UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanisms and engineering of glycosyltransferases Lairson, Luke Lee 2007

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

Item Metadata

Download

Media
831-ubc_2007-318072.pdf [ 26.98MB ]
Metadata
JSON: 831-1.0059750.json
JSON-LD: 831-1.0059750-ld.json
RDF/XML (Pretty): 831-1.0059750-rdf.xml
RDF/JSON: 831-1.0059750-rdf.json
Turtle: 831-1.0059750-turtle.txt
N-Triples: 831-1.0059750-rdf-ntriples.txt
Original Record: 831-1.0059750-source.json
Full Text
831-1.0059750-fulltext.txt
Citation
831-1.0059750.ris

Full Text

MECHANISMS AND ENGINEERING OF GLYCOSYLTRANSFERASES  by LUKE LEE LAIRSON B . S c , University of Guelph, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Chemistry)  T H E UNIVERSITY O F BRITISH C O L U M B I A  October 2007 © Luke Lee Lairson, 2007  ABSTRACT In order to gain insight into the natural evolution of enzyme mechanism and to increase the utility of a class of sugar modifying enzymes known as glycosyltransferases, several representative enzymes were subjected to mechanistic and engineering studies. The chemical and kinetic mechanisms of the GT-A fold inverting sialyltransferase Cst II were investigated by detailed kinetic analysis, protein X-ray crystallography and mutagenesis. This enzyme catalyzes the general SN2-like direct displacement mechanism used by all inverting glycosyltransferases.  However, the chemical strategies utilized to  facilitate reaction are more akin to those of the GT-B fold enzymes, indicating a convergence in mechanism between these two clans of enzymes. By analogy to retaining glycosidases, retaining glycosyltransferases had been thought to use a double displacement mechanism involving an enzymatic nucleophile. However,  a comparison of the  X-ray  crystal structures of multiple retaining  glycosyltransferases indicates a complete lack of conserved structural architecture in the region that would be occupied by this critical catalytic residue. This lack of conserved architecture, precedence for cationic enzymatic mechanisms, and the inherent differences in reactivities of glycosyltransferase and glycosidase substrates all support a notion that the majority of retaining glycosyltransferases utilize a mechanism involving the formation of a short-lived ion pair intermediate species. However, a protein engineering approach was used to explore the possibility of nucleophilic catalysis in the retaining galactosyltransferase LgtC. The results of this work led to the first direct observation of a catalytically  relevant  covalent  glycosyl-enzyme  intermediate  for  a  retaining  glycosyltransferase.  ii  It was demonstrated that catalytically active LgtC could be displayed on the surface of M l 3 bacteriophage as a pill fusion protein. Further, LgtC phage display was successfully performed in the context of a water-in-oil emulsification procedure that may have significant utility in the development of directed evolution screening approaches. A substrate engineering strategy was developed which allowed the substrate specificity of wild type LgtC to be broadened to allow exclusive formation of a-1,2, a-1,3 or a-1,4 linkages at synthetically useful rates with various alternative acceptor substrates. Finally, it was demonstrated that Cst II and LgtC will utilize alternative donor substrates in the presence of their respective natural nucleotide products.  iii  TABLE OF CONTENTS Abstract  ii  Table of Contents  iv  List of Tables  xii  List of Figures  xiii  List of Schemes  xxi  List of Abbreviations  xxii  Acknowledgements  xxv  Co-Authorship Statement  xxvii  C h a p t e r 1. I n t r o d u c t i o n  1  1.1. Carbohydrates in Biology  2  1.2. Biological Glycosyl Group Transfer Reactions  4  1.3. Glycosyltransferases  6  1.3.1. Glycosyltransferase Activities  6  1.3.2. Glycosyltransferase Folds  7  1.3.3. Glycosyltransferase Classification  14  1.3.4. Inverting Glycosyltransferases  17  1.3.4.1. Mechanism of Inverting Glycosidases  17  1.3.4.2. Mechanism of Inverting Glycosyltransferases  18  1.3.4.3. Inverting GT-A Glycosyltransferases (Clan I)  18  1.3.4.4. Inverting GT-B Glycosyltransferases (Clan II)  22  1.3.5. Retaining Glycosyltransferases 1.3.5.1.Mechanism of Retaining Glycosidases  27 27 iv  1.3.5.2. Mechanism of Retaining Glycosyltransferases  30  1.3.5.3. Challenges in Studying the Mechanisms of Retaining Glycosyltransferases  32  1.3.5.4. Retaining G T - A Glycosyltransferases (Clan III)  33  1.3.5.5. Retaining GT-B Glycosyltransferases (Clan IV)  43  1.3.5.6. A n Alternative "SNi-Like" Mechanism  50  1.4. Aims of the Thesis  53  C h a p t e r 2. S t r u c t u r a l a n d M e c h a n i s t i c Investigationso f a n I n v e r t i n g Glycosyltransferase  56  2.1. Summary  57  2.2. Background  57  2.3. Determining the Kinetic Mechanism of the Bifunctional Inverting a-(2,3/8) Sialyltransferase Cst IIOH4384  61  2.3.1. Preliminary Kinetic Analysis  61  2.3.2. Potential Kinetic Mechanisms  63  2.3.3. Distinguishing Between a Sequential or Ping Pong Mechanism  65  2.3.4. Product Inhibition Studies with Cst II  67  2.3.5. Synthesis and Inhibition Studies of a Dead-End Donor Substrate Analogue  71  2.3.6. Conclusions on the Kinetic Mechanism of Cst II  73  2.3.7. Calculating Kinetic Parameters for Cst II  76  2.4. Three-Dimensional X-Ray Crystallographic Analysis of Cst II in Complex with a Stable Donor Substrate Analogue 2.5. Investigating the Role of Putative Catalytic Residues 2.5.1. Tyrl56 and Tyrl62 Facilitate C M P Departure  78 81 82  2.5.2. Hisl88 Acts as the Catalytic Base  84  2.6. Concluding Remarks  90  Chapter 3. Structural and Mechanistic Investigations with Retaining Glycosyltransferases  93  3.1. Summary  94  3.2. Investigating the Catalytic Mechanism of the Retaining a-(l ,4) Galactosyltransferase LgtC  94  3.2.1. Background  94  3.2.2. Engineering the Mechanism of LgtC  98  3.2.3. Kinetic Analysis of LgtC Q189E  99  3.2.4. Intermediate Trapping by ESI-MS  100  3.2.5. Identification of the Site of Labelling  104  3.2.6. Analysis of the D190N Variant of LgtC  107  3.2.7. Crystallographic Analysis of LgtC Q189E  108  3.2.8. Attempts to Accumulate a Covalent Intermediate using Alternative Active Site Modifications  109  3.2.9. Interpretation of the LgtC Q l 89E Labelling Results  Ill  3.3. Investigating the Retaining a-( 1,2) Mannosyltransferase Kre2  115  3.3.1. Background  115  3.3.2. Synthesis of a Non-Reactive Donor Substrate Analogue  116  3.3.3. Kinetic Analysis of Kre2  117  3.3.4. Attempts to Observe a Covalent Glycosyl-Enzyme Intermediate withKre2  117  3.3.5. X-Ray Crystal Structure of Kre2 with Bound GDP 2F Mannose  118  3.4. ESI-MS Analysis of the Retaining cc-(l,3) Galactosyltransferase a3GalT  122  3.4.1. Background  122  3.4.2. Attempts to Observe a Covalent Glycosyl-Enzyme Intermediate with a3GalT  124  3.5. Reasons Why Retaining Glycosyltransferases May Have Evolved a Mechanism Not Involving Nucleophilic Catalysis  126  3.6. Concluding Remarks  135  C h a p t e r 4. Glycosyltransferase Protein E n g i n e e r i n g  142  4.1. Summary  143  4.2. Engineering the Nucleotide Specificity of LgtC  144  4.2.1. Background  144  4.2.2. Mutation of a Conserved Active Site Side Chain Alters the Base Specificity of GTP Binding Proteins  147  4.2.3. Identification and Modification of a Base Specificity Conferring Site Within the LgtC Active Site 4.2.4. Synthesis of CDP Gal  151 152  4.2.5. Testing the Nucleotide Sugar Specificity of Wild Type and N10D LgtC 4.2.6. Concluding Remarks 4.3. Compartmentalized Phage Display of LgtC  153 154 156  4.3.1. Background  156  4.3.2. Phage Display of LgtC as an M13 pill Fusion Protein  163  4.3.3. Compartmentalized LgtC Phage Display  168  4.3.4. A High Throughput Screen for LgtC Directed Evolution Based on Compartmentalized Phage Display  170  4.3.5. Synthesis of a Biotinylated Lactose Conjugate Acceptor Substrate  175  4.3.6. Generating a Library of LgtC Variants  176  4.3.7. Immobilization of the Carbohydrate Binding Domain of Shiga-Like Toxin SLT-IB  177  4.3.8. Multivalent Phage Display of Streptavidin as a pVIII fusion protein ... 179 4.3.9. Simultaneous M13 Phage Display of LgtC and Streptavidin  181  4.3.10. Conclusions and Future Directions  183  C h a p t e r 5. Glycosyltransferase Substrate E n g i n e e r i n g  187  5.1. Siirnmary  188  5.2. Engineering Alternative Acceptor Substrates for Glycosyltransferases  188  5.2.1. Background  188  5.2.2. Exploring the Acceptor Substrate Specificity of LgtC  190  5.2.3. Determining the Ability of LgtC to Transfer to Equatorial Hydroxyl Substituents  192  5.2.4. Exploring the Synthetic Utility of LgtC Catalyzed Galactosyl Transfer to Alternative Monosaccharide Acceptor Substrates  193  5.2.5. A Substrate Engineering Strategy the Uses Aromatic Substituents as Active Site Anchors  194  5.2.6. Using Substrate Engineering to Control Regio-Seletive a3GalT Catalyzed Galactosyl Transfer to a Monosaccharide Acceptor Susbstrate  196  5.2.7. Using Substrate Engineering to Facilitate Cst II Catalyzed Sialyl Transfer to a Monosaccharide Acceptor  198  5.2.8. The Nature of the Substituent on an Alternative Acceptor Substrate can Affect the Regio-Selectivity of LgtC  198  5.2.9. Controlling the Regio-Selectivity of LgtC-Catalyzed Galactosyl Transfer to Mannose  202  viii  5.2.10. Controlling the Regio-Selectivity of LgtC-Catalyzed Galactosyl Transfer to Xylose  205  5.2.11. Concluding Remarks  206  5.3. Engineering Alternative Donor Substrates for Glycosyltransferases 5.3.1. Background  209 209  5.3.2. A n Activated Alternative Donor Substrate for Cst II that Mimics the Product Binding Mode  209  5.3.3. A n Activated Alternative Donor Substrate for LgtC  213  5.3.3.1. Testing a-Linked Aryl Galactosides  213  5.3.3.2. Testing p-Linked Aryl Galactosides  214  5.3.3.3. Mechanistic Implications for LgtC  216  5.3.4. Concluding Remarks  217  C h a p t e r 6. S u m m a r y a n d F u t u r e D i r e c t i o n s  219  C h a p t e r 7. M e t h o d s  226  7.1. General  227  7.2. Chapter 2 Experimental  227  7.2.1. Acknowledgements and Generous Gifts  227  7.2.2. Expression and Purification of Cst II  228  7.2.3. A U V Spectroscopy Based Enzyme Coupled Continuous Assay for Cst II Kinetic Characterization  228  7.2.4. A Capillary Electrophoresis Based Stopped Assay for Cst II Kinetic Characterization  229  7.2.5. Synthesis of C M P 3F NeuAc  230  7.2.6. Generation of Site Directed Mutants  231  7.2.7. Chemical Rescue of the H188A Mutant of Cst II  232  7.3. Chapter 3 Experimental  232  7.3.1. Acknowledgements and Generous Gifts  232  7.3.2. Expression and Purification of LgtC-25  233  7.3.3. Expression of LgtC-25 as a C-terminal 6-Histidine Fusion  234  7.3.4. A U V Spectroscopy Based Enzyme Coupled Continuous Assay for LgtC and Kre2 Kinetic Characterization  235  7.3.5. ESI-MS Analysis of LgtC-25 Q189E  236  6.3.5.1. ESI-MS Conditions  236  6.3.5.2. Labelling of LgtC-25 Q189E  237  6.3.5.3. Turn-over of Covalently Labelled LgtC-25 Q189E  238  7.3.6. Generation of the I143D, N153D, Q185D, and D190N Mutants of LgtC-25  238  7.3.7. Synthesis of GDP 2F Gal  239  7.3.8. ESI-MS Analysis of Kre2  240  7.3.9. Expression and Purification of a3GalT as a MalE Fusion Protein  241  7.3.10. ESI-MS Analysis of a3GalT  242  7.4. Chapter 4 Experimental  242  7.4.1. Acknowledgements and Generous Gifts  242  7.4.2. Synthesis of CDP Gal  243  7.4.3. Generation of the N10D Mutant of LgtC-25  244  7.4.4. Kinetic Analysis of CDP Gal as a Substrate for LgtC-25 and 25 N10D  LgtC244  7.4.5. Subcloning LgtC-25 for Expression as an M13 pill Fusion Protein .... 245 7.4.6. VCSM13 Helper Phage Production  246  X  7.4.7. Small Scale PEG Precipitated LgtC-25 Displaying Phage Preparation  247  7.4.8. Quantification of Phage by U V Spectroscopy  248  7.4.9. Western Blot Analysis of pLLOl and pSJF6 Derived Phage  248  7.4.10. Compartmentalized Phage Display Procedure  249  7.4.11. Testing Infectivity of Emulsified LgtC-25 displaying Phage  249  7.4.12. Synthesis of a Biotin Lactose Conjugate  250  7.4.13. Generation of a Library of LgtC-25 Variants by Error Prone P C R  251  7.4.14. Subcloning Verotoxin for Expression as a C-Terminal 6-Histidine Fusion Protein  252  7.4.15. Subcloning of Streptavidin-pVIII Fusion Protein  254  7.4.16. ELISA Analysis of pSAV and pLL34 Derived Phage  255  7.5. Chapter 5,Experimental 7.5.1. Acknowledgements and Generous Gifts  256 256  7.5.2. LgtC-Catalyzed Galactosylation of Modified Alternative Acceptor Substrates  256  7.5.3. N M R Analysis of LgtC Products  257  7.5.4. a3GalT-Catalyzed Galactosylation of Alternative Acceptor Substrates  261  7.5.5. Cst II-Catalyzed Sialylation of Alternative Acceptor Substrates  262  7.5.6. Testing Alternative Cst II Donor Substrates  263  7.5.7. Testing Alternative LgtC Donor Substrates  264  References  266  Appendix A - I U P A C Recommended Nomenclature for Reaction Mechanisms  296  A p p e n d i x B - Graphical Representation of Kinetic Data  297  LIST OF TABLES T a b l e 2.1.  Apparent Michaelis-Menten kinetic parameters for the transferase and  hydrolase activities of Cst II  61  T a b l e 2.2.  Expected product inhibition patterns for select B i B i enzyme mechanisms . 68  T a b l e 2.3.  Michaelis-Menten kinetic parameters for Cst II catalyzed sialyl transfer to  3'-sialyl lactose  T a b l e 2.4.  Apparent Michaelis-Menten parameters for C M P NeuAc with various Cst II  mutants  T a b l e 3.1.  78  84  Michaelis-Menten parameters for the transferase and hydrolase activities of  WT and various mutants of LgtC-25 with respect to UDP Gal and UDP Glc donor substrates measured at pH 7.5, 37°C  T a b l e 4.1.  Common forms of nucleotide sugar donor substrates found in Nature and  utilized by glycosyltransferases  T a b l e 4.2.  100  144  Michaelis-Menten kinetic parameters for wild type and N10D versions of  LgtC using either CDP Gal or UDP Gal as the varied substrate at a fixed saturating concentration of lactose acceptor (160mM)  T a b l e 5.1.  Kinetic parameters for various LgtC acceptor substrates  153  191  xii  LIST OF FIGURES F i g u r e 1.1.  Overall folds observed for glycosyltransferase enzymes. The G T - A fold (A)  is represented by the inverting enzyme SpsA from Bacillus subtilus (pdb lqgq) and that of the GT-B fold (B) by bacteriophage T4 (i-glucosyltransferase  F i g u r e 1.2.  10  Like glycosidases, glycosyltransferases catalyze glycosyl group transfer with  either inversion or retention of the anomeric stereochemistry with respect to the donor sugar  14  F i g u r e 1.3.  Figure  1.4.  Glycosyltransferase (GT) classification system  A direct displacement S ^ - l i k e reaction results in inverted anomeric  configuration via a single oxocarbenium ion-like transition state  Figure  1.5.  Figure  17  Comparison of the active sites of several metal-dependent inverting  glycosyltransferases from Clan I  F i g u r e 1.6.  16  Active site of wild type T4 bacteriophage BGT from family GT63  20  24  1.7. The double displacement mechanism established for retaining glycosidases  proceeds via two oxocarbenium ion-like transition states with the intermediate formation of a discrete covalently bound glycosyl-enzyme species, resulting in overall retention of anomeric configuration  F i g u r e 1.8.  Proposed double displacement mechanism for retaining  Glycosyltransferases  F i g u r e 1.9.  Clan III  29  31  Comparison of the active sites of several retaining glycosyltransferases from 35  Xlll  Figure  Comparison of the active sites of several retaining glycosyltransferases  1.10.  from Clan IV  Figure  47  The  1.11.  SNI  mechanism for the decomposition of alkyl chlorosulphites is a  special case of an S N I process that leads to complete retention of stereochemistry in the product  Figure  51  Proposed "SNi-like" mechanism for retaining glycosyltransferases and  1.12.  glycogen phosphorylase involving a direct "front-side SN2-like" displacement proceeding through a highly dissociative oxocarbenium ion-like transition state resulting in retention of anomeric configuration  Figure  52  Cleland notation schematic representations of ping pong (A), Random  2.1.  sequential (B), ordered sequential (C), and Theorell-Chance (D) B i B i kinetic mechanisms  F i g u r e 2.2.  S  64  Double reciprocal and secondary plots of initial rate data with respect to  C M P NeuAc at various constant concentrations of 3'-sialyl lactose (3'-SL)  67  F i g u r e 2.3.  Representative C E data used to monitor Cst II sialyltransferase activity ... 6 9  F i g u r e 2.4.  Inhibition of Cst II catalyzed sialyl transfer to F C H A S E 3 ' - S L by the product  C M P with respect to donor (A) and acceptor (B) substrates  Figure  70  Inhibition of Cst II catalyzed sialyl transfer to 3 ' - S L by the dead end  2.5.  substrate analogue C M P 3 F NeuAc with respect to donor (A) and acceptor (B) Substrates  F i g u r e 2.6.  72  Cleland notation schematic representation of the steady state iso ordered B i  B i mechanism  75  xiv  Overall structure of Cst II. Top down (A) and side (B) views of the overall  F i g u r e 2.7.  tetrameric arrangement of Cst II  80  Cst II active site highlighting the hydrogen bond interactions between the  F i g u r e 2.8.  side chain hydroxyls of Y156 and Y162 and the departing C M P leaving group  83  Cst II active site highlighting the side chain imidazole of Hisl88 suitably  F i g u r e 2.9.  positioned on the a-face of the donor substrate to play the role of base catalyst  F i g u r e 2.10.  pH dependence of k /K cat  m  for Cst II. The pKa value was determined by best  fit of the data to a single titration model in Grafit 4.0  Figure 2.11.  86  87  Chemical rescue of nucleophile (A) and acid/base (B) mutant glycosidases  with small molecules  88  F i g u r e 2.12.  Chemical rescue of Cst II H188A  89  Figure 3.1.  LgtC active site with bound substrate analogues UDP 2F Gal and 4'deoxy  lactose highlighting the proximity of the side chain amide of Gin 189 and the 6'-hydroxyl of lactose to the anomeric reaction centre  F i g u r e 3.2.  Mechanism of retaining P-hexosaminidases involving anchimeric assistance  leading to the formation of a stabilized oxazolinium ion intermediate  F i g u r e 3.3.  ESI-MS analysis of labelled and unlabelled species of LgtC Q189E  F i g u r e 3.4.  Decay data for the turn over of the glycosyl-enzyme intermediate species  resulting from the incubation of UDP Glc with LgtC Q189E  F i g u r e 3.5.  96  ESI-MS analysis of the fragmented, purified Gal-labelled peptide  96  102  103  106  XV  F i g u r e 3.6.  Superimposition of the wild type and Q l 89E mutant active sites of  LgtC  Figure  108  3.7. LgtC active site highlighting the side chains of residues within close  proximity to that of Asp 190  F i g u r e 3.8.  Ill  Acyl migration of the galactosyl moiety from the side chain of Q189 to that  of D190 is difficult to rationalize  F i g u r e 3.9.  112  Kre2 active site with bound intact donor substrate analogue GDP 2F  Man  F i g u r e 3.10.  119  Active site of a3GalT with bound non-reactive donor substrate analogue  UDP 2F Glc highlighting the position of Glu317 suitably positioned to play the role of a catalytic nucleophile in a double displacement mechanism  F i g u r e 3.11.  Comparison of the potential differences in the free energy barriers for  124  S N I -  like and SN2-like pathways that could be utilized by (A) retaining glycosidases (GH) and (B) retaining glycosyltransferases (GT)  F i g u r e 3.12.  127  Chemical similarities between the substrate used in the first step of the  glycosyltransferase reaction and the covalent glycosyl-enzyme intermediate turned over in the second step of a (3-retaining glycosidase catalyzed reaction  F i g u r e 3.13.  The reactions catalysed by FPP synthetase  F i g u r e 3.14.  Proposed  glycosyltransferases  DN* A  N S S  J  p  129  133  nucleophilic substitution reaction for retaining 141  xvi  Figure  Hydrogen bonding interactions between guanosine and a side chain  4,1.  carboxylate (A) or xanthosine and a side chain amide (B) determines the base specificity ofEF-TuandFtsY  Figure  148  Hydrogen bonding interactions between guanosine and a side chain  4.2.  carboxylate that are found to control base specificity between guanosine and xanthosine in the case of EF-Tu (A) and FtsY (B)  149  Hydrogen bonding interactions between uridine and a side chain amide (A)  F i g u r e 4.3.  or cytidine and a side chain carboxylate (B) proposed to determine base specificity ... 150  Figure  Active sites of the GT family 8 members glycogenin (A) and LgtC (B)  4.4.  revealing the conserved interaction of a side chain amide with the 4-carbonyl group of the bound uridine ring  Figure  Graphical representation  4.5.  151  of the macromolecular anatomy of an F f  filamentous phage particle  F i g u r e 4.6.  157  Schematic representations of particles resulting from various phage display  systems available for the surface display of foreign peptides or entire proteins  Figure  159  4.7. Recombinant construct within the M C S of pLLOl encoding an LgtC-pIII  fusion protein F i g u r e 4.8.  165  T L C analysis (7:2:1 EtOAc:MeOH:H 0) of the LgtC activity of twice P E G 2  precipitated M l 3 bacteriophage derived from pLLOl phagemid  Figure 4.9.  detection  166  Western blot analysis of various phage particles using anti-c-myc-HRP for 167  xvii  T L C analysis (7:2:1 EtOAc:MeOH:H 0) of the LgtC activity of emulsified  Figure 4.10.  2  M l 3 bacteriophage derived from pLLOl phagemid  169  Proposed screening procedure for the in vitro evolution of LgtC involving  Figure 4.11.  compartmentalized phage display  F i g u r e 4. 12.  174  Three-dimensional x-ray crystal structure of Shiga-like toxin I B subunit  with bound starfish molecule  178  F i g u r e 4.13.  Phage ELISA for S A V display  181  F i g u r e 4.14.  Phage ELISA for S A V display  182  F i g u r e 5.1.  ' H - C O S Y N M R spectrum of /?aro-Nitrophenyl (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l—>4)-2,3,6-tri-0-acetyl-p-D-glucoside  193  Acceptor binding site of LgtC containing a hydrophobic platform (Phel32)  F i g u r e 5.2.  for binding the glucose sugar ring of lactose or an aromatic substituent  F i g u r e 5.3.  1  H-COSY N M R spectrum of Benzyl (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l->3)-2,4-di-0-acetyl-P-D-xyloside  F i g u r e 5.4.  5.5.  'H-COSY  N M R spectrum  of n-Octyl  galactopyranosyl)-(1^3)-2,4,6-tri-0-acetyl-P-D-glucoside  Figure  5.6.  196  ' H - C O S Y N M R spectrum of /?ara-Nitrophenyl (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l-^3)-2,4,6-tri-0-acetyl-P-D-galactoside  Figure  195  'H-COSY  N M R spectrum  of «-Octyl  galactopyranosyl)-(l-^4)-2,3,6-tri-0-acetyl-P-D-galactoside  197  (2,3,4,6-tetra-O-acetyl-a-D199  (2,3,4,6-tetra-O-acetyl-a-D200  xviii  Figure 5.7.  'H-COSY  N M R spectrum  of n-Octyl  (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l-»3)-2,4-di-0-acetyl-P-D-xyloside  201  Figure 5.8. A model for the substrate engineering strategy using 6-OBz Man  202  Figure 5.9. Overlay of the M M 2 minimized structure of 6-OBz Man (green) with the structure of 4-deoxylactose (orange) bound within the active site of LgtC  Figure  5.10.  'H-COSY  NMR  spectrum  203  of (2,3,4,6-Tetra-O-acetyl-a-D-  galactopyranosyl)-(l-^2)-l,3,4-tri-0-acetyl-6-0-benzoyl-a-D-mannose  204  Figure 5.11. A model for the substrate engineering strategy using 4-O-Benzoyl-D-xylose (4-OBz Xyl)  Figure  5.12.  205  'H-COSY  NMR  spectrum  of  (2,3,4,6-Tetra-O-acetyl-a-D-  galactopyranosyl)-( 1 -»2)-1,3 ,-di-0-acetyl-4-0-benzoyl-a/p-D-xylose  206  Figure 5.13. Substrate engineering strategy applied to various alternative LgtC acceptor substrates  208  Figure 5.14. Representation of the product binding mode of Cst II (A) and the proposed mode of binding of a-linked pNP NeuAc that would facilitate in situ generation of the natural donor substrate  Figure 5.15.  211  (A) Monitoring Cst II catalyzed glycosylation of fluorescent fluorescein  lactose conjugate acceptor substrate using /7-nitrophenyl a-sialoside as an alternative donor substrate in the presence or absence of C M P by T L C (4:2:1:0.1 EtOAc: MeOH: H2O: AcOH). (B) Reaction catalysed by Cst II using the alternative donor-substrate pnitrophenyl a-sialoside  212  xix  Figure 5.16.  (A) Monitoring LgtC catalyzed glycosylation of fluorescent fluorescein  lactose conjugate  acceptor substrate using 2,4-dinitrophenyl (3-galactoside as an  alternative donor substrate in the presence or absence of UDP by TLC (7:2:1:0.1 EtOAc: MeOH: H 0 : AcOH). 2  (B) Reaction catalysed by LgtC using the alternative donor  substrate 2,4-dinitrophenyl (3-galactoside  216  ' Figure 7.1. T L C analysis of a-2,3 specific neuraminidase treatment of glycosylation product derived from Cst II using pNP P-galactoside as an alternative acceptor substrate  263  Figure 7.2. T L C analysis of a-2,3 specific neuraminidase treatment of glycosylation product derived from Cst II using /?-nitrophenyl a-sialoside as an alternative donor substrate in the presence C M P monitored by TLC  264  Figure A . l . Michaelis-Menten plots of various Cst II activities  297  Figure A.2. Michaelis-Menten plots of wild type LgtC activity with various acceptor substrates  298  XX  LIST OF SCHEMES S c h e m e 1.1.  Overall reactions catalyzed by (a) glycosidases, (b) glycosyltransferases, (c)  phosphorylases, and (d) polysaccharide lyases  Scheme  2.1.  Monofunctional and bifunctional sialyltransferase activities of Cst  enzymes  Scheme 2.2.  Enzyme coupled continuous assay used to measure the kinetic parameters  S c h e m e 3.2.  71  Galactosylation reaction catalyzed by LgtC  S c h e m e 3.4.  95  Equilibrium for LgtC catalyzed hydrolysis of UDP Gal in the presence of  pyruvate kinase (PK) and phosphoenolpyruvate (PEP)  S c h e m e 3.3.  62  Enzymatic synthesis of the dead end donor substrate analogue C M P 3F  NeuAc  S c h e m e 3.1.  II  60  of sialyltransferases  Scheme 2.3.  5  Mannosylation reactions catalyzed by Kre2  104  116  Synthesis of the non-reactive donor substrate analogue GDP 2F Mannose  via a morpholidate coupling reaction  116  Scheme 3.5.  Galactosylation reaction catalyzed by a3GalT  122  S c h e m e 4.1.  Chemical synthesis of the desired unnatural nucleotide sugar CDP G a l . 152  S c h e m e 4.2.  Synthesis of a biotin lactose conjugate acceptor substrate for LgtC  175  XXI  LIST OF ABBREVIATIONS Ab  antibody  ABTS  2,2'-azino-bis[3-ethylbenziazoline-6-sulphonic acid]  ADP  adenosine 5'-diphosphate  ATP  adenosine 5'-triphosphate  BSA  bovine serum albumin  CDP  cytidine 5'-diphosphate  CDP Gal  cytidine 5'-diphosphogalactose  CE  capillary electrophoresis  CMP  cytidine 5'-monophosphate  C M P NeuAc  cytidine 5'-monophospho-N-acetylneuraminic acid  CTP  cytidine 5'-triphosphate  DCC  dicyclohexylcarbodiimide  DMF  dimethylformamide  DNA  deoxyribonucleic acid  dNTP  deoxynucleoside triphosphates  DTT  dithiothreitol  dNP  2,4-dinitrophenyl  EDTA  ethylenediaminetetraacetatic acid  ELISA  enzyme-linked immunosorbent assay  Et N  triethylamine  ESI-MS  electrospray ionization mass spectrometry  EtOAc  ethyl acetate  FCHASE  6-(5-fluorosceincarboxamido)-hexanoic acid succimidyl ester  GAG  glycosaminoglycan  Gal  galactose  GalNAc  N-acetyl-D-galactosamine  GDP  guanosine 5'-diphosphate  GDP Man  guanosine 5'-diphosphomannose  3  GDP 2F Man  guanosine 5 -diphospho-(2"-deoxy-2"fluoro)-a-D-mannose  Glc  glucose  GlcA  glucuronic acid  GlcNAc  N-acetyl-D-glucosamine  Glc IP  a-D-glucose 1-phosphate  GMP  guanosine 5'monophosphate  HEPES  4-(2-hydroxyethyl)piperazine-1 -ethanesulphonic acid  hGH  human growth hormone  HPLC  high performance liquid chromatography  HR  high resolution  HRP  horse radish peroxidase  ICAT  isotope-coded affinity tags  IMAC  immobilized metal affinity chromatography  IPTG  isopropyl (3-D-thiogalactopyranoside  IUBMB  International Union of Biochemistry and Molecular Biology  IUPAC  International Union of Pure and Applied Chemists  KIE  kinetic isotope effect  LDH  lactate dehydrogenase  ManNAc  N-acetyl-D-mannosamine  MeOH  methanol  MCS  multiple cloning site  NAD  +  P-nicotinamide adenine dinucleotide  NADH  p-nicotinamide adenine dinucleotide, reduced  NDP  nucleoside diphosphate  NMPK  nucleoside monophosphate kinase  NMR  nuclear magnetic resonance  60BzMan  6-O-benzoyl mannose  40Bz Xyl  4-O-benzoyl xylose  PBS  phosphate buffered saline  PEG  polyethylene glycol  PEP  phospho(enol)pyruvic acid  PK  pyruvate kinase  PLP  pyridoxal phosphate  PNP  joara-aminophenyl  RBS  ribosome binding site  RNA  ribonucleic acid  SAV  streptavidin  SDS  sodium dodecyl sulphate  SDS P A G E  sodium dodecyl sulphate polyacrylamide gel electrophoresis  SL  3'-sialyllactose  SSL  3'8"-disialyllactose  TDP  thymidine 5'-diphosphate  TRIS  2-amino-2-(hydroxymethyl)-1,3-propandiol  Tween  polyethylene glycol sorbitan monolaurate  UDP  uridine 5'-diphosphate  UDP Gal  uridine 5'-diphosphogalactose  UDP 2F Gal  uridine 5'-diphospho-(2"-deoxy-2"fluoro)-a-D-galactose  UDP Glc  uridine 5'-diphosphoglucose  UDP 2F Glc  uridine 5'-diphospho-(2"-deoxy-2"fluoro)-a-D-glucose  UDP Glc A  uridine 5'-diphosphoglucuronic acid  UMP  uridine 5'-monophosphate  xxiv  ACKNOWLEDGEMENTS First and foremost I would like to thank Professor Stephen G. Withers for his support during the course of my doctoral studies at U B C . He has provided me not only with tremendous scientific and personal insights, but also with a research environment that is limited only by one's creativity, scientific rigour and self-motivation.  I would  especially like to thank him for allowing me to pursue multiple research projects.  The  diversity of these projects has facilitated the development of skills in a broad range of research techniques and a foundation firmly rooted in the logic of chemistry yet not limited by a classification of academic discipline. I would also like to thank Dr. Warren W. Wakarchuk at the National Research Centre in Ottawa for providing me with clones for multiple enzymes and for allowing me to work under his direct supervision in order to develop molecular biology research skills. In addition, Professor Natalie C.J. Strynadka has provided me with extraordinary insight into the world of protein X-ray crystallography and has shown constant interest in my research with glycosyltransferases.  Prof. Tony Warren has provided me with laboratory  space and with tremendous insights into molecular biology and microbiology techniques. The incredible knowledge base of members of the Withers laboratory has played a critical role in my research. I would especially like to thank Dr. Chris Tarling, Dr. Andrew Watts, Dr. Amir Aharoni, Dr. James MacDonald, and Dr. Omid Hekmat. I must also thank my former research mentors at the University of Guelph for fostering my interest in research science, especially Prof. Bernard Grodzinski and Prof. Adrian Schwan. M y family (Kathy, Lloyd and Jaden Lairson) has always supported me, for which I am truly  XXV  grateful. I would also like to thank my friends, especially Huxley, for their support and well needed intermittent reality checks during the course of my doctoral studies. Finally, I am truly grateful for the financial support of the Natural Science and Engineering Research Council of Canada (NSERC) and the Michael Smith Foundation for Health Research (MSFHR).  xxvi  Co-Authorship Statement The research described in this thesis was identified and designed by Luke L . Lairson under the guidance of Prof. Stephen G. Withers. Unless explicitly stated, Luke L . Lairson performed all of the research and data analyses described in this thesis. Luke L . Lairson wrote the entirety of this thesis with critical readings and suggestions provided by Prof. Stephen G. Withers. Modified versions of portions of previously published text that appear in this thesis were both originally written and subsequently modified by Luke L . Lairson.  xxvii  CHAPTER 1  INTRODUCTION  * A version of portions of this chapter has been published: Lairson, L . L . and Withers, S.G. (2004) Mechanistic Analogies Amongst Carbohydrate Modifying Enzymes. Chemical Communications. (20): 2243. - Reproduced by permission of The Royal Society of Chemistry (RSC)  1  1.1. Carbohydrates in Biology The central dogma of molecular biology consists of a flow of biological information in a template-controlled fashion from D N A to R N A to protein.  The  development of this paradigm has served to revolutionize both the modern view of life and the ways by which a deep understanding of biology are explored. A n appreciation of these precise template-controlled mechanisms has allowed the development of tools that can be used to manipulate one class of molecules based upon an understanding of the other. Examples include recombinant D N A technologies using restriction enzymes and the expression of designer proteins using engineered D N A . However, a more complete view of life at the molecular level incorporates the roles of other major classes of biological molecules including glycans, lipids and small molecules. These classes are not derived in a template-controlled fashion, making their in vitro synthesis a more formidable task.  This has resulted in a relative lag in the progression of our  understanding of their roles in biology compared to that of D N A and proteins. In terms of sheer biomass, carbohydrates are in fact the largest class of biological macromolecules.  Carbohydrates are traditionally known for their roles in providing  structural platforms and in serving as energy storage devices. Cellulose in the cell walls of plants, chitin in the exoskeletons of insects, and glycosaminoglycans (GAGs) in the extracellular matrices of various mammalian tissues are important examples of critical structural roles that carbohydrates play in biology. The base metabolic currencies used to generate and store energy are sugars, with starch and glycogen serving as the reservoirs used by plants and animals respectively to store the initial input of glycolysis (glucose). A coming together of biochemistry, synthetic carbohydrate chemistry, and enzymology  2  led to development of an interdisciplinary field of study that was formally termed Glycobiology in 1988 (Rademacher et al., 1988). From this, a more recent understanding of the important roles that carbohydrates play in molecular recognition events dictating a variety of critical biological phenomena has arisen. Because of the range of available monosaccharide building blocks and the various manners by which individual units can be connected via alternative glycosidic linkages, glycan structures can achieve a level of structural diversity that exceeds that of D N A and even proteins. Whereas both D N A and proteins are linear polymers with a single type of linkage, the individual units of oligosaccharides can in theory be connected by either an a- or P-linkage to any of several positions of another sugar. This point is illustrated by considering the number of variations possible from the combination of three different amino acids, nucleotide bases, or monosaccharides in a peptide, D N A , or oligosaccharide polymer respectively. In the case of both the peptide and D N A only six variations are possible.  In contrast, three hexoses could produce up to 27,648 unique trisaccharide  structures (Laine, 1994). Nature has taken advantage of the potential of this diverse structural space by using glycans as information display systems that are recognized by specific proteins (lectins, microbial carbohydrate binding proteins, GAG-binding proteins, etc.).  These protein-glycan interactions, which occur in the nucleus and  cytoplasm and on both inter- and extracellular membrane surfaces, are involved in processes ranging from cell-cell recoginition and signal transduction to the entry of pathogens into host cells (Rademacher et al., 1988).  3  1.2.  Biological Glycosyl Group Transfer Reactions The enormous complexity of the various oligosaccharide structures found in  nature is derived from a rational orchestration of the enzymatic formation and breakdown of glycosidic linkages achieved by glycosyltransferases, glycosidases, phosphorylases, and polysaccharide lyases (Scheme 1.1). The importance of these classes of enzymes is illustrated by the fact that -1-3% of the proteins encoded by most organisms are those of enzymes that catalyse the synthesis and breakdown of glycosidic bonds (Davies et al., 2005). With the exception of the polysaccharide lyases that utilize mechanisms involving /^-elimination (Ernst et al., 1995; Lee et a l , 2002a; Rye and Withers, 2002) and a recently described mechanism involving transient remote redox and /^elimination steps utilized by a unique family of glycosidases (Yip et al., 2004), these classes of enzymes are generally believed to utilize mechanisms involving straightforward direct nucleophilic substitution at the anomeric reaction centre to catalyse glycosyl group transfer from a donor substrate to an acceptor.  Enzyme catalysed glycosyl group transfer reactions proceed through  transition states similar to those for non-enzymatic nucleophilic substitution reactions at acetal centres. Such transition states have considerable oxocarbenium ion-like character in which both the incoming nucleophile and departing leaving group interact weakly with a reaction centre possessing a significant degree of positive charge that is stabilized by the lone pair of electrons donated by an endocyclic oxygen atom.  4  a  Scheme 1.1.  Overall reactions catalysed by (a) glycosidases, (b) glycosyltransferases, (c)  phosphorylases, and (d) polysaccharide lyases.  5  While the mechanistic strategies used by glycosidases to catalyse glycosidic bond hydrolysis are fairly well understood on both a structural and chemical level (Davies, 1998); (Zechel and Withers, 2000), the characterization and mechanistic understanding of the glycosyltransferases responsible for glycoside bond formation has lagged far behind. In the case of glycosidases, a glycosyl moiety is transferred to water, which acts as the acceptor substrate, while in the case of glycosyltransferases the acceptor is typically the hydroxyl group of an acceptor other than water.  Despite a lack of evolutionary  relatedness, by simple chemical analogy, glycosyltransferases had been thought to use mechanistic  strategies  that  directly parallel those  transglycosidases (Davies, 1998).  used  by  glycosidases  and  However, some distinct differences are becoming  apparent, as will be discussed below. Sandwiched between these two enzyme groups are the phosphorylases; some of which clearly follow a glycosidase type of mechanism while others show much greater mechanistic and structural parallels to the glycosyltransferases.  1.3.  Glycosyltransferases  1.3.1. Glycosyltransferase Activities The I U B M B classification system denotes glycosyltransferase enzymes with an E C number of 2.4.x.y.  However, several of these enzymes utilize disaccharides,  oligosaccharides or polysaccharides as donor sugar substrates.  Examples include  cyclodextrin glucanotransferases (2.4.1.19), dextransucrases (2.4.1.5), and xyloglucan endotransferases (2.4.1.207).  These enzymes are structurally, mechanistically and  evolutionarily related to glycosidases and are more accurately termed transglycosidases and are therefore more appropriately classified along with glycosidases. As such, this  6  thesis will reserve the term glycosyltransferase for those enzymes that utilize an activated donor sugar substrate that contains a (substituted) phosphate leaving group. Donor sugar substrates are most commonly activated in the form of nucleoside diphosphate sugars (e.g. U D P Gal, GDP Man, etc.), however nucleoside monophosphate sugars (e.g. C M P NeuAc) and lipid phosphates (e.g. dolichol phosphate oligosaccharides) are also used. Nucleotide sugar dependent glycosyltransferases are often referred to as Leloir enzymes, in honor of Luis F. Leloir who discovered the first sugar nucleotide and was awarded the Nobel prize in chemistry in 1970 for his enormous contributions to our understanding of glycoside biosynthesis and sugar metabolism.  The acceptor substrates utilized by  glycosyltransferases are most commonly other sugars but can also be a lipid, protein, nucleic acid or antibiotic small molecule.  In addition, while glycosyl transfer most  frequently occurs to the nucleophilic oxygen of a hydroxyl substituent of the acceptor, it can also occur to nitrogen (e.g. the formation of TV-linked glycoproteins), sulphur (e.g. the formation of thioglycosides in plants) and carbon (e.g. C-glycoside antibiotics) nucleophiles.  1.3.2. Glycosyltransferase Folds As has been done for other classes of "Carbohydrate-Active enZYmes" (CAZY), glycosyltransferases are classified into families based on gene sequence similarities (Campbell et al., 1997; Coutinho et al., 2003).  This information is maintained and  continually updated on the C A Z Y website (http://cazy.orgy  Indicative of the pace and  progress of current genome sequencing endeavors, in the ten years since Campbell and colleagues first started the classification of glycosyltransferases, the number of distinct  7  families has grown from 27 to 90. However, many of these sequences are simply open reading frames that encoded putative glycosyltranferases and detailed biochemical characterization of this class of enzymes remains relatively scarce compared to that of the glycosidases.  This relative lag results from the fact that the vast majority of  glycosyltransferases  are membrane-associated  proteins, thereby  causing problems  associated with the production of insoluble and therefore uncharacterizable material when expressed as full-length proteins. Using recombinant D N A technology, this problem is being overcome by the creation of truncated, soluble forms of the enzymes consisting solely of the catalytic domains, thus greater progress should be forthcoming in the near future. Structural and mechanistic characterization of representatives from a large number of the 110 families of glycosidases has been achieved. These studies have revealed an extraordinary degree of diversity in overall fold, despite considerable commonalities in active site features and catalytic mechanisms.  This would indicate a convergence in  mechanism during the course of evolution.  In contrast, a recent burst of reported  glycosyltransferase structures has revealed a quite different situation as only two general folds, called G T - A and GT-B, have been observed for all structures of N D P sugarutilizing glycosyltransferases solved to date (Coutinho et al., 2003; Hu and Walker, 2002; Unligil and Rini, 2000). Further, threading analysis has revealed that the majority of uncharacterized families are also predicted to adopt one of these two folds. This finding is probably a consequence of the structural constraints of a nucleotide-binding motif and may indicate that the majority of glycosyltransferases have evolved from a small number of progenitor sequences. Tantalizing support for this latter notion is derived from the fact  8  that only two glycosyltransferase families (GT2 having the GT-A fold and GT4 having the GT-B fold) are possessed by primitive Archaea, suggesting that these may have served as the origins from which the vast majority of glycosyltranferases have evolved (Coutinho et al., 2003). In addition, within the GT4 family, there exist members that utilize donor substrates activated with a nucleoside diphosphate and others that utilize donor substrates activated with a phosphate group, suggesting an evolutionary link between enzymes that utilize these two substrate forms. The GT-A fold was first described for the inverting enzyme SpsA from Bacillus subtilus, for which both apo- and UDP-bound three-dimensional X-ray crystal structures were obtained (Charnock and Davies, 1999).  Consisting of an open twisted p sheet  surrounded by a helices on both sides, the overall architecture of the G T - A fold is reminiscent of the "Rossman-like" fold of nucleotide-binding proteins.  Two tightly  associated p/a/p domains, the sizes of which vary, abut closely leading to the formation of a continuous central P sheet (Figure 1.1 A). For this reason some describe the GT-A fold as a single domain fold. However, distinct nucleotide and acceptor binding domains are present (Unligil and Rini, 2000). Eukaryotic members possessing the GT-A fold typically have a short N-terminal cytoplasmic domain that is followed by transmembrane and stem regions leading to the globular catalytic region (Breton and Imberty, 1999). It should be noted that not all enzymes that possess a GT-A fold are glycosyltransferases. Indicating the possibility of a divergence in mechanism during the course of evolution, the  sugar-1-phosphate  pyrophosphorylase/nucleotidyl  transferase  superfamiliy  of  enzymes, responsible for the synthesis of nucleoside diphosphate sugars, have been shown to possess the G T - A architecture (Brown et al., 1999).  Finally, most G T - A  9  enzymes possess what is termed a D X D sequence motif, in which the carboxylates coordinate a divalent cation and/or a ribose (Breton et al., 1998; Wiggins and Munro, 1998). These are not sequence-invariant motifs and although frequently described as a determining characteristic of G T - A glycosyltransferases, examples exist of enzymes from this superfamily that do not posses the D X D "signature" (Pak et al., 2006). In addition, as -50% of all protein sequences (119,805) in SwissProt are found to possess a DxD, it certainly cannot be used as a diagnostic for a potential glycosyltransferase (Coutinho et al., 2003).  GT-A  GT-B  Figure 1.1. Overall folds observed for glycosyltransferase enzymes. The G T - A fold (A) is represented by the inverting enzyme SpsA from Bacillus subtilus (pdb lqgq) and that of the G T - B fold (B) by bacteriophage T4 (3-glucosyltransferase (pdb ljg7).  In 1994, the first determined three-dimensional structure for a nucleoside diphosphate utilizing glycosyltransferase was reported and was that of a D N A modifying P-glucosyltransferase from bacteriophage T4 (Vrielink et al., 1994). The overall fold of  10  this protein was found to be homologous to that of glycogen phosphorylase and, now that it has been observed for multiple other glycosyltransferases it is termed the GT-B fold. Like the GT-A fold, the architecture of GT-B enzymes consists of two p/a/(3 Rossmanlike domains, however, in this case the two domains are less tightly associated and face each other with the active site lying within the resulting cleft (Figure L I B ) . Also analogous to the GT-A fold, there are discrete domains associated with the donor and acceptor substrate binding sites. In addition, non-glycosyltransferase enzymes are also known to adopt the GT-B fold.  One such example is the case of U D P GlcNAc 2-  Epimerase (Campbell et al., 2000). A third glycosyltransferase fold termed GT-C was recently predicted based on iterative B L A S T searches (Liu and Mushegian, 2003), however a structure has yet to be solved for any protein from this proposed superfamily.  This report predicted eight  families (GT22, GT39, GT48, GT50, GT53, GT57, GT58, GT59) to possess the GT-C fold. Following the publication of this report, four (GT66, GT83, GT85, GT86) of the 15 glycosyltransferase families subsequently created were also predicted to adopt this fold by the C A Z Y database. The predicted architecture of the GT-C fold is that of a large hydrophobic integral membrane protein located in the E R or on the plasma membrane having between eight and 13 transmembrane helices and an active site located on a long loop region (Maeda et al., 2001; Strahlbolsinger et al., 1993; Takahashi et al., 1996). Consistent with this predicted architecture is the fact that ten of the twelve families (all but GT48 and GT53) predicted to adopt the GT-C fold utilize lipid-phosphate activated donor sugar substrates. Perhaps indicative of the limitations of prediction based solely on B L A S T analysis, a subsequent comparison of glycosyltransferase families using a  11  "profile hidden Markov method" of sequence analysis resulted in classification of family GT48, which uses UDP Glc as donor substrate, within the G T - A superfamily (Kikuchi et al., 2003). Members of family GT53, predicted to adopt the G T - C fold, utilize UDP Larabinose as donor substrates and, as such, it will be interesting to see if this family does in fact adopt the proposed G T - C fold. In addition, not all enzymes that utilize lipid phosphate-activated donor sugars are predicted, based on sequence analysis, to adopt the G T - C fold. For example, members of family GT51 utilize lipid II diphosphate activated donor sugars and family GT76 use dolichol phosphate activated donor sugars and are not predicted to adopt this fold. These are both members of a handful of orphan families that are not predicted to adopt either the G T - A , G T - B or proposed G T - C fold. Finally, certain families, for example GT2 that adopts the G T - A fold, contains both nucleoside diphosphate donor sugar-utilizing members and lipid phosphate donor sugar-utilizing members. A structural and mechanistic study of chitobiose phosphorylase from Vibrio proteolytics revealed that enzymes from family GT36 share more of a structural and evolutionary relationship with glycosidases of clan G H - L having an (ala)(, fold (Hidaka et al., 2004).  As a result, this family was subsequently reclassified as family GH-94.  This may either be interpreted as evidence supporting the notion of a mechanistic continuum between glycosidases and glycosyltransferases or may simply be indicative of inherent limitations of the methods used for classification. much-anticipated  three-dimensional  X-ray  crystal  Similarly, very recently,  structures  of a peptidoglycan  glycosyltransferase from family GT51 were solved with or without a bound substrate analogue/inhibitor (Lovering et al., 2007; Yuan et al., 2007).  As mentioned above,  12  members of this inverting family utilize lipid II phosphate activated donor substrates and the structure of the glycosyltransferase domain was not found to show similarity to either the GT-A, GT-B or proposed GT-C folds. In fact, the a-helical fold and several active site  features  are  more  akin  to  the  lysozyme family  of enzymes  and  this  glycosyltransferase family will likely be reclassified amongst the glycosidase-like phosphorylases in the near future.  This would be the first case of a lipid phosphate-  dependent enzyme being classified amongst the glycosidases and again lends support to the notion of a mechanistic continuum. Future  results  with  the  other  orphan  families  currently  classified  as  glycosyltransferases, but not predicted by sequence analysis to adopt either of the GT-A, GT-B or proposed GT-C folds, will undoubtedly shed more light on the evolutionary origin of glycosyltransferase activities and their relationships to the other major classes of carbohydrate active enzymes. Indeed, the release of the coordinates of a member of one such family (GT70), that have been deposited but not released in the R C S B Protein Data Bank, is much anticipated amongst aficionados. Whether these families will be found experimentally to actually adopt the GT-A, GT-B or putative GT-C folds; to form new glycosyltransferase superfamilies; or simply succumb to reclassifications as described above remains to be determined. In light of the described results with family GT 51, it is possible that those families proposed to adopt a GT-C fold will in fact be found to share more evolutionary and mechanistic commonalities with the glycosidases thereby leading to  an  appropriate  mass  reclassification  of  what  will  be  defined  as  true  glycosyltransferases.  13  1.3.3.  Glycosyltransferase Classification  Two stereochemical outcomes are possible for reactions that result in the formation of a new glycosidic bond: the anomeric configuration of the product can either be retained or inverted with respect to the starting material (Figure 1.2).  As such,  enzymes catalysing glycosyl group transfer are classified as either inverting or retaining, depending on the outcome of the reaction. Logically, it follows that this stereochemical outcome must result from the utilization of different mechanisms by the two classes of enzyme.  o II O-P-OR I-  "O-P-OR l-  o  Figure 1.2. Like glycosidases, glycosyltransferases catalyse glycosyl group transfer with either inversion or retention of the anomeric stereochemistry with respect to the donor sugar.  With the interesting exception of members of glycoside hydrolase (GH) family 4 (Thompson et al., 1998), the enzymes within a given G H family catalyse hydrolysis with the  same stereo-selectivity.  This is also believed to be the  case  amongst  glycosyltransferase families. As such, glycosyltranferase families are classified into clans  14  depending on their fold and the stereochemical outcome of the reactions that they catalyse (Figure 1.3). The overall fold of the enzyme does not dictate the stereochemical outcome of the reaction that it  catalyses as examples of both inverting and retaining  glycosyltransferases have been identified within both the G T - A and G T - B fold superfamilies (Coutinho et al., 2003). Indicative of this phenomenon are recent findings from studies of a mannosylglycerate synthetase, which transfers mannose to the 2-OH of D-lactate, D-glycerate or glycollate with net retention of configuration (Flint et al., 2005). Based on amino acid sequence, this enzyme was initially classified amongst the G T - A inverting  family  GT2, however,  structural  and mechanistic  reclassification amongst the G T - A retaining family GT78.  studies led to  its  To date, all enzymes  predicted to adopt the G T - C fold belong to inverting glycosyltransferase families.  15  GT-A  Inverting - Clan I  2, 7, 12, 13, 14, 16*, 21, 25*. 31, 40*, 42, 43, 49*. 82, 84  Retaining - Clan  6, 8, 15, 24, 27, 34*, 44, 45*. 55, 60*. 62*, 64, 78, 81  Inverting - Clan II  26*, 28, 30,33,41,  Non-GT  1, 9, 10, 17*,19, 23,  47*, 56*. 63, 80  GT-B  Retaining - Clan IV  f ' ' ' *' 4 5  20 32  35,  2  Non-GT  Figure 1.3. Glycosyltransferase (GT) classification system proposed by Coutinho et al. (Coutinho et al., 2003). Families are classified into clans based on their fold and activity. GT family numbers belonging to each clan are indicated on the far right. Bona fide families are indicated in red, having members with solved three-dimensional structures. The remaining families are those predicted to adopt either the G T - A or GT-B fold. Families identified with an asterisk are those with structures predicted to adopt either the GT-A or GT-B fold solely by Liu and Mushegian (Liu and Mushegian, 2003) and those not denoted with an asterisk have G T - A or GT-B structures as predicted by both Liu and Mushegian and the C A Z Y website. This classification system does not include 39 of the 90 glycosyltransferases.  Members from twelve (GT22, GT39, GT48, GT50, GT53,  GT57, GT58, GT59, GT66, GT83, GT85, GT86) of those families not included are predicted to adopt a proposed GT-C fold.  Based on a determined three-dimensional  structure, family GT36 was reclassified amongst the glycosidases as family GH94. Structural characterization of the remaining 26 orphan families (GT11, GT18, GT29, GT37, GT38, GT46, GT51, GT52, GT54, GT61, GT65, GT67, GT68, GT69, GT70, GT71, GT73, GT74, GT75, GT76, GT77, GT79, GT87, GT88, GT89, GT90) will provide insights into the strengths and limitations of predictive bioinformatic tools.  16  1.3.4. Inverting Glycosyltransferases 1.3.4.1. Mechanism of Inverting Glycosidases The mechanistic strategy employed by inverting glycosidases is that of a direct displacement SN2-like reaction. A pair of carboxyl groups exist within the active site, typically separated by 7-11 A, one acting as an acid (A) protonating the glycosidic oxygen, the second acting as a base (B) that activates the incoming water nucleophile facilitating a reaction that proceeds via an oxocarbenium ion-like transition state (Zechel and Withers, 2000)(Figure 1.4).  J V W V A A i  8+ B  5  "  I  ROH  B  OR*  I A A A A A A A  1—  vA/VWW\  Oxocarbenium Ion-Like Transition State  Figure  1.4. A direct displacement SN2-like reaction results in inverted anomeric  configuration via a single oxocarbenium ion-like transition state. For glycosidases, R = a carbohydrate derivative and R ' O H = H 2 O or phosphate (phosphorylases classified as glycosidases). For glycosyltransferases, R = a nucleoside diphosphate, a lipid phosphate, or phosphate (phosphorylases classified as glycosyltransferases) and R ' O H = an acceptor group (e.g. another sugar, a protein, or an antibiotic).  17  1.3.4.2. Mechanism of Inverting Glycosyltransferases Catalysis by inverting glycosyltransferases is believed to directly parallel that of inverting glycosidases wherein a base (B) deprotonates the incoming nucleophile of the acceptor facilitating direct S 2 displacement of the activated (substituted) phosphate N  leaving group (Figure 1.4). Indeed, structural and mechanistic studies clearly point to such a mechanism with the major difference between these two classes simply being the manner by which leaving group activation/departure is facilitated. The key questions in examining the catalytic mechanism of inverting glycosyltransferases are, therefore, the identity of the base catalyst and the method used to facilitate departure of the (substituted) phosphate leaving group.  1.3.4.3. Inverting GT-A Glycosyltransferases (Clan I) Excluding members of the sialyltransferase family GT42 that are the subject of the detailed analysis described in chapter 2, structural and mechanistic studies of nine representatives from six families of inverting glycosyltransferases with the G T - A fold have been undertaken.  These studies have revealed several commonalities and  differences amongst members of Clan I glycosyltransferases. As mentioned, the family GT2 enzyme SpsA from Bacillus subtilis was the first glycosyltransferase experimentally shown to adopt the GT-A fold (Charnock and Davies, 1999). This family represents the largest and evolutionarily most ancient of inverting glycosyltransferases. Unfortunately, the natural donor and acceptor sugars of this enzyme are not known. However, fortuitously, a molecule of the cryoprotectant glycerol was found bound within the SpsA active site in a suitable position to mimic the natural  18  acceptor sugar and hydrogen bonded to the side chain carboxylate of Aspl91, leading to the proposition that this residue played the role of base catalyst in the proposed direct displacement mechanism (Figure 1.5A)(Charnock and Davies, 1999). This notion was later supported by superpositioning of the SpsA structure on the solved G T - A structures of the enzymes lactose synthases (Gal-Tl) from family GT7 (Gastinel et al., 1999; Ramakrishnan et al., 2002; Ramakrishnan and Qasba, 2001, 2002; Ramakrishnan et al., 2006; Ramakrishnan et al., 2001; Ramasamy et al., 2005), GnT-I from family GT13 (Gordon et al., 2006; Unligil et al., 2000), and GlcAT-I from family GT43 (Pedersen et al., 2002; Pedersen et al., 2000; Tarbouriech et al., 2001). These enzymes all have a conserved Asp residue within their active sites occupying a position equivalent to that of Aspl91 in SpsA (Figure 1.5). Similarly, an analogously positioned conserved Asp or Glu residue has been observed in the subsequently obtained structures of the enzymes C2GnT-L from family GT14 (Pak et al., 2006), Mfng from family GT31 (Jinek et al., 2006), and GlcAT-P from family GT43 (Kakuda et al., 2004; Ohtsubo et al., 2000) (Figure 1.5).  Structures with bound acceptor substrates indicate that these conserved  carboxylates are within hydrogen bonding distance of the nucleophilic hydroxyl undergoing reaction, consistent with their roles as base catalysts (Figure 1.5).  