Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The catalytic mechanism of retaining [beta]-glycosidases Vocaldo, David J. 2002

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

Item Metadata

Download

Media
831-ubc_2002-73255X.pdf [ 16.93MB ]
Metadata
JSON: 831-1.0061340.json
JSON-LD: 831-1.0061340-ld.json
RDF/XML (Pretty): 831-1.0061340-rdf.xml
RDF/JSON: 831-1.0061340-rdf.json
Turtle: 831-1.0061340-turtle.txt
N-Triples: 831-1.0061340-rdf-ntriples.txt
Original Record: 831-1.0061340-source.json
Full Text
831-1.0061340-fulltext.txt
Citation
831-1.0061340.ris

Full Text

THE CATALYTIC MECHANISM OF RETAINING p-GLYCOSIDASES by DAVID J. VOCADLO B.Sc, The University of British Columbia, 1994 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CHEMISTRY We accept this thesis as conforming to the required standard: THE UNIVERSITY OF BRITISH COLUMBIA Autumn 2001 ©David J. Vocadlo, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Ctf£/in($t/L<y The University of British Columbia Vancouver, Canada •f DE-6 (2/88) 11 A B S T R A C T Through the detailed characterization of the catalytic mechanisms of four different retaining (3-glycosidases and in consideration of the existing literature a detailed and generally applicable catalytic mechanism is proposed for the very large class of retaining (3-glycosidases. The catalytic mechanism of the family 39 (5-xylosidase (XynB) from Thermoanaerobacterium saccharolyticum is found to proceed through a covalent xylosyl enzyme intermediate flanked by oxocarbenium ion-like transition states. The exo-p-hexosaminidase from Streptomyces plicatus (SpHex) is demonstrated to use a mechanism involving anchimeric assistance from the 2-acetamido moiety of the substrate. The mechanism of this enzyme and, by extension, other family 20 enzymes proceeds through a bicyclic oxazoline intermediate. The functionally related exo-$-hexosaminidase from Vibrio furnisii (ExoII) employs a double-displacement mechanism proceeding through a covalent glycosyl enzyme intermediate. Despite the differences in the nature of the catalytic nucleophile and intermediate the reactions catalyzed by both these two enzymes involve transition states that are remarkable similar to each other, being highly dissociative, with significant oxocarbenium ion character. Lastly, the nature of the contentious intermediate in the hen egg white lysozyme-catalyzed hydrolysis of glycosides is investigated. In three different cases using electrospray ionization mass spectrometry, a catalytically competent covalent glycosyl enzyme intermediate is revealed during the catalytic cycle of HEWL. The three-dimensional structure of this intermediate is also determined by X-ray diffraction analysis. These studies point to the nature of the intermediate in the hydrolysis of glycosides catalyzed by all of these enzymes. On the basis of these results, we conclude that all retaining (3-glycosidases perform catalysis by the formation and subsequent breakdown of a covalent intermediate species and not by the formation of a long-lived ion pair. The covalent intermediate may either be a glycosyl enzyme, as found for most retaining glycosidases, or a bicyclic oxazoline, as found for some classes of enzymes processing sugars bearing an acetamido group at the 2-position of the substrate. This general mechanism unifies the substrate distortion proposed by Phillips, the covalent intermediate first postulated by Koshland, and the electrophilic migration of the anomeric Ill center (C-l) along the reaction coordinate. It is consistent with both the anti-periplanar lone-pair hypothesis and the principle of least nuclear motion. Most importantly, this mechanism is supported by all experimental data including kinetic isotope effects, mass spectrometry, crystal structures and enzyme kinetics for all retaining (3-glycosidases studied, to the best knowledge of the author. iv T A B L E O F C O N T E N T S A B S T R A C T " T A B L E O F C O N T E N T S iv L I S T O F F I G U R E S ix L I S T O F T A B L E S xiii L I S T O F S C H E M E S xiv L I S T O F A B B R E V I A T I O N S xv A C K N O W L E D G E M E N T S xix 1 G E N E R A L I N T R O D U C T I O N 1 1.1 G L Y C O S I D A S E S 2 7 . 7 . 7 Classification systems 2 1.1.1.1 Endo and Exo Action 4 1.1.1.2 Subsite Classification System 5 1.1.1.3 Glycosidase Families 6 7 . 7 . 2 Catalytic Mechanism of Retaining /3-Glycosidases 7 1.1.2.1 Nucleophilic Participation and the Intermediate 10 1.1.2.2 Acid/Base Catalysis 14 1.1.2.3 Substrate Distortion 19 1.1.2.4 Non-covalent Interactions 21 1.2 T O O L S FOR T H E S T U D Y O F G L Y C O S I D A S E S 2 3 7 .2 .7 Competitive inhibitors 23 1.2.2 Affinity Labels 26 1.2.3 Mechanism based inactivators 28 1.2.3.1 Cyclohexane Epoxides 29 1.2.3.2 Glycosylmethyltriazenes 30 1.2.3.3 2- and 5-Fluoro Sugars 31 7 . 2 . 4 Computational Analysis of the Transition State by Multiple Isotope Effects. 34 1.3 A I M S O F THIS THESIS 3 4 2 M E C H A N I S M O F THERMOANAEROBACTERIUMSACCHAROLYTICUM p - X Y L O S I D A S E . 3 6 2 .1 B A C K G R O U N D O F R E T A I N I N G P -XYLOSIDASES 3 6 2 . 2 O B J E C T I V E S O F THIS W O R K 4 1 2 . 3 IDENTIFICATION O F T H E C A T A L Y T I C N U C L E O P H I L E 4 2 2.3.1 Inactivation of the Enzyme 44 2.3.2 Catalytic Competence 45 V 2.3.3 Stoichiometry of Incorporation of Inactivator by ESMS and Identification of the Labeled Active Site Peptide 47 2.3.4 Sequencing of peptide • 49 2.3.5 Sequence Conservation and Alignment : 51 2.3.6 Mutagenesis of the Catalytic Nucleophile 53 2.4 D E T A I L E D M E C H A N I S T I C A N A L Y S I S T H R O U G H K I N E T I C STUDIES O F X Y N B 54 2.4.1 Subcloning, Purification, and Active Site Titration ofXynBH6 57 2.4.2 Substrate Specificity 59 2.4.3 pH Dependence 61 2.4.4 Evidence for a Two Step Mechanism Involving Oxocarbenium Ion-Like Transition States and a Covalent Xylosyl Enzyme Intermediate 63 2 AAA Structure/ reactivity studies 63 2 A A.2 Effect of Exogenous Nucleophiles 65 2.4.4.3 Kinetic Isotope Effects 70 2.4.3.4 Summary 71 2.5 IDENTIFICATION O F T H E A C I D / B A S E C A T A L Y T I C R E S I D U E A N D E V A L U A T I O N O F ITS IMPORTANCE IN C A T A L Y S I S 71 2.5.3 Inactivation of the enzyme 74 2.5.4 Stoichiometry of incorporation of inactivator studied by ESMS 75 2.5.5 Identification of the labeled active-site peptide by ESMS 76 2.5.6 Peptide sequencing 78 2.5.7 Production and Purification of the Mutant Enzyme 80 2.5.8 pH Profiles 80 2.5.9 Comparison of the Substrate Reactivity ofXynBH6 and XynBH6(E160A) 82 2.5.10 Product Analysis 86 2.5.11 Effect of Competitive Nucleophiles on Rates of Hydrolysis by XynBH6(El 60A) 87 2.5.12 Sequence Conservation and Alignment 91 2.6 S U M M A R Y 92 2.7 M E T H O D S A N D M A T E R I A L S 93 2.7.1 General Procedures for Chemical Synthesis 93 2.7.2 Generous Gifts 94 2.7.3 Syntheses. 94 2.7.4 Molecular Biology Techniques 108 2.7A.l General Proceedures 108 2.7.4.2 Amplification and Subcloning of XynB 108 2.7.4.3 Site-Directed Mutagenesis by PCR 110 2.7.4.3.1 Mutagenesis of the Catalytic Nucleophile Glu277 110 2.7.4.3.2 Mutagenesis of the Catalytic Acid/Base Glul60 112 vi 2.7.4.4 Overexpression and purification of His6-tagged XynB, XynB(E277A), and XynB(E160A)... 114 2.7.5 Mass Spectrometry Techniques ll$ 2.7.5.1 Labeling with 2F-DNPX and Proteolysis of T. saccharolyticum XynB 116 2.7.5.2 Labeling with NBX and Proteolysis of T. saccharolyticum XynBH 6 116 2.7.5.3 ESMS Analysis of the Proteolytic Digests 117 2.7.5.4 Aminolysis of the 2F-Xylosyl Labeled Enzyme 118 2.7.6 Enzyme kinetics 2.7.6.1 Active-site Titration of XynBH6 using 2F-DNPX 119 2.7.6.2 pH Dependence of kc.JKm 119 2.7.6.3 Measurement of Kinetic Isotope Effects 120 2.7.6.4 Inactivation of XynB by 2F-DNPX 121 2.7.6.5 Protection from Inactivation by 2F-DNPX 122 2.7.6.6 Reactivation of 2F-Xylosyl Enzyme Intermediate 122 2.7.6.7 Inactivation of XynBH 6 by NBX 122 2.7.6.8 Protection from Inactivation by NBX 123 2.7.6.9 Analysis of the Products of Hydrolysis by XynBH 6 and XynBH6(E160A) 123 2.7.6.9.1 i H N M R Spectrometry of the Products of Enzymatic Hydrolysis 124 2.7.6.10 Determination of the Delta Absorption Coefficient (Ae„) for Aryl Glycoside Substrates 124 3 M E C H A N I S M O F R E T A I N I N G B - H E X O S A M I N I D A S E S 125 3.1 BACKGROUND O N RETAINING (J-HEXOSAMINIDASES 125 3.1.1 Human ^ -Hexosaminidase 126 3.1.2 Mechanism of Retaining ^ -Hexosaminidases 128 3.2 OBJECTIVES O F THIS WORK 133 3.3 NAG-THIAZOLINE; AN N-ACETYL-P-HEXOSAMINIDASE INHIBITOR THAT IMPLICATES ACETAMIDO PARTICIPATION 134 3.3.1 Introduction 134 3.3.2 NAG-Thiazoline is a Potent Inhibitor of Jack Bean ^ -Hexosaminidase 136 3.3.3 MuTAG as a Pseudo-Substrate for Jack Bean ^ -Hexosaminidase 138 3.4 MECHANISM OF ( ACTION A N D IDENTIFICATION O F ASP242 AS T H E CATALYTIC NUCLEOPHILE OF VIBRIO FURNISH P-HEXOSAMINIDASE USING 2-AcETAMiDO-2-DEOXY-5-FLUORO-a-L-iDOPYRANOSYL FLUORIDE 142 3.4.1 Identity of the Catalytic Nucleophile of Family 3 147 3.4.1.1 Identification of the Labeled Active Site Peptide 154 3.4.1.2 Peptide Sequencing 156 3.4.1.3 Multiple Sequence Alignments 157 3.5 COMPARATIVE ANALYSIS O F E X O - P - H E X O S A M I N I D A S E S F R O M FAMILIES 3 A N D 20 158 3.5.1 NAG-Thiazoline as a Probe of Mechanism 159 3.5.2 Requirement for the Substrate 2-Acetamido Group 760 Vll 3.5.3 Electronic Requirement for the 2-Acetamido Group. 163 3.5.4 Exocyclic Oxonium Ion or Oxocarbenium Ion-like Transition States 166 3.5.5 a-Deuterium Kinetic Isotope Effects 171 3.6 SUMMARY 174 3.7 EXPERIMENTAL PROCEDURES 175 3.7.1 General Procedures 175 3.7.2 Generous Gifts 176 3.7.2.1 Syntheses 177 3.7.2.2 Enzyme Kinetics 193 3.7.2.2.1 Enzyme kinetics with ExoII 193 3.7.2.2.2 Enzyme Kinetics with SpHex 193 3.7.2.2.3 Kinetic Isotope Effects 194 3.7.2.2.4 Fluorescence Measurements with Jack Bean P-Hexosaminidase 195 3.7.2.2.5 Determination of the Delta Absorption Coefficient (Ae„) for Aryl Glycoside Substrates... 195 3.7.2.2.6 Labeling and Proteolysis 196 3.7.2.2.7 ESMS Analysis of the Proteolytic Digest 196 3.7.2.2.8 Aminolysis of the Labeled Enzyme Digest 197 4 MECHANISM OF HEN E G G WHITE LYSOZYME .....198 4.1 BACKGROUND OF H E W L '.198 4.2 OBJECTIVES OF THIS WORK 209 4.3 HYDROLYSIS OF A 2-FLUORO-GLYCOSIDE 210 4.4 SYNTHESIS AND TESTING OF POTENTIAL MECHANISM-BASED INHIBITORS OF H E W L 216 4.4.1 2,4-Dinitrophenyl 2-acetamido-2-deoxy-/3-D-glucopyranosyl-(l-4)-2-deoxy-2-fluoro-0-D-glucopyranoside (NAG-2FGlc-DNP) 216 4.4.2 2-Acetamido-2-deoxy-P-D-glucopyranosyl-(l-4)-2-deoxy-2-fluoro-P-D-glucopyranosyl fluoride (NAG-2FGlcF) 219 4.4.2.1 HEWL Turns Over NAG-2FGlcF 220 4.4.2.2 Observation of a Steady-State Glycosyl Enzyme Intermediate 221 4.4.3 Further Observations of Covalent HEWL Glycosyl Enzyme Intermediates 224 4.4.3.1 Testing of 2-Acetamido-2-deoxy-P-D-glucopyranosyl-(l-4)-2-acetamido-2-deoxy-fi-D-glucopyranosyl fluoride (NAG2-F) 224 4.4.3.2 Use of the Glu35Gln Mutant HEWL to Observe a Covalent Glycosyl Enzyme Intermediate by Mass Spectrometry 226 4.4.3.3 A Long-lived Glycosyl Enzyme Observed on HEWL 229 4.4.4 Mass Spectrometric Analysis of the Kinetics of Formation and Hydrolysis of the 2-Fluoro-glycosyl HEWL Intermediate 231 4.4.5 The 3-Dimensional Structure of a HEWL Glycosyl Enzyme Intermediate 235 4.5 SUMMARY 241 viii 4.6 METHODS AND MATERIALS 245 4.6.1 . General methods 245 4.6.2 Generous Gifts 246 4.6.3 Syntheses 246 4.6.4 Enzyme kinetics 255 4.6.5 Electrospray Ionisation Mass Spectrometry 255 4.6.6 Crystallization, Crystal Handling, and Data Collection 256 5 CONCLUSION 257 APPENDIX 1 264 REFERENCES 275 ix L I S T O F F I G U R E S Figure 1.1 Diagram showing the glycone and aglycone moieties of a p-glucoside 2 Figure 1.2 Reactions catalysed by retaining and inverting p-galactosidases 3 Figure 1.3 A depiction of an oligosaccharide showing examples of sites ofendo and exo action 5 Figure 1.4 Depiction of the active site pocket of a glucoamylase 6 Figure 1.5 The 4H3 conformation suggested for the glucosyl cation 13 Figure 1.6 The and orientation of the acid catalytic residue and protonation trajectory 15 Figure 1.7 Potential energy surface showing the extent of C-0 bond cleavage 16 Figure 1.8 Two possible routes for the base catalyzed hydrolysis of an a-glycosyl enzyme 17 Figure 1.9 Orientation of the lone pair of the ring oxygen, 0-5 20 Figure 1.10 4 C i conformation showing steric conflict of the incoming nucleophile 20 Figure 1.11 3,4-Dinitrophenyl 2-acetamido-2-deoxy-fi-D-glucopyranosyl-(l—>4)-2-acetamido-2-deoxy-p-D-glucopyranose-(l—•4)-2-acetamido-2-deoxy-a-L-fucopyranoside 22 Figure 1.12 Examples of some competitive inhibitors for glycosidases 26 Figure 1.13 Chemical structure of 2',3'-epoxypropyl-chitobioside 27 Figure 1.14 Affinity labels for glycosidases 28 Figure 1.15 Representative members of the best-known classes of mechanism based inactivators...29 Figure 1.16 Fluoro sugar inactivators capable of labeling retaining oc-glycosidases 33 Figure 2.1 Structure of xylan showing the (3-1,4-linked D-xylopyranosyl backbone decorated with the pendant L-arabinose and D-glucuronic acid residues 36 Figure 2.2 Structure of 2,4-dinitrophenyl p-D-xylopyranoside 37 Figure 2.3 The pyranose ring of glycosyl enzyme intermediates bound in the -1 subsite adopts a  25B conformation 39 Figure 2.4 Structures of; (a) fS-D-xyloside, (b) P-D-glucoside, (c) a-L-iduronide 40 Figure 2.5 Outline of the basic principles of neutral loss mass spectrometry 43 Figure 2.6 Inactivation of Thermoanaerobacterium saccharolyticum P-xylosidase by 2F-DNPX.. .44 Figure 2.7 Reactivation of 2-deoxy-2-fluoroxylosyl xylosidase 46 Figure 2.8 Structures of; (a) xylobiose, and (b) benzyl-P-D-thio-xyloside (BTX) 46 Figure 2.9 Reconstruct of the mass spectra of; wild-type XynB, and XynB treated with 2FDNPX.... 47 Figure 2.10 ESMS experiments on peptic digest of T. saccharolyticum p-xylosidase 49 Figure 2.11 ESMS/MS daughter ion spectrum of the 2F-DNPX labeled active site peptide 50 Figure 2.12 Partial multiple sequence alignment of the enzymes comprising family 39 of glycosyl hydrolases 52 Figure 2.13 Generic structure of a series of aryl p-D-xylopyranosides 55 Figure 2.14 Br0nsted plot of log k ,^ against pKa of the phenol leaving group found for X the Agrobacterium sp. P-glucosidase catalyzed hydrolysis of a series of aryl glucosides...57 Figure 2.15 Active site titration of XynBH6 using 2F-DNPX 59 Figure 2.16 pH dependence for the hydrolysis ofpNPX catalyzed by T. saccharolyticum p-xylosidase 62 Figure 2.17 Br0nsted plots relating the rate of T. Saccharolyticum (3-xylosidase catalyzed hydrolysis of a series of aryl xylosides with the pKa of the corresponding phenol 65 Figure 2.18 DTT as an exogenous nucleophile increases the rate of XynBH6 catalyzed hydrolysis of 2,5DNPX, but not of PX 66 Figure 2.19 Michaelis-Menten plots for the T. saccharolyticum (3-xylosidase catalyzed hydrolysis of 2,5DNPX and 4BrX 68 Figure 2.20 Rescue of activity through the use of the exogenous nucleophile azide 73 Figure 2.21 Chemical structure of the affinity label N-bromoacelyl-P-D-xylopyranosylamine 74 Figure 2.22 Inactivation of T. Saccharolyticum (3-xylosidasc by NBX 75 Figure 2.23 Reconstructed electrospray mass spectrum of native XynBH6 and XynBH6 after incubation with NBX 76 Figure 2.24 ESMS experiments on a peptic digest of XynBH6 77 Figure 2.25 ESMS/MS daughter-ion spectrum of the labeled peptide and unlabeled peptide 79 Figure 2.26: pH dependence of kcJKm for the wild-type XynBH6 xylosidase and XynBH6(E160A) xylosidase 81 Figure 2.27: Br0nsted relationships for XynBH„ and XynBH6(E160A) 84 Figure 2.28 Effect of azide on the rate of hydrolysis of 3,4DNPX for XynBH6 and XynBH6(E160A) 87 Figure 2.29 Effect of added exogenous nucleophiles on the rate of 3,4DNPX hydrolysis by XynBHs andXynBH6 88 Figure 2.30 Effect of varying concentrations of azide on the kinetic parameters of the XynBH6(E160A) xylosidase catalyzed hydrolysis of 3,4DNPX 89 Figure 2.31 Partial multiple sequence alignment of the enzymes comprising family 39 of glycosyl hydrolases 91 Figure 2.32 Generalized structure of the series of aryl |3-D-xylosides used in this study 97 Figure 3.1 Reaction catalyzed by retaining ^ xo-p-hexosaminidascs 125 Figure 3.2 Structure of ganglioside G M 2 127 Figure 3.3 Structure of some N-acyl substituted glucosaminides 129 Figure 3.4 Structural resemblance of bulgecin, allosamidin, and an oxazoline intermediate 132 Figure 3.5 Chemical structures of 2-methyl oxazoline and 2-methyl thiazoline 135 Figure 3.6 Inhibition of jack bean NAGase by NAG-thiazoline 138 Figure 3.7 Time course of release of 4-methylumbelliferone from MuTag by jack bean NAGase.... 140 Figure 3.8 Chemical structures of para-nitrophenyl 2-deoxy-2-thioacetamido-P-D-glucopyranoside xi and N-acetylglucosamino-l,5-laclone 141 Figure 3.9 Partial multiple sequence alignment of the cloned and putative family 3 p-N-acetylglucosaminidases 144 Figure 3.10 Partial multiple sequence alignment of 15 selected family 3 members including 10 p-N-acetylglucosaminidases 146 Figure 3.11 Chemical structures of 2-acetamido-2-deoxy-5-fluoro-p-D-glucopyranosyl fluoride and 2-acetamido-2-deoxy-5-tluoro-a-L-idopyranosyl fluoride 148 Figure 3.12 Michaelis-Menten plot of initial rates of hydrolysis of SFIdNAcF by V. furnisii ExoII..151 Figure 3.13 Transform of the elcctrospray mass spectrum of (a) native ExoII and (b) ExoII incubated with 5FIdNAcF 153 Figure 3.14 ESMS experiments on peptic digests of V. furnisii ExoII 155 Figure 3.15 ESMS/MS daughter-ion spectrum of the 5FIdNAc-peptide 157 Figure 3.16 Michaelis-Menten plot of the hydrolysis of 2,4DNPG by ExoII 162 Figure 3.17 Substituent analysis of the ExoII and SpHex P-hexosaminidase catalyzed hydrolysis of a series of fluoroacctamido derivatives 164 Figure 3.18 Structures of a) exocylic oxonium ion and b) oxocarbenium ion 167 Figure 3.19 Bronsted plots of the log of the first {kcM) and second (kcJKm) order rate constants for the ExoII and SpHex hydrolysis of a series of aryl glucosaminide substrates 170 Figure 3.20 NAG-thiazoline and glycerol bound to sugar binding subsites 1 and +1 of SpHEX 174 Figure 3.21 Structure of aryl N-acetyl-glucosaminides studied in this work 183 Figure 4.1 Simplified structure of a fragment of the bacterial cell wall of E. coli 198 Figure 4.2 Schematic of chitotriose bound in the active site of HEWL 200 Figure 4.3 Schematic representation of the model based on the non-productive complex of chitotriose with HEWL 201 Figure 4.4 Chitopentaose and 4-methylumbelliferyl chitotrioside 202 Figure 4.5 Transglycosylation reaction catalyzed by HEWL showing the Phillips mechanism 203 Figure 4.6 NAM-NAG-NAM bound to the active site of HEWL 206 Figure 4.7 Tritium-enriched chitotriose 207 Figure 4.8 Progress curve of the reaction of chitotetraosc, 2,4-DNPG, and HEWL 213 Figure 4.9 Structure of 2,4-dinitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranoside-(l-4)-2-deoxy-2-fluoro-p-D-glucopyranoside (NAG-2FGIc-DNP) 216 Figure 4.10 Michaelis-Menten plot of the reaction of NAG-2FGlcF with HEWL 220 Figure 4.11 Mass spectrum of HEWL in the absence and in the presence NAG-2FGlcF 222 Figure 4.12 Chemical structures of NAG-2FGlcF and NAG2-F 224 Figure 4.13 Mass spectra of a control sample containing only Glu35Gln HEWL and a sample containing both Glu35Gln HEWL and NAG 2-F 228 Figure 4.14 Mass spectra of a control sample containing only Glu35Gln HEWL and a sample xii containing both Glu35Gln HEWL and NAG-2FGlcF 230 Figure 4.15 Representative data presented as a stacked plot showing a section of the mass spectra obtained by injecting a sample of the reaction mixture at specific times 232 Figure 4.16 Plot of the kinetic data for the glycosylation of the Glu35Gln HEWL mutant enzyme by NAG-2FGlcF 232 Figure 4.17 Plot of kobs for the formation of the glycosyl enzyme intermediate on the Glu35Gln HEWL mutant enzyme against the concentration of NAG-2FGlcF used 233 Figure 4.18 Plot of the disappearance of the 2-fluoroglycosyl Glu35Gln HEWL intermediate and regeneration of the free enzyme 235 Figure 4.19 Crystals of the 2-fluoroglycosyl Glu35Gln HEWL intermediate complex reveal different morphology 236 Figure 4.20 Maximum-likelihood / aA weighted 2F„|,S-Fcaic electron density for the covalent glycosyl-enzyme intermediate of hen egg white lysozyme contoured at 0.4 electrons/ A 1 238 Figure 4.21 Carboxylate group showing the relative orientations of the syn (white) and anti (shaded) lone pairs 239 Figure 4.22 Overlap of the HEWL-covalent glycosyl enzyme intermediate and the product complex formed with MGM 240 Figure 4.23 Comparison of hexopyranosyl covalent intermediates and Michaelis complexes 242 Figure 4.24 Motion of the C-2 substituent during electrophilic migration of C-l 243 Figure 4.25 A portion of the reaction coordinate showing the difference between a pathway involving transition states (TS-1 and TS-2) surrounding a fleeting oxocarbenium ion and a pathway involving an oxocarbenium ion-like transition state 244 Figure 5.1 Potential energy surface showing the extent of C-0 bond cleavage of the glycosidic oxygen, protonation of the glycosidic oxygen, and bond formation between the nucleophile and the anomeric center 260 Figure 5.2 Cartoon indicating anomeric carbon movement during catalysis 262 xiii LIST OF TABLES Table 2.1 Kinetic parameters for the hydrolysis of a series of para-nitrophenyl glycosides by XynBH6 60 Table 2.2 Kinetic parameters for the XynBH6-catalyzed hydrolysis of a series of aryl P-D-xylopyranosides 64 Table 2.3 Kinetic parameters for the T. saccharolyticum p-xylosidase catalyzed hydrolysis of 2,5DNPX at different DTT concentrations 67 Table 2.4 Kinetic parameter for the hydrolysis of aryl xylosides by XynBH6 and XynBH6(E160A) 83 Table 2.6 The series of aryl P-D-xylosides used in this study 98 Table 2.7 NMR data for the pyranose ring of aryl 2,3,4-tri-O-acetyl-P-D-xylosides 100 Table 2.8 data for the phenyl ring of 2,3,4-tri-O-acetyl-P-D-xylosides 102 Table 2.9 NMR data for the pyranose ring of aryl P-D-xylosides 104 Table 2.10 NMR data for the phenyl ring of aryl P-D-xylosides 106 Table 3.1 Ratio of hydrolytic activity of selected p-hexosaminidases toward 2-acetamido-2-deoxy-p-D-gluco- and galactopyranosides 126 Table 3.2: Inhibiton of p-hexosaminidase catalyzed hydrolysis of pNPGlcNAc by NAG-thiazoline, NAGal-Thiazoline, GlcNAc, andGalNAc 160 Table 3.3: P-Hexosaminidase-catalyzed hydrolysis offluoro-substituted /?ara-nitrophenyl-2-acetamido-2-deoxy-p-D-gIucosides 163 Table 3.4 Kinetic parameters for the hydrolysis of a series of aryl glucosaminides catalyzed bySpHex andExoII 169 Table 3.5 The series of aryl 2-acetamido-2-deoxy-p-D-glucopyranosides used in this study 184 Table 3.6 NMR data for the pyranose ring of the aryl 2-acetamido-2-deoxy-3,4,6-tri-0-acetyl-P-D-glucosides 186 Table 3.7 NMR data for the aryl ring of the aryl 2-acetamido-2-deoxy-3,4,6-tri-(9-acetyl-p-D-glucosides 187 Table 3.8 NMR data for the pyranose ring of the aryl 2-acetamido-2-deoxy-P-D-glucosides 188 Table 3.9 NMR data for the aryl ring of the aryl 2-acetamido-2-deoxy-p-D-glucosides 189 Table 4.1 Cleavage data for selected oligosaccharides 205 Table 4.2 Michaelis-Menten parameters for HEWL-catalyzed hydrolysis of synthetic substrates....212 xiv LIST OF SCHEMES Scheme 1.1 The proposed mechanism of action of inverting glycosidases 7 Scheme 1.2 Three possible mechanisms proposed for hen egg white lysozyme : 9 Scheme 1.3 Ideal mechanism of inactivation of a retaining p-glucosidase by cyclophellitol 30 Scheme 1.4 Ideal mode of inactivation of a P-retaining galactosidase by galactosylmethyl-(4-nitrophenyl)triazene 31 Scheme 1.5 Mechanism of action of 2-deoxy-2-fluoro-p-D-glucosyl fluoride on a retaining P-glucosidase 32 Scheme 2.1 Hypothetical mechanism of retaining p-xylosidases involving a covalent xylosyl enzyme intermediate 38 Scheme 2.2 Transglycosylation reaction shown for a retaining P-xylosidase 41 Scheme 2.3 Simplified scheme showing the labeling and proteolysis to generate a mixture of peptides, one of which bears the label 43 Scheme 2.4 Kinetic scheme outlining the reaction of XynB with 2F-DNPX 45 Scheme 2.5 Kinetic mechanism for XynBH 63 Scheme 3.1 Three possible mechanisms proposed for hen egg white lysozyme (HEWL) 128 Scheme 3.2 Proposed reaction mechanism explaining the pseudo substrate behaviour of (±)-6-acetamido-l,2-anhydro-6-deoxy-wryo-inositol with certain p-hexosminidases 131 Scheme 3.3 Two possible routes for the hydration of 2-acetamido glucal 131 Scheme 3.4 Two possible catalytic mechanisms for retaining N-acetyl-p-hexosaminidases 134 Scheme 3.5 Possible routes leading to the hydrolysis of an oxazoline 136 Scheme 3.6 Synthesis of NAG-thiazoline 137 Scheme 3.7 Synthesis of 4-methylumbelliferyl 2-deoxy-2-thioacetamido-p-D-glucopyranoside 139 Scheme 3.8 Routes to the synthesis of 5FGlcNAcF and 5FIdNAcF 150 Scheme 3.9 Synthesis of deuterium enriched pNPGlcNAc 172 Scheme 4.1 Strategy for the chemical mutagenesis of Asp52 to homoserine 202 Scheme 4.2 Transglycosylation reaction catalyzed by HEWL 211 Scheme 4.3 Synthesis of 2,4-dinitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside-(l-4)-2-deoxy-2-fluoro-p-D-glucopyranoside 217 Scheme 4.4 Synthesis of the alternative mechanism based inhibitor 2-acetamido-2-deoxy-P-D-glucopyranoside-(l-4)-2-deoxy-2-fluoro-p-D-glucopyranosyl fluoride 219 Scheme 4.5 The reaction of HEWL with NAG-2FGlcF 222 Scheme 4.6 The reaction of Glu35Gln HEWL with NAG 2-F 228 Scheme 4.7 The reaction of Glu35Gln HEWL with NAG-2FGlcF 230 LIST OF ABBREVIATIONS 2,3DNPGlcNAc 2,3-Dinitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranoside 2,5DNPX 2,5-Dinitrophenyl p-D-xylopyranoside 2F-DNPG 2,4-Dinitrophenyl 2-deoxy-2-fluoro-p-D-glucopyranosyl fluoride 2F-DNPX 2,4-Dinitrophenyl 2-deoxy-2-fluoro