Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The structure and mechanism of UDP-glucose dehydrogenase Campbell, Robert E. 2000

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

Item Metadata

Download

Media
831-ubc_2000-56519X.pdf [ 14.72MB ]
Metadata
JSON: 831-1.0061448.json
JSON-LD: 831-1.0061448-ld.json
RDF/XML (Pretty): 831-1.0061448-rdf.xml
RDF/JSON: 831-1.0061448-rdf.json
Turtle: 831-1.0061448-turtle.txt
N-Triples: 831-1.0061448-rdf-ntriples.txt
Original Record: 831-1.0061448-source.json
Full Text
831-1.0061448-fulltext.txt
Citation
831-1.0061448.ris

Full Text

THE STRUCTURE AND MECHANISM OF UDP-GLUCOSE  DEHYDROGENASE  Robert E. Campbell B.Sc, University of British Columbia, 1994 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Chemistry  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA July, 2000  © Robert Earl Campbell, 2000  ln  presenting  degree freely  at  the  available  copying  of  department publication  this  of  in  partial  fulfilment  University  of  British  Columbia,  for  this or  thesis  reference  thesis by  this  for  his thesis  and  scholarly  or for  her  of  The University of British Columbia Vancouver, Canada  Date  DE-6  (2/88)  I  I  further  purposes  gain  the  shall  requirements  agree  that  agree  may  representatives.  financial  permission.  Department  study.  of  be  It not  is  that  the  Library  an  granted  by  allowed  the  advanced  shall  permission for  understood be  for  make  extensive  head  that  without  it  of  copying my  my or  written  11  ABSTRACT  UDP-glucose dehydrogenase (UDPGlcDH) catalyses the two-fold oxidation of UDPglucose to UDP-glucuronic acid with sequential hydride transfers to two equivalents of NAD . +  Bacterial UDPGlcDH is essential for formation of the antiphagocytic capsule that protects many virulent bacteria such as Streptococcus pyogenes and Streptococcus pneumoniae type 3fromthe host's immune system. We have carried out a variety of biochemical and X-ray crystallographic studies of UDPGlcDH and have proposed the first detailed and consistent mechanism for this enzyme. UDPGlcDH operates by a Bi Uni Uni Bi ping pong kinetic mechanism in which UDPglucose is bound first by the enzyme and UDP-glucuronic acid is released last. In the first step of the enzyme mechanism, the pro-R hydride of UDP-glucose is transferred to NAD to produce +  the aldehyde intermediate and NADH. The oxidation of the aldehyde intermediate to the carboxylic acid product is initiated by nucleophilic attack of the thiol of Cys 260 to yield a thiohemiacetal. Collapse of the thiohemiacetal intermediate with a second hydride transfer to NAD produces a thioester intermediate that is hydrolyzed in the final step of the mechanism. +  The putative aldehyde intermediate has been chemically synthesized and demonstrated to be kinetically competent to serve as an intermediate in the enzymatic reaction. Evidence for transfer of the pro-R hydride occurring in the first oxidation comes from the observation that the UDPglucose analogue, UDP-6S-6C-methylglucose, is oxidized by UDPGlcDH to the corresponding ketone product but the 6R epimer is not. The involvement of a nucleophilic cysteine thiol is supported by covalent labeling studies with the affinity label, UDP-chloroacetol phosphate. The X-ray structures of wild-type UDPGlcDH and Cys260Ser UDPGlcDH have been determined in ternary complexes with UDP-xylose/NAD and UDP-glucuronic acid/NAD(H) respectively. The +  402 residue homodimeric UDPGlcDH is composed of an N-terminal NAD dinucleotide binding +  domain and a C-terminal UDP-sugar binding domain connected by a long central a-helix. The  Ill  first 290 residues of UDPGlcDH share structural homology with 6-phosphogluconate dehydrogenase, including conservation of an active site lysine and asparagine that are implicated in the enzyme mechanism. Also proposed to participate in the catalytic mechanism are a threonine and a glutamate that hydrogen bond to a conserved active site water molecule suitably positioned for general acid/base catalysis.  iv TABLE OF  CONTENTS  ABSTRACT  ii  TABLE OF CONTENTS  ,  iv  LIST OF FIGURES  viii  LIST OF T A B L E S  xii  ABBREVIATIONS AND SYMBOLS  xiii  ACKNOWLEDGMENTS  xvii  DEDICATION  xviii  C H A P T E R 1: I N T R O D U C T I O N  1  1.1 NAD -DEPENDENT OXIDATION OF ALCOHOLS  3  +  1.1.1 'Short-chain' and 'medium-chain' alcohol dehydrogenase (AlcDH)  3  1.1.2 a-Hydroxy acid dehydrogenase 1.1.3fi-Hydroxyacid dehydrogenase 1.1.4 Aldo-keto reductase  6 •  9 11  1.1.5 Dismutations and alternate reactions catalyzed by AlcDH 1.2 NAD -DEPENDENT OXIDATION OF ALDEHYDES +  12 14  1.2.1 Glyceraldehyde 3-phosphate dehydrogenase  14  1.2.2 Aldehyde dehydrogenase (AldDH)  16  1.3 OTHER NAD -DEPENDENT TWO-FOLD OXIDATIONS +  1.3.1 Histidinol Dehydrogenase 1.3.2 HMG-CoA reductase 1.4 UDP-GLUCOSE DEHYDROGENASE 1.4.1 Biological role of UDPGlcDH.  18 19 21 23 23  1.4.2 Previous Studies on UDPGlcDH  24  1.5 THE PROPOSED MECHANISM OF UDPGlcDH  27  1.6 OTHER RESEARCH IN THIS LABORATORY  28  V C H A P T E R 2: P U R I F I C A T I O N A N D K I N E T I C C H A R A C T E R I Z A T I O N O F U D P G l c D H  31  2.1 INTRODUCTION  32  2.2 MATERIALS AND METHODS  34  2.2.1 General Procedures  34  2.2.2 Specific Procedures  35  2.3 RESULTS  39  2.3.1 Purification of UDPGlcDH  :  2.3.2 Molecular weight determination  39 40  2.3.3 Effect of additives on the activity of UDPGlcDH  42  2.3.4 Test for tightly bound NAD or NADH  43  2.3.5 Initial velocity kinetic studies  45  2.3.6 Product inhibition kinetic studies  47  +  2.3.7 Inhibition by UDP-xylose 2.4 DISCUSSION  2.4.1 Characterization of UDPGlcDH  50 51  51  2.4.2 The kinetic mechanism of UDPGlcDH C H A P T E R 3: S Y N T H E S I S A N D E V A L U A T I O N O F M E C H A N I S T I C P R O B E S O F U D P G l c D H  54 59  3.1 INTRODUCTION  60  3.2 MATERIALS AND METHODS  68  3.2.1 General Synthetic Methods  68  3.2.2 Routine Synthetic Procedures  70  3.2.3 Specific Synthetic Procedures  73  3.2.4 General Enzymatic Methods  84  3.2.5 Specific Enzymatic Methods  85  3.3 RESULTS  89  3.3.1 Synthesis of uridine 5'-(a-D-gluco-hexodialdo-l ,5-pyranosyl diphosphate) (1)  89  3.3.2 Synthesis ofUDP-(6R and 6S)-6C-methylglucose (2a and 2b)  93  3.3.3 Synthesis of UDP-6C-methyl-6-ketoglucose (3)  95  3.3.4 Evaluation of UDC as an irreversible inhibitor of UDPGlcDH  98  vi 3.3.5 Evaluation of the aldehyde 1 as a substrate for UDPGlcDH.  103  3.3.6 Evaluation of the secondary alcohols 2a and 2b as substrates for UDPGlcDH  106  3.3.7 Evaluation of the ketone 3 as both a substrate and inhibitor of UDPGlcDH  108  3.4 DISCUSSION  110  3.4.1 Covalent labeling with UDC  UO  3.4.2 Studies with the putative aldehyde intermediate 1  U2  3.4.3 Studies with C6" methyl analogues (2a, 2b and 3) of UDP-glucose  117  3.4.4 Experiments performed with Cys260Ser and Cys260Ala UDPGlcDH  120  C H A P T E R 4: T H E X - R A Y S T R U C T U R E O F U D P G l c D H  124  4.1 INTRODUCTION  125  4.2 MATERIALS AND METHODS  127  4.2.1 General Procedures  127  4.2.2 Purification and Crystallization  128  4.2.3 Data Collection and Processing  128  4.2.4 Model Building and Structural Refinement  130  4.3 RESULTS 4.3.1 Assessment of the Refined Structure of UDPGlcDH. 4.3.2 Tertiary and Quaternary Structure  130 :  130 134  4.3.3 Structural similarity to other dehydrogenases 4.3.4 Roles of conserved residues 4.3.5 Substrate binding.  138 139 141  4.3.6 Sequestering of reaction intermediates  145  4.3.7 Active site residues of UDPGlcDH.  146  4.4 DISCUSSION C H A P T E R 5: C O N C L U S I O N S A N D F U T U R E D I R E C T I O N S  5.1 THE MECHANISM OF UDPGlcDH 5.1.1 Research summary 5.1.2 Mechanistic relationship to other dehydrogenases 5.1.3 Future mechanistic work  149 155  156 156 160 162  Vll  5.2 THE STRUCTURE OF UDPGlcDH 5.2.1 Summary of UDPGlcDH  structure  168 168  5.2.2 Structural relationship to other dehydrogenases  168  5.2.3 Future structural work  170  5.3 CONCLUDING REMARKS  171  APPENDIX A:THEORETICAL TREATMENT OF ENZYME KINETICS  173  A.l THE RATE EQUATION FOR A SIMPLE ENZYME MECHANISM  174  A.2 REVERSIBLE INHIBITION  175  A.2.1 Competitive Inhibition  176  A.2.2 Uncompetitive Inhibition  177  A.2.3 Noncompetitive inhibition  178  A.3 IRREVERSIBLE INACTIVATION  179  A.4 INITIAL VELOCITY AND PRODUCT INHIBITION PLOTS  180  A.4.1 Rules for predicting initial velocity patterns  181  A.4.2 Rules for predicting product inhibition patterns  181  A.4.3 Distinguishing random from ordered mechanisms  181  A.5 DERIVATION OF THE RATE EQUATION FOR UDPGlcDH  .182  A P P E N D I X B : *H N M R S P E C T R A O F C H A R G E D C O M P O U N D S  187  A P P E N D I X C : C R Y S T A L L O G R A P H I C D A T A F O R C O M P O U N D 8b  192  C.l Experimental Methods C. 1.1 Data Collection  193 193  C.l.2 Data Reduction  193  C.l.3 Structure Solution and Refinement  193  C.l Results REFERENCES  196 199  Vlll  LIST O F FIGURES F i g u r e 1.1 Reaction catalyzed by UDPGlcDH  2  F i g u r e 1.2 Common structure and mechanism of the SDR family of enzymes  5  F i g u r e 1.3 Structure and mechanism for the L-specific dehydrogenases  7  F i g u r e 1.4 The general structure of a D-specific dehydrogenase  8  F i g u r e 1.5 Structure and mechanism of  6PGDH  10  F i g u r e 1.6 The structure of an aldo-keto reductase  12  F i g u r e 1.7 Mechanism of hemiacetal and gew-diol oxidation by AlcDH  13  F i g u r e 1.8 Structure and mechanism of GAPDH  15  F i g u r e 1.9 Structure and mechanism of the AldDH extended family  17  F i g u r e 1.10 Reaction catalyzed by histidinol dehydrogenase  19  F i g u r e 1.11 Proposed mechanism of histidinol dehydrogenase  20  F i g u r e 1.12 Reaction catalyzed by HMG-CoA reductase  21  F i g u r e 1.13 Structure and mechanism of the HMG-CoA reductase  22  F i g u r e 1.14 Kirkwood's proposed mechanism for UDPGlcDH  25  F i g u r e 1.15 The new proposed mechanism of UDPGlcDH  27  F i g u r e 2.1 SDS PAGE of each step in the preparation of UDPGlcDH  40  F i g u r e 2.2 Gel-filtration chromatography UDPGlcDH  41  F i g u r e 2.3 UV-visible spectrum of UDPGlcDH  44  F i g u r e 2.4 Initial velocity pattern with UDP-glucose as the variable substrate  46  F i g u r e 2.5 Initial velocity pattern with N A D as the variable substrate  46  F i g u r e 2.6 UDPGlcA product inhibition pattern with UDP-glucose as the variable substrate  48  F i g u r e 2.7 NADH product inhibition pattern with UDP-glucose as variable substrate  49  +  ix F i g u r e 2.8 NADH product inhibition pattern with N A D as variable substrate  50  F i g u r e 2.9 UDP-xylose inhibition pattern with UDP-glucose as variable substrate  51  +  F i g u r e 2.10 The 6 possible Ter Ter mechanisms in which UDP-glucose and UDPGlcA interact with the same enzyme form  55  F i g u r e 2.11 The three kinetic mechanisms consistent with an intersecting initial velocity plot  56  F i g u r e 2.12 Proposed Bi Uni Uni Bi ping pong kinetic mechanism of UDPGlcDH  58  F i g u r e 3.1 Uridine 5'-diphosphate chloroacetol (UDC)  61  F i g u r e 3.2 Uridine 5'-(a-D-g/Hco-hexodialdo-l,5-pyranosyl diphosphate) (1)  62  F i g u r e 3.3 Uridine 5'-(7-deoxy-L-g/ycero-a-D-g/wco-heptopyranosyl diphosphate) (2a), and uridine 5'-(7-deoxy-Dg/ycero-a-D-g/Mco-heptopyranosyl diphosphate) (2b)  64  F i g u r e 3.4 The diastereotopic hydrogen atoms of UDP-glucose  64  F i g u r e 3.5 The possible alternate oxidation of substrate analogues 2 a or 2b catalyzed by UDPGlcDH  65  F i g u r e 3.6 7-Deoxy-L-g/ycero-D-g/wcoheptopyranose (7a) and 7-deoxy-D-g/ycero-D-g/wco-heptopyranose (7b).. 66 F i g u r e 3.7 Uridine 5'-(a-D-g/wco-hept-6-ulopyranosyl diphosphate) (3)  67  F i g u r e 3.8 7-Deoxy-l,2:3,5-di-0-isopropylidene-L-g/jcera-a-D-g/wco-heptofuranose (10)  67  F i g u r e 3.9 Facile P-elimination in l,2,3,4-tetra-0-acetyl-(3-D-gluco-hexodialdo-l,5-pyranose  89  F i g u r e 3.10 The synthetic route to 1  90  F i g u r e 3.11 l-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-toluenesulfonate (CMC)  91  F i g u r e 3.12 Equilibrium between 1 and the gem-diol of 1 in aqueous solution  93  F i g u r e 3.13 The synthetic route to 2a and 2b  93  F i g u r e 3.14 The X-ray structure of 8b  94  F i g u r e 3.15 The synthetic route to 3  96  F i g u r e 3.16 Attempted MacDonald reaction on the P-anomer of 13  97  F i g u r e 3.17 Inactivation of UDPGlcDH by UDC in the presence of UDPGlcA  101  X F i g u r e 3.18 Inactivation of UDPGlcDH by UDC in the presence of UDPX  101  F i g u r e 3.19 Chloroacetol phosphate  102  F i g u r e 3.20 Determination of kinetic constants for the oxidation of 1 and UDP-glucose  104  F i g u r e 3.21 Time course for the UDPGlcDH catalyzed dismutation of 1 in the presence of NADH  105  F i g u r e 3.22 Time course of UDPGlcDH catalyzed turnover of 2a and 2b  106  F i g u r e 3.23 Time course of UDPGlcDH catalyzed turnover of 2a and 2b in the presence of NH OH  107  F i g u r e 3.24 Inhibition pattern for the ketone 3 with UDP-glucose as variable substrate  108  2  F i g u r e 3.25 Ion-paired reverse phase HPLC experiment to determine the product of UDPGlcDH catalyzed reduction of 3  109  F i g u r e 3.26 Analogous roles for Cys 260 during inactivation by UDC and formation of the thiohemiacetal intermediate  112  F i g u r e 3.27 Three possible routes for the UDPGlcDH catalyzed oxidation of 1  114  F i g u r e 3.28 Schematic representation of the dismutation observed when 1 is incubated with UDPGlcDH in the presence of NADH  116  F i g u r e 3.29 Proposed order of hydride transfer during the normal UDPGlcDH reaction  118  F i g u r e 4.1 Representative electron density for NAD(H) (adenosine not shown), the UDP-sugar, and residues 259260  133  F i g u r e 4.2 Ribbon representation of the ternary complex of the Cys260Ser UDPGlcDH/UDPGlcA/NAD(H) monomer  134  F i g u r e 4.3 Superposition of the a-carbons of residues 1-123 and 310-402 of Cys260Ser UDPGlcDH  135  F i g u r e 4.4 The crystallographic dimer of Cys260Ser UDPGlcDH  137  F i g u r e 4.5 Superposition of the a-carbons of Cys260Ser UDPGlcDH (black coil) and residues 1-300 of 6PGDH (white coil)  139  F i g u r e 4.6 Representative primary sequence alignment of 48 sequences including UDPGlcDH, UDPManNAcDH, and GDPManDH  140  xi F i g u r e 4.7 Schematic representation of interactions and hydrogen bond distances (A) between UDPGlcDH and bound UDPGlcA  143  F i g u r e 4.8 Close-up views of the active site of UDPGlcDH with putative catalytic residues  147  F i g u r e 4.9 Proposed role for Lys 204 in the two-fold oxidation  150  F i g u r e 4.10 An alternative mechanism for the two-fold oxidation  152  F i g u r e 5.1 The revised mechanism of UDPGlcDH  157  F i g u r e 5.2 The methyl ester of UDPGlcA  163  Xll  LIST O F T A B L E S T a b l e 2.1 Summary of the purification of UDPGlcDH  40  T a b l e 4.1 Data collection and phasing statistics  129  T a b l e 4.2 Refinement and model statistics  131  T a b l e 4.3 NAD(H)-UDPGlcDH hydrogen bond distances  144  T a b l e A . l Predicted product inhibition patterns for UDPGlcDH  186  T a b l e C . l Experimental details for the structure determination of compound 8b  194  T a b l e C.2 Atomic coordinates and B for compound 8b  196  T a b l e C.3 Bond lengths (A) for compound 8b  197  T a b l e C.4 Bond angles (°) for compound 8b  198  eq  Xlll  ABBREVIATIONS AND SYMBOLS  2 H  deuterium  6PGDH  6-phosphogluconate dehydrogenase.  A  angstrom (10" meter)  8  chemical shift (ppm)  s  extinction coefficient (M")  pM  micromolar (10" mole liter")  A260  absorbance at 260 nm  Ac  acetyl (CH CO)  AKR  aldo-keto reductase  AlcDH  alcohol dehydrogenase  AldDH  aldehyde dehydrogenase  APT  attached proton test (in reference to NMR)  BSA  bovine serum albumin  cap3a  the gene that encodes for UDPGlcDH in S. pneumoniae  CMC  l-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-  10  1  6  1  3  toluenesulfonate CoA  coenzyme A  Cys260Ala  UDPGlcDH with cysteine 260 replaced by an alanine  Cys260Ser  UDPGlcDH with cysteine 260 replaced by a serine  d  doublet (in reference to NMR)  Da  dalton (the mass of a hydrogen atom)  D2O  deuterium oxide  DCC  AyV'-dicyclohexylcarbodiimide  xiv DCI MS  desorption chemical ionization mass spectrometry  dd  doublet of doublets (in reference to NMR)  DE-52  diethylaminoethylcellulose ion exchange resin  DMF  dimethylformamide ((CH ) NCOH)  DMSO  dimethylsulfoxide ((CH ) SO)  DNA  deoxyribonucleic acid  DTT  DL-dithiothreitol  E. coli  Escherichia coli  EDTA  ethylene diamine tetraacetate (sodium salt)  ESI MS  electrospray ionization mass spectrometry  Et  ethyl (CH CH )  g  gram  g  the acceleration due to gravity at the earth's surface (9.8 meter second")  GAPDH  glyceraldehyde 3-phosphate dehydrogenase  GDPManDH  guanosine-5'-diphosphomannose dehydrogenase  h  hour  hasB  the gene that encodes for UDPGlcDH in S. pyogenes  HMG-CoA  3-hydroxy-3-methylglutaryl-coenzyme A  HPLC  high pressure/performance liquid chromatography  HR LSI MS  high resolution liquid secondary ion mass spectrometry  Hz  hertz (s")  IPTG  isopropyl-1 -fhio-P-D-galactopyranoside  J  coupling constant (Hz)  3 2  3 2  3  2  2  1  turnover number (s") 1  k  cat  XV  kDa  kilodalton (10 dalton)  K\  inhibition constant  K  Michaelis constant  LB  Luria Bertani medium  LSI MS  liquid soft ionization mass spectrometry  M  molar (mole liter")  m  multiplet (in reference to NMR)  MAD  multiwavelength anomalous dispersion  Me  methyl (CH )  MHz  megahertz (10 s")  min  minute  mL  milliliter (IO liter)  mM  millimolar (10" mole liter")  MurD  UDP-7V-acetylmuramoyl-L-alanine:D-glutamate ligase  MW  molecular weight  NAD(H)  oxidized or reduced nicotinamide adenine dinucleotide  NAD  oxidized nicotinamide adenine dinucleotide  3  m  1  3  6  1  3  3  +  1  NADH  reduced nicotinamide adenine dinucleotide  NADP  oxidized nicotinamide adenine dinucleotide phosphate  +  nm  nanometer (10" meter)  NMR  nuclear magnetic resonance  PDB  Protein Data Bank (http://www.rcsb.org)  p.s.i.  pounds per square inch  ppm  parts per million  9  xvi r.m.s.  root mean square  rpm  revolutions per minute  s  singlet (in reference to NMR)  S. pneumoniae  Streptococcus pneumoniae  S. pyogenes  Streptococcus pyogenes  SCHAD  L-3-hydroxyacyl-CoA dehydrogenase  SDR  'short-chain' dehydrogenase/reductase  SDS PAGE  sodium dodecyl sulfate polyacrylamide gel electrophi  SeMet  selenomethionine  t  time  THF  tetrahydrofuran (C H 0)  Trien  triethanolamine (N(C2H40H) )  UDC  uridine 5'-diphosphate chloroacetol  UDP  uridine 5'-diphosphate  UDPGlcA  UDP-glucuronic acid  UDPGlcDH  UDP-glucose dehydrogenase  UDPManNAcDH  UDP-TV-acetylmannosamine dehydrogenase  UTP  uridine 5'-triphosphate  UV  ultraviolet  Vis  visible  4  8  3  XVII  ACKNOWLEDGMENTS  I am forever indebted to my supervisor, Dr. Martin Tanner for giving me the opportunity to work in his research group. He has given me the confidence to pursue my own path with a work ethic that is fueled by curiosity. I am eternally grateful to Dr. Natalie Strynadka for adopting a crystallography neophyte into her group and providing endless technical support and encouragement. Perhaps I owe the most thanks to the people who originally got me hooked on enzymology, namely the professors who taught Biochemistry 402/403 during the academic year of 1993-1994 at the University of British Columbia (UBC). They include: Dr. G. Brayer, Dr. S. Withers, Dr. L. Mcintosh, Dr. G. Mauk, Dr. P. Candido, and Dr. I: Clark-Lewis. All members of the Tanner lab both past and present have shown me great support and I would particularly like to thank Dr. Rafael Sala for teaching me everything I know about organic synthesis. The majority of what I know about enzymology was passed down from Dr. Paul Morgan, Dr. Anne Johnson, and Dr. Suzana Glavas. I would also like to acknowledge Dr. Xue Ge for her contribution to this work, and Jason Galpin for being a great friend and confidant. In the Strynadka lab, Dr. Steve C. Mosimann, Dr. Mark Paetzel, and all the other members have taught me much more about crystallography than I could claim to 'know'. It is understood that I could never have completed an X-ray structure without their help. Numerous members of the Withers group have provided helpful discussion and technical advice over the years and I am grateful to them all. I would like to thank Dr. Ivo van de Rijn for supplying the expression vector for UDPglucose dehydrogenase, Shouming He for performing ESI MS, and Dr. Mark Paetzel for assistance in synchrotron data collection. We are grateful to the Department of Energy and the NSL at Brookhaven as well as the EMBL(DESY) at Hamburg for access to their respective synchrotron radiation sources. The Natural Sciences and Engineering Research Council of Canada (NSERC) and UBC have provided the financial support for this research.  XV111  This thesis is dedicated to my parents (all of them) and my lovely wife Shiho.  As strange as it may sound, I was thinking ofyou all while I wrote this thesis. I know that at least one ofyou is going to sit down and stubbornly try to read it (Mom!). I hope that the casual style and relatively general introductions to each chapter will allow you to gain some appreciation of to what avail all your encouragement and support has been during these six long years.  1  CHAPTER  1:  INTRODUCTION  2  UDP-glucose dehydrogenase (UDPGlcDH) catalyzes the NAD -dependent oxidation of +  UDP-glucose to UDP-glucuronic acid (UDPGlcA).  HO OH  (UDPGlcA)  HO OH  Figure 1.1 Reaction catalyzed by UDPGlcDH. The vast majority of NAD -dependent dehydrogenases catalyze a single oxidation and +  normally two such enzymes would be required to perform the overall transformation catalyzed by UDPGlcDH. For example, an alcohol dehydrogenase (AlcDH) is normally required to oxidize an alcohol to an aldehyde and an aldehyde dehydrogenase (AldDH) to oxidize an aldehyde to a carboxylic acid. There are only a few known NAD -dependent dehydrogenases +  that are able to combine these two activities in a single active site, and UDPGlcDH belongs to this small but privileged class (7). The research described herein positions UDPGlcDH as one of the most thoroughly investigated examples of an enzyme catalyzed two-fold oxidation. The majority of this introductory chapter will consist of an overview of enzymatic NAD -dependent +  oxidation and reduction. In addition, previous research specifically regarding UDPGlcDH will be summarized in order to provide some perspective on the significance of the work presented here. Having read this first chapter, the reader will be equipped with an appreciation of the many unique features of UDPGlcDH from both a biological and chemical perspective. The chapters of this thesis that describe original research are presented in an approximately chronological order beginning with the purification of the enzyme and ending with the first atomic scale structure of UDPGlcDH.  3 NAD -dependent dehydrogenases catalyze the oxidation of a variety of substrates +  ranging from simple organic molecules such as ethanol and formate to large biomolecules such as steroids and sugar-nucleotides. Structural similarities and sequence alignments have allowed the classification of the majority of these enzymes into several large families that will be introduced below. The scope of this overview is limited to enzymes that catalyze the single NAD -dependent oxidation of an alcohol to an aldehyde (or ketone) or an aldehyde to a +  carboxylic acid (or ester). Enzymes that catalyze NAD -dependent oxidative deaminations, such +  as the amino acid dehydrogenases, fall outside this criterion and will not be discussed. A common structural feature that is shared by almost all NAD -dependent dehydrogenases is the +  dinucleotide binding Rossmann fold that is used for binding of the N A D (or NADP ) cofactor 1  +  +  2  (2, 3). Unless otherwise mentioned, all the dehydrogenases discussed below contain a typical Rossmann fold or a slight variation thereof. In order to avoid confusion, every enzymatic reaction will be presented as an oxidation although the primary physiological role for some dehydrogenases may be the reduction of a carbonyl compound.  1.1 N A D - D E P E N D E N T O X I D A T I O N O F A L C O H O L S +  1.1.1 'Short-chain' and 'medium-chain' alcohol dehydrogenase (AlcDH) The two major families of AlcDHs have been designated the 'short-chain' and the 'medium-chain' dehydrogenases (4, 5). The 'medium-chain' dehydrogenases were formally known as the 'long-chain' dehydrogenases, but the discovery of an even longer dehydrogenase  ' As it was first described by Rossmann (2), the 'Rossmann fold' is a mononucleotide binding fold composed of a single (3-a-P-a-P motif. The dinucleotide binding fold is composed of two such motifs related by a 2-fold symmetry axis. However, the term 'Rossmann fold' is now commonly used to describe the dinucleotide binding motif (two mononucleotide binding motifs) and this interpretation will be used in this text.  4  (6) forced the change in convention. Very little is known about the 'long-chain' dehydrogenases and they will not be addressed here. The 'medium-chain' dehydrogenases are typically composed of 350-375 residues and are mechanistically characterized by their requirement for a catalytic Zn  2+  ion and their A face specificity for hydride transfer to the NAD cofactor. The classic 3  +  examples of a 'medium-chain' dehydrogenase are yeast and horse liver AlcDH; two extensively studied enzymes of known tertiary structure with well-understood catalytic mechanisms (5). Due to the central role of a metal ion in the catalytic mechanism, further discussion of the mechanism of these enzymes would provide little perspective towards an understanding of UDPGlcDH, an enzyme whose mechanism does not involve a metal ion. The 'short-chain' dehydrogenases/reductases (SDR) are a large and diverse family of enzymes (7, 8). The SDR enzymes are composed of a single 250-residue domain, have no requirement for a catalytic metal ion, and exhibit B face specificity for hydride transfer. Various SDR enzymes have been found active as monomers, dimers, and tetramers (8). The mechanism of the SDR enzymes features a strictly conserved tyrosine base in which the anionic form of the hydroxyl is stabilized by the side chain of a strongly conserved lysine (Figure 1.2b). At least 77 characterized enzymes of diverse specificity and function have been identified as members of the SDR extended family of enzymes.  Most general discussions of the nicotinamide cofactor will be inclusive of both N A D and NADP , however this distinction will not be explicitly made unless it is relevant. The following convention is used to specify the diastereotopic faces of the nicotinamide ring of NAD . The 'A face' refers to the re face of N A D or the same face of NADH as the pro-R hydrogen. The 'B face' refers to the si face of N A D or the same face of NADH as the pro-S hydrogen. 2  +  3  +  +  +  +  5  a.  Lys Tyr  h  T "O  /  I  H  H O  alcohol  JL  aldehyde o r ketone  NAD+  HO O H  HO O H  b.  Figure 1.2 Common structure and mechanism of the SDR family of enzymes, a) The structure of the Drosophila 'short-chain' AlcDH with bound NAD (PDB identification IB 15). b) The general mechanism of the SDR family of enzymes (R', R" = rest of molecule) with the complete structure of both NAD and NADH. NAD and NADH will be represented by their abbreviations throughout the remainder of the figures. +  +  +  6  The most remarkable example of the diversity in this family of enzymes is UDP-glucose 4-epimerase, an enzyme that catalyzes the interconversion of UDP-glucose and UDP-galactose through the non-specific oxidation and reduction of the C4" hydroxyl of the substrate. At the 4  time of writing, the ever growing number of SDR enzymes analyzed by X-ray crystallography include; 3a,20p-hydroxysteroid dehydrogenase (9), 17P-hydroxysteroid dehydrogenase (10), dihydropteridine reductase (11), mouse lung carbonyl reductase (12), 7a-hydroxysteroid dehydrogenase (13), 1,3,8-trihydroxysteroid naphthalene reductase (14), sepiapterin reductase (15), m-biphenyl-2,3-dihydrodiol-2,3-dehydrogenase (16), tropinone reductase (17), UDPglucose 4-epimerase (18), and Drosophila melanogaster AlcDH (19) As shown in Figure 1.2a, the SDR family of enzymes are typically composed of a single domain that binds both the substrate and the NAD cofactor. +  1.1.2 a-Hydroxy acid dehydrogenase  The a-hydroxy acid dehydrogenases catalyze the oxidation of an alcohol that is on a carbon directly adjacent to a carboxylic acid. All a-hydroxy acid dehydrogenases exhibit A face specificity of hydride transfer, are typically dimers or tetramers, and have an overall topology that features a distinct NAD -binding domain and a catalytic domain (20). This enzyme +  superfamily can be further divided into the L - and D-specific enzymes based on sequence homology, overall structure and substrate specificity (21). The L-specific dehydrogenases include L-lactate dehydrogenase (22) and L-malate dehydrogenase (23), two of the first dehydrogenases to have been structurally characterized by X-ray crystallography. These  The nomenclature of sugar-nucleotides uses a 'double prime' (") to indicate the sugar moiety, a 'single prime' (') for the ribose moiety, and no prime for the base. 4  7  enzymes have been extensively studied by a variety of techniques and numerous X-ray structures with bound cofactors, substrates, and inhibitors have been published. Shown in Figure 1.3a,b is the general structure and mechanism for the L-specific dehydrogenases.  a-L-hydroxyacid  ^  a-ketoacid  Figure 1.3 Structure and mechanism for the L-specific dehydrogenases, a) The structure of Bifidobacterium longum L-lactate dehydrogenase with bound N A D (PDB identification 1LLD). b) The general mechanism of the L-specific dehydrogenases (R = rest of molecule). +  Several X-ray structures are available for the D-specific dehydrogenase family and at the time of writing, the following structures have been published: D-glycerate dehydrogenase (24), D-3-phosphoglycerate dehydrogenase (25), D-2-hydroxyisocaproate dehydrogenase (20), and D-  8 lactate dehydrogenase (26). Shown in Figure 1.4 is the general structure for the D-specific dehydrogenases. The catalytic mechanism employed by these enzymes is identical to the L specific dehydrogenases but the relative orientation of the catalytic residues is 'flipped' in order to accommodate the inverted stereochemistry at the a-carbon of the substrate.  Figure 1.4 The general structure of a D-specific dehydrogenase. The structure shown here is that of Lactobacillus helveticus D-lactate dehydrogenase with bound NAD (PDB identification 2DLD). The enzyme mechanism is analogous to the L-specific dehydrogenases. +  Sequence homology and X-ray crystallographic studies have revealed that formate dehydrogenase is also a member of the D-specific dehydrogenase family although the substrate of this enzyme is not an a-hydroxy acid (27). Formate dehydrogenase catalyzes the simplest NAD -dependent oxidation of a carbonyl compound: hydride transfer from the formate anion to +  give carbon dioxide. A distinguishing feature of the D-specific dehydrogenases is an internal  pseudosymmetry between the NAD -binding domain and the catalytic domain (21). While this +  domain duplication is not unique to this family of enzymes, it does clearly distinguish the Dspecific dehydrogenases from the L-specific dehydrogenases.  1.1.3 P-Hydroxy acid dehydrogenase The p-hydroxy acid dehydrogenases are a small family of NAD -dependent enzymes that +  catalyze the oxidation of a P-hydroxy acid to a p-keto acid. Because only a few examples of these enzymes have been structurally characterized, it is difficult to make generalizations regarding their classification. However, there does appear to be a clear distinction between Phydroxy acid dehydrogenases that require a divalent metal ion and those that do not. Isocitrate dehydrogenase is an example of the former type for which the X-ray structure has been determined (28). Due to the requirement of a catalytic metal ion, the structure and mechanism of these enzymes will not be discussed further. P-Hydroxy acid dehydrogenases that do not require a divalent metal ion for activity include 6-phosphogluconate dehydrogenase (6PGDH) (29) and short-chain L-3-hydroxyacylCoA dehydrogenase (SCHAD) (30) that have both been structurally characterized by X-ray crystallography. These dimeric enzymes are both B face specific and are composed of a N A D +  binding domain and a primarily helical catalytic domain (Figure 1.5a).  10  NADPH  H  / Lys 1 8 3 — N  I  H-O,,  Lys 183  \ H  \\ H  '"0=< \  NADP+ 6-phosphogluconate  p-keto acid  C0  2  Glu 190  H  Glu 190  H  OH  Lys 1 8 3 1 - ^ N - H . . , , , H  Lys 183  'O  -N  H^  HO H \ ^ O H /  "-0  R  ribulose 5-phosphate  b.  1,2-enediol  Figure 1.5 Structure and mechanism of 6PGDH. a) The structure of sheep (Ovis orientalis aries) 6PGDH with bound NAD (PDB identification 1PGO). b) The proposed mechanism of 6PGDH (R = (CHOH) CH OP0 "). +  2  2  2  3  11 Despite their overall structural similarity, 6PGDH and SCHAD differ in their essential active site residues and catalytic mechanism. While SCHAD catalyzes the oxidation of a secondary alcohol to a ketone, 6PGDH actually catalyzes the oxidative decarboxylation of 6phosphogluconate to give ribulose 5-phosphate and carbon dioxide. The mechanism of 6PGDH is believed to proceed through an initial P-oxidation that is similar to the reaction catalyzed by SCHAD. A lysine residue has been implicated as the critical base in 6PGDH (29, 31) while a histidine residue has been proposed to fulfill the analogous role in SCHAD (30). The structure of 6PGDH and the mechanism of the oxidative decarboxylation are shown in Figure 1.5a,b.  1.1.4 Aldo-keto reductase The aldo-keto reductase (AKR) superfamily contains at least 40 known enzymes that catalyze the reduction of a wide variety of aldehyde and ketone substrates that includes molecules as diverse as steroids and monosaccharides (32). These monomeric enzymes are A face specific and consist of a single domain that has an (a/p)g-barrel fold (32, 33) as shown in Figure 1.6. AKR does not utilize a Rossmann fold to bind the nicotinamide cofactor. The mechanism of catalysis of the AKR superfamily is proposed to utilize a tyrosine general base to deprotonate the substrate hydroxyl in the oxidation reaction (33). A neighboring lysine residue may serve to depress the pK of the tyrosine base. This mechanism is closely analogous to the a  mechanism proposed for the SDR family (see Figure 1.2b) and is an apparent example of convergent evolution. Despite their lack of any structural homology, and therefore a common evolutionary ancestor, these two families of enzymes employ the same catalytic residues to perform identical reaction chemistry.  12  Figure 1.6 The structure of an aldo-keto reductase. The structure shown here is human (Homo sapiens) aldose reductase with bound NAD (PDB identification 2ACQ). The mechanism is analogous to the SDR family of enzymes. +  1.1.5 Dismutations and alternate reactions catalyzed by AlcDH While it is conceptually convenient to classify the AlcDHs by their primary activity, it is somewhat misleading due to the broad regioselectivity exhibited by several members of both the 'short-chain' and 'medium-chain' dehydrogenase families (5). The primary activity of 'mediumchain' horse liver AlcDH and Drosophila AlcDH (a member of the SDR family, see Section 1.1.1) is considered to be the oxidation/reduction of alcohols. However, both these enzymes can also catalyze the oxidation of hemiacetals to esters and gem-diols to carboxylic acids. This ability may be a relatively general (and unappreciated) property of dehydrogenases. The ahydroxy acid dehydrogenase, L-lactate dehydrogenase (see Section 1.1.2), has also been shown to catalyze the oxidation of the gew-diol glyoxalate to oxalate (5). It is believed that these 'unnatural' oxidation reactions proceed via a mechanism that is conceptually analogous to the  13 normal alcohol oxidation. With respect to enzyme catalysis, hemiacetals and gem-diols are effectively isosteric to secondary alcohols as seen in Figure 1.7.  HO  H  O  FT^H  NAD+  NADH  primary alcohol  HO R  aldehyde  horse liver or  H CH  R  AlcDH  3  R Enz  H OCH  R  CH  3  ketone  NAD+  B 3  x O  R^X>CH  hemiacetal  HO  x O  Drosophila  secondary alcohol  HO  R^V  R^R'  H  R = alkyl  OH  B = enyzme base  R'= H, alkyl, OCH3, OH  3  ester  O R^O" carboxylic acid  gem-diol  Figure 1.7 Mechanism of hemiacetal and gem-diol oxidation by AlcDH.  These oxidation reactions often manifest themselves when a dehydrogenase is assayed for aldehyde oxidation in the presence of N A D (34). In the three examples mentioned above, +  when this type of experiment was attempted, it was discovered that the aldehyde underwent a dismutation to an equimolar mixture of the corresponding primary alcohol and the carboxylic acid. An aldehyde in aqueous solution will always exist as an equilibrating mixture with the gem-diol form that results from the reversible addition of a water molecule to the aldehyde. Enzymatic oxidation of one molecule of gem-diol to the carboxylic acid will generate one molecule of NADH that then can participate in the reduction of one molecule of the aldehyde to the primary alcohol. Since only a small transient concentration of NADH is necessary for the dismutation to proceed, the reaction is silent to a spectrophotometric assay that detects the formation of NADH.  14  1.2 N A D - D E P E N D E N T OXIDATION O F A L D E H Y D E S +  1.2.1 Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde  3-phosphate  dehydrogenase  (GAPDH)  catalyzes  the  oxidative  phosphorylation of D-glyceraldehyde 3-phosphate to yield 1,3-diphosphoglycerate.  This  ubiquitous enzyme has a central role in the glycolytic pathway and it is therefore not surprising that it has been identified in a variety of organisms ranging from mammals to archea. GAPDH has been extensively studied from both a mechanistic and structural perspective. The tetrameric GAPDH is a B face specific dehydrogenase that consists of an N-terminal NAD -binding +  domain and a C-terminal catalytic domain (see Figure 1.8a). The X-ray structure of lobster GAPDH has been known since 1974 (35) and since then the structure of the enzyme from a variety of other sources, including human muscle (36) and hyperthermophilic bacteria (37), has been determined. The mechanism of GAPDH (Figure 1.8b) proceeds with nucleophilic attack of an active site cysteine residue on the aldehyde substrate to form a covalent thiohemiacetal intermediate (38). Collapse of the tetrahedral thiohemiacetal intermediate with hydride transfer to NAD  +  yields a thioester intermediate that is covalently bound to the active site cysteine. In the final step of the reaction, nucleophilic attack on the acylenzyme intermediate by phosphate anion yields 1,3-diphosphoglycerate. Aldehyde oxidation by this general mechanism is commonly referred to as the 'GAPDH paradigm'. In the currently accepted mechanism, an active site histidine has been proposed to act as a general base to increase the nucleophilicity of the active site cysteine residue (39). An alternative (or additional) role for the active site histidine is stabilization of the tetrahedral transition state through hydrogen bonding to the carbonyl oxygen (40), though this mechanism is not favored by the majority of researchers.  15  Figure 1.8 Structure and mechanism of GAPDH. a) The structure of Leishmania mexicana GAPDH with bound NAD (PDB identification 1GYP). b ) The mechanism of GAPDH (R = CH(OH)CH OP0 "). +  2  2  3  16 Another interesting example of an 'unnatural' dehydrogenase catalyzed oxidation (Section 1.1.5) was found to occur with a site-directed mutant of GAPDH (41). When the essential nucleophilic cysteine of GAPDH (see Figure 1.8b) was replaced with an alanine residue, the mutant enzyme was no longer able to catalyze the normal reaction. However, the mutant enzyme was found to have acquired the ability to catalyze the oxidation of the substrate, glyceraldehyde 3-phosphate, to the corresponding carboxylic acid, 3-phosphoglycerate. Further studies supported a mechanism for this oxidation that involved the direct oxidation of the gem-diol form of the substrate aldehyde.  1.2.2 Aldehyde dehydrogenase (AldDH) The AldDH extended family is a large and diverse collection of enzymes that catalyze the oxidation of an aldehyde to a carboxylic acid via covalent catalysis. The AldDH extended family has been further divided into four classes and at least thirteen families that will not be discussed here since all the families are structurally homologous and probably use the same chemical mechanism. There have been five published X-ray structures of AldDH enzymes: rat aldehyde dehydrogenase (42), mitochondrial aldehyde dehydrogenase (43), sheep liver aldehyde dehydrogenase (44), retinal dehydrogenase (45), and betaine aldehyde dehydrogenase (46). A representative structure of an AldDH and a general catalytic mechanism is shown in Figure 1.9a,b.  17  a.  -H"  O \  /  H  Cys Glu aldehyde  thiohemiacetal  III  o  NADH  .A.O R  H  SH  s I  Cys  Cys  0  I  H0  Glu  ^ 0 Glu  carboxylic acid  thioester  Figure 1.9 Structure and mechanism of the AldDH extended family, a) The structure of rat (Rattus norvegicus) aldehyde dehydrogenase with bound N A D (PDB identification 1AD3). b) The general mechanism of aldehyde dehydrogenase (R = rest of molecule). The same water molecule may or may not be involved in both deprotonation of the thiol and hydrolysis of the thioester. +  18 The AldDH enzymes are composed of a N-terminal dinucleotide binding domain and a C-terminal catalytic domain. Interestingly, both domains have a core topology that is a 5stranded variation of the Rossmann fold (47). All members of the AldDH extended family are A face specific and exhibit either dimeric (class III) or tetrameric (class I and II) quaternary structure (45). In a recent review, 145 sequences of AldDH were aligned and 16 residues were identified that are conserved in at least 95 % of all members of the extended family (48, 49). Among the 16 conserved residues is the active site cysteine residue that participates in covalent catalysis and is present in all AldDHs that have been shown to have catalytic activity. Although it is known that AldDH follows the GAPDH paradigm, the exact roles for the active site residues are still under debate. The mechanism represented in Figure 1.9b represents the minimal mechanism that is in general agreement with most proposals.  1.3 OTHER NAD -DEPENDENT TWO-FOLD OXIDATIONS +  UDPGlcDH  and  its  homologues  UDP-yV-acetylmannosamine  dehydrogenase  (UDPManNAcDH) (50) and GDP-mannose dehydrogenase (GDPManDH) (51, 52) has traditionally been classified as an NAD -dependent four-electron transfer dehydrogenase along +  with the only two other known examples: histidinol dehydrogenase and 3-hydroxy-3methylglutaryl-coenzyme A (HMG-CoA) reductase (1). The defining feature of this class of enzymes is their ability to perform a two-fold oxidation without release of an aldehyde intermediate. Several other characteristics of this class of enzymes include the use of NAD as a +  hydride acceptor and an important thiol in the catalytic mechanism. These relatively superficial empirical correlations have been superceded in the last decade by the wealth of primary sequence data that has resulted from the explosive growth in genetic research and molecular  19 biology. UDPGlcDH, histidinol dehydrogenase, and HMG-CoA reductase lack significant primary sequence homology and probably did not evolve from a common evolutionary ancestor.  1.3.1 Histidinol Dehydrogenase Histidinol dehydrogenase catalyzes the two-fold oxidation of histidinol to histidine as shown in Figure 1.10.  