal] (mM)"  1  [Enzyme]=0.000595 mg/ml, A=340 nm, As=6.22 mM~ cm~ 1  Replots from Lineweaver-Burk Plots  1  171  B.VIII: Effect of MnCI Concentration on the Efficiency of Fluoride Ion 2  Electrode Measurements of F" Concentration at 30°C, pH 7.0  1.1 Aooooooooooo O O O O O O O O O O O O O O O O O  §- 0.9 E.  ••••••••••••••••••DnrjOODDDOOOO  o  control  •  1mM MnCI2  •  5mM MnCI2  •  10mM MnCI2  \L 0.8  20mM MnCI2 A^/Sl\  A A A A A A  ^/VAAAAAAAAAAAAAAAAAA  30mM MnCI2  •  0.7 0.6  j  L  J  200  L  300  400  time (s)  Measured Fluoride Ion Concentration as a function of Time for a Series of 1mM Standard Solutions Containing Variable Concentrations ofMnCI 2  10  20  30  40  [Mn ] (mM) 2t  Measured Fluoride Ion Concentration as a Function ofMnCh  172  B.IX : Mass Spectra of Reaction Mixtures Conta ining FCHASE-lactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  to 1.4  1  to 1.4 1.0M  FCHASE-lactose standard  M  FCHASE-lactose standard  I  t.OM  f  •«  4 M  «u.a  •1 1 . u  ,f  1•00  '*?•»  TOO  000  000  1000  1  1100  1 1  1  «...  •00  •00  TOO  1000  1101  M i.4 •  {  (-) CONTROL •4*4 •  1 too  •00  TOO  000 mfl witv  »00  1000  Ul J  . . J . .•00» v 1W O  1100  TOO  ooo mv  mn,  .1. too  10JT.2  11U.4  1000  11W  104 7.0  t.t*B l.teS  5 1  (+) CONTROL  {  i.Oaft  •0*0  TEST REACTION  4.0M  • 0*4  1  4 7? .2  J., 1 . •00  .!., •00  » » TOO  T 100 mfi. Mnu  DONOR: UDP-galactose  fOO  t.o*s  «... 1000  1100  •T»4 •00  TOO  000  000  DONOR: a-galactosyl fluoride  1000  ">« « 7  1100  173  B.X : Mass Spectra of Reaction Mixtures Conta ining FCHASE-galactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  1 5.  "  l.t  FCHASE-galactose standard  F C H A S E - g a l a c t o s e standard •  4H<  M M UI.l MO  110.4 tOO  MO  TOO  1£M  **».0 10M  11U.1  i  1101  «M.4  Uft  MO  [ l  111*1  •11.4  MS 4 MO  IM  I.I  T4S  MO  101  Lt  I M |  rOM  & «.0M  (-) C O N T R O L  0.0M  u  4.0*1  1.0M  1.0»B  ft 4J*t  473.3 800  »M.O 000  .1 700  •00 MO. M  * * .2  4« 1.2  M i l  M .0 000  1000  4M.4  1118.0  2.MS 2.1«5  ft mmwA%  U4.t  •M.4  ihi^ •00  •In i 1  000  DONOR:  700  TOO  •  1 MO  UDP-galactose  • 0  1M7.4  1000  1100  f  tTM  *  MOB'  4S. 1.3  TEST REACTION  I.OM | 000  0*3.0  M M  •00 mn. mm  M ..4  (+) C O N T R O L  7.0»4  000  M 4.M9  g  .1 •00  1100  S.MS  f  (-) C O N T R O L  743.3  1004.4 10OC  1100  1  1 8 1 2  MO  « 0  700  M4.4  MO mft, rnnu  D O N O R : a-galactosyl  , I tO0  fluoride  100,3.0 1000  1100  174 B.XI: Mass Spectra of Reaction Mixtures Containing FITC-lactose as the Acceptor Substrate for Galactosyl Transfer by LgtC-19  100  000  TOO  MO  1OO0  ' H  1100 MO  000  ,  700  t  UJ 000  •00  Mt 1 10M  11^0 • 1100  ar 1.1 l.taft  (-) CONTROL  I.MS  H1.1  DONOR: UDP-galactose  1  *"•».»!.<  DONOR: a-galactosyl fluoride  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items