Further  support comes from site-directed mutagenesis studies of the family GT2 enzyme ExoM, in which it was shown that mutation of Aspl87 (structurally homologous to Aspl91 in SpsA) abolished all in vitro glycosyltransferase activity (Garinot-Schneider et al., 2000).  19  A -SpsA  B -Gal-TI  (pdb 1qgq) (GT2)  (pdb 2fyd) (GT7)  Glycerol  D254  C -GnTI  D -Mfng  (pdb 1foa) (GT13)  (pdb 2j0b) (GT31)  D232  i  UDP GlcNAc  E -GlcAT-l (pdb 1fgg) (GT43)  D281  F -GlcAT-P (pdb 1v84) (GT43)  D284  D194  UDP GlcNAc  Figure 1.5.  Comparison of the active sites of several metal-dependent inverting  glycosyltransferases from Clan I. A conserved side chain carboxylate is located in a near identical relative position within all active sites on the P face of the donor substrate and (when present) within hydrogen bonding distance of the nucleophilic hydroxyl of the acceptor sugar and plays the role of base catalyst in a direct displacement mechanism. Mn  cations are shown as magenta spheres.  20  , The vast majority of enzymes from glycosyltransferase Clan I that have been subjected to biochemical analysis use an essential divalent cation (usually M n  2 +  or M g ) , 2+  coordinated by the so-called D X D motif, to faciliate departure of the nucleoside diphosphate leaving group by electrostatically stabilizing the developing negative charge. A notable exception to this strategy has recently been reported for the metal ionindependent (3-1,6-GlcNAc transferase C2GnT-L from family GT14 (Pak et al., 2006). This enzyme uses the basic amino acids Arg378 and Lys401 to electrostatically stabilize the nucleoside diphosphate leaving group. This strategy is reminiscent of that used by metal ion-independent glycosyltransferases possessing a GT-B fold (as discussed below), indicative of a convergence of mechanism amongst these two superfamilies. Theoretical support for a concerted SN2-like displacement mechanism for inverting glycosyltransferases  from Clan I is derived from a hydbrid quantum  mechanical/molecular mechanical (QM/MM) study of the p-l,2-GlcNAc transferase GnT-I (Kozmon and Tvaroska, 2006). Using the 1.8 A resolution three-dimensional X ray crystal structure of GnT-I with bound U D P GlcNAc (Unligil et al., 2000) and a knowledge of the ordered sequential B i B i kinetic mechanism of this enzyme (Nishikawa et al., 1988), a theoretical model of the Michaelis complex was generated and subjected to Q M / M M analysis. The results supported a catalytic mechanism involving a concerted SN2-type transition state involving near simultaneous nucleophilic attack, facilitated by proton transfer to the catalytic base (Asp291), and leaving group dissociation steps. A n activation energy of ~19 kcal/mol was estimated for the proposed transition state.  21  1.3.4.4. Inverting GT-B Glycosyltransferases (Clan II) To date, structural and mechanistic studies have been reported  for 14  representatives from seven families of inverting glycosyltransferases that have been shown to possess a GT-B fold.  Despite sharing homologous three-dimensional  architectures, members of this superfamily seem to display a greater degree of diversity in the selected modes of catalysing glycosyl group transfer between, and in some cases even amongst, families compared to what has been observed for the GT-A superfamily. This greater diversity in mechanism is perhaps a reflection of the greater diversity of chemistries catalysed by this superfamily of enzymes, compared to the G T - A superfamily. The prototype of this superfamily is the P-glucosyltransferase (BGT) from T4 bacteriophage, the first nucleoside diphosphate utilizing glycosyltransferase for which a three-dimensional X-ray crystal structure was obtained (Vrielink et al., 1994).  This  unique enzyme, the sole member of family GT63, transfers glucose from UDP Glc to the hydroxymethyl substituents of modified cytosine bases.  It is proposed that this  mechanism has evolved as a defence mechanism to prevent genome degradation from phage and host nucleases (Kornberg et al., 1961). In addition to the initially reported structures with bound metal ions and U D P present (Morera et al., 2001; Vrielink et al., 1994), a subsequent ternary complex X-ray crystal structure with bound U D P and D N A was obtained (Lariviere and Morera, 2002). The ternary complex structure revealed the ability of this glycosyltransferase to facilitate selective glycosylation by inducing the D N A to "flip out" in a fashion reminiscent of D N A glycosylases, methyltransferases and endonucleases.  Analogous to what has been found with the G T - A superfamily,  22  mutagenesis studies revealed Asp 100 to be a likely candidate for base catalyst. This is supported by the observation that co-crystallization of the wild-type enzyme with U D P Glc results in complete hydrolysis of the donor sugar resulting in the obtainment of UDP product complexes, while analogous co-crystallization procedures with the D100A mutant results in the observation of intact donor sugar within the active site (Lariviere et al., 2003). It should be noted that the D100A mutation prevents the formation of an interdomain salt bridge with Argl91, preventing complete formation of the active site, which could also be the cause of the observed decrease in substrate turnover. A crystal structure with bound intact donor substrate was also obtained by a brief soaking of wild-type B G T crystals with UDP Glc (Figure 1.6A). Interestingly, this donor substrate complex had no metal cation bound within the active site despite being soaked with high concentrations of 2+  Mg  (Lariviere et al., 2003).  Instead, positively charged side chains were found to  neutralize the negative charges of the pyrophosphate group of the bound donor substrate (Figure 1.6A). In contrast, in the U D P product complex, obtained by co-crystallization with M n  2 +  and UDP, an M n  2 +  cation was found coordinating the pyrophosphate group  and was located in the region occupied by the glucose moiety in the UDP Glc complexed structure (Figure 1.6B). This led the authors to propose that the divalent cation plays a role in facilitating product release and not necessarily in the cleavage of the glycosidic linkage of the donor substrate.  This alternative mode of leaving group activation,  compared to the G T - A superfamily, has also been proposed for several other GT-B enzymes (described below).  23  A  B D100  D100 2+  Figure  1.6. Active site of wild-type T4 bacteriophage B G T from family GT63. (A)  Donor substrate complex with bound UDP Glc (pdb 1J39). Note the triad of positively charged Arg residues as well as side chain hydroxyls suitably positioned to stabilize developing charge on the UDP leaving group in the absence of a divalent cation. (B) Product complex with bound M n  cation partially occupying the glucose binding site  (lnvk).  Glycosyltransferases from family GT1 adopt the GT-B fold and are responsible for the glycosylation of various organic structures such as terpenes, anthocyanins, cofactors, steroids, peptide antibiotics, and macrolides. The various products produced by this family of enzymes serve critical biological functions, making this one of the most intensely studied families of glycosyltransferases.  Three family GT1 enzymes (GtfA  (Mulichak et al., 2003); GtfB (Mulichak et al., 2001); and GtfD (Mulichak et al., 2004) involved in the biosynthesis of the peptide antibiotic vancomycin have been subjected to structural and biochemical analysis.  The results of these studies suggest that these  enzymes use Asp 13 as the catalytic base and that leaving group departure is facilitated in a metal ion-independent fashion using a helix dipole and interactions with side chain  24  hydroxyl and imidazole groups to stabilize the developing negative charge. A second group of related enzymes from family GT1 have also been characterized, revealing interesting differences from the Gtf enzymes.  These include the multifunctional  terpene/flavonoid glycosyltransferase UGT71G1 (Shao et al., 2005); the flavonoid glucosyltransferase VvGTI (Offen et al., 2006); the macrolide glycosyltransferases OleD and Olel (Bolam et al., 2007); and the human drug metabolizing glucuronyltransferase UGT2B7 (Miley et al., 2007). Although these enzymes were found to use a metal ionindependent mechanism, analogous to that described above, to facilitate leaving group departure, the catalytic base was determined to be a side chain imidazole that interacts with an adjacent conserved side chain carboxylate group. These conclusions were based on comparisons to a three-dimensional X-ray crystal structure with bound acceptor substrate (Offen et al., 2006), in which case the putative side chain imidazole was found within hydrogen bonding distance of the reactive acceptor substrate hydroxyl, and confirmed by kinetic analysis of enzyme mutants. Other inverting GT-B enzymes that have been subjected to structural and mechanistic analysis include the heptosyltransferase WaaC from family GT9 (Grizot et al., 2006); the fucosyltransferases FucT from family GT10 (Sun et al., 2007) and FUT8 from family GT23 (Ihara et al., 2007); and the GlcNAc transferase essential for bacterial cell wall synthesis known as MurG (Ha et al., 2000; Hu et al., 2003). These enzymes all appear to use metal ion-independent methods for stabilizing the departure of nucleoside diphosphate leaving groups analogous to that described above. In the cases of WaaC and FucT, structural and kinetic analysis of enzyme mutants supports the roles of the side  25  chain carboxylates of Asp 13 and Glu95 (respectively) as base catalysts. In the cases of FUT8 and MurG, the identity of the base remains less clear. A  rather  intriguing  story  is  unfolding  sialyltransferase PmSTl from family GT80.  involving  the  multifunctional  This metal ion-independent enzyme has  been reported to possess four distinct enzymatic activities with differing pH optima consisting of: an a-2,3-sialyltransferase activity with an optimal pH of 7.5-9.0; an a-2,6sialyltransferase activity with an optimal pH of 4.5-6.0; an a-2,3-sialidase activity with an optimal pH of 5.0-5.5; and an a-2,3-trans-sialidase activity with an optimal pH of 5.5-6.5 (Yu et al., 2005). It should be noted that the presence of C M P could be the cause of the observed sialidase and trans-sialidase activities, as this enzyme, like most other glycosyltransferases, presumably catalyses the hydrolysis of its donor sugar substrate. At the high concentrations of enzyme used in determining these activities, back reaction in the presence of C M P would lead to the formation of C M P NeuAc in situ, which could then be hydrolysed (leading to the observed sialidase activity) or used as a substrate in the catalysed transfer of sialic acid to an alternative acceptor (leading to the observed transsialidase activity). The three-dimensional X-ray crystal structure of this enzyme has been solved under conditions (pH 7.5) that favour the a-2,3-sialyltransferase activity in apo and product complex form with bound C M P (Ni et al., 2006).  Subsequent co-  crystallization with C M P 3F NeuAc, followed by soaking with lactose acceptor, led to the production of a ternary complex structure (Ni et al., 2007). This revealed the presence of an Asp/His pair, reminiscent of several of the previously described family GT1 enzymes, whose side chain functional groups were suitably positioned for either residue to play the role of base catalyst. The D141A mutant was found to have a transferase activity 20,000-  26  fold lower than wild-type enzyme, compared to a 350- to 1000-fold lower activity measures for the HI 12A mutant. As such, in contrast to what has been suggested with the analogous family GT1 enzymes, Aspl41 was proposed to play the role of base catalyst (Ni et al., 2007).  1.3.5. Retaining Glycosyltransferases 1.3.5.1. Mechanism of Retaining Glycosidases With the notable exception of members from family G H 4 that use a recently described elimination mechanism involving transient remote oxidation (Yip et al., 2004) and retaining p-hexosaminidases that use an intramolecular nucleophile (reviewed in (Mark and James, 2002), the typical mechanism of retaining glycosidases is that of a double-displacement reaction involving a covalently bound glycosyl-enzyme intermediate species (Zechel and Withers, 2000). This mechanism was first put forward by Koshland (Koshland, 1953), who realized that, because inversion of configuration is a fundamental and universal property of bimolecular displacement at saturated  carbon centres,  glycosidases that gave sugar products with the same anomeric configuration as the substrate must catalyse the reaction using two distinct nucleophilic displacement steps involving an enzymatic catalytic nucleophile. A n alternative S N I mechanism, involving the formation of a discrete enzyme-stabilized oxocarbenium ion intermediate species that is shielded on one face by the enzyme, thereby preventing nucleophilic attack from the opposite face of the reaction centre and leading to complete retention of anomeric configuration in the product, was subsequently proposed for the retaining glycosidase hen egg white (HEW) lysozyme by Phillips (Phillips, 1966, 1967). It should be noted that  27  these two mechanisms are the extremes of a range of possible nucleophilic substitution reactions.  The free energy of the intermediate, and therefore the associated transition  states, for the Koshland mechanism would be lower than for those associated with the Phillips mechanism.  The glycosidic linkage between two sugars is very stable.  For  example, the half-lives for the spontaneous hydrolysis of starch and cellulose at room temperature are in the range of 5 million years (Wolfenden et al., 1998). As such, the fact that glycosidases are observed to catalyse the hydrolysis of these materials with rate constants of up to 1000 s" would lead one to believe that this class of enzyme has 1  evolved a mechanism that would involve an intermediate species with the lowest possible free energy, thereby facilitating as much of a decrease in the activation barrier as possible, in order to achieve such a formidable task with such proficiency. With some notable exceptions (Fersht, 1999), this logic, a large array of mechanistic data and relentless touting from its supporters, has led most in the field to accept the Koshland mechanism as the mechanism utilized by virtually all retaining glycosidases. Analogous to the inverting glycosidases, a pair of carboxylates exist within the active site of retaining glycosidases, in this case 5 A apart, with one acting as a general acid/base catalyst (A/B) while the other acts as a nucleophile (Nuc), the reaction again proceeding via oxocarbenium ion-like transition states (Figure 1.7).  Amongst the  glycosidases, notable exceptions as to the nature of the catalytic nucleophile exist amongst certain hexosaminidases in cases where an iV-acetamido group acts as an intramolecular nucleophile (Mark and James, 2002) and a recently characterized transsialidase in which case a tyrosine plays the role (Watts et al., 2003).  -1*  6+  Nuc  | 5-  Nuc  16+ A/B  ROH  A/B" Glycosyl-Euzyme Intermediate  v A A A A A A A  Nuc  Oxocarbenium Ion-Like Transition State  OR' A/B  H'  \+  •  Nuc  R'O A/B  A/B  v/XA/vvvxn.  Oxocarbenium Ion-Like Transition State  Figure 1.7. The double displacement mechanism established for retaining glycosidases proceeds via two oxocarbenium ion-like transition states with the intermediate formation of a discrete covalently bound glycosyl-enzyme species, resulting in overall retention of anomeric configuration. For glycosidases, R = a carbohydrate derivative and R ' O H = H 2 O or phosphate (phosphorylases classified as glycosidases).  Evidence supporting an SN2-like mechanism involving enzymatic nucleophilic catalysis comes from the observation of normal secondary a deuterium KIEs for both transition states (indicating rehybridization of CI from sp to sp )(Kempton and Withers, 3  2  1992; Sinnott and Souchard, 1973; Vocadlo et a l , 2002), as well as the observation of primary al.,  1 3  C KIEs >1.01 consistent with S 2 pathways (Berti and Tanaka, 2002; Huang et  1997).  N  Additionally,  the  covalent  glycosyl-enzyme  intermediates  from  29  representatives of multiple families have been isolated and characterized using mass spectrometry and X-ray crystallography (Zechel and Withers, 2000). This includes the characterization of the intermediate of lysozyme, a result that contradicts the classical textbook mechanism involving the formation of a discrete oxocarbenium ion intermediate species (Vocadlo et al., 2001).  1.3.5.2. Mechanism of Retaining Glycosyltransferases Again by direct comparison to retaining glycosidases, the mechanism of retaining glycosyltransferases has been proposed to be that of a double displacement mechanism involving a covalently bound glycosyl-enzyme intermediate (Figure 1.8), demanding the existence of an appropriately positioned nucleophile within the active site (Davies, 1998). A divalent cation or suitably positioned positively charged side chains or helix dipoles would presumably play the role of the Lewis acids as was described above for the inverting glycosyltransferases. The leaving diphosphate group itself probably plays the role of a base catalyst activating the incoming acceptor hydroxyl group for nucleophilic attack.  The role of substrate/product phosphates acting as base catalysts has been  previously postulated in the mechanism of farnesyl diphosphate synthase (FPPS) from E. coli (Hosfield et al., 2004) and T. Cruzi (Gabelli et al., 2006).  30  OR  Oxocarbenium Ion-Like Transition State  \ HO  9  H  —O  ,0—P=0 2 +  M'''  OR  ^  Glycosyl-Enzyme Intermediate  Oxocarbenium Ion-Like Transition State Figure  1.8. Proposed double displacement mechanism for retaining glycosyltransferases.  For glycosyltransferases, R = a nucleoside diphosphate (e.g. U D P , GDP), a lipid phosphate, or phosphate (phosphorylases classified as glycosyltransferases) and R ' O H = an acceptor group (e.g. another sugar, a protein, or an antibiotic).  In addition, for  glycosyltransferases the role of base catalyst is likely played by the departing (substituted) phosphate leaving group.  31  1.3.5.3. Challenges in Studying the Mechanisms of Retaining Glycosyltransferases Mechanistic characterization of this class of enzymes has proven to be a challenging task. The conclusive identification of a catalytic nucleophile and observation of a true, kinetically and catalytically covalent intermediate has yet to be reported for any retaining transferase despite years of exhaustive studies using techniques that have been successfully applied to the characterization of retaining glycosidases. Although this may be interpreted as evidence against the double displacement mechanism, it could also be the result of the inapplicability of these techniques to the study of transferases due to inherent differences in the nature of the substrates being studied. The most successful approach used for the characterization of retaining glycosidases has involved the use of fluorinated substrate analogues (Withers and Aebersold, 1995). The introduction of an electronegative fluorine at either the 2 or 5 position of a pyranose ring inductively destabilizes the oxocarbenium ion-like transition states through which both steps of the double displacement reaction proceed and in some cases also removes key hydrogenbonding interactions, resulting in a significant decrease in the rate of the overall reaction. By introducing a good leaving group (e.g. dinitrophenol or fluoride), the first step is "rescued", resulting in the accumulation of the intermediate species with a significant lifetime that allows mass spectrometric and X-ray crystallographic characterization. Alternatively, or used in combination with the fluoro-sugar approach, removal of the acid/base catalyst of a retaining glycosidase by mutagenesis also leads to a decrease in the rates of both steps of the reaction and again, by using a substrate with a highly activated  32  leaving group, the glycosylation step can be "rescued" leading to the accumulation of the glycosyl-enzyme intermediate. However, because of the strict requirement of the glycosyltransferases for their NDP leaving group, the relative leaving group ability cannot be manipulated, thus the relative rates of the glycosylation and deglycosylation steps cannot be altered. In order to prevent hydrolysis of metabolically expensive (substituted) phospho donor sugar substrates, it would be expected that glycosyltransferases would have evolved a mechanism in which the step involving cleavage of the bond between the sugar and its activated leaving group would be rate limiting thereby preventing the accumulation of an observable intermediate species which could be easily hydrolysed. In addition, because the leaving groups are themselves believed to play the role of base catalyst, reactions cannot be slowed down for this class of enzyme by mutagenesis of the protein catalyst. This inability to alter the relative rates of glycosylation versus deglycosylation has rendered the fluoro sugar approach ineffective in the trapping of intermediates on retaining transferases (Ly et al., 2002).  1.3.5.4. Retaining GT-A Glycosyltransferases (Clan III) To date, structural and mechanistic studies of twelve representatives from seven families of retaining glycosyltransferases that have been shown to possess the G T - A fold have been undertaken. The results of these studies support generally utilized strategies amongst members of Clan III glycosyltransferases for facilitating leaving group departure and activating the nucleophilic group of the incoming acceptor. However, the observed structures have not revealed a conserved structural architecture on the (3-face of the donor  33  sugar substrate in the region that would be expected to be occupied by the obligate enzymatic nucleophile of a proposed double displacement mechanism. The family GT8 galactosyltransferase LgtC from Neisseria meningitidis was the first retaining nucleoside diphosphate-utilizing retaining glycosyltransferase for which a three-dimensional X-ray crystal structure was obtained (Persson et al., 2001). Using nonreactive U D P 2F Gal and 4'-deoxy lactose as donor and acceptor substrate analogues respectively, a well-resolved ternary complex structure was obtained (Figure 1.9A). As was seen with the inverting G T - A enzymes, this structure revealed the presence of a M n  2 +  cation, coordinated within the active site by the carboxylate side chains of a D X D motif and the donor substrate diphosphate leaving group. It is presumed that the M n  2 +  acts as  Lewis acid to facilitate leaving group departure. In the region that would be occupied by the reactive axial 4'-hydroxyl of the acceptor substrate analogue, the only functional group that is suitably positioned to activate the incoming nucleophile is the leaving group oxygen of the diphosphate leaving group. This suggests that the departing diphosphate moiety acts as the base catalyst. Surprisingly, the only functional group that is suitably positioned within the active site on the p-face of the donor substrate to play the role of the catalytic nucleophile is that of the side chain amide of Gin 189.  The amide carbonyl  oxygen of this residue is perfectly positioned 3.5 A away and with an ideal trajectory for nucleophilic attack on the anomeric reaction centre. However, the results of mutagenesis studies were not consistent with an essential role for this residue as the catalytic nucleophile, since the corresponding alanine mutant retained 3% activity. This enzyme was the subject of the mechanistic studies described in section 3.2 and a detailed description of previous mechanistic studies are provided therein.  34  Figure  1.9. Comparison of the active sites of several retaining glycosyltransferases from  Clan III. Descriptions of conserved structural features, and a lack thereof, are provided in the text. M n  2 +  cations, which play the role of a Lewis acid catalyst that facilitates leaving  group departure, are shown as magenta spheres. Distances are indicated in Angstroms.  35  Subsequently, the three-dimensional X-ray crystal structure of rabbit muscle glycogenin, another family GT8 enzyme, with intact U D P Glc donor substrate bound within the active site was reported (Gibbons et al., 2002). This enzyme catalyses selfglucosylation of a tyrosine residue in another monomer of the enzyme, in a process that is the initial step of glycogen biosynthesis, plus successive glycosylations of the glucosyl tyrosyl formed. Following the formation of an a-l,4-linked oligosaccharide covalently bound to glycogenin, glycogen synthase and the branching enzyme complete the process of polysaccharide synthesis.  The positioning of the tyrosyl hydroxyl that undergoes  glucosylation, 21 A away from the anomeric reaction centre on the surface of the same monomer, suggests that the initial glucosylation reaction occurs in an inter-subunit process with further glucose elaboration possibly occurring in subsequent intra-molecular reactions (Gibbons et al., 2002). This would be consistent with previous mechanistic studies. As was the case with LgtC, a M n  2 +  cation was observed within the active site of  glycogenin coordinated by the side chain carboxylates of a D X D motif positioned to act as a Lewis acid that activates the departing diphosphate leaving group (Figure 1.9B). However, unlike LgtC, the side chain most suitably positioned on the P-face of the donor substrate to act as the catalytic nucleophile is that of the carboxylate of Asp 163, albeit at a distance of 6.1 A from the reaction centre (Figure 1.9B).  This conserved residue  corresponds to Aspl88 of LgtC, which is in close proximity to the positively charged side chain of Lys86. As will be discussed, this residue pairing constitutes the only conserved structural motif that can be identified on the P-face of the donor substrate binding sites of retaining glycosyltransferases. In LgtC the side chain carboxylate of Asp 188 is within hydrogen bonding distance of the C6 and C4 hydroxyls of the galactose moiety of the  36  donor substrate and the positively charged side chain of Arg86 (Figure 1.9A). In the glycogenin structure, the side chain amide of Gin 164, which corresponds to Gin 189 in LgtC, is within hydrogen bonding distance of the C4 hydroxyl of the glucose donor substrate (Figure 1.9B). These observations indicate that either different donor sugar binding modes exist amongst enzymes of the same family and/or that significantly different ground state structures exist, indicating a high degree in structural plasticity for this class of enzyme, for which snapshots can be observed by crystallography. Bovine a3GalT from family GT6 is another retaining  galactosyltransferase  possessing a G T - A fold that has been the subject of significant structural and mechanistic investigations.  The initial crystal structure reported suggested the observation of a  covalently bound p-galactosyl-enzyme intermediate but this was dogged by considerable disorder and limited resolution (Gastinel et al., 2001). Confidence in the interpretation of this putative covalently modified species is significantly compromised by the fact that a second donor galactose residue from a non-covalently bound UDP Gal species was also present. In addition, the electron density for the "covalently bound" galactose was weak. Higher resolution structures were subsequently reported of complexes of a3GalT with UDP (Boix et al., 2002), U D P and acceptors (Zhang et al., 2003) as well as the nonreactive donor substrate analogue U D P 2F Gal (Jamaluddin et al., 2007). The structure with bound UDP 2F Gal clearly reveals the side chain carboxylate of Glu317, structurally equivalent to Glnl 89 in LgtC, to be suitably positioned to play the role of the proposed catalytic nucleophile in a double displacement mechanism (Figure 1.9C). Interestingly the U D P 2F Gal is bound in an essentially identical "folded" conformation to that seen in LgtC. The conserved structural motif of the side chain carboxylate from residue Asp316  37  within hydrogen bonding distance of donor sugar hydroxyls and the positively charged side chain of residue Arg202 is also present (Figure 1.9C). Once again these structures all reveal the presence of a bound divalent cation within the active site suitably positioned to play the role of a Lewis acid that facilitates leaving group departure. product complex with bound UDP contained a second M n  2 +  In addition, a  bound within the active site  in a way that suggests a role in facilitating product release (Jamaluddin et al., 2007). Bovine a3GalT is the subject of the mechanistic investigations described in section 3.4 and a discussion of previous mechanistic studies is provided therein. The glycosyltransferases responsible for addition of the specificity-determining Gal/GalNAc moieties to the cell surface glycolipids and glycoproteins that constitute the ABO(H) blood groups are another group of family GT6 enzymes that have been subjected to intensive structural scrutiny (Watkins, 1957; Yamamoto et al., 1990). Glycosyltransferase A (GTA) specifically uses U D P GalNAc as a donor substrate to modify the 3-OH of the terminal galactose residue of the H-antigen (terminal Fuc-a(l,2)Gal-P-R acceptors)  thereby creating the A blood group antigen.  Conversely,  glycosyltransferase B (GTB) specifically uses U D P Gal to modify the same hydroxyl of H-antigen to generate the B blood group antigen. Amazingly, these two enzymes differ from each other by only 4 out 354 amino acids. The three-dimensional X-ray crystal structures of these two enzymes with both U D P and H-antigen bound revealed that distinction between the two donor substrates is achieved by the substitutions of residues Leu266 and Gly268 in G T A with the more bulky side chains of Met266 and Ala268 in GTB (Patenaude et al., 2002). Similarly, a crystal structure of the inactive mutant of this enzyme, expressed by individuals of the O blood group wherein terminal H-antigens are  38  not modified, revealed the ability of a single mutation, that changes Gly/Ala268 with the bulky side chain of Arg, to block donor substrate binding resulting in the ablation of catalytic activity (Lee et al., 2005). The role of intramolecular hydrogen bonding within acceptors (Nguyen et al., 2003) and the role of the two residues (Gly/Ser235 and Leu/Met266) (Letts et al., 2006) in the assembly of Type I and II H antigens by G T A and GTB has also been revealed by X-ray crystallographic analysis. Finally, the role of various missense mutations leading to the production of enzymes of weak dual specificity has been investigated (Hosseini-Maaf et al., 2007; Persson et al., 2007). Family  GT27  contains  a  series  of  UDP  GalNAc:polypeptide a-TY-  acetylgalactosaminyltransferases (ppGalNAcTs) that initiate the formation of mucin-type 0-linked glycans by catalysing the transfer of GalNAc to serine or threonine residues on protein surfaces, thereby generating the Tn antigen (GalNAc-a-O-Ser/Thr).  These  enzymes are unique amongst the glycosyltransferases in that they possess C-terminal lectin domains. The majority of ppGalNAcTs, termed peptide transferases, are able to catalyse transfer to both unmodified peptides as well as to glycopeptides. Others, termed glycopeptide transferases, are only able to transfer to peptides that have been modified by an initial GalNAc residue (Ten Hagen et a l , 2001; Ten Hagen et al., 1999). The threedimensional X-ray crystal structures of murine ppGalNAcT-1 isozyme with only M n  2 +  bound (Fritz et al., 2004), human ppGalNAcT-10 isozyme in complex with either hydrolysed donor or a GalNAc-serine acceptor (Kubota et al., 2006) and human ppGalNAcT-2 isozyme with UDP and an acceptor peptide bound (Fritz et al., 2006) have been reported.  These structures indicate that activation of the U D P leaving group is  achieved in the usual fashion by the presence of a coordinated M n  2 +  cation within the  39  active site poised to act as a Lewis acid (Figure 1.9D). In addition, the conserved motif of a side chain carboxylate (Glu345 in ppGalNAcT-10 and Glu334 in ppGalNAcT-2) positioned within hydrogen bonding distance of donor substrate hydroxyls and a positively charged side chain (Arg221 in ppGalNAcT-10 and Arg208 in ppGalNAcT-2) is again observed (Figure 1.9D). The side chain amides of Gln346 of ppGalNAcT-10 and Asn335 of ppGalNAcT-2 are located in a structurally analogous position to that of Gin 189 in LgtC and are therefore the most likely candidate to act as the catalytic nucleophile based simply on the available structural information (Figure 1.9D). These structural studies have also revealed how dynamic association between the lectin and catalytic domains, facilitated by a flexible tether, permits the formation of a range of binding modes for various glycopeptide acceptor substrates, thereby facilitating the production of the characteristic high density glycosylation patterns associated with mucins. The catalytic domain of Clostridium difficile Toxin B (ToxB) is a retaining family GT44 glucosyltransferase possessing a GT-A fold that modifies host Rho proteins by glycosylating key threonine residues (Just et al., 1995).  A three-dimensional X-ray  crystal structure of the catalytic domain with bound M n , U D P and glucose has been 2 +  reported (Reinert et al., 2005). This structure revealed the conserved retaining G T - A architectural features of a bound M n that acts as a Lewis acid activator, as well as a side chain carboxylate (Asp270) within hydrogen bonding distance of the donor sugar hydroxyls and a positively charged side chain (Arg273) (Figure 1.9E). The side chain amide of Asn384 is the most suitably positioned candidate for a putative catalytic nucleophile (Figure 1.9E).  40  Mannosylglycerate synthase (MGS) from Rhodothermus marinus is a family GT78 mannosyltransferase that is responsible for the synthesis of the stress protectant 20-a-D-mannosylglycerate.  A n X-ray crystal structure of MGS with bound M n  2 +  and  intact GDP Man bound within the active site also revealed the presence of a coordinated divalent cation Lewis acid activator and a side chain carboxylate (Asp 192) within hydrogen bonding distance of donor sugar hydroxyls and a positively charged side chain (Lys72) (Figure 1.9F)(Flint et al., 2005). The main chain amide of Leul63 is the only functional group suitably positioned on the P-face of the anomeric reaction centre of the mannosyl donor substrate to act as a catalytic nucleophile (Figure 1.9F). Three-dimensional X-ray crystal structures have been reported for the retaining enzymes Kre2 (Lobsanov et al., 2004), a yeast mannosyltransferase from family GT15, and E X T L 2 (Pedersen et al., 2003), a mouse N-acetylhexosaminyltransferase involved in heparan biosynthesis from family GT64. Both of these enzymes were shown to possess a G T - A fold, however the absence of bound donor sugar substrates limits the utility of these structures in identifying candidate catalytic nucleophiles. Both reported structures had bound divalent cations coordinating the nucleoside diphosphate product. A ternary complex structure of E X T L 2 with bound U D P and an acceptor substrate supports a role for the phosphate leaving group in acting as the base catalyst that activates the incoming nucleophile. Kre2 is the subject of the structural and mechanistic studies described in section 3.3. A comparison of representative retaining glycosyltransferases, possessing the GTA fold, from various families illuminates some conserved structural features within some regions of the active site and a lack thereof in other regions for this class of enzyme. The  41  mode of leaving group activation is similar to that used by the majority of inverting G T - A enzymes. A divalent cation, coordinated by the side chains of a conserved D X D motif and the side chain imidazole of a structurally conserved histidine residue, acts as a Lewis acid that facilitates leaving group departure.  In contrast to the conserved features in the  region surrounding the leaving group and incoming acceptor (the a-face of the donor substrate), there appear to be very few conserved architectural features within the active site region that accommodates the (3-face of the donor.  The only strictly conserved  feature observed on this face of the reaction centre is a side chain carboxylate (from an Asp/Glu residue typically situated on helix 6) that is within hydrogen bonding distance of the donor sugar hydroxyls and a positively charged side chain. The side chain of this Arg/Lys residue is in turn hydrogen bonded to a side chain carboxylate, derived from a residue of the D X D motif, which is in turn coordinated to the bound divalent cation. Interestingly, in the case of glycogenin, the side chain from this (3-face residue (Asp 163) is the most suitably situated group to act as the catalytic nucleophile in a double displacement mechanism, indicating how subtle differences in the mode of donor sugar binding can influence our interpretation of mechanism when based solely on structural information. In most cases, a side chain or main chain amide is most suitably positioned in an appropriate trajectory and at a reasonable distance from the anomeric reaction centre to play the role of the catalytic nucleophile in a putative double displacement mechanism. A n exception to this observation is found with family GT6 enzymes that have a side chain carboxylate at this position. This apparent lack of conserved active site architecture amongst retaining G T - A glycosyltransferases, all of which appear to have diverged from a common ancestor, is in stark contrast to what has been observed for the analogous  42  retaining glycosidases.  In those cases, for which strong evidence has been provided  supporting a double displacement mechanism, highly conserved pairs of carboxylates separated by a defined distance are situated within the active site with one being clearly positioned to play the role of catalytic nucleophile. A vast multitude of observed threedimensional folds all having the same disposition of catalytic active site carboxyl groups indicates a convergence in mechanism amongst unrelated retaining glycosidases. By comparison,  i f all retaining  glycosyltransferases  utilize  the  analogous  double  displacement mechanism, it would seem likely that during the course of divergent evolution the most stringently conserved active site feature would have been the relative positioning of the most important component of the catalytic machinery. Based on the available crystal structures, the conservation of this active site feature appears to be absent. This observation may very well suggest that retaining glycosyltransferases utilize a mechanism that differs from that of retaining glycosidases.  1.3.5.5. Retaining GT-B Glycosyltransferases (Clan IV) To date, structural and mechanistic studies of eleven representatives from five families of retaining glycosyltransferases  possessing the GT-B fold have been  undertaken. As was the case with the inverting glycosyltransferases possessing a GT-B fold, these enzymes also use a metal ion-independent method for facilitating leaving group departure. Crystal structures with substrates or inhibitors bound within the donor sugar site reveal the conserved presence of two active site functional groups on the P-face of the donor substrate.  43  Members of family GT35 catalyse the phosphorolysis of a(l,4)-linked glucans leading to the production of a-D-glucose-1-phosphate  (GlcIP), which is in turn  isomerised to glucose-6-phosphate and fed into the glycolytic pathway.  Although the  equilibrium constant favours the synthetic direction of the reaction, under physiological conditions the reaction is driven in the disfavoured phosphorolysis direction by the high cellular levels of phosphate and the metabolic flux of glucose-1-phosphate into the glycolytic pathway.  The enzymes of this family play an essential role in energy  storage/mobilization in organisms ranging from bacteria to mammals. This family of enzymes has a noteworthy history of investigation, with a Nobel prize being awarded to Carl and Gerty Cori in 1947 for the discovery of rabbit muscle glycogen phosphorylase (Green et al., 1942). Given the key role that these enzymes play in metabolism, it is not surprising that the activities of the eukaryotic enzymes are found to be under strict allosteric regulation by covalent modification (serine phosphorylation) and by the binding of  small molecule  activators  (e.g.  A M P ) and  inhibitors  (e.g.  glucose  and  caffeine)(Madsen, 1986). In contrast, the activity of the bacterial enzymes is controlled at the level of expression (Raibaud and Schwartz, 1984).  Because of their potential  application as targets of therapeutics in the treatment of type II diabetes, the development of inhibitors of human liver glycogen phosphorylase is an active area of research (for example (Rath et al., 2000a). The most intensively studied member of family GT35 is that of rabbit muscle glycogen phosphorylase (rmGP), with over 30 structural investigations having been reported to date (for example see (Mitchell et al., 1996; Watson et al., 1994). However, despite a long history of exhaustive and laborious research with rmGP, a detailed  44  understanding of the catalytic mechanism remains elusive. Crystal structures are also available for bacterial maltodextrin phosphorylase (MalP) (Geremia et al., 2002; O'Reilly et al., 1999; O'Reilly et al., 1997; Watson et al., 1999); yeast glycogen phosphorylase (Lin et al., 1996); human muscle glycogen phosphorylase (Lukacs et al., 2006); and human liver glycogen phosphorylase (Ekstrom et al., 2002; Klabunde et al., 2005; Rath et al., 2000b; Wright et al., 2005). The enzymes of this family are unique amongst the glycosyltransferases in that they contain a pyridoxal phosphate (PLP) group covalently bound within their active site via a Schiff base to a lysine residue. The exact role of PLP in the mechanism of catalysis has been a topic of heated debate. A proposed role as an acid catalyst was disfavoured by pH-dependence studies using PLP analogues with altered electronic demands at the phosphate position (Stirtan and Withers, 1996; Withers et al., 1982). The current view is that PLP acts as a surrogate for the nucleoside monophosphate portion of a nucleoside diphosphate  during  the  evolution  from  an  NDP  sugar-dependent  GT-B  glycosyltransferase to a GT-B glycosyltransferase-like phosphorylase.  This notion is  supported  with synthetic  by the fact that glycogen phosphorylase  reconstituted  "pyridoxaldiphosphoglucose" is able to regio- and stereo-specifically transfer a glucose residue to the non-reducing end of a glycogen acceptor (Withers et al., 1981). A ternary complex three-dimensional X-ray crystal structure of rmGP, with the transition state analogue nojirimycin and phosphate bound within the active site, revealed the main chain amide of His377 to be the functional group most suitably positioned to play the role of catalytic nucleophile (Mitchell et al., 1996). More recently, soaking of (Glc)4 or (Glc)5 malto-oligosaccharides into crystals of E. coli MalP with bound GlcP  45  resulted in the production o f ternary complex structures with both transfer product and phosphate bound within the active site (Geremia et al., 2002)(Figure 1.1 OA). L i k e r m G P , the main chain amide o f His345 (corresponding to His377 in rmGP) was found to be the most suitably positioned candidate to act as the catalytic nucleophile (Figure 1.1 OB). The formation o f a hydrogen bond between the side chain imizadole o f H i s 345 and the C 6 hydroxyl o f what would have been the glucose donor sugar results in the closure o f a flexible loop region that makes up a significant portion o f the maltose  acceptor  recognition site (Figure 1.1 OA). The only other suitable functional group positioned on the p-face o f the donor substrate to act as a catalytic nucleophile is the side chain carboxylate o f Asp307 (Figure 1.1 OA).  However, this carboxylate is situated within  hydrogen bonding distance o f the C 2 - and C3-hydroxyls o f the acceptor glucose residue and does therefore most likely serve to recognize and orient the incoming acceptor.  A  conformational change away from the position found in the observed structure would have to occur for this residue to play a nucleophilic role. The ternary complex structure with bound G l c P substrate is virtually identical to the product complex shown in Figure 1.9A with a T R I S molecule acting as an acceptor surrogate making similar interactions with the side chain o f Asp307. This indicates that i f such a conformational change were to occur it would have to happen during the course o f catalysis. The bound phosphate product is located within hydrogen bonding distances o f the positively charged side chains o f Arg534 and Lys539, indicating a role for these residues in stabilizing the departing phosphate leaving group (Figure 1.1 OA).  46  Figure  1.10.  Comparison of the active sites of several retaining glycosyltransferases  from Clan IV. Distances are indicated in Angstroms.  Three-dimensional  X-ray  crystal  structures  with  bound  intact  donor  substrates/analogues are also available for retaining GT-B fold glycosyltransferases from families GT4, GT20 and GT 72. The structure of E. coli oc-(l,3) glucosyltransferase WaaG from family GT4, involved in lipopolysaccharide biosynthesis, was solved with bound non-reactive donor sugar substrate analogue UDP 2F Glc (Figure 1.1 OB) (Martinez-Fleites et al., 2006). Similarly, structures of E. coli OtsA from family GT20,  47  responsible for the biosynthesis o f the very interesting stress response molecule a , a - l , l trehalose-6-phosphate, were obtained with either U D P 2F G l c or U D P G l c (Figure 1.10C) bound (Gibson et al., 2004). GT72,  the  retaining  Finally, ternary complex structures o f A G T from family  D N A modifying  glucosyltransferase  counterpart  from  T4  bacteriophage o f inverting B G T (described in section 1.3.5.3.), were obtained with U D P G l c and various D N A acceptor substrates bound (Lariviere et al., 2005)(Figure 1.1 OD). L i k e B G T , A G T was shown to use a "base-flipping" mechanism to active D N A acceptors.  Crystal structures o f bacterial (Buschiazzo et al., 2004)  and  archaeal  (Horcajada et al., 2006) glycogen phosphorylases from family G T 5 are also available. However, the absence o f the sugar moiety o f the donor substrate limits their utility i n identifying candidate nucleophilic residues. A comparison o f these structures, along with those o f the r m G P and M a l P ternary complexes, facilitates an initial identification o f several conserved active site structural features amongst G T - B fold glycosyltranferases. O n the p-face o f the bound donor sugar substrate, a main chain amide is most suitably positioned, at an appropriate distance and trajectory, to act as the catalytic nucleophile in a double displacement mechanism.  The  conserved presence o f hydrogen bonding partners, donated from main chain carbonyl or side chain amide groups, in proximity to the main chain N H group could help facilitate this catalytic role. The only other candidate for the role o f nucleophile is that o f the side chain carboxylate o f a structurally conserved aspartate (Asp307, AsplOO, A s p l 3 0 , A s p l 15 i n M a l P , WaaG, OtsA and A G T respectively, Figure 1.10) situated ~6 A from the anomeric reaction centre.  The product complex structure o f M a l P suggests that this  residue is more likely involved in acceptor substrate recognition and orientation (Figure  48  1.1 OA).  The plausibility o f a nucleophilic role for this residue is further limited by  mutagenesis investigations with A G T , in which it was found that the D 1 1 5 A displayed - 1 0 % residual transferase activity (Lariviere et al., 2005). Another conserved feature on the P-face o f the donor is the formation of a hydrogen bond between the donor sugar C 6 hydroxyl and, with the exception of W a a G i n which case A s p 19 is the bonding partner, the side chain imidazole of the H i s residue whose main chain amide is positioned to act as a nucleophile (His354/377, His 154 and H i s l l 4  in MalP/rmGP,  OtsA and A G T  respectively, Figure 1.10). A s mentioned, the formation o f this interaction is believed to cause the closure o f a flexible loop leading to the complete formation o f the acceptor recognition site. O n the a-face of the donor substrate, a structurally conserved Arg/Lys pair o f positively charged side chains is present within hydrogen bonding distance o f the phosphate leaving group, poised to replace the role of a divalent cation i n G T - A fold enzymes, electrostatically stabilizing leaving group departure (Figure 1.10). The ternary complex structure o f M a l P with bound products suggests that the substrate/product phosphate group is most suitably positioned to play the role o f the base catalyst that activates the functional group o f the incoming acceptor (Figure 1.1 OA).  49  1.3.5.6. An Alternative "S i-Like" Mechanism N  Albeit rather rare, reactions involving a single nucleophilic displacement step can lead to the formation of a product in which the stereochemistry of the reaction centre is completely retained.  This is a special case of the S N I mechanism that involves the  formation of discrete ion pair intermediates, with lifetimes longer than a molecular vibration, which can collapse to give back the starting material or a product. This unique form of S N I reaction, termed S>ji wherein the " i " indicates internal return, was first described to account for the observed products and the nature of the reaction involved with the decomposition of alkyl chlorosulphites (Cowdrey, 1937; Hughes, 1941). Support for this proposed mechanism was later provided by studies showing that the decompositions of alkyl chlorosulphites were in fact, under certain conditions, first order reactions that lead to the formation of alkyl halides with completely retained stereochemistry (Lewis and Boozer, 1952). The observed stereochemical outcome is proposed to result from a multistep mechanism involving the formation of two discrete intermediate species of intimately interacting ion pairs (Figure 1.11). Multiple studies have since been reported providing support for the formation of ionic species in this S N I process (Boozer and Lewis, 1953; Cram, 1953; Schreiner et a l , 1993).  The unique feature of the S ^ mechanism that  accounts for the exclusive retention of stereochemistry is the fact that the leaving group undergoes decomposition leading to the production of a nucleophile that is held as an ion pair on the same face as the leaving group. Decomposition of the initial intermediate species and attack by the formed nucleophile occurs at a rate that exceeds that of solvent and ion pair reorganisation required for nucleophilic attack from the back face.  The  50  formation o f intimate ion pairs makes the rate o f attack by a nucleophile on the back face slow, as the effective concentration o f the formed nucleophile would be low.  Ion Pair Intermediate  Figure 1.11.  Ion Pair Intermediate  The S i mechanism for the decomposition o f alkyl chlorosulphites is a N  special case o f an S N I process that leads to complete retention o f stereochemistry i n the product.  In light o f structural, mutagenesis  and mechanistic results  alternative mechanism to the double displacement, termed  with L g t C ,  S ^ - l i k e , was  an  proposed  (Persson et al., 2001). This was described as involving a single transition state i n which attack by the incoming nucleophile o f the acceptor and departure o f the leaving group o f the donor occur on the same face, contradicting one o f the most classic tenets o f physical organic chemistry (Figure 1.12). The single exploded four-centre transition state for such a process was proposed to be cyclic and late with highly developed oxocarbenium ion character.  It must be noted that this four-centre transition state is formally forbidden by  the rules o f Woodward and Hoffman (Moss et al., 2007; Schreiner et al., 1994; Woodward and Hoffman, 1969). In such a mechanism, the enzyme acts as a scaffold that precisely orients substrates i n close proximity, decreasing the energy o f the transition state by stabilizing the oxocarbenium ion-like species and activating the leaving group. To prevent antibonding interactions and to allow for the development o f density in the a * orbital between the anomeric carbon and the leaving group o f the donor, departure o f the  51  leaving group and front side attack must occur in an asynchronous fashion.  Chemical  precedent for this type o f a mechanism comes from detailed conformational studies o f the solvolysis o f glucose derivatives in mixtures o f ethanol and trifluoroethanol by Sinnott and Jencks (Sinnott and Jencks, 1980). They suggest that a sufficiently "open" transition state can allow for nucleophilic push on the reaction centre by the entering solvent molecule on either the opposite or the same face as the leaving group, thereby directly challenging the classical Ingold view o f S 2 displacement reactions involving Walden N  inversion.  Figure 1.12.  Proposed "SNi-like" mechanism for retaining glycosyltransferases and  glycogen phosphorylase involving a direct "front-side S 2 - l i k e " displacement proceeding N  through a highly dissociative oxocarbenium ion-like transition state resulting in retention o f anomeric configuration.  The  SNi-like  mechanism  had  previously  been  proposed  for  glycogen  phosphorylase (Klein et al., 1986) and has since been proposed for the structurally defined retaining transferases Kre2 (Lobsanov et al., 2004), T o x B (Reinert et al., 2005), Extl2 (Pedersen et al., 2003), M g s (Flint et al., 2005), W a a G (Martinez-Fleites et al., 2006), O t s A (Gibson et al., 2002) and A G T (Sommer et al., 2004) based on the lack o f an appropriately positioned nucleophile within their active sites.  52  1.4.  Aims of the Thesis Despite their undeniable biological significance and potential applicability in the  synthesis of defined oligosaccharide structures, glycosyltransferases are a relatively poorly characterized class of enzymes that have yet to see deserved widespread practical applications: a situation that this thesis seeks to address. The work described in this thesis is divided into two parts.  In the first part, the structural features and mechanisms of  representative inverting and retaining glycosyltransferases are explored. Following these structure/function  investigations, glycosyltransferases  are subjected to engineering  strategies with the goal of increasing their synthetic utility. A  detailed  understanding  of  the  chemical  mechanisms  by  which  glycosyltransferases function not only provides a rational basis for their engineering and application in both the development and synthesis of new classes of therapeutic agents, but also provides insight into the role of convergence in the natural evolution of enzyme function.  Has nature converged on analogous strategies for catalysing reactions in  opposite directions using different unrelated enzymes (i.e. glycoside hydrolysis by glycosidases and glycoside formation by glycosyltransferases) and/or what is the nature of this mechanistic continuum? The kinetic and chemical mechanism of the inverting bifunctional family GT42 sialyltransferase Cst II will be explored guided by protein X-ray crystallography using the techniques of site directed mutagenesis, detailed Michaelis-Menten kinetic analysis, and chemical rescue. The kinetic mechanisms of glycosyltransferases are generally found to be ternary complex ordered B i B i , consistent with the generally observed feature that complete formation of an acceptor binding site requires the ordering of a flexible loop  53  region caused by the binding o f a donor substrate within the nucleotide binding domain. The Afunctional nature o f Cst II, being able to catalyse sialyl transfer reactions using two significantly different acceptor substrates, makes the study o f its kinetic mechanism a particularly  interesting  pursuit  that could  provide insights  into how this  useful  bifunctional activity is achieved. The general mechanism o f both inverting glycosidases and glycosyltransferases is that o f a direct displacement SN2-like pathway.  Such a  mechanism requires an appropriately positioned base catalyst to activate the nucleophile of the incoming acceptor and functional groups that facilitate leaving group departure. The identities o f these catalytic groups w i l l be determined. Following studies o f this inverting enzyme, the chemical mechanisms o f the retaining enzymes LgtC, a 3 G a l T and Kre2 (from families G T 8 , G T 6 and G T 1 5 respectively) w i l l be investigated.  The mechanism o f retaining glycosyltransferases  remains unclear. B y analogy to the retaining glycosidase enzymes, a double displacement mechanism has been proposed. However, to date no definitive evidence in support o f this proposition has been provided and an alternative SNi-like mechanism, with little chemical precedence, has become the generally quoted mechanism o f choice.  The potential for  nucleophilic catalysis amongst this class o f enzyme w i l l be explored by altering the reactivity o f an appropriately positioned enzymatic functional group within the active site of L g t C .  The resulting engineered enzyme in combination with substrate and substrate  analogues w i l l be subjected to E S I - M S and protein X-ray crystallographic analysis.  In  addition, a 3 G a l T , an enzyme which seems to be one o f the most likely to use a double displacement mechanism amongst characterized glycosyltransferases, w i l l be subjected to E S I - M S analysis with the goal o f directly observing the formation o f a covalently bound  54  glycosyl-enzyme intermediate. Finally, Kre2 will be subjected to both ESI-MS analysis and protein X-ray crystallographic analysis with the goal of identifying potential enzymatic catalytic nucleophiles. Stringent substrate specificity, the high cost of nucleotide sugar donor substrates, and the general lack of availability of soluble/active enzymes limit the synthetic utility of glycosyltransferases. To overcome these issues, glycosyltransferases will be subjected to engineering strategies involving either a modification of the protein catalyst or in a novel approach, modification of substrates. In addtion to their potential practical applications, engineering results may provide information about mechanism.  With the goal of  developing a generally applicable screening strategy for directed evolution, the ability to display LgtC on the surface of M l 3 bacteriophage in the context of "water-in-oil" emulsions will be explored.  In addition, a generally applicable rational engineering  approach will be used to change the nucleotide specificity of LgtC.  55  CHAPTER 2  STRUCTURAL AND MECHANISTIC INVESTIGATIONS OF A N INVERTING GLYCOSYLTRANSFERASE  * A version of portions of this chapter has been published: Chiu, C.P.C., Watts, A . G . , Lairson, L . L . , Gilbert, M . , Lim, D., Wakarchuk, W.W., Withers, S.G., and Strynadka, N.C.J. (2004) Stuctural Analysis of the Sialyltransferase Cst II from Campylobacter jejuni in Complex with a Substrate Analog. Nature Structural and Molecular Biology. 11:163.  56  2.1. Summary Inverting glycosyltransferases are generally believed to catalyse group transfer via a direct displacement SN2-like mechanism. The kinetic and catalytic mechanism of the bifunctional sialyltransferase Cst II was investigated. A three-dimensional X-ray crystal structure of the enzyme was obtained in complex with a non-reactive donor substrate analogue. The resulting crystal structure was used to identify putative catalytic residues within the enzyme active site.  