O  2NAD+  H  histidinol  2 NADH  H  histidine  Figure 1.10 Reaction catalyzed by histidinol dehydrogenase. In addition to UDPGlcDH, histidinol dehydrogenase is the only other known NAD +  dependent enzyme that catalyzes the oxidation of a primary alcohol to a free acid product. Although both enzymes have a conserved active site cysteine, two independent groups have reported that site-directed mutagenesis of the active site cysteine of histidinol dehydrogenase to an alanine residue had essentially no effect on the kinetic constants for the reaction (53, 54). This result conclusively proves that histidinol dehydrogenase does not follow the GAPDH paradigm. In recent years it has also been convincingly shown that histidinol dehydrogenase is a Zn 2+  metalloenzyme (55, 56) that may follow the mechanism shown in Figure 1.11 (57, 58).  20 Enzyme  NAD  HO  T  Kl,H R^iTOH  H  NADH ) HO-fn  7  + 2  w  R ^ O  -  , C' 2  Enzyme  Enzyme  histidinol  n  histidinal  u  NADH  ^ O  Zn Enzyme  histidine  ,  ^  NAD  +  HO \ H  R  X  ^  Zn  Enzyme gem-diol monoanion of histidinal  Figure 1.11 Proposed mechanism of histidinol dehydrogenase (B = putative enzyme base and R = CH(NH )CH2(C3N2PI3)). The structure of histidinol dehydrogenase is unknown. 2  A zinc bound hydroxide ion may be the catalytic base in the first oxidation and could accept a proton from the substrate hydroxyl to form a zinc coordinated water molecule. Displacement of this water molecule by the oxygen atom of the aldehyde intermediate would increase the electrophilic character at the carbonyl carbon. Nucleophilic addition of a hydroxide ion would yield the monoanion of the gem-diol of histidinal that could be directly oxidized to the carboxylic acid product, histidine. These somewhat unexpected conclusions and proposals contributed in part to our desire to investigate the mechanism of UDPGlcDH with modern techniques, such as site-directed mutagenesis, that were not available to previous researchers. It is interesting to note that the crystallization and some preliminary crystallographic studies of  21 histidinol dehydrogenase have been reported (59, 60), though no published report of the tertiary structure of histidinol dehydrogenase has appeared in the literature to date.  1.3.2 HMG-CoA reductase HMG-CoA reductase is a NAD -dependent four-electron transfer dehydrogenase that +  catalyses the oxidative addition of mevalonate and coenzyme A (CoA) to form HMG-CoA as shown in Figure 1.12. Although the primary physiological role for this enzyme is the reduction of HMG-CoA to give mevalonate and CoA, oxidation will be considered the forward direction for consistency.  2 NAD+  2 NADH  HMG-CoA reductase  (R)-mevalonate  coenyzme A  (S)-HMG-CoA  Figure 1.12 Reaction catalyzed by HMG-CoA reductase. This enzyme has recently been the subject of great interest due to its role as the critical regulatory step in the biosynthesis of cholesterol (61). The X-ray structure of HMG-CoA reductase  in nonproductive  ternary complexes  with  both  HMG-CoA/NAD  +  and  mevalonate/NADH has been determined (61, 62). HMG-CoA reductase is composed of a large domain and a smaller NAD -binding domain that shares no comparable structural features with +  the classic Rossmann dinucleotide binding fold (61). The enzyme is B face specific and forms a tightly associated dimeric structure with an extensive subunit interface. The tertiary structure of the binary complex of HMG-CoA reductase with NAD is shown in Figure 1.13a and the +  proposed catalytic mechanism is shown in Figure 1.13b.  22  Figure 1.13 Structure and mechanism of the HMG-CoA reductase, a) The structure of Pseudomonas mevalonii HMG-CoA reductase with bound NAD (PDB identification 1QAX). b) The proposed mechanism of HMG-CoA reductase (R - CH COH(CH )CH C0 ~). +  2  3  2  2  23 The detailed chemical mechanism proposed for HMG-CoA reductase is based on both structural and biochemical evidence (63, 64). In the first step of the proposed mechanism, mevalonate is oxidized to mevaldehyde with consumption of one equivalent of NAD . The +  intermediate aldehyde, mevaldehyde, is oxidized to the product, HMG-CoA, by a mechanism that is fundamentally analogous to the GAPDH paradigm. However, there are two features that distinguish the mechanism of this aldehyde oxidation from that of GAPDH: 1) the nucleophilic thiol is provided by the substrate CoA, and 2) the thioester that results from oxidation of the thiohemiacetal intermediate is not hydrolyzed. Despite these unique aspects of the chemical mechanism, HMG-CoA reductase is the only other known metal-independent enzyme that catalyzes an NAD -dependent two-fold oxidation and therefore serves as excellent precedent for +  the catalytic mechanism of UDPGlcDH.  14 . UDP-GLUCOSE DEHYDROGENASE 1.4.1 Biological role of UDPGlcDH Since the original report of this enzymatic activity in preparations of calf liver more than 40 years ago (65), UDPGlcDH has been identified in many other sources and recently the human enzyme has been cloned and sequenced (66). The product of UDPGlcDH, UDPGlcA, is the activated donor of the glucuronic acid moiety and serves many critical roles in a variety of organisms ranging from mammals to bacteria. In mammals, UDPGlcA is the substrate for UDPglucuronosyl transferases in the liver that catalyze the formation of glucuronide conjugates with various substances such as bilirubin and thereby aid in their excretion (67). UDPGlcA is also essential for the biosynthesis of hyaluronan and various glycoaminoglycans such as chondroitin sulfate and heparan sulfate (68). Mutation of the UDPGlcDH gene of Drosophila (designated sugarless), disrupts biosynthesis of the heparan sulfate side chains on proteoglycan core proteins  24 and is identical in phenotype to the classical wingless mutation (69). In plants, UDPGlcDH may be an important regulatory enzyme in the carbon flux towards cell wall and glycoprotein biosynthesis due to feedback inhibition from UDP-xylose (70). In many strains of pathogenic bacteria such as group A streptococci (71) and Streptococcus pneumoniae type 3 (74 % similarity to Streptococcus pyogenes), UDPGlcDH  provides the UDPGlcA necessary for construction of the antiphagocytic capsular polysaccharide (72). Mutations responsible for altered capsule formation in mutant pneumococcal capsular type 3 strains have been found to map to the gene encoding UDPGlcDH (72). Since proper formation of the capsular polysaccharide is essential for virulence in many pathogenic bacteria, UDPGlcDH is a logical target for the development of new drugs to directly combat these virulent organisms.  1.4.2 Previous Studies on UDPGlcDH The majority of previous studies on UDPGlcDH have focused on the beef liver enzyme (1, 73), and until recently very little was known about the bacterial enzyme (50, 74, 75). Unless otherwise noted, all discussion of previous research on UDPGlcDH will refer to the mammalian enzyme and therefore caution must be taken when extending the interpretation of these results to the bacterial enzyme. The majority of the previous work towards elucidation of the enzyme mechanism was done by Kirkwood during the 1970's (76, 77, 78, 79, 80). His research culminated in the mechanism proposed in Figure 1.14. The original impetus behind this current investigation stemmed from the purely academic desire to resolve some long-standing inconsistencies in Kirkwood's proposed mechanism. This proposal first appeared in the literature nearly 20 years ago and until now it has remained unchallenged, despite its apparent flaws. Kirkwood proposed that thefirstoxidation occurs by the unusual SN2 displacement of a hydride, leading directly to a covalently bound imine intermediate. In the second chemically unusual step,  25 the enzyme bound imine is 'transferred' to an active site cysteine to form a thiohemiacetal intermediate. The proposed oxidation of the thiohemiacetal and subsequent hydrolysis of the resulting thioester is well precedented and follows the GAPDH paradigm as discussed above (Section 1.2.1).  UDP-glucose  imine intermediate  hemiacetal intermediate  NADH  UDP-glucuronic acid (UDPGA)  thioester intermediate  Figure 1.14 Kirkwood's proposed mechanism for UDPGlcDH.  Kirkwood's proposed mechanism provides a very clear explanation of why the aldehyde intermediate had never been detected: there is no such species. The evidence that led to this conclusion came from experiments with UDPGlcDH in which the thiol of the essential active site cysteine residue had been converted to a thiocyanate and was therefore unable to act as a nucleophile (79). When the modified enzyme was preincubated with an impure mixture containing the putative aldehyde intermediate and then treated with sodium borohydride, a covalent enzyme adduct was formed. The adduct was identified as the product resulting from reduction of a Schiff s base that was formed between the aldehyde and an active site lysine  26 residue. This observation led Kirkwood to propose that the imine was a true intermediate in the enzymatic reaction as shown in the mechanism in Figure 1.14. One inconsistency with this aspect of the proposed mechanism comes from an earlier experiment that examined solvent 1&  oxygen isotope incorporation (81). When the enzymatic reaction was performed in O labeled water, only one solvent derived oxygen atom was incorporated into the final product. An imine intermediate would require that both of the oxygen atoms in the final product be derived from water. In order to maintain consistency with the proposed mechanism, the water molecule resulting from the elimination of the substrate hydroxyl oxygen must be sequestered in the active site and delivered back during hydrolysis of either the imine or thioester intermediates. The first reports detailing the purification and partial characterization of bacterial UDPGlcDH from Escherichia coli were published by Feingold in the mid 1970's (74, 75). The relevant results from those initial studies are summarized as follows: 1) the K values are 1.0 m  mM for UDP-glucose and 0.050 mM for NAD , 2) a reducing agent such as 2-mercaptoethanol +  is necessary to maintain maximum activity, and 3) the enzyme is a homodimer of subunit weight 47 kDa. The first UDPGlcDH gene to be cloned and sequenced was hasB from S. pyogenes, an encapsulated group A streptococcal strain (71). Through a collaboration with Ivo van de Rijn at Wake Forest University, the recombinant S. pyogenes UDPGlcDH was made available to us in a suitable expression vector and it is this enzyme that is the subject of the work described in the following chapters. UDP-glucose dehydrogenase from S. pneumoniae type 3 has since been cloned and overexpressed in E. coli (82). This 47 kDa protein was shown to have 74 % sequence similarity and several characteristics in common with UDPGlcDH from S. pyogenes (72). The enzyme showed no dependence on divalent metal ions for activity, the pH optimum was about 9.0, and iodoacetate inhibited the enzyme activity, indicating that there is probably an important cysteine residue.  27  1.5 THE PROPOSED MECHANISM OF UDPGlcDH A chemically reasonable mechanism for UDPGlcDH is proposed in Figure 1.15. This mechanism is consistent with the vast majority of the results obtained by previous researchers and all the chemical transformations involved are well precedented.  Enz  UDP-glucuronic acid (UDPGlcA)  Enz  thioester, ester, or amide intermediate  Figure 1.15 The new proposed mechanism of UDPGlcDH.  In the new proposed mechanism of UDPGlcDH, the initial NAD -dependent oxidation of +  the primary alcohol of UDP-glucose yields a tightly bound aldehyde intermediate. Oxidation of this aldehyde intermediate proceeds in a manner consistent with the well precedented GAPDH paradigm. No assumptions regarding the identity of the nucleophile are made, but reasonable candidates for this role include a thiol (cysteine), an alcohol (serine, threonine, or tyrosine), or an amine (lysine or histidine). Nucleophilic attack of the thiol, alcohol, or amine on the aldehyde intermediate generates a covalently bound thiohemiacetal, hemiacetal, or hemiaminal respectively. Collapse of this tetrahedral intermediate with transfer of the remaining hydride to a  28 second molecule of NAD generates a thioester, ester, or amide intermediate respectively. The +  final irreversible step of the reaction is hydrolysis of the covalent intermediate to yield the free carboxylic acid product.  1.6 O T H E R R E S E A R C H I N T H I S L A B O R A T O R Y  During the course of the research described in the following chapters, a second researcher in our laboratory, Xue Ge, was undertaking a simultaneous investigation of the mechanism of UDPGlcDH (83, 84). The results from her research have contributed immensely to our understanding of UDPGlcDH and provide important background for the experiments described in the following chapters. The investigations of Xue Ge focussed on two mutant enzymes that were constructed by Ivo van de Rijn at Wake Forest University but expressed and purified in this lab. Sequence alignments had indicated that cysteine 260 was the only conserved cysteine in UDPGlcDH from S. pyogenes and was therefore the best candidate for the catalytic nucleophile. The two mutants of interest were the cysteine 260 to serine mutant (Cys260Ser; thiol replaced by a hydroxyl), and the cysteine 260 to alanine mutant (Cys260Ala; thiol replaced by a hydrogen atom). With respect to the overall structure of the enzyme, these are very conservative mutations and should have little or no effect on the rate of catalysis unless they are involved in the catalytic mechanism. Both mutant enzymes show less than 0.1 % of the wild-type activity when assayed under conditions similar to the wild-type enzyme for the oxidation of UDP-glucose to UDPGlcA. However, prolonged incubation of Cys260Ser UDPGlcDH in the presence of UDP-glucose gave a stable covalent adduct that could be observed by electrospray ionization mass spectrometry (ESI MS) (83). Tryptic digestion of the labeled enzyme followed by neutral loss ESI MS or high resolution mass spectrometry of the peptide containing the Cys260Ser mutation, confirmed that the adduct was UDPGlcA.(minus H2O) attached to the protein through an ester linkage.  29 Xue Ge has also synthesized UDP-6"- H, H -glucose and tested it as a substrate with the 2  2  wild-type enzyme in order to quantitate the primary isotope effect for hydride transfer. If either hydride transfer step in the normal reaction mechanism is rate-limiting, the rate of oxidation of this deuterium labeled substrate should be diminished. It was observed that there is effectively no isotope effect (kulkv = 1.1 ±0.1), suggesting that neither hydride transfer is rate-limiting. UDP-6"- H, H-glucose was also used to probe whether or not Cys260Ala UDPGlcDH is 2  2  able to perform the first oxidation. As mentioned above, both mutants showed less than 0.1 % of the wild-type activity. In the case of Cys260Ser UDPGlcDH, this low activity was attributed to the greatly diminished rate of hydrolysis that resulted in accumulation of the ester intermediate. In the case of Cys260Ala UDPGlcDH, it was expected that the low activity was due to the inability of the mutant enzyme to form the hemiacetal intermediate necessary for the second oxidation step (aldehyde to carboxylic acid). However, Cys260Ala UDPGlcDH should be fully competent to perform the first oxidation (alcohol to aldehyde) because Cys 260 is probably not involved in this step. To experimentally test this hypothesis, the deuterium labeled substrate was incubated with Cys260Ala UDPGlcDH in the presence of NAD and an excess of NADH. +  Reversible oxidation of UDP-6"- H, H-glucose should give the tightly bound deuterated 2  2  2  2  aldehyde intermediate and one equivalent of enzyme bound NAD( H). If the bound NAD( H) can exchange with free NADH, a 'washing out' of deuterium at the C6" of the substrate should occur to give the monodeuterated UDP-6"-H, H-glucose. However, this 'washing out' was not 2  observed indicating that either: 1) Cys260Ala UDPGlcDH can not perform the oxidation of UDP-glucose to the aldehyde intermediate, or 2) NADH is not released from UDPGlcDH until after formation of the thiohemiacetal intermediate. In other work, Xue Ge has performed the normal enzymatic reaction with wild-type enzyme in H 0 and analyzed the UDPGlcA product by C NMR (84). It was found that only 18  2  13  30 one equivalent of solvent derived 0 was incorporated into the carboxylic acid product. As 18  discussed in Section 1.4.2, this result is consistent with an imine intermediate (see Figure 1.14) only if the original alcohol-derived water molecule is delivered back to an intermediate before release of the UDPGlcA product. Due to the obvious overlap between the work of Xue Ge and the results presented herein, great care has been taken to clearly demarcate her experiments. In Chapter 3, there is another section (Section 3.4.4) dedicated specifically to conclusions from her investigations. Due to the interdependence of our research, those results would lack context if disclosed at this point.  31  CHAPTER 2: PURIFICATION AND KINETIC CHARACTERIZATION OF UDPGlcDH  32 2.1 I N T R O D U C T I O N  This is truly a privileged time for enzymology. Progress in the field of recombinant DNA technology has made a great variety of enzymes available in quantities that were previously unimaginable. The advent of revolutionary techniques such as the polymerase chain reaction (PCR) has allowed practically any given protein to be routinely expressed at high levels. Prior to the widespread application of these techniques, the only sources of many proteins were the native organisms. Purification of the native proteins was laborious and obtaining only a few milligrams of pure protein was an admirable task and a credit to the skills of the researcher. This aspect of enzymology probably biased early researchers towards studying enzymes that could be found in a relatively high native abundance. UDPGlcDH is a good example of an enzyme that can be found in reasonable quantities in native organisms due to the requirement for high concentrations of the product, UDPGlcA, in several critical biological processes (Section 1.4.1). The classic source of UDPGlcDH, bovine liver, has a high concentration of this enzyme because an abundance of the product, UDPGlcA, is necessary for efficient detoxification through the process of glucuronidation (67). UDPGlcDH can be purified from calf liver homogenate in seven steps with a typical purification yielding 5 mg of enzyme from 1.8 liters of soluble liver homogenate (85, 86). The bacterial UDPGlcDH resisted purification until a mutant strain of E. coli was discovered that produced high levels of this enzyme (74). Using this mutant strain of bacteria, a six-step purification could yield 14 mg of UDPGlcDH starting from 100 g of E. coli cells. Considering that both these purifications began with the natural abundance of UDPGlcDH, these yields are very respectable, although an easier method of obtaining quantities of UDPGlcDH would be desirable. Fortunately, these very difficult and time-consuming purifications are no longer necessary for the routine preparation of pure enzyme. It is now possible to take a gene from a  33  native organism and insert it into an appropriate host organism that can be forced to express large quantities of the protein encoded by the foreign gene. The overexpressed protein may typically constitute 10 % of the total soluble protein in the cell lysate, thus greatly expediting the purification. The details of the techniques involved in this process are beyond the scope of this discussion and a brief overview will suffice. A preparation of the genomic DNA of the organism of interest is partially digested with an endonuclease to give a library of large DNA fragments. The library of DNA fragments can be inserted into an appropriate plasmid (circular non-genomic DNA) that contains a gene encoding for antibiotic resistance. Transforming this plasmid library into E. coli and screening for the combination of antibiotic resistance and the enzymatic activity of interest, will reveal which transformed bacteria contains the target gene. Sequencing of the plasmid DNA will reveal the exact nucleotide sequence of the target gene that could then be specifically excised and inserted into an expression plasmid. An expression plasmid contains a promoter region that allows the expression of a gene to be 'turned on' only in the presence of an external stimulus. A common promoter is the T7 promoter that allows a gene to be specifically expressed in the presence of added isopropyl-l-thio-p-D-galactopyranoside (IPTG). The basic cloning strategy described above was used to obtain both hasB, the S. pyogenes UDPGlcDH gene (77), and later cap3A, the S. pneumoniae UDPGlcDH gene (72). Through a collaboration with Ivo van de Rijn at Wake Forest University, the recombinant S. pyogenes UDPGlcDH was made available to us in an expression vector that allowed the preparation of significant quantities of the enzyme. In this chapter the purification and some preliminary characterization of the recombinant UDPGlcDH from S. pyogenes is described. In addition, a thorough kinetic analysis has been performed and the results from that study have been used to establish the order of substrate  34 binding and product release. The majority of the original research described in this chapter has been published (87).  2.2 M A T E R I A L S A N D  METHODS  2 . 2 . 1 General Procedures Plasmid pGAC147 that contains the hasB gene encoding for UDPGlcDH from S. pyogenes  was provided by Ivo van de Rijn at Wake Forest University, Winston-Salem, North  Carolina (77). Plasmid pGAC147 is an overexpression vector containing the gene for chloramphenicol resistance that can be transformed into competent E. coli JM109(DE3) and induced with IPTG. All chemicals and reagents were of enzyme grade or better. UDP-glucose, UDPGlcA, UDP-xylose, NAD , NADH, and dithiothreitol (DTT) were purchased from Sigma. +  Tryptone, yeast extract, and agar were purchased from Difco. Unless otherwise noted, all manipulations were performed at 4 °C. Protein concentrations were determined by the method of Bradford (88), using bovine serum albumin (BSA) in the normal UDPGlcDH buffer solution (50 mM triethanolamine (Trien), adjusted to pH 8.7 with hydrochloric acid (HC1)) as the standard. Protein purity was assessed using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE) stained with Coomassie blue as described by Laemmli (89). Molecular weight markers for SDS PAGE were carbonic anhydrase (29 kDa) and BSA (66 kDa). High pressure/performance liquid chromatography (HPLC) was performed with a Waters 625 LC system with a Waters 486 tunable absorbance detector. All enzyme kinetic assays and ultraviolet (UV) spectra were recorded on a Cary 3E UV-visible spectrophotometer. The program Grafit (Erithicus Software Ltd.) was used for the construction of all double reciprocal plots and determination of the least squares linear fit to the kinetic data. A unit of UDPGlcDH enzyme activity is defined as the amount of enzyme necessary to produce 2 u,M NADH min" at 30 °C. 1  35 For all kinetic assay experiments, an extinction coefficient of s = 6220 M" at 340 nm was 1  assumed for NADH. An extinction coefficient of 8 = 8900 M" at 260 nm was used to determine 1  the concentration of all UDP-sugars. ESI MS was performed by Mr. Shouming He on a PerkinElmer Sciex API300 electrospray mass spectrometer.  2.2.2 Specific Procedures UDPGlcDH expression and purification. Calcium chloride competent E. Coli JM109(DE3) cells were transformed by heat shocking for 2 min in the presence of pGAC147 at 42 °C and then plated onto LB-agar containing 25 pg mL" chloramphenicol. Plates were 1  incubated at 37 °C overnight and then stored at 0 °C for a maximum of two weeks. Single colonies of transformed E. coli were inoculated into 500 mL of autoclaved TYPG media (8 g tryptone, 8 g yeast extract, 2.5 g sodium chloride, 1.25 g dibasic potassium phosphate, and 2.5 g D-glucose), containing 25 pg mL" chloramphenicol. The inoculated broth was allowed to shake 1  at 37 °C and 290 rpm until the absorbance at 600 nm was between 0.8 to 1.2. IPTG was added to afinalconcentration of 0.4 mM and the bacteria were allowed to grow for 3 hours at 37 °C and 290 rpm. Cells were pelleted by centrifugation at 5000 rpm to give 1.9 g of cell pellet. The cell pellet was resuspended in 6 mL of buffer A (50 mM Trien-HCl pH 8.7, glycerol (10 % by volume),  DTT (2 mM), pepstatin  (1  mg liter"), aprotinin (1 1  mg liter"), and 1  phenylmethanesulfonyl fluoride (1.5 mM)). The cells were lysed by two passes through a French pressure cell at 20 000 p.s.i. and insoluble cell debris was pelleted by ultracentrifugation at 30 000 rpm (60 000 x g) for 1 h. The supernatant (4.9 mL) was loaded onto a column (15 mL) of diethylaminoethylcellulose (DE-52) that had been preequilibrated with buffer A. The column was washed with buffer A (50 mL) and eluted with a linear gradient of sodium chloride in buffer A (0 to 200 mM over 400 mL). Fractions containing UDPGlcDH activity were pooled and  36 concentrated (to 6.7 mL) using Millipore concentrators before dialysis against 1 liter of buffer B (50 mM Trien-HCl pH 8.7, glycerol (10 % by volume), and DTT (2 mM)) for 48 hours. The resulting solution (6.8 mL) was frozen in liquid nitrogen and stored at -75 °C. The protein solution was quickly thawed and applied (approximately 27 mg of total protein in each injection) to a Waters AP-1 Protein-Pak Q HPLC column (10 x 100 mM) that had been equilibrated with buffer B (at 22 °C). The column was washed with buffer B (10 min) and eluted with a linear gradient of sodium chloride in buffer B (0 to 120 mM over 35 min at 1.3 mL min"). The active 1  fractions were pooled, frozen in liquid nitrogen, and stored at -75 °C. Analysis of the protein by SDS PAGE indicated that a single protein with a molecular weight of approximately 45 kDa accounted for greater than 90 % of the total protein in the final solution.  Standard assay and kinetic characterization of UDPGlcDH. UDPGlcDH was assayed using a modification of the previously published procedure (74). All assays were performed in 50 mM Trien-HCl (pH 8.7, 1 mL) containing DTT (2 mM) and were preincubated for 5 min at 30 °C. For routine assays of UDPGlcDH, saturating concentrations of both substrates (200 uM UDP-glucose and 500 uM NAD ) were included in the assay buffer. Initial velocity studies were +  performed by including varying concentrations of both UDP-glucose and NAD in the assay +  buffer. For product inhibition studies, changing fixed concentrations of either UDPGlcA or NADH were included in the assay buffer and the concentration of either UDP-glucose or NAD  +  was varied in the presence of saturating concentrations of the other substrate (200 uM UDPglucose or 500 pM NAD ). The reaction was initiated by adding UDPGlcDH (2.0 x 10" units in +  2  0.010 mL) via syringe (Hamilton). To ensure thorough mixing, the polystyrene cuvette was covered in parafilm and inverted three times. Initial velocities were determined by following the  37 production of NADH at 340 during the first 40 seconds after initiation with UDPGlcDH and calculating the slopes using a least squares analysis with Cary 3 software version 3.0.  Molecular weight determination. The subunit molecular weight was determined by ESI MS. Gel-filtration studies to estimate the native molecular weight of UDPGlcDH were performed with a Protein-Pak 300SW column (Waters). The column was run in potassium phosphate buffer (100 mM, pH 7.0) at a flow rate of 1.0 mL min". The molecular weight 1  standards were prepared in the running buffer at a concentration of 1.0 mg mL" . UDPGlcDH 1  was prepared in the same buffer with the inclusion of DTT to maintain stability. For each run, 0.050 mL of the sample was applied to the column. The molecular weight standards used to construct the calibration curve were cytochrome c (12.4 kDa), ovalbumin (43 kDa), and BSA (66 kDa). Blue dextran (2000 kDa) was used to determine the void elution volume of the column.  Effect of additives on the activity of UDPGlcDH. To test for the presence of an essential catalytic divalent metal ion, the enzymatic reaction was performed in the presence of ethylene diamine tetraacetate (EDTA). Purified UDPGlcDH was assayed using the routine assay procedure described above with EDTA included in the assay buffer at a range of concentrations (0 to 5 mM). In order to determine the effect of carbonyl trapping reagents on the enzymatic reaction, the stoichiometric ratio of UDP-glucose consumed to NADH produced by UDPGlcDH was examined. UDPGlcDH was incubated under the normal assay conditions with 20 uM UDPglucose and 500 uM NAD in the presence of semicarbazide or hydroxylamine (0 to 100 mM). +  The reaction was allowed to proceed to completion as indicated by the cessation of increase in the absorbance at 340 nm. The net increase in absorbance at 340 nm was used to calculate the amount of NADH produced.  38  UV spectra of UDPGlcDH. UDPGlcDH purified as described above (0.4 mL, 2.5 mg mL") was dialyzed (2 x 24 h) under an atmosphere of argon against degassed potassium 1  phosphate buffer (10 mM, pH 8.7, and 2 mM DTT, 1 liter). Maintaining the buffer under an atmosphere of argon was necessary for obtaining a clean UV spectra without interference from oxidized DTT. Prior to addition of DTT, the buffer was thoroughly degassed under vacuum and sparged with argon. Solid DTT was added to the degassed buffer and the stirring solution was allowed to equilibrate under argon at 4 °C. Following dialysis, particulate matter was removed from the enzyme solution by centrifugation at 10 000 rpm for 15 min. A sample of the dialysis buffer was treated the same and used as the blank. A UV spectra of the enzyme solution (1.15 mg mL" or 25 uM) was recorded at 30 °C. To this solution was added a stock solution of UDP1  glucose in dialysis buffer to afinalconcentration of 500 uM and a second spectra was recorded. Finally, one equivalent of NAD in dialysis buffer was added to afinalconcentration of 25 uM +  and a third spectra was recorded.  Test for tightly bound NAD . A sample of UDPGlcDH (3.8 mg, 0.8 mL) was dialyzed +  overnight against phosphate buffer (10 mM, pH 7.0, 2 mM DTT, and 1.0 mM EDTA, 1 liter). The resulting UDPGlcDH solution (2.1 mg mL") was boiled for 3 min and treated with trypsin 1  (0.5 mg mL") and chymotrypsin (0.5 mg mL" ). A control sample containing an equivalent 1  1  concentration of both BSA (2.1 mg mL") and NAD (46.2 uM) was prepared and treated in an 1  +  identical manner. The samples were maintained for 24 h at room temperature and any precipitated material was gently redissolved. The portion of each sample (0.70 mL) was added to a concentrated solution of the assay buffer (0.275 mL, 330 mM glycine, pH 9.0, 1.4 mg mL"  1  semicarbazide). The UV spectrum of the sample was obtained, horse liver AlcDH was added  39 (0.5 mg), and a second spectrum obtained. Finally, EtOH (0.015 mL) was added, and a third spectrum obtained.  2.3 R E S U L T S  2.3.1 Purification of UDPGlcDH The first goal of this project was the purification of recombinant S. pyogenes UDPGlcDH that had been cloned and overexpressed in E. coli. After several unsuccessful approaches were attempted, a strategy was found that gave UDPGlcDH of satisfactory purity in just two purification steps. This procedure involved applying the crude UDPGlcDH to a column of anion exchange resin (DE-52) and eluting with a linear gradient of sodium chloride. Following buffer exchange and concentration, the partially purified UDPGlcDH was applied to an HPLC column of anion exchange resin and again eluted with a linear salt gradient. Collection of the major peak that contained UDPGlcDH activity gave the protein solution that was used for all experiments described in Chapters 2 and 3. For a summary of the protein purification and the yields obtained in each step, refer to Table 2.1. At each step of the purification, the purity of the protein was assessed by SDS-PAGE as shown in Figure 2.1. The purified protein shows a single major band of approximately 45 kDa when stained with Coomassie blue indicating that UDPGlcDH accounts for greater than 90 % of the total protein. A second faint band at approximately 40 kDa is also visible in the SDS PAGE gel. ESI MS confirmed that at least 90 % of the total protein in the final preparation was a single species with a molecular mass of 45 489 ± 4 Da.  40  Table 2.1 Summary of the purification of UDPGlcDH Preparation"  Total protein  Specific activity  Total activity  Yield  (mg)  (units mg")  (units)  (%)  Crude lysate (lane 2)  300  0.23  70  100  DE-52 column (lane 3)  180  0.34  62  89  HPLC column (lane 4)  35  1.5  51  73  a  1  Lane number refers to the appropriate lane of the gel shown in Figure 2.1.  1  I 2  3 Lane  4  5  Figure 2.1 SDS PAGE of each step in the preparation of UDPGlcDH. An equivalent number of units of UDPGlcDH (0.031 units) was loaded in each lane (2-4). Lane 1, markers; lane 2, crude lysate (103 pg total protein); lane 3, after DE-52 column (31 pg total protein); lane 4, after HPLC (16 pg total protein); lane 5, markers.  2.3.2 Molecular weight determination The recombinant gene for UDPGlcDH from S. pyogenes translates to a polypeptide chain of 402 amino acids with an expected molecular weight of 45 487 Da. ESI MS has confirmed a molecular weight of 45 489 i 4 Da for the UDPGlcDH monomeric subunit. In order to investigate the possibility that the enzyme may exist as a noncovalent dimer or higher order  41 complex in solution, gel-filtration chromatography was performed on UDPGlcDH. Gel-filtration chromatography separates molecules based on size and for proteins with a generally globular or spherical shape, it can be assumed that the size is directly proportional to the molecular mass. Based on this assumption, a gel-filtration column can be calibrated using globular proteins of known molecular mass and thus the apparent molecular mass for an unknown protein can be determined. The results from the gel-filtration experiment with UDPGlcDH are shown in Figure 2.2.  5 4.8  •2 4.4  4.2  4  1  1.1  1.2  1.3 Ve/V  1.4  1.5  0  Figure 2.2 Gel-filtration chromatography UDPGlcDH. The vertical line represents the measured V /V  for UDPGlcDH. The calibration curve was constructed using cytochrome c (A, 12.4 kDa), ovalbumin (•, 43 kDa), and BSA (•, 66 kDa). e  0  The data from the gel-filtration experiment is plotted as the logarithm of the molecular weight (log MW) against the ratio of the elution volume to the void volume (V /V ). According e  0  to this calibration curve, UDPGlcDH has a native molecular weight of 52 kDa. Since the polypeptide chain of UDPGlcDH corresponds to a 45.5 kDa protein, it was concluded that the enzyme exists as a monomer under the conditions of the gel-filtration experiment. Subsequent Xray structural studies have revealed that UDPGlcDH is actually a dimer in its active form  42 (discussed in Section 2.4.1 and Section 4.3.2). UDPGlcDH activity was present in the eluent from the gel-filtration column.  2.3.3 Effect of additives on the activity of UDPGlcDH When EDTA was included in the assay buffer, the rate of the normal enzymatic reaction was observed to increase in a linear fashion with respect to EDTA concentration. In the presence of 5 mM EDTA, the initial velocity of UDPGlcDH was 115 % of the initial velocity under normal assay conditions. This effect may be due to the sequestering of divalent counterions that could compete with UDPGlcDH for binding of the diphosphate moiety of the substrate. However, this possibility is not consistent with the fact that the enzyme was assayed under saturating conditions and an increase in the concentration of available substrate should have had no effect on the observed initial velocity. Alternatively, divalent metals could inhibit catalysis by binding to important residues on the enzyme itself. Increasing the concentration of active enzyme by removing the inhibitory metal ions would explain the apparent increase in activity. Regardless of the reason for the observed increase in activity, these results strongly suggest that a divalent metal ion is not required for catalysis. UDPGlcDH was assayed in the presence of either semicarbazide or hydroxylamine with the expectation that these reagents could change the apparent stoichiometry of the reaction if the putative aldehyde intermediate was accessible to the bulk solvent. Both of these reagents could form a relatively stable imine with the aldehyde intermediate, thereby preventing the aldehyde from undergoing the second oxidation to the product carboxylic acid. If significant trapping of the aldehyde intermediate did occur, the expected ratio of 2 NADH formed from every 1 UDPglucose would decrease. It was found that the ratio of NADH formed to UDP-glucose consumed remained constant at 1.9:1 over a concentration range of 0 to 50 mM for each additive. This  43 result indicates that if an aldehyde intermediate is formed in the reaction pathway, it is inaccessible to the bulk solvent and is not released into solution to any significant extent. It was found that DTT was important for maintaining high activity of the enzyme. Incubation of UDPGlcDH at 30 °C (50 mM Trien HC1, pH 8.7) in the absence DTT resulted in the loss of 75 % of the activity after 1.6 h. An identical experiment performed in the presence of 2 mM DTT resulted in the loss of 40 % activity in the same time period. The enzyme retained over 90 % of the activity when incubated for 24 h at 5 °C in the presence of DTT. The stability of UDPGlcDH in buffer solution could be improved by the addition of UDP-glucose (1 mM). Glycerol (10 %) was necessary in order to maintain enzyme activity through repeated rapid freeze and thaw cycles. All buffers used during the purification and assay of UDPGlcDH contained both DTT (2 mM) and glycerol (10 %).  2.3.4 Test for tightly bound NAD or NADH +  The UV-visible spectrum of UDPGlcDH (see Figure 2.3, trace 1) did not show any significant absorption past 300 nm that might be indicative of a bound chromophore such as NADH. Upon addition of an excess of UDP-glucose, there was little change in this region of the spectrum (Figure 2.3, trace 2) indicating no NADH was formed. This result suggests that bound NAD (if present) can not be converted to NADH (either bound of free) to any measurable +  extent. Upon addition of one equivalent of NAD to the same cuvette, an absorbance at 340 nm +  corresponding to the production of one equivalent of NADH was observed (see Figure 2.3, trace 3). This study indicates that only substoichiometric amounts, if any, of NADH are present in the purified sample of UDPGlcDH. However, this study can not rule out the possibility that the predominant form of purified UDPGlcDH contains a tightly bound NAD cofactor. It is possible +  that the bound equilibrium favors the substrates NAD /UDP-glucose over the intermediate +  44 complex of NADH/aldehyde intermediate (discussed in Section 3.4.4). Even in the presence of excess UDP-glucose, it may not be possible to form concentrations of bound NADH significant enough to be readily observed in the UV-spectra.  -1  l  250  300  T ™  350 Wavelength (nm)  I  400  Figure 2.3 UV-visible spectrum of UDPGlcDH. Trace 1, 1.15 mg mL" UDPGlcDH; trace 2, UDPGlcDH with 500 uM UDP-glucose; trace 3, UDPGlcDH with 500 uM UDP-glucose and 1 equivalent of NAD . 1  +  A tightly bound NAD cofactor was ruled out by assaying the proteolytic digest of a +  sample of UDPGlcDH with horse liver AlcDH in the presence of ethanol (90). When a control sample containing the proteolytic digest of an equivalent concentration of BSA and NAD was +  subject to the same assay conditions, an increase in absorbance at 340 nm was observed. This absorbance increase could be attributed to the formation of an approximately stoichiometric concentration of NADH. No significant increase in absorbance at 340 nm was observed with the  45 UDPGlcDH sample, indicating that substoichiometric amounts of NAD , if any, are present in +  the purified enzyme preparation.  2.3.5 Initial velocity kinetic studies An initial velocity kinetic analysis of UDPGlcDH was performed by determination of the initial rates of the enzymatic reaction at varying concentrations of both NAD and UDP-glucose +  (91, 92, 93). A thorough discussion of the rules used to interpret initial velocity and product inhibition plots is provided in Appendix A (Section A.4). When UDP-glucose was the variable substrate and NAD the changing fixed substrate, a double-reciprocal plot of the data (see +  5  Figure 2.4) revealed a set of linear intersecting lines. A similar plot resulted when NAD was the +  variable substrate and UDP-glucose was the changing fixed substrate (see Figure 2.5).  For all discussion of the results of the kinetic analysis, the variable substrate will refer to the species for which the reciprocal of the concentration is plotted along the X-axis. The changing fixed substrate (or inhibitor) will refer to the species whose concentration remains constant on each line that appears on the double reciprocal plot. 5  46  [UDP-glucose]" (mM)" [NAD+]" (mM)" Figure 2.4 Initial velocity pattern with UDP-glucose as the variable substrate. NAD concentrations are 10 pM (A), 16 uM (•), 23 uM (•), 41 uM (•), and 160 uM (O). The graph on the right is a replot of intercepts versus [NAD""]". 1  1  1  1  +  1  1  [NAD ]"' (mM)" [UDP-glucose]" (mM)" Figure 2.5 Initial velocity pattern with NAD as the variable substrate. UDP-glucose concentrations are 6 uM (A), 8 uM (•), 11 uM (•), 20 uM (•), and 70 uM (O). The graph on the right is a replot of intercepts versus [UDP-glucose]" . +  1  1  1  +  1  These intersecting plots are indicative of a kinetic mechanism in which binding of UDPglucose and at least one NAD to the enzyme are reversibly linked. Under initial velocity +  47 conditions in the absence of products, an irreversible link would result if a product were released from the enzyme before the second substrate bound to the enzyme (93).  A replot of the  intercepts against the reciprocal of the concentration of each changing fixed substrate was extrapolated to the x-axis to reveal the K values for each substrate. It was found that for NAD , +  m  K = 60 pM and for UDP-glucose, K = 20 uM. The y-intercept of both replots reveals the 6  m  m  reciprocal of the maximal velocity for the enzymatic reaction. Division of V  max  by the enzyme  concentration gives the turnover number for UDP-glucose of k =1.7 s". The derivation of the 1  cat  rate equation for UDPGlcDH and an explanation of the replot method of determining the kinetic constants is provided in the Appendix A (Section A.5).  2.3.6 Product inhibition kinetic studies When UDPGlcA was the inhibitor and UDP-glucose the variable substrate in the presence of saturating concentrations of NAD , a double reciprocal plot of the initial velocity +  data (see Figure 2.6) revealed competitive inhibition. A competitive inhibition pattern results if the inhibitor and the variable substrate are interacting with the same enzyme form (93). This  apparent competitive inhibition is consistent with an ordered kinetic mechanism in which UDPglucose binds first to UDPGlcDH and UDPGlcA is released last. Although the set of lines in Figure 2.6 should intersect on the y-axis, they appear to intersect slightly to the right of the yaxis . This apparent discrepancy may be due to substrate inhibition although the effect is very 7  Relative error values for kinetic constants are not provided as they have little relevance to this discussion. It will suffice to state that every experiment was repeated a minimum of three times and the resulting kinetic constants never differed by more than 10 to 20 %. An error value would poorly represent the expected reproducibility of these values in another laboratory due to the overwhelming factors of quality of enzyme preparation and particular assay conditions. Each line (constant changingfixedsubstrate or inhibitor) of the initial velocity and inhibition plots was fit individually by a linear least squares analysis and thus the points of intersection (competitive and noncompetitive 6  7  4 8  slight and thus the conclusions from this experiment and the K\ for UDPGlcA remain valid. A replot of the slopes versus UDPGlcA concentration gave an inhibition constant (K\) for UDPGlcA of200 uM.  i  0  i  I  20  i  I  i  40  I  60  i  I  80  i  I  100  ii i -200  i  i  I  0  I  200  400  [UDP-glucose]" (mM)" [UDPGlcA] (uM) Figure 2.6 UDPGlcA product inhibition pattern with UDP-glucose as the variable substrate. UDPGlcA concentrations are 0 uM (V), 50 uM (A), 110 uM (A), 170 uM (•), 230 pM (•), 290 uM (#), and 350 uM (O). The graph on the right is a replot of slopes versus [UDPGlcA]. 