p-D-xylopyranoside 2F-X 2-fluoro-D-xylose 3,4DMeX 3,4-Dimethylphenyl P-D-xylopyranoside 3.4DNPX 3,4-Dinitrophenyl P-D-xylopyranoside 3,5DC1PX 3,5-Dichlorophenyl p-D-xylopyranoside 3,5DNPGlcNAc 3,5-Dinitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside 3C1X /neta-Chlorophenyl p-D-xylopyranoside 4BrX para-Bromophenyl P-D-xylopyranoside 5FGlcNAcF 2-Acetamido-2-deoxy-5-fluoro-P-L-glucopyranosyl fluoride 5FIdNAc-enzyme 2-Acetamido-2-deoxy-5-fluoro-a-L-idopyranosyl enzyme intermediate 5FIdNAcF 2-Acetamido-2-deoxy-5-fluoro-a-L-idopyranosyl fluoride AIBN 2,2'-Azo-bis-isobutyrylnitrile ALPH Antiperiplanar lone pair hypothesis BSA Bovine serum albumin BTX Benzyl P-D-thio-xylobioside Bz Benzene Cex Cellulomonas fimi xylanase CI Chemical ionisation C-terminus Carboxy terminus of a peptide Da Dalton DABCO l,4-Diazabicyclo[2.2.2]octane DAP Diaminopimelic acid residue DCI Desorption chemical ionisation DMF Dimethylformamide DNA Deoxyribonucleic acid DNP 2,4-Dinitrophenol DNPX 2,4-Dinitrophenyl P-D-xylopyranoside DTT Dithiothreitol E Enzyme E.C. Enzyme Commision ES Enzyme-substrate complex xvi ESMS Electrospray mass spectrometry ExoII Vibrio furnisii P-hexosaminidase FDNB Fluoro-2,4-dinitrobenzene h Hours HEWL Hen egg-white lysozyme. HEWL(E35Q) Acid/base catalytic mutant (Glu to Gin) of HEWL His6 Hexahistidine peptide HPLC High Pressure Liquid Chromatography I Inhibitor ISV Ion-source voltage Kan Kanamycin kb KilobasesofDNA fccat Catalytic rate constant (Turnover Number) ku/kD Ratio of any rate constants for protio and deteuro substrates k; Maximum first order rate constant for inactivation Km Michaelis constant of a substrate Afm a p p Apparent Michaelis constant of a substrate (in the presence of an inhibitor) kobs Pseudo-first order rate constant LC/MS Tandem liquid chrorhatography/mass spectrometry m/z Mass to charge ratio MCAC Metal chelate chromatography min Minutes mNHAcGlcNAc meta-Acetamidophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside mNPGlcNAc meta-Nitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside mNPX meta-Nitrophenyl p-D-xylopyranoside MS/MS Tandem mass spectrometry MuGlcNAc 4-Methylumbelliferyl 2-acetamido-2-deoxy-P-D-glucopyranoside MuTAG 4-Methylumbelliferyl 2-deoxy-2-thioacetamido-P-D-glucopyranoside NAG N-Acetyl-P-D-glucosamine NAG-2dGlc-pNP para-Nitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranosyl-(l-4)-P-D-2-deoxy-glucopyranoside NAG 2-F 2-acetamido-2-deoxy-P-D-glucopyranosyl-( 1 -4)-2-acetamido-2-deoxy-p-D-glucopyranosyl fluoride NAG-2FGlc-DNP 2,4-Dinitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranosyl-(l-4)-p-D-2-deoxy-2-fluoro-glucopyranoside NAG-2FGlcF 2-Acetamido-2-deoxy-P-D-glucopyranosyl-( 1 -4)-p-D-2-deoxy-2-fluoro-glucopyranosyl fluoride NAGal-Thiazoline l,2-dideoxy-2'-methyl-a-D-gaIactopyranoso-[2,l-<fl-A2'-thiazoline NAG-Glu-pNP /wzra-Nitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranosyl-( 1 -4)-p-D-glucopyranoside NAG-NAG-pNP para-Nitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranosyl-( 1 -4)-2-acetamido-2-deoxy-P-D-glucopyranoside NAG-oxazoline l,2-dideoxy-2'-methyl-a-D-galactopyranoso-[2,l-cT|-A2'-oxazoline NAG-Thiazoline l,2-dideoxy-2'-methyl-a-D-glucopyranoso-[2,l-J]-A2'-thiazoline NAM N-Acetyl-p-D-muramic acid residue NapX 2-Napthyl P-D-xylopyranoside NBS N-bromosuccinimide NBX N-Bromoacetyl-p-D-xylopyranosyl amine NMR Nuclear magnetic resonance oNHAcX orf/io-acetmidophenyl P-D-xylopyranoside oNPX ort/io-nitrophenyl p-D-xylopyranoside OR Orifice energy , P Product PCR Polymerase chain reaction PES Potential energy surface PGlcNAc Phenyl 2-acetamido-2-deoxy-p-D-glucopyranoside PLNM Principle of least nuclear motion pNAcGlcNAc para-Acetamidophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside pNPGalNAc para-Nitrophenyl 2-acetamido-2-deoxy-P-D-galactopyranoside pNPGlcNAc para-Nitrophenyl 2-acetamido-2-deoxy-p-D-glucopyranoside pNPGlcNAcp para-Nitrophenyl 2-deoxy-2-fluoroacetamido-P-D-glucopyranoside pNPGlcNAcF2 para-Nitrophenyl 2-deoxy-2-difluoroacetamido-p-D-glucopyranoside pNPGlcNAcF3 para-Nitrophenyl 2-deoxy-2-trifluoroacetamido-P-D-glucopyranoside pNPTAG para-Nitrophenyl 2-deoxy-2-thioacetamido-p-D-glucopyranoside pNPX para-nitrophenyl p-D-xylopyranoside pOMeX para-methoxyphenyl P-D-xylopyranoside PX Phenyl p-D-xylopyranoside Q Quadrapole R f Retention factor SDS-PAGE Sodium dodecyl polyacrylamide gel electrophoresis SpHex Streptomyces plicatus P-hexosaminidase TBAB Tetrabutylammonium bromide TFA Trifluoroacetic acid t X V l l l TIC Total ion chromatogram TLC Thin layer chromatography TYP Tryptone-yeast-peptone growth media UV Ultaviolet light v Reaction velocity Vis Visible light V m a x / K m Specificity constant for enzymatic reactions. The second-order rate constant in units of: A\ s"1 mg enzyme"1 XynB Thermoanaerobacterium saccharolyticum fi-xylosidase XynBFL XynB bearing a hexhistine tag at the C-terminus XynBHgCE 160A) Acid/base catalytic mutant (Glu to Ala) of XynBHg XynBH6(E277A) Catalytic nucleophile mutant (Glu to Ala) of XynBHg A Heating to reflux a D (V) aD-KIE measure for the first-order rate constant a D (V/K) aD-KIE measure for the second-order rate constant aD-KIE Alpha-deuterium kinetic isotope effect Pig Slope from Br0nsted linear free energy analysis xix ACKNOWLEDGEMENTS I would like to extend heartfelt thanks to many people who have provided advice and assistance. Special thanks to Prof. Stephen Withers for his exceptional enthusiasm and his kind mentoring. I also thank those who contributed materials used in the studies described in this thesis: their names are mentioned in each chapter. I would also like to thank Dr. Davies for a rewarding collaboration on the HEWL structure as well as Profs. Knapp and James and Brian Mark for a close collaboration on aspects of the (3-hexosaminidase projects. I was fortunate to supervise a talented summer student, Jacqueline Wicki, who carried out a significant portion of the work on XynBH6 kinetic characterization. Special thanks also goes out to Karen Rupitz and Shouming He for expert technical assistance with enzyme kinetics and mass spectrometry. Profs. Warren and Strynadka deserve thanks for helpful advice and discussions and Drs. Tarling and Skupinska as well as Hoa Ly for valuable assistance and advice. Drs. Boraston and McLean are gratefully acknowledged for advice with molecular biology techniques and the Elm Street Chemistry Club for valuable and pleasant discussions. The UBC NMR staff, Microanalytic service, and Mass Spectrometry laboratories provided expert technical assistance and for this service I thank those staff members. The tolerance of my former bench mates for my territorial encroachments also deserve heartfelt thanks. Lastly I thank past and present members of the Cellulase and Withers laboratories who have contributed to making my time in these laboratories both memorable and enjoyable. Chapter 1 1 General Introduction General Introduction 1 Carbohydrates are the most abundant organic molecules on earth. They are best known in a structural capacity as the polymeric materials of cellulose and chitin or as a form of stored energy for biological systems such as starch or glycogen. Carbohydrates, however, are also used to encode complex information, allowing cells and organisms to communicate with, or battle, one another. Nature clearly has recognized both the great chemical stability and the rich information content of oligosaccharides. The formidable stability of oligosaccharides under physiological conditions is underscored by the five million year half-life of a common glycosidic linkage,1 while the vast potential for the storage of information is well illustrated by considering that the union of three simple hexose units through glycosidic linkages provides 38016 possible trisaccharides.2 Such information encoding far exceeds that from the relatively few possible combinations available to both polypeptides and oligonucleotides. Unsurprisingly, given the enormous structural diversity of oligosaccharides and glycoconjugates, a formidable array of enzymes is required for their metabolism. The catabolism of glycosides, in particular, is a challenging task because of the great kinetic stability of the glycosidic linkage. To catalyze the breakdown of such linkages on a useful biological time scale these enzymes, termed glycosidases, must accelerate the hydrolysis reaction on the order of 1017 fold. This tremendous rate acceleration is among the largest accelerations attained by any enzyme3 and corresponds to a lowering of the energy barrier leading to hydrolysis by 22.1 kcal/mol.1 The understanding of how these highly proficient enzymes catabolize the glycosidic linkage is therefore of fundamental interest simply in understanding the tremendous catalytic prowess of enzymes in general. On a more practical level the use of bioengineered glycosidases for the conversion of the abundant carbohydrate biomass to desirable products for human consumption is actively pursued by industry. To this end a detailed understanding of the mechanism of action of glycosidases is a prerequisite for certain industrial applications. Chapter 1 1.1 Glycosidases General Introduction 2 Glycosidases comprise a large superfamily of enzymes that catalyze the breaking of glycosidic linkages through the hydrolytic cleavage of the C - 0 bond between the glycone, which is always a carbohydrate moiety, and the aglycone that may or may not be a carbohydrate group (Figure 1.1). Most commonly, for natural substrates, the aglycone is another saccharide unit although it may also be an alcohol, phosphate, or thiol group. These enzymes often demonstrate high tolerance for the structure of the aglycone and thus may be active on a wide range of synthetic substrates making them readily amenable to detailed kinetic study. Glycone Aglycone Figure 1.1 Diagram showing the glycone and aglycone moieties of a 3-glucoside with the arrow indicating the glycosidic bond. 1.1.1 Classification systems Many methods have been devised to classify the action of glycosidases. The International Union of Biochemistry and Molecular Biology (IUBMB) classifies glycosidases on the basis of their glycone specificity. Occasionally, the IUBMB classifications may indicate the nature of the substrate aglycone although this is not always true. While extremely simple, the IUBMB rules are of limited value in so far as they do not always provide a clear indication of what the preferred substrates are, nor do they provide any idea of the reaction mechanism. A more informative classification scheme that is widely used in the literature classifies the glycosidases on the basis of the Chapter 1 General Introduction 3 configuration of the glycone and the stereochemical outcome of the reaction the enzyme catalyses. The most commonly used method divides them according to the follow criteria: glycone moiety over all others and is classified on this basis. Thus a glucosidase is most active on glucosides yet may also cleave galactosides at a lower rate. ii) Anomeric configuration of the glycone: each glycosidase is highly specific for the anomeric configuration of the glycosidic bond. A s such an cc-glycosidase w i l l cleave only oc-glycosides and not p-glycosides to any extent. iii) Stereochemical outcome of the reaction: the glycosidase may cleave the glycosidic bond with either retention or inversion of configuration at the anomeric centre. For example, an enzyme that cleaves a P-glycosidic linkage to yield a hemi-acetal with the 3 configuration is termed a retaining glycosidase. Conversely, i f the P-glycosidic linkage is cleaved to generate an a hemiacetal then the enzyme is said to be an inverting glycosidase. i) Structure of the glycone: a particular glycosidase has optimal activity on one HO OH OH Figure 1.2 Reactions catalysed by retaining and inverting [3-galactosidases. To summarize with an example; i f a glycosidase cleaves methyl P-D-galactoside more efficiently than other methyl p-D-glycosides, to yield P-D-galactose and methanol, the enzyme is defined as a P-retaining galactosidase. Chapter 1 General Introduction 4 Another, more methodical, approach for the classification of glycosidases that was suggested by Sinnott, has never been widely adopted. This approach defines the orientation of the leaving group of the substrate and the anomeric hydroxyl of the product as being either in an axial (a) or equatorial orientation (e) and then indicating the enzymatic conversion with an arrow between the letters. The enzyme in the above example (Figure 1.2) would thus be defined as an (e—»e) galactosidase. The significant advantage of this system is that it avoids the confusion associated with a and P not clearly reflecting the orientation of the glycosidic bond. Typically, for D-sugars, which are the most common carbohydrates in Nature, an equatorial leaving group is in a (3 configuration. For some L sugars however, a few of which occur naturally, an equatorial group is typically termed a. Even here the complexities associated with carbohydrate nomenclature are resistant to simplification because the relative orientation of the anomeric group is dependent on the conformation of the pyranose ring. The system suggested by Sinnott thus also has its limitation as its application presupposes knowledge of the conformation of the carbohydrate that researchers may not agree upon or even be aware of. One possible method for conciliation of this limitation could be to adopt a standard wherein the reactive glycone hexopyranoside unit is only considered in its 4 Ci conformation, regardless of its true, lowest energy conformation. Regardless of the relative merits and disadvantages of the major systems for classifying glycosidases the convention using the anomeric configuration and stereochemical outcome will be adopted throughout this thesis. Thus all of the enzymes studied during the course of this thesis will be referred to as P-retaining glycosidases and, according to the method of Sinnott, are e—>e glycosidases. 1.1.1.1 Endo and Exo Action An additional distinction used to classify the glycosidases stems from the fact that many of these enzymes act on polymeric substrates such as cellulose. Some of these enzymes have evolved to cleave their substrates somewhere in the middle of the oligosaccharide chain and these are known as endo-glycosidases. Others have evolved to Chapter 1 General Introduction 5 cleave their substrates at either end of the oligosaccharide chain, typically releasing mono- or disaccharides, and these enzymes are known as exo-glycosidases. This distinction serves only to define the substrate specificity of a glycosidase and has no bearing on the catalytic mechanism of the enzyme. However, this definition has been found to correlate with the tertiary structure of glycosidases: the era-acting enzymes commonly have a pocket shaped active site that accommodates only the end of a polymeric substrate while ercdo-acting enzymes typically have a groove shaped active site into which a long polysaccharide chain may bind. C K X X X K ) . a o a o o EXO ENDO EXO Figure 1.3 A depiction of an oligosaccharide showing examples of sites of endo and exo action of glycosidases. 1.1.1.2 Subsite Classification System Unsurprisingly, some glycosidases (both endo and exo) have evolved to bind several saccharide residues of their sometimes-large substrates. These residues are accommodated in a number of subsites and, on occasion, the 3-dimensional structure of a glycosidase can provide a good estimate as to the number of subsites present in an enzyme. Some glycosidases have a binding site that interacts with only one residue while others may have six or more subsites. Davies has advanced a useful nomenclature for the description of the number of binding sites and the relative positioning of the catalytic machinery.4 According to this system the catalytic residues are positioned between the -1 and +1 subsites, the subsites increasing in number +1, +2, +3... towards the reducing end of the substrate and decreasing in number -1,-2,-3... towards the non-reducing end of the substrate. Glucoamylase, for example, is an exo-acting enzyme with a pocket shaped binding site containing four subsites (-1, +1, +2, and +3, Figure 1.4). This 6 General Introduction Chapter 1 enzyme thus prefers oligosaccharides containing four residues or more and it cleaves off one saccharide unit from the non-reducing end of the oligosaccharide chain. \ ^OH / 0 h < ^OH I OH OH 0 H O / W W W OH -1 + 1 +2 +3 Figure 1.4 Depiction of the active site pocket of a glucoamylase containing four binding subsites with the catalytic machinery positioned between the -1 and +1 subsites. 1.1.1.3 Glycosidase Families Henrissat has developed an extremely useful classification system for glycosidases that divides these enzymes into numerous families on the basis of sequence similarities. Currently, over 5000 glycosidases have been sequenced and these have been categorized into over 70 unique families. Members within each family are predicted to have the same three-dimensional fold and catalyze reactions with the same stereochemical outcome. These expectations have been borne out where studies on more than one glycosidase from a family are available for comparison.5"11 The great similarity of the tertiary structures of different family members and the markedly high primary sequence similarity often found in the active site region permits the prediction of key catalytic residues. This classification system is therefore extremely useful as experimental results that unambiguously reveal the identity of catalytic residues can almost always be easily extrapolated to all members of a particular family. Several families that have the same tertiary fold, despite having low overall sequence similarity, have been grouped into several different clans. Presumably, these clans contain Chapter 1 General Introduction 7 evolutionarily related families and so, while extrapolation of active site residues is a more risky endeavor within a clan, useful insights may still be gleaned. 1.1.2 Catalytic Mechanism of Retaining p-Glycosidases Two major classes of catalytic mechanisms have been described for the glycosidases and are clearly distinguished on the basis of the stereochemical outcome of the catalyzed reaction, vide supra. The inverting glycosidases, which will only be introduced briefly, are believed to use a one step displacement mechanism in which an enzymic general base catalytic residue facilitates the attack of water on the anomeric center. An enzymic acid catalytic residue (Scheme 1.1) facilitates concomitant departure of the aglycone. These catalytic residues have in all cases to date been found to be two carboxyl residues positioned approximately 11 A apart.12 Scheme 1.1 The proposed mechanism of action of inverting glycosidases. The catalytic carboxyl groups are shown with their relative separation. The large majority of retaining glycosidases have also been found to employ two carboxyl residues in their catalytic mechanism. However, in this class of glycosidase the catalytic groups are found to be approximately 5.5 A apart and have roles that differ from those in inverting glycosidases.12 Surprisingly, the catalytic mechanism of retaining glycosidases has remained an issue of contention for over 30 years.13"15 Hen egg white lysozyme (HEWL) was the first Chapter 1 General Introduction 8 enzyme to have its tertiary structure solved using X-ray diffraction methods.16 Shortly after this tremendous achievement, Phillips proposed a mechanism on the basis of model building studies and the structure of the em/o-glycosidase HEWL in a non-productive complex with the oligosaccharide chitotriose.13-17 According to its great historic importance, this catalytic mechanism has gained a place in many biochemistry and organic chemistry textbooks as a paradigm for the mechanism of p-retaining glycosidases. In the textbook mechanism proposed by Phillips (Figure 1.6, Path B) the enzyme cleaves the polysaccharide bacterial cell wall substrate by first binding the substrate in a cleft six saccharide units in length (-4 to +2).13 The substrate was proposed to bind in a distorted conformation in which the saccharide unit bound in the -1 subsite adopts a half-chair conformation. It was suggested that a catalytic acid/base residue, Glu35, protonates the glycosidic oxygen while Asp52 acts to stabilize the developing oxocarbenium ion intermediate through electrostatic interactions.13-18'19 It was also proposed that Asp52 acted to shield the a-face of the glucosyl oxocarbenium ion intermediate so that it could only be intercepted from the P-face to yield the hydrolysed product with retained stereochemistry at C - l . 2 0 Earlier Koshland, however, had speculated that the glycosyl enzyme intermediate of retaining P-glycosidases may be a covalent species.15 The debate continues and the features of this mechanism that have been most contentious are the role of substrate distortion and the nature of the intermediate.21-23 An alternative catalytic mechanism has been advanced for HEWL wherein the 2-acetamido group of the substrate acts as an intramolecular nucleophile to displace the leaving group, forming a bicyclic oxazolinium ion intermediate (Scheme 1.2, Path C). 2 4 Glu35 has the same role as in the Phillips mechanism while the suggested role of Asp52 is to stabilize the oxazolinium ion intermediate. Such a mechanism has been demonstrated to operate in the acid-catalyzed hydrolysis of 2-acetamido-2-deoxy-P-D-glucopyranosides.25-26 Interestingly, studies on the enzyme catalysed hydrolysis of aryl 2-acetamido-2-deoxy-P-D-glucopyranoside substrates in which the methyl group of the acetamido moiety has been replaced by mono-, di- or trifluoromethyl have provided some evidence that such a mechanism may operate in a P-retaining-N-acetyl-glucosaminidase.27 Chapter 1 General Introduction 9 E35 D52 Scheme 1.2 Three possible mechanisms proposed for hen egg white lysozyme (HEWL). A, Koshland double displacement mechanism; B, Phillips ion-pair mechanism; C, anchimeric assistance mechanism. A catalytic mechanism for HEWL involving endocyclic bond cleavage has also been proposed on the basis of computer calculations.14 This hypothesis, however, lacks any experimental support and is at odds with a considerable body of literature.28 Despite the number of mechanistic alternatives the large majority of retaining glycosidases appear to utilize a mechanism initially hypothesized by Koshland in 1953. This mechanism has two distinct steps, each of which involves a Walden inversion of the anomeric center,15 and so the mechanism is often termed the double displacement mechanism. In the first step of this mechanism, the glycosylation step, (Scheme 1.2, Path A) one carboxylate residue acts as a nucleophile to attack the anomeric center, displacing the aglycone, and leading to a covalent a-glycosyl enzyme intermediate. In the second Chapter 1 General Introduction 10 step, the deglycosylation step, water attacks the anomeric center displacing the enzymic carboxylate to yield the hydrolyzed product with retained configuration at the anomeric center and the regenerated enzyme. The second enzymic carboxyl group acts in the same manner as that suggested in the Phillips mechanism; in the first step as a general acid to facilitate cleavage of the glycosidic bond and in the second step as a general base to promote the attack of water on the anomeric center of the glycosyl enzyme intermediate. Both of the steps leading to the formation and the breakdown of the glycosyl enzyme intermediate proceed through transition states that have significant oxocarbenium ion-like character. Regardless of which mechanistic alternative is followed, several key features define the enzyme-catalyzed reaction and which of the three realistic catalytic mechanisms operate for any particular (3-retaining glycosidase. These main points have been the focus of most of the mechanistic work elucidating the mechanism of action of (3-retaining glycosidases and they naturally center on the nature of the intermediate and the transition state, issues at which these mechanistic alternative differ: i) The extent of nucleophilic participation and the nature of the intermediate. Is the intermediate covalent or ionic? ii) Considerations regarding the role and nature of acid / base catalysis. iii) The role of substrate distortion in catalysis. iv) The importance of non-covalent interactions in the transition state 1.1.2.1 Nucleophilic Participation and the Intermediate The progress of physical organic chemistry is often well illustrated by the incorporation of its tenets into the mainstream of the biochemical literature. This theory especially holds true for the glycosidases. The pre-1980 carbohydrate chemistry literature is rife with the depiction of stable, solvated, oxocarbenium ions as discrete intermediates.29 Thus the proposal of Phillips was entirely consistent with the in vogue physical organic theories of that time. The assumption that the oxocarbenium ion is a Chapter 1 General Introduction 11_ discrete entity in water was based on considerable experimental evidence including normal a-secondary deuterium kinetic isotope effects (aD-KIE's) of approximately 1.1, solvent dependencies, and structure reactivity studies all of which were consistent with an S N I reaction.30-31 Several groups, however, found that methoxymethyl derivatives (R-CH2-O-CH3), which undergo hydrolysis with classical features of an SN1 type reaction, in fact, exhibited bimolecular reaction kinetics in the presence of nucleophiles.32-33 This observation made it clear that the reaction was actually SN2, as is expected given that the estimated lifetime