Site directed mutagenesis was used to investigate the  catalytic importance of the identified residues, thereby facilitating a detailed account of the reaction mechanism.  2.2. Background Glycoproteins, glycolipids and polysaccharides present on cell surfaces are known to play central roles in multiple biological processes including cell-cell recognition events controlled by receptor-ligand interactions.  Many biologically active glycans are  terminated by the 9-carbon sugar sialic acid, or JV-acetyl-neuraminic acid (NeuAc), that bears a negative charge at physiological p H values and is added to glycoconjugates by a class of glycosyltransferases known as sialyltransferases.  A common feature of  malignant transformation and tumor progression is a change in surface glycosylation profile.  These carbohydrate profile changes represent a significant portion of known  tumor markers. A classic example is the observed increase in cell surface sialic acid content of many animal tumor types and a positive correlation of this change with metastatic potential (Dennis et al., 1982; Ogata et al., 1995; Yogeeswaran and Salk,  57  1981).  Sialic acid is related to cancer pathology for several reasons: sialylated cell  surfaces can either prevent cell-cell interactions through non-specific charge repulsion or may facilitate specific cell adhesion events by binding to adhesion molecules of the selectin or siglec families; terminal sialic acids can conceal underlying oligosaccharide structures and prevent recognition by other lectin-like molecules; sialyltransferase regulatory elements can be the target of specific cell signaling pathways as is seen with the upregulation of ST6Gal I by ras oncogene (Dall'Olio and Chiricolo, 2001). Ganglioside glycosphingolipids constitute the major sialic acid-bearing glycans of vertebrate nerve cells, as well as being the major sialic acid-bearing glycoconjugates in the brain (Vyas and Schnaar, 2001). In the bacterial world, Campylobacter jejuni, a human mucosal pathogen, has been shown to express variable cell surface carbohydrate mimics of gangliosides associated with virulence (Guerry et al., 2002; Penner and Aspinall, 1997). Molecular mimicry between C. jejuni LOS outer core structures and gangliosides is believed to act as a trigger for autoimmune mechanisms in the development of Guillain-Barre syndrome (Endtz et al., 2000).  Compared to their  eukaryotic counterparts, such bacterial enzymes are more convenient systems to study mechanistically since both expression and crystallization are more feasible. A detailed understanding of sialyltransferase structure and mechanism is clearly a topic of intense interest within the glycobiology community. The genes responsible for the biosynthesis of the ganglioside mimics in C. jejuni have been identified (Gilbert et al., 2000; Gilbert et al., 2002). Cst II is a membraneassociated enzyme from family GT42 that possesses an inverting a-2,3-sialyltransferase activity. A sialic acid moiety is transferred to terminal P-galactose-containing acceptor  58  substrates using C M P N e u A c as the donor with net inversion o f anomeric configuration with respect to the donor substrate (Scheme 2.1). Interestingly, multiple versions o f Cst II from different strains o f C. jejuni were identified. Two o f these enzymes, Cst IIo i9 and :  Cst  IIOH4384  from serotypes 0:19 and OH4384 respectively, possess very high sequence  identity (97.3%) and yet the Cst II from serotype OH4384 was found to have bifunctional activity.  Both enzymes possess the a-2,3-sialyltransferase  activity, however,  the  bifunctional enzyme is able to use the initially formed ct-2,3-linked sialic acid as an acceptor substrate in a second transfer step, as shown in Scheme 2.1 (Gilbert et al., 2000). Despite the clear importance o f sialylated glycans in normal and pathological physiologies,  to  date  very  few  structural  or  mechanistic  sialyltransferases from any organism have been undertaken.  investigations  of  This lack o f understanding  results from the fact that members o f this class o f enzyme are generally membrane associated or membrane bound which limits the expression o f soluble protein for characterization. A designed soluble and catalytically active construct o f the bifunctional Cst  IIOH4384,  in which 32 basic and hydrophobic residues are deleted from the predicted C -  terminal membrane association domain have been removed (here after referred to simply as Cst II), was the subject o f the detailed structural and mechanistic investigations described below.  59  Scheme 2.1.  Monofunctional and bifunctional sialyltransferase activities of Cst II  enzymes.  60  2.3. Determining the Kinetic Mechanism of the Bifunctional Inverting cc-(2,3/8) Sialyltransferase Cst IIOH4384 2.3.1. Preliminary Kinetic Analysis The donor substrate used by sialyltransferases is C M P NeuAc.  In order to  determine kinetic parameters by monitoring the release of the C M P leaving group, the enzyme-coupled continuous assay typically used to characterize nucleoside diphosphateutilizing glycosyltransferases had to be adapted. This was accomplished by addition of nucleoside monophosphate kinase (NMPK) and A T P to the typical coupling scheme. Using this coupled assay, glycosyltransferase activity is monitored as a function of C M P release via the consumption of N A D H and the resulting decrease in absorbance at X = 340 nm (Scheme 2.2). Michaelis-Menten parameters were determined for the transferase and hydrolase activities of Cst II (Table 2.1).  Table 2.1.  Apparent Michaelis-Menten kinetic parameters for the transferase and  hydrolase activities of Cst II. Substrate Varied  k (min ) l  cat  C M P NeuAc (hydrolase)  K (mM) m  k t/K ca  (mM" min" ) 1  m  2.0  0.50  4.0  45  0.50  90  40  35  1.1  55  3.5  16  1  3  C M P NeuAc (transferase)  13  Lactose  0  3'-Sialyl Lactose * a b c  -  0  error range in data is between 5-15%. determined in the absence of acceptor substrate determined at a constant lactose concentration of 160 m M . determined at a constant C M P NeuAc concentration of 1 m M .  61  Careful attention was required during kinetic analysis to ensure that rates being determined were in fact those for the transferase activity o f Cst II. A significant rate o f spontaneous background hydrolysis o f C M P N e u A c was observed and subtracted from all enzyme catalysed activities and determined the lower limit o f detection for all assays. This significant rate o f spontaneous donor substrate hydrolysis also limits the maximum concentration o f C M P N e u A c that could be assayed.  Above ~1 m M C M P N e u A c , the  rate o f spontaneous hydrolysis is too great to allow for measurement o f sialyltransferase activity. A s such, when K  m  values are greater than 1 m M , errors are significantly larger  because enzyme activity cannot be measured under truly saturating concentrations o f donor substrate.  OCMP  s i a |  y  COO  Transferase  O ^ / OR  COO  AcHN^  6H  H  0  +  \ ^ °  ATP  CDP  +  ADP H O ^ O  PK NTP  + 2  CMP  H0  CMP  NDP  +  0  3  P O ^  H O ^ O  LDH  NADH  NAD+  H O ^ O  HO'  Scheme 2.2. Enzyme-coupled continuous assay used to measure the kinetic parameters o f sialyltransferases.  62  In addition to catalysing the transfer of sialic acid to an acceptor sugar, Cst II was also found to catalyse the hydrolysis of C M P NeuAc at an appreciable rate, as determined by incubating Cst II with C M P NeuAc in the absence of acceptor (Table 2.1). To ensure that the observed hydrolytic activity was not due to a transfer of sialic acid from C M P NeuAc to an additional molecule of donor, a result that would not be completely unexpected considering the observed bifunctional acceptor specificity, T L C and mass spectroscopic analysis was performed and confirmed that no such transfer product was formed.  The quantification of each of these three possible transformations of C M P  NeuAc proved particularly important during the analysis of several Cst II mutants.  2.3.2. Potential Kinetic Mechanisms In both of the non-hydrolytic reactions catalysed by Cst II, two substrates are converted to two products. As such, Cst II is designated a B i B i enzyme according to the notation developed by Cleland for multisubstrate enzymes (Cleland, 1963).  The  mechanism of a multisubstrate enzyme is designated as either ping pong i f the reaction proceeds with the release of one or more products prior to the association of all substrates or as ternary complex (or sequential) i f all substrates bind to the enzyme prior to the formation of the first product. Sequential reactions are further designated as being either random i f there is no obligatory order of substrate association and product release or as ordered i f the substrates bind to and products are released from the enzyme in a defined order. A special case of the ordered mechanism is called the Theorell-Chance mechanism in which there is an obligatory order of substrate association and product release but without the accumulation of the ternary complex under the reaction conditions.  63  Schematic representations o f these four basic types o f kinetic mechanism are shown i n Figure 2.1 using the notation developed by Cleland.  Figure 2.1.  Cleland notation schematic representations  o f various B i B i kinetic  mechanisms.  ( A ) ping pong, (B) random sequential, (C) ordered sequential, and (D)  Theorell-Chance.  A further distinction o f a kinetic mechanism is made depending on the effect that substrate association and product release have on the overall observed rate o f reaction. Reactions are termed rapid equilibrium, i f the chemical steps are slower than all substrate association and product release steps, or steady state, i f the rates o f substrate association and product release contribute to the overall rate o f reaction.  64  2.3.3. Distinguishing Between a Sequential or Ping Pong Mechanism The rate equations for steady state ping pong and ordered Bi Bi mechanisms are shown in double reciprocal form in equations 2.1 and 2.2. respectively. In the absence of products, the equation for the Theorell-Chance "hit and run" Bi Bi mechanism is identical to that of the steady state ordered Bi Bi mechanism.  K  v  0  V  1 v  °  max  [A]  1  V  o  V  m a x  V  +  K  V  m a x  1 v  +  A m  +  max  [B]  B  [A]  A s  K  B m  K [A] s  K  V  B  max  1  +  A  K  V  m  m  max  +  K  max  (2.1)  +  B m  [B]  +  V  K Km A  B  S  (2.2)  V [A][B] max  K  m  max  [B]  B  +  K  A S  K  B (  (2.3)  V [A][B] max  The steady state random Bi Bi mechanism is exquisitely complicated with 37 additional terms in the denominator compared to the analogous ordered mechanism (Segel, 1975). In many cases for random Bi Bi mechanisms, both substrates and both products are found to be in rapid and independent equilibrium with the enzyme, making the interconversion of E A B to EPQ rate limiting.  In this special case, the mechanism is called rapid  equilibrium random Bi Bi and the rate equation reduces to that shown in equation 2.3 in double reciprocal form.  65  Inspection of the above equations indicates that a ping pong mechanism can be readily distinguished from a sequential mechanism by plotting l / v versus 1/[A] at 0  various constant concentrations of substrate B . For a ping pong mechanism, such a plot will yield a series of straight lines for which the slopes are equal to (K  A m  intercepts are equal to ( K  m  /V  max  [ B ] ) + (1 / V  m a x  IV  m a x  ) and the  ) . Because the slope of each of the lines  is not dependent on the concentration of substrate B, the resulting lines will be a family of parallel lines.  Such a series of parallel lines is diagnostic of a ping pong B i B i  mechanism. In contrast, inspection of equations 2.2 and 2.3 indicates that a plot of l / v  0  versus 1/ [A] at various constant concentrations of substrate B will yield families of straight lines for which the slope of each line is dependent on the concentration of substrate B. Double reciprocal plots for sequential mechanisms yield a family of straight lines that intersect to the left of the l / v axis. 0  To distinguish between a ping pong and a sequential mechanism, initial rates were measured for Cst II with increasing concentrations of C M P NeuAc at various constant concentrations of 3'-sialyl lactose using the enzyme-coupled continuous assay (Figure 2.2). The resulting family of straight lines is clearly not parallel and converges to the left of the l / v axis and approximately on the 1/[CMP NeuAc] axis. A ping pong mechanism 0  for Cst II can therefore be ruled out. For a B i B i mechanism, this data alone cannot be used to distinguish between the various types of sequential mechanisms. It is not surprising that Cst II does not follow a ping pong mechanism. As it is an inverting enzyme, such a mechanism would require the intermediate formation of a stabilized oxocarbenium ion species and release of C M P prior to association and attack  66  by the incoming nucleophilic acceptor. Indeed, under standard aqueous conditions, such a mechanism would result i n minimal transfer to a sugar acceptor.  0  Figure 2.2.  1 2 3 4 1 /[3'-SL](1 / pM)  Double reciprocal and secondary plots o f initial rate data with respect to  C M P N e u A c at various constant concentrations o f 3'-sialyl lactose (3'-SL).  2.3.4. Product Inhibition Studies with Cst II Product inhibition studies can be used to distinguish whether substrates bind to and products are released from a sequential B i B i enzyme in an ordered or random fashion (Cornish-Bowden, 2004; Fromm, 1975; Segel, 1975). For ordered reactions, the patterns o f product inhibition can also identify the order o f substrate association and product release. The product inhibition patterns for several sequential B i B i mechanisms are shown i n Table 2.2.  For a more complete list o f inhibition patterns the reader is  directed to the detailed analysis o f Segel (Segel, 1975).  67  Table 2.2.  Expected product inhibition patterns for select B i B i enzyme mechanisms  (Segel, 1975). Varying A Mechanism  Ping Pong Bi Bi  Steady State Ordered Bi Bi  Steady State Iso Ordered Bi Bi  Steady State Theorell-Chance Rapid Equilibrium Random Bi Bi  Steady State Random Bi Bi  Product Inhibitor  Varying B  Unsat B  SatB  Unsat A  Sat A  P  Mixed  -  Comp  Comp  Q  Comp  Comp  Mixed  -  P  Mixed  Uncomp  Mixed  Q  Comp  Comp  Mixed  P  Mixed  Uncomp  Mixed  Mixed  Q  Mixed  Mixed  Mixed  Uncomp  P  Mixed  _  Comp  Comp  Q  Comp  Comp  P  Comp  Q  Comp  -  Comp  P  Mixed  Mixed  Mixed  Mixed  Q  Mixed  Mixed  Mixed  Mixed  Mixed -  Mixed  -  Comp  _  Unsat - initial rates measured at a constant unsaturating concentration of substrate Sat - initial rates measured at a constant saturating concentration of substrate Comp - Competitive inhibition Uncomp - Uncompetitive inhibition Mixed - Mixed inhibition  In order to determine patterns o f inhibition with respect to the product C M P , because C M P release is the mode o f detecting transferase activity i n the enzyme-coupled continuous assay, an alternative assay was required. A stopped capillary electrophoresis (CE)  based assay was established in which product formation is detected using  fluorescein 3'-sialyl lactose conjugate ( F C H A S E 3 ' - S L ) as the acceptor  substrate.  Representative C E traces for product formation at various concentrations o f C M P N e u A c and C M P are shown i n Figure 2.3. Because the presence o f the fluorescein substituent limits the aqueous solubility o f the conjugated acceptor, initial rate measurements could  68  only be obtained under non-saturating concentrations of the acceptor substrate.  In  addition, because of the instability of C M P NeuAc and resulting difficulty in obtaining donor substrate free of contaminating C M P (-10% contamination following H P L C purification), initial rates could also only be obtained under non saturating concentrations of the donor substrate.  Figure 2.3. Representative C E data used to monitor Cst II sialyltransferase activity. Product formation was determined based on percent conversion of F C H A S E SL acceptor substrate using the peak areas of starting material and product.  Initial rates were  determined using 24 u.g/mL Cst II, various constant concentrations of C M P (0, 100, 250, 500, and 1000 uM), 500 u M F C H A S E SL, and varied concentrations of C M P NeuAc (50, 100,250, and 500 uM).  Using C M P as a product inhibitor o f the Cst II reaction, mixed inhibition patterns were observed with respect to both the donor substrate C M P N e u A c (Figure 2.4A) and the acceptor substrate F C H A S E 3 ' - S L (Figure 2.4B) when the non varied substrate was held at a fixed non saturating concentration. Competitive (Kj ) and uncompetitive (Kj ) C  u  inhibition constants were determined by direct fit o f the data to the equation for mixed inhibition using the program Grafit 4.0 and are shown on the double reciprocal plots. Somewhat surprisingly, these modes o f inhibition are not consistent with either o f the expected steady state ordered or Theorell-Chance B i B i mechanisms (Table 2.1). M i x e d product inhibition patterns for both substrates i n the presence o f fixed non saturating concentrations o f the other substrate are consistent with either a steady state random or an iso ordered B i B i mechanism.  A  [CMP]  B  [ C M P ]  1 /[FC-3'-SL] (1 / UM)  Figure 2.4. Inhibition o f Cst II catalysed sialyl transfer to F C H A S E 3 ' - S L by the product C M P with respect to donor ( A ) and acceptor (B) substrates.  70  2.3.5. Synthesis and Inhibition Studies of a Dead-End Donor Substrate Analogue Inhibition studies using dead end substrate analogues can be used to distinguish between the mechanisms of bireactant enzyme systems (Piszkiewicz, 1977; Segel, 1975). To obtain further insight into the kinetic mechanism of Cst II, the dead end donor substrate analogue C M P 3F NeuAc was synthesized and subjected to detailed kinetic characterization. As described in chapter 1, introduction of the highly electronegative fluorine atom adjacent to the anomeric reaction centre inductively destabilizes oxocarbenium ion-like transition states through which glycosyl group transfer reactions are thought to proceed. Indeed, in the case of glycosyltransferases, N D P 2-fluoro donor sugar analogues have been found to act as competitive inhibitors with respect to the donor substrate and do not act as slow substrates (Ly et al., 2002; Murray et al., 1997). In the case of Cst II, the desired analogue is C M P 3F NeuAc and facile access was obtained in two steps using the enzymes ManNAc aldolase and C M P NeuAc synthetase (Scheme 2.3).  Scheme 2.3. Enzymatic synthesis of the dead end donor substrate analogue C M P 3F NeuAc.  71  Due to limitations in the quantity of available inhibitor at the time, using the method of Dixon, C M P 3F NeuAc was initially reported to act as a competitive inhibitor of Cst II with respect to C M P NeuAc (Chiu et al., 2004). However, subsequent more rigorous analysis clearly indicates that using C M P 3F NeuAc as a dead end substrate analogue leads to mixed inhibition patterns with respect to both C M P NeuAc (Figure 2.5A) and 3'-sialyl lactose (Figure 2.5B) when the non-varied substrate is held at a fixed non-saturating concentration.  Competitive (K, ) and uncompetitive (Kj ) inhibition c  u  constants were determined by direct fit of the data to the equation for mixed inhibition using the program Grafit 4.0 and are shown on the double reciprocal plots.  ao 1 / [CMP NeuAc] (1 /uM)  Figure 2.5.  0.1  0.2  a3  11 [3'-SL] (1 / mM)  Inhibition of Cst II catalysed sialyl transfer to 3'-SL by the dead end  substrate analogue C M P 3F NeuAc with respect to donor (A) and acceptor (B) substrates.  As was observed with C M P product inhibition studies, these patterns of inhibition are not consistent with either the steady state ordered or rapid equilibrium random B i B i kinetic mechanisms.  For both of these mechanisms, competitive inhibition behaviour  would be expected for C M P 3F NeuAc with respect to C M P NeuAc (Piszkiewicz, 1977). However, the observed modes of inhibition for C M P 3F NeuAc are understandable in  72  light of the bifunctional nature of Cst II. A n ability to associate with two distinct stable forms of the enzyme would make C M P 3F NeuAc a mixed inhibitor with respect to C M P NeuAc. This mode of inhibition is therefore consistent with a kinetic mechanism that is more complex than that of the steady state ordered B i B i mechanism and, although by no means conclusive in itself, supports an iso ordered B i B i mechanism.  2.3.6. Conclusions on the Kinetic Mechanism of Cst II Of the few glycosyltransferases that have been subjected to detailed kinetic analysis, to date the vast mojority have been shown to catalyse their respective reactions using an ordered B i B i ((Bendiak and Schachter, 1987) - GnTII; (Kearns et al., 1991) XylT; (Murray et al., 1996) - FucT V ; (Chen et al., 2002) - MurG; (Ly et al., 2002) LgtC; (Zhang et al., 2006b) - TagA; (Morrison and Ebner, 1971) - Lactose Synthase; (Boix et al., 2002) - oc3GalT) or Theorell-Chance ((Kamath et al., 1999) - GTB) mechanism involving the obligate binding of the nucleotide sugar donor substrate prior to the association of the acceptor.  This finding has been rationalized on the basis of  obtained three-dimensional X-ray crystal structures of glycosyltransferases with bound intact donor substrates.  Upon binding the donor substrate, a conformational change  occurs in which a flexible loop closes over the donor substrate and forms a significant portion of the acceptor substrate binding site, consistent with the observed ordered kinetic mechanisms. Interestingly, oleandomycin glucosyltransferase has also been shown to use an ordered B i B i mechanism, but in this case the acceptor substrate associates prior to the donor (Quiros and Salas, 1995). In addition, a polypeptide GalNAc transferase (Wragg et al., 1995) and O-GlcNAc transferase (Kreppel and Hart, 1999) were found to use rapid  73  equilibrium random B i B i mechanisms. These two enzymes both use peptide acceptor substrates. Binding of substrates and release of products does not occur in a defined order but does contribute to the overall observed rate of reaction in the steady state random B i B i mechanism.  In contrast to the frequently observed rapid equilibrium random B i B i  mechanism, few examples have been reported for the steady state random mechanism. The most frequently quoted example, yeast hexokinase, required years of argument and reports of seemingly contrasting data before an authoritative study by Rudolph and Fromm closed the case (Rudolph and Fromm, 1971). This mechanism is very difficult to unambiguously identify for several reasons.  The rate equation is painfully complex,  making the derivation of expected modes of inhibition very difficult to predict. The intial rate data for this mechanism should in theory be non-hyperbolic, although there are several reasons why obtained data can appear hyperbolic (Segel, 1975), which can lead to erroneous interpretations of data. Unfortunately, for reasons described above, in the case of Cst II initial rates cannot be measured under truly saturating concentrations of donor substrate and thus curve fitting to non-hyperbolic equations cannot be used to support this mechanism. A n additional complexity of the steady state random mechanism comes from the fact that certain portions of the mechanism can be in rapid equilibrium while others contribute to the overall rate, further complicating the derivation of the actual rate equation and the prediction of modes of inhibition. The steady state iso ordered B i B i mechanism requires an obligate order of substrate binding and product release but also involves an isomerization between two stable enzyme forms that contributes to the overall observed rate of reaction (Segel,  74  1975). In such a mechanism, product Q will be a mixed inhibitor with respect to substrate A , as was found to be the case with Cst II, i f the two can bind to different stable forms of the enzyme (Figure 2.6).  Although Cleland predicted such a mechanism and the  anticipated modes of inhibition in 1963 (Cleland, 1963), an example with experimental evidence was not reported until 1993 for Amaranth leaf betain aldehyde dehydrogenase (Valenzuelasoto and Munozclares, 1993).  Because Cst II is a bifiinctional enzyme  catalysing two distinct sialyl transfer reactions, it seems quite reasonable that the enzyme could exist in two distinct stable forms. It is possible that, while monitoring transfer to 3'-sialyl lactose, C M P binds to the form of the enzyme that catalyses transfer to lactose. This would give rise to a mixed inhibition pattern with respect to C M P NeuAc i f release of C M P and formation of the other stable enzyme conformation are partially rate limiting. The existence of two stable enzyme forms of Cst II is also consistent with the fact that C M P 3F NeuAc acts as a mixed inhibitor with respect to C M P NeuAc.  CMP NeuAc A  E'  E  CMP Lactose  SL  B  EA  P  EAB ^  EPQ  CMP  NeuAc  Q  EQ  A  E ^  E'  SL  SSL  C  E'A  R  E'AC =^  E'QR  CMP Q  E'Q  E"  E  Figure 2.6. Schematic representation of the steady state iso ordered B i B i mechanism of Cst II. E and E ' represent stable enzyme forms that catalyse sialyl transfer to lactose and 3'-sialyl lactose (SL) respectively. A represents C M P NeuAc, B represents Lactose and C represents 3'-sialyl lactose (as substrate). P represents 3'-sialyl lactose (as product), Q represents C M P and R represents the disialylated product (SSL). Product Q can associate with either stable forms of the enzyme E and E ' which would make it a mixed inhibitor with respect to A . A dead end analogue of A could associate with both stable forms of the enzyme E and E ' which would make it a mixed inhibitor with respect to substrate A .  75  Unfortunately, the described kinetic data alone cannot distinguish between the steady state random and steady state iso ordered B i B i mechanisms with complete certainty. However, in light of the bifunctional nature of Cst II and the precedence for the utilization of ordered mechanisms by glycosyltransferases, the iso ordered mechanism seems the most reasonable.  Evidence disfavoring the steady state random mechanism  could come from structural investigations indicating the occurrence of a conformational change resulting in the formation of the acceptor binding site upon donor substrate association.  2.3.7. Calculating Kinetic Parameters for Cst II Assuming an iso ordered B i B i mechanism, kinetic parameters for Cst II can be calculated using the data presented in Figure 2.2. Isomerization between stable enzyme forms results in the appearance of additional terms containing the concentrations of products P and Q in the denominator of the initial velocity equation that leads to changes in product inhibition patterns.  However, in the absence of products the intial velocity  equation of the steady state ordered and iso ordered B i B i mechanisms are identical (Segel, 1975).  The initial rate equation of these mechanisms (equation 2.4) can be  rearranged to take the general form of the Michaelis-Menten equation (equation 2.5). If the concentration of substrate A is varied while that of substrate B is held at a constant concentration, terms that do not contain A are constants. Michaelis-Menten constants V  a p p m a x  and K  a p p w  Values of the apparent  obtained at a given fixed concentration of  substrate B are therefore functions of the concentration of B (equations 2.6 and 2.7, respectively). These terms can be assembled to take the form of equation 2.8. Because  76  equations 2.6 and 2.8 have the general form of the Michaelis-Menten equation, secondary plots of V  a p p m a x  or V  a p p m a x  / K  against B will be rectangular hyperbolas through the  a p p O T  origin that can be analysed by standard methods to obtain the four parameters V K s , and K . A  B  m  m a x  ,  K  A OT  ,  Because K s does not appear in equation 2.4, this parameter cannot be B  obtained from the analysis of secondary plots.  V ax[A][B] m  K K A  s  (2.4)  + K [A]  B  + K [B] + [A][B]  B  r o  A  m  /  V  m  [B]  max  \ [A]  K +[B] B  m  (2.5) K  A s  K  + K [B]  B  \  A  m  m  + K  [A]  + [B]  B m  V x[B] ma  (2.6)  K +[B] B  m  K Km  A s  K  +  B m  K [B] A  m  app =  (2.7) K  B m  + [B]  (Vmax/OtB] (K K A  s  B m  (2.8)  / K ) + [B] A  m  Direct fit of the data in Figure 2.2 to rectangular hyperbola equations 2.6 and 2.8, shown in double reciprocal format in Figure 2.2, was used to calculate the kinetic parameters shown in Table 2.3. It is worth noting that the actual kinetic parameters calculated from this extensive kinetic analysis of Cst II are quite similar to the apparent values shown in Table 2.1.  Table 2.3.  Michaelis-Menten kinetic parameters for Cst II catalysed sialyl transfer to 3'-  sialyl lactose. Values in parentheses represent the standard error.  Constant  Value  K m (CMP NeuAc)  0.40 (0.078) mM  K  0.38 (0.064) mM  j a  K / n (SL)  kcat  3.7 (0.40) mM 40 (1.4) min  1  2.4. Three Dimensional X-Ray Crystallographic Analysis of Cst II in Complex with a Stable Donor Substrate Analogue In collaboration with Ms. Cecilia Chiu in Prof. Natalie Strynadka's laboratory at U B C , a co-crystal structure of the soluble truncated version of Cst II was obtained with either C M P or an intact molecule of C M P 3F NeuAc bound within the active site. Ms. Chiu found this soluble form of the enzyme to form a tetramer in solution by static light scattering as was also observed in the resulting crystal structure (Figure 2.7 A & B ) . The individual monomers of Cst II were found to be organized into two domains (Figure 7.6 C&D). The first domain, consisting of residues 1 to 154 and 189 to 259, is composed of a mixed a/p fold in which a central parallel seven-stranded twisted beta sheet (topology P 8 , P 7 , p i , P 2 , P 4 , P 5 , P6) is surrounded by five helices on one side (D, E , F, I and J) and four helices (A, B, C and K ) on the other. A Rossman fold nucleotide-binding domain is  78  formed by helices F, I and part of the p-sheet (p8, P7, p i , p2, and p4). Suggesting that protein oligomerization has no direct role in catalysis, the active site of Cst II is not located at the interface between monomers. The second domain, consisting of residues 155 to 188, is composed of a coil extending from P7 followed by helices G and H and finally a coil that connects back to helix I. As discussed in chapter 1, this architecture is that of a G T - A fold and Cst II therefore belongs to glycosyltransferase Clan I. It is however worth noting that significant differences from other members of the G T A group and Cst II are observed in terms of secondary structure connectivity and the absence of a conserved D X D motif.  79  Figure 2.7. Overall structure of Cst II. Top down (A) and side (B) views of the overall tetrameric arrangement of Cst II. Individual monomers are colored differently and the active site location is indicated by the presence of the stick representation of C M P 3F NeuAc. Top down (C) and side (D) views of an individual monomer of Cst II with C M P 3F NeuAc bound within the active site. The two domains of the monomer are colored differently in blue and red.  A comparison of the structures of Cst II with either C M P or C M P 3F NeuAc bound reveals little change in the overall structure of the enzyme (-0.2 A r.m.s deviation for 246 C a atoms). However, the presence of the sialic acid moiety leads to an ordering  80  of residues 175-187 and the formation of an ordered loop that forms a lid that effectively buries the donor substrate. Formation of this lid presumably protects the labile donor substrate from bulk solvent and forms a substantial portion of the acceptor binding site. Formation of the acceptor binding site following binding of the donor substrate is consistent  with  previously obtained  glycosyltransferase  structures  and  provides  substantial support for the proposed iso ordered B i B i mechanism over the alternative random mechanism.  2.5. Investigating the Role of Putative Catalytic Residues To date, all structural and kinetic evidence for inverting glycosyltransferases supports a direct displacement mechanism proceeding through an oxocarbenium ion-like transition state facilitated by base assistance.  Cst II catalyses two such direct  displacements. In the first reaction the 3-hydroxyl of galactose acts as the nucleophile, directly attacking the anomeric centre (C2) of C M P NeuAc to form a NeuAca-2,3-Gal linkage. In the second reaction the 8-hydroxyl from the sialyl moiety of the acceptor acts as the nucleophile, attacking the anomeric centre of a second C M P NeuAc bound in the active site (Scheme 2.1). Key questions about this mechanism then are the identities of the base catalyst that deprotonates the attacking nucleophile and of any acid catalyst that assists C M P departure.  The presence of a carboxylate group at the anomeric centre of  sialic acid alters the electronic and steric nature of the reaction centre and is likely to impose differences in the mechanism and the mode of substrate binding compared to other glycosyltransferases. The carboxylate substituent also provides a potential role in substrate-assisted catalysis.  81  2.5.1 Tyrl56 and Tyrl62 Facilitate CMP Departure Unlike other glycosyltransferases with a G T A fold, bound metal was not observed in the active site of Cst II. In addition, Cst II lacks the ' D X D ' sequence motif found in a wide range of glycosyltransferases that coordinates the divalent cation involved in binding of the nucleotide sugar via interaction with the diphosphate moiety (Ten Hagen et al., 1999; Wiggins and Munro, 1998). Bound metal is generally considered to act as a Lewis acid catalyst that stabilizes the departing nucleoside diphosphate.  Mutagenesis  studies of the conserved Asp residues in the D X D motif of glycosyltransferases from various species indicate that removal of the Asp residues, and hence the coordination of metal ions, completely eliminates the transferase activity (Busch et al., 1998; Ten Hagen et al., 1999; Wiggins and Munro, 1998). Kinetic studies with Cst II indicate that although magnesium and manganese enhance the activity somewhat (~2-fold increase for both metal ions), they are not essential for catalysis.  This difference  from other  glycosyltransferases is not surprising given that the donor substrate in this case is a nucleoside monophosphate sugar (CMP NeuAc). Cst II therefore uses other means to activate/stabilize the leaving nucleotide. Inspection of the structure of Cst II in complex with C M P 3F NeuAc indicates that departure of the phosphate leaving group could be facilitated by the hydrogen bonding interactions of Tyrl56 and Tyrl62 with the non-bridging pro S oxygen (Figure 2.8). In contrast to the alcohol leaving group with pKa «16 released by glycosidases, the second ionization of the monophosphate leaving group has a pKa of around neutrality and discrete acid catalysis is not required for Cst II. In fact, the situation is more  82  Figure 2.8.  Cst II active site highlighting the hydrogen bond interactions between the  side chain hydroxyls of Y156 and Y162 and the departing C M P leaving group. Indicated distances are measured in Angstroms.  equivalent to the second step of a retaining glycosidase wherein the leaving group is an enzymatic carboxylate and a tyrosine residue is often found to play a role in stabilizing the departing leaving group (Gebler et a l , 1995). Mutational analysis of Cst II shows that k  cat  values for the Y156F and Y162F mutants are 75-fold and 360-fold lower than those  for the wild-type enzyme (Table 2.4).  Further, mutation of both of these residues to  phenylalanines results in the complete loss of any enzymatic activity above the rate of spontaneous C M P NeuAc hydrolysis. In the case of the retaining Agrobacterium sp. figlucosidase, Tyr298 plays an equivalent role of interacting with the carboxylate leaving group and mutation to Phe likewise causes a 2500-fold decrease in its k value (Gebler et cat  al., 1995). This metal ion-independent method of facilitating leaving group departure  83  utilized by Cst II is more akin to the mode used by GT-B enzymes, as discussed in section 1.3.4.4., and may therefore indicate a convergence in mechanism.  Table 2.4. Apparent Michaelis-Menten parameters for C M P NeuAc with various Cst II mutants.  Cst II  K  kcat  m  (min' )  (mM)  (mM" min" )  WT  45  0.50  90  WT Hydrolysis"  2.0"  0.50  4.0  R129A  C  <0.20  a  0.25  0.80  H188A  C  <0.20  a  0.30  0.50  1  a  1  1  Y156F  0.60  a  0.20  2.5  Y162F  0.10  a  0.45  0.30  nd  -  Y156/162F  d  <0.040  a  * - error range in data is between 5-20% a - Kinetic parameters were determined at a constant lactose concentration of 160 mM b - For enzyme catalysed hydrolysis kinetic parameters were determined in the absence of acceptor c - No significant activity above that of enzyme catalysed hydrolysis (measured in the absence of acceptor) was observed d - No activity above that of the significant rate of spontaneous background hydrolysis was detected up to an enzyme concentration of 500 ug/mL  2.5.2 Hisl88 Acts as the Catalytic Base As mentioned above, a possible candidate for base catalyst is the carboxylate moiety of the substrate itself. This contention is supported by the complete absence of any basic amino acid side chains in the immediate vicinity of the carboxylate, in contrast to the three arginines seen in sialidases (Amaya et a l , 2003; Burmeister et al., 1992; Crennell et al., 2000; Luo et al., 1998).  The observed in-plane conformation of the 84  carboxylate is not consistent with a role as base catalyst, as such a role would require that the carboxylate be bound perpendicular to the ring plane (Figure 2.9). While it is still possible that the conformation would change upon formation of a ternary complex with the acceptor sugar, it is interesting to note that this near-planar conformation of the carboxylate is what has been predicted by ab initio calculations and supported by kinetic isotope effect experiments to exist during the transition state for the formation of a sialyloxocarbenium ion during the spontaneous hydrolysis of C M P NeuAc (Horenstein, 1997).  Additionally, K I E experiments have cast serious doubt on the ability of the  carboxylate of CMP-NeuAc to function as an intramolecular base catalyst during spontaneous hydrolysis (Horenstein and Bruner, 1998). In glycosidases, including sialidases, and inverting glycosyltransferases, excluding certain GT-B fold family GT1 inverting enzymes that are also thought to use the imidazole of a His residue, the role of base is clearly played by the carboxylate group of a Glu or Asp residue. However, in the three-dimensional X-ray crystal structure of Cst II with bound C M P 3F NeuAc there are no suitably disposed acidic groups in the active site. The closest is that of Glu57 which is 14 A away.  Side chains situated close to the  anomeric carbon are those of Asn31 (3.1 A), Asn51 (4.0 A), Serl32 (3.0 A) and Hisl88 (4.8 A). Hisl88 is the only feasible candidate and is situated appropriately on the alpha face of the sugar and adjacent to an open cleft in the enzyme that would be the most obvious site to bind the acceptor sugar (Figure 2.9). The imidazole side chain is suitably positioned to act as the base in catalysis, deprotonating the incoming hydroxyl group of the acceptor during nucleophilic attack at the anomeric carbon of the donor sugar. Precedence for the role of His 188 as the base catalyst derives from the fact that side chain  85  imidazoles are found to play this role in multiple other classes of enzymes (for example (Legler et al., 2002; Whiteson et al., 2007).  In addition, studies on several human  sialyltransferases (ST3Gal I, ST8Sia II and ST8 Sia IV) have revealed the identity of an invariant His residue that is essential for enzymatic activity (Close et al., 2003; Jeanneau et al., 2004).  Finally, the recently described X-ray crystal structures  of the  sialyltransferases Cst I (Chiu et al., 2007) and PmSTl (Ni et al., 2007) reveal the existence of active site His residues with near identical relative positioning to that of Hisl88inCstII.  Figure 2.9. Cst II active site highlighting the side chain imidazole of His 188 suitably positioned on the a-face of the donor substrate to play the role of base catalyst. Indicated distances are measured in Angstroms.  Optimal Cst II transferase activity occurs at approximately pH = 8 (Figure 2.10). At this pH, the imidazole of His 188 would be expected to be deprotonated, i f it has a  86  "normal" histidine pKa, as required for a base catalyst.  Indeed, the experimentally  determined pKa of Cst II was found to be 6.45, consistent with the typical pKa of a side chain imidazole (pKa ~7), providing support for a catalytic role for Hisl88. Mutation of His 188 to alanine results in a loss of all detectable transferase activity indicating a key catalytic role in Cst II activity (Table 2.4).  2 -,  6  8  10  PH  Figure 2.10. pH dependence of k /Km for Cst II. The pKa value was determined by best cal  fit of the data to a single titration model in Grafit 4.0. The value in parenthesis represents the standard error.  Described as chemical rescue, reactivation of inactive mutants by the addition of a nucleophilic small molecule is a well established mechanistic probe for retaining glycosidases (Macleod et al., 1994; Viladot et al., 1998; Zechel and Withers, 2000). Exogenous addition of a small nucleophile (e.g. azide or formate) to a mutant enzyme, in which the catalytic nucleophile or the acid/base catalyst has been replaced by alanine, restores the activity of the mutant as a function of small molecule concentration and results in the formation of a covalent glycosidic adduct (Figure 2.11). Rescue indicates  87  that the putative catalytic side chain is suitably positioned within the active site during the course of catalysis to play the proposed catalytic role.  Figure 2 . 1 1 . Chemical rescue of nucleophile (A) and acid/base (B) mutant glycosidases with small molecules. When presented with a small exogenous nucleophile (e.g. azide), the nucleophile mutant of a (^-retaining glycosidase will catalyse a group transfer reaction leading to a product of inverted anomeric configuration. In the case of the acid/base mutant of a P-retaining glycosidase, when deglycosylation is rate limiting, the presence of an exogenous nucleophile recovers enzymatic activity and leads to the formation of a covalent adduct with overall retention of anomeric configuration.  To further investigate the role of His 188 as the base catalyst of Cst II, chemical rescue of the HI88A mutant was attempted using sodium azide and sodium formate. Addition of either formate (Figure 2.12A) or azide (Figure 2.12C) to the HI88A mutant resulted in significant reactivation of activity as determined using the N M P K / P K / L D H enzyme-coupled assay. Rates in the presence of acceptor were virtually identical to those in the absence of acceptor at all concentrations of exogenous nucleophile for the HI88A mutant. This indicates that the observed activated reaction is not that of the transferase activity of the enzyme and presumably corresponds to the formation of a sialic acid derivative.  T L C and ESI-MS was used to confirm the formation of a sialyl azide  88  derivative in the HI88A reaction mixture. Importantly, increasing concentrations of either formate or azide did not result in a significant increase in the activity of wildtype Cst II (Figure 2.12 B and 2.12D, respectively).  200  400  GOO  [HCO2N1] (mM)  [HC0 Na] (mM) 2  230  300  [NaNj] (mM)  Figure 2.12.  Chemical rescue of Cst II H188A.  [NaN3] (mM)  Rates were determined using 500  Hg/mL of HI 88 A mutant or 25 p.g/mL of WT enzyme, 500 \iM C M P NeuAc and standard Cst II assay conditions. (A) Activity of Cst II HI88A in the presence of increasing concentrations of formate.  (B) Activity of WT Cst II in the presence of increasing  concentrations of formate. (C) Activity of Cst II HI88A in the presence of increasing concentrations of azide.  (D) Activity of WT Cst II in the presence of increasing  concentrations of azide.  A conserved arginine residue (Argl29) lies 5.9 A away from Hisl88 and could act to provide an electrostatic shield favouring the neutral deprotonated form of the putative base Hisl88 (Figure 2.9). This residue is also suitably positioned within the active site to  89  potentially play a role in orienting the incoming acceptor substrate. shows that mutation of Arg 129 to alanine decreases the k  cat  Kinetic analysis  value by 210-fold and,  similarly to mutants of His 188, that the residual activity results from enzyme-catalysed hydrolysis (Table 2.4).  2.6. Concluding Remarks A combination of kinetic, chemical, structural and mutagenesis techniques were used to investigate the catalytic activities of the bifunctional sialyltransferase Cst II. Kinetic analysis revealed that the overall rate (k /K ) cat  m  for sialyl transfer to a 3'-sialyl  lactose acceptor substrate is approximately an order of magnitude greater than that for transfer to a lactose acceptor. In addition, the enzyme was found to catalyse substantial hydrolysis of the labile C M P NeuAc donor substrate. The measurement of initial rates under a matrix of donor and acceptor substrate concentrations revealed the enzyme to use a sequential mechanism and facilitated the accurate determination of kinetic parameters. Encouragingly, these kinetic parameters were found to be virtually identical to the apparent parameters determined using the less rigorous analysis used to characterize a number of active site mutants. Inhibition studies using either the product C M P or the dead end donor substrate analogue C M P 3F NeuAc were used to further refine the kinetic mechanism of Cst II and were found to be consistent with either a steady state random or steady state iso ordered B i B i mechanism. The bifunctional nature of Cst II makes the existence of two stable enzyme forms a reasonable supposition and the iso ordered B i B i mechanism the preferred choice. Evidence disfavoring the random mechanism comes from the obtained crystal structure of Cst II with an intact molecule of C M P 3F NeuAc  90  bound within the active site. As has been observed with other glycosyltransferases, upon binding the donor substrate, a conformational change occurs leading to the formation of a lid region over the active site that completes the formation of the acceptor binding site. This finding is consistent with an obligate order of substrate binding and product release, thereby disfavoring the alternative random mechanism. Inspection of the X-ray crystal structure of Cst II with bound donor substrate analogue facilitated the identification of putative catalytic residues. Cst II catalyses sialyl transfer with net inversion of anomeric configuration via a SN2-like direct displacement mechanism.  In such a mechanism, the enzyme catalyses the reaction by facilitating  departure of the leaving group and providing a base catalyst that activates the incoming nucleophile.  In contrast to the majority of glycosyltransferases, which use a divalent  cation to facilitate departure of a nucleotide leaving group, Cst II activity was not found to be metal dependent. Instead, Cst II appears to use two tyrosyl hydroxyl groups to stabilize departure of the C M P leaving group. With the exception of two other recently characterized sialyltransferases and certain members of the GT-B fold family GT1 enzymes  that are  also thought  to use  a side chain imidazole, all inverting  glycosyltransferases investigated to date have a side chain carboxylate that is believed to play the role of the base that activates the incoming nucleophilic substituent of the acceptor substrate. In contrast, in the case of Cst II a side chain carboxyl group was not found to be appropriately located within the active site. Kinetic and chemical rescue analysis of the alanine mutant indicates that in the case of Cst II the side chain imidazole of His 188 plays the role of base catalyst. Interestingly, a metal ion-indepent method of leaving group activation and the use of an active site His residue as the catalytic base are  91  strategies utilized by GT-B fold enzymes.  As such, the finding that Cst II utilizes a  mechanism involving these chemical strategies and yet possesses a G T - A fold indicates a possible convergence in mechanism between these two classes of enzymes.  92  CHAPTER 3  STRUCTURAL AND MECHANISTIC INVESTIGATIONS WITH RETAINING GLYCOSYLTRANSFERASES*  * A version of portions of this chapter has been published: Lairson, L . L . , Chiu, C.P.C., Ly, H.D., He, S., Wakarchuk, W.W., Strynadka, N.C.J., Withers, S.G. (2004) Intermediate Trapping on a Mutant Retaining aGalactosyltransferase Identifies an Unexpected Aspartate Residue. Journal of Biological Chemistry. 279: 28339.  93  3.1. Summary Despite  decades  of  exhaustive  glycosyltransferases remains ambiguous.  research,  the  mechanism  of retaining  In order to investigate the potential for  enzymatic nucleophilic catalysis in a double displacement mechanism, the active site of the retaining galactosyl transferase LgtC was engineered in an attempt to drive the reaction along such a coordinate. In addition, ESI-MS analysis was used to try to observe a putative glycosyl-enzyme intermediate with the galactosyltransferase a3GalT and the mannosyltransferase Kre2.  Finally, the non-reactive substrate analogue G D P 2F Man  was synthesized and used to obtain a three-dimensional X-ray co-crystal structure of Kre2.  3.2. Investigating the Catalytic Mechanism of the Retaining a-(l,4) Galactosyltransferase LgtC 3.2.1. Background LgtC (EC 2.4.1.X ) is a family GT8 a-1,4 galactosyltransferase from Neisseria meningitidis that catalyses the transfer of galactose to the 4'-hydroxyl group of terminal lactose-containing acceptor substrates, using UDP Gal as the donor substrate, with overall net retention of anomeric configuration (Scheme 3.1). The Gal-a(l,4)-lactose product of the LgtC reaction is known as the P antigen that terminates P blood group glycolipids in k  humans. Formation of this antigen on the surface of Neisseria meningitidis serves as a biomimicry strategy that allows the pathogen to evade an immune response from the host during infection.  94  Scheme 3.1. Galactosylation reaction catalysed by LgtC.  A ternary complex three-dimensional X-ray crystal structure with both intact donor and acceptor substrate analogues bound is available for LgtC (Persson et al., 2001). Surprisingly, this structure revealed that the only active site side chain suitably positioned to act as the nucleophile in a double displacement mechanism is that of Gin 189 (Figure 3.1).  Although not generally considered a good candidate as an enzymatic nucleophile,  an amide can indeed play this role as is seen with retaining hexosaminidases from families 18, 20 and 56, in which the acetamido substituent of the substrate functions as an intramolecular nucleophile leading to the formation of an oxazolinium ion intermediate (Mark and James, 2002; Mark et al., 2001) (Figure 3.2). In addition, a possible role of an amide as a nucleophile had been previously proposed for glycogen phosphorylase on the basis of a very similar active site interaction, though in that case involving a main chain amide (Mitchell et al., 1996). Subsequently, multiple other examples of amide functional groups positioned to play this role within the active sites of retaining glycosyltransferases have arisen as discussed in sections 1.3.5.4. and 1.3.5.5..  95  Figure 3.1. LgtC active site with bound substrate analogues U D P 2F Gal and 4'deoxy lactose highlighting the proximity of the side chain amide of Gin 189 and the 6'-hydroxyl of lactose to the anomeric reaction centre (pdb lga8) .  Figure 3.2. Mechanism of retaining p-hexosaminidases involving anchimeric assistance leading to the formation of a stabilized oxazolinium ion intermediate.  96  However, despite this precedence, considerable doubt concerning a nucleophilic role for Glnl89 was raised by the finding that the Q189A variant of LgtC possesses relatively high residual activity, with a k  cat  enzyme (Persson et al., 2001).  value that is 3% of that of the wild-type  Doubt concerning a standard double displacement  mechanism also derives from the lack of success in attempts to observe a covalently bound intermediate on LgtC, glycogen phosphorylase and other glycosyltransferases, despite exhaustive studies using techniques that have been successfully applied to observe the glycosyl-enzyme intermediates of glycosidases (Ly et al., 2002). The possibility of a double displacement mechanism involving nonenzymatic nucleophiles has also been investigated with LgtC.  In order to achieve overall net  retention of anomeric configuration, a hydroxyl group from one of the substrates could act as a nucleophile that displaces the UDP leaving group in an initial step. Subsequent attack on the resulting p-linked galactosyl intermediate by the 4'-hydroxyl of the lactose acceptor from the a-face would yield the resulting Gal-(oc-l,4)-Gal linkage. Candidate intramolecular nucleophiles include the 4"- and 6"-hydroxyls of the U D P Gal donor. The plausibility of such roles was discounted on the basis of the inability of LgtC to turnover either of the expected P-1,4 or P-1,6 anhydrogalactose intermediates (Ly et al., 2002). This conclusion is further supported by a report that LgtC is able to use both UDP 4"-deoxygalactose and UDP 6"-deoxygalactose as substrates (Sujino et al., 2000), which would clearly not be possible i f either of the above mechanisms were true. Another candidate nonenzymatic nucleophile is that of the 6'-hydroxyl of the acceptor substrate lactose, which is in fact well positioned for such a role (Figure 3.1).  A mechanism  involving the lactose acceptor in this way would have the advantage of minimizing  97  unwanted donor substrate hydrolysis, as it would obligate ternary complex formation prior to the production of a reactive intermediate species. Limiting the plausibility of such a mechanism is the fact that the resulting intermediate would be that of an inherently unreactive glycoside species. The plausibility of this mechanism was discounted based on the finding that the chemically synthesized putative intermediate (galactosyl P-1,6lactose) was not turned over by the enzyme (Persson et al., 2001). These exhaustive studies led to the proposition that LgtC uses a mechanism termed "SNi-Hke" (Figure 1.12). Such a mechanism has limited chemical precedence and would prove difficult to distinguish experimentally from a classic SNl-like mechanism involving the formation of a discrete oxocarbenium ion intermediate.  3.2.2. Engineering the Mechanism of LgtC To investigate the degree of nucleophilic character contributed by Gin 189 during catalysis, a Q189E variant of LgtC was created, thereby replacing the side chain amide with a more nucleophilic carboxylate. This substitution would also introduce a worse leaving group for the putative covalent glycosyl-enzyme intermediate (pKa of -0.5 to -1.0 for protonation of the oxygen of an amide versus a pKa of -4.0 for that of a carboxylate), which may lead to a rate limiting deglycosylation step and the potential accumulation of the intermediate species. The mutant protein was subjected to kinetic, mass spectrometric and X-ray crystallographic analysis.  Making the transferase more "glycosidase-like"  might result in intermediate accumulation or even perhaps a more efficient enzyme with respect to the rate of transfer, albeit with the potential cost of an increased relative rate of unwanted donor substrate hydrolysis (as discussed later).  98  3.2.3. Kinetic Analysis of LgtC Q189E Expression and purification of the Q189E variant of LgtC-25 proceeded without difficulties yielding protein that displayed the same purification behaviour, yield and C D spectrum as the wild-type enzyme. Kinetic analysis indicates that, while the KM for UDP Gal is approximately unchanged, the measured k  cat  value of this mutant is approximately  3% of that of the wild-type enzyme (Table 3.1). Interestingly, this is very similar to the residual activity of the Q189A mutant observed previously (Table 3.1).  It might be  expected that substitution of the weakly nucleophilic side chain of glutamine with that of a carboxylate-containing side chain could increase the turnover rate of an enzyme that catalysed a reaction via a double displacement mechanism, i f that residue was in fact acting as the catalytic nucleophile. The observed decrease in turnover rate could be interpreted as indicating that this residue is not the catalytic nucleophile, that introduction of a worse leaving group for turnover of the glycosyl-enzyme intermediate leads to a change in the rate limiting step that leads to an overall decrease in the rate of turnover, or that an uncharged nucleophile is required for optimal activity on a charged substrate. The recent finding that an active site tyrosine acts as the nucleophile in the double displacement mechanism of a trans-sialidase, which uses charged sialic acid-containing substrates, supports this latter interpretation (Watts et al., 2003).  Alternatively, this  finding is consistent with the notion that this enzyme uses a mechanism that differs from that involving the discrete nucleophilic catalysis of a double displacement mechanism.  99  T a b l e 3.1.  Michaelis-Menten parameters for the transferase and hydrolase activities of  WT and various mutants of LgtC-25 with respect to UDP Gal and UDP Glc donor substrates measured at pH 7.5, 37°C.  U D P Galactose ^cat  LgtC  Activity  WT  1  ts" ) 1  (uM)  20  1.2  0.007  N/A  Hydrolase  0.02  4.5  0.004  0.0003  N/A  Transferase Hydrolase  0.68 0.003  23 N/A  0.030  0.002 0.001  N/A N/A  Transferase  0.43  25  nd  nd  0.008 0.010  N/A N/A  nd  nd  3  3  d  1  24  13  D190N  (uM'-s )  Transferase b  Q189A  ^cat /  (uM)  (s )  3  Q189E  U D P Glucose  Transferase Hydrolase  3  b  -  c  0.017 c  _  c  -  c  c  c c  * - standard error range in data is from 5 - 20% a - determined at saturating concentrations of lactose acceptor (240 mM) b - determined in the absence of acceptor substrate c - values could not be obtained as the high concentrations of enzyme required to obtain significant rates approached those of unsaturating donor substrate concentrations in violation of MichaelisMenten assumptions. Apparent k values were obtained from rates observed at saturating donor substrate concentrations (100 uM). d - the D190N mutant was generated in a C128/174S background nd - not determined cat  3.2.4. Intermediate Trapping by ESI-MS To further investigate the contribution of nucleophilic catalysis to the mechanism of the Q189E mutant, the enzyme was incubated with various combinations and concentrations of donor and acceptor substrates to see if any covalently modified protein could be observed by ESI-MS.  