1  1  When NADH was the inhibitor and UDP-glucose the variable substrate in the presence of saturating concentrations of NAD , a parallel pattern of lines consistent with uncompetitive +  inhibition was observed (see Figure 2.7). Apparent uncompetitive inhibition results if the  inhibitor and the substrate are interacting with different enzyme forms that are not reversi connected (93). This inhibition pattern is consistent with a kinetic mechanism where one or both NAD bind to the enzyme after UDP-glucose and before the first NADH is released. NAD is +  +  inhibition) were not constrained. The program used to construct these plots (Grafit) provides the option of fitting all data points to the expected inhibition pattern, but this option was not used.  49  present in saturating concentrations so binding of this substrate is an essentially irreversible step. A replot of the intercepts against NADH concentration gave a K\ value for NADH of 100 uM.  [UDP-glucose]" (mM)" [NADH] (uM) Figure 2.7 NADH product inhibition pattern with UDP-glucose as variable substrate. NADH concentrations are 0 pM (A), 15 pM (A), 29 pM (•), 59 pM (•), 110 pM (•), and 180 uM (O). The graph on the right is a replot of intercepts versus [NADH]. 1  1  When NADH was the inhibitor and NAD the variable substrate in the presence of +  saturating concentrations of UDP-glucose, an apparent noncompetitive inhibition pattern was observed (see Figure 2.8). A noncompetitive inhibition pattern results if the inhibitor and  substrate interact with different enzyme forms and the points of interaction are reversibly lin (93). This is consistent with a mechanism where the binding of at least one NAD and the +  release of at least one NADH are not interrupted by any irreversible steps such as binding of UDP-glucose or release of UDPGlcA. A replot of the slopes against NADH concentration gave a K\ for NADH of 130 uM.  50  O.lrrTT  ! ~ 0.6 OH  <u  OJ -4-»  e  I—I  0.05  0i iiI iiiii iii i i ii • • i -100 0 100 10 0 10 20 30 40 50 60 70 [NAD ]"' (mM)[NADH] (pM) Figure 2.8 NADH product inhibition pattern with NAD as variable substrate. NADH concentrations are 0 uM (V), 25 uM (A), 45 uM (A), 65 pM (•), 87 uM (•), 110 pM (#), and 140 uM (O). The graph on the right is replot of intercepts versus [NADH]. +  1  +  No inhibition was observed when UDPGlcA was the inhibitor and N A D the substrate at +  saturating concentrations of UDP-glucose. This is consistent with any mechanism in which UDPGlcA is a competitive inhibitor with respect to UDP-glucose. Since UDP-glucose and UDPGlcA interact with the same enzyme form, saturation with UDP-glucose will ensure that the interacting form of the enzyme has no opportunity to bind UDPGlcA and thus the inhibitor has no effect.  2.3.7 Inhibition by UDP-xylose UDP-xylose is a competitive inhibitor with respect to UDP-glucose, indicating that this inhibitor binds to the free enzyme (see Figure 2.9). A replot of slopes against the concentration of UDP-xylose revealed that the K\ for UDP-xylose is 2.7 uM.  51  [UDP-glucose]" (mM)" [UDP-xylose] (pM) Figure 2.9 UDP-xylose inhibition pattern with UDP-glucose as variable substrate. UDP-xylose concentrations are 0 pM (O ), 0.5 p M 0 ), 1.5 pMTJl ), 3.0 uAB( ), 4.5 pM ( ), 7.0 u M ( ), and 10 p3M ( ). The graph on the right is a replot of slopes versus [UDP-xylose]. 1  2.4  1  D I S C U S S I O N  2.4.1 Characterization of UDPGlcDH The efficient purification of UDPGlcDH described in this chapter permitted many experiments to be performed with little concern for the quantities of enzyme that were being consumed. The most significant concern was for the stability of the purified enzyme preparation. It was found that UDPGlcDH from S. pyogenes is a particularly sensitive enzyme that rapidly loses activity (ti/2 of approximately 1 h) unless certain precautions are taken. This sensitivity probably stems from susceptibility towards oxidation of a critical cysteine thiol that may be important in the catalytic mechanism. Fortunately, the rate of inactivation can be significantly diminished, and the stability of the enzyme activity maintained, by the inclusion of the reducing agent DTT in the purification, storage, and assay buffers. Not surprisingly, UDPGlcDH is much more stable at 5 °C than at 30 °C so great care was taken during all purification steps and subsequent experiments to maintain the enzyme preparation on ice.  52 Several other features of S. pyogenes UDPGlcDH were investigated and found to be consistent with published results with bovine enzyme (1). These two enzymes share 31 % sequence identity so they have probably evolved from a common ancestor and therefore share a common structure and mechanism. The enzymatic reaction was not inhibited by EDTA, suggesting that a divalent metal ion is not involved in catalysis. Carbonyl trapping reagents have no effect on the overall reaction, consistent with the proposal that if an aldehyde intermediate is formed during the course of the enzymatic reaction, it is sequestered from the bulk solvent. One of the first goals with the purified enzyme was to determine its quaternary structure. ESI MS confirmed that the purified enzyme has the expected amino acid composition (45 487 Da) and gel-filtration chromatography suggested that in solution, the enzyme behaves as a globular protein of 52 kDa. Taken together, these results provide evidence that S. pyogenes UDPGlcDH exists as a monomer in solution. The quaternary structure of the hexameric bovine UDPGlcDH has been studied with a variety of techniques including partial denaturation (94, 95), chemical modification (107, 111), electron microscopy (96), and resonance energy transfer (97). Results from one partial denaturation study were interpreted as evidence for two separate active sites: one that performed the oxidation of UDP-glucose to the aldehyde, and one that performed the oxidation of the aldehyde to UDPGlcA (94). It was further proposed that these two active sites were situated on different subunits of the hexamer. The results from the resonance energy transfer experiments refuted this hypothesis by demonstrating that the active sites were remote from each other (97). If the previous evidence in favor of an active site at the dimer interface had been more convincing, the gel-filtration evidence in support of a monomer would have been more critically examined. Although the E. coli UDPGlcDH has also been demonstrated to have dimeric quaternary structure (75), there is precedent for closely related enzymes with high sequence identity differing in quaternary structure. The SDR family of dehydrogenases (Section  53 1.1.1) is an example of a family of enzymes that exists in a variety of quaternary structures ranging from monomers to tetramers. Subsequent X-ray structural studies have revealed that S. pyogenes UDPGlcDH is actually a dimer in its active form (discussed in Section 4.3.2). Based on the overall structure, there is no reason to suspect that the dimer of UDPGlcDH would not behave as a generally spherical protein, and thus the fundamental assumption of gel-filtration chromatography is valid. The discrepancy may be indicative of a monomer/dimer equilibrium that is dependent on the presence of substrate or inhibitors (see Section 4.3.2). The UV-visible spectra of purified UDPGlcDH revealed no significant absorbance past 300 nm that could be attributed to a tightly bound chromophoric cofactor such as NADH (Figure 2.3). Furthermore, treatment of the purified enzyme with excess UDP-glucose did not result in the formation of stoichiometric quantities of NADH. The fundamental assumption in this latter experiment is that saturation with UDP-glucose could significantly shift the bound equilibrium from the bound substrates (tightly bound NAD and UDP-glucose) towards the ternary complex +  of NADH and the aldehyde intermediate. However, this assumption may not be valid so the possibility of a tightly bound NAD cofactor was further investigated by assaying the proteolytic +  digestion of UDPGlcDH for free NAD . No evidence for stoichiometric concentrations of NAD +  +  was obtained from this experiment, eliminating the possibility that UDPGlcDH contains a tightly bound NAD cofactor. Strategies similar to those described above have previously been +  successful in detecting tightly bound cofactors with 5-adenosylhomocysteinase (90) and ADP-Lglycero-D-mannoheptose 6"-epimerase (98). Identical experiments have been used to rule out the possibility of a tightly bound NAD or NADH cofactor with UDP-/V-acetylglucosamine 2"+  epimerase (99).  54 2.4.2 The kinetic mechanism of UDPGlcDH There are six theoretically possible kinetic mechanisms for an enzyme with three substrates (UDP-glucose and two molecules of NAD ) and three products (UDPGlcA and two +  molecules of NADH) (91). These mechanisms (shown schematically in Figure 2.10) have all been considered and evaluated for their consistency with the results from the initial velocity kinetic analysis. The mechanisms shown in Figure 2.10 are already constrained by the observation that UDPGlcA is a competitive inhibitor with respect to UDP-glucose and therefore these two species must interact with the same enzyme form. Without this constraint, each of these 6 mechanisms would have 9 permutations resulting in a sum of 54 possible sequences of product binding and substrate release. The observed intersecting initial velocity plots (Figure 2.4 and 2.5) demands that the binding of UDP-glucose and at least one molecule of NAD are reversibly connected in the +  UDPGlcDH kinetic mechanism. As shown in Figure 2.10, a reversible connection between these binding events exists only in the Ordered Ter Ter, Bi Uni Uni Bi ping pong, and Bi Bi Uni Uni ping pong mechanisms. In the three remaining mechanisms (Uni Bi Bi Uni ping pong, Uni Uni Bi Bi ping pong, and Hexa Uni ping pong), NADH is released from the enzyme between binding of UDP-glucose and NAD . Release of NADH under initial velocity conditions is an effectively +  irreversible step and therefore parallel initial velocity plots would be expected if one of these three mechanisms were operative.  55 UDPG  NAD  NAD  NADH  NADH U D P G A Ordered Ter-Ter  UDPG  NAD  NADH  NAD  NADH U D P G A Bi Uni Uni Bi Ping Pong  UDPG  NAD  NADH  NADH  NAD U D P G A Bi Bi Uni Uni Ping Pong E'NADH  E'NADH UDPG  E'NADH  NADH NADH  A  A  NAD  NAD  Uni Bi Bi Uni Ping Pong E'NADH-;  2  UDPG  UDPGA  NADH  NAD  NAD  NADH U D P G A Uni Uni Bi Bi Ping Pong E'NADH  E'NADH UDPG  NADH  NAD  NADH  NAD  UDPGA Hexa Uni Ping Pong  E'NADH  E'NADH  Figure 2.10 The 6 possible Ter Ter mechanisms in which UDP-glucose and UDPGlcA interact with the same enzyme form.  The three remaining kinetic mechanisms that are consistent with the intersecting initial velocity plot (Ordered Ter Ter, Bi Uni Uni Bi ping pong, and Bi Bi Uni Uni ping pong) are represented in Figure 2.11. Unfortunately, it is not possible to further distinguish these three mechanisms based on the product inhibition studies. The observed uncompetitive inhibition (UDP-glucose versus NADH, see Figure 2.7) requires that binding of UDP-glucose and NADH are not reversibly connected. This constraint cannot be used to distinguish the mechanisms in  56  Figure 2.11 because in all cases NAD (present at saturating concentration) binds to the enzyme +  between the two binding events in consideration. Binding of a substrate present at saturating concentrations is an effectively irreversible step. Finally, the observed noncompetitive inhibition (NAD versus NADH, see Figure 2.8) requires that binding of at least one N A D and at least one +  +  NADH are reversibly connected. This is true in all three of the mechanisms represented in Figure 2.11, so once again this experiment cannot distinguish the mechanistic possibilities.  UDPG  UDPG  NAD  NAD  V  *  NAD  NADH  NADH  NADH U D P G A  Ordered Ter-Ter  NAD  NADH U D P G A  A  UDPG  NAD  NADH  NADH  A  A  A  Bi Uni Uni Bi Ping Pong  NAD U D P G A Bi Bi Uni Uni Ping Pong E-NADH  E-NADH  Figure 2.11 The three kinetic mechanisms consistent with an intersecting initial velocity plot.  In an effort to distinguish the three remaining kinetic mechanisms, the effect of NADH on the initial velocity plot with NAD as the variable substrate was further examined. As +  discussed in Appendix A (Section A.4), if a substrate interacts with the enzyme at more than one point in the kinetic mechanism, the effects at each point of interaction must be examined separately. Multiple slope effects will manifest themselves as parabolic double reciprocal plots if the two points of interaction are reversibly connected. Applying these general rules to UDPGlcDH, it is apparent that the presence of parabolic effects in the reciprocal plots is a potential method of distinguishing the kinetic mechanisms. Of the three remaining mechanisms  57  (Bi Uni Uni Bi ping pong, Bi Bi Uni Uni ping pong, and ordered Ter Ter), only the ordered Ter Ter has a reversible connection between NAD binding events in the absence of NADH. If the +  ordered Ter Ter mechanism is operative, the initial velocity plots (with NAD as the variable +  substrate) should be parabolic. Parabolic effects were not observed in the initial velocity experiment, providing evidence against the ordered Ter Ter mechanism. If either the Bi Uni Uni Bi ping pong or Bi Bi Uni Uni ping pong mechanism is operative, it should be possible to induce parabolic effects by including NADH in the assay buffer at less than saturating concentrations. This type of experiment has previously been used to rule out an ordered Ter Ter mechanism for histidinol dehydrogenase (100). An extensive effort to demonstrate this effect for UDPGlcDH was unsuccessful, indirectly supporting the ordered Ter Ter mechanism. However, the inability to observe parabolic effects under initial velocity conditions or induce them by addition of NADH, effectively nullifies any arguments that depend on either their presence or absence. An important caveat is that parabolic effects can be difficult to observe because the effect may be slight. In the study with histidinol dehydrogenase mentioned above (100), and a previous study with bovine UDPGlcDH (101), the parabolic effects were slight and the deviation from a straight line was probably smaller than the experimental error. Fortunately, several distinctive differences between the three possible mechanisms provide a means to experimentally differentiate them that does not involve kinetic experiments. If UDPGlcDH follows the Bi Bi Uni Uni ping pong mechanism, the enzyme must never exist in the free state and must always have a bound NAD or NADH as shown in Figure 2.11. As +  described above (Section 2.4.1), attempts to observe a tightly bound cofactor in the purified enzyme preparation were unsuccessful. In addition, both the Bi Bi Uni Uni ping pong and the ordered Ter Ter mechanisms require UDPGlcDH to simultaneously bind two molecules of  58 NAD . An analysis of the amino acid sequence reveals only one probable dinucleotide binding +  site (2,102). The initial velocity and product inhibition experiments were sufficient to limit the number of possible mechanisms for UDPGlcDH to three. The absence of a tightly bound cofactor and the presence of a single dinucleotide binding site in UDPGlcDH argue in favor of the Bi Uni Uni Bi ping pong mechanism. A Bi Uni Uni Bi ping pong mechanism has previously been proposed for both bovine UDPGlcDH (101) and histidinol dehydrogenase (100,103,104).  UDPG  NAD  NADH  NAD  NADH U D P G A  A  A  Figure 2.12 Proposed Bi Uni Uni Bi ping pong kinetic mechanism of UDPGlcDH.  The derivation of the rate equation describing this particular Bi Uni Uni Bi ping pong mechanism is provided in Appendix A (Section A.5). This equation has previously been derived for histidinol dehydrogenase (103), but is included in this work so several typographical errors in the published equation could be corrected.  59  CHAPTER 3: SYNTHESIS AND EVALUATION OF MECHANISTIC PROBES OF UDPGlcDH  60  3.1 INTRODUCTION The primary goal of this project was to establish the enzymatic mechanism of UDPGlcDH using the methods of bioorganic chemistry. Bioorganic chemistry may best be described as the marriage of synthetic organic chemistry, physical organic chemistry, and traditional biochemical techniques to answer questions about biological systems (105). An important aspect of this multifaceted approach to understanding enzyme mechanisms is the synthesis of alternate substrates or inhibitors and the testing of these rationally designed probes with the enzyme of interest. These mechanistic probes are tailored to answer a specific question about the enzymatic mechanism and will often provide a qualitative, rather than quantitative, result. For example, a 'yes or no' answer is probably sufficient to answer the main question of interest which might typically be phrased as, 'is the compound of interest an alternate substrate?' Determination of the kinetic constants and/or potential products associated with a substrate analogue or inhibitor is often secondary to the main question of interest. However, the thorough analyses required to address these secondary issues can often provide additional insight into the enzyme mechanism. The utility and elegance of bioorganic chemistry is the ability to design and synthesize rational probes to target a specific aspect of an enzyme mechanism and provide answers to questions that are inaccessible to any other technique. Upon initiation of this project, there were several major questions of interest that we felt could be answered using substrate analogues or potential inhibitors of UDPGlcDH. A critical feature of the proposed enzyme mechanism is an active site nucleophile that is necessary for formation of the thioester intermediate. Precedent from work with the bovine UDPGlcDH strongly supported the presence of a critical active site cysteine thiol, but this had never been verified for the bacterial enzyme. A potential probe of this aspect of the enzyme mechanism, uridine 5'-diphosphate chloroacetol (UDC), was first synthesized by Flentke and  61 Frey (106). UDC was originally prepared as a potential affinity label for an active site general base in UDP-glucose 4-epimerase (106). Although UDC was a potent and effectively irreversible inhibitor of the epimerase, this effort to identify an active site general base was ultimately unsuccessful. Inhibition was actually occurring through formation of a new chromophoric species that was generated by alkylation of the nicotinamide ring of the tightly bound NAD  +  cofactor.  O  HO OH  Figure 3.1 Uridine 5'-diphosphate chloroacetol (UDC).  In the normal UDPGlcDH reaction, a nucleophile (possibly a thiol) is proposed to attack the carbonyl carbon of the intermediate aldehyde (see Figure 1.15) that is four bond lengths away from the P-phosphate of the UDP moiety. The UDP moiety of UDC is a convenient and effective handle that could provide recognition and binding to UDPGlcDH in an orientation similar to the normal substrate, UDP-glucose. UDC also contains an electrophilic carbon atom three bond lengths away from the P-phosphate that could be positioned in the active site of UDPGlcDH in a location similar to the carbonyl carbon of the aldehyde intermediate. Nucleophilic attack by an active site nucleophile on the a-chloroketone of UDC would result in the irreversible formation of a covalent bond with the enzyme. Previous efforts to identify the active site nucleophile of bovine UDPGlcDH relied on non-specific thiol modification reagents such as: iodoacetate (107), iodoacetamide (107), 5-[[(iodoacetamide)ethyl]amino] napthalene-1sulfonic acid (108), iodoacetamidoflourescein (109), 5,5'-dithiobis-(2-nitrobenzoate) (78, 110),  62 and 6,6'-dithionicotinate (777). Inactivation of UDPGlcDH upon exposure to these reagents indicates that a cysteine is required for activity, and the observation that UDP-glucose or UDPxylose affords protection from inactivation indicates that the cysteine is probably located in the active site (707, 770). Proteolytic digestion of bovine liver UDPGlcDH that had been selectively modified with radiolabeled iodoacetate, followed by purification and sequencing of the modified peptide, has identified Cys 275 of the mammalian enzyme as the reactive thiol (109, 73). The important caveat that must be considered with these previous studies is that the labeling reagent was nonspecific and therefore was not targeted rationally or specifically to the active site. In contrast, the UDP 'handle' of UDC has the potential to orient the reactive moiety in close proximity to the catalytic nucleophile in the active site. The use of nonspecific thiol labeling reagents has generated misleading results in both AldDH (772) and histidinol dehydrogenase (53, 54). Preliminary reports have suggested that a compound related to UDC, uridine 5'diphosphate />bromoacetamidophenyl, is a potent inhibitor of UDPGlcDH (773, 774). An obvious synthetic target for an organic chemist investigating this enzyme mechanism would be the putative aldehyde intermediate, uridine 5'-(a-D-g/uco-hexodialdo-l,5-pyranosyl diphosphate) (1).  HO  OH  Figure 3.2 Uridine 5'-(a-D-g/uco-hexodialdo-l,5-pyranosyl diphosphate) (1).  This intermediate is the major difference between the new proposed mechanism and Kirkwood's previously proposed mechanism, making it a target of critical importance to this  63 investigation (refer to Figures 1.14, 1.15). If UDPGlcDH can oxidize 1 to the acid product at a rate greater than or equal to the rate of normal alcohol oxidation, it would provide very strong evidence that the aldehyde 1 is a true intermediate. In addition, the aldehyde 1 should be a substrate for both oxidation (with NAD ) and reduction (with NADH) by UDPGlcDH and thus +  the enzyme mechanism could be dissected into two separate oxidation reactions that could be individually investigated. Finally, the aldehyde 1 is an interesting synthetic target in itself since it is an unknown molecule that may find utility in applications that are far beyond the scope of this one project. For example, bacteria grown in the presence of 1 may incorporate the reactive aldehyde moiety into their cell-surface polysaccharides and thus provide a convenient and specific handle for attaching other groups through an imine linkage (115). It had previously been shown that 1 could be obtained through an inefficient enzymatic synthesis that was not satisfactory for our purposes (76, 79). The reported procedure employed galactose oxidase to oxidize the C6" hydroxyl of UDP-galactose to the corresponding aldehyde. The final product 1 was generated by enzymatic epimerization of the C4" hydroxyl by UDPglucose 4-epimerase to generate an inseparable mixture of compounds for which no spectroscopic characterization or indication of purity was given. Presumably, the mixture would contain UDP-glucose, UDP-galactose, the aldehyde 1, and the galacto analogue of aldehyde 1. Any experiments performed with this mixture of compounds would require many controls and assumptions to be made; particularly since it was discovered in the course of this project that UDP-galactose is a reasonable substrate for UDPGlcDH (k JK approximately 10 fold less 3  c  m  than UDP-glucose). Due to these concerns, the stepwise chemical synthesis of 1 was undertaken. An organic chemist may also speculate on the specificity and versatility of the enzymatic oxidation. Is the enzyme capable of oxidizing alcohols other than its normal primary alcohol substrate and if so, what can this tell us about the enzyme mechanism? In an attempt to answer  64 these questions, the secondary alcohol analogues of UDP-glucose, uridine 5'-(7-deoxy-Lg/ycero-a-D-g/wco-heptopyranosyl diphosphate) (2a), and uridine 5'-(7-deoxy-D-g/vcero-a-Dg/wco-heptopyranosyl diphosphate) (2b) were synthesized.  HO OH  HO OH  Figure 3.3 Uridine 5'-(7-deoxy-L-g(ycero-a-D-g/wco-heptopyranosyl diphosphate) (2a), and uridine 5'-(7-deoxy-D-g/j>cero-a-D-g/wco-heptopyranosyl diphosphate) (2b).  If one or both of these molecules are indeed substrates for UDPGlcDH, they could undergo a single oxidation to the corresponding ketone product but a second oxidation would require the energetically unfeasible cleavage of a carbon-carbon bond. This single oxidation would be conceptually and mechanistically analogous to the normal alcohol to aldehyde oxidation. If this analogy holds, it would be expected that only one of the two diastereomers would be a substrate since the removal of the first hydride will be dictated by the stereospecificity of the enzyme. As shown in Figure 3.4, the two C6" hydrogen atoms of UDPglucose can be distinguished as either pro-R or pro-S.  Figure 3.4 The diastereotopic hydrogen atoms of UDP-glucose.  65 As shown in Figure 3.5, the substrate analogues 2a and 2b each retain a C6" hydrogen atom that is analogous to the pro-R and pro-S hydrogen atoms respectively. Since only one of the two C6" hydrogen atoms of UDP-glucose is transferred to NAD in thefirstoxidation, only +  one of either 2a or 2b is expected to be an alternate substrate for UDPGlcDH. The ketone product resulting from the oxidation of either 2a or 2b may be a good inhibitor of UDPGlcDH (Figure 3.5). The ketone product could take advantage of the same interactions with UDPGlcDH that are responsible for the suspected tight binding of the putative aldehyde intermediate. Specifically, the ketone product could form a covalent hemiacetal intermediate with the putative active site nucleophile (see below for further discussion).  'pro-R  like' hydrogen  Enz  NADH  ketone (3)  hemiacetal  2b  Figure 3.5 The possible alternate oxidation of substrate analogues 2a or 2b catalyzed by UDPGlcDH.  66 The introduction of a stereocenter at the C6 of glucose is a problem that had traditionally been approached by the use of organo-metallic reagents to attack an aldehyde generated at this center (116, 117, 118). This approach has had some success and both 7-deoxy-L-glycero-Dg/wco-heptopyranose  (7a) and 7-deoxy-D-g/ycero-D-g/uco-heptopyranose (7b) have been  prepared in this manner through a long and non-selective synthesis (118). A much more efficient and elegant preparation of both 7a and 7b from D-glucuronolactone in 7 steps has been recently published (119, 120). This non-diastereoselective synthesis allows preparation of both epimers since they are separable by flash chromatography when protected as diacetonides. Deprotection with aqueous trifluoroacetic acid affords the free sugars 7a and 7b that serve as the starting material for the synthesis of 2a and 2b reported here.  Figure 3.6 7-Deoxy-L-g/ycero-D-g/wco-heptopyranose (7a) and 7-deoxy-D-g/ycero-D-g/wcoheptopyranose (7b). As discussed in Section 1.4.1, UDPGlcDH is essential for virulence in certain pathogenic bacteria and inhibitors of this enzyme may lead to novel drugs to combat diseases such as pneumonia. UDC has the potential to be a very good irreversible inhibitor of UDPGlcDH, but the chloroacetol moiety limits the utility of this inhibitor due to its high reactivity towards nucleophiles. For this reason, it was of interest to synthesize a compound that had the potential to act as a potent and specific inhibitor of UDP-glucose dehydrogenase. In Section 2.3.7 the inhibition constants for the commercially available UDP-xylose (K\ = 2.7 pM) were presented. UDP-xylose can take advantage of all the productive binding interactions possible with the  67 normal substrate UDP-glucose, and only lacks the C6" carbon at which the oxidation occurs. We reasoned that a potent inhibitor of UDPGlcDH might be a compound that retained the basic structure of UDP-xylose but also took advantage of the putative ability of UDPGlcDH to form a hemiacetal species within the active site (see Figure 3.5). For this reason, the chemical synthesis of uridine 5'-(7-deoxy-a-D-g/uco-hept-6-ulopyranosyl diphosphate) (3) was undertaken.  HO OH  Figure 3.7 Uridine 5'-(a-D-g/wco-hept-6-ulopyranosyl diphosphate) (3). The enzymatic oxidation of either of the secondary alcohols 2a or 2b may generate the ketone 3 in the active site of UDPGlcDH. However, to obtain meaningful inhibition constants for the ketone 3, it would be critical to independently synthesize it. A second potential use of the ketone 3 would be as a substrate analogue that could undergo reduction by UDPGlcDH and NADH to form one of the secondary alcohols 2a or 2b. The starting material (10) for the synthesis of 3 can be derived from D-glucuronolactone in 6 steps using similar chemistry to that used in the preparation of 7a and 7b (120).  D-glucuronolactone  10  Figure 3.8 7-Deoxy-l,2:3,5-di-0-isopropylidene-l-gfycero-a-D-g/wco-heptofuranose (10).  68 The majority of the original research described in this chapter regarding U D C (57), the putative aldehyde intermediate 1 (121), and the analogues 2a, 2b, and 3 (122), has been published.  3.2 M A T E R I A L S A N D M E T H O D S 3.2.1 General Synthetic Methods A l l chemicals were purchased from Sigma, Aldrich, Fisher Scientific, or Lancaster. 1,2,3,4-tetraacetyl-P-d-glucose 4 was prepared as previously described (123). 7-Deoxy-lg/ycero-d-g/wco-heptopyranose 7a, 7-deoxy-d-g/ycero-d-g/wco-heptopyranose 7b, and 7-deoxyl^iS^-di-O-isopropylidene-l-gfycero-a-d-g/wco-heptofuranose 10 were prepared according to the procedures of Bleriot et al.(120). U D C was prepared in this lab by Rafael Sala, following the preparation of Flentke and Frey (106). U D C was stored as a lyophilized solid and dissolved in distilled water before use. Concentration was determined using A260 and an extinction coefficient of e = 8900 M " . The cyclohexylamine salt of the dimethyl acetal of chloroacetol phosphate was 1  kindly provided by Dr. Anne Johnson. Flash column chromatography was performed with Silica Gel 60 (230-400 mesh, E. Merck, Darmstadt). Pyridine, acetonitrile (CH CN), dichloromethane 3  (CH2CI2),  and triethylamine (NEt ) were freshly distilled over CaEb under N 2 . Methanol 3  (MeOH) was freshly distilled over magnesium methoxide under N2. Tetrahydrofuran (THF) and diethyl ether (Et20) were freshly distilled from sodium and benzophenone. Dimethyl formamide (DMF) and dimethyl sulfoxide (DMSO) were distilled under reduced pressure and stored over 3 A molecular sieves. A l l reactions (unless indicated otherwise) were performed under an atmosphere of argon. Analytical thin layer chromatography was performed on aluminum-backed sheets of silica gel 60F254 (Merck) of thickness 0.2 mm. Visualization of compounds was achieved by spraying the plate with a prepared solution of H S04 (31 mL), ammonium 2  69 molybdate (21 g), and Ce(S04)2 (1 g) dissolved in water (500 mL) and then heating with a heat gun until charring occurred. 'H Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker WH-400 MHz instrument at a field strength of 400 MHz. ' H correlation spectroscopy (COSY) were obtained on either a Bruker WH-400 M H z or a Bruker AC-200E spectrometer.  1 3  C N M R and attached  proton tests (APT) were recorded on a Varian XL-300 at 75 M H z with broad band decoupling. 1 3  C N M R spectra obtained in CDC1 were referenced to the solvent peak (5 = 77.0). P N M R 3 1  3  spectra were recorded on Varian XL-300 at 121.5 MHz with *H decoupling and were referenced to H3PO4 (8 = 0.00), as an external standard. Chemical shifts are reported using the 8 scale in ppm and all coupling constants (J) are reported in hertz (Hz). Mass spectrometry (except ESI MS) was performed at the Mass Spectrometry lab of University of British Columbia (U.B.C.) Chemistry Department. Desorption chemical ionization mass spectrometry (DCI MS) was performed with a Delsi Nermag R10-10C mass spectrometer with NH3 as the ionization gas. Liquid secondary ion mass spectrometry (LSI MS) and high resolution LSI MS (HR LSI MS) was performed with a Kratos Concept II HQ mass spectrometer. Elemental analyses were performed by Mr. Peter Borda in the Microanalytical Lab of the U.B.C. Chemistry Department. Optical  rotations  were  measured  with  a  Perkin-Elmer  241  M C polarimeter.  X-ray  crystallographic analysis was performed by Mr. Steve Rettig in the Crystallography Lab of the U.B.C. Chemistry Department as described in Appendix C. All reported compounds were characterized by ' H NMR, COSY,  1 3  C NMR, and mass  spectrometry (DCI MS or LSI MS). A l l neutral compounds were further characterized by elemental analysis. Charged compounds are not readily amenable to microanalysis so their elemental composition was routinely assessed by high resolution mass spectrometry. A *H N M R of each charged compound is included in Appendix B to serve as an indication of compound  70 purity. The tabulated NMR data for each reported compound includes tentative assignments for each peak and coupling constant. The 'H NMR assignments were based on the COSY spectra and literature precedent from closely related compounds. The C NMR assignments were based 13  on attached proton test (APT) experiments, and literature precedent. P NMR assignments were 31  based solely on literature precedent.  3.2.2 Routine Synthetic Procedures Cation counterion exchange. The compound (10 to 20 mg) was dissolved in distilled water (1-2 mL) and passed through a column of Amberlite IR-120(+) resin (5 cm x 1.5 cm). Typically the sodium form of the resin would be used (wash resin with 1.0 M NaOH until basic to litmus, then water until neutral), though it was sometimes convenient to use the pyridinium form (wash resin with 10 % pyridine until basic to litmus, then water until neutral). In order to exchange to the triethylammonium counterion, the protonated form of the resin was used and excess NEt3 was added to the eluent. Water was removed by lyophilization to afford the pure product. Preparative thin layer chromatography. Preparative thin layer chromatography (TLC) of various modified a-D-glycopyranosyl phosphates was performed on glass-backed plates of either 1 mm or 2 mm thick silica gel 60F254. The sample was spotted with a capillary tube onto a horizontal line 2.5 cm from the bottom of the plate. The running solvent (CHCh/MeOH/HaO) was prepared in a ratio that gave an appropriate Rf (0.20) for the desired compound. Preparative plates were visualized by covering the plate with aluminum foil, leaving 1 cm of a vertical edge visible. The exposed portion of the plate was sprayed with the molybdate/ceric sulfate solution described above, and carefully heated with a heat gun until bands corresponding to product became apparent. This procedure could be repeated on the second edge of the preparative plate if  71 necessary. The appropriate area of the preparative plate was removed, finely ground with mortar and pestle, and suspended in the running solvent (50 mL, 1 h). Several drops of NEt were added 3  to ensure conditions remained neutral or slightly basic. The silica gel was removed by filtration and solvent evaporated in vacuo. This procedure was typically followed by counterion exchange to the sodium salt as described above. Purification of uridine 5'-(a-D-glycopyranosyl diphosphate^. The crude uridine 5'(a-D-glycopyranosyl diphosphate) was applied to a column (1.5 cm x 18 cm) of DE-52 anion exchange resin pre-equilibrated with distilled water. The product was eluted with a linear salt gradient (0 to 300 m M L i C l over 800 mL total volume) at a flow rate of 1.5 mL min" . The 1  product was detected at 254 nm with a S P E C T R U M Spectra/Chrom Flow Thru U V detector. The relevant fraction (30-40 mL) was loaded directly onto a size-exclusion column (45 cm x 2.5 cm or 110 cm x 2.5 cm) of Biogel P-2 (200-400 mesh), and eluted with distilled water in order to desalt the product. The U V active fractions were exchanged to the sodium counterion and lyophilized to dryness. The product was redissolved in distilled water (2-3 mL) and applied to a second Biogel P-2 column. The major U V absorbing peak was collected, the counterion exchanged to sodium, and lyophilized to give a white powder suitable for characterization and further experiments. A l l uridine 5'-(a-D-glycopyranosyl diphosphate^ were stored over desiccant at -20 °C.  0-Deacetylation with sodium methoxide or triethylamine. A solution of sodium methoxide (NaOMe) in M e O H (1 M ) was prepared by slow addition of sodium metal (0.60 g, 26 mmol) to M e O H (26 mL). The acetylated a-D-glycopyranosyl phosphate was dissolved in distilled M e O H (20 mg mL" ) and NaOMe (1 M ) was added to a final concentration of 50 m M . 1  Alternatively, the acetylated a-D-glycopyranosyl phosphate was dissolved in 10 % NEt (20 mL, 3  10 % NEt /45 % MeOH/45 % H 0 ) and stirred at room temperature for 3 h or until T L C 3  2  72 analysis indicates the reaction was complete. Amberlite IR-120(+) resin (proton or pyridinium form) was added until the pH was neutral to litmus. The resin was removed by filtration and washed with MeOH. Solvent removal in vacuo yielded the crude product that was purified by preparative TLC as described above. Formation  of  uridine  5'-(a-D-gIycopyranosyl  diphosphate^  from UMP-  morpholidate. The fully deprotected a-D-glycopyranosyl phosphate was exchanged to either its pyridinium or triethylammonium counterion form and maintained in vacuo overnight. Trioctylamine (1 equivalent) and pyridine (2-5 mL) was added and the solution was concentrated to dryness in vacuo. A second portion of pyridine was added, and the process of dissolution and evaporation was repeated two more times with dry argon being used to equalize the pressure in the flask. UMP-morpholidate (1.5 equivalents) and 177-tetrazole (3.2 equivalents) were added to the flask and the process of dissolution and evaporation with pyridine was repeated twice. The final volume of pyridine (5 mL mmol") was added and the solution stirred at room temperature 1  for 2 days. Solvent was removed in vacuo and the residue was dissolved in water and washed twice with Et20. The product was purified by anion-exchange chromatography and sizeexclusion chromatography as described above. Formation of uridine 5'-(a-D-glycopyranosyl diphosphate^ by enzymatic coupling. This procedure was adapted from Singh et al (124). The a-D-glycopyranosyl phosphate (1.0 mM) was dissolved in the reaction buffer (30-50 mL, 2.0 mM uridine 5'-triphosphate (UTP), 5.0 mM MgS04, 1.0 mM DTT, and 100 mM Hepes, pH 7.5). The reaction was initiated by addition of UDP-glucose pyrophosphorylase (0.81 unit mL" ) and inorganic pyrophosphatase (1.2 unit 1  mL" ). The reaction was monitored by ion-paired reverse phase HPLC with a Radial-pak C-18 1  column with detection at 260 nm (125). At timed intervals, the sample was applied to the column equilibrated with potassium phosphate buffer (100 mM, pH 7.0) containing 5.0 mM  73 tetrabutylammonium hydrogen sulfate. The column was eluted with a linear gradient of acetonitrile (0-50 %) in the same buffer. Once the reaction was complete (as indicated by the HPLC assay), the product was purified by a combination of anion exchange and size-exclusion chromatography as described above.  3.2.3 Specific Synthetic Procedures l,2,3,4-Tetra-0-acetyl-6,7-dideoxy-P-D-g/wc0-hept-6-enopyranose (5). To a solution of alcohol 4 (3.0 g, 8.6 mmol) in dry, distilled DMSO (45 mL) was added l-cyclohexyl-3-(2morpholinoethyl) carbodiimide metho-p-toluenesulfonate (CMC) (11.0 g, 26.0 mmol) and the solution was stirred at room temperature until dissolution was complete. Dichloroacetic acid (1.1 mL, 13.4 mmol) was added via syringe and the solution was stirred for 3 h in a cool water bath. The reaction mixture was poured into ice cold water (250 mL) and thoroughly extracted with ice cold EtOAc (4 x 100 mL). The organic extract was washed with ice cold water (3 x 100 mL) and the pooled aqueous washes were back-extracted once with EtOAc. The organic extracts were pooled, washed (ice cold brine), dried (Na2S04), and concentrated in vacuo to give 1,2,3,4-tetra<9-acetyl-P-D-g/wc0-hexodialdo-l,5-pyranose as a pale yellow foam (2.4 g, 6.8 mmol, 79 %). The product was > 90 % pure by NMR with starting material as the major impurity. The intrinsic instability of the aldehyde due to a facile a,P-elimination prevented further purification so the product was carried on to the next step immediately. The crude aldehyde (2.4 g, 6.8 mmol) was dissolved in C H 2 C I 2 (50 mL) and cooled to 0 °C. Lombardo's reagent (see below), was added via syringe and the reaction was stirred 30 min at 0 °C, poured into saturated N a H C 0 3 (250 mL), diluted with EtOAc (500 mL), and stirred for 1 h. The organic layer was separated and the aqueous layer was filtered through celite before being extracted twice with EtOAc (100 mL). Organic layers were pooled, washed (water then brine), dried (Na2S04), and concentrated in  74 vacuo to give a yellow oil (1.9 g). The crude product was purified by flash chromatography (25 % EtOAc/75 % hexanes) to give the pure alkene 5 as a white solid (360 mg, 12 % from 4). 'H NMR (400 MHz, CDC1 ) 5 5.72 (d, IH, J,, = 8.3, HI), 5.69 (m, H6), 5.33 (d, IH, 3  H7cis), 5.26 (d, IH, J IH, J 6  2>  3  = 9.6, J  2>  1  7trans  2  J cis,6= 7  17.2,  , = 10.5, H7trans), 5.23 (dd, IH, J ,2 = 9.5, J 4 = 9.5, H3), 5.09 (dd, 6  3  3>  = 8.3, H2), 4.93 (dd, IH, J , 5 = 9.7, J , 3 = 9.6, H4), 3.99 (dd, IH, J 4  4  = 7.0, H5), 2.07 (s, 3H, Ac), 2.00 (s, 3H, Ac), 1.98 (s, 3H, Ac), 1.96 (s, 3H, Ac);  l 3  = 9.8, J ,  54  5  C NMR (75  MHz, CDC1 ) 6 170.1 169.4 169.3 168.9 (COCH3, 4Ac), 132.1 (C6), 120.6 (C7), 91.5 (Cl), 76.0 3  72.5 70.9 70.5 (C2, C3, C4, C5), 20.8 20.6 20.6 20.6 (COCH , 4Ac); DCI(+) MS (NH ) m/z 362 3  3  (M + N H , 23 %), 285 (M - OAc, 100 %). Anal, calcd for C 1 5 H 2 0 O 9 : C, 52.33; H, 5.85. Found: +  4  C, 52.36; H, 5.82.  Lombardo's reagent (126). To a suspension of zinc dust (11.5 g, 17.6 mmol) in dry, distilled THF (100 mL) and CH Br (4.05 mL, 8.0 mmol) cooled to -40 °C (acetonitrile/liquid 2  2  N2) is added neat TiCU (4.6 mL, 5.7 mmol) dropwise. The mixture is stirred for 2 h at -40 °C then 2-3 d at 4 °C. The active reagent can be stored for up to several months at -20 °C.  6,7-Dideoxy-a-D-g/KC0-hept-6-enopyranosyl phosphate, disodium salt (6). Compound 5 (150 mg, 0.44 mmol) and Et20 (0.5 mL) were quickly added to crystalline phosphoric acid (300 mg) that had been dried overnight in vacuo over P2O5. The mixture was rapidly stirred and the ether carefully removed in vacuo with slight heating. The neat mixture was maintained in vacuo at 50-55 °C with stirring for 2 h. THF (4 mL) was added, the solution was cooled to -10 °C, and N H 4 O H was added dropwise (11-12 drops) until the pH was neutral to litmus. Ammonium phosphate precipitate was removed by filtration, washed with THF (20 mL), and the solvent removed under reduced pressure to give a dark syrup. The crude product was dissolved  75 in water and washed with  CHCI3.  Several drops of pyridine were added and the water was  removed in vacuo followed by dissolution in toluene and solvent removal in vacuo. This Q  procedure yielded a brown solid (140 mg) which was purified by preparative TLC (60 % % MeOH/5 % H 0) to give 2,3,4-tri-0-acetyl-6,7-dideoxy-a-D-g/«co-hept-6-  CHCI3/35  2  enopyranosyl phosphate as a clear glass (94 mg). This compound was immediately subject to Odeacetylation (NaOMe/MeOH). Following purification by preparative TLC (45 % CF£Cl /45 % 8  3  MeOH/10 % H 2 O ) , and exchange to the sodium counterion , the product was lyophilized to give 6 as a white powder (47 mg, 39 % from 5). *H NMR (400 MHz, D 0) 8 5.86 (ddd, 1H, 2  17.3, J , 6  7tran  s  = 10.3, J ,5 = 7.8, H6), 5.47 (dd, 1H, J,, = 7.2, J,, = P  6  17.1, H7cis), 5.39 (d, 1H, J  7t  rans,6=  (dd, 1H, J , = 9.6, J , = 9.4, 3  2  1H, J , = 9.6, 4  5  3  J ,3 4  4  H-3),  2  3.5,  J , i = 6  7 c  S  HI), 5.46 (d, 1H, J , = 7cis  6  10.3, H7trans), 4.25 (dd, 1H, J 5 ) 4 = 9.4, J 5 , 6 = 8.0, H5), 3.76  3.57 (ddd, 1H, J , = 9.8, J ,, = 3.3, J 2  3  = 9.4, H4); C NMR (75 MHz, D 0) 13  2  2  6  2 )P  = 2.9, H2), 3.32 (dd,  134.3 (C6), 121.2 (C7), 94.7 (d, J , p c  =  5.9, CI), 74.0 73.2 72.7 (C3, C4, C5), 72.0 (d, J , = 7.8, C2); P NMR (121.5 MHz, D 0) 5 31  c  P  2  0.373 (s); HR LSI(-) MS (thioglycerol matrix) m/z calcd for C Hi 0 P 255.02698, found 7  2  8  255.02786 .  