of the methoxymethyl cation (CH2=0+-CH3) in water is in the range of 10"12 to IO" 1 5 s.34 Arguments, however, were made that, owing to the greater stability of the glycosyl oxocarbenium ion, the hydrolysis of methyl a- and (3-glucopyranosides proceeds by an S N I reaction mechanism.35 With the seminal work of Jencks and co-workers it became increasingly clear that these "intermediates," in particular the glucosyl cation in water, had a lifetime on the order of a single picosecond!34 This lifetime approaches the time scale of a single molecular vibration (10"14 seconds) and is less than the time required (10 ps) for the rearrangement of the solvent shell around the reactant.36 From the work of Banait and Jencks it later became clear that the barrier to the collapse of the oxocarbenium ion in the presence of an anion is essentially nonexistent.37 The fatal blow to the concept of a relatively long-lived, water-solvated, oxocarbenium ion intermediate arrived innocuously, from studies on the spontaneous hydrolysis and azide catalyzed cleavage of the C-F bond of a-glucosyl fluoride. Working with this system Banait and Jencks found clear bimolecular reaction kinetics coupled with complete inversion of stereochemistry at the anomeric centre arising from displacement of the fluoride ion leaving group with azide.38 Sinnott has reconciled the complexities in the literature by asserting that "the glucopyranosyl cation is so finely on the border of a real existence that whether you get nucleophilic participation or not depends on the leaving group (anionic or neutral) and anomeric configuration"39'40 as well as perhaps the nature of the nucleophile (anionic or neutral). Such a view can certainly account for the complex results in the literature and is consistent with the lifetime of the glycosyl oxocarbenium ion estimated by Amyes and Jencks.34 Chapter I General Introduction 12_ Regardless of the extent of nucleophilic participation in the transition state, most physical organic chemists sensibly abandoned the concept of a solvent equilibrated glycosyl oxocarbenium ion intermediate and turned in favour of the idea of an "exploded" transition state.41 The key feature of such a transition state is that the combined bond orders of both the C-l-leaving group bond (breaking) and C-l-nucleophile bond (forming) are less than one. A consequence of such an exploded transition state is the significant positive charge developed on C-l in the transition state. Such a charge build up must be accompanied by development of a p-orbital at C-l in the transition state and indeed this expectation is accounted for by the fairly large, normal, a-secondary deuterium KIE's that have been observed in both spontaneous and glycosidase-catalyzed hydrolyses of glycosides.u.35,40 There has also been substantial experimental evidence in the form of studies on cc-deuterium kinetic isotope effects to support the existence of oxocarbenium ion-like transition states in the mechanism of retaining glycosidases. In several cases it has been possible to obtain kinetic isotope effects on both the glycosylation and the deglycosylation steps of glycosidases that use a double displacement mechanism, through the careful selection of the aglycone.42-46 For a hypothetical ion pair intermediate the second step of the reaction would necessarily involve rehybridization of the anomeric center from sp2 in the ion pair to sp3 in the product and so, in cases where the second step is rate determining, only an inverse kinetic isotope effect (&H/&D < 1) would be expected. Such a case has never been observed. Indeed, in every case known the ku/ku values determined for the glycosidase-catalyzed hydrolysis of substrates have been normal (kn/ku > 1) regardless of which step is rate determining. This indicates significant rehybridization of the anomeric center from sp3 in the ground state to sp2 in the transition state of both reaction steps. This rehybridization has considerable implications for the charge distribution in the transition state as well as for the geometry of the sugar ring at the transition state and the nature of the intermediate.47 In order to stabilize the developing positive charge it is believed that the sugar assumes a geometry that enables overlap of the nearly empty p-orbital at the anomeric center with a newly formed p-orbital of the endocylic oxygen thereby distributing the positive charge across both C-l and 0-5 (Figure 1.5). Recent Chapter 1 General Introduction 13 studies on the solvolysis of 2,4-dinitrophenyl P-D-glycopyranosides suggest that the majority (>75%) of the accumulating positive charge resides on the endocyclic oxygen.48 Figure 1.5 The 4/Y 3 conformation suggested for the glucosyl cation. The geometric requirements of the oxocarbenium ion-like transition state have been substantiated for the spontaneous hydrolyses of a- and P-methyl glucosides 4 9 In these studies the pyranose ring was found to adopt a AH3 conformation allowing such orbital overlap to occur (Figure 1.5) where C-5, 0-5, C-l , and C-2 are approximately coplanar. The normal isotope effects observed for both steps in the catalytic mechanism of P-retaining glycosidases also clearly indicate that the intermediate must be sp3 hybridized and that the anomeric centre must be attached by a covalent bond to the enzymic carboxylate at least in the cases studied. Site-directed mutagenesis has been applied extensively to evaluate the effect of deleting the nucleophilic carboxyl group.50 Replacement of the carboxyl group of the catalytic nucleophile (Glu358) of Agrobacterium sp. P-glucosidase with an amide (the Glu358Gln mutant) results in at least a 106 fold reduction in the rate of hydrolysis of aryl P-glucosides.51-52 Similar formidable rate reductions have been observed in several other careful studies with P-retaining glycosidases, suggesting a tremendously important role for this residue (>7.8 kcal/mol of transition state stabilization) that is consistent with its function as a nucleophile.50 In contrast to the significant body of literature pointing to covalent participation of an enzymic nucleophile there is no positive evidence pointing to an ion-pair intermediate in the catalytic mechanism of any glycosidase, including HEWL. OH Chapter 1 1.1.2.2 Acid/Base Catalysis General Introduction 14 The spontaneous hydrolysis of acetals is greatly facilitated by acid catalysis.29-31-49 In the best model systems intramolecular acid catalysis can provide 7 kcal/mol worth of transition state stabilization.49b Likewise, inverting and retaining glycosidases benefit from efficient acid/base catalysis during the hydrolysis of the glycosidic linkage. As mentioned earlier for retaining glycosidases, in the glycosylation step an enzymic carboxyl group acts as an acid catalyst and facilitates departure of the leaving group. From the 3-dimensional structures of several glycosidases this residue has been found in one of two positions, relative to the substrate, where it forms a favorable hydrogen bond with the glycosidic oxygen. The concept of directional protonation of the glycosidic oxygen during glycosidase catalyzed hydrolysis of glycosides was originally proposed by Vasella and coworkers in order to rationalize the great discrepancy in the efficacy of competitive triazole inhibitors (vide infra) with enzymes from different families of glycosidases (Figure 1.6).53 The two possible positions of the acid/base residue are dictated by the orientation of the protonated lone pair on the glycosidic oxygen of the Michaelis complex and have been termed syn or anti. The syn and anti positions of the proton donor are defined by the position of this group relative to the ring oxygen (0-5) of the pyranose sugar in the -1 subsite (the glycone moiety). If the acid/base residue is on the side of a plane formed from C-l , 0-1, and H-1 opposite to the ring oxygen then the proton trajectory is defined as anti. When the proton donor is on the same side of this plane as the endocyclic oxygen then the proton trajectory is defined as syn. Unsurprisingly, because there is considerable similarity of the active site within glycosidase families, every member within one family utilizes the same protonation trajectory.54 The pKa of the conjugate acid of the glycosidic oxygen is approximately -5 and in comparison the pKa of a common carboxylic acid is 4.5. However, the pKa of the acid catalyst of many glycosidases has been found to be significantly perturbed upward to between 6 and 8. The significant pKa perturbation of the acid catalytic residue is presumed to arise from the architecture of the enzyme active site where the nucleophilic carboxyl group is held in relatively close proximity to the acid/base residue. Figure 1.6 The anti orientation of the acid catalytic residue and protonation trajectory as the basis for rationalizing the relative potency of two triazole inhibitors. A, Anti protonation of the glycosidic oxygen of a Michaelis complex; B, absence of any possible hydrogen bonding to the 1,2,3 triazole inhibitor; C, possible "side-on" or anti protonation of the 1,3,4 triazole inhibitor. In this close position it exerts an electrostatic field on the acid/base residue and gives rise to a change in its pKa. The glycosidases have presumably evolved to shift the pKa of the general acid/base catalytic residue above the pH of the medium so as to maintain a favorable protonation state. With this perturbation, the pKa mismatch between the glycosidic oxygen (pKa = -5) and the perturbed acid catalyst (pKa = 6) is still about 11 pKa units. Thus it is unlikely that efficient equilibrium protonation of the glycosidic oxygen to generate substantial amounts of a discrete oxonium ion (vide infra Point B, Figure 1.7) will occur in the Michaelis (enzyme-substrate) complex as has been suggested.18 Protonation of the glycosidic oxygen therefore most likely occurs concomitantly with bond cleavage in the transition state. This hypothesis is supported by small molecule studies showing that the base catalyzed hydrolysis of glycosyl fluorides involves incomplete deprotonation of the incoming solvent at the transition state.37 Earlier studies into the acid-catalyzed hydrolysis of glycosides suggested that equilibrium protonation of the glycosidic oxygen occurred prior to breaking of the glycosidic bond.31 However, whether equilibrium protonation occurs or whether there is general acid catalysis depends on a number of factors. A general rule is that if the fully protonated intermediate is too unstable to exist (i.e. its lifetime is < 10"12 s) then general acid catalysis will occur.39 The lifetime of the protonated oxonium ion (vide supra Point B, Figure 1.7) intermediate is, of course, closely linked to the height of the barrier leading to Chapter 1 General Introduction 16 breakage of the glycosidic linkage. For example, in cases where there is a good leaving group or relief from strain, then general acid catalysis will be typically be observed. In the enzyme catalyzed hydrolysis of glycosides, where the barrier is greatly lowered, it is clear that protonation of the glycosidic oxygen is incomplete in the transition state. Perhaps the best support comes from the analysis of several Br0nsted plots that have been constructed for the cleavage of aryl glycosides by a number of retaining (3-glycosidases. 43,44,55,56 These studies have suggested that protonation of the glycosidic bond is not complete in the transition state and is often not even significantly advanced.44-56 When coupled with the large, normal, aD-KIE's commonly found for the glycosylation step (~ 1.1) and solvent isotope effects of a relatively small magnitude (1.3-1.7) it is clear that cleavage of the glycosidic bond is more advanced than is protonation of the glycosidic oxygen.57 A graphical illustration of this view can be seen in the potential energy surface (PES) diagram below. Figure 1.7 Potential energy surface showing the extent of C-0 bond cleavage versus the extent of protonation of the glycosidic oxygen. No inference is made about the extent of nucleophilic participation. Phillips proposed a mechanism involving a path from point A to point B continuing to point D with no nucleophilic participation. A path typical of many glycosidases studied in depth is shown as the curved line, with the transition state depicted as TS. The glycosyl enzyme intermediate forms as the glycosidic bond is cleaved. The efficient breakdown (deglycosylation) of this acylal enzyme intermediate requires the Chapter 1 General Introduction 17 base catalyzed attack of water. The non-enzymatic base catalyzed hydrolysis of acylal esters can proceed by attack of water at either the acetal or the ester center. In the first case the product formed has undergone a Walden inversion resulting in an inversion of stereochemistry at the acetal center while in the second case the product retains its stereochemistry (Figure 1.8). Thus, for retaining glycosidases a stereochemical requirement of the reaction they catalyse is that attack of water on the glycosyl enzyme intermediate must occur at the acetal center.11 This complete selectivity is apparently accomplished by protecting the acyl center with other amino acid residues thereby hindering the approach of water.8-58 However, an intriguing possibility is that certain inverting glycosidases may utilize a mechanism involving the formation of a covalent intermediate followed by attack of water at the ester center to yield a product with inverted stereochemistry at the anomeric center (Figure 1.8). Some precedent for such a mechanism can be found in studies on the mechanism of action of the epoxide hydrolases. These enzymes act to hydrolyze oxiranes to their corresponding diols by a mechanism involving nucleophilic attack of an enzymic carboxylate on the electrophile.59 This covalent enzyme ester intermediate is then broken down by the base catalyzed attack of water at the ester center in a reaction analogous to that shown below (Figure 1.8, Path 2). Figure 1.8 Two possible routes for the base catalyzed hydrolysis of an a-glycosyl enzyme intermediate. 1, Attack at the anomeric center; 2, attack at the ester center. Possible sites of 1 8 0 incorporation are indicated. Chapter 1 General Introduction 18 Just as the first step, glycosylation, requires acid catalysis, the second step, deglycosylation, similarly benefits from base catalysis. The very same residue that earlier acted as a general acid catalyst is now deprotonated and poised to act in the capacity of a general base catalyst. Indeed, the X-ray crystal structures of several glycosidases in which the glycosyl enzyme intermediate has been observed have revealed that the general acid/base catalytic residue is hydrogen bonded to a well-ordered water molecule positioned near the anomeric center.60-61 How does the very same residue that only moments ago functioned as a general acid with a pKa of 6-8 now act as an efficient general base? Early work by Inoue et al, studying the protonation state of the general acid/base residue (Glu35) of hen egg white lysozyme, pointed the way.62 These researchers found that the pKa of the general acid/base residue was 6.1. However, upon mutagenesis of the catalytic nucleophile, Asp52Asn, the pKa dropped to 5.0, pointing to the influence of an electrostatic field on the pKa of Glu35. Another, similar study, found a change of pKa from 6.3 in the native enzyme to 4.7 in the Asp52Ala mutant.63 More elegant and convincing work supporting an electrostatic basis for pKa cycling of the acid/base residue involved direct 1 3 C NMR titrations of a xylanase from Bacillus circulans that had been 1 3 C labeled in the carboxyl group side chains of both the acid/base and the nucleophilic residues.64 In this study the pKa of the catalytic acid/base residue in the free enzyme was found to be 6.7. However, upon formation of a stable xylobiosyl enzyme intermediate the pKa of this residue dropped by 2.5 units to 4.2 where it could maintain the ionization state required to act as a general base. Site directed mutagenesis studies also designed to evaluate the role of the acid/base catalytic residue have been carried out with several enzymes. Some of the most rigorous studies have focused on the -^retaining xylanase from Cellulomonas fimi. Upon deletion of the carboxyl group of the general acid/base catalytic residue a reduction in the second order rate constant (kcat/Km), a parameter that is dependent on the pKa of the leaving group of the aryl glycoside substrate, is observed.65 Thus, the glycosylation step is significantly slowed (200-2000 fold) for substrates having poor leaving groups that require effective acid catalysis. However, activated substrates, not requiring acid catalysis, are cleaved at rates close to that of the wild type enzyme, generating the glycosyl enzyme intermediate. When the general acid/base catalytic residue is mutated Chapter 1 General Introduction 19 the deglycosylation step is also slowed by 200-2000 fold. Thus when reactive substrates are hydrolyzed by a mutant glycosidase lacking the acid/base catalytic carboxyl group the net result is that there is an accumulation of the intermediate that is reflected in kinetic analysis by an initial "burst" of the activated phenolate leaving group and an extremely low Km value. Both observations are a kinetic consequence of accumulation of the glycosyl enzyme intermediate.66 1.1.2.3 Substrate Distortion The original proposal for the catalytic mechanism of HEWL as advanced by Phillips suggested that the pyranoside ring of the substrate was distorted when bound in the Michaelis complex.17 This remarkable hypothesis was posited only on the basis of model building and the 3-dimensional structure of a non-productive complex of chitotriose bound in the active site of HEWL. The lowest energy conformation of 2-acetamido-2-deoxy-D-glucopyranose is one in which the sugar adopts a 4 Ci conformation (Figure 1.9, A). Phillips proposed that the sugar was distorted into a 4E or 4H3 conformation (Figure 1.9, B) upon binding to the enzyme.17 This hypothesis has resulted in considerable debate, in part because for many years it lacked any experimental support. Some have argued that substrate distortion does not need to be invoked to explain catalysis while others have maintained that such distortion is an absolute requirement for effective catalysis by retaining 3-glycosidases. For many years these two camps were divided according to their belief in the antiperiplanar lone pair hypothesis (ALPH). 6 7 - 6 9 ALPH is a stereoelectronic theory that holds that in a system in which a heteroatom bearing lone pairs is geminal to a leaving group, the reaction is favored when the lone pair of the hetero-atom is in an orientation antiperiplanar to the leaving group. The application of this hypothesis is that P-retaining glycosidases should react through a boat, twist boat, or half-chair conformation (Figure 1.9). A complementary theory that also accommodates the experimental evidence in support of ALPH is the principle of least nuclear motion (PLNM).7 0 This theory states quite simply that the motion of large masses is what provides the greatest barrier to any Chapter 1 : General Introduction 20 chemical reaction and therefore reactions are expected to proceed with a minimum of nuclear motion. 7 1 Figure 1.9 Orientation of the lone pair of the ring oxygen, 0-5, relative to the glycosidic oxygen of the leaving group. A, the AC{ ground state conformation of many common glycosides; B, the 4Hi or 4E conformation proposed by Phillips as the conformation of the saccharide bound in the -1 subsite; C, the 4 I B conformation and the twist-boat conformation D, required for the optimal effect of the antiperiplanar lone pair hypothesis (ALPH). While conclusive experimental support for the P L N M has remained elusive this theory has considerable appeal. Also, regardless of how appealing A L P H may be, a gradual consensus has been reached that this effect in acetals is quite small. 7 1- 7 2 Another argument that has been advanced is that nucleophilic displacement of an equatorial leaving group from an sp hybridized center is impossible on the basis of sterics. Presumably distortion of the pyranose ring into an envelope or boat conformation allows in-line attack of the nucleophile which would otherwise be hindered by the axial 3 and 5 position hydrogen atoms on the bottom face of the ring (Figure 1.10).1 0-4 0-7 3 R Figure 1.10 A, 4C\ conformation showing steric conflict of the incoming nucleophile with H-3 and H-5; B, lAB conformation showing absence of steric conflicts. Regardless of the origin for the substrate distortion hypothesized by Phillips evidence is mounting that such distortion does indeed occur in P-retaining glycosidases. The hypothesis derived from the initial model building studies of Phillips was debated Chapter 1 General Introduction 21 back and forth on the basis of the 3-dimensional structures of (^-retaining glycosidases in complex with both inhibitors and products. 17,21,74,75 Only with the recent elucidation of the 3-dimensional structures of four different retaining P-glycosidases in complex with various substrates and substrate analogues where these molecules span the +1 and -1 subsites has the role of substrate distortion truly gained any substantive support. The X -ray structures of these complexes all reveal that the pyranose ring in the active site is distorted and adopts a 4Hj, or 4E conformation. 61,76-78 The precise conformation of the sugar ring in each case remains elusive owing to the limited resolution of the structures solved and the great similarity of the half chair and envelope conformations. 1.1.2.4 Non-covalent Interactions Also of great importance to catalysis are the less heralded enzymic residues that serve to stabilize the geometry and charge of the transition state. In a visionary manner, Pauling proposed that the majority of the catalytic power of an enzyme is derived from favorable non-covalent interactions between the enzyme and transition state. 7 9 ' 8 0 Few studies have attempted to quantify the effects of these other "accessory residues" which define critical features of the enzyme active site, including its dielectric properties, electrostatics, and complementary Van der Waals surfaces. One notable exception from the perspective of both rigor and history is that of tyrosyl t -RNA synthase.8 1 The high resolution X-ray structures of the free enzyme and the enzyme in complex with an intermediate allowed Fersht and coworkers to be the first to utilize the powerful, and at that time, novel tactic of protein engineering to study this system. By systematically deleting residues in the active site they were able to demonstrate that the enzyme utilizes a complex network of hydrogen bonds to both bind the substrates and stabilize the transition state. In this case the interactions studied account for all of the catalytic power of the enzyme, as there are no residues that participate in traditional catalytic roles as nucleophiles, electrophiles, or acid/base catalysts. A complementary approach, utilizing modified substrates in which each hydroxyl group has been replaced by a hydrogen or fluorine, has been used to probe the importance of non-covalent interactions in several Chapter 1 General Introduction 22 enzymes that utilize sugar-based substrates including glycogen phosphorylase,82 phosphoglucomutase,83 E. coli (3-galactosidase,84 intestinal lactase,85 and Agrobacterium sp. |3-glucosidase (ABG).86 In the ABG substrate complex, contacts with each hydroxyl group of the substrate are fairly weak (<0.8 kcal/mol) and are not critical for the binding of the substrate. During the chemical steps however, these interactions are heightened considerably with the 3- and 6-hydroxyl groups each providing 2.2 kcal/mol towards stabilization of the transition state. The interaction of the 2-hydroxyl group is particularly strong and provides 5 kcal/mol worth of transition state stabilization. In E. coli P-galactosidase and Aspergillus wentii P-glucosidase interactions with the 2 position contribute > 8 kcal/mol and > 7 kcal/mol, respectively, towards transition state stabilization. Another particularly important interaction is seen in HEWL, where replacement of the 6 hydroxyl of the reducing end sugar of the trisaccharide glycoside substrate (Figure 1.11) with hydrogen results in a 1300 fold rate reduction relative to the hydroxylated substrate. Figure 1.11 3,4-Dinitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranosyl-(1^4)-2-acetamido-2-deoxy-P-D-glucopyranosyl-(1^4)-2-acetarnido-2-deoxy-a-L-fucopyranoside. The arrow indicates the deleted hydroxyl group. Another important and common non-covalent interaction is found in the interactions between certain glycosidases and their occasionally large oligosaccharide substrates. As indicated earlier, glycosidases may have several subsites, each one acting to bind a saccharide residue of a polysaccharide substrate. These subsites are often very important for ground state binding and for formation of the Michaelis complex. Indeed, these distant interactions can provide the finely tuned substrate specificity found in some glycosidases. However, in addition to their contribution to ground state binding, these Chapter 1 General Introduction 23 subsites also often provide enormous stabilization of the transition state.87 Although Wolfenden and coworkers are investigating the thermodynamic basis for this effect in other enzymes88 the precise mechanism by which distant subsites contribute to transition state stabilization remains one of the great mysteries of enzyme catalysis and is certainly not limited to glycosidases. 