Interestingly, incubation with 1 m M UDP Gal for 2  minutes followed by quenching in 3 M urea resulted in what appeared to be a steady state  100  population (-10%) of protein for which the mass had increased by 164 mass units, corresponding to a covalently bound galactosyl moiety (Figure 3.3.B). In an attempt to increase the level of labelling, UDP Glc was employed in place of UDP Gal since it has a lower KM value and slower rate of turnover. As hoped, UDP Glc labelled the mutant to a slightly higher level (-30%) (Figure 3.3.C). Incubation in the presence of higher concentrations of donor substrate or in the presence of various concentrations of the incompetent acceptor analogue 4'-deoxy lactose did not increase the relative amount of labelled protein. Surprisingly, quenching the reaction with acidic conditions (phosphate buffer pH 1.6) resulted in the loss of any observable labelled protein. This finding is unusual in that glycosidase intermediates covalently bound by ester linkages are typically stable under acidic conditions. A covalently modified intermediate of the Q189A variant of LgtC-25 was not observed under similar conditions. Supporting the notion that the observed labelled portion of the mutant enzyme was in fact that of a catalytically competent steady state population, rather than a 'dead-end' inhibited population, was the observation that removal of excess U D P Gal from such mixtures by centrifugal dialysis resulted in complete loss of the labelled species as determined by ESI-MS. Presumably the galactosyl-enzyme intermediate was hydrolysed as a result of the significant rate of background hydrolytic activity (Table 3.1). However, because of the higher level of labelling and lower turnover rate, removal of excess U D P Glc from enzyme labelled with that nucleotide sugar allowed the isolation and observation of a labelled species that could be turned over in a time-dependent fashion by incubation in the presence of saturating amounts of lactose acceptor.  This finding  indicates that the observed intermediate is a catalytically relevant species. Indicating a  101  significant kinetic competence for the covalent species, fitting the rate o f decay o f labelled enzyme (as determined by E S I - M S ) to a first order rate equation allowed the calculation o f a first order rate constant o f 0.0014 s" for the turnover o f this intermediate 1  via transglycosylation onto lactose, i n close agreement with the observed k  cat  value o f  0.002 s" for the U D P G l c transferase activity o f this mutant (Table 3.1 and Figure 3.4). 1  A  B  32 432  32 432  Q. O  32595  1n .  Mass amu  ...  k  .  Mass amu D  32 432  32 432  32595  c  Mass amu  Figure  3.3.  32595  Mass amu  E S I - M S analysis o f labelled and unlabelled species o f L g t C Q189E. ( A )  unlabelled, (B) labelled with G a l , (C) labelled with G l c , and (D) labelled with G a l in the presence o f pyruvate kinase and phosphoenolpyruvate.  102  Parameter Initial value Rate constant  Value  SW. Error  35.0453 0.0014  0.8803 0.0001  B y= -OJ0013X • 3 5433 R = 0.989 2  CO  200 600  300 1ime(s)  Time (s)  Figure 3.4. Decay data for the turnover of the glycosyl-enzyme intermediate species resulting from the incubation of U D P Glc with LgtC Q189E.  Following removal of  excess U D P Glc, enzyme was incubated in a solution containing saturating lactose acceptor and left to incubate for various time periods prior to quenching with 6 M urea. First order rate constants were calculated by direct fit of the data to a single exponential first order decay equation (A) or from the slope of a plot of the natural logarithm of concentration versus time (B).  Finally, in an attempt to 'pull' the labelling equilibrium over, the Q189E mutant enzyme was incubated with U D P Gal in the presence of pyruvate kinase and phosphoenolpyruvate to remove U D P formed (Scheme 3.2). Indeed, the introduction of this coupling system resulted in a dramatic increase of the relative proportion of the labelled enzyme species (Figure 3.3.D).  103  UDP-Gal + E  UDP-Gal-E  UTP + Pyruvate + Gal-E  UDPGal-E  PK  UDP + Gal-E  H 2  °»  H C. 2  UDP + Gal + E  UDP + Gal + E  PEP  Scheme 3.2. Equilibrium for LgtC-catalysed hydrolysis of UDP Gal in the presence of pyruvate kinase (PK) and phosphoenolpyruvate (PEP).  3.2.5. Identification of the Site of Labelling To determine i f the Q189E mutant was labelled at the anticipated position (the introduced carboxylate at position 189) following incubation with UDP Gal, the labelled mutant enzyme was digested with trypsin at pH 7.5 and the resulting peptide fragments were subsequently analyzed by reverse phase HPLC/ESI-MS.  Peptide fragments from  both labelled and native protein digests were subjected to comparative mapping analysis, allowing the identification of two peptides that eluted at the same retention time of 34.10 min having m/z ratios of 964 and 1046 respectively. The difference in m/z of 82 is consistent with these being doubly charged peptides carrying a label of mass 164 (the mass of a galactosyl moiety). From the trypsin cleavage sites contained within the LgtC Q189E  protein  sequence,  only  one  peptide  (182DVMQYQDEDILNGLFK197) is expected.  with  a  mass  of  1928  Da  The sequence of this peptide was  subsequently confirmed by fragmentation of the isolated unlabelled species. To determine which residue within the identified peptide contained a covalently bound galactosyl moiety, isolated native and labelled peptides were individually injected  104  and fragmented by increasing the orifice voltage to 65 V . For the labelled peptide species, in addition to the fragmentation pattern found for the unlabelled peptide, peaks were observed having m/z values of 1617, 1600, 1436, 1327, 1309, and 1065 corresponding to peptide fragments containing a covalently bound galactosyl moiety (Figure 3.5).  Most significantly, the peak at m/z = 1065 can only correspond to the  peptide fragment  190DILNGLFK197  lacking ammonia (m/z 902) and containing the  galactose label (m/z 164). To our considerable surprise, therefore, the site of labelling has to be Aspl90.  Substitution of Glnl89, suitably positioned in the ground state crystal  structure to play the role of a catalytic nucleophile, with a more nucleophilic Glu residue therefore results in the accumulation of a steady-state population of covalently modified protein. However, the site of labelling is not the mutated side chain, as was anticipated, but rather a sequentially adjacent aspartate residue.  105  E\D I L N G L F K-197  182-1  1455+163 -NH,= 1600 1291 + 163-NH= 1436 *• 1164+ 163- NH= 1309 920 + 163 - NH = 1065 5  3  3  x3 965.S  1.306  o> CL U  1 1ee  CO c cu  1046.3  1436.8 {1600.2 1617.3  ^ JllL Ji A; 700  900  r  ~" iibo  1  1700  m/z, amu Figure  3.5.  ESI-MS analysis of the fragmented, purified Gal-labelled peptide. The  peptide species bearing the galactosyl label was identified by comparative mapping of tryptic digests of native and labelled LgtC Q189E enzyme species. The sequences of the unlabelled and labelled peptide species were deduced from the observed fragmentation patterns and the enzyme primary sequence to be that of residues 182 to 197 as shown. The location of covalent attachment can conclusively be determined to be Aspl90 from the observed peak at m/z 1065, as no other fragment corresponding to this mass is possible from the known peptide sequence and as this peak is only observed from the fragmented labelled peptide species. The region corresponding to m/z ratios greater than 1100 has been magnified three-fold.  106  3.2.6. Analysis of the D190N Variant of LgtC To probe the mechanistic importance of Asp 190, the D190N mutant of LgtC was generated.  Initial attempts to generate this variant based upon the wild-type enzyme as  template resulted in yields of purified protein that were insufficient for characterization, presumably due to a destabilization of the protein structure. To counter this problem, the mutation was introduced into a double cysteine mutant of LgtC-25 (C128/174S) that had previously been shown to be more stable while maintaining kinetic characteristics that are indistinguishable from those of the wild-type enzyme (Persson et al., 2001). This allowed sufficient enzyme to be purified for kinetic characterisation. The k  cstt  value of the D190N  variant, measured using UDP Gal as variable substrate at a fixed saturating concentration of lactose, was found to be 3000-fold less than that of the wild-type enzyme (Table 3.1). This is consistent with a potential role for this residue as the catalytic nucleophile in a double displacement mechanism.  To ensure that the observed decrease in catalytic  proficiency was not simply the result of unfolded or misfolded enzyme, the D190N variant was subjected to C D analysis, yielding a spectrum indistinguishable from that of the Q189E and wild-type versions of the enzyme. A n exact KM value could not be obtained for the D190N mutant of LgtC-25, as the enzyme appeared to be saturated by substrate concentrations (5 |iM) approaching that of the relatively high enzyme concentration needed to measure significant rates of galactosyl transfer. A n upper limit of 5 ^iM, however, seems reasonable and this significantly lower apparent  than that for the wild-type enzyme ensures that the rates measured are indeed  those of the mutant and not those of contaminating wild-type enzyme (Ly and Withers, 1999).  107  3.2.7.  Crystallographic Analysis of LgtC Q189E  A three-dimensional X-ray crystal structure of the Q189E mutant of LgtC was obtained at 2.6 A resolution with a molecule of UDP 2F Gal bound in the active site by Ms. Cecilia Chiu in Prof. Natalie Strynadka's laboratory at U B C . The Q189E structure can be overlaid with the wild-type structure with a r.m.s deviation of 0.25 A for 269 C a atoms with near identical relative positioning of the carboxylate oxygen of D190 and the anomeric carbon of the donor (Figure 3.6).  Figure 3.6.  Superimposition of the wild-type and Q189E mutant active sites of LgtC.  Side chains for the active site residues are labelled and depicted in C P K coloring with carbon atoms in gray (wild-type) or yellow (mutant), nitrogen atoms in blue, and oxygen atoms in red. U D P 2F Gal is shown with carbon atoms in cyan (wild-type complex) or magenta (mutant complex), nitrogen atoms in blue, oxygen atoms in red, and fluorine atoms in light gray. M n ion is shown as a violet sphere and the water molecule in blue.  108  The near identity of the structures of the wild-type and Q189E mutant indicates that the observed X-ray crystal structures are not of those of the conformation that would lead to the formation of covalent glycosyl-enzyme species involving Asp 190. As such, i f Asp 190 were to act as the nucleophile in a double displacement mechanism, a conformational change from that of the observed ground state would have to occur during the course of the reaction.  3.2.8. Attempts to Accumulate a Covalent Intermediate using Alternative Active Site Modifications If the side chain carboxylate of Asp 190 does in fact act as the catalytic nucleophile in a double displacement mechanism, why does the modification of an adjacent residue result in the accumulation of an observable steady state population of the covalent glycosyl-enzyme intermediate? Because no such intermediate is observed with the wild-type enzyme, one possible explanation is that the modification causes a change in the relative rates of the glycosylation and deglycosylation steps of the mechanism. The Q189E mutation results in the introduction of a charged side chain in close proximity to that of Asp 190, thus producing a potentially unfavourable charge-charge interaction. Formation of a glycosyl-enzyme intermediate would neutralize one of these charges, which could change the relative rates of the glycosylation versus deglycosylation steps leading to an increased life time for the intermediate species. To explore this possibility, other modifications of the LgtC active site were made to determine if this would result in a similar outcome to that of the Q189E mutation. The ternary complex X-ray crystal structure was used to identify candidate uncharged residues  109  close to the side chain of Asp 190. Several uncharged residues were identified that were either in direct proximity (Ilel43, Glnl85) or involved in a hydrogen bond network (Asnl53) with Asp 190 (Figure 3.7). Site directed mutants I143D, N153D and Q185D were generated in the double cysteine background. Unfortunately, the I143D and N153D mutations resulted in a complete loss of soluble protein expression, despite exhaustive efforts to optimize expression conditions. This was not completely unexpected as both mutations result in the introduction of a charged group within the core of the protein, which is frequently found to severely compromise structurally stability. Expression of the Q185D variant proceeded with expression levels comparable to that of the wild-type enzyme. However, no covalent glycosyl-enzyme species was observed despite testing a variety of labelling conditions. This could be a consequence of the side chain of this residue being close to the protein surface where the effect of the introduced charge could be neutralized by bulk solution.  110  Q185  Figure 3.7.  LgtC active site highlighting the side chains of residues within close  proximity to that of Asp 190 (pdb lga8). Distances are indicated in Angstroms.  3.2.9. Interpretation of the LgtC Q189E Labelling Results Several interpretations of the surprising LgtC Q189E labelling results are possible. The first possibility is that labelling does in fact occur at Glul89, but that a subsequent migration results in the label being observed at the alternate Asp 190 site. However, the relative positioning of both of these residues within the active site (both on the (3 face of the galactose ring) would prevent glycosyl migration from occurring in the intact enzyme (Figure 3.6 and 3.7). Additionally, it is hard to envision a chemically sound mode of acyl migration on the denatured or digested protein (Figure 3.8).  Ill  Figure 3.8. A c y l migration o f the galactosyl moiety from the side chain o f Q189 to that o f D190 is difficult to rationalize.  It must therefore be considered possible that A s p 190 is i n fact the catalytic nucleophile o f L g t C and that retaining glycosyltransferases catalyse reactions v i a a double displacement mechanism involving discrete nucleophilic catalysis. The Q189E mutation may lead to a slowing down o f the deglycosylation step relative to the glycosylation step, allowing the accumulation o f an observable steady state population o f labelled enzyme.  Indeed,  formation o f the  glycosyl-enzyme intermediate  would  neutralise the negative charge o f A s p 190, which is i n close proximity to the introduced carboxylate o f G l u l 8 9 , eliminating charge-charge repulsion and thereby altering the relative rates o f intermediate formation and breakdown. This interpretation is supported by the observed pH-dependent stability o f the labelled species. A t low p H the introduced carboxylate would be protonated, preventing unfavourable charge interaction with D190 and thus resulting in a decrease in the stability o f the glycosyl-enzyme intermediate. This argument is i n accordance with the observed lack o f discernible labelling with the Q 1 8 9 A mutant, i n which case the mutation does not introduce a negatively charged species.  112  It is worth noting that previous suggestions regarding the mechanism of LgtC, and retaining glycosyltransferases in general, have been based primarily on the proximity of candidate nucleophiles to the anomeric carbon of a donor sugar within a crystal structure. Indeed, on this basis Asp 190 was discounted as a potential catalytic nucleophile due to the distance (8.9 A) separating the carboxylate oxygen and the reaction centre in the crystal structure of LgtC with bound donor analogue (PDB lga8). If Asp 190 were in fact the catalytic nucleophile, a conformational change would need to occur during catalysis that would bring it into an appropriate position and orientation with respect to the donor sugar substrate. The need for a significant conformational change during catalysis is also suggested by the observed crystal structure of the self-glucosylating enzyme glycogenin (EC 2.4.1.186), the other family GT8 enzyme for which a structure with a bound intact donor sugar substrate exists (Gibbons et al., 2002). This enzyme transfers a glucosyl moiety from U D P Glc to the hydroxyl of Tyrl95 on the surface of another monomer. The carboxylate oxygen of the proposed nucleophile of this enzyme (Asp 163) was found to be 6.1 A from the anomeric carbon of the donor sugar while that of Asp 160 (structurally equivalent to Asp 190 of LgtC) was found to be 6.9 A away.  Equally  significant is the observed distance of -12 A separating the anomeric reaction centre of the donor substrate from the hydroxyl oxygen of the Tyrl95 acceptor. It is also possible that the observed site of labelling is simply an artefact arising from an altered active site configuration that differs significantly from that of the wildtype enzyme in which Asp 190 becomes suitably positioned to attack the UDP-Gal donor as a result of the introduced mutation.  Substitution of a neutral Gin residue with a  charged Glu might introduce significant charge repulsion leading to an altered active site  113  conformation.  To determine whether this latter interpretation was the source of the  observed unexpected site of labelling, a crystal structure of the Q189E variant was obtained at 2.6 A with U D P 2F Gal bound in the active site and compared to the equivalent structure of the wild-type enzyme (Figure 3.6). The observed nearly identical observed active site configurations of these two structures render it unlikely that the unexpected site of labelling was the result of an altered active site conformation or mode of donor sugar binding. Mutation of the catalytic nucleophiles of retaining P-glycosidases typically results in decreased observed k  values by factors of 10 - to 10 -fold (Ly and Withers, 1999). 5  cat  8  However, equivalent mutations in a-retaining enzymes have been shown to result in less dramatic effects.  For example, mutation of the catalytic nucleophile (Asp229) of the  family GH13 cyclodextrin glycosyltransferase (CGTase, E C 2.4.1.19), the enzyme responsible for the conversion of starch and other a-l,4-linked glucopyranosyl chains into cyclic oligosaccharides, to asparagine led to observed decreases in activity on the order of 2000-fold (Knegtel et al., 1995; Mosi et al., 1997). The kinetic analysis of the D190N variant may therefore be interpreted as being consistent with the role of this residue as the catalytic nuclophile in a double displacement mechanism. Recently, it was found that mutation of the proposed nucleophile of the retaining glycosyltransferase oc3GT (Glu317 - structurally equivalent to Gin 189 of LgtC) to glutamine reduced the observed & t for galactosyltransferase activity 2400-fold (Zhang et al., 2003). ca  114  3.3. Investigating the Retaining a-(l,2) Mannosyl Transferase Kre2 3.3.1. Background Kre2 ( M n t l p ) is an a-1,2 mannosyltransferase from Saccharomyces ( E C 2.4.1.131) belonging to family G T 1 5 .  cerevisiae  W i t h overall net retention o f anomeric  configuration, it catalyses the transfer o f mannose to the 2-hydroxyl o f various a - D mannopyranosides using G D P M a n as the donor substrate (Scheme 3.3). A type II G o l g i membrane glycoprotein, Kre2 is involved in the mannosylation o f both N - and O-linked glycoproteins.  Three-dimensional X-ray crystal structures o f this enzyme have been  reported with either G D P or G D P and the acceptor methyl a-D-mannopyranoside bound in the active site (Lobsanov et al., 2004). Kre2 has the overall architecture o f the G T - A fold and is therefore classified amongst Clan III glycosyltranferases. Using the ternary complex structure o f L g t C (pdb lga8) as a reference, the donor mannose moiety was modelled into the active site o f Kre2 structure with bound G D P .  Based on this model,  Tyr220 was proposed to act as the catalytic nucleophile in a double displacement mechanism. Kinetic analysis o f the Y 2 2 0 F mutant revealed a k  cat  value 3000-fold lower  than that o f the wild-type enzyme, which the authors interpreted as support for this catalytic role. A nucleophilic tyrosyl hydroxyl had been reported for a trans-sialidase just prior to the publication o f this Kre2 structural study (Watts et al., 2003).  115  OH HO  Kre2  HO OGDP  OR  GDP OR  Scheme 3.3.  Mannosylation reactions catalysed by Kre2.  In vitro studies indicate  acceptors with R = H, methyl, or a-D-mannose to be efficiently utilized by Kre2 (Lobsanov, 2004).  3.3.2 Synthesis of a Non-Reactive Donor Substrate Analogue To overcome the problem of donor substrate hydrolysis, the non-reactive analogue GDP 2F Man was synthesized using traditional morpholidate coupling methodology (Scheme 3.4).  The ability of this compound to act as either a slow substrate, an  inactivator or a competitive inhibitor of Kre2 was then tested. In addition, GDP 2F Man was subjected to co-crystallization trials with the enzyme.  OH  OH  Carboxamidine salt  Scheme 3.4.  Synthesis of the non-reactive donor substrate analogue GDP 2F Mannose  via a morpholidate coupling reaction.  116  3.3.3 Kinetic Analysis of Kre2 Kinetic analysis of Kre2 varying the concentration GDP Man at a fixed saturating concentration of methyl a-D-mannoside  acceptor  continuous enzyme-coupled assay, yielded K  M  and k  substrate, using the  PK/LDH  values of 110 n M and 28 s"  1  cat  respectively. These values are in approximate agreement with those obtained using a stopped radio-labelled assay (Lobsanov et al., 2004). Consistent with previous findings for the reactivity of nucleotide 2-fluoro sugar derivatives with glycosyltransferases, GDP 2F Man proved not to be a substrate for Kre2 and was found to act as a competitive inhibitor with respect to the GDP Man donor substrate with an observed K; value of 470 \\M. Time dependent inactivation of Kre2 activity, indicative of covalent modification associated with the accumulation of a covalent glycosyl-enzyme intermediate, was not observed.  3.3.4 Attempts to Observe a Covalent Glycosyl-Enzyme Intermediate with Kre2 In an effort to observe a covalently bound glycosyl-enzyme intermediate, thereby providing support for the proposed double displacement mechanism, Kre2 was subjected to ESI-MS analysis in the presence or absence of various combinations of saturating concentrations of donor and acceptor substrates.  Presumably due to the problem of  heterogeneous glycosylation, the catalytic domain of wild-type Kre2 did not give an ESIM S signal that could be deconvoluted to obtain an accurate mass measurement. Comparative peptide mapping of tryptic and peptic digests of Kre2 incubated under a variety of conditions, in which incubation time, quenching pH and combinations of  117  saturating concentrations o f G D P M a n , G D P 2F M a n and methyl oc-D-mannose were varied (see methods section), did not reveal the formation o f a covalent glycosyl-enzyme intermediate species.  3.3.5 X-Ray Crystal Structure of Kre2 with Bound GDP 2F Mannose In collaboration with Dr. Y u r i Lobsanov in Lynne Howell's laboratory at Toronto's Hospital for Sick Children, a co-crystal structure was obtained for Kre2 with the donor substrate analogue G D P 2F M a n bound within the active site. This structure, with the intact substrate analogue bound within the active site, reveals that the previously proposed nucleophilic side chain o f Tyr220 is in fact situated on the alpha face o f the mannose sugar ring plane (Figure 3.9A).  This indicates that the described previous  interpretation based on modeling was i l l conceived. In order for this residue to play the role o f catalytic nucleophile in a double displacement mechanism, it would need to attack the anomeric reaction centre from the P-face o f the sugar ring in order to form the proposed beta-linked covalent glycosyl-enzyme intermediate.  From this structure it is  apparent that the side chain hydroxyl o f Tyr220 is within hydrogen bonding distance o f both the bridging and a non-bridging oxygen o f the phosphate leaving group. A s such, a more plausible catalytic role for this residue is that o f electrostatic stabilization that facilitates leaving group departure, as was similarly observed for T y r l 5 6 and T y r l 6 2 in the case o f C s t l l (Chiu et al., 2004). In the case o f Cst II, these tyrosyl hydroxyl groups were proposed to stabilize departure o f the C M P leaving group and mutation o f either o f these residues to Phe resulted in significant impairment o f both transferase and hydrolase  118  activities, albeit to a less severe degree than that observed with the Y 2 2 0 F mutation o f Kre2.  Figure  3.9.  Kre2 active site with bound intact donor substrate analogue G D P 2F M a n .  Potential candidate catalytic nucleophiles Y 2 2 0 (A), S326 (B), E329 (C) and E247 (D) are highlighted as discussed in the text.  Based on the complexed crystal structure with intact donor substrate analogue bound, amino acid side chains can be identified that exist on the beta face o f the mannose sugar ring lying closest to the anomeric reaction centre in a position to act as a potential catalytic nucleophile.  The side chain hydroxyl o f Ser326 is situated 4.9 A from the  reaction centre within hydrogen bonding distance o f the 2-hydroxyl o f mannose and the side chain carboxyl group o f Asp361 (Figure 3.9B). Interestingly, somewhat reminiscent o f the classical catalytic triad motifs o f serine proteases, albeit i n a differing order, the  119  side chain carboxylate of Asp361 in turn forms hydrogen bonding interactions with the side chain imidazole of His365. Based simply on structure, it could be proposed that the enzyme uses an uncharged side chain hydroxyl as a nucleophile, which is activated by a neighboring carboxylate residue, to attack the charged GDP Man substrate. This proposal is reminiscent of the finding that a similarly activated tyrosyl hydroxyl serves as the nucleophile of a trans-sialidase that acts on a charged sialic acid-containing substrate (Watts et al., 2003). The Kre2 structure in complex with GDP 2F Man also reveals the identity of two possible carboxylate nucleophiles. The side chain carboxylate of Glu247 is close (3.4 A) to the anomeric reaction centre making it a possible candidate, however, the likelihood of a nucleophilic role for this residue is limited by several facts. Firstly, this residue is a component of the D X D motif that is responsible for coordination of the obligate M n within the active site of the enzyme.  2 +  ion  In addition, in the observed conformation, the  carboxylate is not configured at an angle appropriate for nucleophilic attack on the reaction centre (Figure 3.9D). It is difficult to envisage a conformational change that would allow a loss of coordination of M n  2 +  and reconfiguration to an angle that would  facilitate nucleophilic attack. Finally, the side chain carboxylate of Glu329 of Kre2 is located 9.0 A from the anomeric reaction centre of the donor substrate in a similar relative orientation and at a similar distance to that of Asp 190 of LgtC (Figure 3.9C). Also analogous to Asp 190 of LgtC (Figure 3.6), the side chain carboxylate of Glu329 hydrogen bonds to two tyrosyl hydroxyl groups. As has been the case with the majority structural investigations with retaining glycosyltransferases, crystallographic analysis of Kre2 with bound intact donor substrate  120  analogue GDP 2F Man does reveal the clear identity of a potential catalytic nucleophile required by the proposed double displacement mechanism.  As discussed in section  1.3.5.4., the only structurally conserved motif within the active sites of retaining glycosyltransferases that adopt the GT-A fold, on the P-face of the donor sugar substrate, is that of a side chain carboxylate located within hydrogen bonding distance of donor sugar hydroxyls and a positively charged side chain donated by either a lysine or arginine residue, which is in turn involved in hydrogen bonding interactions with a side chain carboxylate derived from a residue of the D X D motif. This motif is observed with Kre2, consisting of the side chains of Asp361, Arg245, and the D X D side chain of Glu247 (Figure 3.9B and 3.9D). However, unlike most other glycosyltransferases wherein the carboxylate hydrogen bonds with the C3, C4 or C6 hydroxyls, in the case of Kre2 the carboxylate of Asp361 interacts with what would be the C2 hydroxyl of the mannose donor. Another unique feature is the observation that the side chain of Arg245 is located in very close proximity to the anomeric reaction centre.  121  3.4. ESI-MS Analysis of the Retaining <x-(l,3) Galactosyl Transferase <x3GalT 3.4.1. Background The Golgi body-residing type II transmembrane family GT6 galactosyltransferase a3GalT (EC 2.4.1.87) is expressed in New World primates and many non-primate mammals (Galili and LaTemple, 1998). Using U D P Gal as donor substrate, a3GalT catalyses the transfer of galactose to the 3'-hydroxyl of terminal LacNAc-containing acceptors with overall net retention of anomeric configuration (Scheme 3.5).  Scheme 3.5. Galactosylation reaction catalysed by a3GalT.  The bovine version of this enzyme has been the subject of various structural and mechanistic investigations. Based on an initial three-dimensional X-ray crystal structure, Glu317 (structurally equivalent to Gin 189 of LgtC) was proposed to act as the catalytic nucleophile in a double displacement mechanism (Gastinel et al., 2001). In fact, in this initial report, the authors claimed to observe a covalently bound intermediate species. However, a lack of order in the active site region and the presence of two crystal forms for the structure containing the proposed covalent intermediate led to a general lack of 122  acceptance o f this interpretation. Later, a detailed study o f the binding affinities o f w i l d type and E317Q enzyme for donor and acceptor substrates led the authors to propose that this residue is required for proper acceptor substrate orientation and does not play the role o f a catalytic nucleophile (Zhang et al., 2003). However, this same study revealed that mutation o f Glu317 led to a 2400-fold decrease in the observed turnover rate, indicating the critical importance o f this residue (Zhang et al., 2003). Very recently, a structure with the non-reactive donor substrate analogue U D P 2F G a l bound within the active became available and, like the initial structural report, also revealed the side chain o f Glu317 to be most suitably positioned to play the role o f a catalytic nucleophile (Jamaluddin et al., 2007) (Figure 3.10). A l s o quite recently, it has been shown that the activity o f the E 3 1 7 A mutant o f a 3 G a l T can be rescued by the addition o f small exogenous nucleophiles like azide and a P-azido derivative was isolated and characterized (Monegal and Planas, 2006). A s described in section 2.5.2, this chemical rescue strategy has been successfully employed to identify the acid/base and nucleophile catalytic residues o f retaining glycosidases. A s such, the chemical rescue o f Glu317 is suggestive that this side chain is suitably positioned to play such a catalytic role in the mechanism o f a 3 G a l T . However, it must be kept i n mind that, while chemical rescue does indicate that a side chain is suitably positioned to play a catalytic role, it does not prove that it plays such a role i n the natural enzyme mechanism. It is not surprising that a mutant enzyme is found to act as a scaffold that catalyses an unnatural reaction simply by proximity effects, when the mutation leads to the introduction o f a space that can accommodate a good nucleophile i n proximity to an electrophilic center. Chemical rescue results alone cannot be used to deduce the natural catalytic mechanism o f an enzyme.  Chemical rescue experiments  123  prove particularly useful in the absence of structural information or, perhaps more relevant to retaining glycosyltransferases,  if the catalytic significance of structural  information is in question.  Figure 3.10.  Active site of a3GalT with bound non-reactive donor substrate analogue  UDP 2F Gal highlighting the position of Glu317 suitably positioned to play the role of a catalytic nucleophile in a double displacement mechanism (pdb ljcf).  3.4.2. Attempts to Observe a Covalent Glycosyl-Enzyme Intermediate with ct3GalT In light of the recently described chemical rescue of the E317A mutant of bovine oc3GalT, providing potential support for the notion that this residue might play the role of a catalytic nucleophile in the proposed double displacement mechanism, it was deemed  124  prudent to determine i f a covalent glycosyl-enzyme intermediate could be observed by ESI-MS. The catalytic domain was cloned as a MalE fusion protein in the laboratory of Dr. Warren Wakarchuk at the National Research Centre Institutes for Biological Sciences in Ottawa. This fusion protein was readily expressed in E. coli, leading to the production of active and homogenous enzyme.  This non-glycosylated enzyme gave an ESI-MS  signal that could be deconvoluted to obtain a mass of 75,172 Da, consistent with the theoretical mass of 75,163 Da for the MalE fusion version of the protein. However, despite exhaustive efforts using a range of labelling conditions (incubation time, quenching pH, combinations of donor and acceptor substrates), a covalent glycosylenzyme intermediate species of a3GalT was not observed.  In addition, comparative  peptide mapping of tryptic and peptic digests of a3GalT incubated with or without U D P Gal under a variety of conditions did reveal the formation of a glycosyl-enzyme adduct.  125  3.5. Reasons Why Retaining Glycosyltransferases May Have Evolved a Mechanism Not Involving Nucleophilic Catalysis By contemplating the inherent differences in the reactivities of the substrates utilized by glycosidases and glycosyltransferases, a rationale can be developed as to why these two classes of enzymes may have evolved differing mechanistic strategies for catalysing glycosyl group transfer reactions with net retention of anomeric configuration. Glycosidases catalyse the hydrolysis of one of the most stable linkages found in nature, the glycosidic bond between two sugars, with the half lives for spontaneous hydrolysis being on the order of ~5 million years (Wolfenden et al., 1998). The stability of this linkage, under neutral conditions, is imparted by the relatively poor leaving group ability (pKa ~14) of natural glycosidase substrates. This stability leads to a large energy barrier to bond cleavage in the first step of the glycosidase reaction (Figure 3.11 A). It would seem logical that retaining glycosidases would have evolved a catalytic mechanism involving an intermediate, and in accordance with the Hammond postulate therefore a transition state, of the lowest possible free energy.  In other words, by selecting a  mechanism with a lower energy intermediate species, catalytic efficiency is easier to achieve because the rate determining transition state being stabilized will also have a lower free energy. By this logic, the Koshland S ^ - l i k e double displacement mechanism, involving the formation of a covalently bound glycosyl-enzyme intermediate, would have been preferred over the Phillips S N I - l i k e mechanism, involving the formation of a higher energy oxocarbenium ion intermediate, during the evolution of retaining glycosidase mechanism (Figure 3.11 A).  126  B  T  Oxocarbenium Ion Intermediate (S 1-like Pathway) N  GT Substrate Complex  Oxocarbenium Ion Intermediate r(S 1-like Pathway)/ N  Covalent GlycosylEnzyme Intermediate (S 2-llke Pathway)  Covalent GlycosylEnzyme Intermediate (S 2-like Pathway)  N  N  GH Product Complex  Figure 3.11. Comparison of the potential differences in the free energy barriers for SNIlike and S ^ - l i k e pathways that could be utilized by (A) retaining glycosidases (GH) and (B) retaining glycosyltransferases (GT). Because glycosyltransferases utilize high energy (substituted) phospho-sugar donor substrates containing a very good leaving group (pKa of ~2 for the coordinated species), the free energy barrier for the first step is much less than that for the glycosidases that utilize a very stable substrate containing a poor leaving group (pKa of ~14). This smaller free energy barrier for the first step of the reaction may have led to the SNl-like pathway being sufficiently effective to be selected for during the course of the evolution of retaining glycosyltransferase mechanisms.  In addition, for  glycosyltransferases, because of the near equality in leaving group ability for both steps, utilization of a side chain carboxylate as a catalytic nucleophile (as is the case for the vast majority of retaining glycosidases) in a double displacement mechanism might lead to the deglycosylation step of the reaction being rate limiting.  This would increase the  probability of unwanted donor substrate hydrolysis.  In contrast, glycosyltransferases utilize high energy (substituted) phospho-sugar donor substrates. The leaving group ability for the donor substrate is much greater to start with, compared to glycosidase substrates, with a pKa of ~7 for the second ionisation of phosphate.  In addition, it is easy to imagine glycosyltransferases readily increasing this  127  leaving group ability electrostatically, using a coordinated divalent cation or two positively charged amino acid side chains (as discussed in sections 1.3.5.4. and 1.3.5.5.), to pKa values of ~ 2 (the pKai value for a phosphate monoester). Because of the higher free energy of the starting material and lower free energy of the first transition state, the inherent barrier for bond cleavage in the first step of the glycosyltransferase reaction is less than that of the glycosidase reaction (Figure 3.1 IB).  Because there is not such a  large barrier for the initial bond cleavage step compared to that for the analogous glycosidase reaction, an SNI-like pathway involving the formation of an intermediate oxocarbenium ion may have been accessible during the course of the evolution of the mechanisms of retaining glycosyltransferases and sufficient to have been selected for (Figure 3.1 IB). In addition, in order to minimize the unwanted side reaction of donor substrate hydrolysis, the relative free energies of the starting material and intermediates might even suggest that the SN2-like pathway may have been selected against. Because of the low energy barrier for the first step of the reaction, formation of a covalent glycosyl-enzyme species could lead to an energy well that would make the second deglycosylation step rate limiting (Figure 3.1 IB).  This can be rationalized on the basis of the relative pKa values  for the leaving groups of the starting (substituted) phospho-sugar donor (pKa of ~2 for the coordinated species) versus that of a proposed side chain carboxylate acting as a catalytic nucleophile (pKa of ~4).  Indeed, the chemistry involved in the first step of the  glycosyltransferase-catalysed reaction is rather analogous to that of the deglycosylation step of a P-retaining glycosidase catalysed reaction (Figure 3.12).  In the case of  glycosyltransferases, the first reaction consists of carbon-oxygen glycosidic bond  128  cleavage with an axially-linked (substituted) phosphate leaving group. In the case of a Pretaining glycosidase the second step consists of carbon-oxygen glycosidic bond cleavage with an axially-linked carboxylate leaving group.  Figure 3.12.  Chemical similarities between the substrate used in the first step of the  glycosyltransferase reaction and the covalent glycosyl-enzyme intermediate turned over in the second step of a p-retaining glycosidase catalysed reaction.  An increased lifetime for the intermediate glycosyl donor species could lead to an increased probability for unwanted attack by water leading to donor substrate hydrolysis and increased cellular metabolic demand. At first approximation this argument may seem to violate reactivity/selectivity concepts and product partitioning principles for reactions in free solution.  However, this argument is being made exclusively  for a reaction  occuring within the confines of an enzyme active site. In the case of a reaction in free solution, the selectivity of a reaction does in fact decrease with increased reactivity and the partitioning of a reactive intermediate between two competing processes will be the same whether formation of the intermediate is the rate limiting step or not.  In the  confines of the active site of a retaining glycosyltransferase the intermediate formed following the first bond cleavage step, whether cationic or covalently bound, has three possible fates. Attack by the (substituted) phosphate leaving group will give back starting material, attack by the nucleophilic group of the acceptor will give the desired product,  129  and attack by water will give the undesired hydrolysis product. Clearly, in solution, with a water concentration of 55 M , donor substrate hydrolysis would be the statistically and thermodynamically favoured outcome. However, the enzyme catalyses group transfer in part by suppressing the hydrolysis reaction by excluding water from the active site, thereby decreasing the effective concentration of water. The structural data available for this class of enzymes indicates that this anhydrous microenvironment is achieved in part by using an acceptor as a molecular "plug" that blocks bulk sovent from the region where glycosyl group transfer occurs. Association of the acceptor with the enzyme is a dynamic process with associated on and off rates. A n enzymatic mechanism involving a longerlived intermediate would not be able to suppress the hydrolytic reaction as efficiently, because this would increase the probability of the acceptor dissociating from enzyme with bound intermediate thereby leading to an increased effective concentration of water and an increased rate of catalysed hydrolysis. This argument could also be the reason why, i f retaining glycosyltransferases were to utilize a double displacement mechanism, an amide group would be the nucleophile of choice due to its better leaving group ability (pKa of ~ -0.5 to - 1 ) , resulting in a shorter lived intermediate species. Of course, the pKa value of a functional group, indicative of its leaving group ability for similar chemical types, varies greatly for species bound within an enzyme active  site.  However, a consideration of the  values outside the enzymatic  microenvironment indicates the starting points for the evolution of protein function. These are therefore useful values to consider when discussing the evolution of enzyme mechanism.  130  Precedence for enzymatic mechanisms involving SNI-like pathways, and the intermediate formation of cationic species, can be derived from enzymes involved in terpenoid biosynthesis - isopentenyl diphosphate (IPP) isomerases, farnesyl diphosphate (FPP) synthetases, and the terpene cyclases. IPP isomerase catalyses the conversion of IPP to dimethylallyl diphosphate (DMAPP) using M g  2 +  and Z n  2+  in a pathway that has  beeri shown to occur through an SNI-like mechanism involving the intermediate formation of a tertiary carbocation intermediate. Support for the intermediate formation of a carbocationic species comes from studies quantitating decreased reactivities for fluorinated analogues of IPP (Muehlbacher and Poulter, 1988) and D M A P P (Reardon and Abeles, 1986) and potent non-covalent inhibition by ammonium-containing analogues that mimic the cationic intermediate species (Muehlbacher and Poulter, 1985, 1988; Reardon and Abeles, 1986). The cationic intermediate is well sequestered from solvent within a highly hydrophobic active site and is stabilized by the presence of an aromatic indole side chain, donated by an appropriately positioned tryptophan residue, via a pication interaction (Durbecq et al., 2001). A similar mechanism, and the use of active site aromatic amino acid side chains to stabilize cationic intermediate species, has been defined for multiple terpene cyclases (Christianson, 2006; Shishova et al., 2007; Starks et al., 1997). FPP synthetases catalyse the synthesis of the terpenoid biosynthetic Cio and C 1 5 building blocks geranyl diphosphate (GPP) and farnesyl diphosphate (FPP), respectively, using the C 5 building blocks IPP and D M A P P (Figure 3.13). The mechanism of FPP synthetases is generally believed to be that of a three-step "ionisation-condensationelimination" process, as shown in Figure 3.13 for the condensation of IPP and D M A P P  131  leading to the formation of GPP. Like Cst II, this enzyme is bifunctional in the sense that it also catalyses the synthesis of FPP in a second condensation reaction between the initially formed GPP product and an additional molecule of IPP (Figure 3.13). Evidence supporting the S ^ - l i k e mechanism shown in Figure 3.13 comes from studies with fluorinated analogues of D M A P P (Poulter and Satterwhite, 1977) and GPP (Poulter et al., 1977,  1978) that undergo, at significantly depressed rates, FPP synthetase-catalysed  condensation reactions with IPP.  The rate depressions observed for the fluorinated  (versus non-fluorinated) species for these enzyme-catalysed processes were compared to the rate depressions observed for the solvolysis of the identically fluorinated (versus nonfluorinated) analogues of the methanesulphonate derivatives of D M A P P and GPP. The solvolyses of these methanesulphonate derivatives are known to occur via S N I processes involving the intermediate formation of a cationic species.  The finding that a similar  degree of rate depression, resulting from introduction of electronically destabilizing fluorine atoms, is observed for enzyme-catalysed and spontaneous solvolysis processes indicates that the FPP synthetase reaction proceeds via a similar SNI-like mechanism.  132  OPP  OPP  OPP  B IPP  HOPP  + OPP  OPP  OPP  DMAPP  OPP GPP IPP HOPP  OPP FPP  Figure 3.13.  The reactions catalysed by FPP synthetase.  A three-step "ionisation-  condensation-elimination" mechanism involves the intermediate formation of cationic species that are stabilized by appropriately positioned carbonyl moieties donated by conserved main chain and side chain amide groups. The majority of characterized retaining glycosyltransferases also have an analogous amide group positioned above the anomeric reaction centre to stabilize the development of positive charge.  As with  retaining glycosyltransferases, a substrate/product phosphate is believed to play the role of a base catalyst (B). FPP synthetase is bifunctional, being able to catalyse both the condensation of D M A P P with IPP leading to the production of GPP and the condensation of GPP with IPP leading to the production of FPP.  133  The chemistry involved in FPP synthetase-catalysed reactions is analogous to that involved in glycosyltransferase catalysis.  In both cases, a nucleophilic substitution  reaction occurs at a carbon centre with a (substituted) phosphate acting as the leaving group. Stereochemically, they are more like inverting enzymes in that they preorganize the nucleophilic alkene in a position to directly trap the carbocation and do not suffer the same stereochemical impediment of the retaining enzymes. However, analogies between the overall chemistries being catalysed provide a foundation for a comparison of the transition states chosen to be stabilized and the possibility of a convergence in mechanism during the course of evolution. As is observed for glycosyltransferases, FPP synthetases use a divalent cation and the positively charged side chains donated from a conserved Arg/Lys pair to stabilize charge development and facilitate departure of the phosphate leaving group (Gabelli et al., 2006; Hosfield et al., 2004).  Intriguingly, conserved  electrostatic interactions with neutral carbonyl groups, derived from the amides of a main chain (Lys202) and a side chain (Gln241), plus a side chain hydroxyl (Thr203), are positioned within the FPP synthetase active site to stabilize the positive charge distributed over the C l and C3 atoms of the cationic intermediate (Gabelli et al., 2006; Hosfield et al., 2004). This is similar to the observed positioning of the carbonyl oxygen of a main chain or a side chain amide in close proximity (on the p-face) to the anomeric reaction centre within the active sites of the vast majority of structurally characterized retaining glycosyltransferases (Figures 1.9 and 1.10).  In both cases, formation of a cationic  intermediate is also stabilized by inherent properties of the substrates.  In the case of  glycosyltransferases,  centre  positive  charge  formation  at  the  anomeric  of  an  oxocarbenium ion is stabilized by a lone pair of non-bonding electrons donated by the  134  endocyclic oxygen of the sugar ring. In the case of FPP synthetases, an allylic tertiary centre stabilizes positive charge development at the carbon centre that undergoes bond fission.  Also, as is proposed for retaining glycosyltransferases (sections 1.3.5.4 and  1.3.5.5.), FPP synthetases appear to situate the phosphate leaving group in such a manner that it is suitably positioned to play the role of the base catalyst (B) required for proton abstraction in the elimination step of the reaction (Figure 3.13).  These described  similarities in the chemistries catalysed and the observed conservation of structural architecture, suggests that retaining glycosyltranferases utilize an S N I - l i k e mechanism analogous to that of FPP synthetase.  3.6. Concluding Remarks Presently, the catalytic mechanism of retaining glycosyltransferases is a topic of debate with no clear and definitive answers.  By analogy to the well-characterized  glycosidases, a double displacement mechanism involving discrete nucleophilic catalysis and the formation of a covalently bound glycosyl-enzyme intermediate seems a likely possibility.  Such a mechanism necessitates an appropriately positioned enzymatic  nucleophile within the active site on the P-face of the donor substrate in close proximity to the anomeric reaction centre. As described in sections 1.3.5.4 and 1.3.5.5, with the exception of enzymes from family GT6 wherein a side chain carboxylate is appropriately situated, the most suitably positioned functional group found within the active sites of structurally characterized retaining glycosyltransferases is usually a main chain or side chain amide group. The results of crystallographic analysis of Kre2 with bound intact donor substrate analogue indicate that, in this case, neither of these functional groups are  135  found and a side chain hydroxyl from Ser326 is present in the approximate location occupied by the amide or carboxylate functional groups in other retaining enzymes. In the case of the family GT6 retaining enzyme a3GalT, crystallographic analysis reveals a side chain carboxylate donated by Glu317 perfectly positioned within the active site to play the role of a catalytic nucleophile in a double displacement mechanism. The suitabilty of the positioning of this group for a catalytic role is further supported by chemical rescue experiments with the E317A mutant. The catalytic importance of this residue is supported by kinetic analysis of the E317A mutant, for which turnover is significantly decreased with respect to the wild-type enzyme. However, whether this enzyme utilizes an S N I - l i k e or Sw2-like pathway cannot be deduced by structural, rescue, and kinetic experiments alone. The structural and rescue results serve only to identify candidate catalytic residues and the mechanistic relevance of these studies should never be overstated.  The kinetic results with the E317A mutant illustrate the catalytic  importance of the Glu317, but do not indicate what that role is. Indeed, i f the enzyme were to use an SNI-like pathway, this side chain carboxylate could play a critical role in stabilizing the positively charged oxocarbenium ion intermediate and its deletion could lead to the observed decrease in rate. Conversely, the inability to observe a covalently modified version of a3GalT by ESI-MS cannot be legitimately used as evidence against the SN2-like pathway. These experiments relied on the optimistic expectation that, i f the enzyme were to use a pathway involving the formation of the glycosyl-enzyme intermediate, the deglycosylation step of the reaction would be rate-limiting and/or the lifetime of this species would be long enough for it to be observed by M S analysis. As such, the mechanism of a3GalT remains unclear.  136  Crystallographic analysis of the family GT8 retaining enzyme LgtC reveals the side chain amide of Glnl89, structurally equivalent to Glu317 in a3GalT, to be perfectly situated to play the role of a catalytic nucleophile in a double displacement mechanism. Kinetic analysis of the Q189A mutant revealed an unexpectedly modest decrease in observed rates, inconsistent with a critical catalytic role for this residue. These kinetic data generated considerable doubt as to the likelihood of a double displacement mechanism for LgtC. However, even i f the SNI-like pathway involving formation of an oxocarbenium ion intermediate were the actual LgtC mechanism, it would be expected that Gin 189 would play a critical role in stabilizing the high energy intermediate species and a larger impact on turnover would be expected when this side chain amide is removed.  This may suggest that the observed ground state crystal structures are  misleading in assigning the active site features essential during the course of reaction. The observed labelling of the Q189E variant of LgtC-25 represents the first report of the direct observation of a glycosyl-enzyme intermediate covalently bound to the active site of a retaining glycosyltransferase. While it is possible that this is simply an artefact resulting from the creation of an active site mutant, it rekindles the controversy surrounding the mechanism of LgtC, which has served as a model system for the mechanistic study of retaining glycosyltransferases. Although by no means conclusive, the surprising site of labelling implicates Asp 190 as an alternative candidate catalytic nucleophile. This notion is supported by kinetic analysis of the D190N mutant and by the catalytic relevance of the intermediate species. If this residue were to serve that function, a conformational change from that observed in the ground state crystal structure would be required during catalysis to allow for an appropriate positioning relative to the donor  137  sugar substrate. A n alternative explanation for the observed labelling with the  Q189E  mutant could be that LgtC does in fact utilize an SNI-like pathway and that the altered electrostatic environment of this mutant leads to the highly reactive electrophilic oxocarbenium ion intermediate being quenched by a nearby nucleophilic side chain. Regardless, an active site configuration other than of the observed crystal structure would have to be accessible in order to account for the observed remote site of labelling at the Asp 190 position. One last possibility is that LgtC does utilize an SNI-like mechanism and that the introduced mutation changes the mechanism to that of an SN2-like pathway. Frustratingly, the exact catalytic mechanism of LgtC remains ill defined. The distinct possibility that retaining glycosyltransferases utilize a mechanism that differs from the double displacement mechanism of the analogous retaining glycosidases must be considered. This notion is supported by the observed lack of conserved structural architecture in the region expected to be occupied by a catalytic nucleophile in the active sites of structurally defined glycosyltransferases (as discussed in sections 1.3.5.5.),  1.3.5.4.  and  in stark contrast to what has been observed with otherwise unrelated retaining  glycosidases, which generally have a side chain carboxylate at this position. However, in the absence of supporting evidence, should a mechanism with little chemical precedence, that contradicts fundamental principles of organic chemistry and is formally forbidden based on orbital symmetry considerations (i.e. the previously proposed SNi-like mechanism) be suggested? Is there a more rational alternative?  Consideration of the  reactivity and associated free energy differences derived from the inherent differences between glycosyltransferase and glycosidase substrates, as well as analogies of chemistry and conserved active site structural features of many retaining glycosyltransferases with  138  FPP synthetases that use an S N I - l i k e mechanism, support the possibility of an ionic mechanism. It is possible that some retaining glycosyltransferases use a distinctively SN2-like mechanism involving formation of a covalently bound glycosyl-enzyme intermediate and others use a distinctively S N I - l i k e mechanism involving formation of an ionic oxocarbenium ion intermediate. More likely, however, a mechanistic continuum exists for this class of enzymes analogous to that of non-enzymatic nucleophilic substitution reactions.  In non-enzymatic processes, the S N 2 direct displacement and S N I ionic  mechanisms are simply the extremes of a continuum.  These classic terms used to  describe the extremes of nucleophilic substitution reactions are defined simply by the order of the reaction. In the extreme ionic mechanism, there is absolutely no interaction of the incoming nucleophile with the reactant in the transition state for leaving group bond cleavage, making it a first order reaction, and nucleophilic attack is possible from either face of the ionic intermediate leading to the formation of a mixture of stereochemical products. In the extreme S N 2 mechanism, the covalent bond formation with the incoming nucleophile and the breaking of bonds to the leaving group occur in a synchronous fashion, with nucleophilic "push" facilitating the reaction, making it a second order reaction and leading to complete inversion of stereochemistry in the products. In between these extremes are the nucleophilic substitution mechanisms, which should more accurately be strictly defined using the IUPAC recommended designations for  reaction mechanisms (Appendix A)(Guthrie and Jencks, 1989; Jencks, 1996),  involving ion pair formation that can have the kinetic properties of an S N 2 process yet  139  produce stereochemical outcomes consistent with an SNI reaction (and vice versa)(Carey, 1990; Jencks, 1980; Winstein et al., 1956) One  such  mechanism,  that  seems  quite  likely  for  many  retaining  glycosyltransferases including LgtC, is that of an ionization mechanism leading to the formation of a positively charged intermediate species, in intimate contact with an anionic leaving group counter ion and having a lifetime longer than that of a molecular vibration, that requires strong back side nucleophilic participation (without discrete covalent interaction) for formation (Figure 3.14). This ion pair could collapse to give back starting material or, following a slight shift in the active site positioning of the cation, could undergo front side attack by an incoming nucleophile leading to a product with retained stereochemistry.  This mechanism would be defined as  D *A sip N  N s  !  where the asterisk  represents the formation of a short lived intermediate and "ssip" represents the formation of a "solvent separated" ion pair. Although the term "solvent separated" seems unusual in the context of an enzymatic mechanism, it is required to differentiate from an "intimate" ion pair that would not allow for attack from the front face.  Indeed,  considering the reactivities of the substrates undergoing reaction, the observed lack of conserved structural architecture amongst retaining glycosyltransferase P-face active site regions and the biological precedence for cationic mechanisms in chemically similar systems, such a mechanism, involving the formation of an oxocarbenium ion species stabilized on the back face by an enzymatic functional group and on the front side by the anionic (substituted) phosphate leaving group seems rather attractive for retaining glycosyltransferases.  140  J W V W U  NH  2  «+  6+  NH  «+ NH  ©  2  .O  2  HO  HO  0  -  H  I  —O \  ,o—P=0 *»''  H-0  ©O  I  .0—P=0  f  ,0—P=0  •  OR  OR  Enzyme Substrate Complex  Oxocarbenium Ion-Like Transition State  I OR  Stabilized Oxocarbenium(Substituted) Phosphate Short-Lived Ion Pair Intermediate Shift In Position of Cation Within Active Site  iAJVUW\<  </\AAAA/b  8-1 OS  8+ NH  8+ NH  X  :  2  ©  HO  HO  0 O—P=  .0—P—o OR  Figure  M  Enzyme Product Complex  3.14.  