Uridine 5'-(a-D-g/«co-hexodialdo-l,5-pyranosyl diphosphate), disodium salt (1). The counterion of compound 6 (10.4 mg, 0.037 mmol) was exchanged to triethylammonium and lyophilized. The compound was dissolved in MeOH (10 mL), cooled to -78 °C (acetone/dry ice and ozone was bubbled through the solution until it turned blue. Argon was bubbled through the solution until the blue color disappeared, dimethyl sulfide (1 mL) was added and the solution was stored overnight at -20 °C. Solvent was removed in vacuo and the product was subject to the  This procedure is described in Section 3.2.1 'Routine Chemical Procedures'.  76 enzymatic coupling reaction with UTP . After 7.5 h, the solution was separated into 2 equal 8  portions and frozen at -20 °C until purification. The aldehyde 1 was purified by a combination of anion-exchange and size-exclusion chromatography followed by exchange to the sodium counterion. The pure aldehyde 1 (sodium salt) was a white powder (6.3 mg, 27 % from 6). 'H NMR (400 MHz, D 0) 5 7.93 (d, IH, J , = 8.1, H6), 5.97-5.96 (m, 2H, HI', H5), 5.63 (dd, IH, 2  JP,P=  6  5  7.6, Ji-,2»= 3.4, HI"), 5.19 (d, IH, J ",5"= 1.5, H6"), 4.36-4.16 (m, 5H, H2\ H3*, H4', H5'a, 6  H5'b), 3.83 (dd, IH, J -, » =10.1, J -, - = 1.4, H5"), 3.76 (dd, IH, J „ „ , J ., „ = 9.5, 9.5, H3"), 5  4  5  6  3  4  3  2  3.54-3.46 (m, 2H, H2", H4"). C NMR (75 MHz, D 0) 5 166.7 (C4), 152.3 (Cl), 141.8 (C6), 13  2  102.9 (C5), 95.7 (d, J , P = 6.3, Cl"), 88.6 (Cl'), 88.2 (C6"), 83.5 (d, J p = 9.1, C4'), 74.0 74.0 C  C>  72.8 70.3 69.9 (C2', C3\ C3", C4", C5"), 71.7 (d, J , p = 8.4, C2"), 65.2 (d, J , = 5.6, C5'); P 31  c  NMR (121.5 MHz, D 0) 6 -10.8 (d, IP, J 2  P > P  c P  20.5, Pa), -12.4 (d, IP, J  =  P  , = P  20.8, Pf3); HR LSI(-  ) MS (thioglycerol matrix) m/z calcd for Ci5H oN Oi P2Na 585.01349, found 585.01370. 2  2  7  l,2,3,4,6-Penta-0-acetyl-7-deoxy-L-g/jcero-P-D-g/Mco-heptopyranose  (8a).  To a  solution of 7a (0.28 g, 1.4 mmol) in Ac 0 (12 mL) was added NaOAc (0.79 g, 9.6 mmol), and 2  the solution refluxed 1 h, cooled to room temperature, and once again heated to reflux in a thermostat controlled oil bath (120 °C). The solution was poured into saturated NaHC03, extracted with EtOAc (3x) and the organic layer was washed (brine), dried (Na S0 ), and 2  4  concentrated in vacuo. Purification by flash chromatography (50 % Et2O/50 % hexanes) yielded an analytically pure mixture of anomers as a white solid (4p:la by 'H NMR, 478 mg, 83 %). Successive recrystallizations from Et 0 gave the pure P-anomer 8a (274 mg, 48 %): [OC]°D = 2  2  15.2 (c = 0.798, CHC1 ); 'H NMR (400 MHz, CDC1 ) 5 5.63 (d, IH, J,, = 8.3, HI), 5.20 (dd, 3  IH, J  3  3>2  3  2  = 9.3, J 4 = 9.3, H3), 5.13 (dd, IH, J 3 = 9.2, J , = 8.3, H2), 5.11 3>  2  2 >  (dd, IH, J 5 = 9.6, J ,  = 9.4, H4), 5.03 (dq, IH, J ,7 = 6.6, J , = 2.0, H6), 3.56 (dd, IH, J , = 9.8, 6  6  5  5  4  4>  J  5 )6  4  = 2.0, H5), 2.10  77 (s, 3H, Ac), 2.04 (s, 3H, Ac), 2.01 (s, 3H, Ac), 1.98 (s, 3H, Ac), 1.98 (s, 3H, Ac), 1.27 (d, 3H,  J  7 >  = 6.6, H7); C NMR (75 MHz, CDC1 ) 8 170.4 170.1 169.3 169.2 169.0 (COCH , 5Ac), 92.2 13  6  3  3  (Cl), 75.7 73.0 70.3 67.5 65.6 (C2, C3, C4, C5, C6), 21.0 20.8 20.6 20.6 20.5 (COCH3, 5Ac),  15.7 (C7); DCI(+) MS (NH ) m/z 422 (M + N H , 100 %), 345 (M - OAc, 32 %). Anal, calcd for +  3  4  C , O H ; C, 50.48; H, 5.99. Found: C, 50.69; H, 6.04. 7  u  2 4  7-Deoxy-L-g/ycera-a-D-g/wco-heptopyranosyl  phosphate,  disodium  salt  (9a).  Compound 8a (100 mg, 0.25 mmol) was treated according to the modified MacDonald procedure described above for the preparation of 6 and purified by preparative TLC (60 % CHCI3/35 % MeOH/5 % H 0) to give 2  2,3,4,6-tetra-0-acetyl-7-deoxy-L-g/ycero-a-D-g/wco-  heptopyranosyl phosphate (47 mg). This compound was immediately subject to O-deacetylation (NaOMe/MeOH). Preparative TLC (45 % CHCl /45 % MeOH/10 % H 0) and exchange to the 3  2  sodium counterion followed by lyophilization gave 9a as a white powder (21 mg, 29 %): 'H NMR (400 MHz, D 0) 5 5.45 (dd, IH, 2  1.5,  H6),  3.71  (dd,  IH, J  3 ( 4  , h,2=  9.5,  3.45 (m, 2H, H2, H4), 1.22 (d, IH, J 6.0, Cl), 72.0 (d,  J ,p= c  7  ,6  J,,P  = 6.9,  9.4,  H3),  Ji, 2 3.60  = 3.4, HI), 4.10 (dq, IH, J , = 6.7, 6  (dd,  IH, J  5 j 4  = 10.1,  J , = 1-4, 5  6  J  7  6 >  H5),  = 6.6, H7); C NMR (75 MHz, D 0) 5 94.8 (d, 13  5 =  3.49J ,P  2  =  c  7.8, C2), 74.4 73.3 70.0 64.7 (C3, C4, C5, C6), 19.1 (C7); P NMR 31  (121.5 MHz, D 0) 8 0.020 (s); HR LSI(-) MS (thioglycerol matrix) m/z calcd for C H 0 P 2  7  14  9  273.03754, found 273.03680.  Uridine 5'-(7-deoxy-L-g/jcm>-a-D-g/wc0-heptopyranosyl diphosphate), disodium salt (2a). Compound 9a (22 mg, 0.074 mmol) was subject to coupling with UMP-morpholidate . 8  Following purification, exchange to the sodium counterion, and lyophilization, this procedure yielded 2a as a white powder (17 mg, 36 %): *H NMR (400 MHz, D 0) 8 7.91 (d, IH, J , = 8.1, 2  6  5  78 H6), 5.96-5.95 (m, 2H, HI', H5), 5.58 (dd, 1H, J H2', H3', H4', H5'a, H5*b), 4.11 (q, 1H, J ", 6  = 7.1, J , - = 3.4, HI"), 4.34-4.16 (m, 5H,  r >P  r  2  = 6.6, H6"), 3.71 (dd, 1H, J », » , h\ r = 9.4, 9.2,  r  3  4  H3"), 3.63-3.59 (m, 1H, H5"), 3.53-3.46 (m, 2H, H2", H4"), 1.25 (d, 3H, J ,6"= 6.6, H7"). C 1 3  r  NMR  (75  MHz, D 0)  5 167.7  2  (C4),  153.0  (C2),  141.7  (C6),  103.0  (C5),  95.9  (d, J , = 6.7, C  CI"),  P  88.7 (CI'), 83.4 (d, J , p = 8.9, C4*), 74.9 74.0 73.4 69.9 69.8 (C2', C3', C3", C4", C5"), 71.9 (d, c  J ,p= c  8.8, C2"), 65.1 (d, J  C  , = P  5.4, C5'), 64.6 (C6"), 19.1 (C7"); P NMR (121.5 MHz, D 0) 8 31  2  10.8 (d, IP, J , = 20.8, Pa), -12.6 (d, IP, J P  P  = 20.7, PP); HR LSI(-) MS (thioglycerol matrix)  P> P  m/z calcd for C i 6 H N 0 P N a 601.04479, found 601.04663. 24  2  17  2  l,2,3,4,6-Penta-0-acetyl-7-deoxy-D-g/jcero-P-D-g/«co-heptopyranose  (8b).  Compound 7b (0.412 g, 2.12 mmol) was subject to complete acetylation as described for the preparation of 8a. Flash chromatography (50 % Et O/50 % hexanes) gave an analytically pure 2  mixture of 8b and its a-anomer (3p:la by 'H NMR, 688 mg, 80 %). Successive recrystallizations (Et 0/hexanes) gave the pure P-anomer 8b (429 mg, 50 %): [a] o = +23.4 (c = 20  2  0.798, CHCI3); 'H NMR (400 MHz, CDC1 ) 8 5.68 (d, 1H, J , , = 8.2, HI), 5.20 (dd, 1H, J 3  9.3,  J  3 >  4  = 9.3, H3), 5.05 (dd, 1H,  H4), 4.90 (dq, 1H,  J ,7 6  = 6.7,  J  J ,5 6  2 >  3  2  = 9.4, J , , = 8.2, H2), 5.00 (dd, 1H, 2  J ,5 4  = 10.0,  J  4 )  3  3 )2  =  = 9.2,  = 2.3, H6), 3.76 (dd, 1H, J , = 10.1, J , = 2.4, H5), 2.09 (s, 5  4  5  6  3H, Ac), 2.03 (s, 3H, Ac), 2.03 (s, 3H, Ac), 2.00 (s, 3H, Ac), 1.98 (s, 3H, Ac), 1.23 (d, 3H,  J , 7  6  =  6.7, H7); C NMR (75 MHz, CDCI3) 8 170.2 170.0 169.6 169.3 168.9 (COCH 5Ac), 91.7 (CI), 13  3  75.4 72.9 70.2 68.7 68.5 (C2, C3, C4, C5, C6), 21.1 20.8 20.6 20.5 20.5 (COCH , 5Ac), 13.5 3  (C7); DCI(+) MS (NH ) m/z All (M + N H , 100 %): 345 (M - OAc, 41 %). Anal, calcd for +  3  4  C 0 , , H ; C, 50.48; H, 5.99. Found: C, 50.56; H, 6.05. 17  24  79 7-Deoxy-D-g/jcero-a-D-g/wco-heptopyranosyl  phosphate,  disodium  salt  (9b).  Compound 8b (100 mg, 0.25 mmol) was treated according to the modified MacDonald procedure described for the preparation of 6. Preparative TLC gave 2,3,4,6-tetra-0-acetyl-7deoxy-D-g/ycero-a-D-g/wco-heptopyranosyl  phosphate (73 mg) which was immediately subject  to O-deacetylation and purification. This procedure gave 9b as a white powder (23 mg, 31 %): 'HNMR (400 MHz, D 0) 5 5.43 (dd, IH, J,, = 7.1, J , , = 3.7, HI), 4.09 (dq, IH, J , = 6.7, J 2  P  2  6  7  6 ) 5  = 2.4, H6), 3.90 (dd, IH, J , = 10.1, J , = 2.3, H5), 3.71 (dd, IH, J , = 9.4, J  3>4  3.48-3.45 (m, IH, H2), 3.32 (dd, IH, J , = 10.0, J , = 9.3, H4), 1.16 (d, 3H, J  = 6.8, H7); C  5  4  5 6  4  NMR (75 MHz, D 0) 8 94.4 (d, J 2  C P  3  5  4  2  3  = 9.4, H3), 13  7 6  = 5.3, Cl), 72.1 (d, J , = 6.9, C2), 74.2 73.6 71.0 66.7 (C3, C  P  C4, C5, C6), 15.5 (C7); P NMR (121.5 MHz, D 0) 8 1.18 (s); HR LSI(-) MS (thioglycerol 31  2  matrix) m/z calcd for C Hi 0 P 273.03754, found 273.03725. 7  4  9  Uridine 5'-(7-deoxy-D-g/ycera-a-D-g/«co-heptopyranosyI diphosphate), disodium salt (2b). Compound 9b (16 mg, 0.053 mmol) was subject to coupling with UMP-morpholidate . 8  Following purification, exchange to the sodium counterion, and lyophilization, this procedure yielded 2b as a white powder (8.6 mg, 26 %): *H NMR (400 MHz, D 0) 8 7.93 (d, IH, J 2  8.1, H6), 5.99-5.98 (m, 2H, HI', H5), 5.60 (dd, IH, J , = 7.3, J r  P  r > 2  6 5  =  - = 3.4, HI"), 4.37-4.17 (m,  5H, H2', H3', H4', H5'a, H5*b), 4.13 (dq, IH, J -, » = 6.7, J ,, - = 2.6, H6"), 3.93 (dd, IH, J », , = 6  10.2,  J -,6" = 5  7  6  5  5  2.4, H5"), 3.75 (dd, IH, J . » = 9.6, J ,. „= 9.3, H3"), 3.52 (ddd, IH, 3 >2  3  4  4  = 9.8, J »,  h-,y  2  I- = 3.3, J „ = 3.1, H2"), 3.39 (dd, IH, J ,, - =10.1, J -, - = 9.2, H4"), 1.21 (d, 3H, J , &• = 6.6, 2  H7").  13  P  4  5  4  3  T  C NMR (75 MHz, D 0) 8 166.6 (C4), 152.1 (C2), 141.8 (C6), 102.9 (C5), 95.8 (d, J 2  6.5, Cl"), 88.6 (CP), 83.5 (d, J , c  C5"), 71.8 (d, J  P  =  = 8.9, C4 '), 74.9 74.0 73.3 70.8 69.9 (C2', C3', C3", C4",  = 8.2, C2"), 65.2 (d, J , P = 5.0, C5'), 66.8 (C6"), 15.6 (C7"); P NMR (121.5 31  C ) P  C > P  c  80 MHz, D 0) 5 -10.8 (d, IP, J , = 20.3, Pa), -12.5 (d, IP, J , 2  P  P  P  P  = 20.8, Pp); HR LSI(-) MS  (thioglycerol matrix) m/z calcd for C ^ ^ O ^ N a 601.04479, found 601.04608.  6,7-Dideoxy-l,2:3,5-di-0-isopropylidene-6-C-methyl-a-D-g/aco-hept-6-enofuranose (11). To a stirring solution of oxalyl chloride (0.182 mL, 4.0 mmol) in CH2CI2 (20 mL), was added DMSO (0.31 mL, 4.0 mmol) in CH C1 (5 mL). This solution was stirred for 2 min at -60 2  2  °C before alcohol 10 (0.50 g, 1.8 mmol) in CH2CI2 (10 mL) was added via dropping funnel and stirred for 15 min at -60 °C. NEt (1.27 mL, 9.1 mmol) was added via syringe and the solution 3  was stirred for 5 min at -60 °C and allowed to warm to room temperature. Water (50 mL) was added, the aqueous solution was extracted (CH2CI2), and the organic layer was washed (brine), dried (Na2S04), and solvent was removed in vacuo. This procedure gave 7-deoxy-l,2:3,5-di-0isopropylidene-a-D-g/uco-hept-6-ulofuranose as a pale yellow oil (0.45 g). *H NMR indicated that this material was > 90 % pure but due to instability to silica gel, this material was carried onto the next step immediately and further purification and characterization were not pursued. To a rapidly stirring slurry of Zn dust (2.6 g, 40 mmol) in THF (40 mL) was added CH2I2 (1.8 mL, 22 mmol). After 30 min, the gray slurry was cooled to 0 °C and 1.0 M TiCl in CH C1 (2.9 4  2  2  mL, 2.9 mmol) was added via syringe. This solution was stirred for 30 min at room temperature before the ketone (610 mg, 2.2 mmol) in THF (7 mL) was added dropwise via syringe. After 1 h, the reaction was diluted with Et20 and 5 % HCI/H2O. The solution was extracted (Et20), and the organic layer was washed (saturated NaHC0 , brine), dried (Na2S04), and solvent removed in 3  vacuo. The residue was purified by flash chromatography (hexanes then 90 % hexanes/10 % EtOAc) to yield 11 as a clear oil (344 mg, 52 %): H NMR (400 MHz, CDC1 ) 5 5.97 (d, 1H, J !  3  = 3.7, HI), 5.05, (s, 1H, H7cis), 4.88 (d, 1H,  J rans,N.D. = 7t  1.0, H7trans), 4.53 (d, 1H, J ,1 = 3.7,  H2), 4.31 (dd, 1H, J , = 7.2, J , = 3.7, H4), 4.17 (d, 1H, J 4  5  4  3  u 2  2  3 > 4  = 3.8, H3), 3.89 (d, 1H, J , = 7.2, 5  4  81 H5),  1.76 (s, 3H, H8), 1.44 (s, 3H, Me C), 1.33 (s, 3H, Me C), 1.32 (s, 3H, Me C), 1.28 (s, 3H, 2  Me C); 2  13  C NMR  (75 MHz,  2  2  CDC1 ) 8 142.5 (C6), 111.9 (CMe , dioxolane ring), 111.6 (C7), 3  2  106.3 (CI), 100.8 (CMe , dioxane ring), 83.7 82.0 75.0 74.3 (C2, C3, C4, C5), 27.1 26.5 24.0 2  23.7 (4Me C), 18.6 (C8); DCI(+) MS (NH ) m/z 271 (M + H , 33 %). Anal, calcd for C 0 H ; +  2  3  14  5  22  C, 62.20; H, 8.20. Found: C, 62.15; H, 8.28.  6,7-Dideoxy-6-C-methyl-D-g/uco-hept-6-enopyranose (12). Compound 11 (330 mg, 1.22 mmol) was dissolved in 50 % TFA/H 0 (30 mL), stirred at room temperature for 3 h, and 2  solvent was removed in vacuo. The residue was redissolved and solvent removed (H 0 then 2  toluene) before purification by flash chromatography (EtOAc then 95 % EtOAc/5 % MeOH then 90 % EtOAc/10 % MeOH). This procedure gave 12 as a white foam (225 mg, 97 %). 'H (400 MHz, D 0) 8 5.20 (d, 0.35H, J , , a  2  (d, 0.65H,  J i  P  )  2  p  2 a  - 3.6, Hla), 5.12-5.10 (m, 2H, H7a/p/cis/trans), 4.65  = 7.8, Hip), 4.18 (d, 0.35H,  3.69 (dd, 0.35H, J  3  a  ,  2  a  = 9.3,  J  3  a  ,  4  a  J ,4a 5 a  = 10.0, H5a), 3.83-3.78 (m, 0.65H, H4p),  = 9.2, H3a), 3.56 (dd, 0.35H, J  3.49-3.43 (m, 1.65H, H3p, H4a, H5P), 3.25 (ddd, 0.65H, H2P),  NMR  J  2 p  ,  2 ( X ; 3 a  , = 7.5, p  - 9.5,  J  J  = 7.5,  2  p  >  3  p  2 a  ,  , - 3.7, H2a), a  J  2  p  >  4  p  = 2.4,  1.74, 1.74 (2s, 3H, H8a/p); C NMR (75 MHz, D 0) 8 141.1 (C6a), 140.9 (C6P), 117.9 13  2  (C7a), 117.7 (C7P), 96.2 (Clp) , 92.5 (Cla), 80.2 75.8 74.5 70.8 (C2p, C3p, C4p, C5p), 75.6 72.9 71.8 71.0 (C2a, C3a, C4a, C5a), 16.7 (C8a/p); DCI(+) MS, (NH ) m/z 208 (M + N H , 39 +  3  4  %). Anal, calcd for C 0 H ; C, 50.52; H, 7.42. Found: C, 50.55; H, 7.48. 8  5  14  l,2,3,4-Tetra-0-acetyl-6,7-dideoxy-6-C-methyl-D-g/«co-hept-6-enopyranose  (13).  The free sugar 12 (270 mg, 1.42 mmol) was subject to complete acetylation as described above for the preparation of 8a. Flash chromatography (70 % hexanes/30 % EtOAc) gave 13 as a white solid (la:2p by *H NMR,  390 mg, 77 %). Recrystallization (Et 0/hexanes) gave the pure P2  82 anomer of 13 (153 mg, 31 %). The P-anomer prepared in this manner was used for all characterization and was subject to the MacDonald procedure described for the preparation of 6. This attempt to introduce the a-phosphate was unsuccessful and it was later determined that the anomeric mix was sufficient for the alternate route described herein, *H NMR (400 MHz, CDC1 ) 5 5.70 (d, IH, J,, = 8.3, HI), 5.23 (dd, IH, J 2 = 9.5, J 3  2  3>  3 > 4  = 9.5, H3), 5.07 (dd, IH, J 3 2>  = 9.5, h, 1 = 8.3, H2), 5.03 (dd, IH, J ,5 = 9.7, J ,3 = 9.6, H4), 4.94 (s, 2H, H7cis/trans), 3.95 (d, 4  IH, J  5>4  4  = 9.8, H5), 2.05 (s, 3H, Ac), 1.99 (s, 3H, Ac), 1.97 (s, 3H, Ac), 1.93 (s, 3H, Ac), 1.70 (s,  3H, H8); C NMR (75 MHz, CDCI3) 8 170.0 169.3 169.3 168.9 (COCH , 4Ac), 139.2 (C6), 13  3  117.0 (C7), 91.5 (Cl), 78.8 72.6 70.4 68.9 (C2, C3, C4, C5), 20.8 20.5 20.5 20.5 (COCH , 4Ac), 3  16.6 (C8); DCI(+) MS (NH ) m/z 376 (M + N H , 34 %). Anal, calcd for C16O9H22; C, 53.63; H, +  3  4  6.19. Found: C, 53.71; H, 6.07.  2,3,4-Tetra-0-acetyl-6,7-dideoxy-6-C-methyl-P-D-g/«co-hept-6-enopyranose  (14). A  solution of the peracetylated sugar 13 (60 mg, 0.17 mmol) and hydrazine acetate (18.6 mg, 0.20 mmol) in DMF (0.6 mL) was heated at 50 °C for 1 min. After 20 min at room temperature, the solution was diluted (EtOAc), washed (brine), dried (Na2S0 ), and the solvent was removed in 4  vacuo. Flash chromatography (60 % hexanes/40 % EtOAc) gave 14 (45 mg, 85 %): H NMR !  (400 MHz, CDCI3) 8 5.54 (dd, 0.7H, J , 4 a , ha, 2a = 9.9, 9.8, H3a), 5.42 (d, 0.7H, J , 3a  Hla), 5.24 (dd, 0.3H, J (dd, 0.7H, J 0.3H,  J  i p  , p 2  2a  ,3a=  3  p  ,  4  p =  9.6,  J p, 3  2  p =  l a  2  a  = 3.6,  9.6, H3p), 5.03-4.94 (m, 3H, H4a/p, H7aba/p), 4.87  7.4, J , .«= 3.6, H2a), 4.84 (dd, 0.3H, J , = 9.7, 2a  2p  = 10.1, Hip), 4.42 (d, 0.7H, J , 5 a  4 a  i p  J  2 P ) 3  = 8.0, H2P), 4.73 (d,  p  = 10.1, H5a), 3.88 (d, 0.3H, J  5  p  > 4  p  = 9.9, H5P),  3.30 (broad s, IH, ClOHa/p), 2.06 (s, 3H, Ac), 1.99 (s, 3H, Ac), 1.94 (s, 3H, Ac), 1.94 (s, 3H, Ac), 1.74 (s, 0.9H, H8p), 1.72 (s, 2.1H, H8a); C NMR (75 MHz, CDC1 ) 8 170.9 170.3 170.2 l 3  3  169.6 165.5 (COCH3, 3Aca/p), 140.3 (C6a), 139.8 (C6P), 116.7 (C7a), 116.6 (C7p), 95.3  83 (Cip), 90.1 (Cla), 78.4 73.4 72.1 69.4 (C2p, C 3 p , C 4 p , C 5 p ) 73.2 71.3 69.6 69.6 (C2a, C 3 a , C 4 a , C 5 a ) , 20.7 20.7 20.6 (COCH , 3Aca/p), 16.7 (C8a/p); DCI(+) MS (NH ) m/z 334 (M + 3  3  N H , 38 %). Anal, calcd for Ci40 H o; C, 53.16; H, 6.37. Found: C, 53.32; H, 6.47. +  4  8  2  7-Deoxy-D-g/MC0-hept-6-ulopyranosyl phosphate, disodium salt (15). To a stirring solution of 14 (82 mg, 0.26 mmol) in C H 2 C I 2 (5 mL) was added lif-tetrazole (73 mg, 1.0 mmol), and dibenzyl A'.A'-diisopropylphosphoramidite (0.22 mL, 0.65 mmol). After 2 h at room temperature, the reaction mixture was diluted (Et 0), washed (ice-cold brine), dried (Na S0 ), 2  2  4  and solvent removed in vacuo. The residue was immediately dissolved in CH C1 (20 mL), 2  2  cooled to -78 °C, and ozone was bubbled through the solution until it turned blue. Argon was bubbled through the solution until the color disappeared, dimethyl sulfide (1 mL) was added and the solution was stored overnight at -20 °C. Solvent was removed in vacuo and the residue was redissolved in 50 % MeOH/50 % EtOAc (20 mL) and 10 % Pd/C (40 mg) was added. The solution was repeatedly degassed in vacuo with H being used to equalize the pressure each time. 2  After stirring 1 h under H , a small amount of cation exchange resin (pyridinium form) was 2  added, the solution was filtered, and concentrated in vacuo. The residue was O-deacetylated with 10 % NEt in MeOH/H 0. Two successive rounds of preparative TLC (45 % MeOH/45 % 3  2  CHC1 /10 % H 0), exchange to a sodium counterion, and lyophilization gave 15 as a white 3  2  powder (3ct:2p by *H NMR, 33.1 mg, 40 %). Material prepared in this manner contained a persistent impurity (10 mol % by H NMR) that is proposed to be iPr NP0 Na : 'H NMR (400 ]  2  MHz, D 0) 8 5.48 (dd, 0.6H, J 2  7.3,  Hip),  4.41  (d, 0.6H,  J  5  a  ,  4  a  l a  ,  P  = 6.7,  = 10.2,  Ji , a  H5a),  2 a  3  = 3.5, Hla), 4.92 (dd, 0.4H, J i 4.06  (d, 0.4H,  J  5  M  P  - 9.7,  2  P  ,  2 p  = 8.0,  H 5 p ) , 3.78  Ji , P  (dd,  P  = 0.6H,  J3«, 2a, ha, 4a = 9.5, 9.3, H 3 a ) , 3.58-3.44 (m, >2H, H2a, H4a, H 3 p , H 4 p , (Me) CH) NP0 Na ), 2  3.35 (dd, 0.4H, J  2 p  ,  3 p  = 8.7,  J , 2 p  2  3  2  , = 8.2, H2p), 2.31 (s, 3H, H7a/p); C NMR (75 MHz, D 0) 8 13  p  2  84 211.1(C6a), 210.3 (C6p), 97.7 (d, J , = 3.7, Cip), 94.8 (d, J , = 4.4, Cla), 79.9, 75.8, 71.4 c  P  c  P  (C3p, C4P, C5p), 76.4, 73.2, 71.6 (C3a, C4a, C5a) , 74.1(d, J , = 5.4, C2p), 71.8 (d, J p = c  5.8, C2a), 28.3 (C7a), 28.1(C7p), 52.8 (d, J , = 4.1, c  P  P  C(  ((Me) CH) NP0 Na2); 2  2  3  31  P NMR (121.5  MHz, D 0) 5 3.07 (s, 0.4P), 2.56 (s, 0.1P), 0.26 (s, 0.6P); HR LSI(-) MS (thioglycerol matrix) 2  m/z calcd for C Hi 0 P 271.02189, found 271.02214. 7  2  9  Uridine 5'-(7-deoxy-a-D-g/MC0-hept-6-ulopyranosyl diphosphate), disodium salt (3). o  Compound 15 (18.5 mg, 0.059 mmol) was subject to enzymatic coupling with UTP . Ion-paired reverse-phase HPLC indicated the reaction had gone to completion after 48 h, at which time the reaction mixture was diluted two-fold with water and purified by anion-exchange and sizeexclusion chromatography. Sodium exchange and lyophilization afforded 3 as a white powder (5.4 mg, 15 %): 'H NMR (400 MHz, D 0) 5 7.91 (d, 1H, J , = 8.1, H6), 5.98-5.93 (m, 2H, HI', 2  H5), 5.62 (dd, 1H, J , = 7.3, J , r  P  r  6  5  = 3.4, HI"), 4.46 (d, 1H, J .., - = 10.2, H5"), 4.35-4.12 (m,  T  5  4  5H, H2', H3', H4*, H5'a, H5'b), 3.80 (dd, 1H, J ", " = 9.6, J " "= 9.3, H3"), 3.57 (ddd, 1H, J ", " = 3  2  3  >4  2  3  9.8, J ", i" = 3.3, J - , = 3.2, H2"), 3.48 (dd, 1H, J „ „ = 10.2, J ", - = 9.2, H4"), 2.31 (s, 3H, (H/D 2  2  P  exchange gives m, <1H), H7");  4  l3  5  4  3  C NMR (75 MHz, D 0) 8 142.0 (C6), 103.1 (C5), 95.7 (d, J ,p 2  c  = 5.0, CI"), 88.8 (CP), 83.6 (d, J , = 6.7, C4'), 76.6 74.2 73.1 71.4 60.1 (C2\ C3', C3", C4", c  P  C5"), 71.6 (d, J , p = 6.3, C2"), 65.3 (d, J P = 4.2, C5'), (C2, C4, C6", and C7" were not c  c>  observed; P NMR (121.5 MHz, D 0) 5 -10.9 (d, IP, J , = 20.2, Pa), -12.7 (d, IP, J 31  2  P  P  Pp); HR LSI(-) MS (thioglycerol matrix) m/z calcd for C e H ^ O ^ N a  PjP  = 20.1,  599.02914, found  599.02682.  3.2.4 General Enzymatic Methods All procedures were performed as described in Section 2.2 unless otherwise noted.  85  3.2.5 Specific Enzymatic Methods Test for irreversible inactivation by UDC. To a solution of UDPGlcDH (1.3 mg mL" ) 1  was added UDC (0.2 mM) and the solution was incubated for 0.5 h at 30 °C before quenching by addition of excess DTT (10 mM). A control sample was treated identically with the exception that no UDC was added. The normal UDPGlcDH assay revealed that the sample incubated with UDC had a specific activity of less than 0.04 units mg" . ESI MS of the UDC treated enzyme 1  showed that the mass of the protein had increased from 45 493 ± 4 to 45 958 ± 4. Both samples were dialyzed (2 d) in separate containers against 1 liter of buffer (50 mM Trien-HCl pH 8.7, 2 mM DTT). After dialysis, the control sample had maintained a specific activity of 1.4 units mg"  1  whereas the sample incubated with UDC still showed a specific activity of less than 0.04 units mg" . 1  Identification of the residue labeled by UDC. To a sample of UDPGlcDH (2.5 mg mL" ) was added UDC (1 mM) and the sample (0.040 mL) was incubated for 0.5 h at 30 °C. A 1  second sample of UDPGlcDH was treated identically but no UDC was added. A saturated solution of urea (pH adjusted to 8.7) was prepared, and an equivalent volume (0.040 mL) was added to both samples. After 1 h at room temperature, iodoacetate (pH adjusted to 8.7) was added to both samples (final concentration 1.0 mM), and the solution was incubated 1 h at room temperature in the dark. DTT was added to afinalconcentration of 5 mM in order to quench the iodoacetate and the samples were stored at 0 °C until ESI MS was performed. The mass of the control sample had increased from 45 493 ± 4 to 45 606 + 4 Da. The mass of the sample treated with UDC had increased from 45 493 ± 4 to 46 009 ± 4 Da.  86  Kinetic characterization of irreversible inactivation by UDC. The inactivation mixture (0.495 mL) containing UDP-xylose (75 uM) and concentrations of UDC ranging from 5 to 17 uM, was prepared in a buffer solution (50 mM Trien-HCl, pH 8.7) that had previously been degassed with argon. This inactivation mixture was added to a small tube equipped with a septum and incubated under argon at 30 °C (5 min). UDPGlcDH (0.005 mL of 2.5 mg mL" ) was 1  added to the sample via syringe, the tube was inverted several times, a 0.050 mL aliquot was removed, and the duration was noted. The aliquot was diluted into the standard assay buffer (0.900 mL) that was lacking NAD and after incubation at 30 °C for at least 5 min, the assay was +  initiated by addition of N A D (500 u.M). A control sample treated in an identical manner +  provided the time zero data point. The initial velocity data was plotted in the form of percentage activity versus time. Rate constants for the decay curves were determined using Grafit and a plot of the observed rate constant versus [UDC] was constructed. This experiment was repeated under identical conditions with the exception that the UDPGlcA (5.06 mM) was included in the inactivation mixture instead of UDP-xylose. The validation of the Bradford assay was performed under similar conditions with 1.3 mg mL" UDPGlcDH and concentrations of UDC ranging from 1  15 to 25 uM.  Inactivation by chloroacetol phosphate. The preparation of chloroacetol phosphate has been previously described (127). The cyclohexylamine salt of the dimethyl acetal of chloroacetol phosphate (2.3 mg) was exchanged to the protonated form, lyophilized, and incubated in D 2 O for 24 hours at 50 °C. *H NMR confirmed that the compound had cleanly converted (> 95 %) to the ketone. Assuming quantitative recovery, a stock solution of 110 uM was prepared. Inactivation kinetics were performed using the same procedure as provided above for UDC.  87 Kinetic characterization of the putative aldehyde intermediate 1. The K  m  and k  cat  values for oxidation of aldehyde 1 were determined under the standard assay conditions described in Section 2.2.2. The reaction was initiated by the addition of UDPGlcDH (0.01 mg) to standard assay buffer that contained the aldehyde 1 (instead of UDP-glucose) at concentrations ranging from 3 to 150 pM. The experiment was repeated under identical conditions with UDPglucose as the substrate. The stoichiometry of the enzymatic oxidation of 1 was assessed by addition of UDPGlcDH (0.05 mg) to the standard assay buffer (1 mL) containing either 1 or UDP-glucose (40 pM). The sample was monitored at 340 nm until the reaction reached completion and the total change in absorbance was used to calculate the concentration of NADH produced. Aldehyde 1 was also tested as a substrate for reduction by UDPGlcDH in the presence of NADH. To a buffered solution (50 mM Trien-HCl, pH 8.7, 2 mM DTT) containing NADH (150 pM) and 1 (74 pM), was added UDPGlcDH (0.04 mg mL" ). The progress of the reaction 1  was monitored at X = 340 nm and was monitored until no further change in absorbance was observed (> 100 min). An attempt to determine the kinetics constants for the reduction was performed under similar conditions with varying concentrations of 1 (5 to 200 pM) and a lower concentration of UDPGlcDH (0.025 mg mL"). 1  Identification of the product of UDPGlcDH catalyzed oxidation of 1. The UDPGlcDH catalyzed oxidation of 1 was monitored by ion-paired reverse phase HPLC with a Radial-pak C-18 column with detection at 260 nm (125). The sample was applied to the column equilibrated with potassium phosphate buffer (100 mM, pH 7.0) containing 5.0 mM tetrabutylammonium hydrogen sulfate and eluting with a linear gradient of 0-50 % acetonitrile in the same buffer. UDPGlcDH (0.05 mg) was added to standard assay buffer (0.43 mL) containing 1 (200 pM) and NAD (500 pM). Aliquots (0.050 mL) were removed at timed intervals and +  88 analyzed as described. Once the reaction had gone to completion, the sample was spiked with UDPGlcA (200 pM) and NADH (200 pM) in order to determine the identity of the new peaks in the chromatogram.  Kinetic characterization of 2a and 2b. In order to test 2a and 2b as substrates, UDPGlcDH (1.4 mg mL- ) was added to assay buffer (0.60 mL) containing either 2a or 2b and 1  NAD (10.0 mM) and the reaction monitored at 340 nm. Both 2a and 2b were also tested as +  substrates under similar conditions with NH OH (25 mM) included in the assay buffer. In order 2  to measure the kinetic constants for turnover of 2a, UDPGlcDH (0.48 mg mL") was added to 1  assay buffer (0.60 mL) containing NH OH (25 mM), NAD 2  +  (10 mM), and varying  concentrations of 2a (0.020 to 1.0 mM).  K\ determination for 3. For the K, determination of 3, the enzymatic reaction was initiated by addition of UDPGlcDH (0.01 mg mL") to assay buffer (1.0 mL) containing NAD 1  +  (500 pM), changing fixed concentrations of 3 (7.5 to 29 pM), and varying concentrations of UDP-glucose.  Identification of the product of UDPGlcDH catalyzed reduction of 3. In order to determine the product of the enzymatic reduction of 3, UDPGlcDH (1.1 mg mL") was incubated 1  at 30 °C with NADH (10 mM) and 3 (1.0 mM) in a total volume of 0.380 mL assay buffer. Aliquots were removed at timed intervals and analyzed by ion-paired reverse-phase HPLC as described above for the identification of the products resulting from oxidation of 1 (125). Peaks were identified by separately spiking aliquots with 3, 2a, 2b, and NAD . +  89  3.3 RESULTS For the sake o f clarity, all discussion o f synthetic procedures w i l l be confined to the R E S U L T S section. The D I S C U S S I O N section w i l l be reserved for conclusions drawn from the testing and application o f the synthetic compounds that are relevant to the mechanism o f UDPGlcDH.  3.3.1 Synthesis of uridine 5'-(a-D-g/«co-hexodialdo-l,5-pyranosyl diphosphate) (1). The strategy for the synthesis o f 1 was inspired by a relatively recent paper that described the synthesis o f thymidine diphospho-6-deoxy-a-D-n°Z?o-3-hexulose and utilized a terminal alkene as the 'protecting group' for a ketone (128). In the final step o f the synthesis, the ketone could be selectively generated from the terminal alkene by ozonolysis in the presence o f a thymine base that differs from uridine only by a single methyl group. This precedent suggested that masking the aldehyde as a terminal alkene and deprotecting either before or after nucleotide coupling, would be a suitable strategy for the synthesis o f 1. A commitment to acetyl protecting groups for all other hydroxyl groups limited the suitable methods for both oxidation and carbonyl methylenation due to the well precedented facile P-elimination shown in Figure 3.9 (see Refs 129,130 and references therein).  AcO  AcO  Figure 3.9 Facile p-elimination in l,2,3,4-tetra-0-acetyl-P-D-gluco-hexodialdo-l,5-pyranose.  A common route to convert a primary alcohol to an aldehyde and then mask it as a terminal alkene might involve a Swern oxidation (131) followed by a Wittig methylenation  90 (132). Both of these reactions are performed under basic conditions so an alternative was sought. Fortunately, a literature example was found in which the Moffat procedure (133, 134, 135) had been utilized to generate an aldehyde at C6 of a glucose derivative with a P-acetyl protecting group (136). This aldehyde was subsequently transformed to the terminal alkene using the methylenation procedure of Oshima (137). The precedent provided by these literature examples served as the basis for the synthetic route to 1 that is shown in Figure 3.10.  HO OH  Figure 3.10 The synthetic route to 1. (i)CMC, DMSO, Cl HCCOOH, (ii) CH Br -Zn-TiCl , (iii) H 3 P O 4 , 55 °C then NaOMe/MeOH, (iv) 0 , -78 °C then Me S, (v) UTP, UDP-glucose pyrophosphorylase, pyrophosphatase. 2  3  2  2  4  2  Oxidation of the C6 primary hydroxyl of the known tetracetylated glucose 4 was achieved by a modification of the standard Moffat procedure (133, 134, 135). A side product of this Moffat procedure is the urea derivative of AyV'-dicyclohexylcarbodiimide (DCC) that is  91 notoriously difficult to separate during the purification. The water soluble coupling reagent 1cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-/>toluenesulfonate (CMC, see Figure 3.11) was used in place of DCC because the resulting water soluble urea derivative can be readily separated by a simple extraction (138).  Figure (CMC).  3.11 l-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide  metho-p-toluenesulfonate  Using this modified Moffat procedure the aldehyde 1,2,3,4-tetra-O-acetyl-p-D-glucohexodialdo-l,5-pyranose could be prepared in approximately 80 % yield with purity > 90 % as judged by 'H NMR. While researching the Oshima methylenation (137), it was found that two variants of the original procedure had since been developed; the Takai procedure (139,140), and the Lombardo procedure (141). Both of these modifications result in more reactive reagents that have been demonstrated to be useful for the methylenation of base sensitive and easily enolizable ketones (142). The method of Lombardo was chosen as this procedure has been suggested to be most appropriate for the methylenation of base sensitive carbonyls (142). Treatment of the crude acetylated aldehyde with Lombardo's reagent followed by purification by flash chromatography yielded the acetylated alkene 5 in 12 % yield from the starting material. In order to introduce the a-phosphate, a modification (143) of the method of MacDonald (144) was utilized as it was found that alkene moiety was compatible with the harsh conditions of this transformation (neat phosphoric acid, 55 °C). This modified procedure allowed the stereospecific preparation of the fully acetylated a-phosphate that was immediately deacetylated with NaOMe/MeOH to give the alkene 6 in 39 % yield. The aldehyde functionality was  92 unmasked by subjecting 6 to ozonolysis at -78 °C followed by overnight treatment with dimethyl sulfide at -20 °C. The most common synthetic route to UDP-sugars was developed by Moffat and Khorana (145) and is achieved by coupling of the phosphorylated sugar with UMP-morpholidate in pyridine. Improved yields for this procedure are obtained in the presence of lif-tetrazole (146). However, due to concerns about the stability of the aldehyde moiety under the conditions of the chemical coupling reaction, an alternate means of performing the same transformation was sought. The mildest route to the product 1 took advantage of the UDP-glucose pyrophosphorylase catalyzed coupling reaction with UTP (124). The normal reaction catalyzed by this enzyme is coupling of glucose 1-phosphate and UTP to form UDP-glucose and pyrophosphate. It was expected that the aldehyde 6 could replace glucose 1-phosphate in this reaction because the steric differences between them are slight and remote from the atoms involved in the chemical transformation. As expected, when the enzymatic coupling reaction was followed by reverse phase HPLC (125), it was observed that one equivalent of UTP was cleanly converted into a new species that was later identified as the desired product 1. Purification by a combination of anion-exchange and size-exclusion chromatographies gave the  final product 1 in a yield of 44 % from 5. The 'H NMR of 1 in D 2 O clearly shows the alde hydrogen as a doublet (J = 1.3 Hz) at 8 - 5.17 (see Appendix B). This indicates that > 95 % of the aldehyde is present as the gera-diol species (see Figure 3.12). This conclusion is further supported by previous study that concluded that l,2:3,4-di-0-isopropylidene-a-D-galactohexodialdo-l,5-pyranose exists primarily as the gem-diol in D 2 O based on a *H NMR chemical shift of 8 = 4.95 (doublet, J = 1.7 Hz) for the aldehyde hydrogen atom (147).  93  H 0  HO-^V^  2  °\  HO-\^--^A HO I OUDP  HO I  w  OUDP  g e m - d i o l of 1 (> 9 5 %)  1 (< 5 %)  Figure 3.12 Equilibrium between 1 and the gem-diol of 1 in aqueous solution.  3.3.2 Synthesis of UDP-(6tf and 6S)-6C-methylglucose (2a and 2b) As mentioned above, the efficient preparation of both 6S, and 6R, -6C-methylglucose (7a and 7b) has been reported by Bleriot et al. (119, 120). This non-diastereoselective synthesis allows preparation of both epimers since they are separable by silica gel chromatography when protected as diacetonides. Deprotection with aqueous trifluoroacetic acid affords the free sugars 7a and 7b that serve as the starting material for the synthesis of both 2a and 2b reported here (see Figure 3.13).  K 2  -z  OH  0)  HO-^\-~-^-^°\ HO  (ii) OAc  OH  AcO  7a,b  OP0 Na  8a (48 %) 8b (50 %) R  a, R, =H, R = C H 2  .OH  HO^V^-^°\ H O A ^ A HO n  u  2a (36 %) 2b (26 %)  3  2  oii)  3  b. R ^ C H . 3 , R = H 2  2  9a (28 %) 9b (31 %)  NH  o  O II -P O — P - O —IP - 0 ONa ONa  1  11  HO O H  Figure 3.13 The synthetic route to 2a and 2b. (i) NaOAc, Ac 0, (ii) H P0 , 55 °C then NaOMe/MeOH, (iii) UMP-morpholidate, l//-tetrazole, pyridine. 2  3  4  94 The free sugars were fully acetylated by refluxing with sodium acetate in acetic anhydride and the anomeric (a:p = 1:3) mixture of each sugar was purified by silica gel chromatography. In both cases, the P-anomer could be selectively crystallized from the anomeric mixture to yield 8a (48 %) and 8b (50 %). To confirm the stereochemical assignment made in the previous work, the structure of 8b was established by X-ray crystallographic analysis (119, 120) as described in Appendix C. The X-ray structure confirmed the R stereochemistry at the C6 position of 8b as shown in Figure 3.14.  Figure 3.14 The X-ray structure of 8b. The a-phosphate was introduced by subjecting both 8a and 8b to the modified MacDonald procedure (143) immediately followed by (9-deacetylation with NaOMe/MeOH. This procedure gave the epimeric fully deprotected a-phosphates 9a and 9b in yields of 49 % and 60 % respectively. Thefinalcoupling to form the UDP sugar was performed using the  95 modified Moffat and Khorana procedure (145) with li/-tetrazole as a catalyst (146). Purification by ion-exchange and size-exclusion chromatographies followed by exchange to the sodium salt gave 2a (26 %) and 2b (36 %). The UDP-glucose pyrophosphorylase catalyzed enzymatic coupling of either 9a or 9b with UTP was not attempted although this alternate route would have likely resulted in similar yields.  3.3.3 Synthesis of UDP-6C-methyl-6-ketoglucose (3) The strategy for the synthesis of the ketone 3 (Figure 3.15) involved protecting the ketone functionality as a terminal alkene in a manner similar to that employed for the synthesis of the aldehyde 1 (refer to Figure 3.10). The known alcohol 10 (120) was oxidized to the methyl ketone using Swern conditions (131) and the crude product was immediately converted to alkene 11 in 52 % yield using the Takai procedure (139, 140). The purification and characterization of the intermediate methyl ketone was not pursued due to an apparent instability to silica gel. Alkene 11 was deprotected with aqueous trifluoroacetic acid to give 12 in 97 % yield. Interestingly, it had previously been reported that an immediate protected precursor to 12 is generated by reacting the tris(tert-butyldimethylsilyl) derivative of D-glucuronolactone with 2 equivalents of Tebbe reagent (148).  96  15 (40%)  HO OH  Figure 3.15 The synthetic route to 3. (i) Swern oxidation (ii) Zn, CH2I2, TiCU, (iii) TFA, H2O, (iv) NaOAc, Ac 0, (v) NH NH AcO", (vi) /-Pr NP(OBn) , 1/Y-tetrazole, then 0 , then Me S, then H2, Pd/C, then 10 % NEt , (vii) UTP, UDP-glucose pyrophosphorylase, pyrophosphatase. +  2  2  3  2  2  3  2  3  The free sugar 12 was peracetylated with acetic anhydride and sodium acetate to give the anomeric mixture of 13 (a:|3 = 1:2) in 77 % yield. The (3-anomer was selectively crystallized from this anomeric mixture and a modified MacDonald procedure (143) was attempted in an effort to introduce the a-phosphate. Unfortunately, this approach was unsuccessful, possibly due to addition of a proton to the alkene resulting in formation of a tertiary carbocation at the C6 position as represented in Figure 3.16.  97  H  3  C ^  C  H  H P0 ,55°C  2  AcO  3  OAc  4  H C 3  —*• A c O - ^ AcOOP0 " 3  (3-anomerof 13  H*'9 Decomposition AcO  •OAc  Figure 3.16 Attempted MacDonald reaction on the P-anomer of 13. As it was not possible to directly introduce the cc-phosphate via the MacDonald procedure, a much more laborious and less elegant route of introducing the anomeric phosphate was undertaken. The anomeric mixture of 13 was treated with hydrazine acetate to selectively deprotect the Cl hydroxyl and give the hemiacetal 14 in 85 % yield. Compound 14 was treated with dibenzyl 7V,A-diisopropylphosphoramidite in the presence of l//-tetrazole to form the r  glycosyl dibenzylphosphite (149). Ozonolysis of the crude glycosyl phosphite was used to achieve both the unmasking of the ketone at C6 as well as oxidation of the dibenzylphosphite to the corresponding dibenzylphosphate. Immediate deprotection of the dibenzylphosphite by hydrogenation over Pd/C followed by O-deacetylation with aqueous NEt3 and purification by preparative TLC gave the anomeric mixture of 15 (a:P = 3:2) in 40 % yield from 14. Separation of the a and p anomers of 15 was not possible and it was very probable that the anomer mixture of UDP-sugars (3 and its P-anomer) resulting from a non-selective chemical coupling procedure would be equally difficult to separate. For this reason, the enzymatic coupling catalyzed by UDP-glucose pyrophosphorylase was used to form the final product 3. Our experience with this enzyme suggested that it could tolerate substrate modifications at the C6 position of the glucosyl a-phosphate as demonstrated by its ability to catalyze the formation of the aldehyde 1. It also  98 seemed probable that due to the exquisite stereoselectivity of enzymatic reactions, the ct-anomer could be selectively coupled out of the anomeric mixture of 14 to yield 3 but not the corresponding (3-anomer. When the enzymatic coupling procedure was attempted with 14, reverse phase HPLC indicated a single new uridine containing species that was later identified as the desired target ketone 3. Purification by a combination of ion-exchange and size-exclusion chromatographies gave 3 in a yield of 15 %.  3.3.4 Evaluation of UDC as an irreversible inhibitor of UDPGlcDH Incubation of UDC with UDPGlcDH resulted in the rapid and irreversible loss of enzymatic activity that could not be recovered by extensive dialysis. ESI MS analysis of the inactivated UDPGlcDH showed the mass of enzyme incubated with UDC is 465 ± 8 Da greater than the mass of UDPGlcDH alone. This increase in mass correlates with the addition of one molecule of UDC (MW = 495 Da) accompanied by loss of one chlorine atom (MW = 35.5 Da) and one proton of UDPGlcDH, resulting in an expected mass increase of 459 Da. These results demonstrate that UDC specifically alkylates UDPGlcDH at a residue that is critical for the enzymatic activity. The thiol of a cysteine residue is the most nucleophilic functional group that is commonly found in proteins and thus a likely candidate for labeling with UDC. It is well established that haloacetate groups (such as the ot-chloroketone of UDC) can readily alkylate the thiol of cysteine (150). UDPGlcDH from S. pyogenes has only two cysteine residues, Cys 162 and Cys 260. When a sample of UDPGlcDH was denatured in 5 M urea and incubated with excess iodoacetate (MW =186 Da), the mass of the enzyme increased by 113 + 8 Da as analyzed by ESI MS. This additional mass is consistent with attachment of two iodoacetate units accompanied by loss of two iodine atoms (MW = 126.90 Da) and two enzyme protons to give an  99 expected mass increase of 116 Da. When the same experiment was performed with a sample of UDPGlcDH that had previously been incubated with UDC, the mass of the enzyme increased by 516 ± 8 Da. The additional mass can be attributed to addition of both UDC (expected mass increase of 459 Da) and one unit of acetate (expected mass increase of 58 Da) resulting in an expected mass increase of 517 Da. Taken together, these results demonstrate that UDPGlcDH contains two reactive nucleophiles, one of which is the site at which UDC becomes covalently attached to the enzyme. The well precedented reactivity of iodoacetate towards cysteine thiols (150) supports the conclusion that these two nucleophilic sites are probably Cys 162 and Cys 260, the two cysteine residues that are present in UDPGlcDH. In order to determine the kinetics of UDC inactivation, it was necessary to take certain measures to ensure that a reasonable estimate of kinetic parameters could be obtained. The difficulties stemmed from the very reactive nature of the a-chloroketone moiety and the limits of the kinetic assay. Thefirstproblem that had to be overcome was the reactivity of UDC towards the reducing agent, DTT that is included in all enzyme buffers to maintain reducing conditions. To circumvent this undesired side reaction, DTT was excluded during the inactivation experiments. To maintain a reducing environment in the absence of DTT, all buffer solutions were thoroughly degassed and the experiment was performed under an atmosphere of argon. The second and more fundamental problem was that the exceptional potency of UDC resulted in a rate of inactivation that exceeded the limitations of the assay. An irreversible inactivation experiment is usually performed by incubating the enzyme of interest with the affinity label and assaying aliquots of the mixture at timed intervals. The activity of the enzyme should decrease in a time dependent fashion that obeys the rate law for afirstorder reaction. When UDPGlcDH was incubated with UDC, it was found that even at concentrations of UDC less than the concentration of enzyme, the inactivation was too rapid to obtain meaningful  100  kinetic values. The quantity of UDPGlcDH present in the inactivation mixture was limited to the minimum value that would provide a significant initial velocity once diluted into the normal assay buffer. One option to circumvent this problem would involve performing the inactivation at a lower temperature, thereby slowing the rate of inactivation. A second option that would allow the rate of inactivation to be decreased in a predictable fashion was to perform the inactivation in the presence of a known competitive inhibitor that would act as a 'protecting agent'. The kinetic equations describing irreversible inactivation both in the absence and in the presence of a 'protecting agent' (P) with an inhibition constant of K\, are explained in detail in Appendix A (Section A.3). If inactivation by the affinity label (I) occurs by a reversible binding step described by K„ followed by an irreversible step with rate constant k , a 'protecting agent' x  will lower the observed rate of inactivation (k bs) according to Equation 0  kbb.  -  1.  E  [I] + K,  1 +  q  l  [Pf  The inactivation was performed in the presence of both 5.1 mM UDPGlcA (K? = 200 pM) and 75 pM UDP-xylose (K, = 2.7 pM) at concentrations of UDC rangingfrom5.5 to 17 p  pM. The data was plotted as percentage remaining activity against time as shown in Figure 3.17 and Figure 3.18. Thefirstorder rate constant (& bs) was extracted from a best-fit line and plotted 0  against the concentration of UDC. The effect of the 'protecting agent' diminishes as the concentration of UDC is increased and thus it was only possible to obtain meaningful kinetics for concentrations of UDC that were well below the value of K\. The slope of the straight line that results when k bs is plotted against [UDC] (see inset, Figure 3.17 and Figure 3.18) is the 0  observed second order rate constant for inactivation, k\l(K (\ + [P]/K{ ). x  ?  101  Time (min) Figure 3.17 Inactivation of UDPGlcDH by UDC in the presence of UDPGlcA. The inactivation mixture contained 5.1 mM UDPGlcA and UDC concentrations of 5.5 pM (O), 7.5 uM (•), 10 uM (•), 13 pM (•), and 17 uM (A). Shown in the inset is a replot of & b s versus [UDC]. 0  T  Time (min) Figure 3.18 Inactivation of UDPGlcDH by UDC in the presence of UDPX. The inactivation mixture contained 75 uM UDP-xylose and UDC concentrations of 5.5 uM (O), 7.5 pM (#), 10 uM (•), 13 uM (•), and 17 pM (A). Shown in the inset is a replot of & b s versus [UDC]. 0  Correcting for the effect of the 'protecting agent' resulted in second order rate constants (k/Ki) of 1.6 x 10 mM" min" and 2.1 x 10 mM" min" with UDPGlcA and UDP-xylose 3  1  1  3  1  1  102 respectively. Considering the intrinsic difficulties in obtaining these values of k\IK\, they are in reasonable agreement and serve as a reliable estimate of the rate of inactivation. To ensure that UDC is an affinity label and the UDP portion is serving as a handle for recognition, the compound chloroacetol phosphate (127) was tested as an inactivator of UDPGlcDH. Chloroacetol phosphate retains the reactive a-chloroketone moiety and the (3phosphate of UDC but lacks the UMP 'handle' that is likely critical for binding and recognition of UDPGlcDH.  O O  O"  Figure 3.19 Chloroacetol phosphate.  To directly compare the efficacy of chloroacetol phosphate and UDC, UDPGlcDH was incubated in the presence of 17 uM chloroacetol phosphate and 5.1 mM UDPGlcA. Over a period of 10 min, there was no detectable loss in activity compared to a control sample. In comparison, the analogous experiment performed with UDC results in loss of 50 % of the enzyme activity within 1 min (see Figure 3.17). Even when the concentration of chloroacetol phosphate was increased to 2.0 mM (in the presence of 5.1 mM UDPGlcA), only 25 % of the activity was lost over a 10 min period. In the absence of UDPGlcA, 2 mM chloroacetol phosphate caused the inactivation of UDPGlcDH with a half-life of approximately 1 min. These results show that the terminal UMP functionality is critical for the potency of UDC and support the conclusion that UDC is a specific affinity label of UDPGlcDH as opposed to a non-specific alkylating agent. An unexpected result that was obtained with UDC was the validation of the Bradford assay (88) that is routinely used to determine the concentration of UDPGlcDH. This assay can be  103 calibrated for almost any buffer system and will provide reproducible results for most soluble proteins. However, the accuracy of this assay is normally difficult to assess and not questioned, even though in rare cases the estimated concentration for the protein may be in error by as much as an order of magnitude. When UDPGlcDH (1.3 mg mL" , 29 pM, as determined by the 1  Bradford assay) was incubated with 15, 20, and 25 uM UDC, the rapid loss of 58 %, 74 %, and 92 % respectively, of the total activity was observed. This result validates the Bradford assay for use with UDPGlcDH and suggests that for this particular system, enzyme concentrations determined with this assay are accurate to within approximately 10 %.  3.3.5 Evaluation of the aldehyde 1 as a substrate for UDPGlcDH. In order to identify the product of the UDPGlcDH catalyzed oxidation of 1, the reaction products were analyzed by ion-paired reverse phase HPLC. A solution containing 1 and a saturating concentration of NAD was prepared in the normal assay buffer. Prior to addition of +  UDPGlcDH, the sample eluted as a single peak (t = 11 min) corresponding to both 1 and NAD . +  UDPGlcDH was added to the assay and after 30 min the sample eluted as 3 peaks (t = 11, 16, and 20 min), one of which corresponded to the unreacted NAD (t = 11 min). UDPGlcA (200 +  uM) was added to the sample and another aliquot was applied to the column. The sample again eluted as three peaks, however the second peak (t = 16 min) had doubled in area, identifying this peak as UDPGlcA and the third peak (t = 20 min) as NADH. The stoichiometry of the UDPGlcDH catalyzed oxidation of 1 was assessed by quantitating the total NADH produced when a known concentration of 1 was completely converted to product. It was found that for each molecule of 1 that is oxidized, one molecule of NAD is reduced to form NADH. Accordingly, the same experiment performed with UDP+  glucose gives a ratio of two molecules of NAD reduced for every molecule of UDP-glucose +  104 oxidized. The kinetic constants for the oxidation of 1 were determined under the normal assay conditions. The oxidation of aldehyde 1 to the UDPGlcA appeared to follow typical saturation kinetics with k 9  =1.0 s" and K = 14 pM. The kinetic constants for oxidation of UDP-glucose 1  caX  m  under the identical conditions were k = 1.2 s" and K 1  cat  m  = 14 pM. These experiments were  repeated three times for both 1 and UDP-glucose, and the kinetic constants were reproducible to within 10 %. A direct plot of the kinetic data for a typical experiment is shown in Figure 3.20.  [UDP-glucose or 1] (mM) Figure 3.20 Determination of kinetic constants for the oxidation of 1 (O) and UDP-glucose (•). Data is represented in the form of a direct plot with the initial velocity (y-axis) uncorrected for reaction stoichiometry.  It is interesting to note that there was no evidence for a dismutation (Section 1.1.5) when 1 was assessed as a substrate for oxidation by UDPGlcDH in the presence of NAD . However, +  when UDPGlcDH was tested for its ability to catalyze the reduction of 1 to UDP-glucose in the presence of NADH, a dismutation did occur as shown in Figure 3.21.  Marginal improvements in the fit of the kinetic data could be made if a correction for potential weak allosteric effects was included. These subtle corrections (Hill coefficient n = 1.5 for UDP-glucose and n = 1.3 for 1) would not change the values of A: by more than 10 %. 9  cat  105 160  T  |  r—r—i  |—i—|—i  |  i  |  i  r  J  i  I  I  i  I  i  L  150  S 140 ffi  120 110  0  20  I  i  i  I  i  40  60 80 100 120 140 Time (min) Figure 3.21 Time course for the UDPGlcDH catalyzed dismutation of 1 in the presence of NADH. The y-axis has been corrected to [NADH] concentration. It appears that when 1 is incubated with UDPGlcDH and NADH, every molecule of NAD that results from the reduction of 1 is then used in the oxidation of a second molecule of +  1. As the reaction proceeds, NAD is initially being produced at a faster rate than it is being +  consumed in the oxidation of 1, and therefore it accumulates in solution. In the experiment shown in Figure 3.21, the concentration of NADH decreases for approximately 10 min as the concentration of NAD increases to a maximum concentration of about 30 pM. At this point the +  rate of oxidation and reduction of 1 are momentarily equivalent. The concentration of NADH then increases back to its initial value. An effort was made to measure the kinetics constants for the UDPGlcDH catalyzed reduction of 1, however the dismutation precluded the reliable determination of these values. As discussed below (Section 3.4.4), the dismutation is not a significant factor with Cys260Ala UDPGlcDH, and the kinetic constants obtained for the reduction of 1 with this mutant may be a more reliable estimate of the actual values (83, 84).  106 3.3.6 Evaluation of the secondary alcohols 2a and 2b as substrates for UDPGlcDH When both 2a and 2b were tested as substrates for UDPGlcDH, it was observed that 2a was a substrate while 2b showed no appreciable turnover even at very high enzyme concentrations (1.4 mg mL ). The product of this oxidation is presumably the corresponding -1  ketone 3. As shown in Figure 3.22, the relatively slow turnover of 2a appeared to reach equilibrium after less than 5 % of the substrate was consumed.  U '  I  '  I  i  I ' I ' I ' I ' I ' I -I  Time (min) Figure 3.22 Time course of UDPGlcDH catalyzed turnover of 2a and 2b. UDPGlcDH (0.9 mg mL ) was incubated with NAD (10 mM) and either 2a, 2b (1 mM) or buffer (control). The extent of turnover of 2a corresponds to approximately 30 pM from 1.0 mM initial concentration (3 %). -1  To further investigate this low extent of turnover, 25 mM hydroxylamine ( N H 2 O H ) was included in the assay buffer to shift the equilibrium for the reaction towards the putative ketone product 3, through formation of an imine. As expected, only 2a was a substrate for UDPGlcDH (see Figure 3.23) when the assay was performed in the presence of  NH2OH.  The change in  absorbance remained linear for over 3 hours and the reaction appeared to go to greater than 10 % completion, eliminating the possibility that the observed turnover was due to an impurity in 2a. As discussed in Section 3.4.3, it was ultimately concluded that the low apparent extent of  107 turnover was due to formation of a product that was a good inhibitor of UDPGlcDH and could compete with 2a for binding to the enzyme.  20  0  40  60  80  Time (min) Figure 3.23 Time course of UDPGlcDH catalyzed turnover of 2a and 2b in the presence of NH OH. UDPGlcDH (0.5 mg mL" ) was incubated with NAD (10 mM) and either 2a, 2b (0.5 mM) or buffer (control) in a buffered solution containing NH OH (25 mM). 1  +  2  2  Interestingly, under these assay conditions there appears to be a reproducible 'burst' of NADH (6-8 pM) over the first 10 to 20 min of the kinetic assay followed by a steady state production of NADH. The 'burst' is approximately stoichiometric with the UDPGlcDH concentration (11 uM), particularly when one considers the extended time period of this assay and the expected poor stability of the enzyme under these conditions (see Section 2.3.3). The presence of this 'burst' suggests that product release (presumably of the ketone 3) is the ratelimiting step of the reaction. The kinetic constants for the turnover of 2a were determined in the assay buffer containing NH OH. These experiments were complicated by the 'burst' phase that 2  could extend over hundreds of minutes at low substrate concentrations. The best estimate of the 10  Due to limited quantities of 2a, this experiment was not repeated. A cautious estimate of the relative error value for the values of k and K , is ± 50 %. 10  cat  m  108 kinetic constants for the UDPGlcDH catalyzed turnover of 2a are k t — 4 x 10" s" and/C =100 4  1  ca  m  uM.  3.3.7 Evaluation of the ketone 3 as both a substrate and inhibitor of UDPGlcDH The ketone 3 was determined to be good inhibitor (K\ = 6.7 pM) of UDPGlcDH, though suprisingly not as potent as UDP-xylose (K\ = 2.7 pM). The reversible inhibition plot for 3 is depicted in Figure 3.24.  0  20 40 60 80 [UDP-glucose]" (mM)" 1  1  -10 0 10 20 30 40  [3]  (MM)  Figure 3.24 Inhibition pattern for the ketone 3 with UDP-glucose as variable substrate. Fixed concentrations of ketone 3 are 0 pM (A), 7.5 uM (A), 10 pM (•), 15 uM (•), 21 pM (•), and 29 uM (O). The graph on the right is a replot of slopes versus [3].  Due to the low extent of UDPGlcDH catalyzed oxidation of 2a, it was not possible to identify the product of the reaction, though the ketone 3 was the prime suspect. It was reasoned that it should be possible to confirm that 3 is the product by running the reaction in the reverse direction (reduction of 3) and analyzing the products by reverse phase HPLC. This experiment would not have been feasible if authentic standards of the two possible products (2a and 2b) of reduction of 3 had not been available. The results from this experiment are shown in Figure 3.25a,b.  109  0.3  a) t =  0  a §  tN +-»  0.2  NAD  C3  <D o C j3 0.1  +  o C/3  0.0 1.4  a  b) t = 12h  1.2  c o 1.0 0.8 O  NAD /2a +  0.6  o 0.4 C/3  0.2 0.0 5.0  J. V 6.0  7.0  8.0  9.0  10.0 11.0  12.0  Time (min) Figure 3.25 Ion-paired reverse phase HPLC experiment to determine the product of UDPGlcDH catalyzed reduction of 3. a) Prior to addition of enzyme the assay solution contains 3, excess NADH (elutes at 20 min), and a minor impurity of NAD (<" 2 %). b) After 12 h, the ketone 3 has been converted exclusively to 2a as confirmed by spiking experiments with both 2a and 2b. +  A minor impurity of NAD (< 2 %) was present in the NADH (10 mM, elutes at 20 min) +  used in this experiment (Figure 3.25a). Spiking experiments confirmed that NAD coelutes with +  2a, and 2b coelutes with ketone 3. However, this apparent complication did not affect the outcome of the experiment because the disappearance of the peak at 11 min (Figure 3.25b) unambiguously demonstrated that 2a, and not 2b is the product of the UDPGlcDH catalyzed reduction of 3. As expected, this result is fully consistent with the observation that 2a, and not 2b is a substrate for oxidation by UDPGlcDH.  110  3.4 DISCUSSION 3.4.1 Covalent labeling with UDC  The affinity label UDC has provided very strong support for the involvement of a cysteine thiol in the enzymatic mechanism of UDPGlcDH. When UDPGlcDH is incubated with UDC, a rapid loss of enzyme activity is observed and ESI MS analysis indicates that the protein has increased by the mass of UDC minus a chlorine atom. This result suggests that UDC is alkylating an enzyme nucleophile that is involved in the catalytic mechanism. The presence of known competitive inhibitors of UDPGlcDH provides protection against inactivation by UDC, indicating that UDC binds in the normal UDP-glucose binding site. The critical importance of this specific binding for the potency of UDC is supported by the observation that chloroacetol phosphate, a truncated form of UDC, is a much less effective inhibitor. Although the alkylated residue has not been unambiguously identified, labeling studies with iodoacetate have shown that it is probably a cysteine thiol that is modified during the inactivation with UDC. UDPGlcDH from S. pyogenes contains only one cysteine residue, Cys 260, that is strictly conserved in all known amino acid sequences and is therefore the best candidate for the active site residue that is modified by UDC. This conclusion, based solely on results obtained with UDC, preceded Xue Ge's discovery of the stable covalent ester intermediate that accumulates when Cys260Ser UDPGlcDH is incubated with substrate (see Section 1.6). That observation clearly provides the strongest evidence for the role of Cys 260 as the nucleophilic thiol but prior to that result, the inactivation with UDC was the best evidence in support of this hypothesis. Based on the results with UDC, Cys 260 is the best candidate for the nucleophilic residue that undergoes addition to the aldehyde intermediate to form a thiohemiacetal. Subsequent oxidation of the thiohemiacetal with hydride transfer to NAD yields the thioester intermediate +  Ill that is hydrolyzed in the final step of the reaction. As discussed above, UDC is bound to UDPGlcDH in the normal UDP-glucose binding site and therefore the nucleophilic carbon atom of UDC should reside in approximately the same location as the carbonyl carbon of the putative aldehyde intermediate. The apparent proximity of the thiol of Cys 260 to this location argues for its participation in covalent catalysis. The rough analogy between inactivation by UDC and formation of the thiohemiacetal intermediate in the normal UDPGlcDH reaction is graphically represented in Figure 3.26. This proposed role for Cys 260 is consistent with the 'GAPDH paradigm' for aldehyde to thioester oxidation. As discussed in Section 1.2.1, GAPDH was the first example and is the most thoroughly studied dehydrogenase that exploits a nucleophilic cysteine thiol to form a thiohemiacetal that is subsequently oxidized to a thioester. Perhaps a better comparison is with the AldDH super family of dehydrogenases (Section 1.2.2) that follow the GAPDH paradigm but also catalyze the hydrolysis (rather than phosphorolysis as in GAPDH) of the thioester.  112  Figure 3.26 Analogous roles for Cys 260 during inactivation by UDC and formation of the thiohemiacetal intermediate, a) Nucleophilic attack by Cys 260 at the electrophilic carbon of UDC results in alkylated and inactive UDPGlcDH. b) Nucleophilic attack by Cys 260 at the carbonyl carbon of the putative aldehyde intermediate 1 generates the covalently bound thiohemiacetal intermediate. 3.4.2 Studies with the putative aldehyde intermediate 1 When an enzymatic reaction is proposed to involve a stable intermediate, it is often possible to chemically synthesize the putative intermediate and test it as an alternate substrate (151). If the putative intermediate is converted to the product at a rate that is as fast as, or faster than that for the normal substrate, it is described as 'kinetically competent'. In the most simplistic approach to determining if a putative intermediate is 'kinetically competent', the steady state kinetics for its conversion are measured, and the first order rate constant (k ) compared with cat  that of the normal substrate. A more rigorous determination would have to account for the fact that this comparison is not strictly valid because the binding of the putative intermediate to the enzyme is not an event that happens in the normal mechanistic pathway (151, 152). Unfortunately, this type of treatment is not readily amenable to UDPGlcDH because it requires  113 knowledge of the equilibrium constants for both bound and unbound substrates as well as the dissociation constants for the bound substrates and products (151). It was found that the kinetic constants for the UDPGlcDH catalyzed oxidation of 1 (k = cal  1.0 s" and K - 14 pM) were very similar to those obtained for UDP-glucose under identical 1  m  conditions (k = 1.2 s" and K = 14 pM). The agreement in the values for k indicates that the 1  cat  m  cat  aldehyde is 'kinetically competent' to serve as an intermediate in the normal reaction pathway. The fact that the k for oxidation of 1 is equivalent to that for UDP-glucose, suggests that the cat  rate determining step for the normal reaction occurs after formation of the aldehyde intermediate. As mentioned in Section 1.6, another researcher in this laboratory, Xue Ge, has demonstrated that there is no primary deuterium kinetic isotope effect for hydride transfer during the UDPGlcDH reaction (84). Since the rate limiting step occurs after formation of the aldehyde intermediate and does not involve hydride transfer, formation of the thiohemiacetal or breakdown of the thioester are the only possibilities for the slowest chemical step in the normal reaction. While it is not possible to firmly rule out either of these options, the very rapid rate for the alkylation of UDPGlcDH by UDC suggests that nucleophilic attack is a fast step in the normal enzymatic mechanism. This leaves hydrolysis of the thioester intermediate as the most likely rate-limiting step in the UDPGlcDH catalyzed oxidation of UDP-glucose to UDPGlcA. One complication in interpreting the results from this experiment is that the aldehyde exists in solution primarily as the corresponding gem-diol (Section 3.3.1). As shown in Figure 3.27, there are three possible routes by which the UDPGlcDH catalyzed oxidation of 1 could be occurring. We believe that Path A is the operative mechanism for the UDPGlcDH catalyzed oxidation of 1 and that Paths B and C do not contribute to any significant extent. All the arguments against Paths B and C and in favor of Path A, come from investigations with the Cys260Ser and Cys260Ala mutants of UDPGlcA (83, 84). The complexity of these arguments  114 and the fact that the experiments are more closely associated with the work of Xue Ge requires that they be discussed in a separate section below (Section 3.4.4) and only the conclusions will be presented here.  HO  H-V°  H 0 2  H  HO-^V^ -^' \ HO-^—""Vi HO ' OUDP gem-diol of 1 V  HOA-—"*Yn rapid equilibrium HO ' OUDP 1 u  0  n  u  Path B: UDPGlcDH catalyzed ir dehydration of the gem-diol HO  0  H-V°  H  Enzyme Bound  H O " - ^ ^ ^ ° v  HOA^--^--A HO I OUDP n  HO  I  OUDP gem-diol of 1 n  u  1  Path A: Oxidation via normal pathway  w  Path C: Direct oxidation of gem-diol to UDPGlcA HO-^  HO ' n  u  OUDP  UDPGlcA  Figure 3.27 Three possible routes for the UDPGlcDH catalyzed oxidation of 1. Path A, UDPGlcDH binds the minor unhydrated aldehyde 1 from solution and oxidizes it by the normal pathway. Path B, UDPGlcDH binds the major hydrated gem-diol of 1 from solution and catalyzes its dehydration and subsequent oxidation by the normal pathway. Path C, UDPGlcDH binds the gem-diol of 1 and directly oxidizes it to UDPGlcA by an alternate mechanism. 1  The evidence supporting Path A as the major route of oxidation of 1 comes from the observation that the Cys260Ser mutant, when incubated with N A D and 1, forms a stable ester +  adduct. This result provides conclusive evidence that when residue 260 can participate in nucleophilic catalysis (as in both wild-type and Cys260Ser UDPGlcDH),  oxidation of  exogenously added 1 proceeds through a thioester intermediate (Path A). However, what may  115 appear to be a conflicting result has been obtained from studies with Cys260Ala UDPGlcDH. Cys260Ala UDPGlcDH, which lacks a potential nucleophile at residue 260, can efficiently catalyze the direct oxidation of the gem-diol of 1 to UDPGlcA. It would be very difficult to eliminate the possibility that the direct oxidation of the gem-diol of 1 (Path C) does occur to some small extent in the wild-type enzyme, but this is clearly not a significant contribution. The second important result from studies with the Cys260Ala mutant is that while either 1 or the gem-diol of 1 can reversibly bind to UDPGlcDH, these two species can not interconvert once they are bound to the enzyme. Extending this conclusion to the wild-type enzyme suggests that Path B could not be operative. The observation that the putative aldehyde intermediate is 'kinetically competent' is the most convincing experimental evidence against Kirkwood's proposed mechanism for UDPGlcDH (refer to Section 1.4.2 and Figure 1.14). The aldehyde is not a true intermediate in this proposed mechanism and it is improbable that the 'unnatural' binding of the aldehyde and formation of the imine linkage could occur at a rate as fast as the overall reaction. This result is fully consistent with the new proposed mechanism (refer to Section 1.5 and Figure 1.15) and provides very strong support in favor of the aldehyde 1 as a true intermediate in the normal UDPGlcDH catalyzed reaction. When 1 was tested as a substrate for reduction by UDPGlcDH, an unanticipated dismutation to UDP-glucose and UDPGlcA was observed (Figure 3.28). Two important factors dictate the time course for this dismutation. The first factor is that the concentration of 1 is asymptotically approaching zero, and at equilibrium, the solution will contain equal concentrations of both UDP-glucose and UDPGlcA. The second factor is that the first oxidation (UDP-glucose to 1) is freely reversible while the second oxidation (1 to UDPGlcA) is effectively irreversible. Initially, a huge excess of NADH drives the reaction towards NAD and UDP+  116 glucose and the ratio of (rate of oxidation of l)/(rate of reduction of 1) is less than one. The ratio of (rate of oxidation of l)/(rate of reduction of 1) increases throughout the time course of the dismutation, but the absolute values of the rates decrease as 1 is slowly exhausted and eventually depleted. Although the overall reaction is driven towards UDPGlcA and NADH by a large equilibrium constant, the NAD concentration is limited to that transiently formed during the +  dismutation and therefore a 50:50 mixture of UDP-glucose and UDPGlcA results. The instant of minimum NADH (and therefore maximum NAD ) concentration corresponds to the point at +  which the ratio of (rate of oxidation of l)/(rate of reduction of 1) is unity.  1  OUDP  UDP-glucose  NADH  NAD  UDPGlcA  1  +  Figure 3.28 Schematic representation of the dismutation observed when 1 is incubated with UDPGlcDH in the presence of NADH. The final solution will contain a 50:50 mixture of UDPglucose and UDPGlcA.  117 The observation of this dismutation is very interesting but provides little insight into the mechanism of UDPGlcDH. The only requirements for the dismutation to occur and have the observed time course are: 1) the first oxidation is freely reversible, and 2) the overall equilibrium constant for the oxidation of UDP-glucose strongly favors the product, UDPGlcA. If the first condition was false, and the first oxidation was effectively irreversible in the direction of oxidation, clearly no dismutation would have been observed. If the second condition were false and the equilibrium did not strongly favor UDPGlcA, the dismutation could still occur but the final ratio of the concentrations of UDP-glucose to UDPGlcA would be greater than one. An interesting situation presents itself when the same experiment is performed with the Cys260Ala mutant as described below in Section 3.4.4. In this case, the first oxidation is effectively irreversible in the direction of reduction, and 1 is primarily converted to UDP-glucose with only a partial dismutation.  3.4.3 Studies with C6" methyl analogues (2a, 2b and 3) of UDP-glucose The UDP-glucose analogues 2a and 2b are useful tools that can probe the order of the hydride transfer steps in the UDPGlcDH reaction. The substrate analogues 2a and 2b can be conceptualized as UDP-glucose with each of the diastereotopic hydrogens at C6" selectively replaced with methyl groups. During the course of the normal UDPGlcDH catalyzed oxidation of UDP-glucose, both the pro-R and pro-S hydrogen atoms at C6" (see Figure 3.4) must be sequentially transferred to two molecules of NAD . It was discovered that only 2a, the analogue +  that retains the pro-R-like' hydrogen of UDP-glucose, is a substrate for UDPGlcDH. Since covalent catalysis could not be involved in the oxidation of 2a, the reaction is most likely analogous to the normal oxidation of UDP-glucose to the aldehyde intermediate 1. Based on this assumption, this result suggests that the pro-R hydride of UDP-glucose is transferred during the  118  normal UDPGlcDH catalyzed oxidation of UDP-glucose to the aldehyde intermediate 1. The remaining pro-S hydride must be transferred during the second oxidation of the thiohemiacetal to the thioester intermediate. The order of hydride transfer is schematically represented in Figure 3.29. These conclusions are in agreement with a previous report that bovine liver UDPGlcDH transfers the pro-R hydride in the first oxidation step (77).  Cys 260  NADH  NAD+  R  HO^V-^-°\ HO-\^--- ^-A ,  I  HO  OUDP UDP-glucose H  0  n  K  OUDP  J  aldehyde  1k  NADH NAD  R  +  Cys 260  Cys 260  H  NAD"  1  NADH  S  H O ' - ^ ^ - ' HO n  u  C  \  HOA----^A  H O A ^ - ^ - A  I  HO  OUDP  n  thioester  u  I OUDP  thiohemiacetal  hydrolysis UDPGlcA  Figure 3.29 Proposed order of hydride transfer during the normal UDPGlcDH reaction.  Analogue 2a was a very poor substrate for UDPGlcDH, exhibiting a second order rate constant (k t/K ) of approximately 10 fold less than that for UDP-glucose when assayed in the ca  4  m  presence of 25 mM N H 2 O H . Furthermore, in the absence of N H 2 O H the enzyme catalyzed oxidation of 2a only proceeded a fraction of the way to completion. The observed turnover was  119 not due to an impurity in the preparation of 2a because the extent of reaction increased in the presence of N H 2 O H . The total concentration of NADH produced was several times greater than the concentration of UDPGlcDH and changing the concentration of UDPGlcDH did not affect the extent of turnover, suggesting that 2a was not a mechanism-based inhibitor. If each molecule of UDPGlcDH only oxidized a single molecule of 2a and thereby generated a potent irreversible inhibitor in the active site, the observed extent of conversion to product should be equivalent to the enzyme concentration. The possibility that the reaction was reaching its external equilibrium was dismissed because increasing the concentration of N A D had no effect on the extent of +  reaction. It was found however, that the extent of turnover of 2a was directly proportional to the concentration of 2a. The only consistent explanation for these observations is that turnover of 2a generates a relatively good reversible inhibitor of UDPGlcDH that accumulates in solution and competes with 2a for binding to the enzyme. The observed extent of turnover of 2a must correspond to several times the K\ value of this inhibitor for it to be able to compete with the excess substrate and cause the reaction to slow to an effective halt. The most likely candidate for the product of the oxidation of 2a, the ketone 3, has been chemically synthesized and demonstrated to be a good competitive inhibitor of UDPGlcDH (K\ - 6.7 uM). In addition, it has been definitively shown that 2a is the product of the UDPGlcDH catalyzed reduction of 3 and thus, by the principle of microscopic reversibility, 3 must be the product of the UDPGlcDH catalyzed oxidation of 2a. When 2a (1 mM) was incubated with UDPGlcDH in the presence of excess NAD , the extent of the reaction was limited to the +  formation of approximately 30 pM of the ketone product 3. This concentration is 4 to 5 fold greater than the K\ of 3 and could therefore cause significant product inhibition at subsaturating concentrations of the substrate 2a. Unfortunately, the K of 2a has not been reliably determined m  under similar experimental conditions so this argument should be treated with some speculation.  120 The ketone 3 was not as potent an inhibitor as had originally been hoped. The K\ value for 3 (6.7 pM) is only a factor of 3 less then the K for UDP-glucose (20 pM) and is a factor of 2 m  greater than the K\ for UDP-xylose (2.7 pM). If the inhibitor 3 was reversibly binding to UDPGlcDH as a covalent thiohemiacetal, analogous to that formed by the aldehyde in the normal enzymatic mechanism, several orders of magnitude greater affinity might have been expected. The fact that the K\ is greater than that for UDP-xylose suggests that any potential improvements in binding affinity have been offset by the introduction of the methyl group. The methyl group could cause steric clashes in the enzyme active site or could lower the propensity for addition of the nucleophilic cysteine thiol through stabilization of the carbonyl form of 3.  3.4.4 Experiments performed with Cys260Ser and Cys260Ala UDPGlcDH In addition to the results described above, the compounds 1 and 2a have been important tools in the studies of the two mutant enzymes, Cys260Ser and Cys260Ala UDPGlcDH. This work was performed in collaboration with Xue Ge in this laboratory and a more in-depth discussion is available in her Ph.D. thesis (84) and the published report of these results (83). As discussed in Section 1.6, the Cys260Ser mutant forms a stable ester adduct when incubated with UDP-glucose and NAD . Xue Ge has discovered that the same adduct is formed when the +  Cys260Ser mutant is incubated with 1 and NAD . Formation of this ester intermediate from +  aldehyde 1 requires consumption of only one equivalent of NAD , compared to two equivalents +  of NAD when UDP-glucose is the substrate. +  The aldehyde 1 has been the primary tool that has allowed the investigation of the Cys260Ala mutant and has provided some very interesting results. The most surprising result was that Cys260Ala is able to efficiently oxidize 1 (& = 0.2 s , K = 260 mM) to UDPGlcA -1  cat  m  even though it lacks the critical catalytic nucleophile (83). This oxidation is probably occurring  121 through the direct oxidation of the gem-diol form of the aldehyde and could best be described as an 'artifact' of the mutation that has little relevance for the normal enzymatic mechanism. The observation of the stable ester adduct that forms with the Cys260Ser mutant is overwhelming evidence for a thioester intermediate in the normal reaction mechanism. As discussed in Section 1.1.5, there are various examples of AlcDHs that are capable of the direct oxidation of gem-diols although this is not their normal substrate. It was also discovered that the Cys260Ala mutant can efficiently catalyze the reduction of aldehyde 1. Only a partial dismutation was observed for this reduction (compared to the complete dismutation with wild-type UDPGlcDH), and the final product ratio was approximately 4 UDP-glucose to 1 UDPGlcA. The kinetic constants for this Cys260Ala catalyzed reduction of 1 were determined to be k = 1.9 s" and K = 58 pM. These 1  cat  m  values may be a more accurate estimate (compared to the same experiment with wild-type enzyme) of the actual rate of reduction of 1 because the complication of the dismutation reaction is much less significant. 2  2  In Section 1.6 an experiment was described in which UDP-6"- H, H-glucose was incubated with Cys260Ala UDPGlcDH in the presence of both NAD and excess NADH. No +  hydrogen/deuterium exchange at the C6" position was observed. Two possible explanations for this result were provided: 1) the Cys260Ala mutant can not catalyze the first oxidation, or 2) NADH is not normally released until after formation of the hemithioacetal (in the normal reaction) or the gem-diol of 1 (in the Cys260Ala reaction). Based on the principle of microscopic reversibility, if Cys260Ala UDPGlcDH can catalyze the reduction of 1 to UDP-glucose, it can catalyze the oxidation of UDP-glucose to 1 and the latter explanation provided above must be true. In dismissing the former explanation, one is confronted with an apparent paradox regarding the catalytic abilities of the Cys260Ala mutant. This mutant enzyme can oxidize 1 to UDPGlcA (presumably through the gem-diol) and can reduce 1 to UDP-glucose (presumably through the  122 unhydrated form of the aldehyde 1). However, the Cys260Ala mutant can not oxidize UDPglucose to UDPGlcA, an overall reaction that appears to be the sum of two reactions this enzyme can catalyze. In order to resolve this apparent paradox, the following two statements must be true:  1. Cys260Ala UDPGlcDH  can not catalyze the reversible hydration/dehydration that  would convert the bound aldehyde 1 into the corresponding bound gem-diol. 2. The dissociation constant of 1 from Cys260Ala UDPGlcDH is very small.  Statement 2 addresses the possibility that 1 could be released from Cys260Ala UDPGlcDH, spontaneously hydrated to the gem-diol in solution, bound as the gem-diol, and directly oxidized to UDPGlcA. The principle of microscopic reversibility demands that this process must be occurring at a finite (albeit very slow) rate since Cys260Ala has been shown to bind both 1 and NADH and therefore must also release these same species. These results support the conclusion that the bound equilibrium greatly favors UDP-glucose and NAD over the +  aldehyde 1 and NADH. The secondary alcohols 2a and 2b were tested as substrates with the Cys260Ala mutant of UDPGlcDH. As with the wild type enzyme, only 2a was a substrate for the enzyme, and the total substrate turnover was limited to less than 5 % in the absence of N H 2 O H . This experiment provides very strong evidence that the thiol of Cys 260 is not involved in thefirstoxidation. As discussed above, the apparent inability of Cys260Ala to catalyze the oxidation of UDP-glucose to the aldehyde intermediate 1 was troubling because this mutant enzyme can efficiently catalyze the reduction of 1 to UDP-glucose. The observation that 2a is an equally good substrate in both the presence and absence of the thiol of Cys 260 provides evidence that the Cys260Ala mutant  123 has not been chemically 'handicapped' with respect to thefirstoxidation. This result is consistent with the arguments provided in the previous paragraph regarding the apparent inability of Cys260Ala UDPGlcDH to catalyze the two-fold oxidation of UDP-glucose to UDPGlcA.  124  CHAPTER 4: THE X-RAY STRUCTURE OF UDPGlcDH  125  4.1 INTRODUCTION Up to this point of the thesis, UDPGlcDH has been treated like a mysterious 'black box' that opens to receive the substrate, performs its catalytic function with the lid closed, and then opens once again to release the product. An understanding of the chemistry that occurs inside this 'black box' is the goal of this project but as the metaphor implies, it is very difficult to obtain that information. For many researchers in enzymology, it is the challenge of working with a 'black box' that holds the greatest appeal and potentially the greatest rewards. However, it is also very difficult because a thorough understanding of the details of the chemical mechanism is only obtainable through the consolidation and interpretation of results from a variety of cleverly designed experiments. In addition, there eventually comes a point of diminishing returns when the proposed experiments will no longer provide enough insight into the mechanism to justify the necessary effort. At this point, it may be necessary to resort to what some may refer to as 'unsporting techniques' such as X-ray crystallography. X-ray crystallography is the most popular and efficient method of obtaining an atomic resolution 'snapshot' of a protein and thereby revealing the inner workings of the 'black box'. There is a misconception though, probably due to the fundamental appeal of a visual image, that a X-ray structure can provide answers to all the questions. This attitude may not be as prevalent as it was in the early days of protein crystallography, with the majority of present day researchers understanding that structural information can only be reasonably interpreted in the context of other mechanistic studies. X-ray crystallographic studies provide a foundation for many further experiments because a refined structure will create many more questions than it will answer. Specifically, an X-ray structure will identify interesting active site residues that are candidates for site-directed mutagenesis in an effort to determine the mechanistic role of that particular residue. In addition, the overall fold of the protein and its active-site architecture may  126 reveal unsuspected distant evolutionary relationships to other known proteins. These relationships could suggest further experiments to specifically investigate whether aspects of the chemical mechanism are shared between the structurally related enzymes and if so, can the results from one system be (cautiously!) extended to the second. As described in the previous chapters, biochemical investigations of UDPGlcDH in this laboratory and others have been very successful in determining the overall mechanism of catalysis. However, having exhausted the majority of biochemical experiments originally planned for this project, it was felt that an atomic resolution structure was necessary in order to continue advancing our understanding of UDPGlcDH. In an effort to obtain this information, I moved to the research group of Dr. Natalie Strynadka (Department of Biochemistry, U.B.C.) in 1998 and began work on the crystallization and structural elucidation of UDPGlcDH. In this chapter, the X-ray structure of wild-type UDPGlcDH with bound UDP-xylose and NAD is +  presented. The structure of Cys260Ser UDPGlcDH that was crystallized in the presence of UDPglucose and NAD has also been solved. As discussed in the first chapter (Section 1.6) another +  researcher in this laboratory, Xue Ge, had discovered that Cys260Ser UDPGlcDH is able to catalyze the two-fold oxidation of UDP-glucose to the corresponding ester intermediate. Hydrolysis of the ester intermediate is very slow so this species accumulates. It was anticipated that growing crystals of Cys260Ser UDPGlcDH in the presence of substrate would allow the observation of the covalent ester intermediate in the high-resolution crystallographic structure. This attempt to directly observe the ester intermediate was unsuccessful, however this molecular structure did reveal the ternary complex of UDPGlcDH with the bound products UDPGlcA and NAD(H). It is likely that the ester adduct of Cys260Ser UDPGlcDH was the species that actually crystallized, but subsequent hydrolysis in the crystalline state gave UDPGlcA (see Section 4.3.1 for further discussion). It was not possible to distinguish whether the oxidized (NAD ) or +  127 reduced (NADH) form of the nicotinamide cofactor was bound to Cys260Ser UDPGlcDH, therefore this species will be referred to as NAD(H) to reflect this uncertainty. The majority of the research in this chapter has been published (153).  4.2 M A T E R I A L S A N D  METHODS  4.2.1 General Procedures. Ammonium sulfate and magnesium chloride were purchased from Fisher Scientific (Fair Lawn, NJ). For all reported procedures Cys260Ser, selenomethionine (SeMet) substituted and wild-type UDPGlcDH were treated in an identical manner unless otherwise noted. SeMet was incorporated into UDPGlcDH according to the procedure of Ramakrishnan (154) that is available on  the  Internet  (Ramakrishnan,  V.,  and  Graziano,  V.,  http:  //snowbird.med.utah.edu/~ramak/madms/segrowth.html). UDPGlcDH was purified as described in Section 2.2.2, although an additional step of chromatography was necessary in order to obtain UDPGlcDH of sufficient quality for crystallization. The primary sequence alignment included 48 sequences of UDPGlcDH, UDPManNAcDH, and GDPManDH that were retrieved from the SWISS-PROT database. The initial alignment was performed with the program CLUSTAL W (755) and then manipulated by hand to maximize conservation of secondary structural elements as identified by PROMOTIF (156). The CCP4 (166) programs AREAIMOL and BAVERAGE were used to determine solvent exposed surface areas and average B factors. All ct-carbon superpositions were performed in the program O (757). Automated analysis of the dimer interface was performed by the Protein-protein interaction server that is freely available on the Internet (Jones, S., and Thornton, J. M., http://www.biochem.ucl.ac.uk/bsm/PP/server). All figures of UDPGlcDH were made with BOBSCRIPT (755) and are presented in divergent stereoview for 'wall-eyed' viewing.  128  4.2.2 Purification and Crystallization. Partially purified UDPGlcDH was applied to a column of ceramic hydroxyapatite (BioRad) in 1 mM magnesium chloride containing 10 % glycerol and 2 mM DTT. The column was eluted with a linear gradient of 30-300 mM sodium phosphate (pH 6.8) in the loading buffer. To crystallize SeMet and wild-type UDPGlcDH (differences for Cys260Ser are noted in parentheses), purified enzyme was dialyzed against 20 mM magnesium chloride and an appropriate stock solution was added to give 5.0-5.5 mg mL UDPGlcDH, 0.2 mM UDP-xylose -1  (1 mM UDP-glucose) and 2 mM NAD (10 mM). Crystals were grown by hanging-drop vapor +  diffusion from 2.0-2.1 M ammonium sulfate (1.6-1.7 M), 6-8 % glycerol and 100 mM Tris-HCl, pH 7.8. Crystals were grown at ambient temperature and generally appeared in less than 48 h and continued to grow for up to 5-6 days. Crystals had orthogonal faces with edges typically ranging from 0.05 mm to 0.3 mm in length. All crystals belonged to space group P42(l)2 with a Matthews coefficient of 2.5, consistent with a single copy in the asymmetric unit.  4.2.3 Data Collection and Processing. Prior to data collection, crystals were transferred to a cryoprotectant solution of mother liquor supplemented with 25 % glycerol before being flash cooled in a stream of N (100 K). 