1.2 Tools for the study of glycosidases Glycosidases have been the focus of considerable mechanistic investigations. Indeed, the structure of HEWL, the first enzyme structure determined, stimulated a tremendous amount of experimentation specifically on the catalytic mechanism of (3-retaining glycosidases. In accord with the huge body of literature that has developed on this topic a large number of different investigative methods have been applied to solving the catalytic mechanisms of these enzymes. A brief summary of the chemical methods employed in this study is provided below and some salient examples from the literature are discussed. This description should provide the non-specialist reader with an adequate background for a critical reading of the following chapters and is not intended as an exhaustive or even a thorough review of the literature. For a rigorous discussion of the literature regarding the mechanistic enzymology of glycosidases, curious readers are directed to the excellent reviews by Capon (1969),29 Sinnott (1990),49 McCarter and Withers (1994),12 Davies, Withers, and Sinnott (1998),11 and Zechel and Withers (2000).10 Perusal of these reviews will also provide the reader with a historical perspective on the evolution in the understanding of glycosidase mechanisms. 1.2.1 Competi t ive inhibitors The insights of Pauling laid the foundation for the development of transition state analogues as potent enzyme inhibitors. At the center of Pauling's simple and elegant hypothesis resides the assumption that at the transition state many interactions within the Chapter 1 General Introduction 24_ activated complex are optimized.79 Therefore, analogues that mimic the structure of the fleeting transition state should bind the enzyme tightly. This idea has come to be known as the concept of transition state analogy and was first espoused by Pauling in 1946 and later but more rigorously and independently by Wolfenden89 and Lienhard.90 A quantitative approach to the evaluation of whether an inhibitor is a transition state analogue has been developed although the detailed basis for this theory remains murky.91-92 Key design features of glycosidase inhibitors include; positive charge at 0-5, positive charge at C-l , a sp2 hybridized anomeric center, and a half-chair or envelope conformation. In short, any feature of the oxocarbenium ion-like transition state that is incorporated into the inhibitor typically increases its affinity. The inhibition of glycosidases by gluconolactone (1.1) was first observed by Ezaki who noted marked inhibition of P-glucosidase from almonds and Aspergillus (Taka-diastase).93 The closely related gluconolactam (1.2) is also an inhibitor of many retaining fi-glucosidases and has the added benefit of being stable in aqueous solution.94 Despite the similarities of their half-chair conformations and charge distributions to the oxocarbenium ion-like transition state these compounds typically bind only 100 to 1000 times more tightly than does the parent sugar.95 Nature offered up more potent glycosidase inhibitors when, in 1965, nojirimycin (1.3) was isolated from Streptomyces roseochromogenes R-468.96 The naturally occurring nojirimycin and 1-deoxy-nojirimycin97 (1.4) are both potent glucosidase inhibitors. They bear a nitrogen atom in the place of the endocyclic oxygen that presumably confers a positive charge on these molecules when they bind in the active site. They have long been supposed to mimic the positive charge on the ring oxygen of the oxocarbenium ion-like transition state although results from Namchuck and Withers suggest that these inhibitors are not transition state analogues.98 Nature has evolved a still more potent class of inhibitor in which the basic nitrogen has replaced the anomeric carbon and not the endocyclic oxygen as was the case for nojirimycin (1.3). Siastatin B 9 9 (1.5) was found to be a very potent inhibitor of N-acetyl-neuraminidase and this discovery prompted the synthesis of isofagomine (1.6). 1 0 0 Isofagomine and this class of nitrogen containing inhibitors are now known as the most potent inhibitors of retaining (3-glycosidases.100 The only transition state feature that these molecules bear is a positive charge on the protonated ring nitrogen that has replaced Chapter 1 General Introduction 25 the anomeric carbon. However, this shortcoming does not impede their potency as inhibitors and they are often found to bind 10000 to 100000 times more tightly than the parent sugar they emulate.101 A further refinement on the isofagomine class of inhibitors involves the inclusion of the 2-hydroxyl group into the inhibitor. The incorporation of this moiety to generate a compound that has been named Noeuromycin1 0 2 (1.7) reiterates the importance of the 2-hydroxy group in catalysis (vide supra). Another effective class of competitive inhibitors has also evolved from the inspiration of Nature. Nagstatin 1 0 3' 1 0 4 (1.8) was found to be a highly potent inhibitor of certain ^-hexosaminidases and prompted the independent groups of Tatsuta and Vassella to prepare a series of synthetic nojirimycin imidazoles 1 0 5' 1 0 6 (1.8 and 1.9) and triazoles 1 0 7- 1 0 8 (1.11 and 1.12). Vasella had already, prior to the discovery of Nagstatin, devised the closely related tetrazole inhibitors109 (1.10). All of these inhibitors incorporate features of Nagstatin and are relatively potent inhibitors of many different glycosidases.54 The generally observed decrease in inhibitory activity on going from the imidazole class (1.8 and 1.9) to the tetrazole class (1.10) of inhibitor is one case where adding more nitrogen atoms has not proven to be advantageous.54 Indeed, this is one case where fewer nitrogen atoms in an inhibitor work better. The imidazole 1.9 is a better inhibitor than triazole 1.11, which is still better than tetrazole 1.10. This difference is believed, however, to stem from increasing basicity of the azole ring that favors a charge-dipole interaction with the catalytic nucleophile.54 Shuffling nitrogen atoms also does not increase inhibitory potency as was found for the nojirimycin (1.3) and isofagomine (1.6) class of inhibitors. In fact, as the reader may recall the 1,2,3-triazoles (1.12) are much less potent inhibitors of anti protonating enzymes as compared to the 1,3,4-triazoles (l . l l) . 5 3 Chapter 1 General Introduction 26 °H / > % „0H H O — V ^ - 0 \ ^ 0 H 0 . C H t - O H r 0 H H 0 H C ^ ^ H 0 H - c 5 ^ OH OH OH " O H OH Gluconolactone Gluconolactam Nojirimycin 1-Deoxy-nojirimyciin 1.1 1.2 1.3 1.4 OH COOH OH < 0 H H 0 A - ^ O N H " - H O \ ^ ^ N H " H 0 \ ^ \ H NHAc OH Siastatin B Isofagomine Noeuromycin 1 5 1.6 1.7 NHAc OH OH Nagstatin 1,4-Glucoimidazole Glucotetrazole 1.8 1.9 1-10 0 H - ^ 0 H K , - N . N -»K H O - — f - ^ ^ N v HO H O - \ ^ ^ = N H 0 OH OH 1,3,4-Glucotriazole 1,2,3-Glucotriazole 1.11 1 .12 Figure 1.12 Examples of some competitive inhibitors illustrating the incorporation of certain features of the oxocarbenium ion-like transition state into transition state analogues for glycosidases. 1.2.2 Affinity Labels Affinity labels have proved to be of use in studies designed to identify residues that form, or are near, the catalytic machinery of many glycosidases. In many cases where the three dimensional structure of the studied protein is unavailable or the active site has not been identified these compounds occupy an important place in the arsenal of the mechanistic enzymologist. Affinity labels for glycosidases are composed of two parts; a sugar moiety that provides the specificity for the enzyme active site and the reactive group that can form adducts with nucleophilic residues. Owing to their inherent reactivity there are several cases where affinity labels have been found to label multiple residues within a single protein. Thus, other than indicating what residues are close to the active site they are of limited use in assigning the catalytic role of enzymic residues and are best used in tandem with site-directed mutagenesis. Better-known examples of Chapter 1 General Introduction 27 these inactivators include the epoxyalkyl glycosides and N-bromoacetyl glycosylamines (Figure 1.14). The first use of the epoxyalkyl glycosides somewhat appropriately involved the inactivation of H E W L using 2',3'-epoxypropyl-chitobioside (Figure 1.13). 1 1 0 Figure 1.13 Chemical structure of 2 ',3 "-epoxypropyl-chitobioside. In general, the epoxyalkyl glycosides (1.13 and 1.14) have enjoyed wider use than the N-bromoacetyl glycosylamines (1.15), having been used to identify catalytic residues in several glycosidases including the cellulases from Bacillus macerans,ni Bacillus amyloliquefaciens,112 and Oxyporus s p . 1 1 3 The three dimensional structures of several epoxyalkyl glycoside-inactivated glycosidases have also been solved including those from H E W L , 1 1 4 human lysozyme, 1 1 5 Fusarium oxysporum E G 1 , 1 1 6 Bacillus macerans,xn and Trichoderma reesei.ni In some cases the labeled residue was later determined to act as the acid/base catalytic residue while in other cases the enzymic nucleophile was labeled. The N-bromoacetyl glycosylamines have enjoyed some use in the labeling of the acid/base catalytic residue of Cellulomonas fimi P-xylanase using N -bromoacetyl-P-D-cellobiosylamine 1 1 8 and the acid/base catalytic residue of the P-glucosidase from Cassava using N-bromoacetyl-p-D-glucosylamine (1.15).119 The use of the N-bromoacetyl class of affinity labels has not, however, been without problems. When E. coli P-galactosidase was incubated with N-bromoacetyl-P-D-galactosylamine a methionine at the mouth of the active site was labe led 1 2 0 and Agrobacterium sp. P-glucosidase was multiply labeled by N-bromoacetyl-P-D-glucosylamine. 1 2 1 1.13 Chapter 1 General Introduction 28 HO 'O 1.14 1.15 1.16 1.17 Figure 1.14 Affinity labels for glycosidases. (1.14) epoxyl-alkyl P-D-glucoside; (1.15) N-bromoacetyl P-D-glucosylamine; (1.16) oc-C-bromo-ketone-mannoside; (1.17) l,2-epoxy-3-(a-D-glucopyranosyl)propane. Recent arrivals to the armoury of affinity labels include the oc-C-linked bromo-ketone-mannoside (1.16) that was used to identify the acid/base catalytic residue of the cc-glucosidase from Saccharomyces cervisiae122 and the C-linked l,2-epoxy-3-(a-D-glucopyranosyl)propane (1.17) that was used to inactivate two different cc-glucosyltransferases both from Streptococcus mutans.12^ These two compounds represent the first generation of affinity labels available for a-glycosidases. The Stick group at the University of Western Australia has prepared a number of p-C-linked epoxyalkyl compounds as potential labels for glycosidases124-125 although these compounds have not yet been tested. All of these C-linked affinity labels are inert to cleavage by glycosidases, providing a distinct advantage over the O-linked epoxyalkyl glycosides, which are readily hydrolyzed by glycosidases. 1.2.3 Mechanism based inactivators For the identification of key active site residues in glycosidases, an approach using affinity labels may be considered as the fundamental equivalent to archery; useful but not always accurate. In this context mechanism-based inhibitors must be considered as riflery, or in some cases, even laser technology. The key design feature of mechanism-based inactivators is that they require some form of activation that can be obtained only in the active site. The typical trigger with mechanism-based glycosidase inactivators involves protonation of the inactivator to increase its electrophilic character followed by Chapter 1 General Introduction 29 nucleophilic attack of an enzymic residue at the activated electrophilic center. Three major designs exist for mechanism-based inactivators of P-retaining glycosidases. The substituted cyclohexane oxides (1.18) developed by Legler, 9 5 the triazenes (1.19) created by Sinnott, 1 2 6 and the fluoro sugar approach conceived by Withers (1.20).127 Figure 1.15 Representative members of the best-known classes of mechanism based inactivators of f3-retaining glycosidases. 1.2.3.1 Cyclohexane Epoxides Humans have often styled their creations after the designs of Nature. Occasionally, however, human creations precede the discovery of the related blueprint within Nature. The development of the cyclohexane epoxide class of mechanism based inactivators of retaining P-glycosidases is one such case. Conduritol B epoxide 1 2 8 (1.18) was devised prior to the knowledge of the existence of the closely related natural product cyclophellitol (1.21) that is isolated from the fungus Phellinus sp. 1 2 9 The conduritol B epoxide was designed to mimic a P-D-glucopyranoside glycone, providing affinity for the active site of glucosidases. The orientation of the epoxide presumably positions the ring oxygen in a similar position to that of the glycosidic oxygen of a normal p-glucoside substrate. Owing to the symmetry of conduritol epoxide B, however, it inactivates both a- and P-glucosidases. While conduritol B epoxide lacks a hydroxymethyl group at the 5-position the natural product cyclophellitol bears this group making it a more faithful mimic of a P-glucopyranoside and breaks the symmetry of the molecule. As a consequence of this small difference in structure, cyclophellitol is highly specific for P-glucosidases and does not inactivate a-glucosidases. In both cases it is expected that 1.18 1.19 1.20 Chapter 1 General Introduction 30 activation of the inactivator occurs by the protonation of the epoxide by the catalytic machinery of the glycosidase, particularly by the acid/base catalytic residue. Once activated, the nucleophile residue attacks the epoxide at the center corresponding to the anomeric carbon of a regular substrate resulting in an alkylated enzyme (Scheme 1.3). ENZ Scheme 1.3 Ideal mechanism of inactivation of a retaining P-glucosidase by cyclophellitol. While these epoxide based inactivators have been found to correctly label the nucleophilic residue in the P-glucosidases from Aspergillus wentiin% and the archaeon Sulfolobus solfataricus130 they have not met with uniform success. Indeed, when the appropriately hydroxylated cyclohexane epoxides were incubated with E. coli |3-galactosidase131 and sweet almond P-glucosidase132 it is known that in neither case was the catalytic nucleophile alkylated but rather another carboxyl group in the active site. 1.2.3.2 Glycosylmethyltriazenes Another, more chemically sophisticated approach involves the decomposition of glycosylmethyltriazenes to generate extremely reactive carbenium ions adjacent to the active site machinery of P-retaining glycosidases. The careful design of these P-C linked glycosides involves the inclusion of the electron withdrawing /?-nitrophenyl moiety to ensure that the compound does not spontaneously decompose.126 This group also allows the reaction to be readily monitored by spectrophotometry. When protonated, these compounds rapidly decompose to yield a carbenium ion (Scheme 1.4) therefore limiting Chapter 1 General Introduction 31 their use in solutions where the pH is acidic (pH < 5). Barring this limitation they have been demonstrated to be of use although they do not consistently label catalytic residues perhaps because the site of protonation is somewhat far removed from the normal protonation trajectory. Although not investigated, it is interesting to speculate that these two classes of inactivator (the cyclohexane epoxides and the triazenes) may be successful only with enzymes having one of the two possible protonation trajectories. ENZ ENZ Scheme 1.4 Ideal mode of inactivation of a P-retaining galactosidase by galactosylmethyl-(4-nitrophenyl)triazene. 1.2.3.3 2-and 5-Fluoro Sugars By far the most successful mechanism based inactivators are the fluoro sugar inhibitors that act to form a stabilized glycosyl-enzyme intermediate, which turns over only very slowly. These compounds are designed on the understanding that the putative covalent glycosyl-enzyme intermediate is bracketed by transition states both of which have considerable oxocarbenium ion-like character.45 The incorporation of a highly electron withdrawing fluorine substituent adjacent to the site of positive charge Chapter 1 General Introduction 32 development therefore slows both steps of the reaction by destabilizing both transition states.133 Such destabilization is not beneficial on its own as the result is simply that the overall reaction is slowed. However, upon incorporation of a good leaving group such as fluoride ion or 2,4-dinitrophenolate the barrier to the formation of the intermediate is lowered while the barrier for its breakdown is unaffected. When these two substitutions are both incorporated into the inhibitor the net result is the generation of a relatively stable glycosyl-enzyme intermediate that breaks down only very slowly (Scheme 1.5). Scheme 1.5 Mechanism of action of 2-deoxy-2-fluoro-P-D-glucosyl fluoride on a retaining P-glucosidase. The first generation of such mechanism based inhibitors were the 2-deoxy-2-fluoro-P-D-glycosyl fluorides133 and the 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-glycosides (Figure 1.16).134 The 2-fluoroglycosyl-enzymes formed from these compounds have been shown, first by using 1 9 F NMR 1 3 5 and later, X-ray crystallography,60 to be a-linked to a carboxyl residue of the enzyme (Scheme 1.5).136 The replacement of the 2-hydroxyl with an "isosteric" 2-fluorine substituent has an additional effect on the energy of the transition states as alluded to earlier. The 2-hydroxyl is of great importance for the stabilization of the transition states in the P-retaining (3-glycosidase catalysed hydrolysis of glycosides and by replacing this group with fluorine these important contacts are lost in the transition state.10 The fluorine group therefore destabilizes the transition state both through its significant inductive effects and also by removing these important interactions. Thus the glycosyl-enzyme intermediates Chapter 1 General Introduction 33 formed using the 2-fluoro strategy are typically very stable, having half-lives ranging from 5 to 100 hours.127 The great stability of these intermediates allows the site of attachment to be identified by proteolytic digestion and sequencing of the isolated, labeled peptide.136 Regardless of the stability of these intermediates they have all been shown to be kinetically competent, slowly reactivating by the normal mechanism involving attack of water or another acceptor at the anomeric centre to yield the regenerated enzyme. Interestingly, the 2-fluoro approach fails with a-retaining enzymes. Instead these reagents act as slow substrates for which the rate-determining step is formation of the glycosyl enzyme intermediate. In order to further slow the glycosylation step, however, complementary approaches have been devised. In one strategy, the incorporation of an additional fluorine at the 2 position and the simultaneous incorporation of a still better leaving group such as 2,4,6-trinitrophenyl allows these 2-deoxy-2,2-difluoro compounds (1.22, Figure 1.16) to form extremely stable glycosyl enzyme intermediates on a-retaining enzymes. The other, more delicate, approach involves the incorporation of a fluorine substituent adjacent to the ring oxygen that is expected to bear considerable positive charge in the transition state. These 5-fluoro-glycosyl fluorides (1.23) have been found to be effective at trapping the glycosyl enzyme intermediate of both a- and P-retaining glycosidases (Figure 1.16). Interestingly the C-5 epimers (1.24) of the normal glycone are typically better inhibitors than the "correctly" configured glycone and form glycosyl enzyme intermediates that are significantly more stable (Figure 1.16). As in the case of the 2-fluoro compounds this effect may stem from further destabilization of both transition states as a result of the axial hydroxymethyl group. Figure 1.16 Fluoro sugar inactivators capable of labeling retaining a-glycosidases. Chapter 1 General Introduction 34 1.2.4 Computational Analysis of the Transition State by Multiple Isotope Effects While affinity labels, inhibitors, and mechanism based inactivators can provide useful information regarding the identity of key catalytic residues, they yield only limited insight into the nature of the transition states. Schramm and coworkers have developed a sophisticated approach to evaluating the geometry of the transition states.136b By measuring a number of primary, cc-secondary, and -^secondary kinetic isotope effects, insight can be gained into the environment of individual atoms that are in proximity to the reactive center. In the case of some N-ribosyl and N-deoxyribosyl transfer reactions such measurements have provided sufficient information to solve the geometry of their transition states, including the extent of participation of both the incoming nucleophile and leaving group.136b In the most favorable situations, the structure of the transition state can be determined to less than 1 A . 1 3 6 b Such calculations, however, are complicated in so far as they require precise knowledge of the structure of the Michaelis complex. Additionally, transition state structures so determined are only models and as such, require careful interpretation. Regardless of the technical pitfalls of this approach, there have been several remarkable successes.136b Indeed, the models of the transition states for the reactions catalyzed by several N-glycosylases have been determined and on their basis several potent transition state analogues have been developed.136b Without question this approach will increase in popularity as the interpretation of isotope effects advances. 1.3 Aims of this Thesis Despite the enormous body of literature published on the mechanism of (3-retaining glycosidases, there has been no universally accepted theory as to their precise mechanism of action. All such enzymes, however, catalyze a fundamentally similar reaction; the hydrolysis of the equatorial glycosidic linkage of a hydroxylated pyranose with net retention of configuration at the anomeric center. Thus, it is not unreasonable to speculate that all of these enzymes use a common catalytic mechanism involving very similar intermediates and transition states. Indeed, as discussed earlier, every aD-KIE Chapter 1 General Introduction 35 measured to date on a retaining (3-glycosidase indicates that each transition state has significant oxocarbenium ion-like character. The aims of this thesis are to elucidate the detailed mechanisms of several retaining P-glycosidases that are either suspected to, or currently believed to, utilize disparate catalytic mechanisms. The particular enzymes selected for these studies are all known to be retaining glycosidases and include the Thermoanaerobacterium saccharolyticum P-xylosidase from family 39, the Streptomyces plicatus P-hexosaminidase from family 20, a novel, recently cloned, unclassified Vibrio furnisii P-hexosaminidase, and the well known hen egg white lysozyme from family 22. Glycosidases from family 11 that cleave pentopyranosides have been postulated to direct reactions through a transition state with a different geometry than do glycosidases that cleave hexopyranoside substrates.137 The P-xylosidase from family 39 also cleaves pentopyranosides substrates and so we have decided to investigate its mechanism. Some reports in the literature have suggested the possibility that the catalytic mechanism of P-hexosaminidases,27 and HEWL, 2 4 may involve anchimeric assistance from the 2-acetamido substituent contained within the substrates of these enzymes. Thus we have opted to study the P-hexosaminidase from family 20, the novel P-N-acetyl-glucosaminidase, and HEWL to investigate this possibility. Lastly, the Phillips mechanism for HEWL, which has long been considered the paradigmatic mechanism of P-retaining glycosidases, is considered to proceed via an ion-pair intermediate and thus any attempt at the formulation of a general catalytic mechanism for these enzymes necessitates a reevaluation of the catalytic mechanism of HEWL. From these studies we hope to identify underlying commonalities that may clarify the catalytic mechanisms of these particular enzymes and also allow the formulation of a generally applicable and detailed catalytic mechanism for the whole class of P-retaining glycosidases. Chapter 2 Mechanism ofT. saccharolyticum B-Xvlosidase 36 2 Mechanism of Thermoanaerobacterium saccharolyticum p-Xylosidase 2.1 Background of Retaining p-xylosidases The major component of plant cell wall hemicelluloses is the polymer xylan. This molecule is composed of a backbone of (3-1,4-linked D-xylopyranosyl units decorated with a variety of sugars including L-arabinose and D-glucuronic acid (Figure 2.1). RO—V -°. R= xylan chain OR 'OH Figure 2.1 Structure of xylan showing the (3-1,4-linked D-xylopyranosyl backbone decorated with the pendant L-arabinose and D-glucuronic acid residues. On a practical level the enzymatic degradation of xylans has attracted the attention of the pulp and paper industry because it has been shown that the "bio-bleaching" of paper-pulp is effective in terms of decreasing both cost and environmental impact.138 The complete digestion of xylan requires the action of many enzymes including xylanases, which hydrolyze the xylan backbone to yield shorter, soluble oligosaccharides, and P-xylosidases that liberate xylose from the non-reducing termini of these soluble xylo-oligosaccharides. The focus of this chapter will be the elucidation of the detailed mechanism of a retaining P-xylosidase. On a more fundamental level, P-xylosidases have been the subject of few mechanistic studies and so relatively little is known about their detailed mechanism of action. They are currently classified into families 3, 39, 43, 52, and 54 of glycoside hydrolases.139 The enzymes from family 43 use an inverting mechanism while all other Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 37 known families of P-xylosidase are believed to use a retaining mechanism.139 Detailed mechanistic studies of P-xylosidases have been limited to the inverting enzymes from family 43,140>141 and recently, the retaining p-xylosidase from Trichoderma koningii G-39 that has been classified into family 54 of glycoside hydrolases.142 Li and coworkers found that the cc-deuterium secondary kinetic isotope effect (aD-KIE) measured on the second step (deglycosylation) of the T. koningii catalyzed hydrolysis of 2,4-dinitrophenyl P-D-xylopyranoside (DNPX, Figure 2.2) was 1.02 ± 0.01. Figure 2.2 Structure of 2,4-dinitrophenyl P-D-xylopyranoside This small value suggests that the transition states of the catalytic mechanism for this enzyme may be very SN2-like. If this result can be validated it represents a break from precedent as all other glycosidases studied to date have been found to operate by a very dissociative SN2-like transition state. Indeed, to the best of this authors' knowledge, in no case has such a small aD-KIE been measured on any chemical step of a glycosidase-catalyzed reaction. This aD-KIE value is so close to unity that it is somewhat ambiguous, and alternative interpretations cannot be entirely ruled out on the basis of the study by Li et al. One alternative explanation for this low KIE value is that the chemical step is, in fact, not rate limiting for DNPX and another step such as substrate diffusion, a protein conformational change, or proton transfer is rate determining. This author feels that it is most likely, despite Li's results, that the catalytic mechanism of retaining P-xylosidases is in keeping with the 'exploded' SN2 (Scheme 2.1) transition states determined for many other glycosidases (as outlined in the introduction, vide supra) and not a strict SN2 transition states. Neither the reports of Li et al, suggesting a strict SN2-like transition state, nor the ion pair hypothesis originally posited by Phillips, however, can be discounted in the absence of firm evidence. Consequently, the nature of the Chapter 2 Mechanism ofT. saccharolyticum B-Xvlosidase 38 intermediate and the precise mechanism in this class of enzyme remains unclear and requires clarification. ENZ XYLOSYLATION STEP XYLOSYL-ENZYME INTERMEDIATE H ENZ >=o H H HO HO DEXYLOSYLATION STEP HO HO O O 0 H ^ E N Z I 0 - ^ H \ 0 ^ ° ENZ ENZ / Q >=o H .H 5 - Q O ENZ Scheme 2.1 Hypothetical mechanism of retaining P-xylosidases involving a covalent xylosyl enzyme intermediate. Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 39 The retaining xylanases from family 11 of glycosyl hydrolases are a functionally related class of enzyme. The retaining |3-xylosidases are exo-acting pentopyranosidases while the xylanases are endo-acting retaining P-xylosidases. The xylanases have been much more rigorously studied than the P-xylosidases and one noteworthy feature is the geometry of the 2-fluoroxylobiosyl enzyme intermediate. The proximal saccharide moiety of the 2-fluoroxylobiosyl enzyme intermediates of the family 11 xylanases from Bacillus circulans131 and Bacillus agaradhaerens143 does not adopt a AC\ conformation as has been observed for all other glycosyl enzyme intermediates11 including the bifunctional xylanase/cellulase from family 10 (Figure 2.4).60 Instead, the pyranose ring bound in the -1 subsite of the xylanase adopts a 2'5B conformation (Figure 2.3). Figure 2.3 The pyranose ring of glycosyl enzyme intermediates bound in the -1 subsite adopts a (a) 25B conformation as seen in the family 11 xylanases or (b) the 4Ci conformation seen in all other covalent glycosyl enzyme intermediates. This unusual conformation suggests that a mechanism distinct from other glycosidases may operate for these family 11 xylanases. Indeed, the 25B intermediate observed here already has the C-2, C-l , 0-5, and C-5 atoms in a coplanar arrangement suggesting that the geometry of the transition state for these enzymes may differ from that of typical retaining P-glycosidases. Should a different mechanism operate for these enzymes it may be a consequence of the properties of their substrates. Xylosides, owing to the absence of a hydroxymethyl group at the 5 position, have greater conformational flexibility than many hexopyranosides (Figure 2.4). Chapter 2 Mechanism ofT. saccharolyticum (i-Xylosidase 40 a b HO' M U — , H 0 ^ « * - ^ \ - - - -c HO HO HO OR HO. OR OH OH Figure 2.4 Structures of; (a) P-D-xyloside, (b) 3-D-glucoside, (c) a-L-iduronide. Although no glycosyl enzyme intermediate has yet been structurally characterized for a retaining P-xylosidase one can suppose that, owing to the similarity of their substrates, their catalytic mechanism may be very similar to that of the family 11 xylanases. If this premise is true it is easy to speculate that this geometric difference may account for the atypical isotope effects observed by Li et al. In light of the conformation of the glycosyl enzyme intermediate in the family 11 xylanases and the unusual aD-KIE value measured for the dexylosylation step of the T. koningii P-xylosidase the precise catalytic mechanism of the retaining P-xylosidases remains uncertain. Does the enzyme-catalyzed reaction involve a covalent intermediate, as precedent from other studies suggests and as this author expects, an ion pair intermediate as is still invoked in the literature for p-retaining glycosidases,144'146 or an SN2-like mechanism as suggested by Li et all142 Family 3 9 of glycoside hydrolases, of which Thermoanaerobacterium saccharolyticum P-xylosidase (XynB) is a member, contains both bacterial P-D-xylosidases and mammalian oc-L-iduronidases.147 All members of this family, by extension from stereochemical studies with T. saccharolyticum P-xylosidase, utilize a retaining catalytic mechanism.148 The inclusion of both xylosidases and iduronidases within this same family is, on first consideration, somewhat surprising as iduronide substrates bear a carboxyl group at the 5 position while xylosides bear only a hydrogen substituent (Figure 2 .4) . The common thread is perhaps that both of these sugars are known to have significant conformational flexibility thereby making it conceivable that this feature plays some role in a common catalytic mechanism. One remarkable property of the Thermoanaerobacterium saccharolyticum P-xylosidase is its selectivity in partitioning of the glycosyl enzyme intermediate. The typical route, hydrolysis, leading to the breakdown of a glycosyl enzyme intermediate Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 41 involves the attack of water at the anomeric center. In such a role the incoming water molecule is defined as an acceptor of the saccharide moiety because the sugar is transferred to the water molecule. Other acceptors such as alcohols or thiols can also intercept the intermediate to generate a glycoside or thio-glycoside. Indeed, certain glycosidases are particularly efficient at promoting the collapse of the glycosyl enzyme intermediate using a saccharide acceptor to generate a polysaccharide product (Scheme 2.2) and these enzymes are termed transglycosylases. Such a reaction is specifically termed transglycosylation and this process is the microscopic reverse of the first step of the reaction (glycosylation). ENZ ENZ Scheme 2.2 Transglycosylation reaction shown for a retaining P-xylosidase with xylose as an acceptor. 2.2 Objectives of this Work We hope to unambiguously establish the mechanism of T. saccharolyticum (3-xylosidase and, by extension, the other enzymes within family 39 of glycoside hydrolases. Specifically, we will attempt to; (i) identify the catalytic nucleophile using a mechanism based inhibitor, (ii) establish the nature of the glycosyl enzyme intermediate (covalent or ion-pair) using detailed steady state kinetic analysis of the enzyme, Chapter 2 Mechanism ofT. saccharolyticum 3-Xylosidase 42 (iii) identify the acid/base catalytic residue using an affinity label in conjunction with site directed mutagenesis and detailed mechanistic studies, (iv) establish the importance to catalysis of the identified residues by detailed enzyme kinetic studies of site directed mutants, These results will serve to clarify the catalytic mechanism of this poorly investigated class of enzyme and also to test the underlying hypothesis of this thesis; are the family 39 retaining P-xylosidases simply one of many retaining P-glycosidase families that use the same catalytic mechanism? 2.3 Identification of the Catalytic Nucleophile In this section we hope to identify the catalytic residue of T. saccharolyticum P-xylosidase (XynB) that acts either in a nucleophilic capacity, to form a covalent glycosyl enzyme intermediate, or simply as a counter ion, to stabilize a long lived ion-pair glycosyl enzyme intermediate. After establishing the identity of this residue we intend to delete this group using site directed mutagenesis in order to evaluate its importance in the catalytic mechanism. 2,4-Dinitrophenyl 2-deoxy-2-fluoro-p-D-glycosides have proven extremely successful reagents for derivatisation of active site nucleophiles in several P-retaining glycosidases.149"151 These inactivators are mechanism-based as the C-2 fluorine destabilizes the transition states for both the glycosylation and deglycosylation steps while the excellent leaving group accelerates the first step (xylosylation) so as to permit trapping of a 2-deoxy-2-fluoro-cc-D-glycosyl enzyme intermediate.134 Digestion of the labeled enzyme using a non-specific protease yields a mixture of short peptides one of which has the fluorosugar attached (Scheme 2.3). Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 43 Activated 2-fluorosugar Non-specific Proteolysis Scheme 2.3 Simplified scheme showing the labeling and proteolysis to generate a.mixture of peptides, one of which bears the label. Identification of the peptide bearing the label can be accomplished using neutral loss tandem mass spectrometry.149-150 This well established technique involves the use of an instrument that has a minimum of three quadrupoles. The first analyzer scans through the mass range permitting a range of ions to enter into the second mass analyzer. This second quadrupole acts as a collision chamber where the analyte collides with an inert gas, resulting in the homolytic fragmentation of the analyte. In the work conducted here the glycosyl enzyme linkage, being the sole acylal ester in the complex, is uniquely fragile and fragments more readily than most other bonds, making the neutral loss method possible. The third mass analyzer scans through the mass range at the same rate as the first analyzer but such that it only permits passage of analyte molecules having lost a fragment corresponding to the mass of the label (Figure 2.5). Following isolation of the labeled peptide, it can be sequenced and the identity of the nucleophilic residue can be established by further collision induced fragmentation using tandem mass spectrometric analysis. of X pass through the first mass analyzer. Those peptides bearing the label with a m/z of Y lose this fragment in the collision cell. Only those peptides of mass (X-Y) pass through the third mass analyzer and are detected.. The label is depicted as the solid circle (•). Detector Figure 2.5 Outline of the basic principles of neutral loss mass spectrometry. Peptides with a m/z Chapter 2 Mechanism ofT. saccharolyticum /3-Xvlosidase 44 2.3.1 Inactivation of the Enzyme As expected, incubation of T. saccharolyticum P-xylosidase (XynB) with 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-xylopyranoside (2F-DNPX, Figure 2.8) resulted in psewdo-first-order inactivation of the enzyme in a time-dependent manner (Figure 2.6). Figure 2.6 Inactivation of Thermoanaerobacterium saccharolyticum P-xylosidase by 2F-DNPX. a) Semilogarithmic plot of residual activity vs. time at the indicated inactivator concentrations: (O) 36.5 uM (•) 107 uM (•) 213 uM (•) 320 uM (A) 532 uM. b) Inactivation with 213 uM 2F-DNPX in the absence (•) and presence (O) of 11.3 mM benzyl p-D-thioxylopyranoside (BTX) c) Replot of the first order rate constants from panel a. The observed rate of inactivation of the enzyme with 2F-DNPX was shown to be dependent on the inactivator concentrations in a saturable manner, consistent with inactivation according to the scheme: Chapter 2 Mechanism ofT. saccharolyticum 6-Xvlosidase 45 H , 0 v ^ E + 2F-X-0H K, fc 2^ E + 2F-DNPX = ^ E - 2 F - D N P X —<r*- E-2F-X DNP Sugar 2F-X-Sugar Scheme 2.4 Kinetic scheme outlining the reaction of XynB (E) with 2F-DNPX. K-, is the dissociation constant for 2F-DNPX. k\ is the rate constant governing the inactivation step. kK is the rate constant governing the spontaneous hydrolysis of the 2-fluoroxylosyl enzyme intermediate (E-2F-X) and k^^ is the rate constant governing the reactivation of the enzyme through transglycosylation to a sugar moiety. Analysis of the inactivation data as described in the methods section of this chapter allowed calculation of the inactivation rate constant (k, = 0.089 ± 0.001 min"1) and the dissociation constant (K\ = 66 ± 4 uM) governing the binding of the inactivator to XynB. Incubation of the enzyme (0.68 mg/mL) with 2F-DNPX (213 uM) in the presence of the competitive inhibitor benzyl P-D-thio-xyloside (BTX; 11.33 mM, K, = 5.3 mM) resulted in a decrease in the apparent inactivation rate constant from 0.060 min"1 in the absence of BTX to 0.047 min"1 in its presence (Figure 2.6b). This observed protection from inactivation in the presence of BTX is consistent with these two ligands competing for the same site, strongly suggesting that inactivation is active site directed.135 These results suggest, by analogy to observations with Agrobacterium sp. [3-glucosidase, that inactivation is a consequence of accumulation of a stable, on the time scale of hours, covalent 2-deoxy-2-fluoro-a-D-xylosyl enzyme intermediate.134 This conclusion is supported by the mass spectral analysis of the inactivated enzyme (vide infra). 2.3.2 Catalytic Competence Further evidence supporting the existence of a covalent 2-fluoroxylosyl enzyme intermediate arises from demonstration of the catalytic competence of the trapped intermediate (Figure 2.7). Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 46 -2.2 > e > a -2.6 -3.4 -3.2 -2.4 -2.8 -3.8 -3.6 -3 0 1000 2000 3000 4000 5000 Time (min) Figure 2.7 Reactivation of 2-deoxy-2-fluoroxylosyl xylosidase. Semilogarithmic plot of activity vs. time in buffer alone (•) with 45 mM xylose (O), with 45 mM BTX (•), or with 45 mM xylobiose (V). Following removal of excess inactivator from the labeled enzyme the sample was incubated at 37 °C in the presence of phosphate buffer alone, with xylose (45 mM), xylobiose (45 mM) or benzyl P-D-thio-xyloside (BTX, 45 mM) and the recovery of activity associated with the regeneration of the free enzyme was monitored (Figure 2.7). The reactivation kinetics of the 2-fluoroxylosyl enzyme intermediate, in buffer alone, followed a first order process with an apparent rate constant of krs = 0.00001 min"1 (ti/2 = 67000 min). Rate constants for reactivation by transglycosylation (&trans) were found to be 0.00005 min"1 (t>/2 = 13000 min) with xylose, 0.0008 min"1 (t1/2 = 900 min) with xylobiose and 0.0004 min"1 (U/2 = 1800 min) with BTX. This observed increase in the reactivation rate when the sample is incubated in the presence of xylose (5 fold), xylobiose (75 fold) or BTX (37 fold) over the spontaneous reactivation rate suggests that reactivation is significantly accelerated by transglycosylation to an acceptor sugar. This observation is consistent with the known selectivity of XynB favoring partitioning of the glycosyl enzyme intermediate toward transglycosylation. a b c Figure 2.8 Structures of; (a) xylobiose, (b) benzyl-p%D-thio-xyloside (BTX), and (c) 2F-DNPX Chapter 2 Mechanism of T. saccharolyticum /3-Xvlosidase 47 Furthermore, the variation in the rate of reactivation seen for different acceptor sugars suggests that increased acceptor-enzyme interactions in the aglycone site promote reactivation by transglycosylation presumably by increasing the effective concentration of the "acceptor and stabilizing the transition state leading to glycosyl transfer. This process has been fully analyzed for the Agrobacterium sp. P-glucosidase by Street et al. 134 2.3.3 Stoichiometry of Incorporation of Inactivator by ESMS and Identification of the Labeled Active Site Peptide The mass of the native xylosidase was found by ESMS to be 58666 + 6 Da (Figure 2.9a). After inactivation with 2F-DNPX only one peak with a mass of 58800 + 6 Da was observed (Figure 2.9b). The difference between the native and inactivated enzyme is 134 Da, which is consistent within error, with the addition of a single 2-fluoroxylosyl label (135 Da). S 0) > u 58666 58800 58000 58500 59000 mass (Da) — i 1 1 59500 60000 Figure 2.9 Reconstruct of the mass spectrum of; (a) wild-type XynB, and (b) XynB treated with 532 uM 2F-DNPX for 120 minutes. Chapter 2 Mechanism ofT. saccharolyticum 6-Xvlosidase 48 Peptic hydrolysis of 2F-xylosyl-xylosidase resulted in a mixture of peptides that was separated by reverse-phase HPLC using the ESMS as detector. When scanned in the normal LC/MS mode the total ion chromatograph (TIC) showed a large number of peaks each corresponding to one or more peptides in the digest mixture (Figure 2.10a). The 2F-xylose labeled peptide was then located in a second experiment using a tandem mass spectrometer set up in the neutral loss mode. The resulting neutral loss of the label leaves the peptide with an unchanged charge state but a decrease in mass of 135 Da. Thus, the two quadrupoles were scanned in a linked mode so as to permit only the passage of ions losing a mass of 135 Da in the collision cell. No significant peaks were observed during this experiment. However, when the spectrometer was scanned in the neutral loss tandem MS/MS mode searching for a mass loss of m/z 67.5 a single predominant peak was observed in the total ion chromatogram (Figure 2.10b). No such peak was observed in the neutral loss spectrum of the unlabeled xylosidase (Figure 2.10c). These results indicate that a doubly charged peptide bearing the 2F-xylose label is being preferentially detected. This doubly charged labeled peptide was measured at m/z 1103 (Figure 2.10d) indicating that the mass of the peptide is 2204 [2(1103)-2 H]. Since the mass of the label is 135 the unlabeled peptide must have a mass of 2070 [2204-135+1 H]. Aminolysis of the isolated peptide resulted in a new single peak in the ion chromatogram of m/z 1035.5 (data not shown), while no peak corresponding to the labeled peptide was observed. The loss of mass 135 [1103-1035.5]*2 resulting from aminolysis of the labeled peptide is consistent with the expected mass loss from cleavage of the ester-linked 2-fluoroxylosyl label. Candidate peptides of this mass were identified by inspection of the known amino acid sequence152 of the enzyme and searching for all possible peptides of this mass. Eighteen possible peptides with a mass of 2070 + 2.0 were identified. The peptide was then identified unambiguously by determining the complete amino acid sequence by ESMS/MS. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 49 100 50 100 i 50 J cs > 'S3 p 100 50 0 16 24 32 Time (min) 4 0 48 u ^ U J L | ^ Ana . , . ! 16 24 32 4 0  Time (min) 48 16 24 32 40 Time (min) 11103 48 600 800 1000 1200 m/z (amu) 1400 1600 Figure 2.10 ESMS experiments on peptic digest of Thermoanaerobacterium saccharolyticum p-xylosidase. (a) labeled with 2F-DNPX, TIC in normal MS mode, (b) labeled with 2F-DNPX, TIC in the neutral loss mode, and (c) unlabeled in the neutral loss mode, (d) Mass spectrum of peptide at 39.4 minutes. 2.3.4 Sequencing of peptide Information on the sequence was obtained, without a need for further purification, by additional fragmentation of the peptide of interest (m/z 1103) in the daughter ion scan mode (Figure 2.11). Both the labeled parent ion (m/z 1103) and the unlabeled intact peptide arising from loss of the 2-fluoroxylosyl label (m/z 1036) appear as doubly charged species. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 50 +label 1979 1850 1736 1650 1512 1365 Y" ions 1601 1230 o - - P h e - - H i ^ l e ^ h r ^ l u - - T y r 1051 1165 1409 1546 1659 B ions C/5 CD 100 75 50 25 J xlO rM+2H] + + 1103 1365 [M+2H3++ 1036 1409 840 1546 1165 J 1051 \ ^230 J Iik.Ul > Li -•I 400 800 1200 1600 2000 m/z (amu) Figure 2.11 ESMS/MS daughter ion spectrum of the 2F-DNPX labeled active site peptide {m/z 1103, in the doubly charged state). Observed B and Y" series fragments shown above. Peaks resulting from B ions correspond to fragments II (m/z 227), UK (m/z 355), IIKN (m/z 469), IIKNSH (m/z 693), IIKNSHF (m/z 840), IIKNSHFPN (m/z 1051), IIKNSHFPNL (m/z 1165), IIKNSHFPNLPF (m/z 1409), IIKNSHFPNLPFH (m/z 1546) and IIKNSHFPNLPFHI (m/z 1659). Interestingly, more labeled than unlabeled Y" ions are observed in the spectrum. Peaks arising from Y" ions bearing the label include KNSHFPNLPFHITEY (m/z 1979, NSHFPNLPFHITEY (m/z 1850), SHFPNLPFHITEY(m/z 1736), HFPNLPFHITEY(m/z 1650), FPNLPFHITEY(m/z 1512) and PNLPFHiTEY(m/z 1365). Unlabeled Y" ions include only SHFPNLPFHITEY (m/z 1601) and PNLPFHITEY (m/z 1230). Doubly charged fragments also appear in the spectrum but, aside from the labeled and unlabeled parent ions, are unassigned. This information, in conjunction with the mass of the labeled peptide and the primary sequence of the enzyme, permits identification of the peptide containing the active site nucleophile as IIKNSHFPNLPFHITEY (261-278). Aminolysis, mentioned above, of the Chapter 2 ; Mechanism of T. saccharolyticum 0-Xylosidase 51 peptide strongly suggests that the label is attached via a covalent ester linkage. On the basis of analogy to other glycosyl hydrolases, in which the nucleophile has been, without exception, an acidic residue,11 it is likely that Glu277 is the catalytic nucleophile since this is the only carboxylic acid present. Sequence alignments and site directed mutagenesis of this residue should be able to provide conclusive evidence confirming or negating this assignment. 2.3.5 Sequence Conservat ion and Al ignment Alignment of the amino acid sequences of functionally related proteins is commonly used to establish the identity of highly conserved residues. Often, completely conserved residues have some catalytic function. Indeed, among glycosidases the catalytic nucleophile and acid/base residue are nearly always conserved throughout all members of a family. Using the Basic Local Alignment Search Tool (BLAST) 1 5 3 the amino acid sequence incorporating the nucleophile and flanking regions showed significant similarity only to Caldocellum saccharolyticum (3-xylosidase and little similarity to other members of family 39 (Figure 2.12). In light of the relatively poor alignments of the sequence containing the nucleophile and flanking regions, inclusion of both cc-L-iduronidases and p-D-xylosidases into family 39 requires further consideration. Indeed, the overall sequence similarity to iduronidases in this family is less than 16%. Interestingly, it is primarily a few discrete regions of high similarity between family members that account for this overall sequence similarity. One notable region of high similarity occurs approximately 100 residues before the catalytic nucleophile and contains a glutamic acid in the sequence Asn-Glu-Pro. This sequence is reminiscent of the Asn-Glu-Pro sequence154 within which the acid/base catalytic residue is found in several other (3-glycosidases including, for example, Agrobacterium sp. (3-glucosidase from family l 1 5 5 and the family 10 xylanases.156 This would suggest that Glu 160 is the aciaVbase catalyst as suggested previously.147 Chapter 2 Mechanism ofT: saccharolyticum 0-Xylosidase 52 xyrib_thesa xynb_calsa idua_human idua_mouse idua_canfa consensus 82 IDRIFDSFLEIGIRPFVEIGFMPKKIASGTQTVFYVffiGNVTPPKDYEKWSDLVKAVLHHF 8 5 IDSIIDFLLEIGMKPFIELSFMPEAIASGTKTVFHYKGNITPPKSYEEWGQLIEELARHL 114 LDGYLDLLRENQLLPGFELMGSA SGHFTDFEDK QQVFEWKDLVSSLARRY 104 LDAFLDLLMENQLLPGFELMGSP. . . .SGYFTDFDDK QQVFEWKDLVSLLARRY 113 LDGYLDLLRENQLLPGFELMGSP. . . .SQRFTDFEDK RQVLAWKELVSLLARRY * * * * * * * * * * xynb_thesa xynb_calsa idua_human idua_mouse idua_canfa consensus O 142 ISRYGIEEVLKWPFEIWNEPNLKEF..WKDADEKEYFKLYKVTAKAIKEVNENLKVGGPA 14 5 ISRYGKNEVREWFFEVWNEPNLKDF.FWAGTME.EYFKLYKYAAFAIKKVDSELRVGGPA 164 IGRYGIAHVSKWNFETWNEPDHHDFDNVSMTMQ.GFLNYYDACSEGLRAASPALRLGGPG 154 IGRYGLTHVSKWNFETWNEPDHHDFDNVSMTTQ.GFLNYYDACSEGLRIAS PTLKLGGPG 163 IGRYGLSWSKWNFETWNEPDHHDFDNVTMTLQ.GFLNYYDACSEGLRAASPALRFGGPG * *** * * ** **** * * * *** xynb_thesa xynb_calsa idua_human iduajnouse idua_canfa consensus 200 ICGGAD..Y...W..IEDFLN....FCYEENVPVDFVSRHATTSKQGEYTPHLIYQEIMP 203 TAIDA W. . IPELKD. . . . FCTKNGVPIDFISTHQYPTDLA. FSTSSNMEEAMA 223 DSFHTPPRSPLSWGLLRHCHDGTNFFTGEAGVRLDYISLHRKGARSS.ISILE.QEKWA 213 DSFHPLPRSPMCWSLLGHCANGTNFFTGEVGVRLDYISLHKKGAGSS.IAILE.QEMAW 222 DSFHPWPRS PLCWGLLEHCHNGTNFFTGELGVRLDYISLHKKGAGSS.IYILE.QEQATV xynb_thesa xynb_calsa idua_human idua_mouse idua_canfa consensus 249 SEYMLNEFKTVREIIKNSHFPNLPFHITEYNTSYSPQNPVHDTPFNAAYIARILSEGGDY 249 KAKRGELAERVKKALEE.AYP.LPVYYTEWNNSPSPRDPYHDIPYDAAFIVKTI1DIIDL 281 QQIRQLFPKFADTPIYN. DEA. DPLVG. . WS . LPQPWR. . ADVTY . AAMWKVIAQHQNL 271 EQVQQLFPEFKDTPIYN. DEA. DPLVG. . WS . LPQPWR. . ADVTY. AALWKVIAQHQNL 280 QQIRRLFPKFADTPVYN. DEA. DPLVG. . WA. LPQPWR. . ADVTY . AAMWKWAQHQNP xynb_thesa xynb_calsa idua_human idua_mouse idua_canfa consensus 309 . VDSFS YWTFSDVFEERDVPRSQFHG GFGLVALNM IPKPTFYTFKF 307 PLGCYS YWTFTDIFEECGQSSLPFHG GFGLLNIHG IPKPSYRAFQI 333 L L A m T S A F P Y A L L S N D N A F L S Y H P H P F A Q R T L T A R F Q V N N T R P P H V Q L L R K P V L T A M G L 323 LFANSSSSMRYVLLSNDNAFLSYHPYPFSQRTLTARSQVNNTHPPHVQLLRKPVLTVMGL 3 32 PRANGSAALRPALLSNDNAFLSFHPHPFTQRTLTARFQVNDTEPPHVQLLRKPVLTAMAL Figure 2.12 Partial multiple sequence alignment of the enzymes comprising family 39 of glycosyl hydrolases. The consensus sequence is shown at the bottom of the alignment, with (*) indicating fully conserved amino acid residues and (.) indicating similar residues. Numbers to the left denote residue positions. The abbreviations used, references to the published sequence and data bank accession numbers are as follows: Xynb-thesa, (3-xylosidase from T. saccharolyticum (152, SwissProt identifier P36906); xynb-calsa, p-xylosidase from Caldocellum saccharolyticum (157, SwissProt identifier P23552); idua-human, a-iduronidase from Homo sapiens ( 1 5 8 , SwissProt identifier P35475); idua-mouse, a-iduronidase from Mus musculus ( 1 5 9 , SwissProt identifier P48441); idua-canfa, a-iduronidase from Canis familiaris ( 1 6 0 , SwissProt identifier Q01634). The previously predicted acid/base catalytic residue is indicated by (o). The nucleophile in p-xylosidase identified in this paper (Glu-277) and the corresponding residue in C. saccharolyticum P-xylosidase (Glu-275) are indicated by (•). Chapter 2 Mechanism ofT. saccharolyticum 6-Xylosidase 53 Further, the local sequence around the nucleophile (Ile-Thr-Glu-Tyr) also recalls the family 1 Agrobacterium sp. p-glucosidase (Tyr-Ile-Thr-Glu) and the family 10 exoglycanase (Tyr-Ile-Thr-Glu). Both of these two enzyme classes, along with families 2,5,17,26,30,35,39,42,53 enzymes have been assigned as members of clan G-H A, a superfamily of glycosyl hydrolases with an a/p barrel fold. Our results concur with this assignment and with the prediction of the active site nucleophile as Glu277,161 but suggest there may be considerable differences between the fine structures of the active sites of P-xylosidases and a-iduronidases within this family. 