Proposed  Oxocarbenium Ion-Like Transition State  DN*AN SJP S  OR  Stabilized Oxocarbenium(Substituted) Phosphate Short-Lived Ion Pair Intermediate  nucleophilic substitution reaction for retaining  glycosyltransferases.  141  CHAPTER 4  GLYCOSYLTRANSFERASE PROTEIN ENGINEERING  142  4.1. Summary The ability to alter the nucleotide specificity of a nucleotide sugar-utilizing glycosyltransferase was examined using a rational engineering strategy with potential general applicability amongst glycosyltransferases. Facile control of glycosyltransferase nucleotide specificity could provide a useful starting point for the development of chemical genetics tools for the in vitro exploration of glycosyltransferase activity. Upon identification and replacement of a suitably disposed conserved side chain within the LgtC active site and chemical synthesis of an alternative nucleotide sugar substrate that was thought would complement the new active site configuration, the donor substrate specificity of wild-type and mutant forms of the enzyme was investigated. For the purpose of the development of a high throughput screen for application in the directed evolution of glycosyltransferases, the ability to display LgtC as a pill fusion protein on M l 3 bacteriophage was explored. In addition, it was tested whether phage display procedures could be conducted in the context of in vitro compartmentalization using water-in-oil emulsification.  143  4.2. Engineering the Nucleotide Specificity of LgtC 4.2.1. Background As discussed in Chapter 1, the majority of glycosyltransferases utilize nucleotide sugars as donor substrates. Interestingly, relatively few of the possible combinations of pairings of nucleotides and monosaccharides commonly exist in nature (Table 4.1).  Table 4.1. Common forms of nucleotide sugar donor substrates found in Nature and utilized by glycosyltransferases (Varki, 1999).  Monosaccharide  Activated Nucleotide Form  Glucose Galactose /v-Acetylglucosamlne ^-Acetylgalactosamine  UDP sugar  Glucuronic acid Xylose Mannose  GDP sugar  Fucose Sialic acid  CMP Sialic Acid  * Listed are the most common nucleotide sugars. Rare forms such as ADP Glucose are typically found In plants and microbes.  A consequence of this phenomenon is the fact that, despite the existence of various genes encoding glycosyltransferases that are responsible for the transfer of a given monosaccharide to a given acceptor in a spatially and temporally specific context, virtually all glycosyltransferase gene products use an identical nucleotide sugar donor  144  substrate. For example, sialyl-linkages terminate mammalian glycoconjugate structures and are limited to one of three possible regio-chemistries (alpha-2,3- to galactose, alpha2,6- to galactose, N-acetylglucosamine, or N-acetylgalactosamine, or alpha-2,8- to another sialic acid). However, despite the relatively low number of possible linkages, at least 20 different sialyltransferases are encoded by the human genome all of which use the same C M P sialic acid donor substrate (Harduin-Lepers et al., 2001). Because of redundancy in acceptor substrate specificity and the fact that each sialyltransferase gene is differentially expressed in a tissue-, cell type- and stage-specific manner, very little is known about the exact role that individual gene products play in normal and pathological physiology in vivo. Studies that make use of non-specific inhibitors or gene deletions have yielded either conflicting results or results that are difficult to interpret. This difficulty in delineating the roles of individual gene products in a stagespecific manner using either chemical or genetic tools individually is a common problem in cell biological studies.  Small molecules typically suffer from a lack of specificity  when they affect the activity of a population of enzymes that use an identical substrate. Conversely, it is often difficult to use gene-inactivation at the level of D N A to address critical temporal issues of function. Chemical genetics strategies that combine both the advantages in target specificity provided by genetics and the temporal control provided by small molecules are of tremendous value in cell biology research (Alaimo et al., 2001; Holt et al., 2002; Ubersax et al., 2003). Chemical complementation of a mutant form of a glycosyltransferase of altered nucleotide specificity with an alternate nucleotide sugar substrate or an inhibitor with an altered purine/pyrimidine scaffold would provide an  145  effective starting point for the development of a chemical genetics strategy to explore glycosyltransferase activities. It is generally believed that glycosyltransferases are highly specific for the nucleotide sugar donor substrates that they use.  The few studies that have actually  explored this issue support such a notion (Baisch et al., 1996; Khaled et al., 2004; Ohrlein, 1999). For example, the GDP fucose-utilizing fucosyl transferase FucT-III was found to utilize A D P fucose -1000 fold less efficiently in terms of k /K cat  M  (Khaled et al.,  2004). A very recent study, describing a GT-B fold polyketide glucosyltransferase (SorF) that possesses promiscuous donor substrate specificity, has challenged this notion (Kopp et al., 2007). Despite a lack of detailed kinetic analysis, the authors report the ability of SorF to utilize dTDP glucose as efficiently as the natural donor substrate U D P glucose. Importantly for the engineering strategy described below, the nucleoside moieties of these two substrates differ only by the presence of a methyl substituent on the pyrimidine ring. Previous examples of rational protein engineering approaches to broadening or changing the substrate specificities of glycosyltransferases have involved modification of residues involved with the binding of substituents on the sugar ring. For example, the R228K (Ramakrishnan et al., 2005) and Y289L (Ramakrishnan and Qasba, 2002) mutations to lactose synthase were found to relax donor specificities with respect to U D P Gal/UDP Glc and U D P Gal/UDP GalNAc relationships, respectively. Computational modeling was used to identify an H308R mutant of GlcAT-I that was found experimentally to utilize U D P Glc, U D P Man, and U D P GlcNAc with nearly identical efficiency of that of the natural U D P GlcA donor (Ouzzine et al., 2002). Structural and sequence data were used to significantly increase the ability of ToxB, for which the  146  natural donor substrate is UDP Glc, to utilize an alternative UDP GlcNAc donor substrate (Jank et al., 2005).  Similar examples exist for GT-B fold glycosyltransferases  (Hoffmeister  2001;  et  al.,  Hoffmeister  et  al.,  2002;  Kubo  et  al.,  2004).  Glycosyltransferase engineering was the topic of an excellent recent review to which the reader is directed (Hancock et al., 2006).  4.2.2. Mutation of a Conserved Active Site Side Chain Alters the Base Specificity of GTP Binding Proteins To explore the base specificity of the GTP binding protein elongation factor Tu (EF-Tu), Hwang and Miller generated an array of point mutants and tested the ability of these variants to bind alternative bases (Hwang and Miller, 1987). The most interesting mutation identified was that of D138N, which was found to induce a complete change in base specificity from guanosine to xanthosine.  These two bases differ only by the  presence of an amino group in the case of guanosine versus a carbonyl substituent in the case of xanthosine at the 2-position of the purine ring (Figure 4.1). The ability to control the nucleotide specificity of EF-Tu was subsequently used to develop a detailed model for the interaction of EF-Tu and the ribosome (Weijland and Parmeggiani, 1993). It was rationalized that the switch in specificity for the D138N mutant results from a change in hydrogen bonding potential.  In the case of the wild-type enzyme, the side chain  carboxylate of Aspl38 can act as a hydrogen bond acceptor and the 2-amino group of guanosine a hydrogen bond donor.  Conversely, in the D138N mutant the side chain  amide of N138 can act as a hydrogen bond donor and the 2-carbonyl of xanthosine as a hydrogen bond acceptor (Figure 4.1).  147  A  B N  <N  N  NH N  R  N H  Figure 4.1.  <N R  NH N H  O *  H l  Hydrogen bonding interactions between guanosine and a side chain  carboxylate (A) or xanthosine and a side chain amide (B) determine the base specificity ofEF-Tuand FtsY.  This supposition is supported by the three-dimensional X-ray crystal structure of GTP bound to EF-Tu, which clearly shows the side chain carboxylate of D138 within hydrogen bonding distance of the 2-amino group of guanosine (Figure 4.2A). In fact, this aspartate residue is completely conserved amongst GTP binding proteins. The generality of this approach for the rational control of nucleotide specificity was subsequently demonstrated by successful application to the study of the E. coli homologue of the eukaryotic signal recognition particle receptor FtsY (Powers and Walter, 1995).  FtsY  contains a GTP binding domain for which the nucleotide specificity could be altered from guanosine to xanthosine by the simple introduction of a D441N point mutation. Again, the hydrogen bonding interaction between the side chain carboxylate of D441 and the 2amino group of guanosine is supported by the X-ray crystal structure of the T. aquaticus homologue of FtsY with bound GDP (Figure 4.2B).  148  Figure 4.2.  Hydrogen bonding interactions between guanosine and a side chain  carboxylate that are found to control base specificity between guanosine and xanthosine in the case of EF-Tu (A) and FtsY (B). Similar hydrogen bond partners are found with other nucleotide binding proteins. The 4-carbonyl oxygen of UDP GlcNAc hydrogen bonds with a main chain amide in the case of UDP Glc NAc enolpyruvyl transferase (C). The 4-amino group of C M P hydrogen bonds with a side chain carboxylate in the case of C M P kinase (D).  Protein Data Bank ID numbers are indicated and distance  measurements are reported in Angstroms.  It would seem reasonable that this approach of using the hydrogen bonding partnering of a carboxylate with an amino group or an amide with a carbonyl group could be extended to control the nucleotide specificity of other proteins.  In the case of  glycosyltransferases, a useful application would be the ability to alter the specificity of UDPbinding domains (Table 4.1). The most logical choice for base switching would be  149  from uridine to cytosine. It would be predicted that in the case of uridine binding proteins an amide from the protein would form a hydrogen bond interaction with the 4-carbonyl oxygen of the uridine ring (Figure 4.3A).  Inspection of the three-dimensional X-ray  crystal structures of various uridine binding proteins reveals that such an interaction is common (Kostrewa et al., 2001; Lariviere et al., 2003; Skarzynski et al., 1998; Thoden et al., 1997)(for example Figure 4.2C). Conversely, in the case of cytosine binding proteins, it would be predicted that a side chain carboxylate would form a hydrogen bond interaction with the 4-amino group of the cytosine base (Figure 4.3B). Again, inspection of structures of cytosine binding proteins with bound ligand reveals the occurrence of such an interaction (Bertrand et al., 2002)(for example Figure 4.2D). Based on the occurrence of these interactions it would be predicted that the nucleotide specificity of a UDP sugar-utilizing glycosyltransferase could be readily altered to that of a CDP sugarutilizing enzyme by changing a single active site amide side chain to a carboxylate. Using the U D P Gal utilizing enzyme LgtC as a model system, the feasibility of this nucleotide specificity switch approach was investigated.  A  R  R  Figure 4.3. Hydrogen bonding interactions between uridine and a side chain amide (A) or cytidine and a side chain carboxylate (B) proposed to determine base specificity.  150  4.2.3. Identification and Modification of a Base Specificity Conferring Site Within the LgtC active site Fortunately, in the case of LgtC a three-dimensional X-ray crystal structure of the enzyme is available with UDP 2F Gal bound within the active site (Persson et al., 2001). In addition, the structure of glycogenin, another GT family 8 member, is available with bound UDP Glc (Gibbons et al., 2002). Inspection of both of these structures reveals the occurrence of a conserved interaction between the 4-carbonyl group of the uridine ring with a side chain amide within the UDP binding domain (Figure 4.4).  Figure  4.4. Active sites of the G T family 8 members glycogenin (A) and LgtC (B)  revealing the conserved interaction of a side chain amide with the 4-carbonyl group of the bound uridine ring.  Protein Data Bank ID numbers are indicated and distance  measurements are reported in Angstroms.  It was therefore proposed that modification of the side chain amide of AsnlO of LgtC to a carboxylate would switch the nucleotide specificity of the enzyme from a UDP Gal to a CDP Gal utilizing enzyme. The N10D point mutation was introduced into LgtC and the resulting variant was expressed in E. coli yielding quantities of protein similar to that of the wild-type enzyme.  151  4.2.4. Synthesis of CDP Gal Initial studies with a U D P Gal synthesizing pyrophosphorylase enzyme from Streptococcus  pneumoniae (spGalU)(Mollerach et al., 1998) indicated that at high  enzyme and nucleotide concentrations, CTP could be substituted for UTP to generate CDP Gal. However, under preparative scale conditions, presumably due to issues with product inhibition, sufficient quantities of the desired product could not be generated using either UTP or CTP. Therefore, the unnatural nucleotide sugar CDP Gal had to be chemically synthesized using a traditional morpholidate coupling protocol.  CMP  morpholidate is not commercially available and had to be generated prior to coupling with a-D-galactose 1-phosphate (Scheme 4.1). Following H P L C purification, the desired nucleotide sugar substrate was obtained in yields typical for this coupling technique.  NH  ^  2  N  ^0 Morpholine, D C C  HO—P —  I  OH  /  ~  \  J J _  Q  1 ^  tBnOH,H 0 reflux 2  OH  OH  "0  OH  OH  13 6(611*)  OH  Galactose 1 phosphate (pyridinium salt) tri-n-octylamine, pyridine, RT4days  ^OH  HO  N  X.  OH  O =P —O —P I I  o-  o-  OH  OH  H P L C purified - 88 m g ( 4 3 J % )  Scheme 4.1. Chemical synthesis of the desired unnatural nucleotide sugar CDP Gal.  152  4.2.5. Testing the Nucleotide Sugar Specificity of Wild-type and N10D LgtC Analysis of analytical scale reactions by T L C indicated that at high enzyme concentrations (~0.5 mg/mL) both the N10D and wild-type versions of LgtC were able to use CDP Gal as an alternative donor sugar substrate. However, it was found that under similar conditions the N10D version of LgtC was able to use the natural donor substrate UDP Gal. Thus it appeared that specificity had been broadened rather than switched. To determine whether a true and useful change in nucleotide sugar preference had occurred as a result of the amide to carboxylate mutation, detailed kinetic analysis of the wild-type and N10D versions of LgtC using either UDP Gal or CDP Gal as the donor substrate was performed using the P K / L D H enzyme-coupled continuous assay (Table 4.2).  Table 4.2. Michaelis-Menten kinetic parameters for wild-type and N10D versions of LgtC using either CDP Gal or U D P Gal as the varied substrate at a fixed saturating concentration of lactose acceptor (160mM). Donor  WT LgtC-25 kcaJK  kcat (min )  (MM)  UDP Gal  960  CDP Gal  n.d.  1  N10D LgtC-25 kcaJK-m  (min" .mM' )  kcat (min")  Km (MM)  25  38  250  90  3.0  n.d.  1.2  n.d.  n.d.  0.25  m  UDP Gal/ CDP Gal  1  32  1  (min" .mM' ) 1  1  12  * error in data is between 5 -15%  153  Inspection of the kinetic data clearly reveals that the desired nucleotide specificity switch from a UDP Gal to a CDP Gal utilizing enzyme was not achieved by mutating the side chain amide of AsnlO to a carboxylate. Both the wild-type and N10D versions of LgtC use U D P Gal in preference to CDP Gal. In fact, the N10D mutation results in a decrease in activity with respect to CDP Gal (-four-fold). A slight (-three-fold) decrease in the preference for U D P Gal over CDP Gal was achieved with the N10D mutation. However this results primarily from the effect of decreasing the natural activity of the enzyme to a greater degree (-13-fold decrease in activity with respect to U D P Gal) than the decrease in activity for the alternative activity (-four-fold decrease in activity with respect to CDP Gal).  4.2.6. Concluding Remarks Although the results of this rational engineering approach for the facile control of glycosyltransferase nucleotide sugar substrate specificity are somewhat disappointing, significant insights and directions for future directions were obtained. Firstly, an essential assumption at the onset was that glycosyltransferases possess a stringent nucleotide sugar donor substrate specificity. The most significant finding of this study is the fact that LgtC is rather promiscuous with respect to the nature of the pyrimidine base moiety of the donor substrate it will tolerate (only a -40-fold preference for UDP Gal versus CDP Gal). Whether this is a general theme amongst glycosyltransferases remains to be investigated, however the results provided here challenge the generally held view of strict donor substrate specificity. It is likely that this strategy for changing nucleotide specificity  154  would be better suited to glycosyltranferases that posses a monogamous relationship with respect to their donor substrates. A possible explanation for the lack of success in altering the nucleotide specificity of LgtC might be the inherent importance of nucleotide binding to the structure and mechanism of the enzyme.  A general theme amongst glycosyltransferases is the  observation that, upon binding the nucleotide sugar donor substrate, significant conformational changes occur that lead to the formation of the active conformation of the enzyme. Perturbing this critical region of the enzyme could lead to the undesirable effect of destabilizing the active form and/or catalytic efficiency of the enzyme. Indeed, the N10D mutation leads to deleterious effects in terms of both K a n d k M  cat  with respect to the  natural reaction using U D P Gal. In order for such a strategy to be successful, it is quite possible that secondary site "suppressor" mutations are necessary to rescue the effect of the specificity switch. Indeed this has been found to be the case with chemical genetics strategies involving several A T P utilizing protein kinases (Zhang et al., 2005). Future strategies to alter the nucleotide specificity of glycosyltransferases will benefit from such considerations.  155  4.3.  Compartmentalized Phage Display of LgtC  4.3.1. Background The F f class (Genus Inovirus) of filamentous phages (fl, fd and M l 3 ) consist of a circular single-stranded D N A (ssDNA) genome housed within a protein capsid cylinder. As their name implies, they specifically recognize the tip of the F conjugative pilus of E. coli containing the F ' plasmid (F' strains).  They are non-lytic phages and upon  productive infection they do not kill their host. Using a mixture of phage- and bacteriumencoded components, single-stranded viral D N A is replicated to a double-stranded intermediate that is used as the template for the production of new phage proteins. In addition, newly synthesized single-stranded viral D N A is generated and maintained by the formation of a complex with multiple copies of single-stranded D N A binding proteins encoded by the phage genome. Viral capsid proteins exist as integral membrane proteins until they are assembled around a newly synthesized phage genome that is extruded out of the bacterial cell through a "membrane-associated assembly site" with concomitant removal of single-stranded D N A binding proteins, thereby giving rise to a newly formed organism. Infected bacteria tolerate the viral replication process with minimal effect, having replication times of just 50% longer than that of uninfected cells and with the generation of -100 particles per generation. Indeed, the replication process can proceed for many generations leading to virus titers of up to 10 per mL of culture. 13  Having a mass of-16.3 M D , an Ff phage particle is a protein-based cylinder (-6.5 x 930 nm) surrounding a closed circular ssDNA genome that consists of 6400 nucleotides.  The entire length of the particle consists of -2700 molecules of the 50-  residue major pVIII coat protein (Figure 4.5). The 78-nucleotide hairpin packaging signal  156  of the genome is always oriented to the capping end that consists of five molecules each of the 33-residue VII and 32-residue pIX coat proteins. The other end of the cylinder consists of five copies each of the 406-residue pill and 112-residue pVI coat proteins. The pill coat protein is required for the F conjugative pili recognition events required for infection. Other proteins encoded by the phage genome are required for D N A replication (pll and pX), phage assembly (pi, pIV and pXI) or binding to ssDNA (pV). A l l F f phage genomes encode virtually identical amino acid sequences for the eleven gene products described above.  Interestingly, two of the gene products (pX and pXI) result from  translation starting at an internal methionine codon (Guycaffey et al., 1992; Rapoza and Webster, 1995). For a detailed description of the F f phage infection process, assembly, life cycle, biology and genetics the reader is directed to the appropriate literature (Barbas III, 2001).  000 Packaging Sequence(PS)  Phage Genome  pVIII  Figure 4.5.  pVH  pIX  PVI  Graphical representation of the macromolecular anatomy of an F f  filamentous phage particle.  157  Peptides or entire proteins can be displayed on the surface of F f phage by fusion to a coat protein. This was first described by George Smith in 1985 wherein fragments of the endonuclease EcoRI were fused to the N-terminal portion of pill, resulting in the production of chimeric proteins that were packaged into phage particles (Smith, 1985). Multiple systems have since been tailored for the display of peptides and proteins on the Ff phage surface (Smith and Petrenko, 1997; Smith and Scott, 1993). Phage display systems are classified based on the resulting arrangement of coat proteins (Smith, 1993). This classification system is illustrated using pill and pVIII fusions in Figure 4.6 and a brief description for fusions to pill follows. In a type 3 system, the phage genome is modified to contain a single version of recombinant gill existing as a fusion to an amino acid sequence of interest. In theory, the foreign amino acid sequence is displayed on all five copies of the pill coat protein, thereby greatly reducing infectivity. In a type 3 3 system, the phage genome contains two copies of the gill, encoding two different versions of the pill protein.  One version is the wild-type gene and the other is a  recombinant encoding a fusion of the amino acid sequence of interest fused to gill. Particles arising from this system contain a mixture of wild-type and fusion pill proteins on their surface. Finally, in the 3+3 system, like in the 3 3 system, two gene versions of gill are present.  However, in this case, a wild-type version is encoded by a phage  genome (described as the "helper" phage) and a recombinant version is encoded by a type of plasmid known as a "phagemid". A phagemid carries an antibiotic resistance gene, both E. coli and filamentous phage origins of replication and the 78-nucleotide packaging sequence that facilitates its incorporation into newly forming progeny. In this system, an F' strain of E. coli is transformed with the phagemid, cultured under antibiotic selection,  158  and then the phage genome is introduced by superinfection with helper phage. In modern helper phage systems, the packaging sequence of the helper phage genome is disrupted to prevent packaging into newly formed progeny allowing for the production of a homogenous population of phage harboring phagemid D N A and displaying copies of both the wild-type and fusion versions of the pill coat protein on their surface. The use of phagemid/helper phage systems is particularly well-suited to the production and screening of large libraries of protein and peptide variants. Usage of a 3+3 system is suited to the "low copy" display (-one fusion protein per particle) of whole proteins, while the 8+8 system is suited for the polyvalent display of shorter peptides.  Type 3  F i g u r e 4.6.  Type 3 3  Type 3+3  Type 8  Type 8 8  Type 8+8  Schematic representations of particles resulting from various phage display  systems available for the surface display of foreign peptides or entire proteins.  Phage display has become an invaluable tool in screening procedures for the in vitro evolution of protein function, wherein active protein variants are selected from a  159  large pool of random mutants thereby mimicking the natural selection process. Although library sizes are still limited by transformation efficiencies (<10 ), unlike traditional 9  screening or selection strategies performed in bacterial or yeast cells, phage display allows complete control over reaction conditions (e.g. buffer, pH, metals, substrates, competing substrates, etc.).  Phage display has proved to be particularly useful in the  engineering of antibodies and other binding proteins with desired binding affinities. In these cases, screening is easily achieved by panning phage displaying protein variants over immobilized ligands and enriching those with desired binding properties. Limiting the application of phage display in the directed evolution of enzymes is the difficulty in obtaining a physical linkage between a phage that displays an active enzyme variant and the catalytic activity (i.e. product formation). The first applications of phage display to the evolution of catalytic activity made use of similar approaches to those used for the evolution of binding affinities.  Variants of catalytic antibodies or protein enzymes  displayed on phage were screened based on their ability to bind immobilized transition state analogues or suicide substrates (Fernandez-Gacio et al., 2003).  Schultz and co-  workers provided the first report of a phage display-based screening methodology for the directed evolution and functional cloning of enzyme activity based on turnover (Pedersen et al., 1998). Since then several other strategies have been reported, all of which involve creating a physical linkage between a substrate and a phage particle and rely on proximity effects to ensure that product formation is the result of the enzyme linked to the same phage particle (Atwell and Wells, 1999; Cesaro-Tadic et al., 2003; Demartis et al., 1999; Fa et al., 2004; Jestin et al., 1999; Strobel et al., 2003; Xia et al., 2002; Y i n et al., 2004).  160  Clearly, a phage display-based strategy that does not rely on proximity effects and is based on turnover would be a useful development in the field. One of the most significant recent developments in the field of directed evolution has been in vitro compartmentalization (IVC) procedures based on water-in-oil emulsions. Such strategies elegantly mimic those used by nature to link a genotype with a phenotype.  In nature, lipid bilayer membranes of cells or organelles are used to  compartmentalize a genotype with gene products, thereby facilitating a physical link between inheritable genotypes and selectable phenotypes. Similarly, the biomimetic IVC strategy involves the dispersion of a water phase in an oil phase and results in the creation of aqueous droplets (~5 fL each) that effectively restrict diffusion between compartments. Concentrations are chosen such that individual compartments contain on average less than a single gene of interest, thereby effectively recreating cell-like conditions in which transcription, translation and protein activity all take place within an isolated compartment.  The system has the advantages of being high throughput  (>10  10  phenotypes per 1 mL), economical, easily performed, and being amenable to a range of reaction and temperature conditions. Application of IVC for the directed evolution of enzyme activity was first reported for the selection of a D N A methyltransferase (Tawfik and Griffiths, 1998) and subsequent reports of application to the evolution of other D N A modifying enyzmes rapidly ensued (Cohen et al., 2004; Doi et al., 2004; Ghadessy et al., 2001; Ghadessy et al., 2004; Lee et al., 2002b). Application of IVC to a non-DNA modifying enzyme has also been reported and involves two compartmentalization steps (Griffiths and Tawfik, 2003). A l l of these  161  procedures required in vitro transcription/translation of the protein variants, a procedure notoriously challenging to perform with a high degree of consistency and reliability. More recently, a screening methodology has been developed that involves the use of double "water-in-oil-in-water" emulsions that can be directly sorted using fluorescence activated cell sorting (FACS) (Aharoni et al., 2005; Bernath et al., 2004). In this F A C S based procedure, product formation is detected by the production of a fluorescent signal. Individual cells expressing a unique protein variant are isolated within a compartment that prevents the diffusion of the fluorescent signal used to sort active from inactive variants. If the fluorescent signal indicating product formation is retained within a cell, direct F A C S can be performed and emulsification is not necessary. Very recently, just such an approach (non phage-based) was described for the directed evolution of the sialyltransferase Cst II using a lactose acceptor conjugated to a Bodipy fluorophore (Aharoni et al., 2006). Because the resulting 3'-sialyl lactose product is not exported by lactose permease, the enzyme product, and therefore the fluorescent signal, is retained within the E. coli cell following extensive washing steps thereby facilitating the direct sorting of cells harboring active protein variants. The FACS-based screening approaches described above require product formation to be coupled to a fluorescent signal either resulting from the release of a fluorescent product or by the use of a substrate that has been modified with a fluorescent substituent. For many enzyme reactions, particularly those catalysed by transferase enzymes, fluorogenic substrates are not available. Considering the adage that "you get what you select for", a strategy that involves the use of substrates modified with a fluorophore has the potential disadvantage of screening for an activity dependent on the alternative  162  fluorophore-containing substrate. In fact, the directed evolution of Cst II resulted in the isolation of a protein variant with an evolved fluorophore binding site and increased activity was only detected using substrates that contained an expensive Bodipy substituent (Aharoni et al., 2006). In addition, in vivo screening procedures have the further limitation of lacking control over reaction conditions.  Finally, FACS-based  screening is a technologically intensive endeavor, limiting its application to only the more affluent of laboratories. Application of IVC to phage display would provide the field of directed enzyme evolution with a generally applicable, inexpensive and facile screening technology. Unlike previously described IVC-based approaches, the robust procedure of phage display would eliminate the need for difficult in vitro transcription/translation and would have the advantage over in vivo procedures of providing complete control of reaction conditions. In addition, the use of IVC would allow the detection of product formation without reliance on proximity effects.  Using the glycosyltransferase LgtC as model  system, it was determined whether phage display methodologies are possible in the context of IVC and i f the coupling of these two procedures could be used to devise a screening strategy for application in directed enzyme evolution.  4.3.2. Phage Display of LgtC as an M13 pill Fusion Protein A modified version of the commercial pBluescript S K (-)™ phagemid designated pSJF6 was obtained from the laboratory of Dr. Warren Wakarchuk at the National Research Council of Canada Institute for Biological Sciences in Ottawa. The multiple cloning site (MCS) of pSJF6 contains a D N A sequence encoding an N-terminal OmpA  163  leader sequence fused to the pill coat protein of M l 3 bacteriophage via a Gly/Ser rich linker that also contains a c-myc detection sequence.  The lack of an appropriate  restriction site between the OmpA sequence and the linker necessitated a subcloning strategy in which LgtC was P C R amplified using a forward primer encoding the entire OmpA leader sequence and then inserted into pSJF6 via EcoRI and Bglll restriction sites. However, presumably due to the large size of the forward primer (-100 bp), the quantity of PCR product obtained was insufficient for subcloning. In order to increase the amount of D N A encoding the OmpA-LgtC fusion, a nested PCR was performed using the small amount of initial P C R product as template and the resulting nested P C R product was successfully cloned into pSJF6. Sequencing the resulting product revealed a seven base pair deletion in the OmpA leader sequence that presumably arose during the initial P C R amplification.  The deleted seven base pairs were readily reintroduced using the  restriction enzyme Aarl.  D N A sequencing was used to confirm the presence of the  desired D N A consisting of a contiguous N'-OmpA-LgtC-linker-pIII-C encoding sequence within the M C S , and the resulting phagemid was designated pLLOl (Figure 4.7).  The introduction of an Ncol cut site at the start of the LgtC sequence makes pLLOl  a useful vector for the creation of other pill fusion proteins.  In fact, pLLOl was  successfully used by Dr. David Fox to generate a phagemid for type 3+3 display of the xylanase Bex by subcloning Bex into pLLOl via Ncol and Bglll  restriction sites  (unpublished results).  164  EcoRI Ncol  OmpA Figure  Bglll  LgtC  Linker  pill  4.7. Recombinant construct within the M C S of p L L O l encoding an LgtC-pIII  fusion protein.  To determine whether LgtC remained catalytically active when expressed as a C terminal p i l l fusion protein, activity tests were performed on cell lysates derived from IPTG-induced cultures o f E. coli harboring p L L O l .  A s determined by T L C analysis,  L g t C remained active upon fusion to the p i l l M l 3 phage coat protein.  A small scale  PEG-precipitated preparation o f putative LgtC-displaying phage was performed on phage derived from the X L 1 Blue ( f ) strain o f E. coli harboring p L L O l superinfected with V C S M 1 3 helper phage.  The resulting phage progeny were quantified using U V  absorbance, and LgtC activity was tested by T L C analysis using fluorescein lactose conjugate as acceptor (Figure 4.8). Encouragingly, this analysis revealed that M l 3 phage derived from E. coli harboring p L L O l did possess LgtC activity comparable to that o f wild-type (non fused) LgtC indicating the successful display o f active enzyme as a p i l l fusion.  It must be noted that the LgtC activity observed from twice P E G precipitated  phage may have resulted from fusion proteins transported to the periplasm but not incorporated into phage that had co-precipitated with phage particles.  Although this  interpretation cannot be completely ruled out, several facts would argue against it.  It  would be a significant coincidence that the level o f LgtC activity resulting from nonincorporated fusion protein would directly correlate with the concentration o f displayed L g t C determined based on the concentration o f phage particles.  In addition, bacterial  165  cells are removed by l o w speed centrifugation prior to P E G precipitation o f phage from the resulting supernatant. In the absence o f lysozyme, it is difficult to rationalize how an appreciable quantity o f fusion protein not incorporated into phage particles would be excreted past the outer membrane. A t the time, this was the first glycosyltransferase to be successfully displayed on the surface o f phage.  Walker and co-workers subsequently  reported the successful display o f the glycosyltransferase M u r G (Love et al., 2006).  «  _  _  »  W  A  Figure  4.8.  B  C  . •*  Fluorescein Lactose  D  T L C analysis (7:2:1 E t O A c : M e O H : H 0 ) o f the L g t C activity o f twice P E G 2  precipitated M l 3 bacteriophage derived from p L L O l phagemid.  Reactions were run  using 100 m M H E P E S (pH 7.5), 5 m M M n C l , 5 m M D T T , 5 m M U D P G a l , 0.5 m M 2  fluorescein lactose and either ~30 n M purified L g t C ( A ) , - 1 0 n M M l 3 phage derived from p L L O l (B), - 1 0 n M M l 3 phage derived from pSJF6 (C), or without enzyme source (D).  Western blot analysis o f p L L O l -derived phage was performed to confirm the display o f LgtC-pIII fusion coat protein using rabbit anti c-myc Ab-horse radish peroxidase ( H R P ) conjugate for detection (Figure 4.9). This analysis reveals the presence o f a p i l l fusion protein (derived from p L L O l ) that differs from that o f the non-fused p i l l protein (derived from pSJF6). It must be noted that the theoretical mass o f the LgtC-pIII fusion protein is 55.2 k D a and the major band derived from the putative fusion protein, although clearly larger than non-fused p i l l , appears less than 50 k D a . This is highly  166  suggestive that the LgtC-pIII fusion has been subjected to proteolytic degradation, a frequent concern in phage display.  Fortunately, intact LgtC-pIII with an apparent  molecular weight o f - 5 5 k D a could also be detected as a major band in other experiments (as discussed i n section 4.3.3). It must also be noted that non-fused p i l l protein, derived from pSJF6, migrates on an S D S P A G E gel with an apparent mass that differs from the actual mass o f 23.6 k D a due to the presence o f a large number o f positively charged residues within the protein.  LgtC-c-myc-plll  c-myc-plll  F i g u r e 4.9.  Western blot analysis o f various phage particles using anti-c-myc-HRP for  detection. Lysate from induced cultures o f E. coli harboring pSJF6 was used as a positive control (lane 1). V C S M 1 3 helper phage was used a negative control (lane 7). Molecular weight standards (lane 2) were written onto the Western blot by overlaying onto the S D S P A G E gel. The molecular weights o f these standards are 108, 90, 50.7, 35.5, and 28.6 kDa.  P E G precipitated phage were derived directly from E. coli harboring either p L L O l  (lane 3) or pSJF6 (lane 4) or in the context o f water-in-oil emulsions from E. coli harboring either p L L O l (lane 5) or pSJF6 (lane 6).  167  4.3.3. Compartmentalized LgtC Phage Display Having demonstrated that catalytically active LgtC can be displayed on M l 3 bacteriophage as a pill fusion, it was determined whether this phage display procedure could be successfully conducted in the context of water-in-oil emulsions. X L 1 Blue (f) E. coli harboring pLLOl were emulsified from an aqueous solution containing all of the components necessary for LgtC activity and VCSM13 helper phage.  As a positive  control, LgtC-displaying phage were generated and left to incubate under identical reaction conditions but without being subjected to the emulsification procedure.  To  obtain conditions in which on average less than one cell, and as such no more than one LgtC gene, was contained within each compartment, ~10 cells were emulsified in ~10 8  9  compartments. Washed cells derived from overnight culture were resuspended in 50 m M HEPES (pH 7.5) containing 5 m M M n C b and 10 ug/mL ampicillin and immediately prior to the addition of ten volumes of an ice cold mixture of AbilEM90™ in light mineral oil (2.9% w/w), a ten-fold concentrate of LgtC reaction components (50 m M DTT, 50 m M U D P Gal, 5 m M fluorescein lactose conjugate) and 2 x 10  10  VCSM13  helper phage were added to the resuspended cells. Emulsification on ice for five minutes at 9500 rpm using an I K A T-25™ homogenizer resulted in the production of a milky white suspension that was incubated at 37°C overnight. The following day, the mixture was centrifuged (14, 000 rpm) and the oil phase was removed. The aqueous layer was diluted five-fold with a volume of 50 m M HEPES (pH 7.5) and washed twice with watersaturated diethyl ether. Remaining ether was immediately removed from the aqueous phase using a Speed Vac. Detection of LgtC activity that occurred within compartments was confirmed by T L C analysis of the aqueous phase immediately following  168  centrifugation (Figure 4.10).  In addition, LgtC activity of phage isolated by P E G  precipitation from aqueous phases was tested by TLC analysis (Figure 4.10). Isolated phage generated in emulsions were also subjected to Western blot analysis as described above to confirm the presence of LgtC-pIII fusion proteins (Figure 4.9).  As was observed with non-emulsified phage, a significant band appears  corresponding to an LgtC-pIII fusion protein that has been subjected to proteolytic degradation. However, a major band is present that corresponds to the expected mass (55 kDa) of an LgtC-pIII fusion protein.  • ??  • • •J J J -  I •0 5 A B C  D  E  Fluorescein Lactose .LgtC Product  F  Figure 4.10. T L C analysis (7:2:1 EtOAc:MeOH:H 0) of the LgtC activity of emulsified 2  M l 3 bacteriophage derived from pLLOl phagemid. Reactions were run using 100 m M HEPES (pH 7.5), 5 m M M n C l , 5 m M DTT, 5 m M UDP Gal, 0.5 m M Fluorescein 2  Lactose and the following enzyme sources. (A) -30 n M of purified LgtC. (B) Aqueous phase of broken emulsions containing phage derived from pLLOl.  (C) Twice P E G  precipitated emulsified phage derived from pLLOl (-10 nM). (D) Aqueous phase of broken emulsions containing phage derived from pSJF6. (E) Twice P E G precipitated emulsified phage derived from pSJF6 (-10 nM). (F) No enzyme source.  169  A near identical number of phages were recovered from the aqueous phase of the emulsions compared to an equivalent volume of culture not subjected to emulsification 12  (~5 x 10  1  mL"). The infectivity of emulsified LgtC-displaying phage was lower (~4 x  10 Tu/mL from an input of ~10 mL" as) than that of non-emulsified LgtC displaying 6  8  phage (~4 x 10  7  1  Tu/mL from an input of ~10  8  mL" ). 1  For this reason, i f  compartmentalized phage display were incorporated into a directed evolution screen, to maximize the size of a library of variants, the enriched gene sequences should be recovered by lysis of enriched phages followed by PCR amplification of the gill fusion sequences. The results described in this section indicate that phages displaying active LgtC can be generated from individual E. coli cells harboring an LgtC encoding phagemid when compartmentalized using water-in-oil emulsions. Catalytic activity was detected within the compartments and the phagemids of isolated phages could be readily obtained by infection of E. coli cells.  4.3.4. A High Throughput Screen for LgtC Directed Evolution Based on Compartmentalized Phage Display A  screening methodology involving compartmentalized phage display was  devised for the directed evolution of LgtC activity. The screen was developed such that it would be high throughput, inexpensive, easy to perform, reliable and readily adaptable to other enzyme systems. Methods for creating a physical linkage between phage harboring a phagemid encoding an active protein variant and the product of the reaction being catalysed, and for enriching phage linked to the product from those linked to the substrate were required.  170  Sidhu and colleagues have reported an evolved version of the major coat protein pVIII of M l 3 bacteriophage that greatly enhances the polyvalent surface display of the oligomeric protein streptavidin (SAV) (Sidhu et al., 2000). Inclusion of this construct in the system used for successful display of LgtC as a pill fusion would result in phages that simultaneously displays LgtC (or a variant thereof) in low copy number as a pill fusion and multiple copies of S A V as a pVIII fusion on its surface. S A V is a tetrameric protein from Streptomyces avidinii that is widely used in molecular biology and biochemistry applications because of its ability to bind the vitamin biotin with one of the strongest known non-covalent interactions (Ka ~10" M)(Chaiet, 1964; Green, 1975). As such, a 15  substrate conjugated to biotin would be essentially irreversibly linked to the surface of a phage displaying S A V . Compartmentalization of an E. coli cell infected with M l 3 helper phage and harbouring a phagemid encoding a variant of LgtC fused to pill and S A V fused to pVIII under reaction conditions that include the presence of a biotin-lactose conjugate would lead to the production of multiple isolated phages simultaneously displaying an identical version of LgtC and having multiple copies of lactose acceptor linked to its surface. If the version of LgtC were catalytically active, the surface of each phage would be modified to have the trisaccharide product of the LgtC reaction on its surface.  Because all of the phage within a particular compartment would harbour the  same phagemid encoding a particular LgtC variant, this system would not rely on proximity effects to link product formation and enzyme variant on the same phage. In addition, this system would be readily adaptable to other enzyme systems by simply generating a conjugate consisting of a substrate and biotin. Following destruction of the compartments and isolation of the aqueous phase, phage harbouring phagemids encoding  171  active LgtC variants could be enriched from inactive variants by panning over a macromolecule that specifically recognizes and binds the trisaccharide product of the LgtC reaction. Derived from pathogenic bacteria, verotoxins, also known as Shiga-like toxins, are able to arrest eukaryotic protein synthesis and are known to damage endothelial cells of the kidney and brain leading to renal failure and deleterious neurological effects (Coia, 1998; Kitov et al., 2000). These toxins consist of two domains. The A domain confers activity by modifying the 28S rRNA of the 60S ribosome such that it can no longer interact with EF-1 and EF-2 elongation factors, thereby preventing effective protein synthesis (Endo et al., 1988).  The B domain is a homopentamer that specifically  recognizes eukaryotic cell receptor globotriaosylceramide (GB3) and facilitates receptormediated endocytosis of the toxic A domain (Sandvig and Vandeurs, 1994).  Each  monomer possesses three saccharide binding sites that recognize the terminal cc-Dgal(l->4)p-D-gal(l-»4)P-D-glc " P " trisaccharide of G B . Interaction of the B-domain k  3  pentamer with the methyl glycoside of the P k trisaccharide is known to be of rather low affinity (A" ~ 1 mM)(St. Hilaire et al., 1994). d  However, because of the multivalent  interaction resulting from a total of 15 protein-carbohydrate interactions on the intact pentamer, the estimated affinity of the pentamer for cells expressing GB3 on their surface is very high (A"d ~ 1 nM)(Fuchs et al., 1986). Because the product of the LgtC reaction is the P  k  trisaccharide, the B domain of a verotoxin would be an ideal candidate as a  molecular recognition device for isolating phage containing the LgtC product on their surface.  172  The proposed screening strategy for the directed evolution of LgtC was to simultaneously display multiple copies of S A V and a variant of LgtC in low copy number on the surface of M l 3 bacteriophage in water-in-oil emulsions in the presence of a desired reaction condition that includes the presence of a biotin-lactose conjugate acceptor substrate. LgtC activity would convert lactose to the P trisaccharide, which k  would be effectively linked to the phage surface via the high affinity biotin-SAV interaction.  Phage harbouring D N A encoding active LgtC variants could be enriched  using a panning procedure involving the B-domain of a verotoxin and the LgtC encoding regions of isolated phagemids could be P C R amplified and subjected to subsequent rounds of screening. A n overview of this scheme is shown in Figure 4.11.  173  Aqueous compartment  Figure 4.11. Proposed screening procedure for the in vitro evolution of LgtC involving compartmentalized phage display. (1) Individual cells of an f strain of E. coli harbouring a phagemid containing a D N A sequence for polycistronic expression of a unique variant of LgtC fused to pill and SAV-pVIII fusion proteins are subjected to water-in-oil emulsification in the presence of the desired reaction conditions including a biotin-lactose conjugate acceptor substrate and VCSM13 helper phage. (2) Following infection with helper phage, phage harbouring phagemid D N A encoding a unique LgtC variant and having both LgtC displayed in low copy number as a pill fusion and S A V displayed in high copy number as a pVIII fusion on their surface are excreted from the bacterial cell. (3) Depending on the level of transferase activity, the lactose acceptor attached to phage surfaces via the high affinity biotin-SAV interaction is converted to a trisaccharide product. (4) Phage with catalytically active enzyme on their surface are enriched from inactive variants by breaking emulsions, isolating phage and panning using a carbohydrate binding protein specific for the trisaccharide product of the LgtC reaction. Enriched gene variants are then isolated and subjected to subsequent rounds of screening.  174  4.3.5. Synthesis of a Biotinylated Lactose Conjugate Acceptor Substrate A biotin lactose conjugate was readily prepared using an advanced precursor generated and provided by Dr. Spencer Williams that has been successfully applied in ICAT-based proteomic investigations of glycosidases (Hekmat et al., 2005; Williams et al., 2006). It should be noted that nearly identical precursors consisting of biotin and a P E G linker terminating in a free acid are commercially available, making the general applicability of this screening methodology feasible. The acid terminating the linker of the biotinylated precursor was activated as the pentafluorophenyl ester prior to reaction with /?-amino-phenyl-P-D-lactoside (Scheme 4.2). The resulting biotin lactose conjugate was subjected to preparative H P L C purification affording -10 mg of pure acceptor substrate.  Scheme 4.2. Synthesis of a biotin lactose conjugate acceptor substrate for LgtC.  175  4.3.6. Generating a Library of LgtC Variants Error-prone P C R (epPCR) is the method most often used to generate libraries of protein variants containing random mutations for directed evolution applications. P C R protocols are modified to alter and enhance the natural error rate of a D N A polymerase (Leung, 1989). Despite its bias towards A T to G C changes, because of its high natural error rate, Taq polymerase is commonly used. Typical conditions used to increase and alter the errors that occur during the P C R process include the use of increased concentrations of M g C l  2  to stabilize non-complementary base pairing (Cadwell and  Joyce, 1994); addition of M n C l to increase the error rate (Lin-Goerke, 1997); varying the 2  ratio of nucleotides (Vartanian et al., 2001); the addition of nucleotide analogues (Fenton, 2002); using alternative D N A polymerase enzymes such as Strategene's Mutazyme™ polymerase; increasing the number of doubling cycles; or changing the initial concentration of template. The length and base composition of the template D N A dictate the resulting rate of mutational frequency. As such, under identical epPCR conditions, two different genes will most likely exhibit different rates and forms of mutation. A library of LgtC variants was generated using epPCR conditions that included a 5:1 GT to A C nucleotide bias and varying concentrations of M n C l . Presumably due to 2  the relatively large size of the vector, initial attempts using LgtC in pET29 (+) as template prevented the generation of libraries with a diversity of greater than 10 . Using LgtC in 5  Q  pUC18 a library size of 10 was achieved. Sequencing of ten random variants revealed that use of 0.25 m M M n C l led to the generation of a balanced mutational rate of -two 2  mutations per gene and use of 0.5 m M M n C l  2  led to the generation of a balanced  mutational rate of more than four mutations per gene. The library generated using 0.25  176  m M M n C b and pUC18 containing wild-type LgtC was transformed into an expression strain of E. coli.  Crude T L C scale activity tests on induced cell lysates revealed an  activity level for the library that was -20% that of the wild-type enzyme. Based on the mutational rate and activity, this library was deemed suitable for the directed evolution of LgtC activity.  4.3.7.  Immobilization of the Carbohydrate Binding Domain of Shiga-Like Toxin SLT-1B  The three-dimensional X-ray crystal structure of the carbohydrate binding domain of Shiga-like toxin SLT-1B derived from E. coli 0157 with bound Pk trisaccharide was used to determine a suitable strategy for immobilizing a functional form of the protein. The C- and N-termini of the protein are located on opposite sides of the overall structure, with the ligand binding sites located on the N-terminal face (Kitov et al., 2000)(Figure 4.12). It was therefore decided to immobilize the protein on an I M A C surface as a Cterminal 6-histidine version of the macromolecule. The D N A encoding the carbohydrate binding domain of this verotoxin protein was subcloned from pVT27 vector (supplied by Dr. Warren Wakarchuk, IBS N R C , Ottawa) into pET29(+) for expression with 6histidines fused to the C-terminus using Ndel and Xhol restriction sites. Sequencing of plasmid D N A isolated from a resulting clone confirmed the generation of the desired sequence. Although over-expression of this construct in E. coli led to the production of very high amounts of insoluble protein, -100 mg/L of soluble protein could be readily purified using a N i  charged I M A C column. Indicating a rather high affinity for the  I M A C column, 250 m M imidazole was required to elute bound verotoxin.  177  In order to test the ability of this immobilized protein to bind the P trisaccharide, k  fluorescein lactose conjugate acceptor was converted to the P product using LgtC. N i charged I M A C beads were saturated with the purified protein and then incubated with either fluorescein lactose or Pk conjugates.  As a negative control, N i  + 2  charged I M A C  beads without bound protein were incubated with fluorescein P . Following washing, k  only I M A C beads with bound protein and incubated with the fluorescein Pk conjugate remained fluorescent.  This indicates the ability of the immobilized construct to  selectively bind the product of the LgtC reaction and its suitability for application in the enrichment step of the described screening strategy for the directed evolution of LgtC.  Figure 4. 12. Three-dimensional X-ray crystal structure of Shiga-like toxin I B subunit with bound starfish molecule (pdb lqnu). The P trisaccharide portions of the bound k  ligand are shown in top down (A) or side on (B) views. The locations of the N - and Ctermini of one of the subunits of the homopentamer are shown in panel B to illustrate why a C-terminal 6-histidine tag was chosen as the construct for functional immobilization on an I M A C surface.  178  4.3.8. Multivalent Phage Display of SAV as a pVIII fusion protein  In practice, polyvalent display on pVIII is difficult.  The levels of display are  highly dependent on fusion length and sequence (Malik et a l , 1996). Indeed, using the type 8 system (i.e. without the incorporation of wild-type pVIII coat protein) fusion of peptides greater than ten amino acids leads to the production of unstable phage (Iannolo et al., 1995). Using phagemid systems in which wild-type pVIII proteins are supplied in trans facilitates the display of larger peptides.  However, even with the use of 8+8  systems, fusion of proteins greater than -100 residues usually results in the display of no more than a single copy per phage particle (Kretzschmar and Geiser, 1995). To overcome the limitations of 8+8 display, Sidhu et al. used directed evolution to isolate pVIII variants that greatly increased the surface display of monomeric and oligomeric proteins (Sidhu et al., 2000). Human growth hormone (hGH), a monomeric protein of M W = 22 kDa, was fused to the N terminus of pVIII and a library of variants in the pVIII sequence was generated. Variants that facilitated increased surface display of hGH were selected by panning over h G H receptor.  After five rounds of selection,  variants were isolated that increased the level of hGH display by -100-fold. To test the limits of type 8+8 display, the ability to display S A V (MW -50 kDa) as a pVIII fusion was explored. Because S A V exists as a tetramer and is poorly secreted by E. coli (Weber et al., 1989), fusion of S A V to wild-type pVIII resulted in an undetectable level of display of functional SAV. To facilitate display of the functional tetrameric form of S A V , an amber stop codon was inserted between the sequences encoding S A V and pVIII. Phage production in an amber suppressor strain of E. coli leads  179  to translation of the majority of S A V protein as a free monomer, but also allows a small amount of read-through leading to the production of SAV-pVIII fusions. This results in the production of phage particles in which tetrameric S A V is likely displayed in a form that consists of one SAV-pVIII fusion with three associated S A V monomers. A library of pVIII variants fused to S A V was subjected to selection for increased display of S A V using anti-SAV polyclonal antibody as the capture target.  Following five rounds of  selection, variants of pVIII were isolated that provided an ~50-fold increase in the level of functional S A V displayed. Using this in vitro evolution approach, S A V display as a pVIII fusion was 'increased from an impractical level to levels that should allow for functional selection' (Sidhu et al., 2000). It was estimated that each phage particle would display 10 copies of tetrameric S A V thereby providing 40 ligand-binding sites (Sidhu, SS personal communication). The phagemid pSAV, containing the construct encoding S A V fused to one of the selected pVIII variants, was obtained from Dr. Sachdev Sidhu (Genentech, San Francisco). A n amber suppressor strain of E. coli (XL1 Blue (F')) was transformed with p S A V and phage particles displaying S A V as a pVIII fusion were generated by superinfection with VCSM13 helper phage. The resulting phage particles were isolated by double P E G precipitation and subjected to ELISA analysis using BSA-biotin conjugate as the capture target and anti-M13 phage monoclonal ab/horseradish peroxidase (HRP) conjugate for detection. Consistent with the published results obtained using this phagemid system (Sidhu et al., 2000), maximal response occurred at a phage concentration of ~10 n M , indicating a virtually identical level of S A V display to that obtained in the Genentech laboratories (Figure 4.13).  180  2 -| 1.8 1:6 -  i pSAV derived phage  E 1.4 c e 1.2 XG> O 1 c  kpLLOl derived phage  nf  •e  0.8 o _o •st 0.6 0.4 02.  -  0 0.001  0.01  0.1  1  10  P h a g e c o n c e n t r a t i o n (nM)  F i g u r e 4.13. Phage E L I S A for S A V display. Phage derived from uninduced cultures o f X L 1 Blue ( f ) E. coli harbouring either p S A V (squares) or p L L O l (diamonds) were twice P E G precipitated and levels o f S A V display investigated by determining their ability to bind immobilized biotin. Biotinylated B S A was immobilized on a 96-well plate as the capture target and following extensive washing, bound phage was detected using a monoclonal anti-M13 phage ab/HRP conjugate.  4.3.9. Simultaneous M13 Phage Display of LgtC and SAV The sequence encoding the optimized S A V - p V I I I fusion protein was subcloned from p S A V into p L L O l using the Hindlll  restriction site.  