2  Diffraction data to 3.2 A for the multiwavelength anomalous diffraction (MAD) experiment with the SeMet UDPGlcDH crystals was collected on beamline XI1 at DESY, Hamburg. Three highly redundant data sets were collected on a single crystal at wavelengths corresponding to the selenium absorption edge (A.1 = 0.9791 A), the peak (12 = 0.9740 A), and a wavelength remote from the selenium absorption edge (A3 = 0.9200 A). High-resolution data sets were collected for wild-type UDPGlcDH (2.6 A) at the same facility, and Cys260Ser UDPGlcDH (2.0 A) at  129 beamline X12C, Brookhaven National labs (see Table 4.1). All diffraction data were processed with DENZO/SCALEPACK (159) and initial phases were determined using SOLVE (160) which was successful in finding 6 of the 7 selenium atoms per asymmetric unit. The SOLVE scaled structure factors and intensities were refined in SHARP (161) and flattened with SOLOMON (162).  Table 4.1 Data collection and phasing statistics. SeMet UDPGlcDH  Wavelength (A) Resolution range (A) Mosaicity (°) Completeness (%)  a  Observations Unique observations •Emerge ( %)"' *  //sigma"  Wild-type  Cys260Ser  M  X2  A3  0.9791  0.9740  0.9200  0.9057  0.9790  19.8-3.20  20.0-3.20  20.0-3.20  24.3-2.30  27.8-1.80  0.3  0.3  0.3  1.2  0.8  100 (100)  100 (100)  100 (100)  95.1 (76.2)  87.3 (42.4)  78 626  62 644  64 705  119 755  176 474  8176  8056  8118  20 014  38 511  12.4 (29.8) 12.6 (30.5)  12.9 (33.4)  8.2 (37.7)  4.5 (51.0)  19.1 (7.7)  18.2 (7.7)  18.5 (7.4)  18.5 (2.7)  28.8(1.7)  2.14/3.18  2.87/3.68  0/0  1.43  1.66  1.34  0.51/0.48  0.53/0.52  0/0  0.89  0.83  0.91  Phasing statistics  Dispersive phasing power  c  (centrics/acentrics) Anomalous phasing power  c  ^cuiiis  dispersive''  (centrics/acentrics) i? iiis cu  anomalous**  Values in parentheses are for the outermost shell which is 3.27-3.20 A for SeMet data, 2.34-2.30 A for wild-type, and 1.83-1.80 A for Cys260Ser. The highest resolution shell with at least 50 % of the data with I/sigma greater than 3 is 2.08-2.03 A for Cys260Ser and 2.73-2.66 A for wild-type. Emerge = 2 | / - <I> \ I £ 1. Phasing power = (r.m.s. Fu) I (r.m.s. E), where F^ is the heavy-atom structure amplitude and E Is the residual lack of closure error. R Mh- 2 \E\ I £ |AF|. a  b  0  d  C  130 4.2.4 Model Building and Structural Refinement. The S O L O M O N flattened structure factors and phases produced a readily interpretable electron density map with well-defined density for both UDP-xylose and N A D . X T A L V I E W +  (163) was used for all model building. A l l 402 residues, UDP-xylose and N A D were fit in the +  initial round of model building and refined at full occupancy against the high resolution Cys260Ser data set (UDP-xylose replaced with UDPGlcA) in C N S - X P L O R (164) utilizing rigid body minimization followed by simulated annealing and restrained B-factor refinement. Templates for UDP-xylose, UDPGlcA and NAD(H) were constructed from fragments obtained from the C N S - X P L O R dictionaries or were modified versions of X P L O R (165) templates. The X-ray coordinates of both wild-type and Cys260Ser U D P G l c D H have been deposited in the Protein Data Bank (http://www.rcsb.org, accession codes 1DLI and 1DLJ respectively) and will be released upon publication of the manuscript describing this research (153).  4.3 R E S U L T S 4.3.1 Assessment of the Refined Structure of UDPGlcDH Following  successive cycles of model building and refinement  of Cys260Ser  UDPGlcDH, all 402 amino acids, UDPGlcA, NAD(H) and 390 solvent molecules were visible in the (2F - F ) S I G M A A (166) map contoured at l a . In addition, 3 molecules of sulfate and 3 0  c  molecules of glycerol were satisfactorily modeled into persistent (F - F ) density in chemically 0  c  reasonable environments. These molecules were also modeled into corresponding density in the wild-type structure though the occupancies had to be adjusted in order to maintain reasonable B factors. Wild-type UDPGlcDH was solved by molecular replacement with the final coordinates of Cys260Ser UDPGlcDH and refined in a similar manner. Both models were evaluated with the program P R O C H E C K (167) and were average or better in all statistical indicators of model  131 quality (Table 4.2) when compared to a representative selection of refined structures at the same / resolution.  Table 4.2 Refinement and model statistics. Wildtype  Cys260Ser  Resolution range ( A )  24-2.3  28-1.8  Observations  20008  38487  #factor (%)  18.6  RfrJ (test set, %)  25.9  21.3  Rm.s.d. bonds ( A )  0.011  0.011  Rm.s.d. angles (°)  1.6  1.5  Fully allowed <>| , v|/ ( % )  89.0  89.0  Disallowed  0.5  0.3  a  17.9  Model Statistics  " -^factor  =  ^  y ( %)  ll-^obsl " l ^ c a l c l l / l-^obsl-  *  Rfree  Was  calculated on 10 % of the reflections that were randomly omitted from the refinement  The Ramachandran plot of both structures indicates a common residue, Arg 316, in the disallowed region. This residue is well ordered (Cys260Ser B  a v  = 28, wild-type B  a v  = 32) and is  tightly packed in the C-terminal domain with the guanidinium moiety forming a salt bridge with the carboxylate of Glu 349. As will be discussed, the Leu 317 to Ser 329 Q-loop is critical for sequestering of the substrate and the main chain conformation of Arg 316 may be relevant to this function. A second disallowed residue in the wild-type structure, Asn 273 ( B  av  = 33 in both  structures), is found on the surface and is stabilized by a hydrogen bond with Arg 228 of a symmetry related molecule. Difficulties were encountered in modeling of the active site structure of the Cys260Ser mutant due to close contacts between the C6" carboxylate of UDPGlcA, the nicotinamide ring of  132 NAD(H) and O G of Ser 260. It was not possible to explain all the difference density in the vicinity of Ser 260 as shown in Figure 4.1a,b. This difference density may be the result of multiple main-chain conformations caused by unfavorable steric contacts. The more ordered nature of the crystals of Cys260Ser UDPGlcDH strongly suggests that the covalent ester form of the enzyme may have been the relevant species undergoing crystallization. However, the rate of hydrolysis in solution (tn.s of approximately 24-48 h at 20 °C) implies that U D P G l c A may have to exchange with UDP-glucose (and N A D H with N A D ) in the crystalline state in order to +  observe the ester intermediate in the X-ray structure. The apparent lack of exchange has resulted in trapping of the poorly binding product UDPGlcA (K\ = 200 p M compared to 2.7 u M for UDP-xylose) (87) and NAD(H) in an unfavorable ternary complex that is stabilized by crystal packing interactions.  133  Figure 4.1 Representative electron density for NAD(H) (adenosine not shown), the UDP-sugar, and residues 259-260. For both structures a (F - F ) map (contoured at 3a) was calculated with the final model coordinates lacking NAD(H), the UDP-sugar, and residues 259-260. a) Wildtype UDPGlcDH shows well-defined density for all omitted residues, b) Cys260Ser UDPGlcDH clearly shows the bound UDPGlcA and NAD(H) but Tyr 259 and Cys 260 have additional density that is not adequately explained by the final model (black ball-and-stick). A proposed alternate conformation of the Tyr 259-Ser 260 peptide bond and the OG of Ser 260 is shown in gray with two associated water molecules. Q  c  134  4.3.2 Tertiary and Quaternary Structure. UDPGlcDH consists of two discrete a/p domains, each of which contains a core P-sheet sandwiched between a-helices as shown in Figure 4.2. These two domains are connected by a long (48 A) central a-helix (a9) that also serves as the core of the dimer interface. The 6-stranded parallel P-sheet (Pl-p4, P7, P8) that is characteristic of the dinucleotide binding Rossmann fold (2) serves as the core of the N-terminal NAD -binding domain (residues 1-196). +  Figure 4.2 Ribbon representation of the ternary complex of the Cys260Ser UDPGlcDH/UDPGlcA/NAD(H) monomer. UDPGlcA, NAD(H), and the side chain Ser 260 (Nterminus of a l 1) are shown in ball-and-stick.  NAD is bound in a typical orientation in the cleft between strands pi and P4, positioning +  the nicotinamide ring in the active site formed at the domain interface. The Rossmann fold is followed by an additional p-a-Punit (P9, a8, BIO) that is antiparallel and contiguous with the  135 central 6-stranded p-sheet. However, only a single main chain hydrogen bond connects strand P8 of the 6-stranded sheet and strand P9 of the antiparallel p-a-Punit so these should probably be considered separate P-sheet structures. Inserted between P4 and P5 is a 14 residue loop (residues 80-93) that forms a small antiparallel P-sheet (P5, p6) which packs against a l 1 of the C-terminal domain. The last strand of the N-terminal domain (P10) leads directly into the long central ahelix (a9). The core of C-terminal domain (residues 229-402) is a 5-stranded parallel p-sheet (P11-P15) with a-helices packed on both sides. The topology of the C-terminal a/p fold (residues 310-395) is identical to the first 5-strands of the N-terminal dinucleotide binding fold (residues 1-115). A superposition of the a-carbons of both domains gives a r.m.s. difference of 2.0 A over 65 atoms as is shown in Figure 4.3.  Figure 4.3 Superposition of the a-carbons of residues 1-123 and 310-402 of Cys260Ser UDPGlcDH. Residues 1-123 and NAD(H) are represented with the black coil and black bonds respectively. Residues 310-402 and UDPGlcA are represented with the white coil and gray bonds respectively.  The predominant differences between the two regions of similar structure are two insertions (residues 41-59 and residues 80-93) in the N-terminal domain, and a single insertion  136 (the Q-loop) in the C-terminal domain. This interdomain pseudo-symmetry of the dinucleotide binding fold is similar to that seen in the D-specific dehydrogenases (Section 1.1.2) and aldehyde dehydrogenase (Section 1.2.2), though it has also been observed in alanine dehydrogenase (168), and UDP-Af-acetylmuramoyl-L-alanine:D-glutamate ligase (MurD) (169). In addition to the dinucleotide binding fold, the C-terminal domain also contains a stretch of 3 a-helices (al0-al2) that is divided by a single long stretch of extended coil (residues 242-258) that wraps around approximately half of the UDP-glucose binding pocket. This region of mixed a-helix and coil leads from the long interdomain a-helix (a9) into the 5-stranded P-sheet region. The C-terminal domain is primarily responsible for binding the UDP moiety of the UDPsugar in a deep pocket approximately adjacent to the middle third of the central a-helix (a9). The UDP-sugar is oriented such that the pyranose sugar ring is positioned at the domain interface with the C6" position of UDP-glucose (by analogy to UDPGlcA and UDP-xylose in the X-ray structures) approximately 2.5 A from C4 (NC4) on the B face of the nicotinamide ring of NAD . +  The catalytic nucleophile, Cys 260, is found in the C-terminal domain at the N-terminus of a l l . The overall structure of both Cys260Ser and wild-type UDPGlcDH are highly similar with a r.m.s. difference of 0.2 A over all a-carbon atoms. As discussed in Section 2.4.1, gel-filtration studies with the S. pyogenes UDPGlcDH have indicated that it may exist as a monomer in solution. However, the X-ray structure of UDPGlcDH reveals a crystallographic dimer with an interface of greater than 2600 A as shown 2  in Figure 4.4 (170). The dimer is clearly the active form of UDPGlcDH, suggesting that the conclusions from the earlier gel-filtration study were false. One possible explanation for this discrepancy is that the equilibrium between monomelic and dimeric UDPGlcDH could be dependent on the presence of substrate or an inhibitor. Under the conditions of the gel-filtration study (no substrate or inhibitor) the equilibrium could favor the monomer, whereas under the  137 conditions of the crystallization (substrate or inhibitor present), the dimer could be favored. No experiments to address this question have been performed yet. The helical portion of the C-terminal domain (al0-al2) contributes the majority (52 %) of the interface solvent inaccessible surface area, followed by the central a-helix a9 (37 %), and the N-terminal domain (12 %).  Figure 4.4 The crystallographic dimer of Cys260Ser UDPGlcDH. Indicated in the figure is the side chain of Arg 244 that forms hydrogen bonds with the substrate of the dimer partner.  There are a total of 24 hydrogen bonds stabilizing the dimer interface, though none of the amino acids involved are strictly conserved. Aromatic residues including Phe 206, Tyr 210, Tyr 217, Tyr 224 and Tyr 272 dominate the dimer interface. Of these five residues, Tyr 210 exhibits the strongest conservation of aromatic character and it is interesting to note that this residue interacts with itself through crystallographic symmetry. Tyr 217 is noteworthy as it exhibits strong conservation of aromatic or hydrophobic character in all bacteria, though higher order species have a serine at this position. There are no residues contributed from the dimer partner that could conceivably have a catalytic role in UDPGlcDH, but there is one residue, Arg 244, that contributes to UDP-glucose binding in the adjacent active site (Figure 4.4). This interaction  138 appears to be important for positioning of the substrate (see Section 4.3.5) and thus provides strong support for the dimer as the active form of UDPGlcDH.  4.3.3 Structural similarity to other dehydrogenases. The program DALI (171) was used to search for homologous proteins and as expected, many dehydrogenases and related proteins containing dinucleotide binding domains were identified as being structurally similar. The proteins with the greatest overall structural homology to UDPGlcDH are short-chain L-3-hydroxyacyl-CoA dehydrogenase (SCHAD) (30), and 6phosphogluconate dehydrogenase (6PGDH) (29). The structure and mechanism of these enzymes was briefly discussed in Section 1.1.3 as examples of metal-ion independent P-hydroxy acid dehydrogenases. Structural alignment of residues 1-290 of UDPGlcDH with each of these two proteins gives a r.m.s. difference of 1.9 A in both cases over 163 and 168 a-carbons respectively. The superposition of UDPGlcDH with 6PGDH is shown in Figure 4.5. All the major secondary structural elements of the N-terminal domain as well as significant portions of the central a-helix (a9) and the C-terminal a-helical region (al0-al2) are common to UDPGlcDH, SCHAD, and 6PGDH. Beyond residue 290 of UDPGlcDH there is no structural homology with either SCHAD or 6PGDH as both of these proteins have a primarily a-helical domain (refer to Figure 1.5) whereas UDPGlcDH has the C-terminal dinucleotide binding fold. While the reactions catalyzed by both SCHAD and 6PGDH are similar (P-oxidation and p-oxidation/decarboxylation respectively), they are fundamentally different from UDPGlcDH (2-fold oxidation followed by hydrolysis) which leads one to speculate on the foundation for this structural relationship. The basis for this relationship may be the conservation of the NAD -binding domain as well as active +  site architecture and specific catalytic residues that are common to both enzymes. Inspection of the active site of 6PGDH superimposed with UDPGlcDH revealed a remarkable conservation of  139 identity and conformation for two active site residues: Lys 204 and Asn 208 (Lys 183 and Asn 187 in sheep liver 6PGDH). These residues have been implicated in the enzymatic mechanism of 6PDGPI (29, 31), and their possible roles in the mechanism of UDPGlcDH will be discussed below.  Figure 4.5 Superposition of the a-carbons of Cys260Ser UDPGlcDH (black coil) and residues 1-300 of 6PGDH (white coil). UDPGlcA and NAD(H) bound to Cys260Ser UDPGlcDH are represented with gray bonds and white atoms. In addition, the side chains of the conserved residues Lys 204 and Asn 208 of UDPGlcDH (gray bonds with white atoms) and Lys 183 and Asn 187 of 6PGDH (white bonds with black atoms) are indicated.  4.3.4 Roles of conserved residues. Alignment of 48 sequences including UDPGlcDH, UDPManNAcDH, and GDPManDH revealed a total of 22 residues that were strictly conserved in 47 or more sequences. A representative selection of sequences from this alignment is shown in Figure 4.6.  140  S. pyogenes U D P G D H S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH £. colt U D P M a n N A c D H  S. pyogenes U D P G D H S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa G D P M a n D H E. coli U D P M a n N A c D H  M M F -  E  M S  Y V E K Y K E V E A L A S V  K  I ;M K F  E  K M A  I I I I  j  K f R T ;  A A C S  V I C I  A A I F  S  V  I  L V  V V C S A G A  P T V C  L L V C A  L  L I  L F  S A A S A  Q H M A R S R  70  1 1  a  K S  L V  G A  K E L E K S C R G K Q Q G R Q T G K T A V E G G V E  S N N  S K E KA AS TL LD DP A F F S T N D S G T T D F K R A S T T P V  L L L L  R F  1  A  A  H V D A K A E A  "ff2~^> f a.2 - - QNEVT I V D I LPSKVDK I - - HHEVKV I 0 V IKDKVES I CPE IRVTVVTBINESRINA W - - GHEV IGVDVSSTKIDL I -QKQVIGVDINQHAVDT I  Y K E Y K D 1 K E V L D -  A E V E A D L V S D V • D R w  s  Y L F F  0> 1 1 A V 1 1 1 S V 1 C V  1  S. pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E. coli U D P M a n N A c D H  F Y A V S  I L L I V  M G  A  a9  T Y L T Y L A F L V W H S F R  V V A F N T T N T W R K T V E V T N S R  i  A A A A D  T  K Q  T K Q V L N V R A P W F  L T Q A V  S . pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E. coli U D P M a n N A c D H  K  E  Q  K P S D K R L  K Q S  L  F  A  V  A  -  -  -  -  N  T V  D C  L  D  K  S R  M  L  N  K  L  E  S  T K  I V  E P  c  D T  R  E  F I  E Y H Q  A L  N V N I  T  R R R I A  Q D R P E  V  L  N Y N Y  K T Y G K K N G K G D -  JL T  K F Q T K F N T I F D A L I E D W L A E  L  d  T  P  - R - - D L M G K G - -  D D N T C S M R P  D  T  Y A E Y S E L C E I A K I C A  270  L  A A Y Y A  N Y  d  C  2>  N R S L C E R A S Q N P  F  d  A  D T Q H V D T S S V D L K Y 1 L D L G Y 1 P D M T Y V  P E -  S  L  Y  K  A A  c  /'  - D H E  E E  T V S A  A  -  R  S N S D S D S D K D T G A G T D  P  -  -  H V  L S  V  s  T L  -TH] I  R  E  I I G A  -0  P V P I K I P  F L E G  I  E C  _2_4C  d  -  -  Q  T  L  I  E  N  I  (  N N D D  D  A  I  Q G I S I V C A I G V I C R L A N  d  Y Y M Q R  D D D D H  D P Q H P  R I R I R I K L R V  G  M - -  G S G N N L  S  R  300  290  3 3  E  S  S  S  S  V I M R  Q D R E  S M S V  D Y I R R F Q K A F W V  d  °  I L  N D L K S L D P Y  s  K  V  S A Q  •  L L  D  •  E  D G Y E  A  P A M E  A  N  F  K  F C  E  V  K N I D G A A S S  L  A  I  E T  V A  I  K Q  T  A  A  K Q  A G A A S R F D L I D Q  £JTT I R  L L  K S K D N Y  1  - D K I I - G K E I V I - G A H L H I  Y L M D E E M L I G K - G Y E L I A Q W H S G  E  F  R  E  T  L  380  D E E A E  Y  d  a13  E S A I K S S A V K E S S S I E S P L V  D  V  a 14 A R  I  c  - R I N L Q V O F G V D V N I  <x10  ID—c  d d  v V  1 1  V A  E E E  I Y R E L D A  E  S R K L N S H M I I V K E L N P K T I I D A T G A D V E E V A T A V G V D G R E V M D D Q G I itt V W E L I  - E N L I T A V E V A R Y W Q Q V - H P M L G S L - A R L I R T A  320  V  D Q  T  N Y F N Q F R A A  - - - - K P N - G K K A G P Q Q V G E Q  L E E A  L  P G L P O L  190  R  Y A G  a12  N I P D V P A L N L Q L D V Q Q - -  P  S  280  n - i o o p ••  K D K I  V V  -  370  S S  S P I  E  E E E  1 00  1  N S D V  230 L  FHE I D T S I Q s ISA F ADE I G N F AQE L S L  L K L V L T I V  V  2?Q FljlE  V V G V Y R L I V V G I Y R L I K I A I L G F A K V G L L G L S K I A C F G L A  I S H P G V K E Y I E L C T L  H  K  E K D E Y P  1 30  JmS.  E D Q S M G Y R Q V V D L R V H G A N P K K L T G  T T T S F  L D K Q T G D L M  R V A Y R V R Y R I S K V T N I A  310 L  M V I V M  N Q G N R G  a3  a D  K V K A D A E K F A L L L K S A A K K N N V E L T K R A W Q F A D L L K G G A I K E E V D E T P E G Q R A V Q A L C A V Y E H W V P  -2-LQ  S  a11  V  E E S V K  %  G P  P P P  a8  Y P R I I Vl Y P R I V V | N P D R V F P P M T K N D R V I G | ll  S. pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E. coli U D P M a n N A c D H  S. pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa G D P M a n D H E. c o " U D P M a n N A c D H  T K E N E  S S  L L K D I  200 S . pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa G D P M a n D H E. coli U D P M a n N A c D H  I T A N T  .Y A  A A N  N N G L S N N R K S N S P T L  90  80  3  L  S. pyogenes UDPGDH S. pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E. coli U D P M a n N A c D H  40  ii  3 0_  a1  A  N  I  I  S D I V H A V V D V D V  L L  V V  DP FDR  Y I V V  E  P  -|3i»-heli»(|  V  T N R A N R C T E L G N G D E M L V D H S V I  Y E P Y E P  -  w  L Q  F  Y D N M N D D M F K V D L V  F  K V  I  E  L  D L E L N K N G D  390 io-helix])—|P15^S . pyogenes UDPGDH S . pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E. coli U D P M a n N A c D H  S . pyogenes UDPGDH S . pneumoniae UDPGDH B o v i n e liver U D P G D H P. aeruginosa GDPManDH E . coli U D P M a n N A c D H  S - R D I F G R D T  -  R  D  L  F  G  R  E  M L K P A F I F D G R R V L D G D L V G F M P H T T T A Q A E G D A K G V W R  L  I  H N E C W -  L O T  I G F Q  I G K K V S S K R  I P Y A P S G E  I P  K F S L Q D M P N K K P R V  Figure 4.6 Representative primary sequence alignment of 48 sequences including UDPGlcDH, UDPManNAcDH, and GDPManDH. Secondary structural elements are shown schematically with cylinders representing a-helices and arrows representing p-strands. The two boxed areas correspond to the regions of similar fold superimposed in Figure 4.3. Positions in the sequence that are highlighted with white text on a black background are strictly conserved in at least 47 of the 48 aligned sequences while positions highlighted in gray exhibit strong conservation. The one letter code directly below the aligned sequences indicates a hydrogen bond to either NAD(H) (n), the bound UDP-sugar (u), or the dimer partner (d). Putative catalytic residues mentioned in the text are also indicated (c). Residue 161 (indicated by *), meets the above criteria for strict conservation but S. pyogenes is the only divergent sequence.  141 A l l secondary structural elements appear to be conserved across the 48 sequences, with the  greatest sequence diversity occurring near the  C-terminus,  as  exemplified by  the  approximately 50 additional residues in the mammalian U D P G l c D H . O f the 22 conserved residues, 6 are primarily involved in binding N A D , 2 are involved in binding UDP-glucose, and +  a total o f 10 residues are either directly involved in catalysis or are critical for the proper positioning o f the catalytic groups. These residues w i l l be discussed below in the context o f substrate binding and the catalytic mechanism. The 4 remaining strictly conserved residues (Ser 161, G l u 201, A s n 219, and A s n 287) are all remote from both the active site and the substrate binding pockets so are likely important for maintaining structural integrity or are necessary for productive folding. Interestingly,  the  relatively  conservative  mutation  Glu201Asp  results  in  an  unencapsulated phenotype in mutant strains o f S. pneumonia {172). The carboxylate o f G l u 201 forms buried hydrogen bonds with the main chain amide nitrogens o f 3 residues ( G l y 122, Phe 123, and He 124) located at the dimer interface. Disruption o f these critical hydrogen bonds would likely interfere with proper formation o f the dimeric species and may additionally alter the position o f Thr 118, a critical active site residue that is close in primary sequence.  4.3.5 Substrate binding. The fold o f the C-terminal UDP-glucose binding domain is homologous to the ubiquitous dinucleotide binding fold so it is extraordinary that the orientation o f substrate binding has no apparent relation to the normal orientation o f N A D  +  binding as clearly shown in Figure 4.3. In  the X-ray structure o f M u r D (169), the substrate (a UDP-sugar) is bound to the N-terminal Rossmann fold in an orientation similar to that o f N A D . In U D P G l c D H , the orientation o f the +  bound UDP-sugar relative to N A D ( H ) is reminiscent o f substrate binding in the abortive complex  142 of UDP-galactose 4-epimerase with UDP-glucose and NADH (173). Despite the lack of structural homology in the UDP-sugar binding regions of these two enzymes, there are several apparent similarities in the protein-substrate interactions. Both enzymes make similar main chain hydrogen bonding interactions with the uridine moiety and utilize a carboxylate group to form a hydrogen bond with the ribose C2' hydroxyl. The UDP-glucose binding pocket of UDPGlcDH which is schematically represented in Figure 4.7, can be divided into two regions: the UMP binding pocket which is composed solely of residues contributed from the C-terminal domain, and the glucose-1-phosphate binding pocket consisting primarily of residues from the N-terminal domain. The UMP binding pocket is lined with a stretch of coil (Tyr 249-Gly 257) that makes 3 main chain hydrogen bonds, 2 side chain hydrogen bonds, and a pi-edge stacking interaction of Tyr 249 with the UMP moiety. The residues involved in these interactions include the strictly conserved Gly 257 that forms a hydrogen bond with the ribose C3' hydroxyl of UMP. There is one strictly conserved charged residue, Lys 320, that forms a salt bridge with the pyrophosphate moiety and is probably critical for sequestering of the substrate from the bulk solvent as will be discussed.  143  Figure 4.7  Schematic representation of interactions and hydrogen bond distances (A) between UDPGlcDH and bound UDPGlcA. The conformations of UDPGlcA bound to Cys260Ser UDPGlcDH and UDP-xylose bound to wild-type UDPGlcDH are very similar (r.m.s. difference = 0.2 A) and so are their interactions with the enzyme. Arg 244 (dashed box) is contributed by the symmetry related dimer partner.  Although the glucose 1-phosphate binding pocket is found at the dimer interface, binding interactions are limited to a small region (Phe 142-Glu 145) between 07 and al of the N-terminal domain that forms 3 main chain hydrogen bonds to the glucose-1-phosphate moiety. The glucose 1-phosphate binding pocket contains one additional residue contributed from the dimer partner. Arg 244 extends into the adjacent active site, forming hydrogen bonds with the pyranose C2" and C3" hydroxyls and may assist in proper orientation of the substrate for oxidation (see Figures 44, 4-7). Arg 244 reveals a clear mechanism for communication between active sites of the dimer pair, and UDPGlcDH from S. pyogenes has been noted to exhibit weak allosteric effects (Section 3.3.5). This is in marked contrast to the mammalian enzyme that exhibits very strong allosteric effects resulting in the observed 'half-of-the-sites' reactivity (/). Interestingly, the sequence alignment shows that Arg 244 may be a determinant of substrate specificity as it appears to be  144 conserved through all UDPGlcDH and UDPManNAcDH, but is replaced by a lysine residue in GDPManDH (see Figure 4.6). NAD(H) is bound to the N-terminal domain of UDPGlcDH in a typical orientation with the nicotinamide ring in a syn conformation. All hydrogen bonds between UDPGlcDH and NAD(H) are listed in Table 4.3.  T a b l e 4.3 NAD(H)-UDPGlcDH hydrogen bond distances Distance (A)  Residue  NAD(H)  Val 11 N  N02  3.1 (3.1)  Asp 29 OD1  A02*  2.8 (2.6)  Asp 29 OD2  A03*  2.7 (2.7)  Lys 34 NZ  A03*  3.0 (3.0)  Thr 83 OG1  N03*  2.5 (2.5)  Thr 118N  N03*  3.2 (3.0)  Glu 145 OE2  NN7  3.1 (3.2)  Lys 263 NZ  N02*  2.9 (2.7)  Arg 327 NH1  AOl  3.1 (3.4)  Arg 327 NH2  NOl  2.9 (3.2)  a  13  Atom names for NAD(H) follow the PDB convention. Distances are given for Cys260Ser/NAD(H) followed by wildtype/NAD in parentheses  a  b  +  The B face of the nicotinamide ring is facing the UDP-sugar and the A face is packed against a hydrophobic patch composed of Val 11, Leu 143, and Glu 141 (CB and CG). Six of the strictly conserved residues of UDPGlcDH are primarily involved in binding NAD : the three +  glycines of the GxGxxG 'fingerprint' of the Rossmann fold, the Thr 81/Pro 82 pair that packs against the adenine ring of NAD , and Arg 327 that forms a salt bridge with the pyrophosphate. +  Asp 29 is a critical residue of the dinucleotide binding 'fingerprint' and exhibits strict  145 conservation of either aspartate or glutamate that can hydrogen bond to both hydroxyls (A02* and A03*) of the adenine ribose.  4.3.6 Sequestering of reaction intermediates. A defining characteristic of the UDPGlcDH family of enzymes is the inability to detect the transient formation of any intermediates at the oxidation level of an aldehyde (7). This negative result had contributed to Kirkwood's proposed mechanism (Section 1.4.2) in which the intermediate at the oxidation level of the aldehyde actually exists as a covalently bound imine. An inspection of the structure of UDPGlcDH reveals a deeply buried active site that only exposes 5 A of the UDP-sugar to the bulk solvent. This is in contrast to NAD(H) which has a solvent 2  accessible area of approximately 54 A . The relatively exposed cofactor binding site suggests that 2  reversible exchange of NADH and NAD could be accomplished during the course of the +  catalytic mechanism with little structural reorganization of UDPGlcDH. To permit reversible binding of the UDP-sugar, UDPGlcDH would require either a significant interdomain movement and/or a repositioning of a loop region that covers the substrate. This is consistent with the observation that extensive screening of crystallization conditions in the absence of UDP-sugars and NAD  +  failed to produce crystals of free UDPGlcDH, suggesting that there is a  conformational ordering upon substrate binding that is conducive to crystallization. The protein surface that buries the UDP-sugar is composed of 4 separate regions that together shield 156 A of UDPGlcA (Cys260Ser UDPGlcDH) from the bulk solvent. These 2  regions are composed of the Q-loop (53 A ), the C-terminus (33 A ), the side chain of Arg 244 (8 2  2  A ), and the Arg 144-Lys 147 P-turn (62 A ). Of these four regions, the Q-loop is the best 2  2  candidate for a mobile 'gate' that opens and closes to allow reversible substrate binding. The Qloop is present in all aligned sequences of UDPGlcDH (see Figure 4.6) and contains two strictly  146 conserved residues, Lys 320 and Arg 327, which form salt bridges with the pyrophosphate moieties of the UDP-sugar and NAD(H) respectively. The salt bridges formed by Lys 320 and Arg 327 are the predominant interactions locking the 'gate' in the 'closed' conformation, and in the absence of substrate, the Q-loop may adopt an alternative 'open' conformation. The Cterminus (Asp 402) makes several contacts with both the UDP moiety and the Q-loop and movement of either of these groups would also require a reorientation of the C-terminus, implying that it may be a second component of the active site 'gate'. As suggested earlier, to explain the apparent lack of substrate exchange in the crystalline state, it is believed that the 'gate' is forced to maintain its 'closed' conformation due to crystal packing interactions. Since there are no direct crystal contacts with the Q-loop, opening of the 'gate' may be associated with a significant global conformational change that is unfeasible within the crystalline lattice.  4.3.7 Active site residues of UDPGlcDH. As illustrated in Figure 4.8a,b, the active site of UDPGlcDH contains the side chains of 6 conserved residues that are contributed from the N and C-terminal domains as well as from the central a-helix (a9). The most notable (and least surprising) conserved active site residue is the catalytic nucleophile, Cys 260 that is contributed from the C-terminal domain. The thiol of Cys 260 is positioned within 3.0 A of the predicted position of C6" of UDP-glucose (by analogy with UDP-xylose and UDPGlcA).  147  b.  Figure 4.8 Close-up views of the active site of UDPGlcDH with putative catalytic residues, a) The active site of wild-type UDPGlcDH with bound UDP-xylose and N A D truncated at the pyrophosphate bond, b) The active site of Cys260Ser UDPGlcDH with bound UDPGlcA and NAD(H). Notice that the C6" carboxylate of UDPGlcA occupies a position similar to a water molecule in the wild-type structure. +  148 Other than Cys 260, Asp 264 is the only other conserved residue in the C-terminal domain that extends into the active site and probably has a direct role in the enzyme mechanism. Both Cys 260 and Asp 264 are situated in the strictly conserved active site signature sequence, GGxCxxxD, (see Figure 4.6) that is characteristic of UDPGlcDH. The strictly conserved glycine pair preceding Cys 260 suggests that main chain conformations inaccessible to non-glycine residues are required for proper orientation of the catalytic nucleophile during catalysis. This is well precedented within the AldDH extended family where a strictly conserved glycine residue precedes the catalytic cysteine and is necessary to allow the polypeptide chain to 'twist back on itself (49). From the N-terminal domain, Thr 118 of the conserved loop between p7 and a6 forms a hydrogen bond to an ordered and conserved active site water molecule (Cys260Ser B = 21, wildtype B = 23) that may be critical for the catalytic mechanism. Flanking Thr 118 in the primary sequence are the strictly conserved Ser 117 and Pro 120 that are probably essential for proper orientation of Thr 118. A strictly conserved pair of residues in the N-terminal domain, Pro 140/Glu 141 (between P8 and a7), position the carbonyl oxygen of Glu 141 such that a hydrogen bond is formed to NZ of the key catalytic residue Lys 204. An additional active site residue and an apparent determinant of substrate specificity, Glu 145, exhibits very strong conservation across all sequences of UDPGlcDH and GDPManDH but is replaced by a proline in UDPManNAcDH. Glu 145 forms a hydrogen bond to a conserved water molecule (Cys260Ser B = 36, wild-type B = 41) that is in turn coordinated to a phosphate (PB) oxygen in both structures and a C6" carboxylate oxygen (06"2) in Cys260Ser UDPGlcDH. The central a-helix (a9) contributes two strictly conserved active site residues, Lys 204 and Asn 208, which coordinate a carboxylate oxygen (06" 1) of UDPGlcA in Cys260Ser and a similarly positioned active site water (B = 39) in the wild-type structure.  149  4.4 DISCUSSION In the first step of the catalytic mechanism, UDPGlcDH oxidizes the primary C6" hydroxyl of UDP-glucose with transfer of the pro-R hydride to C4 (NC4) on the B face of the nicotinamide ring of NAD . It is well precedented that enzymatic alcohol oxidation requires a +  general base: a role performed by a tyrosine hydroxyl in the SDR family of enzymes (Section 1.1.1) and a histidine imidazole in lactate dehydrogenase (Section 1.1.2). HMG-CoA reductase has been proposed to utilize a lysine general base to deprotonate the alcohol (in the direction of oxidation) as well as to stabilize the oxyanion of the hemiacetal intermediate (Section 1.3.2). A conserved arginine residue in the active site of HMG-CoA reductase may serve to depress the lysine pK through electrostatic interactions so it can serve as a base. In UDPGlcDH, the known a  stereochemistry of the first hydride transfer (the pro-R hydride, see Section 3.4.3) and the observed orientation of the UDP-sugar and NAD in the molecular structure fixes the relative +  position of the catalytic base that must deprotonate the substrate alcohol. The C6" hydroxyl of the substrate, UDP-glucose, probably occupies a position similar to the carboxylic acid oxygen (06" 1) of UDPGlcA in the Cys260Ser structure and the similarly positioned water molecule in the wild-type structure (see Figure 4.8a,b). Lys 204 (NZ) is the closest residue (2.8 A) to this position (2.9 A in wild-type/UDP-xylose) and based on the precedence of HMG-CoA reductase, it is tempting to propose that Lys 204 is the general base responsible for deprotonation of the substrate alcohol as shown in Figure 4.9.  150 Cys 260  Lys204  S"  ^ NH,  ~ ~™ Cys 260  X  Lys 204 ^-j—  ^NH  H 0 2  NAD+  H  ^fi';;,....  HN 2  HO A ^ - ^ - A  "  X  O  ^  NADH  H O - ^ ^ ^ °  I  x  HO-Y-—-^-A  A s n 208  Asn 208  I  HO  OUDP  u  R  OUDP  UDP-glucose  aldehyde  1k  NADH NAD  NH ^ -  6  S NADH  °  +  Lys 204  Lys 204  C  R  2  H  ,,H 0 2  d:'.. H N.  .0  H N^ ^ O  NAD+  2  2  S  HO  H^Xi-^X H  O  Asn 208  Asn 208 OUDP  OUDP  thioester  thiohemiacetal  hydrolysis  UDPGlcA  Figure 4.9 Proposed role for Lys 204 in the two-fold oxidation. Lys 204 is responsible for deprotonating the alcohol substrate during the first oxidation and contributing to stabilization of the tetrahedral intermediate during the second oxidation.  It is generally accepted that the pK of free lysine in water is 10.5. However, this value a  can be substantially decreased within the local environment of an enzyme active site through either electrostatic or hydrophobic effects (174). One of the best-studied examples of an extremely perturbed pK is the active site lysine nucleophile of acetoacetate decarboxylase (jpK = a  a  6) that is necessary in formation of an imine intermediate (175). A similar degree of perturbation has been proposed for the catalytic lysine base of mandelate racemase (pK = 6.4) (/ 76). In both a  of these examples, the proximity of additional charged residues (or metal ions) may destabilize  151 the protonated (charged) form of a primary amine, resulting in the extreme shift in pK . A buried a  lysine in a hydrophobic region can exhibit a similar shift in pK that is attributed to the instability a  of a positive charge in a non^polar environment (177). Several other examples of lysine acting as a general base catalyst include; HMG-CoA reductase (see Section 1.3.2), 6PGDH (see Section 1.1.3), P-lactamase (178), and signal peptidase (179, 180). As mentioned above, Lys 204 is structurally analogous to Lys 183 of 6PGDH, and therefore may perform a similar function in the catalytic mechanism. Recent evidence has provided support for Lys 183 of 6PGDH acting as the general base for deprotonation of the substrate alcohol (31). It was proposed that the hydrophobic nature of the 6PGDH active site could perturb the pK of Lys 183 by approximately 2.5 pH units a  to a pK of 8. In the ternary complex of UDPGlcDH, Lys 204 is inaccessible to bulk solvent but a  there are no positively charged residues or extensive hydrophobic regions in the immediate vicinity of Lys 204 (NZ) that should significantly lower the pK . However, UDPGlcDH has a a  relatively basic pH-rate optimum (pH 9) so invoking a substantially depressed pK for Lys 204 a  may not be necessary. An alternative candidate for the role of the general base is the conserved water molecule that is activated due to its hydrogen bond (2.6 A in both structures) to the Asp 264 carboxylate as shown in Figure 4.10. This tetrahedrally coordinated water molecule also forms a hydrogen bond (2.8 A ) with the carboxylate oxygen (06" 1) of UDPGlcA in Cys260Ser (2.4 A to the similarly positioned water in the wild-type/UDP-xylose structure). Additional hydrogen bonds with the strictly conserved residue Thr 118 and a ribose hydroxyl of NAD (N02*) are likely critical for +  proper positioning of this putative catalytic water. The hydrogen bonding environment of Thr 118 does not preclude this residue acting as either a proton donor or acceptor with the putative catalytic water molecule. In this proposed mechanism, Lys 204 NZ and Asn 208 ND1 could participate in thefirstoxidation through electrostatic stabilization of the alkoxide form of the  152 substrate alcohol. It is relevant to note that the only known divergent sequence at position 204 {Shigella sonnei, SWISS-PROT accession number Q55042, not shown in Figure 4.6) is a glutamine residue that could provide electrostatic stabilization but not act as an acid/base catalyst.  Asp 264  thioester  A s p 264  thiohemiacetal  hydrolysis  UDPGlcA  Figure 4.10 An alternative mechanism for the two-fold oxidation. The general acid/base catalyst is the conserved water molecule that is rendered basic due to its coordination with Asp 264.  A strictly conserved mechanistic feature of all aldehyde oxidations that utilize covalent catalysis is the requirement for a nucleophilic thiol. To increase the nucleophilic character of the thiol, enzymes may utilize an adjacent base such as histidine as proposed for both GAPDH  153 (Section 1.2.1) and HMG-CoA reductase (Section 1.3.2). In order for the thiol of Cys 260 to attack the si face of the aldehyde intermediate and generate the appropriate tetrahedral geometry for the second hydride transfer, a rotation from gauche(+) to gauche(-) may be necessary. This rotation could be accompanied by other conformational changes that could orient an appropriate base, such as the water activated by Glu 145, to deprotonate the thiol. It is interesting to note that the UDPGlcDH pH-rate optimum of pH 9 is consistent with the normal pK of a cysteine thiol a  and suggests that an assisting base may not be necessary. The oxidation of UDP-glucose to the aldehyde intermediate generates a proton donor (either Lys 204 or the water molecule activated by Asp 264) that would be strategically positioned to participate as a general acid catalyst during the subsequent attack of the cysteine thiol on the aldehyde intermediate (see Figure 4.9, 4.10). The oxyanion of the tetrahedral thiohemiacetal could be stabilized by Asn 208 and the catalytic moiety (Lys 204 or the water molecule activated by Asp 264) not involved in general acid catalysis. Collapse of the tetrahedral intermediate with hydride transfer to NAD and proton transfer back to the catalytic +  base would yield the covalently bound thioester. The proposed role for Asn 208 is similar to the conserved active site asparagine in the AldDH extended family that has been proposed to coordinate the carbonyl oxygen of the substrate (181). It is reasonable to propose that UDPGlcDH can perform two oxidations and a hydrolysis in a single active site only because the catalytic machinery required for all three steps is similar. As discussed above, the geometry, chemistry, and roles of active site residues in both oxidation steps must be very similar. This conservation of reaction chemistry can been speculatively extended to thefinalhydrolysis because of the particular similarities between these individual steps. Specifically, the developing negative charge on the thioester carbonyl oxygen could be stabilized by the same residues that stabilized the alkoxide in thefirstoxidation and the anionic  154 thiohemiacetal in the second oxidation. In addition, elimination of the thiol of Cys 260 must occur from the 'face' of the thioester that corresponds to the 'face' of the aldehyde to which it added. In order to generate a tetrahedral intermediate with geometry analogous to the thiohemiacetal, the hydrolytic water must add to the re face of the thioester or the face that is blocked by NADH. This argument leads to two possibilities: either NADH dissociates prior to the final hydrolysis or the tetrahedral intermediate formed during hydrolysis is not strictly analogous to the thiohemiacetal and there is a conformational change that moves the covalent thioester away from NADH. In support of the former option, the strictly conserved Glu 141 is located beneath the nicotinamide ring (see Figure 4.8a,b) and in the absence of NADH, could extend into the active site and deliver an activated water molecule to the thioester. The only appropriate base to activate a hydrolytic water molecule for si face attack is Glu 145, although this residue is not strictly conserved and was discussed above as a possible determinant of substrate specificity. It is not known if NADH dissociates from UDPGlcDH before hydrolysis of the thioester intermediate.  155  CHAPTER 5: CONCLUSIONS AND FUTURE DIRECTIONS  156  5.1 THE MECHANISM OF UDPGlcDH 5.1.1 Research summary The research described in this thesis has contributed immensely to our understanding of the mechanism of UDPGlcDH, but more importantly, it has raised many new and exciting questions. The next phase in the investigation of the mechanism of UDPGlcDH will focus on answering increasingly detailed questions concerning the roles of individual residues in the UDPGlcDH mechanism. There is little room left for studies that ask general questions regarding the mechanistic strategy employed by UDPGlcDH. In the introductory chapter, a brief summary of the previous research on UDPGlcDH was provided along with the standing mechanism prior to the work described here. The introductory chapter concluded with the new proposed mechanism of UDPGlcDH, for which there is now ample support. In addition, we now have the first hints of how the catalytic machinery of UDPGlcDH is able to perform a two-fold oxidation in a single active site. In this section, the important conclusions from the previous chapters will be summarized and discussed in terms of the mechanism of UDPGlcDH. In Figure 5.1 the revised mechanism of UDPGlcDH is shown with the details that have garnered support from this research.  157  Cys 260  L  S  ys2°  ^~ NH  2  H H  HO-A—-^-A NAD  H  Lys 204  Cys 260  4  0  R  Asn 208  OUDP  ,-H,0 R  HO^-^-- \ H O A ^ ^ A HO I  J _ Asn 208  0  N  NAD  OUDP  UDP-glucose  aldehyde NADH  ^1  NAD  Lys 204  Lys 204 Cys 260 Cys 260 f V ^  S  \^  d  Q V  ;  ;  ;  ;  R  +  HN 2  NH  2  ..-• ° H2  H  O ^===  —I—  Asn 208 OUDP  M A n u NADHc  NAD  OUDP thiohemiacetal  Lys 204 Cys 260  ^I  +NH  2  S" BH  BH  ???  ???  Asn 208 OUDP tetrahedral intermediate  H  H0 2  HO-/  H N^.O 2  H O ^ \ ^ - ° HO-A---^>-\  |  x  HO  Asn 208  1OUDP  UDPGlcA  Figure 5.1 The revised mechanism of UDPGlcDH.  When UDP-glucose is initially bound by UDPGlcDH, the C6" primary alcohol is positioned in an electrostatically positive pocket formed from the primary amine of Lys 204, the carboxamide nitrogen of Asn 208, and a conserved active site water molecule. Lys 204 probably  158 acts as a general base that deprotonates the substrate alcohol and contributes to stabilization of the alkoxide. A conserved active site water molecule could also fulfil this role; however, for the sake of this discussion it will be assumed that Lys 204 is the general base. NAD is bound by +  UDPGlcDH with C4 of the B face of the nicotinamide ring aligned with C6" and the pro-R hydrogen of UDP-glucose. Transfer of the pro-R hydride to NAD occurs with collapse of the +  alkoxide to form the aldehyde intermediate 1. Addition of the nucleophilic Cys 260 thiolate to the si face of 1 likely occurs before dissociation of NADH. The electropositive pocket could promote nucleophilic attack by polarization of the aldehyde as well as stabilization of the developing negative charge through both electrostatic effects and general acid catalysis involving Lys 204. Once the thiohemiacetal is formed, NADH must dissociate and a second molecule of NAD is bound by UDPGlcDH. The S stereochemistry of the thiohemiacetal +  positions the remaining (originally pro-S) hydride in the same location as the pro-R hydride during the first oxidation. Oxidation of the thiohemiacetal is analogous to thefirstoxidation step, with Lys 204 acting as the general base to assist in formation of the anionic thiohemiacetal that collapses with transfer of the remaining hydride to the second molecule of NAD . Dissociation +  of the second molecule of NADH leaves the covalent thioester intermediate that must be hydrolyzed in thefinaland rate-determining step of the reaction. An unidentified enzyme residue (possibly Glu 141) acts as a general base catalyst to deprotonate a water molecule that adds to the re face of the thioester to generate a second tetrahedral intermediate that has similar geometry to the thiohemiacetal. This would allow UDPGlcDH to again stabilize the negative charge on the oxygen atom through electrostatic effects and general base catalysis by Lys 204. Collapse of the tetrahedral intermediate with elimination of the thiolate of Cys 260 would generate the product, UDPGlcA.  159 The detailed mechanism presented in Figure 5.1 is consistent with all previous studies on both the mammalian and bacterial UDPGlcDH (Section 1.4.2). In addition, it is consistent with all the kinetic, mechanistic, and structural studies described in this work. The kinetic mechanism of UDPGlcDH has been shown to be a Bi Uni Uni Bi ping pong mechanism and thus the order of substrate binding and product dissociation is known (Section 2.4.2). The order of hydride transfer was established using analogues of UDP-glucose (Section 3.4.3) and was consistent with the substrate geometry observed in the X-ray structure (Section 4.4). The geometry of the bound substrates has also confirmed the B face specificity for hydride transfer to NAD (Section 4.3.5) +  that had previously been demonstrated for the bovine UDPGlcDH (1). The involvement of an imine intermediate as proposed by Kirkwood (Section 1.4.2) has been ruled out due to the observation that the aldehyde intermediate 1 is kinetically competent to serve as an intermediate (Section 3.4.2). A mechanistic role for a nucleophilic active-site cysteine thiol gained strong supportfromlabeling studies with UDC (Section 3.4.1), and the presence of Cys 260 in the active site of UDPGlcDH was unambiguously proven by the X-ray structure (Section 4.3.7). Other research in this lab with the Cys260Ser mutant enzyme has provided overwhelming support for a thioester intermediate formed with the thiol of Cys 260 (Sections 1.6, 3.4.4). Finally, the X-ray structure has allowed us to speculate on the possible roles of some conserved active site residues and propose that Lys 204 is a critical base in the enzyme mechanism (Section 4.4). This role for Lys 204 is supported not only by its appropriate positioning in the molecular structure but also on the intriguing conservation of this residue between UDPGlcDH and 6PGDH (Sections 4.3.2, 4.4).  160  5.1.2 Mechanistic relationship to other dehydrogenases UDPGlcDH can perform two oxidations and a hydrolysis within a single active site because similar catalytic machinery can be exploited for all three reactions. It is expected that the active site of UDPGlcDH should contain the catalytic components that normally constitute both an AlcDH and an AldDH. Now that the conserved active site residues of UDPGlcDH have been identified and tentatively assigned to mechanistic roles, comparisons can be made with the various dehydrogenases that were introduced in Chapter 1. A common theme for all the enzymes that catalyze alcohol oxidation is that a general base is required to deprotonate the hydroxyl group and assist in formation of the alkoxide. The residues that have precedent as a general base in this role include tyrosine (SDR, Section 1.1.1; AKR, Section 1.1.4), histidine (a-hydroxy acid dehydrogenase, Section 1.1.2; SCHAD, Section 1.1.3), and lysine (6PGDH, Section 1.1.3; HMG-CoA reductase, Section 1.3.2). It has been proposed that the analogous base in the UDPGlcDH mechanism is a lysine and therefore both 6PGDH and HMG-CoA reductase serve as excellent precedent. However, since the structural homology between 6PGDH and UDPGlcDH was used as evidence in support of similar roles for their conserved lysine residues, further comment on this correlation would be redundant. In contrast, the similar role proposed for lysine in both UDPGlcDH and HMG-CoA reductase is very intriguing. Despite fundamental differences in their overall reactions and sharing no structural homology (Section 1.3.2), both UDPGlcDH and HMG-CoA reductase have independently been proposed to use a lysine in identical roles. Specifically, the lysine residue is proposed to deprotonate the substrate alcohol in the first oxidation and stabilize the thiohemiacetal in the second oxidation. In UDPGlcDH, the proposed role for the lysine residue has been extended to include stabilization of the tetrahedral intermediate during the hydrolysis of the thioester; a step that does not occur in HMG-CoA reductase. It is notable that in 6PGDH, the conserved lysine  161 also participates in stabilization of an enolate during the oxidative decarboxylation (Section 1.1.3). It is interesting to speculate that the chemical properties of lysine make it uniquely suited to this role. In certain environments, the pK of lysine can be sufficiently depressed for it to act a  as a base at physiological pH. In addition, once the primary amine of a basic lysine accepts a proton, it becomes positively charged and therefore able to stabilize a negative charge through both electrostatic effects and general acid catalysis. A possible reason why the only other logical candidate, histidine, is not utilized in this role is because it would have a diminished ability to participate in electrostatic stabilization due to the derealization of charge over the imidazole ring. A common theme in the GAPDH paradigm for oxidation of aldehydes is a general base that assists in stabilization of the thiolate of the nucleophilic cysteine. A histidine is proposed to fulfil this role in both GAPDH (Section 1.2.1) and HMG-CoA reductase (Section 1.3.2) and a glutamate coordinated water molecule may perform the analogous role in AldDH (Section 1.2.2). The active site of UDPGlcDH does not contain a conserved histidine and there are currently no other candidates for a general base to active the cysteine thiol. As previously mentioned (Section 4.4), the pH optimum for UDPGlcDH activity is approximately the same as the normal pK for a thiol and therefore a general base may not be necessary to assist in a  formation of the thiolate. Histidinol dehydrogenase, the only other enzyme that catalyzes the two-fold oxidation of a primary alcohol to a free carboxylic acid, does not utilize covalent catalysis (Section 1.3.1). Although comparisons with histidinol dehydrogenase are of little benefit to deepening our understanding of the mechanism of UDPGlcDH, the comparison does raise one very interesting question. Why does UDPGlcDH follow the GAPDH paradigm rather than catalyzing the formation and oxidation of the gem-diol of 1? Histidinol dehydrogenase has evolved to follow  162 the latter option and is able to catalyze the same overall transformation as UDPGlcDH with somewhat better efficiency (K - 15 pM, k m  = 13 s") (58). In addition, it has been demonstrated 1  Cdtt  that Cys260Ala UDPGlcDH can efficiently catalyze the oxidation of the gem-diol of 1 to UDPGlcA (Section 3.4.4). To definitively answer this question, one would require a thorough understanding of the biological requirements and constraints that governed the evolution of each of these two enzymes. The structural homology between UDPGlcDH and the p-hydroxy acid dehydrogenases 6PGDH and SCHAD (Section 1.1.3) are the first clues to the evolutionary path of UDPGlcDH. Interestingly, neither 6PGDH nor SCHAD catalyze the oxidation of an aldehyde and neither have an important active site cysteine. Through random mutation, it may be 'easier' for an enzyme to acquire an active site cysteine (required for the GAPDH paradigm) then to acquire a catalytic metal ion cofactor (required for the histidinol dehydrogenase mechanism).  5.1.3 Future mechanistic work  Even with the detailed structural information provided by a X-ray structure, it is extremely difficult to predict the effect of mutating a specific catalytic residue. The easiest question to answer with a point mutation is 'is this catalytic residue important for catalysis?' If the residue in question is a conserved active site residue, the answer will undoubtedly be 'yes', as evidenced by the greatly diminished enzymatic activity of the mutant enzyme in the normal assay. However, even if the researcher has evidence that the particular mutation did not cause any gross structural perturbations or protein misfolding, he or she has still learned nothing about the particular role of the mutated residue. Mutagenesis experiments can only provide detailed mechanistic information if the mutant is well characterized and the researcher has tools, other than the normal enzyme assay, at their disposal. A good example of a tool that will facilitate further research into the mechanism of UDPGlcDH is the synthetic aldehyde intermediate 1. If a  163  particular mutant enzyme shows severely compromised activity for the normal oxidation of UDP-glucose, but can efficiently catalyze the reduction of 1, the mutated residue is likely critical only in the second oxidation or the hydrolysis. In the following section, several specific and unanswered questions regarding the mechanism of UDPGlcDH will be discussed and possible routes to answering them will be proposed.  Which residue acts as a general base to activate the hydrolytic water molecule? In theory, this should be the most straightforward question that can be answered by site-directed mutagenesis. The X-ray structure has suggested only two good candidates for this role and mutating these residues should have no effect other than to compromise the hydrolytic step of the enzyme reaction. If the rate of hydrolysis is diminished, the covalent thioester intermediate would accumulate resulting in an adduct that may be observable by ESI MS. It should also be possible to observe a pre-steady state burst phase of two equivalents of NADH formed per equivalent of UDPGlcDH. The mutant enzyme should still be fully competent in the reduction of the aldehyde intermediate 1 to give UDP-glucose. Another possible tool that may assist in answering this particular question is the methyl ester of UDPGlcA.  HO O H  Figure 5.2 The methyl ester of UDPGlcA.  Synthesis of this compound should be relatively straightforward when starting from the P-anomer of the fully acetylated methyl ester of glucuronic acid. Introduction of the a-phosphate  164 via the MacDonald procedure followed by deacetylation (in dry NaOMe/MeOH) and coupling will yield the desired product. Incubation of this compound with wild-type UDPGlcDH would likely result in the rapid formation of the thioester intermediate that would then be hydrolyzed to give UDPGlcA. This reaction would be spectrophotometrically silent so the reverse phase HPLC assay would be required to monitor the reaction. A mutant enzyme that was deficient in its ability to hydrolyze the ester intermediate would simply form the covalent species and be trapped. The advantage of this approach is that it could distinguish residues that were required for more than one step in the enzyme reaction from those that were required only for hydrolysis.  Is NADH released before or after hydrolysis of the thioester intermediate? Based on the crystallographic structural information it was proposed that Glu 141 may act as the general base to activate a water molecule for hydrolysis of the thioester. However, Glu 141 is located beneath the nicotinamide ring of NAD(H) and could only act as the general base in the absence of the cofactor. This question is closely associated with the previous question. If mutating Glu 141 to a nonbasic residue results in an enzyme that can catalyze the two-fold oxidation, but not hydrolysis, this would provide very strong evidence that NADH must dissociate before hydrolysis. In the interest of thoroughness, one would like to perform an experiment to directly test this hypothesis. The methyl ester of UDPGlcA may provide the most efficient test of this. As discussed above, incubating UDPGlcDH with the methyl ester of UDPGlcA should result in the rapid formation of the thioester intermediate. If UDPGlcDH can hydrolyze the thioester at a rate equivalent to or greater than the rate for the normal reaction, the thioester is a true intermediate in the reaction pathway. Performing this same experiment in both the presence and absence of NADH should reveal if the cofactor is normally released before or after hydrolysis.  165 A second (and much more speculative) approach is to take advantage of the fact the UDPGlcDH from S. pyogenes contains no tryptophan residues. It is very likely that a tryptophan residue could be engineered into UDPGlcDH in a position that is spatially near to the NAD  +  cofactor-binding site. Fluorescent quenching of the tryptophan could indicate when NAD (or +  NADH) is bound to enzyme. This approach has been successfully used to investigate coenzyme binding in horse liver AlcDH (182) and GAPDH (183). The first use of the engineered UDPGlcDH would be to experimentally verify that the mechanism is indeed ordered as proposed from the kinetic experiments. Neither NAD nor NADH should bind to the free enzyme (and +  thereby quench tryptophan fluorescence) unless UDP-glucose or UDPGlcA respectively is present. A second interesting experiment would be to create the double mutant with both the tryptophan replacement and the Cys260Ser mutation. Monitoring the tryptophan fluorescence during the burst production of NADH (in the presence of UDP-glucose) could provide strong evidence with respect to whether NADH is released before or after hydrolysis. With the Cys260Ser mutant, a burst of NADH is produced as the enzyme slowly forms the covalent ester intermediate that is the major enzyme form in the steady state. If NADH is released before the final hydrolysis, the fraction of the mutant UDPGlcDH with bound NADH should decrease (as should the amount of quenching) as the enzyme approaches the steady state. If NADH is released after the final hydrolysis, thefractionof enzyme with bound cofactor should remain constant (and so should the degree of quenching). Another interesting possibility is that a tryptophan residue engineered into the putative active site 'gate' (the Q-loop) could be a fluorescent indicator of structural rearrangements that occur upon binding of UDP-glucose.  Which residue acts as the general base to increase the nucleophilicity of Cys 260? There is good precedent that enzymes that utilize a nucleophilic cysteine thiol in their catalytic  166 mechanism require a general base to stabilize the thiolate form. Three good examples include HMG-CoA reductase (Section 1.3.2), aldehyde dehydrogenase (Section 1.2.2), and GAPDH (Section 1.2.1). In the X-ray structure of UDPGlcDH there are no obvious candidates for a residue that could perform the analogous role. However, it is likely that one of the conserved active site residues does perform this role and therefore all appropriate mutants should be screened for their reactivity towards UDC. UDC rapidly and specifically alkylates the active site cysteine through a mechanism that depends directly on the nucleophilicity of the thiol. If a mutant enzyme had a diminished affinity towards UDC, the mutated residue may have contributed towards stabilization of the thiolate. However, one potential problem is that the normal rate of inactivation is very rapid so even a diminished rate may be difficult to accurately measure. Two potential solutions include synthesis of the less reactive uridine 5'-diphospho fluoroacetol or performing the inactivation at a more acidic pH that should diminish the rate.  Is Lys 204 or the water molecule activated by Asp 264 the general base responsible for deprotonation of the substrate alcohol and stabilization of the thiohemiacetal intermediate? This is a very difficult question to answer because both Lys 204 and the activated water molecule probably contribute to stabilization of the alkoxide form of the substrate and the thiohemiacetal. Undoubtedly, one these two moieties is the general base and accepts the proton while the other provides electrostatic stabilization. Specifically replacing Lys 204 with various amino acids may reveal the answer to this question. For example, if the Lys204Arg mutant retains significant activity, it is possible that the primary role for this residue is providing electrostatic stabilization. Another possibility is using the Lys204Ala mutant and attempting 'chemical rescue' experiments with a variety of small organic amines. It is expected that the Lys204Ala mutant should have negligible activity. However, performing the reaction in the  167  presence of relatively high concentrations of ethylamine may 'rescue' the activity because the small amine can diffuse into the active site and effectively reconstitute the original lysine moiety. If this 'rescue' is possible with the methylated quaternary ammonium salt of ethylamine, this would be evidence that the primary role for this residue is electrostatic and it is not a general base.  The questions posed above and the experiments to answer them are only a fraction of the possible experiments that remain to be done with UDPGlcDH. Several other techniques that would be useful in the investigation of this enzyme mechanism are: Analytical ultracentrifugation. Utilizing this technique in the presence and absence of substrates and/or inhibitors would allow one to determine if the monomer to dimer equilibrium is dependent on substrates. Equilibrium dialysis. Equilibrium dialysis could be used to determine the dissociation constant for the aldehyde 1. It has been proposed that 1 is very tightly bound by UDPGlcDH, although this has never been verified. This technique could also be used to verify the ordered mechanism of substrate binding and product release. pH-rate profiles. The pH dependence of both k and K is currently unknown, and this cat  m  information would certainly aid in the identification of putative general acids or bases in the UDPGlcDH mechanism. One minor complication is that the rate of the UDPGlcDH reaction exhibits some dependence on the composition of the buffer.  168 5.2 T H E S T R U C T U R E O F U D P G l c D H  5.2.1 Summary of UDPGlcDH structure The overall fold of UDPGlcDH has been described for the first time and consists of two domains, each of which contains a central ( 3 s h e e t sandwiched between a-helices. The Nterminal domain is primarily composed of a classical NAD -binding Rossmann fold that leads +  into a long central a-helix. The central a-helix extends the full length of the protein, and eventually leads into the C-terminal UDP-sugar binding domain. Interestingly, the C-terminal domain has the same topology as the dinucleotide binding fold, however this structure is suprisingly not exploited for binding of the substrate in a typical orientation. UDPGlcDH is a dimer with an extensive solvent inaccessible interface that is primarily composed of the long central a-helix that joins the N- and C-terminal domains. The active site of UDPGlcDH is found at the interdomain cleft with putative catalytic residues that are contributed from both domains as well as from the central a-helix. A structure-based sequence alignment has confirmed the catalytic importance of several active site residues and has contributed to the first detailed proposal of the mechanism of UDPGlcDH (Section 5.1.1). The bound UDP-sugar is effectively shielded from the bulk solvent by a large loop (the Q-loop) that is proposed to be a component of a mobile 'gate' that can reversibly open and close to allow substrate exchange. The observation of a buried active site may explain the observed inaccessibility of the aldehyde intermediate to carbonyl trapping reagents.  5.2.2 Structural relationship to other dehydrogenases UDPGlcDH contains a typical dinucleotide binding fold and in this respect, it is related to the vast majority of all known dehydrogenases. The only known enzymes that exhibit any structural homology with UDPGlcDH other than the dinucleotide binding fold are 6PGDH and  169 SCHAD (Section 1.1.3). The first 300 residues of these three enzymes are structurally homologous and they must share a single evolutionary ancestor that catalyzed an NAD +  dependent reaction and utilized the conserved lysine in its mechanism (Lys 204 of UDPGlcDH). It would be naive to suppose that any one of these enzymes is the more 'ancient' then the others or to assume a direction for the evolution (i.e. UDPGlcDH evolved from 6PGDH). The first gene that encoded for an enzyme with UDPGlcDH activity probably arose from a rearrangement of the gene fragments that constituted the common ancestor. A second Rossmann fold domain may have been appended to the C-terminus of the common ancestor, resulting in a new protein with a stable tertiary structure and UDPGlcDH activity. The Rossmann fold domain is remarkably common and may serve to not only bind dinucleotides, but also as a convenient structural unit that can be exploited for a variety of purposes. A particularly notable correlation is that the overall domain structure and organization of UDPGlcDH is similar to that observed in both the D-specific dehydrogenases (Section 1.1.2) and AldDH (Section 1.2.2). Although there is no apparent evolutionary relationship between these three disparate classes of enzymes, they all have both an N- and C-terminal domain with the topology of the Rossmann fold. In each case, one Rossmann fold is utilized for binding of the NAD cofactor while the second binds the substrate. Between these three classes of enzymes, the +  substrates represent a quite diverse collection of organic molecules. This correlation suggests that with respect to dehydrogenases in particular, evolution has repeatedly selected for the Rossmann fold domain as a 'convenient' structural unit. The term 'convenient' refers to the ability of this domain to reliably fold into a stable tertiary structure that can be manipulated through evolution to perform a variety of specific functions.  170 5.2.3 Future structural work Now that the atomic resolution structure of UDPGlcDH is known, determining the structure of other mutant enzymes, different substrate or inhibitor complexes, or enzymes from different species, should be greatly facilitated. These structures can be solved by molecular replacement using the wild-type coordinates and it will not be necessary to go through the laborious process of obtaining experimental phases. One unsuccessful goal of this research was an X-ray structure of Cys260Ser UDPGlcDH with the covalent ester intermediate. Future efforts to achieve this goal may require a different mutant enzyme that is even more severely handicapped in the normal hydrolysis of the (thio)ester intermediate. It is possible that mutagenesis studies will discover the general base that delivers the activated water during the hydrolysis and mutations of this residue may be suitable candidates. However, due to the apparent lack of reversible exchange of substrates and products in the crystalline state, a rate of hydrolysis that it is greatly diminished may still be insufficient due to the delay between crystal growth and data collection. An obvious solution to this problem, that was not available at the time of this research, is to flash freeze and store a freshly grown crystal at cryogenic temperatures until data collection. The mutant UDPGlcDH with the covalent (thio)ester would be indefinitely stable under these conditions. Other interesting experiments with UDPGlcDH from S. pyogenes include the determination of the structure crystallized in the presence of the aldehyde intermediate 1 (no NAD or NADH) or the ketone inhibitor 3 (and NAD ). The +  +  structure of the free enzyme (no substrates or inhibitors) could provide some information about the conformational changes that occur upon binding of substrate. Some logical directions for related structural studies would include the X-ray structure determination of the mammalian UDPGlcDH and the bacterial UDPManNAcDH and/or GDPManDH. The structure of the mammalian UDPGlcDH is essential if rational inhibitors or  171 potential drugs are to be designed to specifically bind to the bacterial UDPGlcDH. Structures of UDPGlcDH from other organisms will also greatly assist in assessing the relative importance of the conserved catalytic residues. With atomic resolution structures of related enzymes with different substrate specificity, one would be afforded the opportunity to engineer UDPGlcDH in an attempt to create enzyme specificities that are not found in nature. For example, in UDPGlcDH the binding sites for UMP and glucose 1-phosphate are composed of different regions of the enzyme. Manipulation of the genetic sequences encoding these regions would allow one to create any combination of UMP and GDP binding C-terminal domains with either glucose 1-phosphate, ManNAc 1-phosphate, or mannose 1-phosphate binding N-terminal domains. This could of course be done in the absence of structural information but there would be a much larger element of guesswork and random screening of chimeric proteins.  5.3 C O N C L U D I N G R E M A R K S  UDPGlcDH is a remarkable and unique enzyme that can catalyze three separate transformations within its single active site. With the exception of our own work, over the last two decades there have only been a handful of papers published regarding the mechanism of this fascinating enzyme. It is hoped that the published reports of the work described herein will serve as a reminder to the enzymology community of the unique aspects of this enzyme mechanism. The public availability of the atomic coordinates may stimulate other research groups to initiate new investigations of UDPGlcDH from various organisms. In addition, pharmaceutical companies may be prompted to pursue UDPGlcDH as a target for structure-based inhibitors and potential antibiotics to specifically combat pathogens such as S. pneumoniae and group A streptococci. These are of course very lofty expectations, but at the very least, this work has  172  provided some intriguing insights and clarified some long-standing inconsistencies in the proposed mechanism of UDPGlcDH. For many years to come, the structure and mechanism of UDPGlcDH from S. pyogenes enzyme will be the representative example of enzyme catalyzed two-fold oxidations that utilize covalent catalysis.  173  APPENDIX A: THEORETICAL TREATMENT OF E N Z Y M E KINETICS  174 A.1 T H E R A T E E Q U A T I O N F O R A S I M P L E E N Z Y M E M E C H A N I S M  Although single substrate enzyme kinetics were not described in this thesis, a brief derivation of the Michaelis-Menten rate equation for this simple mechanism is necessary in order to introduce the reader to the terms and conventions used in the text. The basic steady-state model (Briggs-Haldane) for enzyme kinetics has the enzyme (E) and the substrate (S) in a rapid equilibrium between the free species (E + S) and the bound complex (ES). Catalytic turnover of the ES complex occurs in a slower, rate determining step (k ). 2  k  E  + S  i  k  2  ES  ^  E + P  k-1  The model assumes that several conditions are met: 1. Initial velocity conditions so [S] ~ [S] ([S] is the initial concentration of substrate) 2. No product present so k is effectively irreversible 3. [S] » [E]T ([E]T is the total concentration of enzyme, both bound and free) 4. No allosteric or cooperative effects 0  0  2  Under steady state conditions, there is no net change in [ES] and therefore the rate of formation of the ES complex (ki) is equal to sum of the rates for dissociation (k.i) and turnover (k ). 2  d[ES] =0=  dt  k,[E][S] - (k.., + k )[ES] or 2  k [E][S] = (k. + 1  k )[ES]  1  Eq.3  2  The total concentration of enzyme is given by: [E] = [E] + [ES]  Eq.4  T  Rearranging Eq. 4 to solve for [E] and substituting into Eq. 3 gives: k ([E] -[ES])[S] = (k + k. )[ES] or 1  [ES] =  T  [E]T[S]  2  [E] [S] T  =  1  where K =  k.  1  +  m  Ik., + k \ + [S]  K + [S]  2  M  k  2  E q >  5  K L  The initial rate of the reaction (v ) will be equal to: 0  v = k [ES] 0  2  Eq. 6  175 Substituting Eq. 5 into Eq. 6 we arrive at:  Vn  Vmax[S]  =  k-! + k  where K =  2  and V max = k [E] 2  m  Eq.7  T  K + [S] m  The equation derived above (Eq. 7) is based on the Briggs-Haldane steady state model but is equivalent to the original Michaelis-Menten equation. The constant K is the Michaelis constant and is equal to the substrate concentration at half the maximal velocity (V ) of the enzyme reaction. As will be seen in the derivation of a more complex rate equation (Section A.5), the value of K can be a very complicated function of rate constants. However, it is most easily understood as the apparent dissocation constant for all ES complexes in the steady state. m  max  m  K  [E][S]  -  m  Eq.8  Z[ES] V is the maximal velocity for the enzyme reaction. Dividing V by the total enzyme concentration gives the first order rate constant for the overall enzymatic reaction, k (ki). The value of k is commonly referred to as the turnover number, or the number of moles of product formed per unit time per mole of enzyme. m  a  x  m  a  x  cai  cat  /feat  V  —  n  Eq.9  [E]T  To experimentally determine the values for K and k , a Lineweaver-Burke or double reciprocal plot has traditionally been used. Inverting both sides of Eq. 7 gives an equation of the form y = mx + b where m is the slope and b is the y-intercept: m  1  =  K  m  +  cat  1 Eq. 10  VrraxfS]  V max  Plotting v~' against [S]" will give a linear plot with slope equal to /C A^ x and yintercept equal to Vmax" . The x-intercept is equal to -K '\ It is also possible to use a sophisticated graphing program, such as Grafit (Erithicus Software), to apply a best fit curve to a direct plot of the data (v against [S]) and obtain a better estimate of the values for K and V . 1  0  m  ma  1  m  0  m  m  a  x  A.2 R E V E R S I B L E INHIBITION The three types of reversible inhibition are best introduced with the simplest possible model system: an enzyme with one substrate and one product. As with the steady state rate equation derived above, these simple equations do not assist directly in the interpretation of the kinetic results presented in this thesis, but they do provide the necessary foundation for the discussions that will follow.  176 A.2.1 Competitive Inhibition A competitive inhibitor (I) binds reversibly to the free enzyme and prevents access of the substrate into the active site. This can be schematically represented as shown: k  E + S ===== E S + k.-, I  2  *• E + P  Ki = [E][l]  K  [El]  Using a similar treatment to that described for the Michaelis-Menten rate equation derived above, we find that this competitive inhibition model can be described by the following equation: v  0  VmaxIS]  =  [S] +  Eq. 11  K /1 + [I] m  K  Since most kinetic data presented in this thesis is in the form of double reciprocal plots, it is helpful to represent Eq. 10 as its inverse: 1 v  0  = 1 [S]  K  ml V  \  Eq. 12  K n  V  r  If a plot of VQ" against [S]" at changing fixed concentrations of the inhibitor (I) is constructed, a series of linear lines intersecting on the y-axis will result. Changing the concentration of the inhibitor can have no effect on the y-intercept (V x ')• In a double reciprocal plot, the y-intercept corresponds to an infinite concentration of the substrate, against which no finite concentration of competitive inhibitor can compete for binding to the free enzyme. The apparent K (K ) for a competitive inhibitor is K multiplied by a factor of (1 + [I]/ATj). As the [I] increases, so does the ^ as a greater concentration of substrate is required to attain the steady state concentration of [ES] required for half maximal velocity (which is unchanged). To experimentally determine the value of Ki, a replot of the slopes versus [I] is constructed. This gives a linear plot of the form y = mx + b: 1  1  ma  app  m  m  m  a p p  m  Eq. 13 max  The x-intercept of this plot is equal to -K\.  177  A.2.2 Uncompetitive Inhibition An uncompetitive inhibitor can only bind to the ES complex and not to the free enzyme. The simplest model for uncompetitive inhibition is shown below: E  *  +S  1  k.1  k o  - ES  ~+ E +P  +  Ki = [ES][I] K  [ESI] ESI  This model can be described by the following equation: Vmax 1+[l] Vn  Ki/I ]  Eq. 14  S  =  [S] + / K  \  m  1+JI]  Rearranging Eq. 13 to fit the form of a double reciprocal plot results in: 1  = 1 / Km \ + [S]  1  V, max  Eq. 15  The K for an uncompetitive inhibitor is K divided by a factor of (1 + [I]/ATj). Likewise, the V for an uncompetitive inhibitor is V divided by a factor of (1 + [I]/ATj). The reciprocal plot for uncompetitive inhibition is a series of parallel lines. Notice that in the reciprocal plot (Eq. 15), the effects of the inhibitor on K and V cancel out in the term that describes the slope of the line. In uncompetitive inhibition, an equilibrium will exist between the productive ES complex and the unproductive ESI complex that can not be affected by changing substrate concentration and will only vary as a factor [I]. The result is that both AT and V will decrease as [I] increases. This apparent improvement in the value of Km is observed because the equilibrium between the free enzyme and all the bound enzyme complexes shifts towards the bound species. To experimentally determine the value of K a plot of intercepts versus [I] is constructed. This linear plot is defined by the equation: a p p  m  m  a p p  m a x  m a x  m  m a x  app  m  PP  h  a p p  m a x  178 intercepts = [I]/  1  \ V  KiVr,  Eq. 16 max  The x-intercept of this plot is equal to -K\. A.2.3 Noncompetitive inhibition Noncompetitive inhibition results when an inhibitor can reversibly bind to both the free enzyme and the enzyme substrate complex. A schematic representation of noncompetive inhibition is shown below. The assumption is made that binding of inhibitor and substrate are independent events. *1  E + S ===== ES + k.-, +  E + P K = [ES][I] = [E][l] [ESI]  [El]  El + S ===== ESI k-1  Derivation of the rate equation describing this mode of inhibition gives: Vmax  1 + [Q V  0  Eq. 17  75/  =  [S] + K  m  Rearranging Eq. 17 to give the double reciprocal form of the equation results in: 1  = 1  v  [S]  0  /  Km \ +  1  Eq. 18  The value of K is unaffected by noncompetitive inhibition. Both the free enzyme (E) and enzyme substrate complex (ES) have equal affinity for the inhibitor (I) so increasing the concentration of the inhibitor will decrease the concentration of both E and ES by the same ratio. As seen in Eq. 8, decreasing both [E] and [ES] by the same factor will not affect the value of K . As with uncompetitive inhibition, V is decreased by a factor of (1 + [I]/ATi). The reciprocal plot for noncompetitive inhibition is a set of lines that intersect to the left of the y-axis. A replot of intercepts versus [I] will give a line described by Eq. 16, and thus the K\ value can be determined. m  m  m a x  1 7 9  A.3 I R R E V E R S I B L E I N A C T I V A T I O N  Irreversible inactivation of an enzyme by an affinity label can be represented schematically by the following model: E  +  I =5=  El  El*  The affinity label (I) can reversibly bind to the free enzyme (E) to form the enzyme inhibitor complex (EI) with an inhibition constant defined by K\. The bound affinity label irreversibly inactivates the enzyme in a slower chemical step with rate constant k\, to form the covalently bound complex EI*. It is assumed that [I] is much greater than [E] and is therefore constant. The derivation of the rate equation describing the observed rate of inactivation is straightforward. frjobs =  /f|[EI] =  Eq. 19 Ki + [I]  The observed rate of inactivation ( & i ° ) is the first order rate constant obtainedfroma plot of the time-dependent inactivation of the enzyme in the presence of the affinity label. A direct plot of k° against [I] should obey typical saturation kinetics and the values for k\ and K\ can be extracted from this curve. When [I] is much less than K\, a direct plot of & i ° versus [I] will be linear with a slope equal to the second order rate constant for inactivation, k\IK\. If the rate of inactivation is too rapid to accurately measure, a protecting agent can be included in the assay buffer to decrease the rate of inactivation. b  s  bs  b  s  K  E  +  I =5=  El  El*  +  P  K|P  EP  In this schematic representation, the protecting agent (P) is a competitive inhibitor that binds to the free enzyme with an inhibition constant of K\ . The protection afforded by the protecting agent is expressed by the equation:  <obs  -  [Q + K, 1 + [ P ] KP  Eq. 1  180 The protecting agent (P) competes with the affinity label (I) for binding to the free enzyme and thus increases the apparent K\ value by a factor of (1 + [P]/ATj). For a given concentration of P with a known value for K\, one can calculate the expected decrease in k b . P  0  AA  INITIAL V E L O C I T Y A N D P R O D U C T I N H I B I T I O N  s  P L O T S  The theoretical treatment of enzyme kinetics described above dealt with the simplest model systems in order to introduce many of the basic concepts. These treatments rely on mathematical models that can be readily interpreted in terms that correspond with our intuitive understanding of the mode of enzyme action. However, in more complicated multisubstrate systems the interpretation of the rate equations becomes much more difficult so Cleland (91, 92, 93), has devised a general and non-mathematical method for the interpretation of enzyme kinetics. A thorough introduction to this method is beyond the scope of this work, however, enough background will be provided for an understanding of the work presented in Chapter 2. The foundation of the method of Cleland is the representation of the kinetic mechanism of an enzyme using a straight line to represent the enzyme and vertical arrows to represent substrate binding and product dissociation. An example of a kinetic mechanism with three substrates and three products is represented below: A  B  ki k >  E  k  2  f  3  >  EA  P k  C  k  > 4  f /EAB\  k  K  7  >  F  Q k  R  k  > 8  k  9  k-io  kn  k  1 2  f / FC \  ER  E  Substrates binding to the enzyme (down arrows) are denoted sequentially starting with A. Products dissociating from the enzyme (up arrows) are denoted sequentially starting with P. The rate constants on the left of the arrows describe the forward reaction direction (left to right) and those on the right are for the reverse reaction. Stable enzyme forms are indicated by a single letter labeled sequentially beginning with E to represent the free enzyme. An example of stable enzyme form other then the free enzyme would be any transiently formed enzyme complex with a covalently bound intermediate. Species enclosed in brackets represent central enzyme complexes that can not be kinetically distinguished. The nomenclature of a kinetic mechanism is based on the order of substrate addition and product dissociation in the forward direction using the identifiers Uni, Bi, Ter, and Quad. In the mechanism represented above, two substrates add (Bi), one product dissociates (Uni), one substrate adds (Uni), and finally two products dissociate (Bi). Kinetic mechanism can be further distinguished as being either sequential or ping pong. In a sequential mechanism, all substrates must be bound by the enzyme before the reaction can occur. In a ping pong mechanism, one or more products must dissociate before all the substrates have combined with the enzyme. Using this nomenclature, the mechanism represented above would be termed Bi Uni Uni Bi ping pong. As represented, this mechanism is ordered because there is only one sequence by which substrates can be bound (and products released) by the enzyme. If there were more than one sequence by which these events could occur, the mechanism is termed random. Every kinetic mechanism described by the nomenclature above will have a distinguishing set of initial velocity and product inhibition patterns that can be predicted by the application of a small set of rules. The experimental procedure to obtain these patterns involves measuring the  181  initial velocity of the reaction while varying the concentration of one substrate in the presence of changing fixed concentration of a second substrate or inhibitor. The data is plotted in the form of a double reciprocal plot (VQ" versus [variable substrate]") that will result in set of lines that may either not intersect (parallel), intersect on the the y-axis, or intersect to the left of the y-axis. The rules to predict or interpret the observed pattern can be divided into the rules that govern slope effects and those that govern intercept effects. The rules for both initial velocity and product inhibition experiments are the same though they differ slightly in their application. 