The resulting phagemid  (pLL34) was generated for the polycistronic expression o f LgtC-pIII and S A V - p V I I I fusion proteins.  D N A sequencing was used to confirm the presence o f the desired  construct consisting o f the N-terminal-OmpA-LgtC-pIII fusion, a 42 bp spacer, ribosomebinding site, N-terminal-secretion sequence-SAV-amber stop-pVIII fusion within the M C S of pLL34.  181  It was then determined whether pLL34 could be used to generate phage particles that simultaneously displayed -one copy of LgtC as a pill fusion and -ten copies of S A V as a pVIII fusion. XL1 Blue (f) E. coli were transformed with pLL34, pLLOl or p S A V and then superinfected with VCSM13 helper phage.  The resulting phage were P E G  precipitated and subjected to both LgtC activity tests and ELISA analysis to determine the level of S A V display. Phage derived from E. coli harboring pLL34 had an -100-fold decrease in LgtC activity compared to those derived from pLLOl, as determined by T L C analysis. In addition, display of functional S A V was significantly decreased and could barely be detected above background for phage derived from pLL34 (Figure 4.14).  2.5 • pSAV derived phage • pl_L01 derived phage o  ApLL34 derived phage  1.5  0.5 • A " • A « 4, A  0.001  a  0.01  •  A  .  • A - • A .  An  0.1  10  Phage concentration (nM)  Figure 4.14. Phage ELISA for S A V display. Phage derived from uninduced cultures of XL1 Blue (f) E. coli harbouring either pLL34 (triangles), pSAV (diamonds) or pLLOl (squares) were twice P E G precipitated and levels of S A V display investigated by determining their ability to bind immobilized biotin. Biotinylated B S A was immobilized on a 96-well plate as the capture target and, following extensive washing, bound phage was detected using a monoclonal anti-M13 phage ab/HRP conjugate.  182  In an effort to restore LgtC and S A V display on phage derived using pLL34, induction with a range of IPTG concentrations was investigated. Although induction with 0.1 m M IPTG led to a recovery of LgtC activity to within 10% of that of phages derived from pLLOl, increased functional S A V display was not obtained. This is consistent with the reported observation that IPTG induction does not increase the level of S A V display and in fact resulted in a decrease in phage production (Sidhu et al., 2000). This is most likely the result of the fact that the S A V protein itself depletes bacterial cells of the essential vitamin biotin and is toxic. It could very well be that simultaneous phage display of these two different proteins as pill and pVIII fusion is simply too taxing for the production of viable phage. In addition, the choice of S A V as the pVIII fusion partner pushes the limits. Indeed, using wild-type pVIII as the fusion partner, S A V display could not be detected above background.  Even with the optimized version of pVIII, the level of display was  significantly less than the level of display achieved using an optimized version of pVIII fused to hGH.  Whereas a significant ELISA response was observed at a phage  concentration of ~10 p M for hGH, in the case of S A V an equivalent response required a phage concentration of ~10 n M (Sidhu et al., 2000).  4.3.10. Conclusions and Future Directions The most significant finding described in this section is the observation that the process of phage display can be readily performed in the context of water-in-oil emulsions.  This has significant implications for the application of phage display  technology to the directed evolution of enzyme activity. Screening procedures for the  183  directed evolution of enzyme function involving phage display described to date have relied on proximity effects to ensure that the product of enzyme activity is derived from the protein variant attached to the same surface. This approach requires an optimization of the linker distance and, depending on the enzyme class, is limited in the number of substrate molecules accessible to the enzyme. By compartmentalizing a population of identical phage, carrying the same phenotype and displaying the same enzyme variant, multiple turnover of substrates on neighboring phage surfaces can be linked to the encoding genotype of an active variant. In addition, it was shown that LgtC could be readily displayed on the surface of M l 3 phage as a pill fusion protein. Glycosyltransferases are a class of enzyme that is particularly well suited for directed evolution involving phage display screening approaches.  This  applicability arises  from  glycosyltransferases are of membrane-associated  the  fact  that  the  majority  of  species that contain a membrane  association domain that is frequently removed for the purpose of obtaining soluble constructs. As such, the creation of a fusion in which this membrane association domain is substituted with an M l 3 coat protein should result in a catalytically active species. In addition, because glycosyltransferases are frequently transported to different organelles (e.g. Golgi bodies) or membrane surfaces, they should be readily secreted through the inner membrane of E. coli, a key requirement for successful phage display. Future work involving the phage display of LgtC should also address the issue of proteolytic cleavage by incorporating appropriate protease inhitors. Finally, i f LgtC phage display is to be performed in the absence of emulsification, kinetic analysis of phage isolated using a CsCl gradient should be performed to determine the level of activity of displayed enzyme.  184  Unfortunately the final component  of the proposed  screening  procedure,  simultaneous display of active LgtC and S A V on the surface of phage, was not achieved. It is possible that further labor intensive optimization, using directed evolution to optimize pill and pVIII coat proteins or by optimizing transcription/translation levels using a different rbs or different ordering and spacing of gene sequences for polycistronic expression, might overcome the observed dilemma.  However, the finding that S A V  expression is toxic to E. coli makes it an unattractive component for a screening strategy and might lead to artifactual results arising from events leading to the control of S A V expression. As such, an alternative strategy for creating a physical linkage between the glycosyltransferase acceptor and the phage surface is appropriate. One possibility would be the use of a maleimide substrate conjugate. Maleimides react with thiols and, under alkaline conditions, with amino groups, and as such are capable of reacting with multiple sites on the surface of M l 3 phage. Jestin and co-workers demonstrated by SELDI-MS that a maleimide substrate conjugate could be used to generate a covalent adduct with the pVIII coat protein of M l 3 phage, by reaction with either the N-terminal A l a residue or the exposed side chain amino substituent of Lys8, and applied this strategy to develop a screening procedure that relied on proximity effects (Jestin et al., 1999). In order to perform directed evolution on LgtC, a maleimide-lactose conjugate could be readily obtained by reaction of the /?-amino phenyl-(3-D-lactoside with any of a number of commercially available maleimidating agents. Substituting the biotin-lactose conjugate with this reagent during the compartmentalization procedure would negate the need for S A V display as a pVIII fusion, allowing for the direct use the pLLOl phagemid that  185  facilitated robust LgtC display. The work described in this section provides a foundation for the future directed evolution of LgtC activity.  186  CHAPTER 5  GLYCOSYLTRANSFERASE SUBSTRATE ENGINEERING  * Versions of portions of this chapter have been published: Lairson, L.L., Watts, A . G . , Wakarchuk, W.W., and Withers, S.G. (2006) Using Substrate Engineering to Harness Enzymatic Promiscuity and Expand Biological Catalysis. Nature Chemical Biology.  2: 724.  Lairson, L . L . , Wakarchuk, W.W., and Withers, S.G. (2007) Alternative Donor Substrates for Inverting and Retaining Glycosyltransferases. Chemical Communications. (4): 365. - Reproduced by permission of The Royal Society of Chemistry (RSC)  187  5.1.  Summary In an effort to increase the synthetic utility of glycosyltransferases, a process of  "substrate engineering" was investigated. A strategy for expanding acceptor substrate specificity was developed in which a readily removable functional group was attached to alternative glycosyltransferase acceptor substrates to induce productive binding modes, thereby facilitating rational control of substrate specificity and regio-selectivity using wild-type enzymes.  The ability of representatives of both retaining and inverting  glycosyltransferases to use inexpensive alternative donor substrates was also explored. Results with these alternative donors provide more fodder for the mechanistic debate surrounding this class of enzyme and present starting points for engineering efforts to increase their potential in enzymatic glycan synthesis.  5.2. Engineering Alternative Acceptor Substrates for Glycosyltransferases 5.2.1. Background Despite their unparalleled catalytic prowess and environmental compatibility, enzymes have yet to see widespread application in synthetic chemistry. This lack of application and thereby the minimisation of their enormous potential stems not only from a wariness of aqueous biological catalysis on the part of the typical synthetic chemist but also from limitations of applicability that arise from the high degree of substrate specificity possessed by most enzymes.  This latter perceived limitation is being  successfully challenged through rational protein engineering and directed evolution  188  efforts to alter substrate specificity as discussed in section 4.3.1. However, such programs require considerable effort to establish. A n alternative approach involving the chemical modification of alternative substrates using readily removable substituents, in a manner that facilitates productive binding to the enzyme by mimicking a natural binding mode, would serve as a useful method for increasing the range of starting materials that can be transformed using a wildtype enzyme and therefore their inherent synthetic utility.  A n excellent test case for this  approach is that of oligosaccharide synthesis using glycosyltransferases, since the chemical synthesis of such molecules remains arduous. In stark contrast to the situation for peptides and oligonucleotides, a lack of synthetic capacity has delayed our understanding of the important roles that glycan structures play in various biological phenomena. In the case of carbohydrate modifying enzymes, aromatic substituents serve as ideal synthetic mimics of the natural sugar moieties of substrates. Indeed, the use of aryl glycosides in glycosidase kinetic analysis as chromogenic surrogates and as reactive substrates in glycosidase and engineered glycosidase catalysed synthesis is ubiquitous. The ability of aryl substituents to mimic the productive binding modes of sugars is derived by their ability to mimic the hydrophobic platform of natural sugar substrates. Hydrophobic interactions with sugar ring faces are known to provide significant binding affinity and the three-dimensional structures of the sugar binding sites from many carbohydrate-utilizing proteins contain aromatic residues that provide a "hydrophobic platform" onto which a sugar ring face can bind (Nerinckx et al., 2003). In fact, the ability of aromatic substituents to mimic sugar moieties of glycosynthase substrates has  189  been used to alter their regio-selectivity (MacManus and Vulfson, 2000; Stick et al., 2004).  5.2.2. Exploring the Acceptor Substrate Specificity of LgtC Glycosyltransferases are generally believed to be highly selective for both the donor and acceptor substrates, although some exceptions have recently been reported (Bencur et al., 2005; Blanco et al., 2001; Borisova et al., 2004; Durr et al., 2004; Oberthur et al., 2005; Yang et al., 2005).  As described in detail in chapter 3, LgtC catalyses the  transfer of galactose to lactose containing acceptor substrates with net retention of anomeric configuration (scheme 3.1). Previous studies had shown that D-galactose also serves as an acceptor for LgtC, though with greatly reduced efficiency (Ly et al., 2002). To determine the actual degree of acceptor substrate specificity possessed by LgtC, a broad range of sugar and sugar-like substrates were tested as potential substrates for this enzyme. Interestingly, as shown in Table 5.1 and confirmed by ESI-MS and T L C , a range of D-aldose sugars and even /wyo-inositol were found to act as acceptor substrates for LgtC at rates comparable to that for D-galactose.  190  Table 5.1.  Kinetic parameters for various L g t C acceptor substrates.  Acceptor (min.-'.mM" ) 1  Lactose  240 OH  D-Galactose  a  OOH H^OH  0.3  OH^OH OH  A W *»  D- L y x o s e  OH'-QH  3  0.3 OH  D-Arabinose  OH  HO  ^-pL ^ 0  OH  D-Glucose  0.1  a  C  -OH  a H  0.2  ° H - 0 ^ ^ , O H "  0  H  ,-OH  D-Allose  0.05  a  OH  D-Mannose D-Xylose  °  OH  H  ^ O H < O H  a  H  0.3  °H^L2^  1.8  HO  a  -HO^^t, O H "  Myo-inositol  0  H  0.2  a  H O - V ^ A . -OH OH  ,OH  H0  Benzyl-P-D-cellobioside  ^\5^1 -"T^S^l -"• , -O OH  1.7  OH  Benzyl-(3-D-xyloside Octyl-|3-D-glucoside  40 ^ O H  a H O H  1.4  -olg^o.  HO  O H O H  Octyl-p-D-galactoside  H O - 5 ^ ^ 0 .  a  OH  Octyl-p-D-xyloside  V  3.8  '7  22  a W  OH  7  OH  6-OBz-D-Man  4-OBz-D-Xyl  b-r-oH  5.0  °  HO' HO'  a  12  a - no saturation of enzyme activity was observed up to near saturated concentrations. Reliable k  cat  K  m  and K  values were therefore only attainable for Lactose (k , = 160 s" , 1  m  ca  = 40 m M ) , Benzyl-p-D-cellobioside (k  xyloside (k  = 1.0 s" , K 1  cal  cal  m  = 35 m M ) , Benzyl-P-D-  = 10 s" , K = 15 m M ) and 6 0 B z M a n (k =2.5 s" , K 1  1  m  cal  m  = 30 m M ) .  5.2.3. Determining the Ability of LgtC to Transfer to Equatorial Hydroxyl Substituents This wide acceptor range was surprising, especially given that some of these sugars possess only equatorial hydroxyl groups yet LgtC transfers to an axial hydroxyl in its natural substrate. To determine whether LgtC was indeed capable of transferring to an equatorial hydroxyl, the product formed when p-nitrophenyl P-D-glucopyranoside (pNP Glc) acts as acceptor was investigated. Analysis of the single disaccharide product, by ESI-MS and two-dimensional ' H N M R spectroscopy (Figure 5.1), clearly revealed the formation of a Gal-a-(l,4)-Glc linkage, indicating that pNP Glc adopts the same basic binding mode as lactose, but that transfer occurs to the equatorial 4-hydroxyl. Assignments were made by examining the couplings within the ring system and the locations of glycosidic linkages were determined on the basis of chemical shift. Because the acetate group is more electron withdrawing than a glycoside substituent, the proton attached to the carbon involved in the glycosidic bond resonates further upfield (<4.50 ppm) in a region clearly distinct from anomeric protons or protons attached to carbons bearing an acetate group.  192  OAc  ^4 LA.  p 3.8  MO D  4  »i0  4 .8  a  5.2  •j—-4- S . 4  SO  1 .8  Figure 5.1.  '  1 5.6  1 5.4  1 5.2  1 5.0  1 4.8  '  1 4.6  1 4.4  1 4.2  '  1 4.0  '  1 3.8  1 3.6  1-6.0 ppm  ' H - C O S Y N M R spectrum of para-nitrophenyl (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(1^4)-2,3,6-tri-0-acetyl-(3-D-glucoside.  5.2.4. Exploring the Synthetic Utility of LgtC Catalysed Galactosyl Transfer to Alternative Monosaccharide Acceptor Substrates This finding of relatively broad specificity led to an initial hope that wild-type LgtC could be used as a catalyst in a range of useful galactosylation reactions. Disaccharide products derived from two of the alternative acceptor monosaccharides studied, D-mannose and D-xylose, were therefore acetylated, purified and subjected to  193  ESI-MS and N M R structural analysis. In both cases, although M S analysis revealed masses consistent with that of disaccharide product, ' H N M R analysis revealed the purified products to be mixtures of regio-isomers. This finding was quite disappointing since a catalyst that yields a mixture of regio-isomers is clearly of limited utility.  5.2.5. A Substrate Engineering Strategy that Uses Aromatic Substituents as Active Site Anchors The fact that a single regio-isomer was obtained when pNP Glc was used as acceptor suggested a strategy in which the appendage of a conveniently installed and removed substituent onto the acceptor might control its active site binding orientation in a useful manner. In fact, as mentioned, the sugar binding sites from many carbohydrateutilizing proteins contain aromatic residues that provides a sugar binding "hydrophobic platform" (Nerinckx et al., 2003). Inspection of the three-dimensional structure of LgtC (pdb lga8) reveals that the aromatic side chain of Phel32 is positioned to function in such a role, providing a hydrophobic platform onto which the glucose ring of the lactose acceptor or aromatic moieties may bind (Figure 5.2). This process of substrate engineering was further explored using LgtC.  A  convenient appendage at the anomeric centre is a benzyl group, which can be installed by classical glycoside formation chemistry and which can be removed by hydrogenation. Correspondingly, benzyl (3-D-cellobioside and benzyl (3-D-xyloside were synthesized and tested as acceptors.  Both T L C and ESI-MS analysis clearly indicated that they both  functioned well in this role with the formation of a single product as had been seen with pNP Glc. Measurement of reaction rates revealed that, not only did the inclusion of an  194  aromatic substituent control regiochemistry, but also it resulted in substantially improved reaction rates, with the k JK c  m  value of benzyl p-D-xyloside being over 20-fold higher  than that for its parent sugar and within a factor of four of that for the natural acceptor lactose (Table 5.1).  Figure 5.2. Acceptor binding site of LgtC containing a hydrophobic platform (Phel32) for binding the glucose sugar ring of lactose or an aromatic substituent (pdb lga8).  Analysis of the product formed from benzyl P-D-xyloside by two-dimensional H [  N M R spectroscopy revealed the exclusive formation of an cc-l,3-linkage (Figure 5.3), indicating that xylose adopts an alternative binding mode within the LgtC active site. The formation of 1,3-linkages to xylosides has been observed in transglycosylation reactions  195  with several glycosidases and glycosynthases with preferences,  otherwise, for  1,4-  linkages, providing precedence for this alternative binding mode (Mackenzie et al., 1998).  H2  H4  H1  Figure 5.3. H - C O S Y N M R spectrum o f benzyl (2,3,4,6-tetra-O-acetyl-a-D1  galactopyranosyl)-(l->3)-2,4-di-0-acetyl-P-D-xyloside.  5.2.6. Using Substrate Engineering to Control RegioSelective cc3GalT Catalysed Galactosyl Transfer to a Monosaccharide Acceptor Substrate A s described in detail in section 3.4., a 3 G a l T is known to transfer galactose from U D P galactose to terminal A^-acetyllactosamine-containing glycoconjugates producing  196  Gal-a-(l,3)-Gal-p-(l,4)-GlcNAc epitopes (Scheme 3.5). The ability of cc3GalT to use Dgalactose as an acceptor was investigated. A s was observed with LgtC, transfer to an underivatized monosaccharide led to the formation of a mixture of regio-isomers of disaccharide products as determined by H N M R analysis, whereas the presence of an l  aromatic substituent led to the exclusive formation of a pure a-(l,3)-linked regio-isomer when pNP Gal was used as acceptor (Figure 5.4.).  -OAc  *-"  H2  V_|T \ OAc  H1  i  H3  ppm  j  3.9  Til—i  i  4.0  —  4.1  } I  ;  4.2  I  4.3  i  4.4  j  4.5  ;  4.6 4 .7  |  4.8 4.9  :  5.0  i I  J  (  •  .  _  5.1 5.2 5.3  !  5.4  i  5.5  ^  SL  .  5.6 5.7  i—•—i— •  5.8  5.6  i  5.4  1  i  5.2  • i • i •— 5.0  4.8  4.6  4.4  4.2  4.0  ppm  Figure 5.4. ' H - C O S Y N M R spectrum of /?ara-nitrophenyl (2,3,4,6-tetra-O-acetyl-a-Dgalactopyranosyl)-(1^3)-2,4,6-tri-0-acetyl-P-D-galactoside.  197  5.2.7. Using Substrate Engineering to Facilitate Cst II Catalysed Sialyl Transfer to a Monosaccharide Acceptor Substrate Cst II catalyses the transfer of sialic acid (NeuAc) from cytidine 5'-monophosphoN-acetylneuraminic acid (CMP NeuAc) to terminal lactose or 3'-sialyl lactose-containing glycoconjugates yielding a-(2,3) and o>(2,8) linked products respectively (scheme 2.1). Due to the high rate of enzyme catalysed hydrolysis of the donor substrate (Chiu et al., 2004), when underivatized D-galactose was used as acceptor, no product formation was observed. However, when the D-galactose was derivatized with an aromatic substituent (pNP Gal), quantitative product formation was observed in a completely regio-selective manner.  5.2.8. The Nature of the Substituent on an Alternative Acceptor Substrate can affect the Regio-Selectivity of LgtC Unexpectedly, substituents  when  reactions  of monosaccharides  containing  aromatic  were performed using LgtC from crude cell lysates containing the  commercial protein extraction reagent BugBuster™ (Novagen), a disaccharide product was isolated that did not contain aromatic *H signals but rather possessed an eight carbon saturated alkyl chain. Inspection of the H and  C N M R spectra revealed the product to  be derived from galactosylation of the non ionic detergent octyl-P-D-glucoside that is present in the BugBuster™ preparation.  However, in contrast to the Gal-a-(l,4)-Glc  linkage observed when pNP Glc was used as acceptor, the single product obtained from the alkyl-substituted substrate possessed a Gal-a-(l,3)-Glc linkage, as determined by the  198  ' H - C O S Y N M R spectra (Figure 5.5). This result was confirmed using authentic octyl-pD-glucoside and highly purified LgtC and all subsequent reactions were performed with pure enzyme to avoid this complication.  H4 H2  3.4  -3 .6  -.1.8  -4.0  -4.2  4.4  •1. {  4.8  5.0  5.3 5 .i  S. 4  Figure  5.5.  'H-COSY  S.2  S.O  4.8  4.6  4.4  N M R spectrum  4.2  4.0  of H-octyl  3.8  3.6  3.4  ppm  (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l-^3)-2,4,6-tri-0-acetyl-P-D-glucoside.  Thus, while in both cases chemical modification of the alternative substrate with either an aromatic or alkyl substituent provides control of regio-selectivity and leads to the formation of a single product, the nature of the substituent facilitated alternative productive binding modes, resulting in the production of different regio-isomers.  Alkyl  199  galactoside and xyloside derivatives were also tested as alternative acceptors. In the case of octyl-P-D-galactoside, presumably, interactions with the natural galactose moiety dominate, resulting in the exclusive formation o f a Gal-a-(l,4)-Gal linkage (Figure 5.6).  - |  5.5  Figure  5.6.  ,  'H-COSY  ,  ,  ,  1  5.0  ,  ,  ,  ,  1  4.5  N M R spectrum  ,  ,  of  ,  ,  1  t.O  ,  rc-octyl  ,  r—,  p-  3.5  (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l->4)-2,3,6-tri-0-acetyl-P-D-galactoside.  200  As the aromatic derivative of xylose was already found to adopt an alternative binding mode, it was not unexpected that octyl-(3-D-xyloside also yielded a Gal-a-(l,3)X y l linkage (Figure 5.7). The ability of a glycosyltransferase to use these alkylated lipidlike derivatives could prove useful in combinatorial oligosaccharide synthesis and in the production of novel glycolipids (Miura et al., 1996).  ppm  Figure  5.7.  1  H-COSY  5.0  4.5  N M R spectrum  4.0  of /?-octyl  3.5  (2,3,4,6-tetra-O-acetyl-a-D-  galactopyranosyl)-(l-»3)-2,4-di-0-acetyl-P-D-xyloside.  201  5.2.9. Controlling the Regio-Selectivity of LgtC-Catalysed Galactosyl Transfer to Mannose This flexibility of sugar binding mode opened up the intriguing possibility that the regiochemical outcome could be further controlled by appending an aryl group at other positions on the sugar ring as has been reported for glycosidases and glycosynthases (MacManus and Vulfson, 2000; Stick et al., 2004). For aldohexoses, the most convenient substituent is a benzoate ester, which can be installed selectively at the 6-position at low temperatures, with limiting quantities of acylating agent, and removed under a variety of mild conditions. As shown in Figure 5.8, a binding mode for a 6-O-benzoyl sugar could be envisioned in which the sugar 2-hydroxyl group adopts the same position as the 4hydroxyl of the galactose and the aromatic substituent takes the place of the glucose ring of lactose.  Figure  5.8. A model for the substrate engineering strategy using 6-OBz Man. Two  representations of equivalent (=) 6-OBz Man illustrating how 6-OBz Man can approximate («) a lactose binding mode.  202  Indeed, in the case of 6-O-benzoyl-D-mannose (6-OBz Man) the hydroxyl group at this position also has the axial configuration. In fact, the molecular mechanics (MM2) energy-minimized structure of 6-OBz Man overlays well with the X-ray crystal structure of 4'-deoxy lactose bound to LgtC, with the 2-OH of mannose oriented in the position that would be occupied by the 4'-OH that undergoes glycosylation in the LgtC catalysed reaction (Figure 5.9).  Figure  5.9. Overlay of the M M 2 minimized structure of 6-OBz Man (green) with the  structure of 4-deoxylactose (orange) bound within the active site of LgtC (pdb 1GA8).  6-OBz Man was readily prepared in one step from D-mannose by low temperature benzoylation and tested as an alternative acceptor substrate for LgtC, kinetic analysis revealing that the addition of the benzoyl group to mannose improved k /K Qat  approximately 20-fold (Table 5.1).  m  by  To determine whether the appended aromatic  functional group also provides control of regio-selectivity, the single product formed was acetylated, purified and subjected to product analysis as described. ESI-MS revealed the expected mass for a disaccharide product, with *H N M R analysis confirming a single, pure regio-isomer.  1  H-COSY N M R analysis was used to assign all proton signals,  203  identifying the newly formed glycosidic bond as the predicted Gal-oc-(l,2)-Man linkage (Figure 5.10).  Figure 5.10. H-COSY N M R spectrum of (2,3,4,6-tetra-O-acetyl-a-D-galactopyranosyl)1  (1 ->2)-1,3,4-tri-0-acetyl-6-0-benzoyl-a-D-mannose.  204  5.2.10. Controlling the Regio-Selectivity of LgtC-Catalysed Galactosyl Transfer to Xylose To further test the predictability of the model, the relatively symmetric xylose moiety was subjected to further derivatization. With a knowledge of the alternative binding mode of xylosides in hand, as shown in Figure 5.11, aromatic derivatization at C4 should allow a binding mode that leads to the formation of an a-(l,2)-linkage.  HO  F i g u r e 5.11.  A model for the substrate engineering strategy using 4-O-benzoyl-D-xylose  (4-OBz Xyl). Two representations of equivalent (=) 4-OBz Xyl illustrating how 4-OBz X y l can approximate («) the observed binding mode of benzyl P-D-xyloside.  4-OBz X y l was tested as an alternative acceptor substrate for LgtC-catalysed galactosylation. Appendage of the aromatic group again resulted in an increase in the observed rate constant (~7-fold) (Table 5.1), while product analysis (ESI-MS, and ' H , C 1 3  and H-COSY N M R ) revealed exclusive production of the anticipated a-l,2-linked 1  disaccharide (Figure 5.12).  205  H1oc  H1pH3a H3|i  H2o. HZp  OM|  0-  I ill  L  i k » i . .  *-OBr  ppm  3.5  h4.0  4.5  5.0 6  r •  »  8  8  8  5.5  6.0  6.5  6.5  Figure 5.12.  6.0  5.5  5.0  4.5  4.0  3.5  ppm  'H-COSY N M R spectrum of (2,3,4,6-tetra-O-acetyl-a-D-galactopyranosyl)-  (1 -»2)-1,3 ,-di-0-acetyl-4-0-benzoyl-a/(3-D-xylose.  5.2.11. Concluding Remarks By the simple expedient of transient attachment of an alkyl or aryl substituent to the acceptor sugar, the substrate specificity of wild-type LgtC has been broadened to allow exclusive formation of a-1,2, a-1,3 or a-1,4 linkages at synthetically useful rates (Figure 5.13). Besides demonstrating an alternative strategy for expanding the synthetic utility of enzymatic catalysis by transient chemical modification (substrate engineering),  206  the findings presented here illustrate four points of significance. Firstly, they highlight the need for caution in assigning the natural role of glycosyl transferases when using unnaturally modified acceptors.  Indeed, potential acceptor substrates bearing alkyl or  aromatic substituents are frequently used to deduce the natural activities of newly isolated glycosyltransferases. Secondly, they demonstrate that the acceptor substrate specificity of glycosyltransferases is not quite as rigid as previously believed, with significant implications for their synthetic utility.  Thirdly, this promiscuity can be harnessed by  modification of the acceptor sugar with a readily removable aromatic substituent or a simple alkyl chain that facilitates a productive and predictable binding mode providing both control of regio-selectivity and an increase in observed rates. Finally, the described strategy provides facile access to glycosidic linkages for which no 'wild-type' enzymatic activities have yet been identified.  This glycosyltransferase substrate engineering  strategy should prove useful in the development of libraries of oligosaccharide structures or in providing facile synthetic access to a given synthetic target by allowing a single protein catalyst to be used in a broad range of reactions.  207  Aryl glucosides  OH  Alkyl galactosides  4-OBz-xylose  Figure 5.13. Substrate engineering strategy applied to various alternative LgtC acceptor substrates. Derivatization of alternative acceptors is used to generate substrates that mimic the lactose binding mode for LgtC-catalysed galactosyl transfer.  208  5.3. Engineering Alternative Donor Substrates for Glycosyltransferases 5.3.1. Background The application of glycosyltransferases in oligosaccharide synthesis has suffered from both a lack of access to the enzymes themselves, as most exist as membraneassociated species, and also from the cost of high energy glycosyl donor substrates (typically nucleotide sugars). With bacterial enzymes, the former problem is currently being overcome by successful cloning strategies involving truncations to yield recombinant soluble catalytic domains. Attempts to overcome the latter problem involve elaborate coupled enzyme recycling schemes involving in situ generation of the donor substrate (Gilbert et al., 1998; Ichikawa et al., 1991). A recently described alternative strategy takes advantage of the reversibility of natural product glycosyltransferases for in situ generation of desired nucleotide sugars (Zhang et al., 2006a).  The increasing  accessibility of the pyrophosphorylases and nucleotidyltransferases responsible for the formation of nucleotide sugars should further facilitate these approaches, however, the technical complexity and problems of product inhibition provide an impetus for exploring potential alternative donor substrates for glycosyltransferases.  5.3.2. An Activated Alternative Donor Substrate for Cst II that Mimics the Product Binding Mode As described in detail in chapter 2, Cst II is an inverting bifunctional cc-2,3/2,8 sialyltransferase that uses P-linked C M P sialic acid as a donor substrate and transfers the sialic acid moiety with net inversion of anomeric configuration to the 3' hydroxyl of  209  terminal lactose-containing acceptors (Scheme 2.1). Subsequently, Cst II will use this 3'sialyl lactose product as an acceptor and transfer sialic acid to the 8" hydroxyl. The three dimensional structure of this enzyme and the description of a catalytic mechanism that appears to involve a straightforward direct displacement reaction whereby active site residues facilitate departure of the C M P leaving group and activate the incoming nucleophile of the acceptor was presented in detail in chapter 2. To determine whether this enzyme could use an alternative source of activated sialic acid as a donor substrate, we chose to investigate the ct-linked /rara-nitrophenyl derivative.  Providing potential precedence for their use as alternative substrates for  glycosyltransferases, nitrophenyl derivatives of ribose have been successfully employed as alternative substrates and thence used as mechanistic probes for both purine nucleoside phosphorylase and N-ribohydrolase (Mazzella et al., 1996). Because Cst II is an inverting enzyme, it was presumed that the aromatic ring of the nitrophenyl substituent could be accomodated within the active site by taking the place of the galactose ring of the acceptor which must be located on the a face of the sialic acid donor (Figure 5.14). With the pNP a-D-sialoside bound in such a manner, C M P could also be accommodated within the active site on the P face of sialic acid, allowing for the direct displacement of pnitrophenolate by C M P and the in situ formation of P-linked C M P sialic acid. This could then be used as the donor substrate in a subsequent transfer reaction via the normal mechanism following release of pNP.  210  A  B  1 pNP  OH  OCMP  HO  HO  COO" HO  L ^ N 0  2  Figure 5.14. Representation of the product binding mode of Cst II (A) and the proposed mode of binding of a-linked pNP NeuAc that would facilitate in situ generation of the natural donor substrate.  Enzyme-catalysed sialic acid transfer was monitored by T L C using a fluoresceinlactose conjugate acceptor and as is shown in Figure 5.15A, Cst II was indeed found to catalyse transfer to the fluorescent acceptor but only in the presence of C M P . Subsequent preparative scale reactions were performed to determine the efficiency of the reaction (Figure 5.15B). Product formation was confirmed by H R ESI-MS and by treatment of the purified reaction product with an ct-2,3 specific neuraminidase (see methods section). Although reported yield is appreciable, the synthetic utility of this particular process is limited by the high cost of pNP NeuAc, being approxiamately equal to the that of C M P NeuAc. However, application of this approach to other inverting glycosyltransferases, for which the cost of aryl derivatives of donor substrates are substantially less expensive than  211  their nucleoside diphosphate counterparts, should provide an economical alternative approach to syntheses involving this class of catalyst.  A  B  Figure 5.15.  (A) Monitoring Cst II catalysed glycosylation of fluorescent fluorescein  lactose conjugate acceptor substrate using /^-nitrophenyl a-sialoside as an alternative donor substrate in the presence or absence of CMP by T L C (4:2:1:0.1 EtOAc: MeOH: H2O: AcOH).  (B) Reaction catalysed by Cst II using the alternative donor substrate p-  nitrophenyl a-sialoside.  Direct formation of the proposed CMP NeuAc intermediate could not be detected either in the presence or absence of acceptor substrate by either negative ion mode ESIMS or T L C analysis. This is not surprising giving the labile nature of this material and the very low concentrations that would exist in solution. In addition, due to the apparent low affinity of Cst II for the donor substrate analogue and absorbance interference with the pNP subsituent in the U V range used to measure rates with the N A D H dependent enzyme-coupled assay, rates of product or CMP NeuAc formation could not be obtained.  212  5.3.3. An Activated Alternative Donor Substrate for LgtC 5.3.3.1. Testing a-Linked Aryl Galactosides To  further explore the  generality  of this substrate promiscuity amongst  glycosyltransferases, a similar strategy was applied to the retaining a-1,4 galactosyl transferase LgtC. As described in section 3.2, LgtC catalyses the transfer of galactose to the 4' hydroxyl of terminal lactose-containing acceptor substrates using a-linked UDP galactose as the donor with overall net retention of anomeric configuration (Scheme 3.1). It has previously been observed that LgtC will use a-galactosyl fluoride as a donor substrate analogue in the presence of UDP (Lougheed et al., 1999). In this study, the in situ formation of UDP Gal arising from substitution of the reactive fluoride group was demonstrated using an enzyme-coupled assay involving UDP Glc dehydrogenase (Lougheed et al., 1999).  This finding is consistent with either of the proposed  mechanisms of retaining glycosyltransferases.  The observed activity could occur via a  double displacement mechanism involving initial displacement of fluoride ion by a catalytic nucleophile, resulting in the transient formation of a covalent glycosyl-enzyme intermediate, followed by subsequent displacement by an incoming nucleophile (either the acceptor or UDP). Alternatively, an oxocarbenium ion is formed, with a lifetime that allows for a degree of departure by the fluoride ion leaving group sufficient for attack by the incoming nucleophile on the same face. To further test the limits of donor substrate promiscuity, LgtC was tested for its ability to use a-linked pNP or dNP galactosides as the surrogate donor substrate. Again, glycosyl transfer was monitored by T L C using a fluorescein-lactose conjugate acceptor. Under these conditions, no trisaccharide product  213  was observed. This is not surprising as the dNP leaving group would have to occupy the UDP binding site of the enzyme and the binding of UDP to LgtC is known to induce a conformational change that results in the closure of a loop that makes up a significant portion of the active site (Persson et al., 2001). Therefore, an a-linked alternative leaving group must be of sufficiently small size to allow for simultaneous binding of UDP, as is the case with a-galactosyl fluoride.  5.3.3.2. Testing (3-Linked Aryl Galactosides A strategy analogous to that described above for Cst II was then applied to LgtC in which an alternative donor with an activated leaving group of anomeric configuration opposite to that of the natural donor was tested. In this case, because LgtC is a retaining enzyme, the aromatic leaving group would not occupy the acceptor site as both the leaving group and the incoming nucleophile are present on the same face of the donor sugar in the active site of this class of enzyme. As such, success of this approach would rely upon fortuitous accomodation of an aromatic substituent on the p face of the donor within the LgtC active site. Initially, pNP P-D-galactoside was tested for its ability to act as a surrogate donor in the presence or absence of UDP, but no trisaccharide product was observed in either case, even at high enzyme concentrations (0.5 mg/mL) and after extended incubation times (5 days).  Dependence on leaving group ability was then  probed by repeating the above experiments using the more activated derivative 2,4dinitrophenyl P-galactoside.  Using this more activated analogue a trisaccharide was  observed and again glycosyl transfer only occurred in the presence of the natural nucleotide (UDP) (Figure 5.16A). Product formation was insufficient to obtain a yield  214  and was confirmed solely by HR ESI-MS. Analogously to the inverting enzyme, LgtC presumably catalyses the direct displacement of the activated P-linked donor analogue by UDP resulting in the formation of the natural a-linked UDP galactose donor substrate which is then used in a subsequent glycosylation reaction involving the acceptor substrate to yield a trisaccharide product. However, as was described in section 5.3.2., rates of formation of the product or the putative U D P Gal intermediate could not be determined due to the apparent low affinity binding of the donor substrate analogue and interference of the dNP subsituent in the U V range used to measure rates with the N A D H dependent glycosyltransferase and U D P Glc dehydrogenase enzyme-coupled assays. The observed rate of product formation was low and this reaction is of limited synthetic utility due to the high rate of spontaneous hydrolysis of the dNP P-D-galactoside donor substrate analogue. However, the results provide a potential starting point for directed evolution efforts to increase this alternative activity and provides potential insight into the structural plasticity of the active site, with associated mechanistic significance, for this enzyme.  215  A  B  F i g u r e 5.16.  (A) Monitoring LgtC catalysed glycosylation of fluorescent fluorescein  lactose conjugate acceptor substrate using 2,4-dinitrophenyl p-galactoside as an alternative donor substrate in the presence or absence of UDP by TLC (7:2:1:0.1 EtOAc: MeOH: H2O: AcOH).  (B) Reaction catalysed by LgtC using the alternative donor  substrate 2,4-dinitrophenyl P-galactoside.  5.3.3.3. Mechanistic Implications for LgtC As described in section 3.2, a catalytically relevant covalent glycosyl-enzyme intermediate was observed for an active site mutant of LgtC, however the aspartate residue that became labelled is separated from the anomeric reaction centre by a distance of 9 A in the available crystal structure. Even if this residue does not play the role of the catalytic nucleophile in a double displacement mechanism, in order for this covalent species to be formed and turned over a significant change in conformation from that observed in the crystal structure must occur during catalysis, thereby indicating a significant degree of structural plasticity. Support for this idea of significant structural plasticity within the LgtC active site is further provided by the above described  216  accommodation of dNP-P-D-galactoside in a catalytically competent conformation, as it is difficult to imagine how this P-linked aryl substituent could be accomodated within the donor substrate binding site based on the available ternary complex crystal structure without active site reconfiguration. Although neither of these results provides definitive support for a given catalytic mechanism used by retaining glycosytransferases, they do indicate that crystal structures for glycosyltransferase complexes should be interpreted cautiously.  5.3.4. Concluding Remarks Cst II and LgtC were found to use ;?-nitrophenyl a-sialoside and 2,4-dinitrophenyl P-galactoside, respectively, as alternative donor substrates in the presence of their respective natural nucleotide products. The results of this work illustrate how the donor substrate promiscuity of both inverting and retaining glycosyltransferases can be harnessed, providing the starting point of an alternative strategy for their application as synthetic tools. Both classes of enzymes appear to be able to use glycosyl donors with alternative activated leaving groups of opposite anomeric configuration, compared to the natural donor substrate, in the presence of catalytic amounts of the natural nucleotide. Presumably, the enzyme functions by catalysing the in situ formation of the natural nucleotide sugar donor, which is then used in a subsequent catalysed glycosylation reaction involving an acceptor substrate. From the present understanding of structure and mechanism of the inverting Cst II, it is reasonably clear how the enzyme active site accommodates the observed alternate activity. The a-linked pNP substituent could bind within the acceptor substrate binding  217  site ready for displacement by C M P bound in its natural site on the P face of sialic acid. It is therefore expected that this approach should be generally applicable amongst other inverting glycosyltransferases. It is more difficult to rationalize the observed activity with the retaining enzyme LgtC. Based on the available ternary complex crystal structure it is not clear how the P linked dNP substituent can be accommodated within the active site as both the acceptor and leaving group binding sites lie on the a face of the galactosyl donor. These results indicate that alternative catalytically relevant conformation from that of the observed LgtC ternary complex crystal structure are possible, consistent with the described labelling of a remote putative catalytic nucleophile. It would be quite interesting to examine this flexibility, by obtaining a three-dimensional X-ray crystal structure of LgtC with a non-reactive aryl analogue (e.g. phenyl P-D-galactoside) and U D P Gal bound, to determined how this alternative substrate is accomodated within the active site in a catalytically competent manner.  218  CHAPTER 6  Summary and Future Directions  Despite their biological significance and potential applicability in oligosaccharide synthesis, glycosyltransferases remain a relatively poorly characterized and utilized class of enzyme.  The work presented in this thesis sought to address deficiencies  in  mechanistic  understanding  of  glycosyltransferase catalysis.  and  to  increase  the  practical  applicability  Mechanistic studies with glycosyltransferases provided  insight into the role of convergence in the evolution of enzyme mechanism not only between unrelated classes of glycosyltransferases, but also between different classes of carbohydrate modifying enzymes (i.e.  glycosyltransferases  understanding of the mechanisms of glycosyltransferases engineering strategies.  and glycosidases).  An  facilitated the success of  Conversely, the results of engineering studies provided insight  into mechanism. Inverting glycosyltransferases catalyze glycosyl group transfer with net inversion of stereochemistry at the anomeric reaction centre. The general mechanism of this class of enzyme is that of an SN2-like pathway involving direct nucleophilic displacement. The chemical and kinetic mechanism of a representative inverting sialyltransferase known as Cst II was investigated by detailed kinetic analysis, protein X-ray crystallography and site directed mutagenesis. Product and dead-end donor substrate analogue inhibition kinetics, as well as X-ray crystallographic analysis, support a sequential iso-ordered Bi Bi kinetic mechanism for this bifunctional enzyme.  Kinetic analysis of mutants of candidate  catalytic residue, identified by X-ray crystallographic analysis of Cst II with a bound nonreactive donor substrate analogue, revealed that the tyrosyl hydroxyls of Tyrl56 and Tyrl62 play critical roles in facilitating departure of the nucleoside monophosphate leaving group. Chemical rescue and kinetic analysis of the HI 88A mutant, as well as pH  220  dependence studies, implicates His 188 as the catalytic base that activates the incoming nucleophile of the acceptor substrate.  Cst II was found to adopt a G T - A fold.  Interestingly, the identities of the residues utilized to catayze group transfer are akin to what has been observed with the G T - B fold enzymes, indicating a convergence in catalytic mechanism between these two inverting clans. Future work with Cst II should include the obtainment of ternary complex X-ray crystal structures with either lactose or 3'-sialyl lactose bound. This will elucidate the nature of the two stable enzymes forms suggested by the proposed iso ordered Bi Bi kinetic mechanism. Conformational analysis of these two structures will provide an understanding of how this enzyme evolved a bifunctional activity. By simple analogy to the well-characterized retaining glycosidases, retaining glycosyltransferases had been thought to use a double displacement mechanism involving enzymatic nucleophilic catalysis.  Such a mechanism necessitates the appropriate  positioning of a catalytic nucleophile within the enzyme active site.  However, a  comparison of the X-ray crystal structures of representative retaining glycosyltransferases from multiple families, which are believed to have all diverged from a common evolutionary origin, indicates a complete lack of conserved structural architecture in the region that would be occupied by this critical catalytic residue. This is in stark contrast to what has been observed for the vast majority of retaining glycosidases, which have not evolved from a common evolutionary origin yet have all converged to have an identically positioned amino acid side chain carboxylate poised to play the role of a catalytic nucleophile. This revelation in itself may well support an argument against a common catalytic mechanism between these two classes of enzymes.  221  The possibility of a pathway involving nucleophilic catalysis for retaining glycosyltransferases was explored. A representative galactosyltransferase known as LgtC was subjected to a rational protein engineering study with the goal of driving the reaction along a coordinate involving the formation of an intermediate species covalently bound to the enzyme.  This work led to the first direct observation of a catalytically relevant  covalently bound glycosyl-enzyme intermediate for a retaining glycosyltransferase. However, surprisingly, the site of labelling was found to be a residue that is relatively remote from the anomeric reaction centre in the ground state crystal structure, suggesting the possible identity of a previously unidentified candidate catalytic nucleophile. If this residue were to act as a catalytic nucleophile in a double displacement mechanism, a significant conformational change would have to occur during the course of catalysis from that of the ground state ternary complex crystal structure. This would indicate that the active site conformations of retaining glycosyltransferases are highly plastic. Unfortunately, the exact mechanism of LgtC remains ambiguous. The observed labelling result could either be interpreted as evidence supporting a double displacement mechanism involving a previously unidentified nucleophile or as an artefact resulting from an altered active site electrostatic environment. A  lack of conserved architecture amongst related enzymes, precedence for  cationic enzymatic mechanisms, and the  inherent  differences  in reactivities of  glycosyltransferase and glycosidase substrates all support a notion that retaining glycosyltransferases and retaining glycosidases have not converged on an identical catalytic mechanism. A likely catalytic pathway for many retaining glycosyltransferases, including LgtC, is that of a D N * A N  S  SJP  mechanism involving the formation of a short-lived  222  ion pair intermediate species. If LgtC were to use such a pathway, it is possible that the highly reactive electrophilic intermediate could react with a nucleophilic side chain within the altered electrostatic environment of the engineered active site thereby accounting for the observed labelling results. Most likely, a mechanistic continuum exists between members of this class of enzyme in which some use discretely S ^ - l i k e pathways while others use ion pair mechanisms. This would be analogous to what is known for non-enzymatic nucleophilic substitution reactions in which S N I - l i k e or SN2-like pathways are simply the extremes between which a range of ion pair mechanisms exists. Future work with this class of enzyme should involve obtaining an X-ray crystal structure of the covalently bound intermediate of the LgtC Q189E mutant in order to understand how alternative active site conformations facilitate catalysis.  In addition, the protein  engineering strategy employed with LgtC should be applied to other retaining enzymes in order to determine i f other alternative candidate nucleophiles can be identified. It was demonstrated that catalytically active LgtC could be displayed on the surface of M l 3 bacteriophage as a pill fusion protein. Further, LgtC phage display was successfully performed in the context of a water-in-oil emulsification procedure in which multiple identical phage are produced in isolated ~5 fL nanodrops. This procedure has widespread potential in the development of phage display-based directed enzyme evolution screening methodologies that are not reliant on proximity effects. Future work with this emulsified phage display procedure should involve the development of a strategy for covalently modifying the surface of phage with multiple copies of an acceptor substrate. Upon successfully achieving this goal, the procedure should be utilized in directed evolution approaches to alter the activity of glycosyltransferases. This screening  223  procedure should be generally applicable to the directed evolution of various classed of enzyme catalysts. A substrate engineering strategy was developed whereby a readily removable substituent is appended to various locations of an alternative glycosyltransferase acceptor substrate in a manner that facilitates a productive binding mode. The aromatic substituent mimics the hydrophobic surface of natural sugar substituents. Using this technique, the substrate specificity of wild type LgtC has been broadened to allow exclusive formation of a-1,2, a-1,3 or a-1,4 linkages at synthetically useful rates with various alternative acceptors.  Future work with this substrate engineering approach should involve its  application to other carbohydrate modifying enzymes and its use in the synthesis of libraries of oligosaccharide structures. Finally,  the  ability  of representatives  of both retaining and inverting  glycosyltransferases to use inexpensive alternative donor substrates was explored. It was demonstrated that Cst II and LgtC will use />-nitrophenyl a-sialoside and 2,4dinitrophenyl [3-galactoside, respectively, as alternative donor substrates in the presence of their respective natural nucleotide products. In the case of the inverting enzyme, the engineering strategy was based on an understanding of kinetic mechanism.  The  alternative donor substrate binds to the enzyme in a mode that mimics the product complex in which the aromatic substituent takes the place of the sugar ring of the trisaccharide product. In the case of the retaining enzyme, the observations are more difficult to rationalize. Binding of the alternative P-linked donor substrate would rely on a fortuitous catalytically active binding mode. Such a binding mode would result in steric clashes in the observed ground state X-ray crystal structure. This result is consistent with  224  the described labelling results, providing further support for the notion that the active sites o f retaining glycosyltransferases are conformationally dynamic species.  The ability o f  glycosyltransferases to use these inexpensive alternative donor substrates provides a starting point for future engineering strategies to increase the utility and economic viability o f glycosyltransferase catalyzed synthetic routes. It is believed that the approach should be generally applicable to other inverting glycosyltransferases. Directed evolution should be applied to the inverting enzymes to increase the rates at which these alternative substrates are utilized.  225  CHAPTER 7  EXPERIMENTAL METHODS  7.1. General A l l buffers and reagents were from Sigma-Aldrich Chemical Company unless otherwise stated. A l l ' H and C N M R spectra were recorded on Bruker Avance-400 or 1 3  Avance-600 spectrometers. Chemical shifts are reported in 8 units (ppm) using residual ' H and  1 3  C signals of deuterated solvents as reference: 5  H  (CDC1 ) 7.27, 8 3  C  (CDCI3)  77.23. A l l F N M R spectra were recorded on a Bruker Avance-300 spectrometer using 19  CF3CO2H  as reference and all P N M R spectra were recorded on a Bruker Avance-400 3 1  spectrometer using 85% H3PO4 as reference. Unless otherwise stated, all site directed mutations were generated using the method of QuikChange™ (Stratagene) according to the manufacturers' instructions.  A l l standard molecular biology procedures were  performed using the appropriate Qiagen kit according to the manufacturers' instructions.  7.2. Chapter 2 Experimental 7.2.1. Acknowledgements and Generous Gifts Ms. Melissa Schur at the N R C IBS in Ottawa performed C E analysis of Cst II reaction mixtures.  Ms. Cecilia C P . Chiu in Prof. Natalie Strynadka's laboratory  performed crystallization, X-ray diffraction and electron density refinement analysis of Cst II. Mark Chen supplied 3-fluoro sialic acid for the preparation of C M P 3F NeuAc. Dr. A . G . Watts provided Clostridium perfringens sialidase. Dr. Warren W. Wakarchuk supplied the H188A, Y156F and R129A mutants of Cst II, F C H A S E 3'-sialyl lactose for C E analysis, and purified C M P NeuAc synthetase.  C M P NeuAc was supplied as a  generous gift from Neose Technologies Inc. (Horsham, PA).  227  7.2.2. Expression and Purification of Cst II An overnight culture of E. coli (BL21 (DE3)) harboring pET28a(+) containing the Cst II gene was diluted 100 fold in L B media (containing 50 ug/mL kanamycin) and grown to an OD600 of 0.6 at which point IPTG was added to 0.5 m M and the induced culture was incubated overnight at 20°C.  Cells were harvested by light centrifugation  (4000 rpm Beckman J2-21) and the resulting cell pellets were resuspended in binding buffer (20 m M Tris (pH 8.3), 500 m M NaCl) containing Roche EDTA-free protease inhibitor cocktail and then disrupted using a French press. loaded onto a charged N i  2 +  Cleared supernatant was  IMAC column (5 mL G E Healthcare HisTrap Fast Flow)  equilibrated with binding buffer and washed with 10 mL of binding buffer and 10 mL of binding buffer containing 50 mM imidazole. Cst II was eluted using 2 column volumes each of binding buffer containing 100 mM, 250 mM, and 500 mM imidazole. Fractions containing purified Cst II, as determined by SDS P A G E analysis, were combined, concentrated and imidazole was removed using a pre-packed desalting column (PD10 Sephadex G-25M, Pharmacia Biotech) according to manufacturer's instructions. Purified protein was stored in 20 mM Tris (pH 8.3) at 4°C. Typical yields of purified Cst II were -10 mg/L.  7.2.3. A UV Spectroscopy-Based Enzyme-coupled Continuous Assay for Cst II Kinetic Characterization A continuous coupled assay, in which the release of CMP is coupled to the oxidation of N A D H (k = 340 nm, s = 6.22 mM"'cm"'), was used to monitor the activity of all Cst II variants. Absorbance measurements were obtained using Cary 300 and Cary 4000 UV-VIS spectrophotometers equipped with a circulating water bath and a Peltier  228  temperature controller, respectively. GraFit version 4.0 (Erithacus) was used to calculate kinetic parameters by direct fit of initial rates to the respective equations. The assay conditions in a 200 uL quartz cuvette are as follows: 50 m M HEPES (pH 7.5), 100 m M KC1, 10 m M M g C l , 10 m M M n C l , 0.7 m M PEP, 0.25 m M N A D H , 2 2  2  m M ATP, 1.0 mg/mL BSA, 5.9U of PK, 8.4U of L D H , 7.5U of N M P K , and varying concentrations of acceptor and donor substrates. The cuvette is left at 37°C until a stable rate of decreasing absorbance at 340 nm is established. This period is required to deplete the C M P present in the donor solution and any NDPs present in the A T P or N M P K solutions.  The stable rate of change in absorbance is the result of the spontaneous  hydrolysis of the relatively labile C M P NeuAc substrate. Once a stable background rate is established (~20 minutes), 4 uL of transferase solution is added and the solution is mixed thoroughly to initiate the assay. The concentration of enzyme is chosen such that the rate of change in absorbance resulting from transferase activity is constant for a period of more than 10 minutes.  The initial rate of transferase activity ( u M /minute) is  calculated using the millimolar extinction coefficient of N A D H (6220 M^cm" ). 1  7.2.4. A Capillary Electrophoresis Based Stopped Assay for Cst II Kinetic Characterization A stopped assay similar to that described previously was utilized to determine inhibition kinetic parameters for Cst II in the presence of the product inhibitor C M P (Gilbert et al., 1997). Assays were run in 10 (aL volumes in triplicate in 50 m M HEPES (pH 7.5) containing 10 m M M g C l and 24 |j.g/mL Cst II. When varying the concentration 2  of C M P NeuAc (50 - 500 \iM), FCHASE-3'-sialyl lactose acceptor was held at a fixed concentration of 500 u M . When varying the concentration of FCHASE-3'-sialyl lactose  229  (0.25 - 1 mM), C M P NeuAc donor was held at a fixed concentration of 500 u M . C M P concentration was varied from 0.1 to 1 m M . Assays were run for 10 minutes, during which time the rate of product formation was linear over the range of substrates used, and then quenched by the addition of an equal volume of stop solution consisting of 2% (w/v) SDS, 10 m M E D T A and 50% (v/v) MeOH and stored at -20°C prior to analysis by CE. Reaction mixtures were analyzed by C E using a Beckman Instruments P A C E 5510 equipped with a 3 mW Argon-ion laser-induced fluorescence detector (excitation at 488 nm, emission at 520 nm). The capillary was of bare silica (75 |j.m x 57 cm) with the detector at 50 cm. Following capillary conditioning (washing with 0.2 M NaOH for 2 minutes, water for 2 minutes and then equilibration with 20 m M sodium phosphate pH 7.5), samples were injected under pressure for 2-5 s and separation was performed at 18 kV, 75 uA. Peak integration was performed using the Beckman P A C E station software and data was analyzed using Grafit 4.0 (Erithacus software).  