1  1  A.4.1 Rules for predicting initial velocity patterns. a) Intercept effects. The y-intercept of a double reciprocal plot corresponds to the velocity in the presence of an infinite concentration of the variable substrate (V ). An initial velocity plot will almost always show a slope effect because the changing fixed substrate is normally combining with a different enzyme form, the concentration of which is unaffected by saturating concentrations of the variable substrate. b) Slope effects. Slope effects are observed if there is a reversible connection between the points at which the variable substrate and the changing fixed substrate combine with the enzyme. A irreversible connection normally arises because a product dissociates from the enzyme between the two points of substrate addition. Under initial velocity conditions, the concentration of product is essentially zero and thus combination of the product with the enzyme will not occur. An irreversible connection can also arise from addition of a substrate present at saturating concentrations or from a chemical transformation that is effectively irreversible due to a high equilibrium constant in favor of product formation. A parallel (no slope effect) pattern is indicative of a ping pong mechanism. c) Multiple addition of the same substrate. If the same substrate adds more than once in the course of the kinetic mechanism, the intercept and slopes effects for each point of addition must be considered separately. The observed effect will be the combination of all predicted effects. If there is more then one expected slope effect, parabolic reciprocal plots will be observed if the two points of combination are reversibly connected. max  A.4.2 Rules for predicting product inhibition patterns. a) Intercept effects. Intercept effects occur when the variable substrate and the inhibitor combine with different enzyme forms. If they combine with the same enzyme form, saturation with the variable substrate (at the y-axis of the reciprocal plot) will eliminate any effect from the inhibitor and thus no intercept effect is observed. b) Slope effects. Slope effects are observed if there is a reversible connection between the point at which the variable substrate and the inhibitor combine with the enzyme. c) Multiple addition of the same inhibitor. As with substrates, the effects at each point of addition must be considered separately and then combined in order to predict the observed effect. A.4.3 Distinguishing random from ordered mechanisms. Using the rules above it is possible to distinguish if substrate binding and product release is ordered or random. An ordered mechanism has only one sequence for the overall mechanism while a random mechanism has at least two. Shown below is an example of Bi Uni Uni Bi ping pong mechanism in which both substrate binding and product release is random.  182  A random mechanism should always give nonlinear reciprocal plots, however the curvature may not always be obvious, and thus one can not rely on the observation of linear plots to rule out a random mechanism. The following example illustrates how the above rules can be used to distinguish an ordered from a random mechanism. In the ordered Bi Uni Uni Bi ping pong mechanism, R will show competitive product inhibition when A is the variable substrate. There is a reversible connection between release of product R and binding of substrate A, so slope effects are observed. Intercept effects are not observed because A and R combine with the same enzyme form. However, if the random Bi Uni Uni Bi ping pong mechanism is operative, slope effects will still be observed but so will intercept effects. This is because R can now combine with a different enzyme form that will still be present in a finite concentration in the presence of saturating concentrations of A. A.5 D E R I V A T I O N O F T H E R A T E E Q U A T I O N F O R U D P G l c D H  In this section the rate equation for UDPGlcDH is derived using the method of King and Altman (184), as described by Cleland (91). The equation derived here differs from the standard rate equation for a Bi Uni Uni Bi ping pong mechanism because two of the substrates (NAD ) are the same species, as are two of the products (NADH). This rate equation has been derived previously for histidinol dehydrogenase (100, 103). The equation derived here differs only slightly in the definition of several constants and corrects a few typographical errors from that publication. Based on results in our laboratory, the first molecule of NADH is not release from the enzyme until after formation of the thiohemiacetal intermediate. For the purposes of this derivation, it will be assumed that the second molecule of NADH is released before the final irreversible hydrolysis step, although this is not known. These decisions will effect neither the overall form nor the interpretation of the resulting rate equation. +  183 P  p  k  EA  4  k  A  5 6 k  /EAB\  k  F  FP  8  A  Q  9 10  k  k  k  A  11  /FB \  G T * E Q  GP  *13  k  12  E  -= (EAB =F^= FP) =F== F <- \ / kP i i  EA  6  kB  E  7  k  8  kQ 12  k  1 3  EQ-«  G  GP  kioP  :  FB  Shown above is the graphical representation of the kinetic mechanism of UDPGlcDH using the conventions of both Cleland (top) and King-Altman (bottom). The latter is much more useful for the purpose of this derivation. The concentration of substrates, products, and different enzyme forms are each represented by a single letter: A, UDP-glucose; B, NAD ; P, NADH; Q, UDPGlcA; E, free enzyme; F, enzyme with covalently bound thiohemiacetal; G, enyme with covalently bound thioester. All rate constants with odd numbers are for the forward direction while even numbers are for the reverse reaction. The rate constant for the irreversible hydrolysis step is denoted by k^. This mechanism has seven different enzyme complexes so the total enzyme concentration, [E] = (E) + (EA) + (EAB<-»FP) + (F) + (FB<->GP) + (G) + (EQ). The +  T  concentration of each of the enzyme species, relative to the total concentration of enzyme, is sum of rate constants describing all possible reaction sequences that lead to the formation of  given enzyme species. Since there are seven (n) different enzyme species, the number of rate constants included in each sequence is limited to six (n-l). Applying this strategy to each of the seven different enzyme forms gives the equations describing the relative concentration of species. (E) = k k k Pk k Pk + k^kePkfjknk^ + k k k Pk k 3k9 + k^kuk^kgkyB [E]j kzknk^kgkyBks + k ^ ^ k / B I ^ B 2  4  6  8  10  11  2  4  6  11  1  Eq. 20  (E) + (EA) + (EAB=S=FP) + (F) + (FB=S=GP) + (G) + (EQ) (EA) = k k Pk ki Pk Ak ~[ET i^Akukwkgk/Bkjj 4  6  8  0  1  11  l^kePks^Aknk^ + k k Pk Ak k k9 + M^Akuk^kgkyB 4  6  1  11  13  t  Eq.21  (E) + (EA) + (EAB=5=FP) + (F) + (FB=S=GP) + (G) + (EQ) (EAB=i=FP) = kePkgk^PkaB^Akn + kePkgkaB^Aknk^ + kePkaB^Akuk^kg + kaBkiAknk^kgkyB Eq. 22 [Eh (E) + (EA) + (EAB=S=FP) + (F) + (FB=5=GP) + (G) + (EQ) kskioPkskaBkiAku + ksksksB^Aknk^ + ^ B ^ A k ^ ^ g [E]  T  (E) + (EA) + (EAB=S=FP) + (F) + (FB=S=GP) + (G) + (EQ)  Eq. 23  184 (FB=s=GP) =  k Pk Bk5k Bk Ak 10  7  3  1  + k Bk k Bk Ak k 3  11  7  5  3  1  Eq. 24  1  11  (E) + (EA) + (EAB=S=FP) + (F) + (FB^GP) + (G) + (EQ)  [ E J t  (G) = kgkyBkgkaB^Aku [E]j (E) + (EA) + (EAB=S= FP) + (F) + (FB =S=GP) + (G) + (EQ) (EQ) [E] T  Eq. 25  = k 2Qk k k Pk k oP + k Qk k4k Pk k + k Qk2k4k Pk k9 + k Qk k k k Bk + k Qk k9k Bk k B + k 3k k Bk5k Bk A 1  2  12  4  2  6  8  13  9  1  7  12  5  2  12  6  13  8  7  5  13  12  3  6  1  13  7  9  3  k Qk k4k k9k7B + 12  2  1  13  Eq. 26  (E) + (EA) + (EAB=5=FP) + (F) + (FB=S=GP) + (G) + (EQ)  The net rate of production of product (Q) is described by the equation: d(Q) =  k (G)[E] 13  Eq. 27  T  (E) + (EA)+ (EAB=5=FP) + (F) + (FB=S=GP) + (G) + (EQ)  Substituting the relative enzyme concentrations (as described by equations Eq. 20-26) into Eq. 27 and grouping like terms gives the complete rate law for the reaction catalyzed by UDPGlcDH. v =  ABZk^kukgkT-kskaMEj-r  £  ^  B(k k k k k k + k k k k k k2) + E^^u^i^z AB(k k k k k k + k^kukgk^^ + ^ ^ ^ 3 ^ + k^kukgksksk,) + AB2(k k k k k k + k ^ n ^ k ^ + k ^ ^ k ^ k ^ g M ^ ^ ) + AP(k k k k k k + k^kukskek^) + AP2k k k k k k + BQ(k k k k k k + k k kgk k k ) + B2Qk k k k k k + ABP(k k k k k k + k ^ k ^ ^ ^ + k n ^ o M ^ ^ ) + ABP2k k k k k k AB2pk k k k k k + P(ki3knk8kek4k + k^kukg^k^) + P2k k k k k k + QP(k k k k k k + k k k k k4k ) + QP2k k k k k k +  13  11  13  9  7  11  13  13  11  11  13  4  9  7  9  9  12  2  3  4  7  11  9  7  5  1  7  6  9  13  4  1  1  4  11  2  13  13  11  8  6  3  1  11  10  8  6  3  1  12  7  5  2  10  7  10  13  8  12  6  9  4  7  1  5  3  4  2  +  11  5  3  2  13  12  8  6  1  11  4  2  13  12  9  6  10  2  8  12  6  4  10  8  2  6  In this form, the rate equation is unwieldy and very difficult to interpret. The equation can be greatly simplified by defining kinetic constants that are a function of the individual rate constants. For simple kinetic mechanisms, the kinetic constants sufficient to describe the rate law are a maximum velocity and both a Michaelis constant and an inhibition constant for each reactant. For terreactant mechanisms, the equation can be significantly more complex and finding a consistent set of kinetic constants is nontrivial. v  =  V  V  K AK BB M  +  2  K B2 A  m  +  AB2  K™ AB + AB + KPK *KfPP + B  2  3  m  K PK P 2  3  Ki K-3 AP2 + K K B Q + K B Q + K ABP + K PK P KQ K-,0. KP B  B  A  m  2  3  B  A  2  2  m  B  3  n  3  K ABP + AB P + KjAK^KaBKjPP + K i K i K 3 P + K PK P K P K PK P K PK P B  2  2  A  3  2  3  3  2  3  K^K^KaBKjPQP + K K K Q P 2 K K K3 KiQK-jPKrjP 1  Q  n  2  P  P  A  1  B  B  3  B  B  2  3  2  Eq. 29  185 Where the kinetic constants are defined as: Vmax =  k^kukgkstEJT (k k k + k knk + knkgk+ k k k  Ki = k2_ A  ki K,P = k (k + k ) ki(ki knk + k knk + k^kgks* k k k ) kioks K B = k k (k k k + k k k + k k k + k k k ) K P = (k k k + k k k + k k k ) k k (k 3k kg + k knk + knkgk+ k k k ) kiokgkg = k K P = (k knk + k knk + knkgk + k k k ; ks" k nk k 13  K  A  m  11  9  13  5  5  13  9  5  ki3knkgk  =  5  3  M  13  7  9  11  9  1  3  7  13  4  9  11  13  5  7  13  5  13  8  5  9  5  9  5  3  13  9  3  5  5  2  =  2  13  13  k)  K,Q  5  13  9  6  10  k  3  8  6  13  9  6  7  10  8  8  5  9  13  5  5  13  9  5  5  = _kii  kk K B = (k k k + k k k + k k k ) kiok k 13  6  10  2  5  8  -  3  k(k4 +  3  8  5  4  K B  9  12  5  5  Inverting both sides of Eq. 29 provides the double reciprocal form of the rate equation (Eq. 1) that corresponds with the experimental plots in Chapter 3 and will aid in the interpretation of the initial velocity and product inhibition data. J_ V  =_J_  ( K AK B m  Vmax I  + -i + K ^ K ^ K ^ P  + KmA + J ^ B  2  AB  A  B  2  K BK BP2 + K A K B Q 1  3  2  Ki AB  K PK PB 2  B  K P  3  K  3  KI KIBK K-|PQP + A  3  2  Eq.  2  Q  p 2  B 3  Ki P p  3  + K BP  +  3  K / B + K-| Ki K A  K PAB2  B  B 3  P  2  +  K PK PAB2  3  2  3  K AK BK BQP2\  B  1  K QK PK PAB2 1  A  Q  p + K-|AKi K  +  3  M  Ki A  Q  3  K BP2  + K  2  m  K PK PB2  +  K PK PB2  1  3  K!QK PK PAB2 J  3  2  3  Under initial velocity conditions there are no products present so P = Q = 0 and Eq. 1 reduces to a much simpler equation: Km K A  B 2  + K  m  m  AB  Vo  Eq. 30  A+ K B + 1  B  The equation describing a plot of v ~' against [A]" can be mathematically expressed in the form y = mx + b. 1  0  J_  A  Km K B A  2  BV  M  + Km  A  Vmax  K B M  BVmax  + J _  Eq.31  Vmax  In Eq. 31 the concentration of the second substrate (B), is a variable in both the slope and the y-intercept terms so the initial velocity plot will be a set of linear lines that intersect to the left of the y-axis. A replot of intercepts against [B]" will be linear with a x-intercept of (K )"' and a y-intercept of (Vmax)". A similar result is predicted when B is the variable substrate and a replot of intercepts against [A]" will have a x-intercept of (K )'' and the same y-intercept 1  B  m  1  1  A  m  (Vmax)"'. Eq. 1 can also be used to predict the product inhibition patterns. If Q is tested as an inhibitor against substrate A, substrate B is present in saturating concentrations and all terms  186  with B in the denominator become vanishingly small. Product P is absent so all terms containing P are set to zero. J_  V  = 1 0  ( K A + K AQ M  M  A I V™  \  K^Vrraxi  +  1  V™,  Eq. 32  Eq. 32 describes a competitive inhibition pattern because only the slope and not the intercept is dependent on the concentration of Q. A replot of slopes against [Q] will give a straight line with an x-intercept of - K i . K i is the inhibition constant for product Q and will be represented as K ® . Using a similar treatment to that described above for inhibition of product Q against the variable substrate A, the remaining combinations of inhibitors and variable substrates can be examined. Q  Q  Table A.1 Predicted product inhibition patterns for UDPGlcDH Inhibitor Variable Predicted Replot of intercepts substrate Pattern Q "A Competitive NA P A Uncompetitive X-intercept of -K (-K ) Q B No inhibition NA P B Noncompetitive X-intercept of -K p  P  3  p  Replot of slopes X-intercept o f - K (-Ki ) NA NA Complex and nonlinear Q  Q  {  The inhibition constant for NADH ( K ) is K 3 , which is a complex function of rate constants as defined in Eq. 29. This complexity reflects the fact that it binds to the enzyme at two separate points in the enzyme reaction and thus can not be characterized with a simple dissociation constant. It can be shown the rate equation describing inhibition by NADH simplifies to v = Vm /2 when P is set to K.3 . This is consistent with the empirical definition of the inhibition constant. p  P  0  ax  P  187  A P P E N D I X B: *H N M R S P E C T R A O F CHARGED  COMPOUNDS  188  200 MHz  !  H N M R in D 2 0  N02808.102 AU PROB: X00.AU DATE 28-11-96 SF 200.133 8Y 80.0 01 3960.000 81 32768 TD 32768 8M 4000.000 HZ/PT .244  200 MHz H l  N M R in D , 0 O  3E100S.128 All PROG: XOO.lU DAU 10-12-98 SP SV OJ SI  200.233 80.0 39S0.OOO 32768 ^a 32768 SH •1000. 000 H7. PT .2*4 PK  0.0  m  o.o 4.036 32 32 298  RO  MS TK  FX DP  5000 3630.000 631. PO  13 G3 CX  .300 .100 23.00  02  Ft  F2 SH  17.00  '.;.50»p  -i.ooo;-' cy ,Cflsoo.osa .500 2834. C'J  *******  HO  ' OUDP  189  200 MHz  !  H N M R in H  AP3008.104  D0 2  !i/  o H  AU PROS:  9a  X00.AU DATE 30-4-97 8F 200.133 8Y 80.0 01 3990.000 81 32768 TD 32788 SW 4000.000 HZ/PT .244  200 MHz  !  H N M R in  HO  I  OP0 Na 3  D0 9  0C0203.103 AU PROS: XOO.AU DATE 8-10-97 200.133 8Y 80.0 01 3960.000 81 32768 TO 32768 8* 4000.000 HZ/PT .244 PM RD AB R6 N8 TE  0.0 0.0 4.096 32  FN 02 DP  S000 3830.000 63L PO  .300 .100 2B.00 CX 17.00 CY 11.600P Fl -1.000P F2 HZ/CM 100.068 PPM/CH  HO  ' OUDP  2  190  200 MHz R N M R in D,0 l  AP3008.10B AU PROft X00.AU DATE 30-4-97  9b  SF 200.133 SY BO.O 01 3980.000 81 32768 TD 32768 8M 4000.000 HZ/PT .244 PH RD AQ  ne  N3 TE FN 02 DP  OP0 Na 3  0.0 0.0 4.096 20 32 298 5000 3830.000 63L PO  .300 .100 CX 25.00 CY 17.00 Fl 11.BOOP F2 -l.OOOP HZ/CM 100.068 PPM/CM .500 Sfl 2834.03  200 MHz H N M R in D 0 !  2  H .OH H C ^ / 3  H O ^ \ - ^ ° \ 200.133 SY 80.0 01 39S0.000 81 32768 TD 32768 8N 4000.000 HZ/PT .244  2b  HO ' OUDP  2  191 200 MHz 'H NMR in D 0 2  SE300S.106 AU PROG: X00.AU DATE 30-9-98 SF 200.133 SY 80.0 01 3950.000 SI 32768 TD 32768 SN 4000.000 HZ/PT .244 PM RO AO RB NS TE  0.0 0.0 4.098 20 32 298  FM 02 OP  5000 3630.000 63L PO  200 MHz H NMR in D 0 [  2  S* .'OOO. 000 H2.-PT PW  0.0  iii  4.  RG  6-1  Tr,  ?88  :i  tl.O  3?.  10.0  3.0  8.0  7.0  6.0  5.0  4.0  3.0  8.0  1.0  192  APPENDIX C: CRYSTALLOGRAPHIC DATA FOR 1,2,3,4,6-PENTA-OACETYL-7-DEOXY-D-GX YCERO-$-D-GL  HEPTOPYRANOSE (8b)  UCO-  193  C.l Experimental Methods The X-ray structure o f 8b was determined by M r . Steve Rettig o f the Crystallography Lab, Department o f Chemistry, U B C .  C.1.1 Data Collection  A colorless irregular crystal o f C17H24O11 (8b) having approximate dimensions o f 0.30 x 0.35 x 0.45 m m was mounted on a glass fiber. A l l measurements were made on a R i g a k u / A D S C C C D area detector with graphite monochromated M o - K a radiation. C e l l constants and an orientation matrix for data collection based on 14211 reflections with 29 = 5.0- 63.7° corresponded to a primitive orthorhombic cell with dimensions: a = 8.977(1) A b = 12.6645(5) A c = 17.5459(6) A V = 1994.9(2) A  3  For Z = 4 and F . W . = 404.37, the calculated density is 1.35 g / c m . The systematic absences of: 3  hOO: h * 2 n OkO: k * 2 n 001:1 * 2n uniquely determine the space group to be: P2i2,2i (#19) The data were collected at a temperature o f -93 ± 1 °C to a maximum 29 value o f 63.7°. Data were collected in 0.50° oscillations with 70.0 second exposures. A sweep o f data was done using (j) oscillations from 0.0 to 190.0° at x -90° and a second sweep was performed using co oscillations between -22.0 and 18.0° at x ~ -90°. The crystal-to-detector distance was 39.14(3) mm. The detector swing angle was -10.00°. =  C.1.2 Data Reduction O f the 18314 reflections which were collected, 3061 (29 = 60°) were unique (R,„, = 0.049); equivalent reflections were merged. The linear absorption coefficient, ju, for M o - K a radiation is 1.1 cm" . The data were corrected for Lorentz and polarization effects. max  1  C.1.3 Structure Solution and Refinement The structure was solved by direct methods (185) and expanded using Fourier techniques (186). The non-hydrogen atoms were refined anisotropically. Hydrogen atoms were fixed i n calculated positions with C - H = 0.98 A (methyl orientations based on difference map peak  194  positions). The final cycle of full-matrix least-squares refinement was based on all 3061 reflections (20 = 6 0 ° ) and 253 variable parameters and converged (largest parameter shift was 0.0005 times its esd) with unweighted and weighted agreement factors of: 11  max  R = S||Fo | - |Fc ||/E|Fo | = 0.070 R = ( ( Z ^ | F o | - |Fc |) /Z<yFo )) = 0.081 2  2  2  2  2  2  4  05  w  The standard deviation of an observation of unit weight was 2.40. The weighting scheme was based on counting statistics. Plots of Y.OJ(\FO \ - |Fc |) versus |Fo |, reflection order in data collection, sin 9/A, and various classes of indices showed no unusual trends. The maximum and minimum peaks on the final difference Fourier map corresponded to 0.36 and 0.38 e"/A , respectively. Neutral atom scattering factors were taken from Cromer and Waber (187). Anomalous dispersion effects were included in Fcalc (188); the values for A f and A f were those of Creagh and McAuley (189). Absolute configuration is based on 4 known centers. The values for the mass attenuation coefficients are those of Creagh and Hubbel (190). A l l calculations were performed using the teXsan (191) crystallographic software package of Molecular Structure Corporation. 12  2  2  2  2  3  A summary of the experimental details is provided in Table C . l Table C.l Experimental details for the structure determination of compound 8b  Crystal Data Empirical Formula Formula Weight Crystal Color, Habit Crystal Dimensions Crystal System Lattice Type Lattice Parameters  C17H24O11  404.37  Colorless, irregular 0.30 X 0.35 X 0.45 mm Orthorhombic Primitive a = 8.977(1) A b=  12.6645(5) A  c = 17.5459(6) A V =  Space Group Z value  1994.9(2) A  3  P 2 i 2 i 2 i (#19) 4  1.346 g/cm  Dcalc  Fooo p(MoKa)  3  856.00 1.14  cm-  1  Intensity Measurements Diffractometer Radiation  11  12  Rigaku/ADSC C C D M o K a (X = 0.71069 A) graphite monochromated  Least-Squares: Function minimized: E<y(|Fo | - |Fc |) 2  2  2  Standard deviation of an observation of unit weight: where: No = number of observations and Nv = number of variables.  2  2 2  03  (Eft)(|Fo | - |Fc |) /(No - Nv))  195 Detector Aperture Data Images (j) oscillation Range (% = -90) co oscillation Range (% = -90) Detector Position Detector Swing Angle 2&m ax  No. of Reflections Measured Corrections  Structure Solution and Refinement Structure Solution Refinement Function Minimized Least Squares Weights p-factor Anomalous Dispersion No. Observations (2# = 60°) No. Variables Reflection/Parameter Ratio Residuals (on F , all data): R; Rw Goodness of Fit Indicator No. Observations (I > 3a(I)) Residuals (on F , all data): R; Rw Max Shift/Error in Final Cycle Maximum peak in Final Diff. Map Minimum peak in Final Diff. Map max  2  2  94 mm x 94 mm 460 exposures at 70.0 seconds 0.0 - 190.0° -22.0-18.0° 39.14(3) mm -10.00° 63 7° Total: 18314 Unique: 3061 (R = 0.049) Lorentz-polarization Absorption/ scaling (trans, factors: 0.8776 - 0.9990) int  Direct Methods (SIR92) Full-matrix least-squares Xco(\Fo \ - \Fc \f co = (a (Fo) )"' 0.0000 All non-hydrogen atoms 3061 253 12.10 0.070; 0.081 2.40 1783 0.042; 0.039 0.0005 0.36 e"/A -0.38 e"/A 2  2  2  2  196  C.2 Results Table C.2 Atomic coordinates and B for compound 8b Atom* X Z B y 2.54(4) 0.4256(2) 0.52504(9) 0.1969(1) 0(1) 2.71(4) 0.4115 (2) 0.3743(1) 0.53181(9) 0(2) 0.2434(2) 0.3499(2) 3.83(5) 0.6259(1) 0(3) 0.3246(2) 2.45(4) 0.3818(1) 0.37528(9) 0(4) 4.68(6) 0.5515(2) 0.4564(2) 0.3669(1) 0(5) 0.3447(2) 2.65(4) 0(6) 0.1913(1) 0.29533(9) 0.0993(2) 0.2097(2) 0.2706(1) 5.20(6) 0(7) 0.2980(2) 0.0062(1) 0.38089(9) 2.65(4) 0(8) 0.5065(2) -0.0726(2) 0.3384(1) 3.91(5) 0(9) 3.06(4) 0.4274(2) -0.0863 (1) 0. 5158(1) 0(10) 0.5293(2) -0.1342(2) 6.94(8) 0.6646(3) 0(11) 0.2855(2) 0.4942 (1) 2.36(6) 0.3517(3) C(l) 0.2903(2) 2.22(6) C(2) 0.3883 (3) 0.4103(1) 0.1931(2) 0.3168 (3) 0.3760(1) 2.27(5) C(3) C(4) 0.0935(2) 0.4117(1) 2.26(6) 0.3834(3) 0.0973(2) 2.31(6) C(5) 0.3679(3) 0.4981(1) 0.0191(2) 0.5416(1) 2.79(6) C(6) 0.4653 (3) 0.4412 (3) 0.0272(2) 0.6272(1) 3.18(7) C(7) 0.4002 (2) 0.6002(1) 2.74(6) C(8) 0.3439(3) 0.6336(2) 0.4134(4) 0.4952 (2) 3.81(7) C(9) 0.4202 (3) 0.3536 (2) 2.98(7) C(10) 0.4595 (2) 0.3135(2) 3.42(7) C(ll) 0.3406(3) 0.5463 (2) 0.2031(2) 0.2484(2) 3.10(7) C(12) 0.2263(3) 0.2748(4) 0.2091(2) 0.1680(1) 3.86(7) C(13) C(14) 0.3739(4) -0.0705(2) 0.3433(1) 3.01(7) 0.2682(4) -0.1488(2) 0.3094(2) 4.17(8) C(15) 0.5391(4) -0.1573(2) 0.5142 (2) 3.49(7) C(16) 0.4911(2) 4.29(8) 0.4885 (4) -0.2640(2) C(17) B — (8/3)rc (t7n(aa*) + £/ 2(bb*) + t7 (cc*) + 2t/i aa*bb*cosY + a  e q  e q  2  2  e<7  2t7i3aa*cc*cos(3 + 2l723bb*cc*cosa) b  2  2  2  Atoms are named as indicated in Figure 3.14  33  2  197  Table C.3 Bond lengths (A) for compound 8b Atom Atom C(l) 0(1) 0(2) C(l) 0(3) C(8) 0(4) C(10) C(3) 0(6) C(12) 0(7) C(14) 0(8) C(6) 0(10) C(16) 0(11) C(2) C(3) C(4) C(5) C(7) C(6) C(ll) C(.IO) C(14) C(15) O(l) C(5) C(8) 0(2) C(2) 0(4) C(10) 0(5) C(12) 0(6) C(4) 0(8) C(14) 0(9) C(16) 0(10) C(2) C(l) C(4) C(3) C(6) C(5) C(9) C(8) C(13) C(12) C(16) C(17)  Distance 1.411(3) 1.410(3) 1.193(3) 1.360(3) 1.437(3) 1.207(3) 1.358(3) 1.451(3) 1.194(3) 1.513(4) 1.524(3) 1.522(4) 1.488(4) 1.495(4) 1.442(3) 1.385(3) 1.431(3) 1.202(3) 1.353(3) 1.449(3) 1.194(3) 1.347(3) 1.510(3) 1.530(4) 1.525(3) 1.476(4) 1.477(4) 1.482(4)  198  Table C.4 B o n d angles ( ° ) for compound 8b atom Atom C(l) 0(1) C(2) 0(4) C(4) 0(8) 0(1) C(l) 0(2) C(l) 0(4) C(2) 0(6) C(3) C(2) C(3) 0(8) C(4) 0(1) C(5) C(4) C(5) 0(10) C(6) 0(2) C(8) 0(3) C(8) 0(4) C(10) 0(6) C(12) 0(7) C(12) 0(8) C(14) 0(10) C(16) C(16) 0(11) C(l) 0(2) C(3) 0(6) C(6) 0(10) C(l) 0(1) 0(4) C(2) C(2) C(l) 0(6) C(3) 0(8) C(4) C(3). C(4) C(5) 0(1) 0(10) C(6) C(5) C(6) 0(2) C(8) 0(4) C(10) 0(5) C(10) 0(6) C(12) 0(8) C(14) 0(9) C(14) 0(10) C(16)  Atom C(5) C(10) C(14) 0(2) C(2) C(3) C(2) C(4) C(5) C(4) C(6) C(7) 0(3) C(9) C(ll) 0(7) C(13) C(15) 0(11) C(17) C(8) C(12) C(16) C(2) C(l) C(3) C(4) C(3) C(5) C(6) C(5) C(7) C(9) 0(5) C(ll) C(13) 0(9) C(15) C(17)  angle 113.6(2) 117.0(2) 117.5(2) 106.0(2) 109.9(2) 108.6(2) 109.3(2) 110.0(2) 110.3(2) 108.7(2) 115.1(2) 109.6(2) 122.1(3) 127.3(3) 111.3(2) 123.6(2) 125.6(3) 110.4(3) 122.3(3) 125.0(3) 115.3(2) 117.5(2) 116.5(2) 107.6(2) 111.4(2) 105.3(2) 108.8(2) 105.6(2) 110.2(2) 101.4(2) 107.9(2) 111.6(2) 110.6(2) 122.8(3) 125.9(3) 110.8(2) 123.5(3) 126.1(3) 112.7(3)  199  REFERENCES  1) Feingold, D. S., and Franzen, J. S. (1981) Trends Biochem. Sci. 6, 103-105. 2) Rossmann, M. G., Liljas, A., Branden, C.-L, and Banaszak, L. J. (1975) in The Enzymes (Boyer, P. D., Ed.) Vol. 11, 3 ed., pp 61-102, Academic Press, New York, NY. rd  3) Bellamacina, C. R. (1996) FASEBJ. 10, 1257-1269. 4) Jornvall, H., Persson, M., and Jeffery, J. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 4226. 5) Oppenheimer, N. J., and Handlon, A. L. (1992) in The Enzymes (Sigman, D. S., Ed.) Vol. 20, 3 ed., pp 453-505, Academic Press, San Diego, CA. rd  6) Inoue, T., Sunagawa, M., Mori, A., Imai, C , Fukada, M., Takagi, M. and Yano, K. (1989) J.  Bacteriol. 171,3115-3122. 7) Persson, B., Krook, M., and Jornvall, H. (1991) J. Biochem. 200, 537-543. 8) Jornvall, PL, Persson, B., Krook, M., Atrian, S., Gonzalez-Duarte, R., Jeffery, J., and Ghosh, D. (1995) Biochemistry 34, 6003-6013. 9) Ghosh, D., Wawrzak, Z., Weeks, C. M., Duax, W. L., and Erman, M. (1994) Structure 2, 629-640. 10) Ghosh, D., Pletnev, V., Zhu, D. W., Wawrzak, Z., Duax, W. L., Pangborn, W., Labrie, F. and  Lin, S. X. (1995) Structure 3, 503-513. 11) Varughese, K. I., Skinner, M. M., Whiteley, J. M., Matthews, D. A., and Xuong, N. PI. (1992) Proc. Natl. Acad. Sci. U.S.A. 89, 6080-6084. 12) Tanaka, N., Nonaka, T., Nakanishi, M., Deyashiki, Y., Hara, A. and Mitsui, Y. (1996)  Structure 4, 33-45. 13) Tanaka, N., Nonaka, T., Tanabe, T., Yoshimoto, T., Tsuru, D. and Mitsui, Y. (1996)  Biochemistry 35, 7715-7730.  200  14) Andersson, A., Jordan, D., Schneider, G., and Lindqvist, Y. (1996) Structure 4, 1161-1170. 15) Auerbach, G., Herrmann, A., Gutlich, M., Fischer, M., Jacob, U., Bacher, A. and Huber, R. (1997) EMBOJ. 16, 7219-7230. 16) Hulsmeyer, M., Hecht, H. J., Niefind, K., Hofer, B., Lindsay, D. E., Timmis, K H., and Schomburg, D. (1998) Prot. Sci. 7, 1286-1293. 17) Nakajima, K., Yamashita, A., Akama, H., Nakatsu, T., Kato, H., Hashimoto, T., Oda, J., and Yamada, Y. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 4876-4881. 18) Bauer, A. J., Rayment, I., Frey, P. A. and Holden, H. M. (1992) Proteins: Struct. Fund. Genet. 12,372-381. 19) Benach, J., Silvia, A., Gonzalez-Duarte, R., and Ladenstein, R. (1999)/. Mol. Biol. 289, 335355. 20) Dengler, U., Niefind, K., Kieb, M., and Schomburg, D. (1997) J. Mol. Biol. 267, 640-660. 21) Kutzenko, A. S., Lamzin, V. S., and Popov, V. O. (1998) FEBS Lett. 423, 105-109. 22) Abad-Zapatero, C., Griffith, J. P., Sussman, J. L., and Rossman, M. G. (1987) J. Mol. Biol. 198, 445-467. 23) Birktoft, J. J., Rhodes, G., and Banaszak, L. J. (1989) Biochemistry 28, 6065-6081. 24) Goldberg, J. D., Yoshida, T., and Brick, P. (1994) J. Mol. Biol. 236, 1123-1140. 25) Schuller, D. J., Grant, G. A., and Banaszak, L. J. (1995) Nature Struct. Biol. 2, 69-76 26) Stoll, V. S., Kimber, M. S., and Pai, E. F. (1996) Structure 4, 437-447. 27) Lamzin, V. S., Dauter, Z., Popov, V. O., Harutyunyan, E. H., and Wilson, K. S. (1994) J. Mol. Biol. 236, 759-785. 28) Hurley, J. H , Dean, A. M., Koshland, D. E. Jr. and Stroud, R. M. (1991) Biochemistry 30, 8671-8676.  201  29) Adams, M. J., Ellis, G. H., Gover, S., Naylor, C. E., and Phillips, C. (1994) Struct. 2, 651668. 30) Barycki, J. J., O'Brien, L. K., Bratt, J. M., Rongguang, Z., Sanishvili, R., Strauss, A. W., and Banaszak, L. J. (1999) Biochemistry 38, 5786-5798. 31) Zhang, L., Chooback, L., and Cook, P. F. (1999) Biochemistry 38, 11231-11238. 32) Ye, Q., Hyndman, D., Li, X., Flynn, G., and Jia, Z. (2000) Proteins: Struct. Fund. Gen. 38, 41-48. 33) Schlegel, B. P., Jez, J. M., and Penning, T. M. (1998) Biochemistry 37, 3538-3548. 34) Henehan, G. T. M., and Oppenheimer, N. J. (1993) Biochemistry 32, 735-738. 35) Buehner, M., Ford, G. C , Moras, D, D., Olsen, K. W., and Rossman, M. G. (1974) J. Mol. Biol. 90, 25-49. 36) Mercer, W. D., Winn, S. I., and Watson, H. C. (1976) J. Mol. Biol. 104, 277-283. 37) Isupov, M. N., Fleming, T. M., Dalby, A. R., Crowhurst, G. S., Bourne, P. C , and Littlechild, J. A. (1999) J. Mol. Biol. 291, 651-660. 38) Harris, J. I., and Waters, M. (1976) in The Enzymes (Boyer P. D., Ed) Vol. 13, 3 ed., pp. 1rd  49, Academic Press, New York, NY. 39) Soukri, A., Mougin, A., Corbier, C , Wonacott, A., Branlant, C , and Branlant, G. (1989) Biochemistry 28, 2586-2592. 40) Boshi-Muller, S., and Branlant, G. (1999) Arch. Biochem. Biophys. 363, 259-266. 41) Corbier, C , Seta, F. D., and Branlant, G. (1992) Biochemistry 31, 12532-12535. 42) Liu, Z. J., Sun, Y. J., Rose, J., Chung, V. J., Hsiao, C. D., Chang, W. R., Kuo, I., Perozich, J., Lindahl, R., Hempel, J., and Wang, B. C. (1997) Nature Struct. Biol. 4, 317-326. 43) Steinmetz, C. G., Xie, P., Weiner, H , and Hurley, T. D. (1997) Structure 15, 701-711.  202  44) Moore, S. A., Baker, H. M., Blythe, T. J., Kitson, K. E., Kitson, T. M., and Baker, E. N. (1999) in Enzymology and Molecular Biology of Carbonyl Metabolism 7 (Weiner, H., Maser, E., Crabb, D. W., and Lindahl R. Ed.) pp 27-38, Kluwer Academic/Plenum Publishers, New York, NY. 45) Lamb, A. L., and Newcomer, M. E. (1999) Biochemistry 38, 6003-6011. 46) Johansson, K., El-Ahmad, M., Ramaswamy, S., Hjelmqvist, L., Jornvall, FL, and Eklund, H. (1998) Prot. Sci. 7, 2106-2117. 47) Hurley, T. d., Steinmetz, C. G., and Weiner, H., (1999) in Enzymology and Molecular Biology of Carbonyl Metabolism 7 (Weiner, H., Maser, E., Crabb, D. W., and Lindahl R. Ed.) pp 15-25, Kluwer Academic/Plenum Publishers, New York, NY. 48) Perozich, J., Nicholas, H , Wang, B.C., Lindahl, R., and Hempel, J. (1999) Prot. Sci. 8, 137146. 49) Perozich, J., Nicholas, H., Lindahl, R., and Hempel, J. (1999) in Enzymology and Molecular Biology of Carbonyl Metabolism 7 (Weiner, H., Maser, E., Crabb, D. W., and Lindahl R. Ed.) pp 1-7, Kluwer Academic/Plenum Publishers, New York, NY. 50) Kawamura, T., Ishimoto, N., and Ito, E. (1979) J. Biol. Chem. 254, 8457-8465. 51) Deretic, V., Gill, J. F., and Chakrabarty, A. M. (1987) Nucleic Acid. Res. 15, 4687-4581. 52) Roychoudhury, S., May, T. B., Gill, J. F., Singh, S. K., Feingold, D. S., and Chakrabarty, A. M. (1989) J. Biol. Chem. 264, 9380-9385. 53) Teng, H., Segura, E., and Grubmeyer, C. (1993) J. Biol. Chem. 268, 14182-14188. 54) Nagai, A., Kheirolomoom, A., and Ohta, D. (1993) J. Biochem, 114, 856-861. 55) Grubmeyer, C , Skiadopoulos, M., and Senior A. E., (1989) Arch. Biochem. Biophys. 272, 311-317.  203  56) Nagai, A., Daisaku, O., (1994) J. Biochem. 115, 22-25. 57) Teng, H., and Grubmeyer, C. (1999) Biochemistry 38, 7355-7362 58) Teng, H., and Grubmeyer, C. (1999) Biochemistry 38, 7363-7371. 59) Loper, J. C. (1968)7. Biol. Chem. 243, 3264-3272. 60) Yang, H. J., Lee, B., and Haslam, J. L. (1973) J. Mol. Biol. 81, 517-519. 61) Lawrence, C. M., Rodwell, V. W., and Stauffacher C. V. (1995) Science 268, 1758-1762. 62) Tabernero, L., Bochar, D. A., Rodwell, V. W., and Stauffacher C. V. (1999) Proc. Natl. Acad. Sci. 96,7167-7171. 63) Bochar, D. A., Tabernero, L., Stauffacher, C. V., and Rodwell, V. W. (1999) Biochemistry 38, 8879-8883. 64) Bochar, D. A., Stauffacher, C. V., and Rodwell, V. W. (1999) Biochemistry 38, 1584815852. 65) Strominger, J. L., Kalckar, H. M., Axelrod, J., and Maxwell, E. S. (1954) J. Am Chem. Soc. 76, 6411-6412. 66) Spicer, A. P., Kaback, L. A, Smith, T. J., and Seldin M. F. (1998) J. Biol. Chem. 273, 2511725124. 67) Dutton, G. J. (1980) in Glucuronidation of drugs and other compounds, CRC press, Boca Raton, FL. 68) Roden, L. (1980) in The Biochemistry of Glycoproteins and Proteoglycans (Lennarz, W. J., Ed.) pp 267-371, Plenum Publishing Co., New York. 69) Binari, R. C , Staveley, B. E., Johnson, W. A., Godavarti, R., Sasisekharan, R., and Manoukian, A. S. (1997) Development 124, 2623-2632. 70) Dalessandro, G., and Northcote, D. H. (1977) Biochem. J. 162, 267-279.  204  71) Dougherty, B. A., and van de Rijn, I. (1993) J. Biol. Chem. 268, 7118-7124. 72) Arrecubieta, C , Lopez, R., and Garcia, E. (1994) J. Bacteriol. 176, 6375-6383. 73) Hempel J., Perozich, J., Romovacek, H., Hinich, A., Kuo, I., and Feingold, D. S. (1994) Prot. Sci. 3, 1074-1080. 74) Schiller, J. G., Bowser, A. M., and Feingold, D. S. (1973) Biochim. Biophys. Acta 293, 1-10. 75) Schiller, J. G., Lamy, F., Frazier, R., and Feingold, D. S. (1976) Biochim. Biophys. Acta 453, 418-425. 76) Nelsestuen, G. L., and Kirkwood, S. (1971)/. Biol. Chem. 246, 3828-3834. 77) Ridley, W. P., and Kirkwood, S. (1973) Biochem. Biophys. Res. Commun. 54, 955-960. 78) Ridley, W. P., Houchins, J. P., and Kirkwood, S. (1975) J. Biol. Chem. 250, 8761-8767. 79) Ordman, A. B., and Kirkwood, S. (1977) J. Biol. Chem. 252, 1320-1326. 80) Eccleston, E. D., Thayer, M. L., and Kirkwood, S. (1979) J. Biol. Chem. 254, 11399-11404. 81) Schiller, J. G., Bowser, A. M., and Feingold, D. S. (1972) Carbohydr. Res. 25, 403-410. 82) Arrecubieta, C., Garcia, E., and Lopez, R. (1996)/. Bacteriol. 178, 2971-2974. 83) Ge, X., Campbell, R. E., van de Rijn, I., and Tanner, M. E. (1998) J. Am. Chem. Soc. 120, 6613-6614. 84) Ge, X. (2000) Covalent Catalysis in the UDP-Glucose Dehydrogenase Reaction. Ph.D. thesis, The University of British Columbia. 85) Strominger, J. L., Maxwell, E. S., Axelrod, J., and Kalckar, H. M. (1957) J. Biol. Chem. 224, 79-90. 86) Zalitis, J., and Feingold, D. S. (1969) Arch. Biochem. Biophys. 132, 457-465. 87) Campbell, R. E., Sala, R. F., van de Rijn, I., and Tanner, M. E. (1997) J. Biol. Chem. 212, 3416-3422.  205  88) Bradford, M.M. (1976) Anal. Biochem. 72, 248-254. 89) Laemmeli, U. K., (1970) Nature 227, 680-685. 90) Palmer, J. L., and Abeles, R. H. (1976) J. Biol. Chem. 251, 5817-5819. 91) Cleland, W. W. (1963) Biochim. Biophys. Acta 67, 104-137. 92) Cleland, W. W. (1963) Biochim. Biophys. Acta 67, 173-187. 93) Cleland, W. W. (1963) Biochim. Biophys. Acta 67, 188-196. 94) Eccleston, E. D., Thayer, M. L., and Kirkwood, S. (1979) J. Biol. Chem. 254, 11399-11404. 95) Jaenicke, R., Rudolph, R., and Feingold, D. S. (1986) Biochemistry 25, 7283-7287. 96) Franzen, J. S., Ashcom, J., Marchetti, P., Cardamone, J. J. Jr., and Feingold, D. S. (1980) Biochim. Biophys. Acta 614, 242. 97) Franzen, J. S., Marchetti, P. S., and Feingold, D. S. (1980) Biochemistry 19, 6080-6089. 98) Ding, L., Seto, B. L., Ahmed, A., and Coleman, W. G. (1994) J. Biol. Chem. 269, 2438424390. 99) Morgan, P. M., Sala, R. F., and Tanner, M. E. (1997)7. Am. Chem. Soc. 119, 10269-10277. 100)  Gorisch, H., (1979) Biochem. J. 181, 153-157.  101)  Ordman, A. B., and Kirkwood, S. (1977) Biochim. Biophys. Acta 481, 25-32.  102)  Wieranga, R. K., Terpstra, P., and Hoi, W. G. J. (1986) J. Mol. Biol. 187 , 101-107.  103)  Burger, E., and Gorisch, H. (1981) Eur. J. Biochem. 116, 137-142.  104)  Kheirolomoom, A., Mano, J., Nagai, A., Ogawa, A., Iwasaki, G., and Ohta, D. (1994)  Arch. Biochem. Biophys. 312, 493-500. 105)  Dugas, H. (1996) in Bioorganic Chemistry: A practical approach to enzyme action  ed., Springer-Verlag, New York, NY. 106)  Flentke, G. R., and Frey, P. A. (1990) Biochemistry 29, 2430-2436.  206  107)  Franzen, J. S., Ishman, R., and Feingold, D. S., (1976) Biochemistry 15, 5665-5671.  108)  Franzen, J. S., Marchetti, P. S., and Feingold, D. S., (1980) Biochemistry 19, 6080-6089.  109)  Franzen, B., Carrubba, C , Feingold, D. S., Ashcom, J., and Franzen, J. S. (1981)  Biochem. J. 199, 603-609. 110)  Gainey, P. A., Pestell, T. C., and Phelps, C. F. (1972) Biochem. J. 129, 821-830.  111)  Franzen, J. S., Marchetti, P. S., Ishman, R., Ashcom, J., and Feingold, D. S. (1978)  Biochem. J. 173, 701-704. 112)  Farres, J., Wang, T. T. Y., Cunningham, S. J., and Weiner, PI. (1995) Biochemistry 29,  2592-2598. 113)  Wong, Y. H. H., Winer, F. B., and Frey, P. A. (1979) Biochemistry 18, 5332-5336.  114)  Winer, F. B. (1972) Studies of Reversible and Irreversible Inhibition of Uridine  Diphosphate Galactose-4-Epimerase M.Sc. thesis, The Ohio State University. 115)  Mahal, L. K., Yarema, K. J., and Bertozzi, C. R. (1997) Science 276, 1125-1128.  116)  Dasser, M., Chretien, F., and Chapleur, Y. (1990) J. Chem. Soc. Perkin Trans. 1, 3091-  3094. 117)  Czeraecki, S., Horns, S., and Valery, J. M. (1995) J. Org. Chem. 60, 650-655.  118)  Ito, T., Ezaki, N., Niida, T. (1971) Carbohydr. Res. 11, 375.  119)  Bleriot, Y., Smelt, K. H , Cadefau, J., Bollen, M., Stalmans, W., Biggadike, K., Johnson,  L. N., Oikonomakos, N. G., Lane, A. L., Crook, S., Watkin, D. J., and Fleet, G. W. J. (1996) Tetrahedron Lett. 37, 7155-7158.  120)  Bleriot, Y., Masaguer, C. F., Charlwood, J., Winchester, B. G., Lane, A. L., Crook, S.,  Watkin, D. J., and Fleet, G. W. J. (1997) Tetrahedron 53, 15135-15146.  207  121)  Campbell, R. E., and Tanner, M. E. (1997) Angew. Chem. Int. Ed. Engl. 13/14, 1520-  1522. 122)  Campbell, R. E., and Tanner, M. E. (1999) J. Org. Chem. 64, 9487-9492.  123)  Reynolds, D. D., and Evans, W. L. (1955) in Organic Syntheses Vol. 3 (Horning, E.C.  Ed.), pp. 432-434, Wiley, New York. 124)  Singh, A. N., Newborn, J. S., and Raushel, F. M. (1995) Bioorg. Chem. 16, 206-214.  125)  Meynial, I., Paquet, V., Combes, E. (1995) Anal. Chem. 67, 1627-1631.  126)  Lombardo, L. (1982) Tetrahedron Lett. 23, 4293-4296.  127)  Hartman, F. C. (1970) Biochemistry 9, 1776-1782.  128)  Muller, T., and Schmidt, R. R. (1995) Angew. Chem. Int. Ed. Engl. 34, 1328-1329.  129)  Binkley, R. W. (1977) J. Org. Chem. 42, 1216-1221.  130)  Falcone-Hindley, M. L., and Davis, J. T. (1998) J. Org. Chem. 63, 5555-5561  131)  Mancuso, A. J., Huang, S. L., and Swern, D. (1978) J. Org. Chem. 43, 2480-2482.  132)  Maercker, A. (1965) Org. React. 14, 270-490.  133)  Pfitzner, K. E. and Moffat, J. G. (1963) J. Am. Chem. Soc. 85, 3027.  134)  Pfitzner, K. E. and Moffat, J. G. (1965) J. Am. Chem. Soc. 87, 5661.  135)  Pfitzner, K. E. and Moffat, J. G. (1965) J. Am. Chem. Soc. 87, 5670.  136)  Nicoll-Griffith, D., and Weiler, L. J. (1984) Chem. Soc. Chem. Commun. 659-661.  137)  Takai, K., Hotta, Y., Oshima, K., and Nozaki, H. (1978) Tetrahedron Lett. 27, 2417-  2420. 138)  Szeto, P. (1995) in Encylcopedia of Reagents for Organic synthesis, Vol. 3 (Paquette,  A. Ed.) pp 1439-1440, Wiley, Chichester. 139)  Hibino, J., Okazoe, T., Takai, K., and Nozaki, H. (1985) Tetrahedron Lett. 26, 5579.  208  140)  Okazoe, T., Hibino, J., Takai, K., and Nozaki, H. (1985) Tetrahedron Lett. 26, 5581.  141)  Lombardo, L. (1982) Tetrahedron Lett. 23, 4293-4296.  142)  Petasis, N. A., and Staszewski, J. P. (1995) in Encylcopedia of Reagents for Organic  Synthesis, Vol. 3 (Paquette, L. A. Ed.) pp 1565-1568, Wiley, Chichester. 143)  Warren, C. D., Jeanloz, R. W. (1973) Biochemistry 12, 5031-5037.  144)  MacDonald, D. L. (1972) Methods Carbohydr. Chem. 6, 389-392.  145)  Roseman, S., Distler, J. J., Moffatt, J. G., and Khorana, H. G. (1961) J. Am. Chem. Soc.  83, 659-663. 146)  Wittmann, V., and Wong, C-H. (1997) J. Org. Chem. 62, 2144-2147.  147)  Horton, D., Nakadate, M., and Tronchet, J. M. J. (1968) Carbohydr. Res. 7, 56-65.  148)  Ali, M. H., Collin, P. M., and Overend, W. G. (1990) Carbohydr. Res. 205, 428-434.  149)  Sim, M. M., Kondo, H., and Wong, C-H. (1993) J. Am. Chem. Soc. 115, 2260-2276.  150)  Fersht, A. (1985) in Enzyme Structure and Mechanism 2 ed., pp. 248-256, W. H. nd  Freeman and Company, New York, NY. 151)  Cleland, W. W. (1990) Biochemistry 29, 3194-3197.  152)  Anderson, K. S., and Johnson, K. A. (1990) J. Biol. Chem. 265, 5567-5572.  153)  Campbell, R. E., Mossiman, S. C , van de Rijn, I., Tanner, M. E., and Strynadka, N. C. J.  (2000) Biochemistry 39, 7012-7023. 154)  Ramakrishnan, V., Finch, J. T., Graziano, V., Lee, P. L., and Sweet, R. M. (1993) Nature  362, 219-223. 155)  Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) Nucleic Acids Res. 22, 4673-  4680. 156)  Hutchinson, E. G., and Thornton, J. M. (1996) Prot. Sci. 5, 212-220.  209  157)  Jones, T. A., Zou, J.-Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Cyrstallogr.  A 4 7 , 110-119. 158)  Esnouf, R. M. (1997) J. Mol. Graphics 15, 133-138.  159)  Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-325.  160)  Terwilliger, T. C , and Berendzen, J. (1997) Acta Crystallogr. D53, 571-579.  161)  De La Fortelle, E., and Bricogne, G. (1997) Methods Enzymol. 276, 472-494.  162)  Abrahams, J. P., and Leslie, A. G. W. (1996) Acta Crystallogr. D52, 30-42  163)  McRee, D. E. (1992) J. Mol. Graphics 10, 44-46.  164)  Briinger, A. T., Adams, P. D., Clore, G. M., Gros, P., Grosse-Kunstleve, R. W., Jiang, J.  S., Kuszewski, J., Nigles, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr. D54, 905-921. 165)  Briinger, A. T. (1992) X-PLOR, version 3.1, Yale University Press, New Haven, CT.  166)  Collaborative Computational Project Number 4. (1994) Acta Crystallogr. D50, 760-763.  167)  Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl.  Crystallogr. 26, 283-291. 168)  Baker, P. J., Sawa, Y., Shibata, H., Sedelnikova, S. E., and Rice, D. W. (1998) Nature  Struct. Biol. 5, 561-567. 169)  Bertrand, J. A., Auger, G., Fanchon, E., Martin, L., Blanot, D., van Heijenoort, J., and  Dideberg, O. (1997) EMBOJ. 16, 3416-3425. 170)  Jones, S., and Thorton, J. M. (1995) Prog. Biophys. Mol. Biol. 63, 31-65.  171)  Holm, L., and Sander, C. (1993) J. Mol. Biol. 233, 123-138.  172)  Arrecubieta, C , Lopez, R., and Garcia, E. (1994) J. Bacteriol. 176, 6375-6383.  173)  Thoden, J. B., Frey, P. A., and Holden, H. M. (1996) Biochemistry 35, 5137-5144.  210  174)  Paetzel, M., and Dalbey, R. E. (1997) Trends Biochem. Sci. 22, 28-31.  175)  Westheimer, F. H. (1995) Tetrahedron 51, 3-20.  176)  Neidhart, D. J., Howell, P. L., Petsko, G. A., Powers, V. M., Li, R., Kenyon, G. L., and  Gerlt, J. A. (1991) Biochemistry 30, 9264-9273. 177)  Dao-pin, S., Anderson, D. E., Baase, W. A., Dahlquist, F. W., and Mathews, B. W.  (\99\) Biochemistry 30, 11521-11529. 178)  Strynadka, N. C. J., Adachi, H , Jensen, S. E., Johns, K., Siekecki, A., Betzel, C., Sutoh,  K., and James, M. N. G. (1992) Nature 359, 700-705. 179)  Paetzel, M., Strynadka, N. C. J., Tschantz, W. R., Casareno, R., Bullinger, P. R., and  Dalbey, R. E. (1997) J. Biol. Chem. 272, 9994-10003. 180)  Paetzel, M., Dalbey, R. E., and Strynadka, N. C. J. (1998) Nature 396, 186-190.  181)  Hempel, J., Perozich, J., Chapman, T., Rose, J., Boesch, J. S., Liu, Z., Lindahl, R., and  Wang, B. (1999) in Enzymology and Molecular Biology of Carbonyl Metabolism 7 (Weiner, H , Maser, E., Crabb, D. W., and Lindahl R. Ed.) pp 53-59, Kluwer Academic/Plenum Publishers, New York, NY. 182)  Strambini, G. B., and Gonnelli, M. (1990) Biochemistry 29, 196-203.  183)  Gabellieri, E., Rahuel-Clermont, S., Branlant, G., and Strambini, G. B. (1996)  Biochemistry 35, 12549-12559. 184)  King, E. L., and Altman, C. (1956) J. Phys. Chem. 60, 1375.  185)  SIR92: Altomare, A., Cascarano, M., Giacovazzo, C , and Guagliardi, A. (1993). J. Appl.  Cryst. 26, 343.  211  186)  DIRDIF94: Beurskens, P. T., Admiraal, G., Beurskens, G., Bosnian, W. P., de Gelder, R.,  Israel, R. and Smits, J. H. M. (1994). The DIRDIF-94 Program System, Technical Report of the Crystallography Laboratory, University of Nijmegen, The Netherlands. 187)  Cromer, D. T. and Waber, J. T. (1974) "International Tables for X-ray Crystallography",  Vol. IV, The Kynoch Press, Birmingham, England, Table 2.2 A. 188)  Ibers, J. A. and Hamilton, W. C. (1964) Acta Crystallogr. 17, 781.  189)  Creagh, D. C. and McAuley, W. J. (1992) "International Tables for Crystallography",  Vol C, (A.J.C. Wilson, ed.), Kluwer Academic Publishers, Boston, Table 4.2.6.8, pp 219222. 190)  Creagh, D. C. and Hubbell, J. H. (1992) "International Tables for Crystallography", Vol  C, (A. J. C. Wilson, ed.), Kluwer Academic Publishers, Boston, Table 4.2.4.3, pp 200-206. 191)  teXsan: Crystal Structure Analysis Package, Molecular Structure Corporation (1985 and  1992).  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061448/manifest

Comment

Related Items