7.2.5. Synthesis of CMP 3F NeuAc C M P 3F NeuAc was synthesized using 3-fluoro sialic, CTP and C M P NeuAc synthetase according to a modified version of a previously published protocol (Simon et al., 1988); (Ohrui, 1988). 3-Fluoro sialic acid (20 mg, 61 umol) and CTP (50 mg, 92 pmol) were incubated with C M P NeuAc synthetase (1 mg/mL) and inorganic pyrophosphatase (50 units) in 200 m M Tris (pH 8.5) containing 20 m M M g C b and 0.2 m M DTT. Prior to the addition of enzymes, the pH was adjusted to 8.5 by the addition of NaOH. The solution was incubated at room temperature for -18 hours until reaction was complete (as determined by negative ion mode ESI-MS). Following completion of the reaction, enzyme was removed using a 5 kDa molecular weight cut off Centricon filter  230  (with the retained solution being washed twice with water).  The combined aqueous  solution containing the desired product was lyophilized, dissolved in 2 mL of water, filtered with a 0.2 pm syringe filter and loaded onto a preparative HPLC TSKgel Amide80 (Tosoh Bioscience) column. The desired product was purified using a flow rate of 6 mL/minute and a gradient of acetonitrile (A): water (B) (80:20 A:B hold for 20 minutes, linear gradient from 80:20 A : B to 65:35 A:B over 40 minutes, hold 65:35 A : B for 15 minutes).  The 6 mL fractions containing pure C M P 3F NeuAc (as determined by  negative ion mode ESI-MS) were combined, solvent removed under pressure, and the resulting aqueous solution was lyophilized to yield a white fluffy solid. 29.7 mg (77%); 'H-NMR (400 MHz) 5 (D 0): 7.85 (d, 1 H , J 7.6 Hz, H6), 6.01 (d, 1 H , J 7.6 Hz, H5), H  2  5.85 (d, 1 H , J 4.4 Hz, HI'), 4.78 (dd, J 48.4 Hz, J 2.0 Hz, H3"), 4.22 - 3.95 (m, 8 H , H2', H 3 \ H5', H5', H 4 \ H4", H5", H6"), 3.84 (m, 1 H , H8"), 3.76 (dd, 1 H , J 2.4 Hz, J 12 Hz, H9"), 3.50 (dd, 1 H , J 6.8 Hz, J 12.0 Hz, H9"), 3.33 (d, 1 H , J 9.6 Hz, H7"), 1.92 (s, 3 H , NAcCH );  19  3  F - N M R (282 MHz, proton decoupled) 5 (D 0); -131.52; P - N M R (162 31  F  2  MHz, proton decoupled) 5 (D 0); -4.54; HR ESI-MS: m/z = 631.1298 [M]~ (expected for P  2  C oH29FN Oi F: m/z = 631.1300). 2  4  6  7.2.6. Generation of Site-Directed Mutants The Y162F and Y156/162F mutants of Cst II were generated using the following primers. For the Y162F mutant pET28a(+) plasmid D N A containing Cst II was used as template. For the Y156/162F double mutant, pET28a(+) plasmid D N A containing Y156F Cst II was used as template. A l l constructs were sequenced to verify the presence of the mutation of interest.  231  Y162Ffwd: 5'- G A A TTG A T T TTT A T C A A A A T G GGT C A T CTT TTG CTT TTG A T A C T A A A C A A A A A A A T C TTT T - 3' Y162Frev: 5'- A A A A G A TTT TTT TGT TTA G T A T C A A A A G C A A A A G A T G A C C C A TTT T G A T A A A A A T C A A T T C - 3' Y l 56/162Ffwd: 5' - G A A A T T T A T CTT T C G G G A A T T G A T TTT TTT C A A A A T G G G T C A TCT TTT GCT TTT G - 3'  Y156/162Frev: 5'- C A A A A G C A A A A G A T G A C C C A T TTT G A A A A A A A T C A A TTC C C G A A A G A T A A A TTT C - 3'  7.2.7. Chemical Rescue of the H188A Mutant of Cst II Chemical rescue of the HI 88A mutant of Cst II was performed using the coupled continuous assay described in section 6.2.3.. Initial rates for the H188A (500 ug/mL) and wild-type (25 pg/mL) enzyme were determined at 0.5 m M C M P NeuAc with varying concentrations of sodium formate (0 - 1000 mM) or sodium azide (0 - 250 mM). Formation of an a-linked sialyl azide derivative was confirmed by negative ion mode ESI-MS and by the observed ability of Clostridium perfringens sialidase to convert the putative species to sialic acid as determined by TLC analysis (4: 2: 1: 0.1 EtOAc: MeOH: H 0 : AcOH). 2  7.3. Chapter 3 Experimental 7.3.1. Acknowledgements and Generous Gifts Dr. Warren W. Wakarchuk provided pCW plasmid D N A containing WT, Q189E, and Q189A versions of LgtC-25 and pCW-BovlO plasmid containing the gene encoding a3GalT-MalE fusion protein. Ms. Aidha Shaikh performed expression optimization of the I143D, N153D and Q185D variants of LgtC-25, as well the protein expression,  232  purification, kinetic and ESI-MS analysis of the Q185D variant of LgtC-25. Dr. Andrew G. Watts provided 2-deoxy-2-fluoro-a-D-mannose-l-phosphate for the synthesis of GDP 2F Man.  Ms. Cecilia C P . Chiu in Prof. Natalie Strynadka's laboratory performed  crystallization, X-ray diffraction and electron density refinement analysis of LgtC-25 Q189E. Prof. Annette Herscovics provided purified Kre2. Dr. Yuri Lobsanov in Prof. Lynne Howell's laboratory performed crystallization, X-ray diffraction and electron density refinement analysis of Kre2. Mr. Shouming He performed ESI-MS/MS analysis of labelled LgtC-25 peptides, Kre2 peptides, and whole protein and peptides of a3GalT. UDP Gal was supplied as a generous gift from Neose Technologies Inc. (Horsham, PA).  7.3.2. Expression and Purification of LgtC-25 A n overnight culture of E. coli (DH5a) harboring p C W containing the LgtC-25 gene was diluted 100 fold in 2xYT media (containing 100 ug/mL ampicillin) and grown to an OD600 of 0.6 at which point IPTG was added to 0.5 m M and the induced culture was incubated for 6 hours at 37°C Cells were harvested by low speed centrifugation (4000 rpm Beckman J2-21) and the resulting cell pellets were resuspended in 50 m M HEPES (pH 7.5) containing 100 m M NaCl and 1 m M benzamidine and then disrupted using a French press. Following clarification by high-speed centrifugation, NaCl (to 200 mM) and PEG-8000 (to 6% (v/v)) was added. Following incubation on ice for 1 hour, the solution was Centrifuged at 10, 000 rpm for 30 minutes and the supernatant was then dialyzed overnight against 20 m M Tris (pH 7.5). The following morning the dialysate was centrifuged at 25, 000 rpm, filtered with a 0.2 |^m syringe filter, diluted 1:1 with 20 m M Tris (pH 7.5) and loaded onto a 30 mL Macro Prep High Q strong anion exchange column (Bio Rad). Protein was eluted using a gradient of 0% to 30% NaCl (w/v) in 20  233  mM Tris (pH 7.5) over 4 column volumes using a ProSys workstation BioSepra FPLC. Fractions containing LgtC-25 (as determined by SDS PAGE analysis), were combined, concentrated to -3 mL and loaded onto a HiPrep 16/60 Sephacryl S-100 High Resolution size exclusion column equilibrated with 50 mM NH Ac (pH 7.0). Fractions containing 4  purified LgtC-25 were combined and concentrated to -10 mg/mL and stored at 4°C. Typical yields of purified protein were -20 mg per litre of culture.  7.3.3. Expression of LgtC-25 as a C-terminal 6-Histidine Fusion LgtC-25 was subcloned into pET29a(+) for expression as a C-terminal 6-histidine fusion protein using Ndel and Xhol restriction sites.  The LgtC-25 gene was PCR  amplified using pLLOl as template and the following primers: LgtC6HisCtermpET29NdeI: 5' - GTA GCG TGC ATA TGG ACA TCG TAT TTGCGGCAG A C G - 3 ' LgtC6HisCtermpET29XhoI: 5' - CGG TAG GCC TCG AGG TGC GGG A C G GCA AGT TTG CCG CGC C - 3' The resulting PCR product was subjected to Qiagen PCR reaction clean up kit purification according to manufacturer's instructions and then subjected to Ndel and Xhol double digestion overnight. The resulting digested PCR product and similarly digested pET29a(+) vector were purified from a 1.5% agarose gel using Qiagen's gel elute kit according to the manufacturer's instructions. Ligation of insert and vector was achieved using Roche's rapid ligation kit according to manufacturer's instructions and 2 uL of the resulting ligation mixture was then used to transform 50 pL of chemically competent Topp 10 E. coli (Invitrogen). Plasmid DNA from 3 of the resulting clones was subjected  234  to D N A sequencing, revealing the desired incorporation of the LgtC-25 gene into pET29a(+). A n overnight culture of E. coli (BL21 (DE3)) harboring pET29a(+) containing the LgtC-25 gene was diluted 100 fold in 2xYT media (containing 50 ug/mL kanamycin) and grown to an OD oo of 0.6 at which point IPTG was added to 0.5 m M and the induced 6  culture was incubated for 6 hours at 37°C. Cells were harvested by light centrifugation (4000 rpm Beckman J2-21) and the resulting cell pellets were resuspended in binding buffer (20 m M Tris (pH 7.5), 100 m M NaCl) containing Roche EDTA-free protease inhibitor cocktail and 1 m M benzamidine and then disrupted using a French press. Cleared supernatant was loaded onto a charged N i  2 +  I M A C column (5 mL G E Healthcare  HisTrap Fast Flow) equilibrated with binding buffer, washed with 10 mL of binding buffer and 10 mL of binding buffer containing 25 m M imidazole. LgtC-25 was eluted using 2 column volumes each of binding buffer containing 50 m M , 100 m M , 250 m M , and 500 m M imidazole. Fractions containing purified LgtC-25 (as determined by SDS P A G E analysis) were combined, concentrated and imidazole was removed using a prepacked desalting column (PD10 Sephadex G-25M, Pharmacia Biotech) according to manufacturer's instructions. Purified protein was stored in 20 m M Tris (pH 7.5) at 4°C. Typical yields of purified LgtC-25 were -25 mg per litre of culture.  7.3.4. A U V Spectroscopy-Based Enzyme-coupled Continuous Assay for Kinetic Characterization of LgtC and Kre2 A continuous coupled assay similar to that described previously (Gosselin et al., 1994; Ly et al., 2002), in which the release of UDP or GDP is coupled to the oxidation of N A D H (k = 340 nm, 8 = 6.22 mlvT'cm" ), was used to monitor the activity of all LgtC 1  235  variants and Kre2. Absorbance measurements were obtained using Cary 300 and Cary 4000 UV-VIS spectrophotometers equipped with a circulating water bath and a Peltier temperature controller, respectively. GraFit version 4.0 (Erithacus) was used to calculate kinetic parameters by direct fit of initial rates to the respective equations.  The assay  conditions were identical to those described in section 6.1.3. with the omission of N M P K and ATP.  7.3.5. ESI-MS Analysis of LgtC-25 Q189E 7.3.5.1. ESI-MS Conditions Mass spectra were recorded on a PE-Sciex API 300 triple quadrupole mass spectrometer (Sciex, Thornhill, Ontario, Canada) equipped with an Ion spray ion source. Whole proteins and peptides were separated by reverse-phase H P L C on a L C Packing UltiMate Micro HPLC system (Dionex, Sunnyvale, CA) that was directly interfaced with the mass spectrometer. In each of the M S experiments whole protein was loaded onto a C4 column (LC Packings, 300A, 1 x 150 mm) equilibrated with solvent A and then eluted with a step gradient of 60:40 A : B for 10 minutes, 40:60 A : B for 8 minutes, 30:70 for 4 minutes and 20: 80 for 4 minutes followed by 100% B for 5 minutes (solvent A : 0.05% trifluoroacetic acid/2% acetonitrile in water, solvent B : 0.045% trifluoroacetic acid/80% acetonitrile in water).  Solvents were pumped at a constant flow rate of 40 pL/min.  Spectra were  recorded in the single quadrupole scan mode (LC/MS) with the quadrupole mass analyzer scanned over a mass to charge ratio (m/z) range of 500 - 2500 Da with a step of 0.5 Da and dwell time of 1 ms per step. The ion source voltage (ISV) was set at 5.5 k V and the  236  orifice energy (OR) was set at 45 V . Protein masses were deconvoluted from multiply charged species using the Analyst 1.2 software package. For each of the M S experiments the proteolytic digest was loaded onto a C18 column (LC Packings, 100A pepMap, l x 150 mm) equilibrated with solvent A . Elution of the peptides was accomplished using a gradient (0-60%) of solvent B over 60 minutes followed by 85% solvent B over 20 minutes. Solvents were pumped at a constant flow rate of 50 uL/min. A post-column splitter was present in all experiments, splitting off 85% of the sample into a fraction collector and sending 15% into the mass spectrometer. Spectra were recorded in the single-quadrupole mode (LC/MS) with the quadrupole mass analyzer scanned over a m/z range of 300-2200 Da with a step of 0.5 Da and a dwell time of 1.5 ms per step. The ISV was set at 5.5 k V and the OR set at 45 V . Native or labelled peptides were sequenced in the single quadrupole mode by increasing the OR to 65 V .  7.3.5.2. Labelling of LgtC-25 Q189E LgtC-25 Q189E (25 uL, 1 mg/mL, in 20 m M HEPES buffer pH 7.5, 5 m M M n C b , 5 m M DTT) was incubated in the presence of 1 m M UDP Gal for 2 minutes or 1 m M U D P Glc for 2 hours to obtain optimal amounts of labelled protein. A l l samples were quenched by the addition of an equal volume of 6 M Urea and stored frozen prior to mass spectrometric analysis. To increase the relative proportion of labelled enzyme by 'pulling'  the  equilibrium over,  15  units  of  pyruvate  kinase  and  5 mM  phosphoenolpyruvate were added to the above incubations. Proteolysis of LgtC (25 \xL in 20 m M HEPES buffer pH 7.5, native or labelled, 1 mg/mL) was achieved by incubation with trypsin (25 uL in 20 m M HEPES buffer pH 7.5, 1 mg/mL) at room temperature until digestion was complete (30 minutes).  237  7.3.5.3. Turn-over of Covalently Labelled LgtC-25 Q189E Excess U D P Gal or U D P Glc was removed from LgtC-25 Q189E labelled (as determined by M S analysis) with either Gal or Glc by exchanging the labelling buffer conditions (described above) with that of 20 m M HEPES buffer, p H 7.5, 5 m M M n C l  2  and 5 m M DTT using an Ultrafree®-0.5 Centrifugal Filter Device (Millipore) following manufacturer's instructions to obtain a final protein concentration of 2 mg/mL. When labelled with Gal, the peak corresponding to labelled enzyme had disappeared following solvent exchange (~30 minutes).  When labelled with Glc, a significant peak  corresponding to labelled enzyme was observed following solvent exchange.  To this  solution (10 pL aliquots) an equal volume of 20 m M HEPES buffer, p H 7.5, 5 m M M n C l , 5 m M DTT (control) or 20 m M HEPES buffer, pH 7.5, 5 m M M n C l , 5 m M 2  2  DTT, 200 m M lactose was added.  Aliquots were incubated for predetermined time  periods prior to being quenched by the addition of 6 M Urea and stored frozen prior to M S analysis. A n observed first order rate constant for the turnover of LgtC-25 Q189E labelled with Glc was obtained from a plot of the logarithm of the relative peak height of the labelled species versus time.  7.3.6. Generation of the I143D, N153D, Q185D, and D190N Mutants of LgtC-25 The I143D, N153D, Q185D, and D190N mutants of LgtC-25 were generated using the following primers and pCWlgtC-25 C128/174S plasmid D N A as template. A l l constructs were sequenced to verify the presence of the mutation of interest.  238  I143Dfwd: 5'- C A G G A A G G C T A C A A A C A A A A A G A T GGT A T G G C G GAC G G C G A - 3 ' I143Drev: 5'- T C G C C G TCC G C C A T A C C A TCT TTT TGT TTG T A G C C T TCC T G - 3 ' N153Dfwd: 5'- G A C G G C G A A TAT TAT TTC G A T G C C G G C G T A T T G CT-3' N153Drev: 5'- A G C A A T A C G C C G G C A T C G A A A T A A T A T T C G C C G TC-3' Q185Dfwd: 5'- G A A C A A T A C A A G G A C G T G A T G G A T T A T C A G G A T ' C A G G A C ATT TTG - 3 Q185Drev: 5'- C A A A A T GTC C T G A T C C T G A T A A T C C A T C A C GTC CTTGTATTG TTC-3' D190Nfwd: 5'- G A T G C A A T A T C A G G A T C A G A A C A T TTT G A A C G G GCT GTT TA-3'  D190Nrev: 5'- T A A A C A G C C CGT T C A A A A TGT TCT G A T CCT G A T ATT G C A T C - 3 '  7.3.7. Synthesis of GDP 2F Man GDP 2F Man was synthesized using a modified version of a previously published protocol (Stick and Watts, 2002; Wittmann and Wong, 1997). The triethyl ammonium salt of 2-deoxy-2-fluoro-a-D-mannose-l-phosphate (-100 mg) was dissolved in 1 mL of water and loaded onto a 10 mL A G 50Wx2 (Bio Rad) cation exchange column equilibrated in the pyridinum form and then thoroughly eluted using water. The eluant was lyophilized to afford the desired pyridinium salt. To this pyridinium salt (30 mg, 72 umol), tri-n-octylamine (63 uL, 144 (j.mol) and 5 mL of dry pyridine were added and the mixture was evaporated and then co-evaporated 3 times with 5 mL of dry pyridine. G M P morpholidate (35 mg, 48 |jmol) was then added and this mixture was again coevaporated  239  3 times with 5 mL of dry pyridine. Tetrazole (10 mg, 144 pmol) and dry pyridine (5 mL) were added and the reaction left at room temperature under dry nitrogen for 10 days. Pyridine was evaporated and the resulting residue was resuspended in 100 m M NH4HCO3, and washed with diethyl ether. The washed aqueous solution was lyophilized, dissolved in 5 mL of water, filtered with a 0.2 pm syringe filter and loaded onto a preparative H P L C TSKgel Amide-80 (Tosoh Bioscience) column. The desired product was purified using a flow rate of 6 mL/minute and a gradient of acetonitrile (A): water (B) (80:20 A : B hold for 20 minutes, linear gradient from 80:20 A : B to 65:35 A : B over 40 minutes, hold 65:35 A : B for 15 minutes). The 6 mL fractions containing pure GDP 2F Man (as determined by negative ion mode ESI-MS) were combined, solvent removed under pressure, and the resulting aqueous solution was lyophilized to yield a white fluffy solid.  ' H , P , and 3 1  19  F N M R spectra were consistent with those reported previously  (Wong, 1998). 12.9 mg (42%); select data, F - N M R (282 M H z , proton decoupled) 8 19  F  (D 0): -128.2; P - N M R (162 MHz, proton decoupled) 8 (D 0): -11.1 (m, 1 P), -13.6 31  2  P  2  (m, I P ) ; ESI-MS: m/z = 606.2 [M]" (expected for C,6H 3FN Oi5P2": m/z = 606.1). 2  5  7.3.8. ESI-MS Analysis of Kre2 Kre2 (25 pL, 1 mg/mL, in 20 m M HEPES buffer pH 7.5, 5mM M n C l ) was, 2  incubated in the presence of 1 m M GDP Man or 1 m M GDP 2F Man or 1 m M GDP Man and 100 m M methyl a-D-mannoside or 1 m M GDP 2F Man and 100 m M methyl a-Dmannoside for time periods ranging from 2 minutes to 2 hours. In an attempt to increase the relative proportion of labelled enzyme by 'pulling' the equilibrium over, 15 units of pyruvate kinase and 5 m M phosphoenolpyruvate were also added to the above incubation  240  conditions.  Proteolysis of Kre2 (25 pL in 20 m M HEPES buffer pH 7.5, native or  putatively labelled, 1 mg/mL) was achieved by incubation with trypsin (25 pL in 20 m M HEPES buffer pH 7.5, 1 mg/mL) or pepsin (25 pL in 200 m M phosphate buffer pH 2.0, 1 mg/mL) at room temperature until digestion was complete (30 minutes).  ESI-MS of  peptide fragments was performed as described in section 6.2.5.1..  7.3.9. Expression and Purification of a3GalT as a MalE Fusion Protein The pCW-BovlO plasmid was transformed into electrocompetent E. coli (DH5a). Protein overexpression was achieved by standard IPTG (0.5 mM) induction with overnight expression at 20°C in 2xYT media (containing 100 pg/mL ampicillin). Cells were collected by light centrifugation (4000 rpm Beckman J2-21), resuspended in binding buffer (20 m M TRIS (pH 7.5), 200 m M NaCl) containing Roche EDTA-free protease inhibitor cocktail and disrupted with a French Press. Protein purification was achieved by loading clarified supernatant onto a 15 mL Amylose Resin Fast Flow column (New England Biolabs) equilibrated with binding buffer at a flow rate of 0.8 mL/min. Following loading, the column was washed with 12 column volumes of binding buffer and then eluted with 4 column volumes of binding buffer containing 10 m M maltose. During elution, 3 mL fractions were collected and those containing pure a3GalT (as determined by SDS P A G E analysis) were combined, concentrated and maltose was removed using a pre-packed desalting column (PD10 Sephadex G-25M, Pharmacia Biotech) according to manufacturer's instructions.  Purified protein was stored as a 5  mg/mL solution in 20 m M Tris (pH 7.5). Typical yields of purified protein were ~15 mg per litre of culture.  241  7.3.10. ESI-MS Analysis of o3GalT a3GalT (25 uL, 1 mg/mL, in 20 m M HEPES buffer pH 7.5, 5 m M MnCl ) was 2  incubated in the presence of 1 m M U D P Gal or 1 m M U D P Gal and 100 m M lactose for time periods ranging from 2 minutes to 2 hours.  A l l samples were quenched by the  addition of an equal volume of 6 M Urea or 3 M phosphate (pH 2.0) and stored frozen prior to mass spectrometric analysis. In an attempt to increase the relative proportion of labelled enzyme by 'pulling' the equilibrium over, 15 units of pyruvate kinase and 5 m M phosphoenolpyruvate were also added to the above incubation conditions. Proteolysis of a3GalT (25 uL in 20 m M HEPES buffer pH 7.5, native or putatively labelled, 1 mg/mL) was achieved by incubation with trypsin (25 uL in 20 m M HEPES buffer p H 7.5, 1 mg/mL) or pepsin (25 L I L in 200 m M phosphate buffer pH 2.0, 1 mg/mL) at room temperature until digestion was complete (30 minutes).  ESI-MS analysis of whole  protein and peptide fragments was performed as described in section 6.2.5.1..  7.4. Chapter 4 Experimental 7.4.1. Acknowledgements and Generous Gifts Dr. Sachdev Sidhu from the Department of Protein Engineering at Genentech in San Francisco provided the pSAV phagemid for optimized phage display of S A V . Dr. Spencer  Williams  provided  3-[N-(D-biotinyl)-13-amino-4,7,10-  trioxatridecanylaminocarbonyl] propanoic acid for the synthesis of a lactose-biotin conjugate.  Dr. Warren W. Wakarchuk provided pSJF6 phagemid as the starting vector  for 3 + 3 phage display of LgtC and spGalU enzyme.  242  7.4.2. Synthesis of CDP Gal Cytosine  5'-monophosphomorpholidate  4-morpholine-N,N'-  dicyclohexylcarboxamidine salt was prepared essentially as described (Moffatt, 1966, 1961). Briefly, a solution of D C C (3.8g, 18.6 mmol) in t-butyl alcohol (45 mL) was added dropwise to a refluxing solution of C M P (1.0 g, 3.1 mmol) and morpholine (1.6 mL, 18.6 mmol) in t-butyl alcohol and water (30 mL each) over a period of 3 hours. The combined mixture was then refluxed for an additional 2 hours until all of the C M P was consumed. The mixture was then cooled to room temperature, precipitate was removed by filtration, and t-butyl alcohol was removed under reduced pressure from the combined filtrates. The remaining aqueous solution was washed with diethyl ether and lyophilized to yield an off white gum. The gum was triturated twice using diethyl ether to afford a white solid. 1.3 g (61% recovered). The *H and P N M R spectra were consistent with 3 1  those previously published (Moffatt, 1966). CDP Gal was prepared by reacting the above C M P morpholidate reagent (240 mg, 0.35 mmol) with the pyridinium salt of a-D-galactose-1-phosphate (200 mg, 0.53 mmol) as described for the synthesis of GDP 2F Man in section 6.3.7.. Preparative scale H P L C purification using an Amide-80 column as described for the purification of GDP 2F Man afforded the desired CDP Gal product as a white solid. 88 mg (43%); H - N M R (400 MHz): 6 ( D 0 ) : 7.98 (d, 1 H, J 1  H  2  5>6  7.24 Hz, H-6), 6.10 (d, 1  H , J , 7.24 Hz, H-5), 5.82 (d, 1 H , h\r 3.6 Hz, H - l ' ) , 5.50 (dd, 1 H , J - 7.2 Hz, J 5  6  r  P  r > 2  »  3.6 Hz, H - l " ) , 4.27-4.00 (m, 6 H , H-2', H-3', H-4', H-5 ', H-5 ', H-5"), 3.88 (m, 1 H , H a  b  4"), 3.77 (dd, 1 H, J , » 10.2 Hz, J " , " 3.0 Hz, H-3"), 3.69-3.54 (m, 3 H , H-2", H-6 ", r  3  3  4  a  243  H6 ");  3 1  b  P - N M R (162 MHz) 5 (D 0); -10.2 (d, 1 P, J P  2  P a P p  21.0 Hz), -11.7 (d, 1 P, J  P a P p  21.0 Hz); ESI-MS: m/z = 564.3 [M]" (expected for C H24N Oi P2": m/z = 563.1). 15  3  6  7.4.3. Generating the N10D Mutant of LgtC-25 The N10D mutant of LgtC-25 was generated by the method of QuikChange™ (Stratagene) using the following primers and pCWlgtC-25 C128/174S plasmid D N A as template. The resulting construct was sequenced to verify the presence of the mutation of interest. LgtCNIODfwd: 5' - C G G C A G A C G A C G A C T A T G C C G C C T A T C - 3' LgtCNIODrev: 5' - G A T A G G C G G C A T A G T CGT CGT C T G C C G - 3'  7.4.4. Kinetic Analysis of CDP Gal as a Substrate for LgtC-25 and LgtC N10D Kinetic analysis of wild-type and N10D LgtC was performed as described in section 6.2.4 using variable concentrations of CDP Gal or UDP Gal as donor substrate at a fixed saturating concentration (160 mM) of lactose acceptor substrate. For the wildtype enzyme, the concentration of UDP Gal was varied from 5 to 125 u M at an enzyme concentration of 0.20 ng/mL and the concentration of CDP Gal was varied from 25 to 500 u M at an enzyme concentration of 0.43 [ig/mL.  For the N10D mutant, the  concentration of UDP Gal was varied from 25 to 500 u M at an enzyme concentration of 1.0 |j.g/mL and the concentration of CDP Gal was varied from 25 to 1000 u M at an enzyme concentration of 4.0 ug/mL.  244  7.4.5. Subcloning LgtC-25 for Expression as an M13 phage pill Fusion Protein LgtC-25 was subcloned into pSJF6 phagemid for expression as a C-terminal M l 3 phage pill fusion protein containing an N-terminal OmpA leader sequence using EcoRI and BgUl restriction sites. The LgtC-25 gene was P C R amplified using pCWLgtC-25 C128/174S plasmid D N A as template and the following primers: LgtC-phage-5p: 5' - G G G GCG ACC GCG  G G G G A A TTC A T G A A A A A A A C C GCT A T C A T C G C A GTT G C A C T G GCT GGT TTC G C T GTT G C G C A A G C C A T G G A C A T C G T A TTT GCA GAC G - 3 '  LgtC-phage-3p: 5' - G G G G G G A G A TCT G T G C G G G A C G G C A A G TTT GCC GCG C C - 3 ' The resulting P C R product was subjected to Qiagen's P C R reaction clean up kit according to manufacturer's instructions and then subjected to a nested P C R reaction to increase the overall yield of D N A product using LgtC-phage-3p and the following primer: LgtCPh5pSh: 5'- G G G G G G G A A TTC A T G A A A A A A A C C G - 3' The resulting P C R product was subjected to the Qiagen P C R reaction clean up kit according to manufacturer's instructions and then subjected to EcoRI and BgUl double digestion overnight. The resulting digested P C R product and similarly digested pSJF6 vector were purified from a 1.5% agarose gel using Qiagen's gel elute kit according to manufacturer's instructions. Ligation of insert and vector was achieved using Roche's rapid ligation kit according to manufacturer's instructions and 2 uL of the resulting ligation mixture was then used to transform 50 uL of chemically competent Topp 10 E. coli (Invitrogen). Colony PCR using commercial M13fwd and M13rev primers and D N A from 10 isolated transformants as template revealed one potential candidate containing an insert of the appropriate size. However, D N A sequencing of phagemid D N A isolated  245  from this clone revealed the presence of a 7 base pair deletion from the OmpA leader sequence region. These 7 base pairs were reintroduced by P C R amplifying the entire phagemid (containing the insert with the 7 base pair deletion) using Ultra high fidelity pfu D N A polymerase (Stratagene) and the following primers that contain Aarl restriction sites and the missing 7 base pairs: OmpA7infwd: 5'- G G A GGT C G C A C C TGC A T T G T C G C A GTT G C A C T G GCT GGT TTC GCT A C C GTT G C G C - 3' OmpA7inrev: 5 ' - G G A GT C G C A C C TGC T A C TGC G A T C G C G A T A G C GGT TTT TTT C A T G A A TTC T A C C - 3' Following P C R cycling the reaction was subjected to Dpnl digestion to degrade all template D N A (containing the 7 base pair deletion) and then subjected to Qiagen's P C R reaction clean up kit according to manufacturer's instructions. Overnight Aarl digestion was followed by purification from a 1.5% agarose gel using Qiagen's gene elute kit. The resulting linear Aarl digested phagemid was ligated using Roche's rapid ligation kit and transformed into chemically competent one shot TopplO E. coli (Invitrogen). Extensive D N A sequencing of phagemid D N A isolated from one of the transformants revealed the presence of the desired complete LgtC gene (fused via its C-terminus to the pill and containing an N-terminal OmpA leader sequence) within the M C S . The resulting phagemid was designated pLLOl.  7.4.6. VCSM13 Helper Phage Production A modified version of Cold Spring Harbor protocol 10.2 was used (Barbas III, 2001).  A single isolated colony of E. coli (XL1 Blue F') grown on L B agar plates  (containing 10 pg/mL tetracycline) was used to inoculate 10 mL of SB media (containing 10 pg/mL tetracycline). This culture was incubated at 37°C (with shaking) until an OD600  246  of 0.3 was achieved at which point 3 x 10  10  VCSM13 helper phage were added.  Following an infection period of 30 minutes, during which the culture was incubated at 37°C without shaking to allow regrowth of sheared pili, kanamycin was added to a final concentration of 25 pg/mL and the culture was left to incubate overnight at 37°C (with shaking). The following morning, E. coli cells were heat killed at 65°C for 15 minutes and the culture was spun down twice at 4000 rpm in a tabletop centrifuge to remove bacterial debris. The resulting supernatant was aliquoted into 1 mL volumes and stored at 4°C. Titers o f ~ 1 0 pfu/mL were typical. n  7.4.7. Small Scale PEG Precipitated LgtC-25-Displaying Phage Preparation A modified version of Cold Spring Harbor protocol 11.2 was used (Barbas III, 2001). A single isolated colony of E. coli (XL1 Blue F') harboring the pLLOl phagemid was used to inoculate 4 mL of SB media (containing 100 pg/mL ampicillin and 10 pg/mL tetracycline). This culture was incubated for 11 hours at 37°C (with shaking), at which point 50 pL of helper phage solution was added and the culture was incubated at 37°C without shaking for 30 minutes and then for 90 minutes at 37°C with shaking. Following this infection period, kanamycin was added to a final concentration of 50 pg/mL and the culture was incubated overnight at 37°C with shaking.  The following morning, the  culture was spun down at 3500 rpm in a tabletop centrifuge for 15 minutes and to 1.2 mL of supernatent, 0.3 mL of a sterile 5x concentrated PEG/NaCl solution (200 g PEG-8000, 150 g NaCl per liter) was added and the mixture left to incubate on ice for 30 minutes. The solution was then spun down at 14,000 rpm at 4°C for 10 minutes, the supernatant was removed and the resulting white pellet was spun down again at 14,000 rpm at 4°C for  247  1 minute. This pellet was then resuspended in 1.5 mL of l x PEG/NaCl solution and the described precipitation procedure repeated. The resulting pellet was then resuspended in 50 uL of 50 m M HEPES (pH 7.5).  7.4.8. Quantification of Phage by UV Spectroscopy A modified version of Cold Spring Harbor protocol 15.6 was used (Barbas III, 2001). Phage concentrations were determined by measuring absorbance of solutions at 269 nm according to nucleotide content of phage using an averaged extinction coefficient of 1.006 x 10 M - cm" according to the following equation: 4  1  Phage particles per mL =  (adjusted  A269)  x (6 x l O ) 16  number of nucleotides in the phage genome where adjusted  A269 = (A 6o) 2  - (A269-background) - (A3 o) 2  7.4.9. Western Blot Analysis of pLLOl and pSJF6 Derived Phage Solutions saturated with PEG-precipitated phage were diluted 2x with SDS loading solution (containing P-mercaptoethanol) and boiled for 10 minutes prior to being directly loaded onto each of two identically loaded SDS P A G E Gels and proteins were then resolved by standard electrophoresis.  One of the resulting gels was subjected to  Coomassie blue analysis and the other to Western blot analysis using a Bio-Rad semi-dry transfer apparatus according to the manufacturer's instructions.  Detection of c-myc  containing proteins on the resulting blot was achieved by an immunoblotting procedure using Amersham's Anti-myc-HRP antibody according to the manufacturer's instructions.  Optimal detection was achieved by a 1-minute incubation with HRP substrate and a 40 minute exposure.  7.4.10. Compartmentalized Phage Display Procedure E. coli (XL1 Blue F') cells (~10 ) harboring pLLOl were emulsified from an 8  aqueous solution into ~10 aqueous ~5 fL compartments. 9  Washed cells derived from  overnight culture were resuspended in 50 m M HEPES (pH 7.5) containing 5 m M M n C b and 100 ug/mL ampicillin and immediately prior to the addition of ten volumes of an ice cold mixture of AbilEM90™ in light mineral oil (2.9% w/w), a 10 fold concentrate of LgtC reaction components (50 m M DTT, 50 m M U D P Gal, 5 m M fluorescein lactose conjugate) and 2 x 10 VCSM13 helper phage were added to the resuspended cells. 10  Emulsification on ice for 5 minutes at 9500 rpm using an I K A T-25™ homogenizer resulted in the production of a milky white suspension that was incubated at 37°C overnight.  The following day, the mixture was centrifuged (14, 000 rpm) and the oil  phase was removed. The aqueous layer was diluted 5x with 50 m M HEPES (pH 7.5) and washed twice with water-saturated diethyl ether.  Remaining ether was immediately  removed from the aqueous phase using a Speed Vac and the resulting phage-containing aqueous solution was subjected to PEG precipitation as described in section 6.4.7.  7.4.11. Testing Infectivity of Emulsified LgtC-25-Displaying Phage A modified version of Cold Spring Harbor protocol 15.2 was used (Barbas III, 2001). Starved E. coli (XL1 Blue F') cells were generated according to a version of the Cold Spring Harbour protocol 15.1 (Barbas III, 2001). Starved cells were infected with  249  PEG-precipitated phage at a ratio of 50:1 (cells : phage) by incubation for 15 minutes at room temperature (without shaking). Antibiotic resistance was induced by addition of 1 mL of L B media containing a low concentration of ampicillin (2 pg/mL). Following 1 hour incubation at 37°C (with light shaking o f - 1 0 0 rpm), 200 pL of lOx, lOOx, and lOOOx fold dilutions were plated on L B agar plates (containing 100 pg/mL). Following overnight incubation at 37°C the number of resulting colonies was used to determine the infectivity of LgtC-displaying (or control) phages.  7.4.12. Synthesis of a Biotin Lactose Conjugate 3-[N-(D-Biotinyl)-13-amino-4,7,10-trioxatridecanylaminocarbonyl]propanoic acid was activated as the pentafluorophenyl ester as described previously (Williams et al., 2006). Briefly, to a solution of the acid (50 mg, 92 pmol) in dry D M F (2 mL), E t N (20 3  pL, 140 pmol) and pentafluorophenyl trifluoroacetate (24 pL, 140 pmol) were added and the reaction left stirring at room temperature for 5 hours under dry nitrogen gas. The solvent was evaporated and the resulting residue was triturated with diethyl ether twice to afford the pentafluorophenyl ester (53.8 mg, 82% recovery). The white residue was dried under vacuum before direct use in the following coupling reaction. The crude pentafluorophenyl ester (50 mg, 70 pmol) and p-aminophenyl P-Dlactopyranoside (20 mg, 46 pmol)(Toronto Research Chemicals, Toronto) were dissolved in a dry mixture of D M F and pyridine (5 and 30 mL respectively) and left stirring overnight at room temperature. Solvent was removed under reduced pressure and the resulting residue was lyophilized from water, dissolved in 2 mL of water, filtered with a 2 pm syringe filter and then subjected to preparative HPLC purification using an Amide-80 column as described for the purification of GDP 2F Man.  250  17.1 mg (38%); ' H - N M R (400 MHz): 5 (D 0): 7.45-7.38 (m, 2 H , Ar), 7.19-7.12 (m, 2 H  2  H, Ar), 5.14 (d, 1 H , J , 8.0 Hz, H - l ) , 4.61-4.57 (m, 1 H), 4.50 (d, 1 H , J ;2  r > r  7.6 Hz, H -  1'), 4.41-4.37 (m, 1 H), 4.03-3.50 (m, 30 H), 3.35-3.21 (m, 6 H), 2.98 (dd, 1 H , J 5.2 Hz, J 13.2 Hz), 2.80-2.59 (m, 5 H), 2.24 (m, 2 H), 1.85-1.51 (m, 8 H), 1.45-1.32 (m, 2 H); 13  C - N M R (100 MHz) 6 (D 0): 176.8, 174.8, 173.5, 165.5, 154.1, 148.2, 139.8, 132.2, C  2  123.7, 117.3, 103.1, 100.4, 78.2, 75.6, 75.1, 74.4, 72.8, 72.7, 71.2, 69.8 (x2), 69.7, 69.6, 68.8, 68.6, 68.5, 62.3, 61.2, 60.4, 60.1, 55.6, 39.9, 36.5, 35.7, 32.1, 31.5, 31.4, 28.5 (x2), 28.1,  27.9, 25.4;  H R ESI-MS: m/z = 984.4104  [M + N a ]  +  (expected  for  C H67N Oi8SNa : m/z = 984.4100). +  42  5  7.4.13. Generating a Library of LgtC-25 Variants by Error Prone PCR Libraries of LgtC variants were generated using established protocols (Vartanian et al., 2001). Briefly, 2-10 ng of pCW-LgtC-25C138/174S plasmid D N A was used as template to P C R amplify the LgtC-25 gene using Taq polymerase and reaction conditions containing dNTP ratios of 1:5 A C : T G supplemented with 0.25 or 0.5 m M M n C l . The 2  following primers, used to amplify the LgtC-25 gene in these reactions, contained .EcoRI and Sail restrictions site such that the resulting P C R product could be subcloned into pUC18 vector. LgtCfwdEcoRI: 5'- G T A G C G T G G A A T T C A T G G A C A T C G T A T TTG CGG CAG A C G - 3 ' LgtCrevSall: 5'- C G G T A G G C G T C G A C G TGC G G G A C G G C A A G T TTG CCG C G C C - 3 ' The P C R program consisted of 30 cycles (95°C for 0.5 minutes, 53°C for 0.5 minutes, 72°C for 2 minutes). The PCR products were subjected to PCR reaction clean up, double  251  digested with EcoRI and Sail overnight, and purified on a 1.5% agarose gel. The gelpurified double-digested P C R product was ligated into similarly digested pUC18 vector and then transformed into E. cloni™ ultra competent E. coli cells using the UltraClone D N A ligation and transformation kit (Lucigen) according to manufacturer's instructions. Transformation efficiency, which dictates library size, was assessed by plating serially diluted transformed cells. Using the UltraClone kit and pUC18 as vector, libraries of >10  transformants were achieved.  Transformed cells were grown overnight in L B  (containing 100 ug/mL ampicillin) and the library plasmid D N A was isolated. D N A sequencing of plasmid D N A isolated from random colonies revealed that a desired mutation rate of ~2 amino acid mutations per gene was achieved using 2 ng of template and 0.25 m M M n C l . 2  7.4.14. Subcloning Verotoxin for Expression as a C-Terminal 6-Histidine Fusion Protein Verotoxin was subcloned from pVT21 into pET29a(+) for expression as a Cterminal 6-histidine fusion protein using Ndel and Xhol restriction sites. The verotoxin gene was PCR amplified using pVT21 as template and the following primers: VT6HCpETFW: 5'- G T A G C G TGC A T A T G A C G C C T G A T T G T G T A A C T G G A A A G G TG-3' VT6HCpETRV: 5'- C G G T A G G C C T C G A G A C C G G A G C C A G A A C G A A A A A T A A C T T C A C T G A A T C C C CCT C C - 3' The resulting P C R product was subjected to the Qiagen P C R reaction clean up kit according to manufacturer's instructions and then subjected to Ndel and Xhol double digestion overnight. The digested P C R product and similarly digested pET29a(+) vector were purified from a 1.5% agarose gel. Ligation of insert and vector was achieved using  252  Roche's rapid ligation kit according to manufacturer's instructions and 2 uL of the resulting ligation mixture was then used to transform 50 uL of chemically competent Topp 10 E. coli (Invitrogen). Plasmid D N A from 3 of the resulting clones was subjected to D N A sequencing, revealing the desired incorporation of the verotoxin gene into pET29a(+). A n overnight culture of E. coli (BL21 (DE3)) harboring pET29a(+) containing the verotoxin gene was diluted 100 fold in L B media (containing 50 |j.g/mL kanamycin) and grown to an OD oo of 0.6 at which point IPTG was added to 0.5 m M and the induced 6  culture was grown overnight at 25°C. Cells were harvested by light centrifugation and the resulting cell pellets were resuspended in binding buffer (20 m M Tris (pH 7.5), 100 m M NaCl) containing Roche EDTA-free protease inhibitor cocktail and then disrupted using a French press. Cleared supernatant was loaded onto a charged N i  2 +  I M A C column  (5 mL G E Healthcare HisTrap Fast Flow) equilibrated with binding buffer and then washed with 10 mL of binding buffer and 10 mL of binding buffer containing 50 m M imidazole.  Verotoxin was eluted using 2 column volumes each of binding buffer  containing 100 m M , 250 m M , and 500 m M imidazole. Fractions containing purified verotoxin (as determined by SDS P A G E analysis) were combined, concentrated and imidazole was removed using a pre-packed desalting column (PD10 Sephadex G-25M, Pharmacia Biotech) according to manufacturer's instructions. Typical yields of purified verotoxin were in excess of 100 mg/L.  253  7.4.15. Subcloning of SAV-pVIII Fusion Protein The gene encoding the SAV-pVIII fusion optimized for type 8 + 8 phage display (Sidhu, 2000) was subcloned from pSAV into pLLOl for polycistronic expression using a single Hindlll  restriction site. The SAV-pVIII gene was PCR amplified using pSAV as  template and the following primers: SASFwdHind: 5' - TTC G A T A A A A G C T T G G A G GTT A T A T G A A A A T A A A A A C A G G T G C A C G C A TCC T C G C - 3' SASRevHind: 5' - TTC G A T A A A A G C TTA A T A T C A GCT TGC TTT C G A GGT G A A TTT CTT A A A C A G C - 3' The forward primer was designed to include a ribosome binding site downstream of the Hindlll  restriction site prior to the S A V start codon. The resulting P C R product was  subjected to Qiagen P C R reaction clean up kit according to manufacturer's instructions and then subjected to Hindlll  digestion overnight. The resulting digested P C R product  and similarly digested pLLOl vector were purified from a 1.5% agarose gel using Qiagen's gel elute kit according to manufacturer's instructions. Ligation of insert and vector was achieved using Roche's rapid ligation kit according to  manufacturer's  instructions and 2 pL of the resulting ligation mixture was then used to transform 50 pL of chemically competent Topp 10 E. coli (Invitrogen).  Colony PCR using template  derived from 10 of the resulting colonies was performed using commercial M13fwd primer and a custom internal LgtC sequencing primer to determine candidates containing a single insert with the correct orientation. Plasmid D N A from 3 such positive candidates was then subjected to D N A sequencing analysis to confirm incorporation of the S A V gene in the correct orientation into pLLOl.  Extensive D N A sequencing was then  performed on one of the positive clones to ensure that the entire desired sequence within  254  the M C S was present.  This phagemid containing the desired insert was designated  pLL34.  7.4.16. ELISA Analysis of pSAV and pLL34 Derived Phage PEG-precipitated phage derived from either pLLOl, pSAV or pLL34 phagemid were subjected to ELISA analysis to determine the level of S A V display according to a modified version of a previously published procedure (Sidhu et al., 2000). Immobilized biotin was used as the capture target and an Amersham detection module for recombinant phage A b system was used for detection. Immulon 2HB (Nunc) 96-well plates were incubated with 200 uL of BSA-biotin conjugate (10 |ag/mL in 50 m M Na2C03 pH 9.6) for 2 hours at room temperature in a humidified container. Wells were then emptied, blotted on paper towel and blocked by treatment with 300 uL of blocking buffer (0.2% B S A in PBS) for 1 hour. Wells were then emptied, blotted on paper towel and washed 8 times with PBS containing 0.05 % Tween-20. In pre-blocked plates, phage were serially diluted using an equal volume of blocking buffer and incubated for 30 minutes at room temperature in a humidified container. To the wells of another pre-blocked 96-well plate, 200 [iL of the diluted phage solutions were added and left to incubate for 2 hours at room temperature in a humidified container. Following incubation, wells were emptied, blotted on paper towel and washed 8 times with PBS containing 0.05% Tween-20. Detection of immobilized phage was then achieved by treatment of wells with 200 pL of HRP-anti-M13 Ab conjugate (diluted 1:1000 from the commercial stock in blocking buffer). Following incubation for 1 hour at room temperature in a humidified container, the wells were emptied, blotted on paper towel and washed 8 times with PBS containing 0.05% Tween-20 and then twice with  255  PBS. Following washing, 200 pL of ABTS (0.2 mg/mL) / H 0 (0.02 % v/v) solution 2  2  was added and left to incubate for 40 minutes, at which point immobilized phage were quantified by measuring the absorbance at 410 nm.  7.5. Chapter 5 Experimental 7.5.1. Acknowledgements and Generous Gifts Dr. Andrew Watts provided 6 0 B z Man and 4 0 B z X y l . Dr. Warren Wakarchuk provided FCHASE-Lactose. Dr. David Zechel provided dNP oc-D-galactoside and dNP P-D-galactoside.  Dr. Karena Thieme provided pNP a-D-sialoside.  Benzyl P-D-  cellobioside and benzyl P-D-xyloside were obtained from the Withers laboratory collection of custom sugars.  7.5.2. LgtC-Catalysed Galactosylation of Modified Alternative Acceptor Substrates Acceptor substrate (5 mM), U D P galactose (7.5 mM) and LgtC (-0.5 mg/mL) were incubated for -16 hours at room temperature in 20 m M HEPES buffer (pH 7.5) containing 50 m M NaCl, 5 m M M n C l and 5 m M dithiothreitol. Upon completion, 2  reactions were partially purified with SepPak CigPlus Cartridges (0.39 g, Waters) by washing with H P L C grade water (10 mL) and eluting with HPLC grade M e O H (10 mL). Lyophilized  products  were  then  subjected  to  standard  per-O-acetylation  with  pyridine/Ac 0 followed by a final purification by column chromatography (10:1 to 3:1 2  toluene/EtOAc).  A l l fractions containing disaccharide products were combined and  analyzed by ' H N M R .  In contrast to the cases where underivatized monosaccharides  were used as acceptors and mixtures of regio-isomers were observed, when derivatized  256  acceptors were used, reactions were found to be completely regio-selective beyond the limit of U N M R detection. Of potential utility, in the cases where pNP Glc, octyl Glc l  and octyl Gal were used as acceptors, a trisaccharide product was detected (confirmed by ESI MS) accounting for the lack of near quantitative yields.  7.5.3.  NMR Analysis of LgtC Products Specific assignments of individual ' H peaks were unambiguously determined  using two-dimensional *H C O S Y N M R .  Assignments were made by examining the  couplings within the ring system and the locations of glycosidic linkages were determined on the basis of chemical shift. Because the acetate group is more electron withdrawing than a glycoside substituent, the proton attached to the carbon involved in the glycosidic bond resonates further upfield (<4.50 ppm) in a region clearly distinct from anomeric protons or protons attached to carbons bearing an acetate group. Ji, 2 coupling constants were used to assign the configuration of the anomeric linkage (a or P).  para-Nitrophenyl (2,3,4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(l->4)-2,3»6-tri-0acetyl-P-D-glucoside:  3.6 mg (51%); ' H - N M R (400 MHz): 5 (CDC1 ): 8.23 (m, 2 H , Ar), 7.08 (m, 2 H , Ar), H  3  5.49 (d, 1 H , J,, 4.0 Hz, H - l ' ) , 5.46 (d, 1 H , U\y 2.5 Hz, H-4'), 5.35 (t, 1 H , J 2  3 2  = J3.4 8.8  Hz, H-3), 5.29-5.24 (m, 2 H , H - l , H-3'), 5.18-5.12 (m, 2 H , H-2, H-2'), 4.50 (dd, 1 H , J , 2.7 Hz, J 6  5  6s6  12.1 Hz, H-6), 4.30-4.19 (m, 2 H , H-5', H-6), 4.14-4.07 (m, 3 H , H-4, H -  6', H-6'), 3.97-3.90 (m, 1 H, H-5), 2.15 (s, 3 H , Ac), 2.09 (s, 3 H , Ac), 2.08-2.06 (4xs, 12 H , 4xAc), 2.00 (s, 3 H , Ac); C - N M R (100 MHz) 5 (CDCI3): 13  C  171.1, 170.6, 170.4,  170.2, 170.1, 169.8, 161.3, 143.5, 128.9, 126.0 (x2), 116.8 (x2), 97.8, 96.5, 75.3, 73.0,  257  72.7, 71.7, 67.9, 67.7, 67.4, 67.0, 63.1, 61.7, 21.1-20.8 (x7);  H R ESI-MS: m/z =  780.1987 [M + Na] (expected for C 2H39N0 oNa : m/z = 780.1963). +  +  3  2  Benzyl (2,3,4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(l->3)-2,4-di-0-acetyl-P-Dxyloside: 48 mg (86%); H - N M R (400 MHz): 8 (CDC1 ): 7.40-7.13 (m, 5 H , Ar), 5.39 (d, 1 H , 1  H  Ji',2-  3  3.8 Hz, H - l ' ) , 5.29 (d, 1 H , J ',r 2.3 Hz, H-4'), 5.23 (dd, 1 H , h\r 4  10.9 Hz, J  3 ; 4  '  3.3  Hz, H-3'), 5.07-4.98 (m, 2 H , H-2', H-2), 4.87-4.81 (m, 2 H , H-4, BnCH), 4.60 (d, 1 H , Ji, 4.7 Hz, H - l ) , 4.52 (d, 1 H , J 11.9 Hz, BnCH), 4.34-4.28 (m, 1 H , H-5'), 4.19 (dd, 1 2  H, J  5 > 4  4.0 Hz, J ,5 12.2 Hz, H-5), 4.04-3.92 (m, 2 H , H-6',H-6'), 3.86 (t, 1 H , J , = J 5  3  6.3 Hz, H-3), 3.42 (dd, 1 H , J  5 > 4  6.1 Hz, J  5 ; 5  2  3 > 4  12.2 Hz, H-5), 2.10 (3xs, 9 H , 3xAc), 2.00  (3xs, 9 H , 3xAc); C - N M R (100 MHz) 5 (CDC1 ): 170.8, 170.5, 170.3, 170.2, 169.9, 13  C  3  169.4, 137.0, 129.2, 128.8, 128.4, 128.3, 128.2, 125.5, 98.4, 97.3, 75.0, 70.3, 70.2, 68.2, 68.0, 67.3, 66.9, 61.5, 60.5, 21.1, 21.0, 20.9 (x2), 20.8 (x2)  ; H R ESI-MS: m/z =  677.2057 [M + Na] (expected for C H O N a : m/z = 677.2058). +  +  3 0  3 8  1 6  w-Octyl (2,3,4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(1^3)-2,4,6-tri-0-acetyl-P-Dglucoside: 19.9 mg (52 %); ' H - N M R (400 MHz):  8 (CDC1 ): 5.43 (m, 1 H , H-4'), 5.31 (d, 1 H , J , ' H  3  r  3.2 Hz, H - l ' ) , 5.23 (dd, 1 H , h',r 11.0 Hz, J , > 3.3 Hz, H-3'), 5.16 (t, 1 H , J , = J 3 4  Hz, H-4), 5.10-5.00 (m, 2 H , H-2, H-2'), 4.36 (d, 1 H , J  4  u  3  4 5  2  9.60  7.8 Hz, H - l ) , 4.27 (m, 1 H , H -  5'), 4.23-3.98 (m, 4 H , H-6, H-6, H-6', H-6'), 3.95-3.82 (m, 2 H , H-3, C H ) , 3.61-3.54 2  (m, 1 H , H-5), 3.47-3.39 (m, 1 H , C H ) , 2.13 (s, 3 H , Ac), 2.09 (2xs, 6 H , Ac), 2.08 (2xs, 2  6 H , Ac), 2.05 (s, 3 H , Ac), 1.98 (s, 3 H , Ac), 1.60-1.54 (m, 2 H , C H ) , 1.34-1.24 (m, 10 2  H, 5x C H ) , 0.89 (t, 3 H , J 7.0 Hz, C H ) ; C - N M R (100MHz) 8 (CDC1 ): 171.1, 171.0, 13  2  3  C  3  258  170.5, 170.3, 169.8, 169.6, 169.1, 101.2, 96.5, 72.3, 72.1, 70.4, 70.3, 68.6, 68.4, 67.8, 67.1, 67.0, 62.3, 61.8, 32.1, 29.6, 29.5 (x2), 26.1, 22.9, 21.1-20.8 (x7), 14.8; H R ESIM S : m/z = 771.3049 [M + N a ] (expected for C 4H520, Na : m/z = 771.3051). +  +  3  8  /i-Octyl (2,3,4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(l->4)-2,3,6-tri-0-acetyI-P-Dgalactoside: 20.6 mg (54 %); ' H - N M R (400 MHz): 5 (CDC1 ): 5.57 (m, 1 H , H-4'), 5.39 (dd, 1 H , H  3  J ', ' H.O Hz, J . . 3.3 Hz, H-3'), 5.23-5.14 (m, 2 H , H-2, H-2'), 5.01 (d, 1 H , 1 ,T 3.6 3  2  3  v  j4  Hz, H - l ' ) , 4.82 (dd, 1 H , J  3>2  10.8 Hz, J , 2.7 Hz, H-3), 4.54 (m, 1 H , H-5'), 4.50-4.42 3  4  (m, 2 H , H-6, H - l ) , 4.22-4.04 (m, 4 H , H-6', H-6', H-6, H-4), 3.93-3.84 (m, 1 H , C H ) , 2  3.81-3.74 (m, 1 H , H-5), 3.52-3.3.42 (m, 1 H , C H ) , 2.13 (s, 3 H , Ac), 2.10 (s, 3 H , Ac), 2  2.08 (2xs, 6 H , Ac), 2.05 (2xs, 6 H , Ac), 1.99 (s, 3 H , Ac), 1.62-1.54 (m, 2 H , C H ) , 1.342  1.24 (m, 10 H , 5x C H ) , 0.88 (t, 3 H , J 7.0 Hz, C H ) ; C - N M R (100MHz) 5 (CDC1 ): 13  2  3  C  3  171.0, 170.8, 170.7 (x2), 170.4, 170.0, 169.3, 101.5, 99.6, 73.0, 72.0, 70.3 (x2), 69.0, 68.8, 68.1, 67.6, 67.3, 62.2, 60.7, 32.0, 29.6, 29.5, 29.5, 26.1, 22.9, 21.2-20.8 (x7), 14.3; H R ESI-MS: m/z = 771.3055 [M + Na] (expected for C H 20i8Na : m/z = 771.3051). +  +  34  5  «-Octyl (2,3,4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(l->3)-2,4-di-0-acetyl-P-Dxyloside: 27.8 mg (72%); ' H - N M R (400 MHz) 5 (CDC1 ): 5.45 (m, 1 H , H-4'), 5.43 (d, 1 H , h\r H  3  3.8 Hz, H - l ' ) , 5.26 (dd, 1 H , J >, - 10.9 Hz, J > < 3.3 Hz, H-3'), 5.07 (dd, 1 H , J , 3 . 8 Hz, 3  2  3  4  J ', < 10.9 Hz, H-2'), 4.94 (dd, 1 H , J ,i 5.4 Hz, J 2  3  2  2  2>3  r  7.0 Hz, H-2), 4.87-4.81 (m, 1 H , H4),  4.47 (d, 1 H , J i 5.4 Hz, H - l ) , 4.36 (m, 1 H , H-5'), 4.21-4.01 (m, 3 H , H-6', H-5, H-6'), ; 2  3.89 (t, 1 H , J  3>2  = J , 7.0 Hz, H-3), 3.82-3.74 (m, 1 H , C H ) , 3.44-3.37 (m, 1 H , C H ) , 3  4  2  2  3.33 (dd, 1 H , J , 7.1 Hz, J , 12.0 Hz, H-5), 2.13 (s, 3 H , Ac), 2.09 (2xs, 6 H , Ac), 2.07 5  4  5  5  259  (s, 3 H , Ac), 2.06 (s, 3 H , Ac), 1.99 (s, 3 H , Ac), 1.62-1.54 (m, 2 H , CH ), 1.35-1.23 (m, 2  10 H , 5x C H ) , 0.89 (t, 3 H , J 7.0 Hz, C H ) ; 2  13  3  C - N M R (100MHz) 6 (CDC1 ): 170.9, C  3  170.5, 170.4, 170.2, 169.9, 169.3, 100.2, 96.9, 75.2, 70.8, 70.7, 69.5, 68.3, 67.9, 67.4, 66.9, 61.3, 60.9, 32.0, 29.6, 29.5 (x2), 26.2, 22.9, 21.2, 21.0, 20.9 (x3), 20.8, 14.3; H R ESI-MS: m/z = 699.2848 [M + N a f (expected for C 3 i H 0 i 6 N a : m/z = 699.2840). +  48  (2,3,4,6-Tetra-0-acetyl-a-D-galactopyranosyl)-(l->2)-l,3,4-tri-0-acetyl-6-0benzoyl-a-D-mannose: 12.1 mg (84%); ' H - N M R (400 MHz): 8 (CDC1 ): 8.08 (m, 2H, Ar), 7.52 (m, 1 H , Ar), H  7.38 (m, 2 H , Ar), 6.13 (d, 1 H , J  u  3  2.1 Hz, H - l ) , 5.58 (t, 1 H , J , = J , 10.1 Hz, H-4), 4  3  4  5  5.43 (d, 1 H , J ., ' 3.2 Hz, H-4'), 5.36-5.30 (m, 3 H , H - l ' , H-3, H-3'), 5.08 (dd, 1 H , h\v 4  3  3.9 Hz, J ', ' 11.0 Hz, H-2'), 4.48-4.37 (m, 2 H , H-6, H-6), 4.32-4.28 (m, 1 H , H-5'), 2  3  4.16-4.12 (ddd, 1 H , H-5), 4.08 (t, 1 H , J ,i = J , 2.4 Hz, H-2), 4.05-4.00 (m, 2 H , H-6, 2  2  3  H-6), 2.15 (s, 3 H , Ac), 2.10 (s, 3 H , Ac), 2.06 (s, 3 H, Ac), 2.02 (s, 3 H, Ac), 2.00 (3xs, 9 H , 3xAc);  13  C - N M R (100 MHz) 8 (CDC1 ): 171.1, 170.6, 170.2. 170.0, 169.6, 169.3, C  3  168.1, 166.1, 133.0, 129.8 (x2), 129.7, 128.3 (x2), 97.0, 91.9, 72.9, 70.9, 70.7, 67.9, 67.8, 67.3 (x2), 65.9, 62.2, 61.9, 20.9-20.4 (x7); H R ESI-MS: m/z = 763.2063 [ M + N a f (expected for C33H oOi9Na : m/z = 763.2061). +  4  (2,3j4,6-Tetra-0-acetyl-a-D-galactopyranosyl)-(l->2)-l,3?-di-0-acetyl-4-0-benzoylct/P-D-xylose: 5.9 mg (75 %); H - N M R (600 MHz): 8 (CDC1 ): 8.01 (m, 4 H , Ar), 7.61 (m, 2 H , Ar), 1  H  3  7.47 (m, 4 H , Ar), 6.30 (d, 1 H , J , , 3.4 Hz, H - l ) , 5.71-5.66 (m, 2 H , H-lp, H-3 ), 5.54 2  a  a  (d, 1 H , h\r 3.8 Hz, H - l ' ) , 5.49-5.44 (m, 2 H , H - 3 , H4 '), 5.41(m, 1 H , H 4 ' ) , 5.25a  P  P  a  5.19 (m, 4.H, H 3 ' , H - l ' , H - 2 ' , H 3 ' ) 5.16 - 5.10 (m, 2 H , H - 4 , H-4 ), 5.05 (dd, 1 H , a  p  P  P  a  P  260  h-v 3.8 Hz, h\y  10.9 Hz, H-2 '), 4.31-4.18 (m, 4 H , H-5 ', H-5 , H - 5 ' , H-6 '), 4.16a  P  P  a  P  4.04 (m, 3 H , H - 5 , H - 6 ' , H-6 '), 4.03-3.99 (m, 1 H , H-6 '), 3.97 (dd, 1 H , J a  J  3s2  a  a  10.0 Hz, H-2 ), 3.91(t, 1 H , J a  P  2jl  H-5 ), 3.60 (dd, 1 H , J , 9.4 Hz, J a  5  4  =J 5 ; 5  2>1  3.6 Hz,  7.5 Hz, H-2 ) 3.80 (dd, 1 H , J , = J ,s 10.9 Hz,  2>3  P  5  4  5  11.6 Hz, H-5 ), 2.24, 2.17, 2.15, 2.14, 2.09, 2.07, P  2.06, 2.06, 2.05, 2.05, 2.00, 1.98 (12xs, 36 H , 12xAc);  13  C - N M R (100 MHz) 5 (CDC1 ): C  3  170.7, 170.2, 169.9, 169.7, 169.2, 169.1, 165.5, 133.7, 129.9 (2x), 128.9, 128.6 (2x), 96.1, 95.5, 94.0, 88.9, 74.1, 73.8, 71.8, 70.1, 69.7, 69.5, 68.0, 67.7, 67.5, 67.4 (x2), 67.2, 67.0, 66.8, 63.1, 61.1 (x2), 60.8, 21.0-20.6 (x6); H R ESI-MS: m/z = 691.1840 [M + N a ]  +  (expected for C H O i N a : m/z = 691.1850). +  3 0  3 6  7  7.5.4. a3GaiT-Catalysed Galactosylation of Alternative  Acceptor Substrates  Acceptor substrate (10 mM), U D P galactose (15 mM) and a3GalT (-0.5 mg/mL) were incubated for -16 hours at room temperature in 50 m M HEPES buffer (pH 7.5) containing 10 m M M n C l , 0.1% BSA, and 15 U of bovine alkaline phosphatase. Upon 2  completion, reactions were purified as described for LgtC. / M j r a - N i t r o p h e n y l (2,3»4,6-tetra-0-acetyl-a-D-galactopyranosyl)-(l->3)-2,4,6-tri-0acetyl-P-D-galactoside:  41.4 mg (82 %); ' H - N M R (400 MHz): 5 (CDC1 ): 8.21 (m, 2 H , Ar), 7.09 (m, 2 H , Ar), H  3  5.55 (dd, 1 H , J ,i 7.56 Hz, J , 9.98 Hz, H-2), 5.50-5.44 (m, 2 H , H-4', H-4), 5.32-5.27 2  2  3  (m, 2 H , H - l ' , H-2'), 5.22-5.16 (m, 1 H , H-3'), 5.13 (d, 1 H , J  U 2  7.56 Hz, H - l ) , 4.32-  4.26 (m, 1 H , H-5'), 4.26-4.02 (m, 6 H , H-6', H-6', H-6, H-6, H-5, H-3), 2.20 (s, 3 H , Ac), 2.16 (2xs, 6 H , Ac), 2.10 (s, 3 H , Ac), 2.06 (s, 3 H , Ac), 2.03 (s, 3 H , Ac), 1.98 (s, 3 H , Ac); C - N M R (100 MHz) 5 (CDC1 ): 170.5, 170.4 (x3), 169.9, 169.1, 161.4, 143.4, 1 3  C  3  261  129.2, 126.0 (x2), 116.8 (x2), 98.8, 93.9, 72.7, 71.7, 69.2, 67.8, 67.3, 67.1, 66.8, 65.0, 61.8, 61.4, 21.0-20.7 (x7);  H R ESI-MS: m/z = 780.1964 [ M + Na] (expected for +  C 2H39N0 oNa : m/z = 780.1963). +  3  2  7.5.5. Cst II-catalysed sialylation of alternative acceptor substrates Acceptor substrate (10 mM), C M P NeuAc (15 mM) and Cst II (-0.5 mg/mL) were incubated at room temperature in 100 m M HEPES buffer (pH 7.5) containing 10 mM MgCl . 2  Reactions were monitored by T L C (7: 2: 1: 0.1 EtOAc: MeOH: H 0 : 2  AcOH). When Gal was used as acceptor, only donor substrate hydrolysis and no product formation was detected, even after extensive incubation. When pNP Gal was used as acceptor, near quantitative conversion to product was observed following a 6 hour incubation.  This product was partially purified using a SepPak CigPlus Cartridge by  fractionated elution with water and characterized by H R ESI-MS (expected for C 2 H N O i 6 N a : m/z = 615.1650, m/z = 615.1648 [M + Na] found). Regio-selectivity +  3  32  +  2  was confirmed by treatment of 5 nmol of the reaction product with a selective a-2,3sialidase (New England Biolabs).  Under a unit defined time condition, complete  reversion of the product to pNP Gal was observed (Figure 7.1).  262  A  Figure 7.1.  B  C  T L C analysis of a-2,3 specific neuraminidase treatment of glycosylation  product derived from Cst II using pNP P-galactoside as an alternative acceptor substrate (7:2:1:0.1 EtOAc: MeOH: H 0 : AcOH). 2  (A) pNP p-galactoside starting material, (B)  sialylated product derived from Cst II, (C) sialylated product treated with specific neuraminidase.  7.5.6. Testing Alternative Cst II Donor Substrates F C H A S E lactose acceptor substrate (10 mg, 12.6 pmol), pNP a-D-sialoside (17 mg, 39 pmol), C M P (7 mg, 19 pmol) and Cst II (-0.5 mg/mL) were incubated at room temperature in 3 mL of 100 m M HEPES buffer (pH 7.5) containing 10 m M M g C l . 2  Reactions were monitored by T L C (7: 2: 1: 0.1 EtOAc: MeOH: H 0 : AcOH). 2  Upon  completion of the reaction (-36 hours), the product was partially purified using a SepPak CigPlus Cartridge by fractionated elution with water and those fractions containing the sialylated product were combined, lyophilized, resuspended in 2 mL of water and purified by preparative HPLC using an Amide-80 column as described in section 6.3.7.  The  H P L C purified product (10.1 mg, 72%) was characterized by H R ESI-MS (expected for C o H N 0 " : m/z = 1081.2937, m/z = 1081.2931 [M]" found). Regio-selectivity was 5  53  2  25  confirmed by treatment of 5 nmol of the reaction product with a selective ct-2,3-sialidase  263  (New England Biolabs). Under a unit defined time condition, complete reversion of the product to FCHASE-lactose was observed (Figure 7.2).  A  B  C  Figure 7.2. T L C analysis of a-2,3 specific neuraminidase treatment of glycosylation product derived from Cst II using /(-nitrophenyl a-sialoside as an alternative donor substrate in the presence of C M P monitored by T L C (7:2:1:0.1 EtOAc: MeOH: H 0 : 2  AcOH).  (A) Fluorescein lactose conjugate starting material, (B) sialylated product  derived from Cst II using alternative donor substrate, (C) sialylated product treated with specific neuraminidase.  7.5.7. Testing Alternative LgtC Donor Substrates F C H A S E lactose acceptor substrate (5 mg, 6.2 (imol), dNP P-D-galactoside (6.5 mg, 18.8 umol), UDP (4.0 mg, 8.9 ^mol) and LgtC (-0.5 mg/mL) were incubated at room temperature in 3 mL of 100 m M HEPES buffer (pH 7.5) containing 10 m M M n C l , 2  and 10 m M DTT. Reactions were monitored by T L C (7: 2: 1: 0.1 EtOAc: MeOH: H 0 : 2  AcOH).  Following a week long incubation at room temperature, although product  formation could be detected, the majority of the starting acceptor substrate remained and all of the dNP P-D-galactoside donor substrate surrogate had been consumed (presumably  264  due to spontaneous hydrolysis).  As such, yields were insufficient for isolation of  observable masses of product and characterization was limited to H R ESI-MS (expected for  C 5H48N 0,9 : +  4  2  m/z = 954.2668, m/z = 954.2674 [M] found). +  265  References Aharoni, A . , Amitai, G., Bernath, K., Magdassi, S., and Tawfik, D.S. (2005). Highthroughput screening of enzyme libraries: Thiolactonases evolved by fluorescenceactivated sorting of single cells in emulsion compartments. Chemistry & Biology 12, 1281-1289. Aharoni, A . , Thieme, A . A . K . , Chiu, C.P.C., Buchini, S., Lairson, L.L., Chen, H . M . , Strynadka, N.C.J., Wakarchuk, W.W., and Withers, S.G. (2006). High-throughput screening methodology for the directed evolution of glycosyltransferases. Nature Methods 3, 609-614. Alaimo, P.J., Shogren-Knaak, M . A . , and Shokat, K . M . (2001). Chemical genetic approaches for the elucidation of signaling pathways. Current Opinion in Chemical Biology 5, 360-367. Amaya, M.F., Buschiazzo, A., Nguyen, T., and Alzari, P.M. (2003). The high resolution structures of free and inhibitor-bound Trypanosoma rangeli sialidase and its comparison with T-cruzi trans-sialidase. Journal of Molecular Biology 325, 773-784. Atwell, S., and Wells, J.A. (1999). Selection for improved subtiligases by phage display. Proceedings of the National Academy of Sciences of the United States of America 96, 9497-9502. Baisch, G., Ohrlein, R., and Katopodis, A . (1996). Enzymatic fucosylations with purinediphosphate-fucoses (PDP-fucoses). Bioorganic & Medicinal Chemistry Letters 6, 29532956. Barbas III, C.F., Burton, D.R., Scott, J.K., and Silverman, G.J. (2001). Phage Display: A Laboratory Manual (New York, Cold Spring Harbor Press). Bencur, P., Steinkellner, H . , Svoboda, B., Mucha, J., Strasser, R., Kolarich, D., Harm, S., Kollensperger, G., Glossl, J., Altmann, F., et al. (2005). Arabidopsis thaliana beta 1,2xylosyltransferase: an unusual glycosyltransferase with the potential to act at multiple stages of the plant N-glycosylation pathway. Biochemical Journal 388, 515-525. Bendiak, B., and Schachter, H . (1987). Control of Glycoprotein-Synthesis - Kinetic Mechanism, Substrate-Specificity, and Inhibition Characteristics of Udp-NAcetylglucosamine-Alpha-D-Mannoside Beta-1 -2 N-Acetylglucosaminyltransferase-Ii from Rat-Liver. Journal of Biological Chemistry 262, 5784-5790.  266  Bernath, K., Hai, M.T., Mastrobattista, E., Griffiths, A.D., Magdassi, S., and Tawfik, D.S. (2004). In vitro compartmentalization by double emulsions: sorting and gene enrichment by fluorescence activated cell sorting. Analytical Biochemistry 325, 151-157. Berti, P.J., and Tanaka, K.S.E. (2002). Transition state analysis using multiple kinetic isotope effects: Mechanisms of enzymatic and non-enzymatic glycoside hydrolysis and transfer. In Advances in Physical Organic Chemistry, Vol 37, pp. 239-314. Bertrand, T., Briozzo, P., Assairi, L., Ofiteru, A . , Bucurenci, N . , Munier-Lehmann, H . , Golinelli-Pimpaneau, B., Barzu, O., and Gilles, A . M . (2002). Sugar specificity of bacterial C M P kinases as revealed by crystal structures and mutagenesis of Escherichia coli enzyme. Journal of Molecular Biology 315, 1099-1110. Blanco, G., Patallo, E.P., Brana, A.F., Trefzer, A . , Bechthold, A . , Rohr, J., Mendez, C , and Salas, J.A. (2001). Identification of a sugar flexible glycosyltransferase from Streptomyces olivaceus, the producer of the antitumor polyketide elloramycin. Chemistry & Biology 8, 253-263. Boix, E., Zhang, Y . N . , Swaminathan, G.J., Brew, K . , and Acharya, K . R . (2002). Structural basis of ordered binding of donor and acceptor substrates to the retaining glycosyltransferase, alpha-1,3-galactosyltransferase. Journal of Biological Chemistry 277, 28310-28318. Bolam, D . N . , Roberts, S., Proctor, M.R., Turkenburg, J.P., Dodson, E.J., MartinezFleites, C , Yang, M . , Davis, B.G., Davies, G.J., and Gilbert, H.J. (2007). The crystal structure of two macrolide glycosyltransferases provides a blueprint for host cell antibiotic immunity. Proceedings of the National Academy of Sciences of the United States of America 104, 5336-5341. Boozer, C.E., and Lewis, E.S. (1953). The Decomposition of Secondary Alkyl Chlorosulfites. II. Solvent Effects and Mechanisms. Journal of the American Chemical Society 75,3182-3186. Borisova, S.A., Zhao, L.S., Melancon, C.E., Kao, C.L., and Liu, H.W. (2004). Characterization of the glycosyltransferase activity of DesVII: Analysis of and implications for the biosynthesis of macrolide antibiotics. Journal of the American Chemical Society 126, 6534-6535. Breton, C , Bettler, E., Joziasse, D.H., Geremia, R.A., and Imberty, A . (1998). Sequencefunction relationships of prokaryotic and eukaryotic galactosyltransferases. Journal of Biochemistry 123, 1000-1009. Breton, C , and Imberty, A . (1999). Structure/function studies of glycosyltransferases. Current Opinion in Structural Biology 9, 563-571.  267  Brown, K . , Pompeo, F., Dixon, S., Mengin-Lecreulx, D., Cambillau, C , and Bourne, Y . (1999). Crystal structure of the bifunctional N-acetylglucosamine 1-phosphate uridyltransferase from Escherichia coli: a paradigm for the related pyrophosphorylase superfamily. Embo Journal 18, 4096-4107. Burmeister, W.P., Ruigrok, R.W.H., and Cusack, S. (1992). The 2.2-a Resolution CrystalStructure of Influenza-B Neuraminidase and Its Complex with Sialic-Acid. Embo Journal 77,49-56. Busch, C , Hofmann, F., Selzer, J., Munro, S., Jeckel, D., and Aktories, K . (1998). A common motif of eukaryotic glycosyltransferases is essential for the enzyme activity of large clostridial cytotoxins. Journal of Biological Chemistry 273, 19566-19572. Buschiazzo, A., Ugalde, J.E., Guerin, M.E., Shepard, W., Ugalde, R.A., and Alzari, P . M . (2004). Crystal structure of glycogen synthase: homologous enzymes catalyze glycogen synthesis and degradation. Embo Journal 23, 3196-3205. Cadwell, R.C., and Joyce, G.F. (1994). Mutagenic Per. Pcr-Methods and Applications 3, S136-S140. Campbell, J.A., Davies, G.J., Bulone, V . , and Henrissat, B . (1997). A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochemical Journal 326, 929-939. Campbell, R.E., Mosimann, S.C., Tanner, M . E . , and Strynadka, N.C.J. (2000). The structure of UDP-N-acetylglucosamine 2-epimerase reveals homology to phosphoglycosyl transferases. Biochemistry 39, 14993-15001. Carey, F.A., and Sundberg, R.J. (1990). Advanced Organic Chemistry. Part A : Structure and Mechanisms, Third edn (New York, Plenum Press). Cesaro-Tadic, S., Lagos, D., Honegger, A . , Rickard, J.H., Partridge, L.J., Blackburn, G . M . , and Pluckthun, A . (2003). Turnover-based in vitro selection and evolution of biocatalysts from a fully synthetic antibody library. Nature Biotechnology 27, 679-685. Chaiet, L.a.W., F.J. (1964). The Properties of Streptavidin, a Biotin-Binding Protein Produced by Streptomycetes. Archives of Biochemistry and Biophysics 106, 1-5. Charnock, S.J., and Davies, G.J. (1999). Structure of the nucleotide-diphospho-sugar transferase, SpsA from Bacillus subtilis, in native and nucleotide-complexed forms. Biochemistry 38, 6380-6385.  268  Chen, L., Men, FL, Ha, S., Ye, X . Y . , Brunner, L., Hu, Y., and Walker, S. (2002). Intrinsic lipid preferences and kinetic mechanism of Escherichia coli MurG. Biochemistry 41, 6824-6833. Chiu, C.P.C., Lairson, L . L . , Gilbert, M . , Wakarchuk, W.W., Withers, S.G., and Strynadka, N.C.J. (2007). Structural analysis of the alpha-2,3-sialyltransferase cst-i from Campylobacter jejuni in apo and substrate-analogue bound forms. Biochemistry 46, 71967204. Chiu, C.P.C., Watts, A . G . , Lairson, L.L., Gilbert, M . , Lim, D., Wakarchuk, W.W., Withers, S.G., and Strynadka, N.C.J. (2004). Structural analysis of the sialyltransferase Cstll from Campylobacter jejuni in complex with a substrate analog. Nature Structural & Molecular Biology 77, 163-170. Christianson, D.W. (2006). Structural biology and chemistry of the terpenoid cyclases. Chemical Reviews 106, 3412-3442. Cleland, W.W. (1963). The Kinetics of Enzyme-Catalyzed Reactions with Two or More Substrates or Products. I. Nomenclature and Rate Equations. Biochimica et Biophysica Acta 67, 104-137. Close, B.E., Mendiratta, S.S., Geiger, K . M . , Broom, L.J., Ho, L.L., and Colley, K . J . (2003). The minimal structural domains required for neural cell adhesion molecule polysialylation by PST/ST8Sia IV and STX/ST8Sia II. Journal of Biological Chemistry 278, 30796-30805. Cohen, H . M . , Tawfik, D.S., and Griffiths, A . D . (2004). Altering the sequence specificity of Haelll methyltransferase by directed evolution using in vitro compartmentalization. Protein Engineering Design & Selection 7 7, 3-11. Coia, J.E. (1998). Clinical, microbiological and epidemiological aspects of Escherichia coli 0157 infection. Ferns Immunology and Medical Microbiology 20, 1-9. Cornish-Bowden, A . (2004). Fundamentals of Enzyme Kinetics, Third edn (London, Portland Press). Coutinho, P . M . , Deleury, E., Davies, G.J., and Henrissat, B. (2003). A n evolving hierarchical family classification for glycosyltransferases. Journal of Molecular Biology 525,307-317. Cowdrey, W.A., Hughes, E.D., Ingold, C.K., Masterman, S., and Scott, A . D . (1937). Reaction Kinetics and the Walden Inversion. Part IV. Relation of Steric Orientation to  269  Mechanism in Substitutions Involving Halogen Atoms and Simple or Substituted Hydroxyl Groups. Journal of the Chemical Society, 1253-1271. Cram, D.J. (1953). Studies in Stereochemistry. X V I . Ionic Intermediates in the Decomposition of Certain Alkyl Chlorosulfites. Journal of the American Chemical Society 75, 332-338. Crennell, S., Takimoto, T., Portner, A . , and Taylor, G. (2000). Crystal structure of the multifunctional paramyxovirus hemagglutinin-neuraminidase. Nature Structural Biology 7, 1068-1074. Dall'Olio, F., and Chiricolo, M . (2001). Sialyltransferases in cancer. Glycoconjugate Journal 18, 841-850. Davies, G.J., Gloster, T.M., and Henrissat, B. (2005). Recent structural insights into the expanding world of carbohydrate-active enzymes. Current Opinion in Structural Biology 15, 637-645. Davies, G.S., M . L . ; Withers, S.G. (1998). Glycosyl Transfer. In Comprehensive Biological Catalysis, M . L . Sinnott, ed. (Academic Press), pp. 119-208. Demartis, S., Huber, A., Viti, F., Lozzi, L., Giovannoni, L., Neri, P., Winter, G., and Neri, D. (1999). A strategy for the isolation of catalytic activities from repertoires of enzymes displayed on phage. Journal of Molecular Biology 286, 617-633. Dennis, J., Waller, C , Timpl, R., and Schirrmacher, V . (1982). Surface Sialic-Acid Reduces Attachment of Metastatic Tumor-Cells to Collagen Type-Iv and Fibronectin. Nature 300, 274-276. Doi, N . , Kumadaki, S., Oishi, Y . , Matsumura, N . , and Yanagawa, H . (2004). In vitro selection of restriction endonucleases by in vitro compartmentalization. Nucleic Acids Research 32. Durbecq, V., Sainz, G., Oudjama, Y., Clantin, B., Bompard-Gilles, C , Tricot, C , Caillet, J., Stalon, V . , Droogmans, L., and Villeret, V . (2001). Crystal structure of isopentenyl diphosphate : dimethylallyl diphosphate isomerase. Embo Journal 20, 1530-1537. Durr, C , Hoffmeister, D., Wohlert, S.E., Ichinose, K . , Weber, M . , von Mulert, U . , Thorson, J.S., and Bechthold, A . (2004). The glycosyltransferase UrdGT2 catalyzes both C- and O-glycosidic sugar transfers. Angewandte Chemie-International Edition 43, 29622965.  270  Ekstrom, J.L., Pauly, T.A., Carty, M.D., Soeller, W . C , Culp, J., Danley, D.E., Hoover, D.J., Treadway, J.L., Gibbs, E . M . , Fletterick, R.J., et al. (2002). Structure-activity analysis of the purine binding site of human liver glycogen phosphorylase. Chemistry & Biology 9, 915-924. Endo, Y . , Tsurugi, K . , Yutsudo, T., Takeda, Y . , Ogasawara, T., and Igarashi, K . (1988). Site of Action of a Vero Toxin (Vt2) from Escherichia-Coli 0157-H7 and of Shiga Toxin on Eukaryotic Ribosomes - Rna N-Glycosidase Activity of the Toxins. European Journal of Biochemistry 777, 45-50. Endtz, H.P., Ang, C.W., van den Braak, N . , Duim, B., Rigter, A., Price, L.J., Woodward, D. L., Rodgers, F.G., Johnson, W . M . , Wagenaar, J.A., et al. (2000). Molecular characterization of Campylobacter jejuni from patients with Guillain-Barre and Miller Fisher syndromes. Journal of Clinical Microbiology 38, 2297-2301. Ernst, S., Langer, R., Cooney, C.L., and Sasisekharan, R. (1995). Enzymatic Degradation of Glycosaminoglycans. Critical Reviews in Biochemistry and Molecular Biology 30, 387-444. Fa, M . , Radeghieri, A . , Henry, A . A . , and Romesberg, F.E. (2004). Expanding the substrate repertoire of a D N A polymerase by directed evolution. Journal of the American Chemical Society 126, 1748-1754. Fenton, C , Hao, X . , Petersen, S.B., and El-Gewely, M.R. (2002). Random mutagenesis for protein breeding. In In Vitro Mutagenesis Protocols, J. Braman, ed. (Totowa, Humana Press), pp. 231-241. Fernandez-Gacio, A . , Uguen, M . , and Fastrez, J. (2003). Phage display as a tool for the directed evolution of enzymes. Trends in Biotechnology 27, 408-414. Fersht, A . (1999). Structure and Mechanism in Protein Science: A Guide to Enzyme Catalysis and Protein Folding (New York, W.H. Freeman and Company). Flint, J., Taylor, E., Yang, M . , Bolam, D.N., Tailford, L.E., Martinez-Fleites, C , Dodson, E. J., Davis, B.G., Gilbert, H.J., and Davies, G.J. (2005). Structural dissection and highthroughput screening of mannosylglycerate synthase. Nature Structural & Molecular Biology 72, 608-614. Fritz, T.A., Hurley, J.H., Trinh, L . B . , Shiloach, J., and Tabak, L . A . (2004). The beginnings of mucin biosynthesis: The crystal structure of UDP-GalNAc : polypeptide alpha-N-acetylgalactosaminyltransferase-Tl. Proceedings of the National Academy of Sciences of the United States of America 707, 15307-15312.  271  Fritz, T.A., Raman, J., and Tabak, L . A . (2006). Dynamic association between the catalytic and lectin domains of human UDP-GalNAc : polypeptide alpha-Nacetylgalactosaminyltransferase-2 (vol 281, pg 8613, 2006). Journal of Biological Chemistry 281, 25000-25000. Fromm, H.J. (1975). Initial Rate Enzyme Kinetics (New York, Springer-Verlag). Fuchs, G., Mobassaleh, M . , Donohuerolfe, A . , Montgomery, R . K . , Grand, R.J., and Keusch, G.T. (1986). Pathogenesis of Shigella Diarrhea - Rabbit Intestinal-Cell Microvillus Membrane-Binding Site for Shigella Toxin. Infection and Immunity 53, 372377. Gabelli, S.B., McLellan, J.S., Montalvetti, A . , Oldfield, E., Docampo, R., and Amzel, L . M . (2006). Structure and mechanism of the farnesyl diphosphate synthase from Trypanosoma cruza: Implications for drug design. Proteins-Structure Function and Bioinformatics 62, 80-88. Galili, U . , and LaTemple, D.C. (1998). Augmenting immunogenicity of autologous tumor vaccines by the natural anti-Gal antibody. Faseb Journal 12, A280-A280. Garinot-Schneider, C , Lellouch, A . C . , and Geremia, R . A . (2000). Identification of essential amino acid residues in the Sinorhizobium meliloti glucosyltransferase ExoM. Journal of Biological Chemistry 275, 31407-31413. Gastinel, L . N . , Bignon, C , Misra, A . K . , Hindsgaul, O., Shaper, J.H., and Joziasse, D . H . (2001). Bovine alpha 1,3-galactosyltransferase catalytic domain structure and its relationship with A B O histo-blood group and glycosphingolipid glycosyltransferases. Embo Journal 20, 638-649. Gastinel, L . N . , Cambillau, C , and Bourne, Y . (1999). Crystal structures of the bovine beta 4galactosyltransferase catalytic domain and its complex with uridine diphosphogalactose. Embo Journal 18, 3546-3557. Gebler, J . C , S.G. (1995). Identification Biochemistry  Trimbur, D.E., Warren, A.J., Aebersold, R., Namchuk, M . , and Withers, Substrate-Induced Inactivation of a Crippled Beta-Glucosidase Mutant of the Labeled Amino-Acid and Mutagenic Analysis of Its Role. 34, 14547-14553.  Geremia, S., Campagnolo, M . , Schinzel, R., and Johnson, L . N . (2002). Enzymatic catalysis in crystals of Escherichia coli maltodextrin phosphorylase. Journal of Molecular Biology 322, 413-423.  272  Ghadessy, F.J., Ong, J.L., and Holliger, P. (2001). Directed evolution of polymerase function by compartmentalized self-replication. Proceedings of the National Academy of Sciences of the United States of America 98, 4552-4557. Ghadessy, F.J., Ramsay, N . , Boudsocq, F., Loakes, D., Brown, A., Iwai, S., Vaisman, A . , Woodgate, R., and Holliger, P. (2004). Generic expansion of the substrate spectrum of a D N A polymerase by directed evolution. Nature Biotechnology 22, 755-759. Gibbons, B.J., Roach, P.J., and Hurley, T.D. (2002). Crystal structure of the autocatalytic initiator of glycogen biosynthesis, glycogenin. Journal of Molecular Biology 319, 463477. Gibson, R.P., Tarling, C.A., Roberts, S., Withers, S.G., and Davies, G.J. (2004). The donor subsite of trehalose-6-phosphate synthase - Binary complexes with UDP-glucose and UDP-2-deoxy-2-fluoro-glucose at 2 angstrom resolution. Journal of Biological Chemistry 279, 1950-1955. Gibson, R.P., Turkenburg, J.P., Charnock, S.J., Lloyd, R., and Davies, G.J. (2002). Insights into trehalose synthesis provided by the structure of the retaining glucosyltransferase OtsA. Chemistry & Biology 9, 1337-1346. Gilbert, M . , Bayer, R., Cunningham, A . M . , Defrees, S., Gao, Y . H . , Watson, D.C., Young, N . M . , and Wakarchuk, W.W. (1998). The synthesis of sialylated oligosaccharides using a CMP-Neu5Ac synthetase/sialyltransferase fusion. Nature Biotechnology 16, 769-772. Gilbert, M . , Brisson, J.R., Karwaski, M.F., Michniewicz, J., Cunningham, A . M . , Wu, Y . Y . , Young, N . M . , and Wakarchuk, W.W. (2000). Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384 - Identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-MHz H - l and C-13 N M R analysis. Journal of Biological Chemistry 275, 3896-3906. Gilbert, M . , Cunningham, A . M . , Watson, D.C., Martin, A . , Richards, J.C., and Wakarchuk, W.W. (1997). Characterization of a recombinant Neisseria meningitidis alpha-2,3-sialyltransferase and its acceptor specificity. European Journal of Biochemistry 249, 187-194. Gilbert, M . , Karwaski, M.F., Bernatchez, S., Young, N . M . , Taboada, E., Michniewicz, J., Cunningham, A . M . , and Wakarchuk, W.W. (2002). The genetic bases for the variation in the lipo-oligosaccharide of the mucosal pathogen, Campylobacter jejuni - Biosynthesis of sialylated ganglioside mimics in the core oligosaccharide. Journal of Biological Chemistry 277, 327-337. Gordon, R.D., Sivarajah, P., Satkunarajah, M . , Ma, D., Tarling, C.A., Vizitiu, D., Withers, S.G., and Rini, J . M . (2006). X-ray crystal structures of rabbit N -  273  acetylglucosaminyltransferase I (GnT I) in complex with donor substrate analogues. Journal of Molecular Biology 360, 67-79. Gosselin, S., Alhussaini, M . , Streiff, M.B., Takabayashi, K . , and Palcic, M . M . (1994). A Continuous Spectrophotometric Assay for Glycosyltransferases. Analytical Biochemistry 220, 92-97. Green, A . A., Cori, G.T., and Cori, C.F. (1942). C R Y S T A L L I N E P H O S P H O R Y L A S E . Journal of Biological Chemistry 142, 441-44%.  MUSCLE  Green, N . (1975). Avidin. Advances in Protein Chemistry 29, 85-133. Griffiths, A . D . , and Tawfik, D.S. (2003). Directed evolution of an extremely fast phosphotriesterase by in vitro compartmentalization. Embo Journal 22, 24-35. Grizot, S., Salem, M . , Vongsouthi, V . , Durand, L., Moreau, F., Dohi, H . , Vincent, S., Escaich, S., and Ducruix, A . (2006). Structure of the Escherichia coli heptosyltransferase WaaC: Binary complexes with A D P A N D ADP-2-deoxy-2-fluoro heptose. Journal of Molecular Biology 363, 383-394. Guerry, P., Szymanski, C M . , Prendergast, M . M . , Hickey, T.E., Ewing, C P . , Pattarini, D.L., and Moran, A.P. (2002). Phase variation of Campylobacter jejuni 81-176 lipooligosaccharide affects ganglioside mimicry and invasiveness in vitro. Infection and Immunity 70, 787-793. Guthrie, R.D., and Jencks, W.P. (1989). Iupac Recommendations for the Representation of Reaction-Mechanisms. Accounts of Chemical Research 22, 343-349. Guycaffey, J.K., Rapoza, M.P., Jolley, K . A . , and Webster, R.E. (1992). Membrane Localization and Topology of a Viral Assembly Protein. Journal of Bacteriology 174, 2460-2465. Ha, S., Walker, D., Shi, Y . G . , and Walker, S. (2000). The 1.9 angstrom crystal structure of Escherichia coli MurG, a membrane-associated glycosyltransferase involved in peptidoglycan biosynthesis. Protein Science 9,1045-1052. Hancock, S.M., D Vaughan, M . , and Withers, S.G. (2006). Engineering of glycosidases and glycosyltransferases. Current Opinion in Chemical Biology 10, 509-519. Harduin-Lepers, A., Vallejo-Ruiz, V., Krzewinski-Recchi, M . A . , Samyn-Petit, B., Julien, S., and Delannoy, P. (2001). The human sialyltransferase family. Biochimie 83,121-131.  274  Hekmat, O., Kim, Y.W., Williams, S.J., He, S.M., and Withers, S.G. (2005). Active-site peptide "fingerprinting" of glycosidases in complex mixtures by mass spectrometry Discovery of a novel retaining beta-l,4-glycanase in Cellulomonas fimi. Journal of Biological Chemistry 280, 35126-35135. Hidaka, M , Honda, Y . , Kitaoka, M , Nirasawa, S., Hayashi, K . , Wakagi, T., Shoun, H . , and Fushinobu, S. (2004). Chitobiose phosphorylase from vibrio Proteolyticus, a member of glycosyl transferase family 36, has a clan GH-L-like (alpha/alpha)(6) barrel fold. Structure 12, 937-947. Hoffmeister, D., Ichinose, K . , and Bechthold, A . (2001). Two sequence elements of glycosyltransferases involved in urdamycin biosynthesis are responsible for substrate specificity and enzymatic activity. Chemistry & Biology 8, 557-567. Hoffmeister, D., Wilkinson, B., Foster, G., Sidebottom, P.J., Ichinose, K., and Bechthold, A . (2002). Engineered urdamycin glycosyltransferases are broadened and altered in substrate specificity. Chemistry & Biology 9, 287-295. Holt, J.R., Gillespie, S.K.H., Provance, D.W., Shah, K . , Shokat, K . M . , Corey, D.P., Mercer, J.A., and Gillespie, P.G. (2002). A chemical-genetic strategy implicates myosinic in adaptation by hair cells. Cell 108, 371-381. Horcajada, C , Guinovart, J.J., Fita, I., and Ferrer, J.C. (2006). Crystal structure of an archaeal glycogen synthase - Insights into oligomerization and substrate binding of eukaryotic glycogen synthases. Journal of Biological Chemistry 281, 2923-2931. Horenstein, B . A . (1997). Quantum mechanical analysis of an alpha-carboxylatesubstituted oxocarbenium ion. Isotope effects for formation of the sialyl cation and the origin of an unusually large secondary C-14 isotope effect. Journal of the American Chemical Society 119, 1101-1107. Horenstein, B.A., and Bruner, M . (1998). The N-acetyl neuraminyl oxocarbenium ion is an intermediate in the presence of anionic nucleophiles. Journal of the American Chemical Society 120, 1357-1362. Hosfield, D.J., Zhang, Y . M . , Dougan, D.R., Broun, A . , Tari, L.W., Swanson, R.V., and Finn, J. (2004). Structural basis for bisphosphonate-mediated inhibition of isoprenoid biosynthesis. Journal of Biological Chemistry 279, 8526-8529. Hosseini-Maaf, B., Letts, J.A., Persson, M . , Smart, E., LePennec, P.Y., Hustinx, H . , Zhao, Z.H., Palcic, M . M . , Evans, S.V., Chester, M.A., et al. (2007). Structural basis for red cell phenotypic changes in newly identified, naturally occurring subgroup mutants of the human blood group B glycosyltransferase. Transfusion 47, 864-875.  275  Hu, Y . N . , Chen, L., Ha, S., Gross, B., Falcone, B., Walker, D., Mokhtarzadeh, M . , and Walker, S. (2003). Crystal structure of the MurG : UDP-GlcNAc complex reveals common structural principles of a superfamily of glycosyltransferases. Proceedings of the National Academy of Sciences of the United States of America 100, 845-849. Hu, Y . N . , and Walker, S. (2002). Remarkable structural similarities between diverse glycosyltransferases. Chemistry & Biology P, 1287-1296. Huang, X . , Tanaka, K.S.E., and Bennet, A.J. (1997). Glucosidase-Catalyzed Hydrolysis of &#x03Bl;-D-Glucopyranosyl Pyridinium Salts: Kinetic Evidence for Nucleophilic Involvement at the Glucosidation Transition State. Journal of the American Chemical Society 119, 11147-11154. Hughes, E.D., Ingold, C.K., and Whitfield, I.C. (1941). Walden Inversion in the replacement of hydroxl with halogen. Nature 147, 206-207. Hwang, Y.W., and Miller, D.L. (1987). A Mutation That Alters the Nucleotide Specificity of Elongation Factor-Tu, a Gtp Regulatory Protein. Journal of Biological Chemistry 262, 13081-13085. Iannolo, G., Minenkova, O., Petruzzelli, R., and Cesareni, G. (1995). Modifying Filamentous Phage Capsid - Limits in the Size of the Major Capsid Protein. Journal of Molecular Biology 248, 835-844. Ichikawa, Y . , Liu, J.L.C., Shen, G.J., and Wong, C H . (1991). A Highly Efficient Multienzyme System for the One-Step Synthesis of a Sialyl Trisaccharide - Insitu Generation of Sialic-Acid and N-Acetyllactosamine Coupled with Regeneration of UdpGlucose, Udp-Galactose, and Cmp-Sialic Acid. Journal of the American Chemical Society 113, 6300-6302. Ihara, H., Ikeda, Y . , Toma, S., Wang, X . C , Suzuki, T., Gu, J.G., Miyoshi, E., Tsukihara, T., Honke, K . , Matsumoto, A., et al. (2007). Crystal structure of mammalian alpha 1,6fucosyltransferase, FUT8. Glycobiology 17,455-466. Jamaluddin, H . , Tumbale, P., Withers, S.G., Acharya, K.R., and Brew, K . (2007). Conformational Changes Induced by Binding UDP-2F-galactose to [alpha]-1,3 Galactosyltransferase- Implications for Catalysis. Journal of Molecular Biology 369, 1270-1281. Jank, T., Reinert, D.J., Giesemann, T., Schulz, G.E., and Aktories, K . (2005). Change of the donor substrate specificity of Clostridium difficile toxin B by site-directed mutagenesis. Journal of Biological Chemistry 280, 37833-37838.  276  Jeanneau, C , Chazalet, V . , Auge, C , Soumpasis, D . M . , Harduin-Lepers, A . , Delannoy, P., Imberty, A . , and Breton, C. (2004). Structure-function analysis of the human sialyltransferase ST3Gal I - Role of N-glycosylation and a novel conserved sialylmotif. Journal of Biological Chemistry 279, 13461-13468. Jencks, W.P. (1980). When is an intermediate not an intermediate? Enforced mechanisms of general acid-base, catalyzed, carbocation, carbanion, and ligand exchange reaction, pp. 161-169. Jencks, W.P. (1996). Quantitive approaches to reaction mechanisms and catalysis. Substitution at carbon. Journal of Physical Organic Chemistry 9, 337-340. Jestin, J.L., Kristensen, P., and Winter, G. (1999). A method for the selection of catalytic activity using phage display and proximity coupling. Angewandte Chemie-International Edition 38, 1124-1127. Jinek, M . , Chen, Y.W., Clausen, FL, Cohen, S.M., and Conti, E. (2006). Structural insights into the Notch-modifying glycosyltransferase Fringe. Nature Structural & Molecular Biology 13, 945-946. Just, I., Selzer, J., Wilm, M . , Voneichelstreiber, C , Mann, M . , and Aktories, K . (1995). Glucosylation of Rho-Proteins by Clostridium-Difficile Toxin-B. Nature 375, 500-503. Kakuda, S., Shiba, T., Ishiguro, M . , Tagawa, FL, Oka, S., Kajihara, Y . , Kawasaki, T., Wakatsuki, S., and Kato, R. (2004). Structural basis for acceptor substrate recognition of a human glucuronyltransferase, GlcAT-P, an enzyme critical in the biosynthesis of the carbohydrate epitope H N K - 1 . Journal of Biological Chemistry 279, 22693-22703. Kamath, V.P., Seto, N.O.L., Compston, C.A., Hindsgaul, O., and Palcic, M . M . (1999). Synthesis of the acceptor analog alpha Fuc(l -> 2)alpha Gal-0(CH2)(7) CH3: A probe for the kinetic mechanism of recombinant human blood group B glycosyltransferase. Glycoconjugate Journal 16, 599-606. Kearns, A.E., Campbell, S . C , Westley, J., and Schwartz, N . B . (1991). Initiation of Chondroitin Sulfate Biosynthesis - a Kinetic-Analysis of Udp-D-Xylose-Core Protein Beta-D-Xylosyltransferase. Biochemistry 30, 7477-7483. Kempton, J.B., and Withers, S.G. (1992). Mechanism of Agrobacterium BetaGlucosidase - Kinetic-Studies. Biochemistry 31, 9961-9969. Khaled, A., Ivannikova, T., and Auge, C. (2004). Synthesis of unnatural sugar nucleotides and their evaluation as donor substrates in glycosyltransferase-catalyzed reactions. Carbohydrate Research 339, 2641 -2649.  277  Kikuchi, N . , Kwon, Y . D . , Gotoh, M . , and Narimatsu, H . (2003). Comparison of glycosyltransferase families using the profile hidden Markov model. Biochemical and Biophysical Research Communications 310, 574-579. Kitov, P.I., Sadowska, J.M., Mulvey, G., Armstrong, G.D., Ling, H . , Pannu, N.S., Read, R.J., and Bundle, D.R. (2000). Shiga-like toxins are neutralized by tailored multivalent carbohydrate ligands. Nature 403, 669-672. Klabunde, T., Wendt, K . U . , Kadereit, D., Brachvogel, V . , Burger, H.J., Herling, A.W., Oikonomakos, N.G., Kosmopoulou, M . N . , Schmoll, D., Sarubbi, E., et al. (2005). Acyl ureas as human liver glycogen phosphorylase inhibitors for the treatment of type 2 diabetes. Journal of Medicinal Chemistry 48, 6178-6193. Klein, H.W., Im, M.J., and Palm, D. (1986). Mechanism of the Phosphorylase Reaction Utilization of D-Gluco-Hept-l-Enitol in the Absence of Primer. European Journal of Biochemistry 757, 107-114. Knegtel, R.M.K., Strokopytov, B., Penninga, D., Faber, O.G., Rozeboom, H.J., Kalk, K . H . , Dijkhuizen, L., and Dijkstra, B.W. (1995). Crystallographic Studies of the Interaction of Cyclodextrin Glycosyltransferase from Bacillus-Circulans Strain-251 with Natural Substrates and Products. Journal of Biological Chemistry 270, 29256-29264. Kopp, M . , Rupprath, C , Irschik, FL, Bechthold, A . , Elling, L . , and Muller, R. (2007). SorF: A glycosyltransferase with promiscuous donor substrate specificity in vitro. Chembiochem 8, 813-819. Kornberg, S.R., Zimmerman, S.B., and Kornberg, A . (1961). Glucosylation of Deoxyribonucleic Acid by Enzymes from Bacteriophage-infected Escherichia coli. Journal of Biological Chemistry 236, 1487-1493. Koshland, D.E. (1953). Stereochemistry and the mechanism of enzymatic reactions. Biological Reviews of the Cambridge Philosophical Society 28, 416-436. Kostrewa, D., D'Arcy, A . , Takacs, B., and Kamber, N . (2001). Crystal structures of Streptococcus pneumoniae N-acetylglucosamine-1-phosphate uridyltransferase, GlmU, in apo form at 2.33 A resolution and in complex with UDP-N-acetylglucosamine and Mg2+ at 1.96 A resolution. Journal of Molecular Biology 305, 279-289. Kozmon, S., and Tvaroska, I. (2006). Catalytic mechanism of glycosyltransferases: Hybrid quantum mechanical/molecular mechanical study of the inverting N acetylglucosaminyltransferase I. Journal of the American Chemical Society 128, 1692116927.  278  Kreppel, L.K., and Hart, G.W. (1999). Regulation of a cytosolic and nuclear O-GlcNAc transferase - Role of the tetratricopeptide repeats. Journal of Biological Chemistry 274, 32015-32022. Kretzschmar, T., and Geiser, M . (1995). Evaluation of Antibodies Fused to Minor Coat Protein-Iii and Major Coat Protein-Viii of Bacteriophage M13. Gene 155, 61-65. Kubo, A., Arai, Y., Nagashima, S., and Yoshikawa, T. (2004). Alteration of sugar donor specificities of plant glycosyltransferases by a single point mutation. Archives of Biochemistry and Biophysics 429, 198-203. Kubota, T., Shiba, T., Sugioka, S., Furukawa, S., Sawaki, H . , Kato, R., Wakatsuki, S., and Narimatsu, H . (2006). Structural basis of carbohydrate transfer activity by human UDP-GaTNAc: Polypeptide alpha-N-acetylgalactosaminyltransferase (pp-GaINAc-T10). Journal of Molecular Biology 359, 708-727. Laine, R . A . (1994). A Calculation of A l l Possible Oligosaccharide Isomers Both Branched and Linear Yields 1.05x10(12) Structures for a Reducing Hexasaccharide - the Isomer-Barrier to Development of Single-Method Saccharide Sequencing or Synthesis Systems. Glycobiology 4, 759-767. Lariviere, L., Gueguen-Chaignon, V., and Morera, S. (2003). Crystal structures of the T4 phage beta-glucosyltransferase and the D100A mutant in complex with UDP-glucose: Glucose binding and identification of the catalytic base for a direct displacement mechanism. Journal of Molecular Biology 330, 1077-1086. Lariviere, L . , and Morera, S. (2002). A base-flipping mechanism for the T4 phage betaglucosyltransferase and identification of a transition-state analog. Journal of Molecular Biology 324, 483-490. Lariviere, L., Sommer, N . , and Morera, S. (2005). Structural evidence of a passive baseflipping mechanism for A G T , an unusual GT-B glycosyltransferase. Journal of Molecular Biology 352, 139-150. Lee, H.J., Barry, C.H., Borisova, S.N., Seto, N.O.L., Zheng, R.X.B., Blancher, A., Evans, S.V., and Palcic, M . M . (2005). Structural basis for the inactivity of human blood group 0-2 glycosyltransferase. Journal of Biological Chemistry 280, 525-529. Lee, S.S., Y u , S.K., and Withers, S.G. (2002a). alpha-1,4-glucan lyase performs a transelimination via a nucleophilic displacement followed by a syn-elimination. Journal of the American Chemical Society 124, 4948-4949.  279  Lee, Y.F., Tawfik, D.S., and Griffiths, A . D . (2002b). Investigating the target recognition of D N A cytosine-5 methyltransferase Hhal by library selection using in vitro compartmentalisation. Nucleic Acids Research 30, 4937-4944. Legler, P.M., Massiah, M . A . , and Mildvan, A.S. (2002). Mutational, kinetic, and N M R studies of the mechanism of E. coli GDP-mannose mannosyl hydrolase, an unusual nudix enzyme. Biochemistry 41, 10834-10848. Letts, J.A., Rose, N.L., Fang, Y.R., Barry, C.H., Borisova, S.N., Seto, N.O.L., Palcic, M . M . , and Evans, S.V. (2006). Differential recognition of the type I and IIH antigen acceptors by the human ABO(H) blood group A and B glycosyltransferases. Journal of Biological Chemistry 281, 3625-3632. Leung, D.A., Chen, E., and Goeddel, D.V. (1989). A method for random mutagenesis of defined D N A segments using a modified polymerase chain reaction. . Technique i , 1115. Lewis, E.S., and Boozer, C E . (1952). The Kinetics and Stereochemistry of the Decomposition of Secondary Alkyl Chlorosulfites. Journal of the American Chemical Society 74, 308-311. Lin-Goerke, J.L., Robbins, D.J., and Burczak, J.D. (1997). PCR-based random mutagenesis using manganese and reduced dNTP concentration. Biotechniques 23, 409412. Lin, K . , Rath, V . L . , Dai, S . C , Fletterick, R.J., and Hwang, P.K. (1996). A protein phosphorylation switch at the conserved allosteric site in GP. Science 273, 1539-1541. Liu, J., and Mushegian, A . (2003). Three monophyletic superfamilies account for the majority of the known glycosyltransferases. Protein Science 12, 1418-1431. Lobsanov, Y . D . , Romero, P.A., Sleno, B., Yu, B . M . , Yip, P., Herscovics, A . , and Howell, P.L. (2004). Structure of Kre2p/Mntlp - A yeast alpha 1,2-mannosyltransferase involved in mannoprotein biosynthesis. Journal of Biological Chemistry 279, 17921-17931. Lougheed, B., Ly, H.D., Wakarchuk, W.W., and Withers, S.G. (1999). Glycosyl fluorides can function as substrates for nucleotide phosphosugar-dependent glycosyltransferases. Journal of Biological Chemistry 274, 37717-37722. Love, K . R . , Swoboda, J.G., Noren, C.J., and Walker, S. (2006). Enabling glycosyltransferase evolution: A facile substrate-attachment strategy for phage-display enzyme evolution. Chembiochem 7, 753-756.  280  Lovering, A . L . , de Castro, L . H . , Lim, D., and Strynadka, N.C.J. (2007). Structural insight into the transglycosylation step of bacterial cell-wall biosynthesis. Science 315, 14021405. Lukacs, C M . , Oikonomakos, N.G., Crowther, R.L., Hong, L . N . , Kammlott, R.U., Levin, W., L i , S., Liu, C M . , Lucas-McGady, D., Pietranico, S., et al. (2006). The crystal structure of human muscle glycogen phosphorylase a with bound glucose and A M P : A n intermediate conformation with T-State and R-State features. Proteins-Structure Function and Bioinformatics 63, 1123-1126. Luo, Y . , L i , S . C , Chou, M . Y . , L i , Y.T., and Luo, M . (1998). The crystal structure of an intramolecular trans-sialidase with a NeuAc alpha 2 -> 3Gal specificity. Structure 6, 521530. Ly, H.D., Lougheed, B., Wakarchuk, W.W., and Withers, S.G. (2002). Mechanistic studies of a retaining alpha-galactosyltransferase from Neisseria meningitidis. Biochemistry 41, 5075-5085. Ly, H.D., and Withers, S.G. (1999). Mutagenesis of glycosidases. Annual Review of Biochemistry 68, 487-522. Mackenzie, L.F., Wang, Q.P., Warren, R.A.J., and Withers, S.G. (1998). Glycosynthases: Mutant glycosidases for oligosaccharide synthesis. Journal of the American Chemical Society 120, 5583-5584. Macleod, A . M . , Lindhorst, T., Withers, S.G., and Warren, R.A.J. (1994). The Acid/Base Catalyst in the Exoglucanase/Xylanase from Cellulomonas-Fimi Is Glutamic-Acid-127 Evidence from Detailed Kinetic-Studies of Mutants. Biochemistry 33, 6371-6376. MacManus, D.A., and Vulfson, E.N. (2000). Regioselectivity of enzymatic glycosylation of 6-O-acyl glycosides in supersaturated solutions. Biotechnology and Bioengineering 69, 585-590. Madsen, N . B . (1986). The Enzymes 17, 365-394. Maeda, Y . , Watanabe, R., Harris, C.L., Hong, J.J., Ohishi, K . , Kinoshita, K . , and Kinoshita, T. (2001). PIG-M transfers the first mannose to glycosylphosphatidylinositol on the lumenal side of the ER. Embo Journal 20, 250-261. Malik, P., Tarry, T.D., Gowda, L.R., Langara, A . , Petukhov, S.A., Symmons, M.F., Welsh, L . C , Marvin, D.A., and Perham, R . N . (1996). Role of capsid structure and membrane protein processing in determining the size and copy number of peptides  281  displayed on the major coat protein of filamentous bacteriophage. Journal of Molecular Biology 260, 9-21. Mark, B . L . , and James, M.N.G. (2002). Anchimeric assistance in hexosaminidases. Canadian Journal of Chemistry-Revue Canadienne De Chimie 80, 1064-1074. Mark, B.L., Vocadlo, D.J., Knapp, S., Triggs-Raine, B.L., Withers, S.G., and James, M . N . G . (2001). Crystallographic evidence for substrate-assisted catalysis in a bacterial beta-hexosaminidase. Journal of Biological Chemistry 276, 10330-10337. Martinez-Fleites, C , Proctor, M . , Roberts, S., Bolam, D.N., Gilbert, H.J., and Davies, G.J. (2006). Insights into the synthesis of lipopolysaccharide and antibiotics through the structures of two retaining glycosyltransferases from family GT4. Chemistry & Biology 13, 1143-1152. Mazzella, L.J., Parkin, D.W., Tyler, P.C., Furneaux, R.H., and Schramm, V . L . (1996). Mechanistic diagnoses of N-ribohydrolases and purine nucleoside phosphorylase. Journal of the American Chemical Society 118, 2111-2112. Miley, M.J., Zielinska, A . K . , Keenan, J.E., Bratton, S.M., Radominska-Pandya, A . , and Redinbo, M.R. (2007). Crystal Structure of the Cofactor-Binding Domain of the Human Phase II Drug-Metabolism Enzyme UDP-Glucuronosyltransferase 2B7. Journal of Molecular Biology 369, 498-511. Mitchell, E.P., Withers, S.G., Ermert, P., Vasella, A.T., Garman, E.F., Oikonomakos, N.G., and Johnson, L . N . (1996). Ternary complex crystal structures of glycogen phosphorylase with the transition state analogue nojirimycin tetrazole and phosphate in the T and R states. Biochemistry 35, 7341-7355. Miura, Y . , Arai, T., and Yamagata, T. (1996). Synthesis of amphiphilic lactosides that possess a lactosylceramide-mimicking N-acyl structure: Alternative universal substrates for endo-type glycosylceramidases. Carbohydrate Research 289, 193-199. Moffatt, J.G. (1966). Sugar nucleotide synthesis by the phosphomorpholidate procedure. Methods in Enzymology 8, 136-142. Moffatt, J.G., and Khorana, H.G. (1961). Nucleoside Polyphosphates. X . The synthesis and some reactions of nucleoside-5' phosphoromorpholidates and related compounds. Improved methods for the preparation of nucleoside-5'polyphosphates. Journal of the American Chemical Society 83, 649-658. Mollerach, M . , Lopez, R., and Garcia, E. (1998). Characterization of the galU gene of Streptococcus pneumoniae encoding a uridine diphosphoglucose pyrophosphorylase: a  282  gene essential for capsular polysaccharide biosynthesis. Journal of Experimental Medicine 188, 2047-2056. Monegal, A . , and Planas, A . (2006). Chemical rescue of alpha 3-galactosyltransferase. Implications in the mechanism of retaining glycosyltransferases. Journal of the American Chemical Society 128, 16030-16031. Morera, S., Lariviere, L., Kurzeck, J., Aschke-Sonnenborn, U., Freemont, P.S., Janin, J., and Ruger, W. (2001). High resolution crystal structures of T4 phage betaglucosyltransferase: Induced fit and effect of substrate and metal binding. Journal of Molecular Biology 311, 569-577. Morrison, J.F., and Ebner, K . E . (1971). Studies on Galactosyltransferase - Kinetic Investigations with N-Acetylglucosamine as Galactosyl Group Acceptor. Journal of Biological Chemistry 246, 3977-&. Mosi, R., He, S.M., Uitdehaag, J., Dijkstra, B.W., and Withers, S.G. (1997). Trapping and characterization of the reaction intermediate in cyclodextrin glycosyltransferase by use of activated substrates and a mutant enzyme. Biochemistry 36, 9927-9934. Moss, R . A . , Fu, X . L . , and Sauers, R.R. (2007). S(N)i fragmentations of alkoxychlorocarbenes - a perspective. Journal of Physical Organic Chemistry 20, 1-10. Muehlbacher, M . , and Poulter, C D . (1985). Isopentenyl Diphosphate Dimethylallyl Diphosphate Isomerase - Irreversible Inhibition of the Enzyme by Active-Site-Directed Covalent Attachment. Journal of the American Chemical Society 107, 8307-8308. Muehlbacher, M . , and Poulter, C D . (1988). Isopentenyl-Diphosphate Isomerase Inactivation of the Enzyme with Active-Site-Directed Irreversible Inhibitors and Transition-State Analogs. Biochemistry 27, 7315-7328. Mulichak, A . M . , Losey, H . C , Lu, W., Wawrzak, Z., Walsh, C.T., and Garavito, R . M . (2003). Structure of the TDP-epi-vancosaminyltransferase GtfA from the chloroeremomycin biosynthetic pathway. Proceedings of the National Academy of Sciences of the United States of America 100, 9238-9243. Mulichak, A . M . , Losey, H . C , Walsh, C.T., and Garavito, R . M . (2001). Structure of the UDP-Glucosyltransferase GtfB that modifies the heptapeptide aglycone in the biosynthesis of vancomycin group antibiotics. Structure 9, 547-557. Mulichak, A . M . , Lu, W., Losey, H . C , Walsh, C.T., and Garavito, R . M . (2004). Crystal structure of vancosaminyltransferase GtfD from the vancomycin biosynthetic pathway: Interactions with acceptor and nucleotide ligands. Biochemistry 43, 5170-5180.  283  Murray, B.W., Takayama, S., Schultz, J., and Wong, C H . (1996). Mechanism and specificity of human alpha-l,3-fucosyltransferase V . Biochemistry 35, 11183-11195. Murray, B.W., Wittmann, V . , Burkart, M . D . , Hung, S . C , and Wong, C H . (1997). Mechanism of human alpha-1,3-fucosyltransferase V : Glycosidic cleavage occurs prior to nucleophilic attack. Biochemistry 36, 823-831. Nerinckx, W., Desmet, T., and Claeyssens, M . (2003). A hydrophobic platform as a mechanistically relevant transition state stabilising factor appears to be present in the active centre of all glycoside hydrolases. Febs Letters 538, 1-7. Nguyen, H.P., Seto, N.O.L., Cai, Y . , Leinala, E.K., Borisova, S.N., Palcic, M . M . , and Evans, S.V. (2003). The influence of an intramolecular hydrogen bond in differential recognition of inhibitory acceptor analogs by human ABO(H) blood group A and B glycosyltransferases. Journal of Biological Chemistry 278, 49191-49195. N i , L., Chokhawala, H.A., Cao, H., Henning, R., Ng, L., Huang, S., Yu, H., Chen, X . , and Fisher, A . J . (2007). Crystal Structures of Pasteurella multocida Sialyltransferase Complexes with Acceptor and Donor Analogues Reveal Substrate Binding Sites and Catalytic Mechanism. Biochemistry 46, 6288-6298. N i , L.S., Sung, M . C , Y u , H . , Chokhawala, H . , Chen, X . , and Fisher, A . J . (2006). Cytidine 5 '-monophosphate (CMP)-induced structural changes in a multifunctional sialyltransferase from Pasteurella multocida. Biochemistry 45, 2139-2148. Nishikawa, Y . , Pegg, W., Paulsen, H., and Schachter, H . (1988). Control of GlycoproteinSynthesis .15. Purification and Characterization of Rabbit Liver Udp-NAcetylglucosamine-Alpha-3 -D-Mannoside Beta-1,2-N-Acetylglucosaminyltransferase-I. Journal of Biological Chemistry 263, 8270-8281. O'Reilly, M . , Watson, K . A . , and Johnson, L . N . (1999). The crystal structure of the Escherichia coli maltodextrin phosphorylase-acarbose complex. Biochemistry 38, 53375345. O'Reilly, M . , Watson, K . A . , Schinzel, R., Palm, D., and Johnson, L . N . (1997). Oligosaccharide substrate binding in Escherichia coli maltodextrin phosphorylase. Nature Structural Biology 4, 405-412. Oberthur, M . , Leimkuhler, C , Kruger, R.G., Lu, W., Walsh, C.T., and Kahne, D. (2005). A systematic investigation of the synthetic utility of glycopeptide glycosyltransferases. Journal of the American Chemical Society 127, 10747-10752.  284  Offen, W., Martinez-Fleites, C , Yang, M . , Kiat-Lim, E., Davis, B.G., Tarling, C.A., Ford, C M . , Bowles, D.J., and Davies, G.J. (2006). Structure of a flavonoid glucosyltransferase reveals the basis for plant natural product modification. Embo Journal 25,1396-1405. Ogata, S., Ho, I., Chen, A . L . , Dubois, D., Maklansky, J., Singhal, A., Hakomori, S., and Itzkowitz, S.H. (1995). Tumor-Associated Sialylated Antigens Are Constitutively Expressed in Normal Human Colonic Mucosa. Cancer Research 55, 1869-1874. Ohrlein, R. (1999). Glycosyltranferase-catalyzed synthesis of non-natural oligosaccharides. In Biocatalysis - from Discovery to Application, pp. 227-254. Ohrui, H., Meguro, H., Ioa, T. (1988). N-acetyl-3-fluoro-neuraminic acid derivatives and preparation thereof, E. Patent, ed. Ohtsubo, K . , Imajo, S., Ishiguro, M . , Nakatani, T., Oka, S., and Kawasaki, T. (2000). Studies on the structure-function relationship of the HNK-1 associated glucuronyltransferase, GlcAT-P, by computer modeling and site-directed mutagenesis. Journal of Biochemistry 128, 283-291. Ouzzine, M . , Gulberti, S., Levoin, N . , Netter, P., Magdalou, J., and Fournel-Gigleux, S. (2002). The donor substrate specificity beta 1,3-glucuronosyltransferase I toward U D P glucoronic acid is determined by two crucial histidine and arginine residues. Journal of Biological Chemistry 277, 25439-25445. Pak, J.E., Arnoux, P., Zhou, S.H., Sivarajah, P., Satkunarajah, M . , Xing, X . K . , and Rini, J.M. (2006). X-ray crystal structure of leukocyte type core 2 beta 1,6-Nacetylglucosaminyltransferase - Evidence for a convergence of metal ion-independent glycosyltransferase mechanism. Journal of Biological Chemistry 281, 26693-26701. Patenaude, S.L, Seto, N.O.L., Borisova, S.N., Szpacenko, A., Marcus, S.L., Palcic, M . M . , and Evans, S.V. (2002). The structural basis for specificity in human ABO(H) blood group biosynthesis. Nature Structural Biology 9, 685-690. Pedersen, H . , Holder, S., Sutherlin, D.P., Schwitter, U . , King, D.S., and Schultz, P.G. (1998). A method for directed evolution and functional cloning of enzymes. Proceedings of the National Academy of Sciences of the United States of America 95, 10523-10528. Pedersen, L . C , Darden, T.A., and Negishi, M . (2002). Crystal structure of beta 1,3glucuronyltransferase I in complex with active donor substrate UDP-GlcUA. Journal of Biological Chemistry 277, 21869-21873.  285  Pedersen, L . C , Dong, J., Taniguchi, F., Kitagawa, H . , Krahn, J . M . , Pedersen, L . G . , Sugahara, K . , and Negishi, M . (2003). Crystal structure of an alpha 1,4-Nacetylhexosaminyltransferase (EXTL2), a member of the exostosin gene family involved in heparan sulfate biosynthesis. Journal of Biological Chemistry 278, 14420-14428. Pedersen, L.C., Tsuchida, K., Kitagawa, H., Sugahara, K., Darden, T.A., and Negishi, M . (2000). Heparan/chondroitin sulfate biosynthesis - Structure and mechanism of human glucuronyltransferase I. Journal of Biological Chemistry 275, 34580-34585. Penner, J.L., and Aspinall, G.O. (1997). Diversity of lipopolysaccharide structures in Campylobacter jejuni. Journal of Infectious Diseases 176, S135-S138. Persson, K . , Ly, H.D., Dieckelmann, M . , Wakarchuk, W.W., Withers, S.G., and Strynadka, N.C.J. (2001). Crystal structure of the retaining galactosyltransferase LgtC from Neisseria meningitidis in complex with donor and acceptor sugar analogs. Nature Structural Biology 8, 166-175. Persson, M . , Letts, J.A., Hosseini-Maaf, B., Borisova, S.N., Palcic, M . M . , Evans, S.V., and Olsson, M . L . (2007). Structural effects of naturally occurring human blood group B galactosyltransferase mutations adjacent to the D X D motif. Journal of Biological Chemistry 282, 9564-9570. Phillips, D.C. (1966). The three-dimensional structure of an enzyme molecule. Scienctific American 215, 78-90. Phillips, D.C. (1967). The Hen Egg White Lysozyme. Proceedings of the National Academy of Sciences of the United States of America 57, 484-495. Piszkiewicz, D. (1977). Kinetics of Chemical and Enzyme-Catalyzed Reactions (New York, Oxford University Press). Poulter, C D . , Argyle, J . C , and Mash, E.A. (1977). Prenyltransferase - New Evidence for an "Ionization-Condensation-Elimination Mechanism with 2-Fluorogeranyl Pyrophosphate. Journal of the American Chemical Society 99, 957-959. Poulter, C D . , Argyle, J . C , and Mash, E.A. (1978). Farnesyl Pyrophosphate Synthetase Mechanistic Studies of 1 '4 Coupling Reaction with 2-Fluorogeranyl Pyrophosphate. Journal of Biological Chemistry 253, 7227-7233. Poulter, C D . , and Satterwhite, D . M . (1977). Mechanism of Prenyl-Transfer Reaction Studies with (E)-3-Trifluoromethyl-2-Buten-l-Yl and (Z)-3-Trifluoromethyl-2-Buten-1Y l Pyrophosphate. Biochemistry 16, 5470-5478.  286  Powers, T., and Walter, P. (1995). Reciprocal Stimulation o f Gtp Hydrolysis by 2 Directly Interacting Gtpases. Science 269, 1422-1424. Quiros, L . M . , and Salas, J . A . (1995). Biosynthesis o f the Macrolide Oleandomycin by Streptomyces-Antibioticus - Purification and Kinetic Characterization o f an Oleandomycin Glucosyltransferase. Journal o f Biological Chemistry 270, 18234-18239. Rademacher, T . W . , Parekh, R . B . , and Dwek, R . A . (1988). Glycobiology. Annual Review of Biochemistry 57, 785-838. Raibaud, O., and Schwartz, M . (1984). Positive Control o f Transcription Initiation i n Bacteria. Annual Review o f Genetics 18, 173-206. Ramakrishnan, B . , Balaji, P . V . , and Qasba, P . K . (2002). Crystal structure o f beta 1,4galactosyltransferase complex with U D P - G a l reveals an oligosaccharide acceptor binding site. Journal o f Molecular Biology 318, 491-502. Ramakrishnan, B . , Boeggeman, E . , and Qasba, P . K . (2005). Mutation o f arginine 228 to lysine enhances the glucosyltransferase activity o f bovine beta-l,4-galactosyltransferase I. Biochemistry 44, 3202-3210. Ramakrishnan, B . , and Qasba, P . K . (2001). Crystal structure o f lactose synthase reveals a large conformational change in its catalytic component, the beta 1,4galactosyltransferase-1. Journal o f Molecular Biology 310, 205-218. Ramakrishnan, B . , and Qasba, P . K . (2002). Structure-based design o f beta 1,4galactosyltransferase I (beta 4GaI-Tl) with equally efficient Nacetylgalactosaminyltransferase activity - Point mutation broadens beta 4 G a I - T l donor specificity. Journal o f Biological Chemistry 277, 20833-20839. Ramakrishnan, B . , Ramasamy, V . , and Qasba, P . K . (2006). Structural snapshots o f beta1,4-galactosyltransferase-l along the kinetic pathway. Journal o f Molecular Biology 357, 1619-1633. Ramakrishnan, B . , Shah, P.S., and Qasba, P . K . (2001). alpha-Lactalbumin ( L A ) stimulates milk beta-l,4-galactosyltransferase I (beta 4 G a l - T l ) to transfer glucose from UDP-glucose to N-acetylglucosamine - Crystal structure o f beta 4 G a l - T l - L A complex with U D P - G l c . Journal o f Biological Chemistry 276, 37665-37671. Ramasamy, V . , Ramakrishnan, B . , Boeggeman, E . , Ratner, D . M . , Seeberger, P . H . , and Qasba, P . K . (2005). Oligosaccharide preferences o f beta 1,4-galactosyltransferase-I: Crystal structures o f Met340His mutant o f human beta 1,4-galactosyltransferase-I with a  287  pentasaccharide and trisaccharides of the N-glycan moiety. Journal of Molecular Biology 353, 53-67. Rapoza, M.P., and Webster, R.E. (1995). The Products of Gene I and the Overlapping inFrame Gene X I are Required for Filamentous Phage Assembly. Journal of Molecular Biology 248, 627-638. Rath, V . L . , Ammirati, M . , Danley, D.E., Ekstrom, J.L., Gibbs, E . M . , Hynes, T.R., Mathiowetz, A . M . , McPherson, R.K., Olson, T.V., Treadway, J.L., et al. (2000a). Human liver glycogen phosphorylase inhibitors bind at a new allosteric site. Chemistry & Biology 7, 677-682. Rath, V . L . , Ammirati, M . , LeMotte, P.K., Fennell, K.F., Mansour, M . N . , Danley, D.E., Hynes, T.R., Schulte, G.K., Wasilko, D.J., and Pandit, J. (2000b). Activation of human liver glycogen phosphorylase by alteration of the secondary structure and packing of the catalytic core. Molecular Cell 6, 139-148. Reardon, J.E., and Abeles, R . H . (1986). Mechanism of Action of Isopentenyl Pyrophosphate Isomerase - Evidence for a Carbonium-Ion Intermediate. Biochemistry 25, 5609-5616. Reinert, D.J., Jank, T., Aktories, K . , and Schulz, G.E. (2005). Structural basis for the function of Clostridium difficile toxin B. Journal of Molecular Biology 351, 973-981. Rudolph, F.B., and Fromm, H.J. (1971). Computer Simulation Studies with Yeast Hexokinase and Additional Evidence for Random Bi-Bi Mechanism. Journal of Biological Chemistry 246, 6611-&. Rye, C.S., and Withers, S.G. (2002). Elucidation of the mechanism of polysaccharide cleavage by chondroitin A C lyase from Flavobacterium heparinum. Journal of the American Chemical Society 124, 9756-9767. Sandvig, K., and Vandeurs, B . (1994). Endocytosis and Intracellular Sorting of Ricin and Shiga Toxin. Febs Letters 346, 99-102. Schreiner, P.R., Schleyer, P.v.R., and Hill, R . K . (1993). Reinvestigation of the SNi reaction. The ionization of chlorosulfites. Journal of Organic Chemistry 58, 2822-2829. Schreiner, P.R., Schleyer, P.v.R., and Hill, R . K . (1994). Mechanisms of front-side substitutions. The transition states for the SNi decomposition of methyl and ethyl chlorosulfite in the gas phase and in solution. Journal of Organic Chemistry 59, 18491854.  288  Segel, I.H. (1975). Enzyme Kinetics: Behaviour and Analysis of Rapid Equilibrium adn Steady-State Enzyme Systems (New York, John Wiley & Sons, Inc.). Shao, FL, He, X.Z., Achnine, L., Blount, J.W., Dixon, R.A., and Wang, X . Q . (2005). Crystal structures of a multifunctional triterpene/flavonoid glycosyltransferase from Medicago truncatula. Plant Cell 77, 3141-3154. Shishova, E.Y., Di Costanzo, L., Cane, D.E., and Christianson, D.W. (2007). X-ray crystal structure of aristolochene synthase from Aspergillus terreus and evolution of templates for the cyclization of farnesyl diphosphate. Biochemistry 46, 1941-1951. Sidhu, S.S., Weiss, G.A., and Wells, J.A. (2000). High copy display of large proteins on phage for functional selections. Journal of Molecular Biology 296, 487-495. Simon, E.S., Bednarski, M.D., and Whitesides, G.M. (1988). Synthesis of Cmp-Neuac from N-Acetylglucosamine - Generation of Ctp from Cmp Using Adenylate Kinase. Journal of the American Chemical Society 770, 7159-7163. Sinnott, M.L., and Jencks, W.P. (1980). Solvolysis of D-glucopyranosyl derivatives in mixtures of ethanol and 2,2,2-trifluoroethanol. Journal of the American Chemical Society 702, 2026-2032. Sinnott, M.L., and Souchard, I.J. (1973). Mechanism of Action of Beta-Galactosidase Effect of Aglycone Nature and Alpha-Deuterium Substitution on Hydrolysis of Aryl Galactosides. Biochemical Journal 133, 89-98. Skarzynski, T., Kim, D.H., Lees, W.J., Walsh, C.T., and Duncan, K. (1998). Stereochemical course of enzymatic enolpyruvyl transfer and catalytic conformation of the active site revealed by the crystal structure of the fluorinated analogue of the reaction tetrahedral intermediate bound to the active site of the C115A mutant of MurA. Biochemistry 37, 2572-2577. Smith, G.P. (1985). Filamentous Fusion Phage - Novel Expression Vectors That Display Cloned Antigens on the Virion Surface. Science 228, 1315-1317. Smith, G.P. (1993). Surface Display and Peptide Libraries - Articles Occasioned by a Meeting at the Banbury-Center, Cold Spring Harbor Laboratory, April 4-7, 1992 Preface. Gene 128, 1-2. Smith, G.P., and Petrenko, V . A . (1997). Phage display. Chemical Reviews 97, 391-410. Smith, G.P., and Scott, J.K. (1993). Libraries of Peptides and Proteins Displayed on Filamentous Phage. Methods in Enzymology 277, 228-257.  289  Sommer, N . , Depping, R., Piotrowski, M . , and Ruger, W. (2004). Bacteriophage T4 alpha-glucosyltransferase: a novel interaction with gp45 and aspects of the catalytic mechanism. Biochemical and Biophysical Research Communications 323, 809-815. St. Hilaire, P.M., Boyd, M . K . , and Toone, E.J. (1994). Interaction of the Shiga-like Toxin Type 1 B-Subunit with Its Carbohydrate Receptor. Biochemistry 33, 14452-14463. Starks, C M . , Back, K.W., Chappell, J., and Noel, J.P. (1997). Structural basis for cyclic terpene biosynthesis by tobacco 5-epi-aristolochene synthase. Science 277, 1815-1820. Stick, R.V., Stubbs, K . A . , and Watts, A . G . (2004). Modifying the regioselectivity of glycosynthase reactions through changes in the acceptor. Australian Journal of Chemistry 57, 779-786. Stick, R.V., and Watts, A . G . (2002). The chameleon of retaining glycoside hydrolases and retaining glycosyl transferases: The catalytic nucleophile. Monatshefte Fur Chemie 133, 541-554. Stirtan, W.G., and Withers, S.G. (1996). Phosphonate and alpha-fluorophosphonate analogue probes of the ionization state of pyridoxal 5'-phosphate (PLP) in glycogen phosphorylase. Biochemistry 35, 15057-15064. Strahlbolsinger, S., Immervoll, T., Deutzmann, R., and Tanner, W. (1993). Pmtl, the Gene for a Key Enzyme of Protein O-Glycosylation in Saccharomyces-Cerevisiae. Proceedings of the National Academy of Sciences of the United States of America 90, 8164-8168. Strobel, H., Ladant, D., and Jestin, J.L. (2003). In vitro selection for enzymatic activity: A model study using adenylate cyclase. Journal of Molecular Biology 332, 1-7. Sujino, K . , Uchiyama, T., Hindsgaul, O., Seto, N.O.L., Wakarchuk, W.W., and Palcic, M . M . (2000). Enzymatic synthesis of oligosaccharide analogues: UDP-Gal analogues as donors for three retaining alpha-galactosyltransferases. Journal of the American Chemical Society 122, 1261-1269. Sun, H.Y., Lin, S.W., Ko, T.P., Pan, J.F., Liu, C.L., Lin, C.N., Wang, A.H.J., and Lin, C.H. (2007). Structure and mechanism of Helicobacter pylori fucosyltransferase - A basis for lipopolysaccharide variation and inhibitor design. Journal of Biological Chemistry 282, 9973-9982. Takahashi, M . , Inoue, N . , Ohishi, K . , Maeda, Y . , Nakamura, N . , Endo, Y . , Fujita, T., Takeda, J., and Kinoshita, T. (1996). PIG-B, a membrane protein of the endoplasmic  290  reticulum with a large lumenal domain, is involved in transferring the third mannose of the GPI anchor. Embo Journal 15, 4254-4261. Tarbouriech, N . , Charnock, S.J., and Davies, G.J. (2001). Three-dimensional structures of the M n and M g dTDP complexes of the family GT-2 glycosyltransferase SpsA: A comparison with related NDP-sugar glycosyltransferases. Journal of Molecular Biology 314, 655-661. Tawfik, D.S., and Griffiths, A . D . (1998). Man-made cell-like compartments for molecular evolution. Nature Biotechnology 16, 652-656. Ten Hagen, K . G . , Bedi, G.S., Tetaert, D., Kingsley, P.D., Hagen, F.K., Balys, M . M . , Beres, T.M., Degand, P., and Tabak, L . A . (2001). Cloning and characterization of a ninth member of the UDP-GalNAc : polypeptide N-acetylgalactosaminyltransferase family, ppGaNTase-T9. Journal of Biological Chemistry 276, 17395-17404. Ten Hagen, K . G . , Tetaert, D., Hagen, F.K., Richet, C , Beres, T.M., Gagnon, J., Balys, M . M . , VanWuyckhuyse, B., Bedi, G.S., Degand, P., et al. (1999). Characterization of a UDP-GalNAc : polypeptide N-acetylgalactosaminyltransferase that displays glycopeptide N-acetylgalactosaminyltransferase activity. Journal of Biological Chemistry 274, 2786727874. Thoden, J.B., Gulick, A . M . , and Holden, H . M . (1997). Molecular structures of the S124A, S124T, and S124V site-directed mutants of UDP-galactose 4-epimerase from Escherichia coli. Biochemistry 36, 10685-10695. Thompson, J., Pikis, A . , Ruvinov, S.B., Henrissat, B., Yamamoto, H . , and Sekiguchi, J. (1998). The gene glvA of Bacillus subtilis 168