2.3.6 Mutagenesis of the Catalytic Nucleophile Site directed mutagenesis of conserved residues found within families of glycosidases has often been used to identify their key catalytic groups. Here, having obtained good evidence pointing to Glu277 as the catalytic nucleophile, we hope to use site directed mutagenesis to both confirm this assignment and to establish the importance of this residue in the catalytic mechanism of XynB. Mutation of the catalytic nucleophile of various retaining glycosidases has resulted in a range of reductions in &Cat from 10 -fold for Agrobacterium sp. P-glucosidase162 to a reduction of at least 104-fold for hen egg white lysozyme acting on well defined substrates.163 The Glu277Ala mutant generated in this work has very low activity, under conditions where the enzyme (1.5 mg/mL) was incubated at 37°C with 2,5-dinitrophenyl p-D-xylopyranoside (2,5DNPX, 4.55 mM). This apparent rc0bs value of the Glu277Ala mutant is at least 25 000 fold less than the kcal value for the wild type XynB. Unfortunately, complications prevented accurate determination of the activity of this mutant enzyme. The first complication was that in the assay buffer, at higher enzyme concentrations, the enzyme precipitated slowly over time, thereby prohibiting the use of very concentrated enzyme solutions or extended reaction times. The second complication, arising as a consequence of the relatively low concentration of enzyme used, was that under the assay conditions the rate of spontaneous hydrolysis of the substrate was approximately 50% of the observed rate. Consequently, the evaluation of catalytic activity of the Glu277Ala mutant outlined here Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 54 must be viewed only as an upper estimate. This 25 000 fold decrease in activity is consistent with Glu277 having a critical role in catalysis as can be expected of a catalytic nucleophile acting to form a covalent xylosyl enzyme intermediate. Those critics favoring an ion pair mechanism may reasonably argue, however, that destroying the favorable electrostatic interactions between the putative oxocarbenium ion and Glu277 in a low dielectric medium such as the enzyme active site may easily account for such a rate decrease. In view of these two arguments and the limitations inherent in assaying this mutant these results cannot reasonably be claimed to negate either mechanism. Regardless of this uncertainty, the data presented in this section provides compelling evidence that Glu277 is the catalytic group in the labeled peptide identified earlier. The observation, however, of a covalent 2-fluoroxylosyl enzyme intermediate can certainly be considered as strong evidence pointing to a Koshland-like mechanism. Those readers favoring a mechanism involving an ion pair intermediate between a glycosyl oxocarbenium ion and a carefully positioned enzymic carboxylate residue may argue that the 2-fluorine substituent of 2F-DNPX perturbs the system such that the normal ion pair intermediate collapses to form an unnatural covalent xylosyl enzyme intermediate. In consideration of the large precedent favoring a mechanism involving a covalent intermediate and the absence of positive evidence pointing to an ion pair intermediate this author feels the ion pair theory to be most unlikely. Regardless of the strength of the ion pair hypothesis, the nature of the intermediate can be discerned using a-deuterium kinetic isotope effects (aD-KIE's). The following section will involve the detailed kinetic analysis of XynB with a view to establishing the nature of the glycosyl enzyme intermediate using aD-KIE's. 2.4 Detailed Mechanistic Analysis Through Kinetic Studies of XynB This section of chapter 2 is directed toward establishing the similarities and differences of the catalytic mechanism of XynB as compared to other retaining (3-glycosidases. One central issue is, of course, the nature of the reaction intermediate. Is it a covalent species or an ion-pair? Also of interest is the possibility, suggested by Li et al, Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 55 that an SN2 type mechanism may operate for some xylosidases. As discussed in the introduction these mechanistic alternatives can be readily distinguished if one can measure the aD-KIE's using substrates for which the second step is rate determining. A critical aspect in this approach is finding a substrate for which the deglycosylation step is rate determining. There are several possible methods available to modern enzymologists for establishing the rate-determining step for any given substrate. One approach involves using chromogenic substrates in conjunction with presteady-state kinetics to look for a 'burst' of chromophore followed by a steady state turnover of the substrate. The burst stems from an initial chemical step that releases the chromophore from the substrate and generates an enzyme intermediate. The resulting data can be analyzed to determine the value of the rate constant governing the first chemical step (fo) and this can be compared to the observed rate constant for the steady state reaction (font). If fo is considerably greater than foat a subsequent step must be rate determining. An alternative approach, using steady state enzyme kinetics, which has been used successfully with some retaining (3-glycosidases, involves use of a series of systematically varied substrates. The glycone moiety of these substrates remains unchanged but different aglycone leaving groups, with varying pKa values, are incorporated (Figure 2.13). Figure 2.13 Generic structure of a series of aryl (3-D-xylopyranosides. The phenol leaving group is substituted with one or more groups that alter the leaving group ability of the phenol. The values of the rate constants governing the hydrolysis of such a series of substrates are plotted logarithmically against the pKa of the leaving group phenol. In this way a linear free energy relationship (LFER) is established that indicates the extent of charge development on the glycosidic oxygen at the rate determining transition state. Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 56 Such an LFER is commonly known as a Br0nsted plot and the relationship is governed by the equation: log k = (3 pKa + C Where k can be either the first or second order rate constant governing the reaction, pKa is the pKa value of the liberated phenol and the slope, P, is the Br0nsted value indicating the magnitude of the relationship. When plotting the first order rate constant (kcat) a slope that is not equal to zero strongly suggests that the first chemical step (glycosylation) is rate determining. This is necessarily so as this is the only step in the reaction in which the aryl glycosidic linkage is broken. The corollary is that a slope of zero is highly suggestive of some other step being rate determining. Thus, a concave downward break in a LFER of log &cat values, with one phase having a slope less than zero and the other phase (where the pKa of the leaving group is relatively low) having a slope of zero necessarily indicates a change in the rate determining step of the reaction. Such a concave downward slope has been found for several P-retaining glycosidases including the family 1 P-glucosidases from Agrobacterium sp.,44 sweet almonds,56, and Pyrococcus furiosus 5 (Figure 2.14). The rate-determining step for substrates used in the generation of the phase having a slope of zero, however, may be either the second chemical step or a non-chemical step. Other experimental devices that can be used to distinguish between these two possibilities will be discussed later. In this study the LFER approach was used to define the rate-determining step for a series of aryl P-D-xylopyranosides. Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 57 3 I 1 — i — i — i 1—i r 0 I i i i i i i i I 4 6 8 1 0 pKa Figure 2.14 Br0nsted plot of log kca, against pKa of the phenol leaving group found for the Agrobacterium sp. P-glucosidase catalyzed hydrolysis of a series of aryl glucosides. Adapted from reference 44. 2.4.1 Subcloning, Purification, and Active Site Titration of XynBH6 For the detailed kinetic study of any enzyme it is, of course, essential to have sufficient quantities of a pure enzyme preparation. In the work described here we sought to perform a large number of studies that would consume a significant amount of enzyme. In order to avoid the lengthy literature purification process carried out for the work described in the previous section we opted to introduce an affinity tag using the techniques of molecular biology. The use of polyhistidine tags fused to proteins for use in their purification by metal chelate affinity chromatography (MCAC) has been described extensively in the literature and inclusion of the tag does not commonly appear to interfere with enzyme function. We therefore subcloned XynB into the pET29b(+) vector and incorporated several codons encoding a hexahistidine tag at the 3' end of the gene in order to generate the expression vector pET29bXynBH6. The gene product of this plasmid is the fusion protein, dubbed XynBH6, which bears a hexahistidine tag at the C-terminal end of the protein. The Michaelis parameters of this recombinant protein (Km = 36 pM and &cat = 275 min"1) are similar to those of the recombinant enzyme lacking the histidine tag (Km = 26 pM and r c c a t = 248 min"1) indicating that the hexahistine tag does not interfere with catalysis by XynBH6. The purification of XynBH6 was greatly faciliated by the presence of the polyhistidine tag and could be accomplished in one step Chapter 2 Mechanism ofT. saccharolyticum 0-Xvlosidase 58 using MCAC. In the work outlined in this chapter the purity of the wild-type and mutant enzymes has been established using SDS-PAGE in conjunction with silver staining. All enzymes preparations used were estimated to be of greater than 95% purity. One other important consideration when undertaking detailed kinetic studies is that one must also establish whether the entire enzyme sample obtained is active. This can be accomplished using an active site titrant. As outlined in the introduction the 2,4-dinitrophenyl 2-deoxy-2-fluoro-(3-D-glycosides have been shown to be very effective mechanism based inactivators of P-retaining glycosidases.12 As determined in the preceding section of this chapter, for each equivalent of Thermoanaerobacterium saccharolyticum p-xylosidase (XynB) reacted with 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-xylopyranoside (2F-DNPX) one equivalent of 2,4-dinitrophenolate (DNP) is released. Results discussed in section 2.3.2 of this chapter reveal that the resulting fluoroglycosyl enzyme intermediate turns over only very slowly, thereby allowing its complete accumulation.164 Thus, incubation of XynBH6 with 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-xylopyranoside (2F-DNPX) resulted in its stoichiometric labeling as confirmed by ESMS indicating that the entire enzyme preparation is properly folded and functional. The number of active enzyme units and their extinction coefficient can thus be determined by monitoring spectrophotometrically the amount of DNP released at several different enzyme concentrations. A plot of the volume of enzyme solution added against the final change in absorbance at 400 nm arising from stoichiometric release of the DNP moiety is shown in figure 2.15 and allows for the calculation of an experimental extinction coefficient of 3.19 mL mg"1 cm"1 for XynBH6. The enzyme stock was also periodically assayed using the substrate depletion method (vide infra, section 2.4.2) in order to determine whether the second order rate constant for a fixed amount of enzyme remained the same. In this way it was established that the recombinant XynBH6 was stable for at least 6 months at 4 °C when stored in the elution buffer of the metal chelate affinity column. Chapter 2 Mechanism ofT. saccharolyticum 0-Xvlosidase 59 12 0 0 10 20 30 40 50 XynBH 6 (p,L stock solution) Figure 2.15 Active site titration of XynBH6 using 2F-DNPX. 2.4.2 Substrate Specificity Glycosidases are often very versatile catalysts, hydrolyzing a number of substrates having differing aglycone moieties. Occasionally, experimentalists, having cloned a particular glycosidase, will not have thoroughly characterized its substrate specificity and will offer only a tentative assignment of its glycone specificity. In the absence of any experiments directed to clearly establishing the specificity of such enzymes their tentative assignments will slowly tend to mistakenly become accepted as definitive. One example is the enzyme Cex, originally cloned from Cellulomonas fimi. This enzyme was long considered an exo-glucanase.165 Later studies, however, revealed that the enzyme has greater activity on the structurally similar xylan than on glucans.166 The enzyme should therefore, in strict terms, be defined as an endo-xylanase. In order to verify the initial assignment of XynB as a retaining P-xylosidase we have opted to study a series of structurally related substrates. Kinetic parameters for a series of para-nitrophenyl glycoside substrates containing different sugar moieties are presented in Table 2.1. From examination of the rate constants governing the hydrolysis of a series of aryl glycosides it can be seen that Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 60 the enzyme has considerable specificity for xylosides over all other substrates tested, which is consistent with earlier qualitative observations. 1 5 2 Table 2.1 Kinetic parameters for the hydrolysis of a series of para-nitrophenyl glycosides by XynBH6. Phenyl Glycoside Substrate &cat ( s ) Km(mM) W^mCs'mM-1) P-pNPX HO-HO. N02 a-pNPX P-pNPGlc OH HO HO-P-L-pNPFuc HO HO C H 3 P-pNPGal HO - O H HO-Q O OH A\ P-pNPGlcNAc ^ - O H H O — ^ A - — H 0 - \ » « ^ ^ - 0 . NHAc \ \ 3.4 Undetectable = 1.0 0.18 0.082 Undetectable Undetectable 0.058 N/A = 3.0 7.4 = 17 N/A N/A 59 N/A = 0.33 0.024 * 0.006 N/A N/A Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 61 Xylosidases from families 43 1 6 7 and 54 1 4 2 have been shown to have a relaxed specificity for oc-L-arabinosides (the C4 epimer of P-D-xylosides). These enzymes often cleave the cc-L-arabinosides at a similar rate to, or even greater rate167 than, P-D-xylosides. The lower relative activity of XynBH6 toward a-L-arabinosides suggests that XynBH6 recognizes 0-4 more stringently than do xylosidases from other families. Specificity is even greater at other positions, as the enzyme does not readily accommodate any increased steric bulk at C-5 and C-2 as evidenced by the much lower rate constant for the hydrolysis of fuco-, gluco-, and 2-acetamido-glucosides. 2.4.3 pH Dependence Having unambiguously established the substrate specificity of XynBH6, the next requirement for detailed characterization is to examine its pH profile. To this end, values of kcJKm at a series of pH values between 4.5 and 9.5 were determined using the substrate depletion method. In this approach the substrate concentration that is used is much less than Km and consequently the reaction follows second order kinetic behavior but since the enzyme concentration is constant this appears as pseudo-first order behavior. By continuous spectrophotometric monitoring of the reaction for at least five half-lives an asymptotic progress curve is obtained. The second order rate constant kcJKm can therefore be conveniently extracted from the progress curve of the reaction by non-linear regression. Km was also determined at both extremes of the pH range studied and it was found that its value was significantly higher than the substrate concentration used in these studies. A plot of the pH profile (Figure 2.16) reveals that the kcJKm values depend on two ionizations and the apparent p K a values of these titratable groups could be extracted from the data as pKai = 4.1 and p K a 2 = 6.8. The value of pKai is somewhat unreliable as the data for this limb of the pH profile could not be completed owing to instability of the enzyme at lower pH. Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 62 300 250 h O 4 5 6 7 8 9 10 P H Figure 2.16 pH dependence for the hydrolysis of pNPX catalyzed by T. saccharolyticum Presumably, the pH dependence of the second order rate constant kcat/Km for the hydrolysis of pNPX reveals ionizations occurring within the free enzyme. A bell-shaped pH profile similar to the one obtained in this study has been observed for several (3-retaining glycosidases.44'168"171 The large half-height width of the bell-shaped profile suggests that these apparent pKa values likely reflect intrinsic pKa values of the ionizable residues, although the possibility of inverse-protonation172'173 makes firm assignments impossible. Nmr studies, however, in which the ionizable groups in P-retaining glycosidases have been titrated reveal that the two apparent pKa values associated with the inflections in bell shaped profiles of kcJKm versus pH reflect the ionization of two active site carboxyl groups.6 4 , 1 7 4 One of these carboxyl groups has a significantly perturbed pKa value (6.8) compared to that of a normal carboxyl residue (4.8). Such significant perturbations have been seen in many glycosidases. 4 3 . 4 4 > 1 7 5 The other residue has an estimated pKa (4.1) consistent with that expected for a carboxylic acid (4.8). By analogy with previous studies it is very likely that these apparent pKa values reflect the intrinsic pKa values of the nucleophile and acid/base residues. Such an assignment, however, cannot be made conclusively in the absence of further studies owing to the above mentioned possibility of inverse-protonation. If inverse-protonation were operative here the two residues would have the opposite assignments. Regardless of this xylosidase. Chapter. 2 Mechanism ofT. saccharolyticum 0-Xylosidase 63 ambiguity, for the purpose of determining the pH at which the enzyme has optimal activity, this study provides the necessary information to proceed with more detailed mechanistic work. 2.4.4 Evidence for a Two Step Mechanism Involving Oxocarbenium Ion-Like Transit ion States and a Covalent Xylosyl Enzyme Intermediate. 2.4.4.1 Structure/ reactivity studies The XynBH6 catalyzed hydrolysis of a series of aryl xylosides provides a biphasic concave downward Br0nsted plot of log kcat against pKa of the leaving group (Figure 2.17a). As discussed above such a plot strongly suggests a two step mechanism. Similar biphasic Br0nsted plots have been observed with the p-retaining glucosidases from Agrobacterium sp. 4 4 and sweet almonds 5 6 . Conversely, for the retaining p-xylosidase from Trichoderma koningii no dependence on leaving group reactivity was observed 1 4 2 . A two-step ping-pong kinetic mechanism that is commonly found for retaining glycosidases is outlined in Scheme 2.5. Scheme 2.5 Kinetic mechanism for XynBH^. The first step (fo), the xylosylation step, involves cleavage of the glycosidic bond and formation of the xylosyl enzyme intermediate. Substrates with good leaving groups (pKa < 9) show no dependence of their reactivity on phenol leaving group ability. Therefore, for these activated substrates, the xylosylation step is unlikely to be rate determining although it is possible that a non-chemical step, such as product dissociation, is rate determining with these activated substrates. E + R X k k -i Chapter 2 Mechanism of T. saccharolyticum 0-Xyldsidase 64 Table 2.2 Kinetic parameters for the XynBH-j-catalyzed hydrolysis of a series of aryl (3-D-xylopyranosides. Substrate pK a a ^•cat (mM) (s"1 mM"1) 2.5DNPX 5.15 7.3 0.011 670 3,4DNPX 5.36 8.9 0.0099 900 " . pNPX 7.18 9.7 0.036 270 oNPX 7.22 9.9 0.046 220 3.5DC1PX 8.19 7.7 0.035 220 mNPX 8.39 11 0.15 73 3C1PX 9.02 10 0.29 36 4BrPX 9.34 1.7 0.34 5.1 NapX 9.51 2.6 0.31 8.5 PX 9.99 4.0 1.9 2.1 pOMePX 10.20 1.1 0.22 5.2 . 3,4DMePX 10.32 1.6 2.5 0.65 a pKa Values used taken from references Good evidence, however, pointing to a rate determining chemical step is found in the nucleophilic competition experiments when DTT is added as an exogenous nucleophile to the reaction mixture. In this manner we can examine whether the breakdown of the xylosyl enzyme intermediate is rate determining. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase •65 a 1.5 3 at o 0.5 -0.5 4 3 -i 1 r 1 r r _j I i L 10 P#a at 2 u 1 M O 0 -1 _i I i L 10 P*a 12 12 Figure 2.17 Br0nsted plots relating the rate of T. saccharolyticum f3-xylosidase catalyzed hydrolysis of a series of aryl xylosides with the pKa of the corresponding phenol, (a) Plot of log(£ c a t) vs pKa of the aglycon leaving group, (b) Plot of log(£cal/ATm) vs pKa of the aglycon phenol.efefef 2.4.4.2 Effect of Exogenous Nucleophiles If for any given substrate the second chemical step is rate determining and we add a small nucleophilic molecule to the reaction mixture we can.expect a rate increase if the nucleophile acts to intercept the xylosyl enzyme more efficiently than water. In the studies described here DTT was used as the external nucleophile. The values of &cat for the enzyme-catalyzed hydrolysis of 2,5DNPX were determined at varying concentrations 66 Mechanism ofT. saccharolyticum 0-Xylosidase Chapter 2 of DTT and are shown in Figure 2.18. A clear linear increase in foat value is observed with increasing DTT concentration for 2,5DNPX. 0 100 200 300 400 500 [DTT] (mM) Figure 2.18 DTT as an exogenous nucleophile increases the rate of XynBH6 catalyzed hydrolysis of (O) 2.5DNPX, but not of (•) PX. This reasonably suggests that DTT efficiently intercepts the xylosyl enzyme intermediate and accelerates its breakdown by a competing alternate pathway governed by fo and DTT concentration (Scheme 2.5). At DTT concentrations greater than approximately 150 mM there is no additional effect on foat indicating that the rate of dexylosylation has been increased such that it is either greater than, or similar to, that of the xylosylation step. Confirmation of this interpretation can be had by examining the Michaelis-Menten parameters for the hydrolysis of 2,5DNPX by XynBH6 in the presence of 20, 62, and 169 mM DTT (Table 2.3). Km can be expressed as: Km = [E][S] / £ [All enzyme bound species] Therefore as the concentration of DTT is increased and the dexylosylation step, proceeding by pathways A and B (Scheme 2; Path A (fo) and Path B (fo)) increases, the concentration of the enzyme intermediate decreases. The net result is an increase in the value of Km in accord with the above equation. No effect on the first step is expected and this is borne out by the absence of any significant change in kcJKm (Table 2.3), a Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 67 parameter that presumably reflects the first irreversible step of the reaction. The lack of dependence of the rate of hydrolysis of PX on DTT concentration lends further support for this interpretation (Figure 2.18). Table 2.3: Kinetic parameters for the T. saccharolyticum p-xylosidase catalyzed hydrolysis of 2,5DNPX at different DTT concentrations. [DTT] (mM) ^cat ( s ) K m (mM) KJKM (s"' mM"1) 0 4.0 ±0 .2 .22 ± 3 0.18 ±0.03 20 8.2 ±0 .5 49 ± 9 0.17 ±0.03 62 13.1+0.4 57 ± 6 0.23 ± 0.03 169 26.9 ± 0.9 146 ± 16 • 0.18 ±0 .02 Since the xylosylation step is rate hmiting for PX this confirms that DTT only acts to accelerate the dexylosylation step of the reaction by intercepting the xylosyl enzyme intermediate, and has no effect on the xylosylation step. Together, these data clearly show that, under the conditions studied here, only the chemical steps are rate determining and the biphasic nature of the plot is a consequence of a change in rate determining step from xylosylation (for substrates bearing a leaving group with a pKa value of > 9) to dexylosylation (for substrates bearing a leaving group with a pKa value of < 9). As described earlier, a hydroxyl group from another saccharide moiety can also act in the place of water to intercept the covalent xylosyl enzyme (sections 2.1 and 2.3.1). This has been supported by the reactivation studies of the 2-fluoro xylosyl enzyme in the presence of xylose, xylobiose, and BTX, where the rate of breakdown of the intermediate was increased by a factor of 75 fold over the rate of hydrolysis. More tangible evidence for this partitioning is seen on incubating the enzyme in the presence of substrates bearing leaving groups with either high (PX), or low (2,5DNPX) pKa values. Mass spectrometric analysis of these mixtures reveals disaccharide products (data not shown). Kinetic evidence for transglycosylation can be found in the Michaelis-Menten plots of the enzyme-catalyzed hydrolysis of certain aryl xylosides. At higher concentrations, substrates bearing activated leaving groups, with pKa values of less than 9.0, show significant deviation from Michaelian saturation kinetics. However, substrates bearing a leaving group with a pKa value higher than 9.0 show normal saturation kinetics. Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 68 Examples of the Michaelis-Menten and Lineweaver-Burk plots of substrates bearing a leaving group with either a low or a high pKa value are shown in Figure 2.19 (Michaelis-Menten plots for all substrates can be found in appendix 1 of this thesis). Values of fcCat and Km for all substrates showing non-Michaelian kinetics were determined by non-linear fitting of data to substrate concentrations lower than those levels at which significant deviation from Michaelian kinetics is observed. Such data fitting yielded kinetic parameters in good accord with the values obtained by linear regression of the linear portion of the Lineweaver-Burk plots found at low substrate concentrations. Consequently, the Michaelis parameters obtained for those substrates having pKa values less than 9.0 are less accurate but should be considered close approximations of the true values. 0.06 0.04 0.02 0 0 0.02 0.04 0.06 0.08| [2.5DNPX] mM _ j I I L 0.01 ^ 0.008 * | 0.006 | 0.004 h ^ 0.002 0 0 0.1 0.2 0.3 0.4 0.5 [2,5DNPX] mM 1 2 3 4 5 [4BrX] mM Figure 2.19 Michaelis-Menten plots for the T. saccharolyticum p-xylosidase-catalyzed hydrolysis of (a) a substrate for which dexylosylation is rate determining, 2,5DNPX, and the region at low substrate concentration (inset) obeying Michaelian kinetics; (b) a substrate for which xylosylation is rate determining, /?ara-bromophenyl P-D-xylopyranoside (4BrPX). A biphasic Br0nsted plot is also observed if log (hax/Km) is plotted against the pKa value of the phenol leaving group (Figure 2.17b). The region of the Br0nsted plot for substrates bearing relatively poor leaving groups (pKa > 8) reveals a significant correlation between the pKa value of the leaving group phenol and log (kcai/Km) as can be seen from the slope of Pig = -0.97 (n = 8, r = -0.98) in figure 2.17. Such a large negative (3ig indicates that there is significant fission of the glycosidic bond in the transition state Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 69 and relatively little proton donation to the developing phenolate anion. The observation of almost complete breakage of the glycosidic linkage is consistent with the isotope effects (vide infra, section 2.4.3.3) that indicate significant oxocarbenium ion-like character at the transition state. The parameter kcJKm governs the reaction of the free enzyme to the transition state of the first irreversible step along the reaction pathway regardless of the pKa of the leaving group, while kcat only governs this step when substrates bearing poor leaving groups (pKa > 9) are used. We can expect, therefore, that the negative pig value determined from the plot of log (kcat/Km) should correlate with the region of the plot of log (kcat) involving substrates having leaving group pKa values of greater than 9.0 and, indeed, this is the case. Given that kcJKm is expected to reflect the first irreversible chemical step of the reaction it is a rather curious observation that the slope of the Br0nsted plot of log (kca,t/Km) is not linear throughout the entire range plotted. Indeed, pi g becomes less negative for xylosides with good leaving groups. Such non-linear behavior has also been observed in the Br0nsted plots of log (kcJKm) constructed for the hydrolysis of aryl glycosides by other p-retaining glycosidases including the glucosidases from Agrobacterium sp.44 and sweet almonds,56 and a xylanase from Cellulomonas fimi.43 The biphasic nature of this plot may therefore arise from the first chemical step becoming reversible as the pKa of the leaving group decreases. Early studies on the inhibition of sweet almond P-glucosidase showed that the K\ of a series of phenols was dependent on their pKa value, with lower pKa phenols being more effective inhibitors.179 Thus it is possible that with lower pKa phenols the off rate constant for dissociation of the phenol inhibitor from the enzyme is slowed and regeneration of the aryl xyloside by internal return becomes a competing process. This would make the first step reversible and consequently kcJKm under such circumstances might not reflect solely the xylosylation step. Regardless of this complication it is clear that these Br0nsted plots, in conjunction with the partitioning studies using DTT, indicate that the mechanism of XynBH6 involves a two step reaction. The large, negative pi g observed indicates that the transition state leading to xylosylation is very polar with breakage of the glycosidic bond being far advanced and, presumably, little proton donation from the general acid/base catalytic Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase '70 residue. Furthermore, which chemical step is rate determining depends upon which aryl xyloside is being hydrolyzed. This last point is critical, as we now know which substrates can be used to measure kinetic isotope effects on each step of the enzyme-catalyzed reaction. 2.4.4.3 Kinetic Isotope Effects cc-Deuterium kinetic isotope effects have been used to probe changes in hybridization at the anomeric center for both the enzyme-catalyzed and the spontaneous hydrolysis of glycosides. These isotope effects reveal the change in hybridization on proceeding from a stable species (free enzyme and substrate, Michaelis complex, or glycosyl enzyme intermediate) to the subsequent transition state. C-l Deuterium enriched substrates that incorporate leaving groups with a large range of leaving group pKa values were prepared in order to determine the change in hybridization for the xylosylation step (first transition state), the dexylosylation step (second transition state), and a near borderline case. In all three cases a significant and normal isotope effect of ka/kD ~ 1.09 is observed. From the Br0nsted plots we know that kcat for both pNPX and 2,5DNPX must reflect the second chemical step. As discussed above kcJKm reflects the first irreversible chemical step and for oNHAcX (leaving group pKa = 9.95) this is.clearly the xylosylation step. From the studies outlined just above we can be confident that chemical steps are entirely rate-determining for any of these substrates and consequently the measured isotope effects should reflect the intrinsic isotope effect. The KIE values measured are in the range measured for, other glycosidases (in cases where it is certain the chemical step is being measured) on both the glycosylation (1.05 - 1.11) and deglycosylation steps (1.08 - 1.25) 43-45,47,180 The.normal isotope effects observed for both steps of the reaction indicate a rehybridization of C-l from sp3 to sp2 in both steps. Such changes in hybridization are consistent only with a covalent sp3 hybridized xylosyl enzyme intermediate and cannot be reconciled with an oxocarbenium ion intermediate. The isotope effects for the two steps are relatively large and of very similar magnitude, indicating that the transition state for both steps has significant oxocarbenium ion-like Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 71 character. This result, in conjunction with a Pig of -0.97, indicates that the transition states leading to the formation and breakdown of the covalent xylosyl enzyme intermediate are "exploded" with very little nucleophilic participation.41 2.4.3.4 Summary All of the data reported here support a two-step double displacement mechanism for the family 39 XynB enzyme in which a covalent xylosyl enzyme is formed and hydrolyzed with acid/base catalytic assistance. The transition states bracketing the intermediate have considerable oxocarbenium ion-like character and very little nucleophilic participation. Such a mechanism is very similar to the A N D N mechanism found for the hydrolysis of glycosyl fluorides in the presence of anionic nucleophiles.34 The results outlined here contrast with those observed by Li et al who suggest that some retaining P-xylosidases have a reaction mechanism involving very SN2-like transition states. That study, while very intriguing, should be revisited in order to ensure that the chemical step is rate determining rather than some non-chemical step such as diffusion. In summary, the catalytic mechanism of XynBH6 is very similar to the catalytic mechanisms of the very large majority of glycosidases studied to date and involves highly dissociative transition states which flank a covalent glycosyl enzyme intermediate. This conclusion is in disagreement with the ion pair proposal of Phillips.13 2.5 Identification of the Acid/Base Catalytic Residue and Evaluation of its Importance in Catalysis The previous studies on the mechanism of XynB have clearly established the catalytic mechanism of XynB. The intermediate is unambiguously a covalent xylosyl enzyme intermediate and the transition states flanking this species are highly dissociative. These observations are consistent with a nucleophilic role for Glu277 and with the assignment of this enzyme to clan GH-A. The catalytic acid/base residue of this enzyme has been proposed, by us 1 6 4 and by Henrissat,181 to be Glul60. One outgrowth of the studies into Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 72 the mechanism of XynB is the subject of this section. Here we detail studies directed towards the unambiguous identification and detailing of the importance of the catalytic acid/base residue of XynB. The identification of key active-site residues in glycosidases is critical for understanding their catalytic mechanisms, for enzyme classification, and for bioengineering of glycosyl hydrolases with altered properties.182 In the absence of X-ray crystallographic data many catalytically important residues have been identified through labeling studies using affinity labels or mechanism-based inactivators. Mechanism based inactivators such as the 5-fluoro-glycosyl fluorides183'184 and 2-deoxy-2-fluoro glycosides149'150-164 have proved to be of great utility in reliably labeling the catalytic nucleophile. The development of a reliable method for the labeling of the acid/base catalyst has been elusive making the reliable identification of these residues difficult. A variety of affinity labels, which incorporate a sugar moiety to provide specificity for the active site and a reactive group capable of forming stable conjugates with the enzyme, have been developed and have met with varying degrees of success. Epoxyalkyl glycosides have proved useful in a number of cases although in many instances they have labeled the nucleophile.185 As discussed in the introduction, N-bromoacetyl glycosylamines have also met with mixed success. N-Bromoacetyl P-galactosylamine was used to identify a methionine in the active-site of E. coli P-galactosidase120 while the gluco analog was used to label a conserved carboxylate residue in the P-glucosidase from cassava.119 In Agrobacterium sp. P-glucosidase inactivation of the enzyme by N-bromoacetyl P-glucosylamine was found to follow pseudo-first order kinetics but the enzyme was shown by mass spectrometry to be labeled with at least three equivalents of the reagent.121 However, in a study of an exoglycanase (Cex) from Cellulomonas fimi, N-bromoacetyl P-cellobiosylamine was found to label a conserved residue proposed as the acid/base catalyst.118 Indeed, later x-ray crystallographic studies on the 2-fluoroglycosyl enzyme60 and kinetic studies of mutants at this position186 have provided clear evidence to confirm the identity of the residue so labeled as the acid/base catalytic residue. An alternative method for the identification of the catalytic acid/base residue using a combined mutagenesis and chemical rescue strategy has been developed 6 6 - 1 8 6 Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 73 and proved to be of use in several cases 55,187,188 Studies have shown that for those substrates that bear a very good leaving group that does not require acid catalytic assistance, removal of the acid/base catalytic residue has minimal effect upon the glycosylation step. However, the second step, in which water attacks this intermediate, is significantly compromised when the general base has been deleted as the nucleophilicity of the incoming water is no longer enhanced by base catalysis. Both steps, however, are greatly compromised for substrates with poor leaving groups that require acid catalytic assistance. Additionally, replacement of the acid/base catalyst in a retaining P-glycosidase with a smaller residue has been proposed to generate a small cavity at the P-face of the substrate adjacent to the anomeric center. This site may accommodate small molecules, particularly anions.66 In several cases significant increases in the steady state hydrolysis rate have been observed when a substrate with a good leaving group, which does not require acid catalysis, is incubated with the enzyme in the presence of an anion more nucleophilic than water. Such behavior has been shown to result from the interception of the glycosyl enzyme intermediate by the exogenous nucleophile to yield P-glycosyl adducts of the nucleophile (Figure 2.20). ENZ Figure 2.20 Rescue of activity through the use of the exogenous nucleophile azide. In order to identify unambiguously the acid/base catalytic residue we will use a combined approach involving the use of a novel affinity label, site directed mutagenesis, and detailed kinetic analysis. Mechanism ofT. saccharolyticum 0-Xylosidase Chapter 2 2.5.1 Inactivation of the enzyme 74 Incubation of recombinant hexa-histidine tagged Thermoanaerobacterium saccharolyticum P-xylosidase (XynBHe) with N-bromoacetyl P-D-xylosylamine (NBX, Figure 2.21) resulted in pseudo-first-order inactivation of the enzyme in a rapid, time-dependent manner (Figure 2.22a). The enzyme inactivation rate with NBX was shown to be dependent on the concentration of inactivator, although saturation was not observed (Figure 2.22c). Analysis of the data as described in the methods and materials section permitted the calculation of the second order rate constant for the inactivation process (k-JK, = 0.026 min 'mM'1) with NBX. Br Figure 2.21 Chemical structure of the affinity label N-bromoacetyl-/3-D-xylbpyranosylamine (2.1). Incubation of the enzyme with NBX (11.9 mM) in the presence of the competitive inhibitor xylose (80.8 mM, Ki = 20 mM) or benzyl P-D-thio-xyloside (BTX; 10.6 mM, K{ = 5.3 mM) resulted in lower apparent inactivation rate constants. In the absence of xylose or BTX the apparent inactivation rate constant was 0.29 min"1, while in the presence of xylose or BTX this value dropped to 0.08 min -1 and 0.19 min -1 respectively (Figure 2.27b). This observed protection from inactivation indicates that NBX and xylose or BTX are competing for the same site, suggesting that the inactivation occurs at the enzyme active site. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 75 0 5 10 15 20 [NBX] (mM) Figure 2.22 Inactivation of T. Saccharolyticum P-xylosidase by NBX. (a) Semilogarithmic plot of residual enzyme activity versus time at the indicated inactivator concentration: (O) 1.49 mM, (•) 2.99 mM, (•) 5.97 mM, (• ) 11.94 mM, (A) 17.91 mM. (b) Inactivation with 11.9 mM NBX alone (•) and in the presence of 10.6 mM BTX (•) or 80.8 mM xylose (O). (c) Plot of the first-order rate constants from (a). 2.5.2 Stoichiometry of incorporation of inactivator studied by ESMS The mass of the native XynBH6 was found by ESMS to be 59726 ± 6 Da (Figure 2.23a). After inactivation of the enzyme with NBX, a new species was observed with a mass of 59914 ± 6 (Figure 2.23b). The mass difference between inactivated and native XynBH6 is 188 Da and corresponds, within error, to the addition of one N-acetyl (3-xylosylamine label (190 Da). A second, minor peak at 60 088 Da presumably corresponds to a very small amount of doubly labeled enzyme. Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 76 59914 C es 2 59726 58500 59000 59500 60000 60500 61000 Mass (Da) Figure 2.23 Reconstructed electrospray mass spectrum of; (a) native XynBH 6 and (b) XynBH 6 after incubation with 21 mM NBX for 30 minutes. 2.5.3 Identification of the labeled active-site peptide by ESMS Peptic hydrolysis of the N-acetyl xylosylamine-enzyme resulted in a mixture of peptides that were separated by reverse-phase HPLC using the ESMS as a detector. A sample of unlabeled enzyme was also subjected to peptic digestion. When these two mixtures were analyzed in separate experiments by scanning in the normal LC/MS mode the total ion chromatogram (TIC) in both cases showed a large number of peaks, each corresponding to one or more peptides in the digest mixture (Figure 2.24a, 2.24b). The peptide bearing the N-acetyl (3-xylosylamine label was then located by careful comparison of the TIC's of the labeled and unlabeled enzyme digests. It was expected that the masses of relevant labeled and unlabeled active-site peptides would differ by the mass of the N-acetyl f3-xylosylamine label (190 Da). The masses of peptides Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 77 corresponding to peaks found solely in one of the TIC's from either the labeled or unlabeled enzyme were compared. In this way a peptide with a mass of 1075 ± 1 Da was found only in the TIC of the labeled enzyme digest at 29.3 minutes and this peptide was therefore a good candidate for the active-site peptide of interest. A search of the TIC of the unlabeled enzyme digest for a possible unlabeled peptide of mass 885 ± 1 Da, corresponding to the mass difference between the peptide of mass 1075 Da and the mass of the N-acetyl xylosylamine label (190 Da), yielded a peak at 31.3 minutes (Figure 2.24b). The TIC of the labeled enzyme digest was examined for the same peptide and a peak was found at the same position, albeit at a much lower intensity. These results suggest that the peptide of mass 1075 Da is likely the modified peptide and that the appearance of the peptide of 885 Da in the labeled sample was likely due to incomplete inactivation of the enzyme and/or cleavage of the N-acetyl xylosylamine label during proteolysis or chromatography. The labeled peptide was purified by reverse phase C-l8 HPLC using the ESMS as a detector. 20 2'5 30 35 Time (min) 40. 45* 20 22 24 26 28. 30 32 34 36 . 38 40 42 44 Time (min) Figure 2.24 ESMS experiments on a peptic digest of XynBHg. (a) TIC of the unlabeled enzyme digest indicating the position of the unlabeled peptide of m/z 885 (•). .(b) TIC of the enzyme labeled with NBX where (T) indicates the position of the newly observed peak corresponding to the labeled peptide m/z 1075 and (•) indicates the position of the unlabeled peptide of m/z 885. Mechanism ofT. saccharolyticum 0-Xvlosidase Chapter 2 2.5.4 Peptide sequencing 78 Information on the sequence of the peptide bearing the label and its site of attachment was obtained in separate experiments by fragmentation of the two peptides of interest (m/z 1075 and m/z 885) in the daughter ion scan mode (Figure 2.25a and 2.25b). The parent ions in both the labeled peptide and the unlabeled peptide appear as singly charged species. Peaks resulting from B ions of the unlabeled peptide correspond to IW (m/z 300), IWN (m/z 414), IWNE (m/z 543), IWNEP (m/z 640), and IWNEPN (m/z 754). Peaks arising from Y" ions of the unlabeled peptide correspond to NL (m/z 246), PNL (m/z 343), EPNL (m/z 472), NEPNL (m/z 586), and WNEPNL (m/z 772). This information, in conjunction with the mass of the intact peptide and the primary sequence of the enzyme, clearly defines the unlabeled peptide as 158IWNEPNL164. In a separate MS/MS sequencing experiment the purified, labeled peptide was also subjected to fragmentation. The peaks arising from B ions of the labeled peptide (where * denotes the label) include IW (m/z 300), IWN (m/z 414), IWNE* (m/z 733), and IWNE*PN (m/z 942). Peaks arising from Y" ions of the labeled peptide correspond to NL (m/z 246), PNL (m/z 343), E PNL (m/z 662), and NE PNL (m/z 775). Additional peaks arising from the fragmentation of the ester bond linking the label and peptide are marked by ( T ) in Figure 2.30b. Together, this information clearly permits identification of the active-site peptide bearing the label as 157IWNEPNL164 within which the site of attachment is Glul60. The identification of this residue, the sole carboxyl residue within the peptide, was encouraging. In all cases to date where the identity of the acid/base catalyst in a glycoside hydrolase has been determined it has been found to be either an Asp or Glu side chain. During the preparation of this thesis a paper appeared describing the identification of this conserved residue as the acid/base catalytic residue of the Bacillus stearothermophilus P-xylosidase using a similar chemical rescue approach.188 Chapter 2 Mechanism ofT. saccharolyticum 6-Xylosidase 79 a cu .fix 20 15 Y" ions 10 -F 0 775 662 343 246 I - - W - N - / - E - / P - / N T / - L 300 414 733 942 B ions F 2 4 6 . 300 JLll 200 343 / yd 414 662 / 733 942 775 400 600 m/z 800 1075* IM + HY r •« i 1000 1200 b a 5 « d . • • + - » c Y"iohs 772 586 472 343:246 50 40 30 20 10 0 I/W/N-/-E-/P-/-N-/-L 300 246. \ 300 414 543 640 754 343 B ions 1200 Figure 2.25 ESMS/MS daughter-ion spectrum of the (a) labeled peptide (m/z 1075, in the singly charged state) and (b) unlabeled peptide (m/z 885, in the singly charged state). Observed Y" and B series fragments are shown above and below the peptide sequence respectively. (•) Indicates the site of attachment of the label and (•) indicates fragments arising from homolytic cleavage of the ester bond linking the peptide and label. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 80 2.5.5 Production and Purification of the Mutant Enzyme The mutant XynBH6(E160A) was prepared from the pET29b(+)XynBH6 construct described in a preceding section (2.4.1) using simple two primer mutagenesis. Complete sequencing of the gene confirmed the desired DNA sequence. The enzyme was expressed to high levels as revealed by silver staining of SDS-PAGE of a sample of the cell extract (data not shown). During the purification of the mutant enzyme, great care was taken to eliminate the possibility of contamination of XynBH6(E160A) by XynBH6. Purification of the enzyme using metal chelate chromatography column provided a rapid route to a sample of the pure enzymes which, by SDS-PAGE and silver stain analysis migrated as a single band (>95 % by inspection), at the same position as a sample of the recombinant XynB. 2.5.6 pH Profiles Values of kcJKm were determined for the hydrolysis of pNPX by XynBH6 and of 3,4DNPX by XynBH6(E160A) as a function of pH within the pH stability range (4.5-9.5) of the recombinant enzyme (A table of the data can be found in Appendix 1). The mutation had a significant effect on the pH profile of the enzyme as shown in Figure 2.26. Fitting the data to double titration curves yielded the following pKa values: for XynBH6, pKa! =4.1 ± 0.1, pKa2 = 6.8 ± 0.1; for XynBH6(E160A), pKai = 6.5 ± 0.1, pKa2 = 9.0 ±0 .1 . Bell shaped pH profiles are commonly seen for wild-type glycosidases and usually reflect the ionizations of the catalytic nucleophile and the acid/base residue of the free enzyme.189 The approximate 2 pH unit shift of the bell-shaped pH profile upon removal of the acid/base residue is unexpected. Elimination of the acid/base residue in several other cases has resulted in the elimination of the basic limb of the pH profile corresponding to pKa 2 . 5 5 - 1 8 6 Chapter 2 Mechanism ofT. saccharolyticum B-Xvlosidase 81 Figure 2.26: pH dependence of kcJKm for the wild-type XynBH6 xylosidase catalyzed hydrolysis of pNPX (O ) and XynBH6(E160A) xylosidase catalyzed hydrolysis of 3.4DNPX (•). The lines shown represent fits to the data for enzymes with two ionizable residues. Note that the substrates used are different. The precise cause of the change in the pH activity profile observed here cannot be determined from this study. It is possible, however, to speculate that in XynBH6 a third ionizable group having a pKa greater than 6.8 in the active site is required for efficient catalysis. This ionization would be masked in XynBH6 by the ionization of the acid/base catalyst. Thus, only on removal of the acid/base catalyst residue, as in XynBHe(E160A), would the ionization of this third residue become apparent. The interpretation of the shift in pKai that presumably corresponds to the ionization of the catalytic nucleophile is also beyond the scope of the data described here although it is possible to speculate that it may be a consequence of altering the electrostatic field within the enzyme active site. An alternative explanation may be that the shift results from inverse-protonation. The true microscopic pKa of the nucleophile may be quite high and the pKa of the acid/base catalytic residue much lower. On deletion of the carboxyl group of the acid/base residue the true microscopic pKa of the nucleophilic residue is unmasked. Regardless, these unusual effects on the pH profile of the mutant enzyme are consistent with Glu 160 having a significant role in catalysis, likely as the acid/base catalytic residue. Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 82 2.5.7 Comparison of the Substrate Reactivity of XynBH6 and XynBH6(E160A) The acid/base catalyst serves an important function in both steps of the hydrolysis of a glycoside catalyzed by a retaining glycosidase. In the first step of the reaction (glycosylation) the residue acts as a general acid to donate a proton to the glycosidic oxygen in order to facilitate its cleavage. In the second step of the reaction (deglycosylation), the residue functions as a general base catalyst to assist in the attack of an incoming water molecule on the glycosyl enzyme intermediate. Removing the carboxyl group of the acid/base catalytic residue by site directed mutagenesis should therefore slow both steps. However, the degree to which each step is slowed should depend on the leaving group of the substrate studied. For all glycosides of the same sugar, regardless of the leaving group, the deglycosylation rate should be the same, as they share a common intermediate. This must also be true for the dexylosylation step of XynBHe(E160A). The rate of the xylosylation step, however, is expected to vary significantly as the rate constant governing that step should be dependent upon the leaving group ability of the phenol. In cases where the substrate has a leaving group with a pKa significantly greater than that of the acid/base residue the rate of the glycosylation step would be expected to be increased by protonation of the glycosidic bond in the transition state. Conversely, highly reactive substrates bearing a leaving group that has a pKa significantly lower than that of the acid/base residue would not be expected to be able to derive significant benefit from protonic assistance in the glycosylation step. Thus, in the case where the carboxyl group of the acid/base catalytic residue has been removed, we expect significantly lower deglycosylation rates for all substrates and lower values of the rate constants governing the glycosylation step for substrates bearing poor leaving groups. The corollary is that for substrates with an aglycone having a pKa significantly lower than that of the aciaVbase catalyst residue the rate of the glycosylation step should be similar to that observed for the wild-type enzyme, providing that other deleterious effects are not introduced by the mutation. Indeed such a phenomenon has been observed in several cases.55-65'66-187'190 Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 83 The Michaelis-Menten kinetic parameters for the hydrolysis of 2,5DNPX, 3,4DNPX, pNPX, oNPX, and mNPX by both XynBH 6(E160A) and X y n B H 6 were determined (Table 2.4). The mutation of the N B X labeled glutamic acid residue (Glu 160) to an alanine residue results in an enzyme with impaired, yet significant, catalytic activity. The appreciable residual activity and the chemical rescue (vide infra) strongly suggests that the enzyme is properly folded. Moreover, the extent to which the catalytic activity of XynBH6(E160A) is impaired is strongly dependent on the leaving group pKa of the substrate (Table 2.4). Table 2.4: Kinetic parameters for the hydrolysis of aryl xylosides by XynBH 6 and XynBH6(E160A) carried out at pH 6.5. Substrate pKa Enzyme kcat (S ) Ratio kca, K m Ratio (kcat /Km) [Wt/E160A] (uM) (s"1 mM"1) [Wt/E160A] 2.5DNPX 5.15 Wild-type 7.3 52 11 670 48 E160A 0.14 10 14 3.4DNPX 5.36 Wild-type 8.9 66 10 900 93 E l 60 A 0.13 14 9.6 pNPX 7.18 Wild-type 9.7 1.4 x 103 36 270 5.8 x 103 E l 60 A 6.7 x 10"3 1.4 x 102 4.6 x 10"2 oNPX 7.22 Wild-type 9.9 660 46 220 1.2 x 103 E l 60 A 1.5 x 10"2 80 0.19 mNPX 8.39 Wild-type 11 1.3 x 104 150 73 4.2 x 105 E l 60 A 8.6 x IO"4 4.9 x 103 1.8 x 10"4 Conveniently, steady-state determinations of the second order rate constants, kcJKm, permit access to information regarding the first irreversible step along the reaction coordinate; the xylosylation step. Br0nsted plots of l o g ( & c a t / £ m ) against pKa of the aryl leaving group for both enzymes provide good correlations, with Pig = -0.31 for the X y n B H 6 and Pig = -1.4 for XynBH 6(E160A) (n = 5, r = -0.98, Figure 2.27a). From the complete data set for XynBH6 it is already known that there is an unexpected deviation from linearity in this plot that is masked by the absence of sufficient data points (vide supra). Thus any comparative analysis between wild-type and mutant enzyme using these data will be shunned. We can, however, confidently use the Pig value (-0.97, n = 8, r = -Chapter 2 Mechanism ofT. saccharolyticum 6-Xylosidase 84 0.98) determined for XynBH6 using substrates with poor leaving groups in comparative analyses. For the mutant enzyme the large, negative slope of the Br0nsted plot of log(&Cat/^m) against pKa is consistent with the xylosylation step being the first irreversible process on the reaction coordinate. This assumption is reasonable given that upon deletion of the catalytic acid/base residue the interaction between phenol and acid/base residue is lost and so internal return resulting from low off rates of the phenol is unlikely (vide supra). The ratio of the second order rate constants for the wild-type and Glul60Ala mutant are listed on the table so that the increasingly deleterious effect of the mutation as a function of the pKa of the leaving group is made apparent to the reader. As described above, this effect is consistent with the role of Glul60 as the acid/base catalytic residue. a Ml O 4 5 6 7 8 9 10 b O 4 5 6 7 8 9 10 Figure 2.27: Br0nsted relationships for XynBH6 (O) and XynBH6(E160A) (•). The data for the plots of (a) log (kcJKm), and (b) log (&cat) were taken from Table 2.4. We are now in a position where we can estimate the importance of the acid/base catalytic residue to catalysis by XynBH6. By subtracting the Pig values (obtained from plots of log(fcCat/£m) against pKa) for the Glul60Ala mutant (Pig = -1.4) from that of the wild-type enzyme (Pig = -0.97) we can obtain a APig(fcat/A:m) value. This APig(fcat//rm) value of 0.43 informs us about the contribution of the acid/base catalytic residue to stabilizing the transition state leading to the formation of the glycosyl enzyme intermediate per pKa unit of the leaving group. By subtracting the linear functions obtained in each Br0nsted analysis and extrapolating the resulting line to the pKa of a sugar hydroxyl group (pKa = 13) we can roughly estimate the contribution of acid catalysis for disaccharide substrates. Using these rather loose assumptions we can roughly estimate that acid catalysis Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 85 contributes close to 10 kcal/mol toward stabilization of the transition state leading to cleavage of a simple disaccharide. Such an estimate is certainly consistent with the known deleterious effects of deletion of the acid/base catalytic residue in many glycosidases and would result in an approximate 108-fold decrease in rate of cleavage of simple disaccharides. Such a tremendous rate decrease is consistent with the apparent absence of activity in acid/base catalyst mutants of the Bacillus circulans xylanase 1 7 5 on xylan and of hen egg white lysozyme on peptidoglycan.163 This effect is completely consistent with the role of Glul60 as the acid/base catalytic residue, and the significant slope observed (APig^cat/Km) = 0.43) indicates that acid catalysis plays a critical role in the xylosylation step. Analysis of previously published data from kinetic studies of acid/base catalyst residue mutants of the Agrobacterium sp. P-glucosidase and the P-l,4-glycanase from Cellulomonas fimi reveal slopes corresponding to APig(kCat/Km) ~ 0.6 (although for the P-glucosidase the data is fairly scattered). The similarity between the A P i g value found in this study and those extracted from the other two studies mentioned, suggests that acid catalysis is of similar importance for all three of these enzymes and perhaps for all retaining glycosidases. Previous kinetic studies using aryl glycosides with glycosyl hydrolases in which the acid/base catalytic residue has been mutated have revealed that for substrates bearing poor leaving groups the deglycosylation step is slowed to a greater extent than the glycosylation step. However, in this study the opposite appears to be the case. The Br0nsted plot of log kcat for the hydrolysis of a series of aryl xylosides by the wild-type enzyme against pKa of the leaving group reveals no dependence on leaving group ability in the pKa range of 5-8 (Figure 2.27b). This suggests that the dexylosylation step is rate limiting for these substrates and indeed studies outlined in section 2.4 have confirmed this view. With the Glul60Ala mutant the Br0nsted plot reveals a strong dependence on leaving group ability (Pi g = -0.65, n = 5, r = -0.96) indicating that the xylosylation step is partially rate determining (Figure 2.27b) at least for highly activated substrates (pKa ~ 5). Supporting the view that the xylosylation step is only partially rate determining for activated substrates is that the slope p i g for log(&cat) is less negative than the P i g from the plot of log ikcJKm). Therefore, this difference in slope between plots of log (£ c a t) and log {kcJKm) may be a consequence of the strong dependence of the slope on the data from Chapter 2 Mechanism ofT. saccharolyticum 6-Xvlosidase 86 the highly activated substrates skewing the plot of log (kcat). If the xylosylation step is only partially rate determining for these activated substrates such a difference in Pig is what would be expected, because kcJKm reflects only the first irreversible step while &cat reflects a composite of all rate-determining steps. 1/Jfecat = 1/ Jfci + llk2 + 1/Jfc3 From examination of the Br0nsted plots it appears that mutation of Glu 160 results in a change in rate determining step for the substrates having leaving groups of approximately pKa 7.0, from dexylosylation to xylosylation. 2.5.8 Product Analysis As described earlier in this section (2.5) the monitoring of rates of hydrolysis upon addition of exogenous nucleophiles such as azide can be used in conjunction with site directed mutagenesis as a diagnostic tool to establish the identity of the acid/base catalytic residue. 5 0 The azide acts to intercept the glycosyl enzyme and increase the rate of deglycosylation. The net result for retaining (3-glycosidases is the formation of a P-glycosyl azide product and an increase in kcat. Thin layer chromatographic analysis of reaction mixtures containing both 3,4DNPX and 2000 m M azide and either XynBH6 or XynBH6(E160A) revealed the formation of a non-uv active product distinct from xylose (R f = 0.28) and 3,4DNPX (R f = 0.72) but identical in mobility to a standard of P-D-xylosyl azide (Rf = 0.64) prepared by chemical synthesis. 1 9 1 Additionally, analysis of a mixture of the products from both reactions by *H N M R revealed signals diagnostic of P-D-xylosyl azide ( J H N M R (400 M H z , DMSO-d6) 8: 4.42 (1 H , d, JU2 = 8.5 Hz, H- l ) , 3.75 (1 H , dd, /5eq,5ax = 11.1 Hz, 75eq,4 = 5.2 Hz, H-5 e q)). The other xylosyl azide signals were obscured by signals arising from solvent, buffer, and other products. The formation of a P-azide product has been shown to be useful for diagnosing the function of the mutated residue as the catalytic acid/base. In previous studies, no glycosyl azide product was observed when the wild-type enzyme was incubated with both substrate and sodium Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 87 azide. The observation that both the XynBH 6 and the Glul60Ala mutant yield significant quantities of (3-azide product is unexpected and forces careful analysis of the kinetic data. 2.5.9 Effect of Competitive Nucleophiles on Rates of Hydrolysis by XynBH6(E160A) Having established that xylosyl azide is formed during the enzymatic reaction between either XynBH6 or XynBH6(E160A) in the presence of both substrate and azide we can proceed to analyze the kinetic behavior of these systems. To this end, the rate of cleavage of a fixed concentration of 3,4DNPX by both the XynBH6(E160A) and XynBH6 were determined in the presence of varying concentrations (0-2000 mM) of sodium azide (Figure 2.28): In interpreting these data it is important to consider what step is rate determining. For the wild-type enzyme acting on 3,4DNPX the rate determining step is clearly dexylosylation (vide supra) while for XynBH6(E160A) the Br0nsted plot of log(&cat) against pKa of the phenol leaving group clearly indicates that xylosylation is at least partially rate determining. There is some evidence, however, stemming from the Br0nsted plots, that xylosylation is indeed only partially rate limiting for XynBH6(E160A) with the most activated substrates (pKa ~ 5). If the xylosylation step were entirely rate determining for the mutant enzyme we would expect only a very small effect (<2 fold) from added nucleophiles on the rate of hydrolysis of 3,4DNPX that may result from the anion assisting that step by acting as a general acid/base catalyst.66 0.25 0.2 • | 0.15 «J 0.1 0.05 0 0 0.5 1 1.5 2 [Azide] mM Figure 2.28 Effect of azide on the rate of hydrolysis of 3,4DNPX (49 uM) for 2.3 pg/mL XynBH 6 (0) and 31.2 pg/mL XynBH6(E160A) (•). Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 88 The 8-fold increase in the rate of cleavage of 3,4DNPX upon inclusion of azide observed with the mutant enzyme strongly suggests that the deglycosylation step is substantially rate limiting for XynBHe(E160A) with this substrate, consistent with the interpretation of the Br0nsted plots. The increase in rate observed with the wild-type enzyme is surprising and suggests that azide can intercept the xylosyl enzyme despite the presence of the acid/base catalytic group. This is in marked contrast to previous studies on other glycosidases where the absence of any effect of added azide was rationalized on the basis of the charge of the carboxylate side chain of the acid/base residue screening the active site from the negatively charged azide.50 Clearly charge screening does not prohibit azide from competing with water for the XynBH6 xylosyl enzyme intermediate. Why the xylosyl enzyme studied here is more susceptible to attack by azide than the glycosyl enzyme intermediates from Agrobacterium sp. P-glucosidase or the xylanase from Cellulomonas fimi is unclear. Regardless, the effects of other nucleophiles including formate, acetate, and fluoride were also investigated in the hope that greater rate increases could be observed. These alternative anionic nucleophiles, however, were found to give rise to much smaller rate accelerations as compared to azide (Figure 2.29) and thus azide was used for all the detailed chemical rescue studies. 0.28 r 1 0 0.5 1 1.5 2 Nucleophile [M] Figure 2.29 Effect of added exogenous nucleophiles on the rate of 3,4DNPX hydrolysis by XynBH 6 (empty symbols) and XynBH6 (filled symbols). Data corresponding to fluoride addition is indicated by circles; formate addition by squares; and acetate addition by triangles. A pH 6.5 was carefully maintained in all cases. Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 89 Additional experiments using a range of sodium chloride concentrations revealed that there was little effect on the rate of 3,4DNPX hydrolysis by XynBH6 arising from ionic strength. Indeed, 2 M sodium chloride resulted only in a 20% decrease in £ c a t . The superiority of azide as an exogenous nucleophile in chemical rescue studies of glycosidases has been demonstrated in several previous cases.65-66 Having decided on the use of azide for the chemical rescue studies, we conducted a more detailed analysis of the effect of azide on the XynBHe(E160A) mutant enzyme. The rates of cleavage of a range of concentrations of 3,4DNPX (typically 0.3 - 5 x Km) by XynBH6(E160A) were determined in the presence of varying concentrations (0-2000 mM) of sodium azide. Both kcat and Km increased as a function of azide concentration (Figure 2.30), leveling off at higher concentrations. In marked contrast, kcJKm remained essentially unchanged or even slightly decreased across the range of azide concentrations studied. The absence of any increase in kcJKm implies that azide has no effect on the xylosylation step. The lack of any significant effect on kcJKm is reasonable, as it is difficult to envision a situation where azide could act to increase the rate of glycosylation. Indeed, insensitivity of kcJKm to azide concentration has been observed in such studies previously.65-66 Figure 2.30 Effect of varying concentrations of azide on the kinetic parameters of the XynBH6(E160A) xylosidase catalyzed hydrolysis of 3.4DNPX. Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 90 The increase in kcal upon addition of azide and the formation of a P-xylosyl azide product suggests that azide is a more effective nucleophile than water in reacting with the xylosyl enzyme intermediate. The observed increase in Km as a function of azide concentration is consistent with this interpretation. The consequence of azide having an effect only on the second step of the reaction is to decrease the steady-state concentration of the xylosyl enzyme intermediate (see scheme 2.22). An expression for Km, pointed out earlier in this chapter, that clarifies this interpretation directly reveals that as the concentration of enzyme bound species decreases, the Km increases. For convenience the equation is listed here once again: Km = [E][S]/2[Enzyme bound species] The observed increase in kcat with increasing azide concentration is similar to the small effect observed with the acid/base mutant (Glu 127Ala) of the Bacillus circulans xylanase175 (8 fold) and also the family 39 P-xylosidase from Bacillus stearothermophilus (8-fold).188 Most other studies have revealed a much greater effect of azide on rate (up to several hundred fold) but in these cases the rate-determining step was, in all instances, the deglycosylation step. The moderate increase in kC3X observed here with XynBHe(E160A) is therefore consistent with the kcal being governed primarily by the xylosylation step as evidenced by sloped Br0nsted plots of log kcat vs pKa and the relatively small magnitude of the azide effect (8 fold). Consequently, the leveling off of r c c a t as the azide concentration is increased likely arises from the xylosylation step becoming solely, rather than partially, rate determining. Although azide does have an effect on XynBH6 the effect is pronounced with XynBH6(E160A). The effects that are observed lend further support for the assignment of Glu 160 as the catalytic acid/base residue. Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 97 2.5.10 Sequence Conservation and Alignment Multiple sequence alignments of the members of family 39 glycosyl hydrolases revealed sequence similarities between the xylosidases and iduronidases of this family to be less than 16%. This overall sequence similarity is primarily accounted for by a few discrete regions that demonstrate significant sequence similarity across all family members, such as the segment approximately 100 residues before the catalytic nucleophile. This section of the sequence contains the glutamic acid residue (Glul60) that has been labeled and mutated within the studies outlined within this section. This residue is completely conserved throughout all members of family 39 of the glycoside hydrolases and is contained within the short conserved sequence i5oAsn-Glu-Proi6i (Figure 2.31). xynb-thesa 137 WSHFIERYGIEEVRTWLFEVWNEPNLVNFWKDANKQEYFKLYEVT 182 xynd_calsa 93 WKHFIDRYGEKEWQWPFEIWNEPNLNVFWKDANQAEYFKLYEVT 138 xynB_calsa 140 LARHLISRYGKNEVREWFFEVWNEPNLKDFFWAGTMEEYFKLYKYA 185 idua_human 159 LARRYIGRYGLAHVSKWNFETWNEPDHHDFDNVSMTMQGFLNYYDA 204 idua_canfa 158 LARRYIGRYGLSYVSKWNFETWNEPDHHDFDNVTMTLQGFLNYYDA 203 idua_mouse 149 LARRYIGRYGLTHVSKWNFETWNEPDHHDFDNVSMTTQGFLNYYDA 194 CONSENSUS : •_**** * * * * **£*. * . * * . Figure 2.31 Partial multiple sequence alignment of the enzymes comprising family 39 of glycosyl hydrolases. The consensus sequence is shown at the bottom of the alignment, with (*) indicating fully conserved amino acid residues and (:) indicating similar residues. Numbers to the left denote residue positions. The abbreviations used, references to the published sequence and data bank accession numbers are as follows: Xynb-thesa, P-xylosidase from T. saccharolyticum (152, SwissProt identifier P36906); xynb-calsa, P-xylosidase from Caldocellum saccharolyticum (157, SwissProt identifier P23552); idua-human, a-iduronidase from Homo sapiens ( 1 5 8 , SwissProt identifier P35475); idua-mouse, a-iduronidase from Mus musculus (15^, SwissProt identifier P48441); idua-canfa, a-iduronidase from Canis familiaris (1(^, SwissProt identifier Q01634). The acid/base residue in P-xylosidase identified in this work (Glul60) and the corresponding residues in other members of this family are indicated by (•). This sequence is reminiscent of the Asn-Glu-Pro sequence within which the acid/base catalytic residue is found in several other p-glycosidases, including family 1 Agrobacterium sp. P-glucosidase and the family 10 xylanases. Both enzyme classes, Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 92 along with enzyme families 2, 5, 17, 26, 30, 35, 39, 42, 53, have been assigned as members of clan G-H A, a superfamily of glycosyl hydrolases with an a/|3 barrel fo ld . 1 4 7 The results of the kinetic analysis of XynBHe(E160A) and the complete conservation of this residue across all members of this family provide very strong evidence that Glu 160 is indeed the acid/base catalyst. 2.6 Summary The results of this chapter allow the unambiguous assignment of both the catalytic nucleophile as Glu277 and the acid/base catalytic residue as Glu 160. Additionally, the kinetic isotope effect studies carried out with XynBthj provide strong evidence indicating that the reaction mechanism of XynB proceeds via a covalent intermediate in common with all other retaining P-xylosidases studied to date. Kinetic analysis of the XynB acid/base catalytic residue mutant, XynBH6(E160A) has also allowed a crude estimate of the importance of acid catalysis (10 kcal/mol of transition state stabilization energy) in the reaction of XynB with its native substrate. The Br0nsted analysis of XynBH6 is very similar to that found for several retaining P-glucosidases and indicates that protonation of the glycosidic oxygen in the transition state is not highly advanced. The net result of these biochemical investigations is that XynB appears to be a fairly typical retaining exo-(3-glycosidase. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 2.7 Methods and Materials 93 2.1 A General Procedures for Chemical Synthesis All buffer chemicals and other reagents were obtained from the Sigma/Aldrich/Fluka Chemical Co. unless otherwise noted. Solvents and reagents used were either reagent, certified, or spectral grade. Anhydrous solvents were prepared as follows. Methanol was distilled from magnesium turnings in the presence of iodine and tetrahydrofuran was distilled from sodium in the presence of benzophenone. Dichloroethane, toluene, pyridine, benzene, triethylamine, acetonitrile, and dichloromethane were prepared by distillation from calcium hydride. Solvents were distilled immediately prior to use. Dimethylformamide was dried sequentially and stored over 4 A molecular sieves. Synthetic reactions were monitored by TLC using Merck Kieselgel 60 F254 aluminum-backed sheets (thickness 0.2 mm). Compounds were detected by ultraviolet light (254 nm) and/or by charring with 10% ammonium molybdate in 2 M H2SO4 and heating. Flash chromatography under a positive pressure was performed with Merck Kieselgel 60 (230-400 mesh) using the specified eluants. *H NMR spectra were recorded on a Bruker WH-400 spectrometer at 400 MHz, a Bruker AV-300 at 300 MHz, or a Bruker AC-200 at 200 MHz. Chemical shifts are reported on the 8 scale in parts per million from tetramethylsilane (TMS) and were measured relative to CDCI3, CD3OD, or to DSS when taken in D2O. The abbreviations used in describing multiplicity are: s-singlet, d-doublet, t-triplet, m-multiplet, and br-broad. The Mass Spectrometry Laboratory, University of British Columbia performed both high and low resolution mass spectra. Mr. Peter Borda of the Microanalytical Laboratory, University of British Columbia performed elemental analyses. Melting points were recorded using a Laboratory Devices Mel-Temp II melting point apparatus and are uncorrected. Chapter 2 2.7.2 Generous Gifts Mechanism ofT. saccharolyticum (^Xylosidase 94 A number of the aryl 3,4,6-tri-O-acetyl-P-D-xylopyranosides (2.3, 2.4, 2.6, 2.7, 2.9, 2.11, 2.15) and aryl (3-D-xylopyranosides (2.20, 2.22, 2.25, 2.27, and 2.29) were generously provided in pure form, without N M R data, by L l o y d MacKenzie , a doctoral candidate, and Dr. Lothar Ziser, a postdoctoral fellow, both working in the laboratory of Prof. Withers. L l o y d MacKenzie also provided the sample of 2 F - D N P X and B T X . Prof. Greg Zeikus donated the p H X P 3 plasmid bearing the X y n B gene. 2.7.3 Syntheses N-Bromoacetyl /5-D-xylopyranosylamine Bromoacetic anhydride 1 9 2 (121 mg, 2.15 mmol) was added to a solution of (3-D-xylosylamine 1 9 3 (152 mg, 1.02 mmol) in D M F (1.5 mL) over 15 min. The mixture was stirred for 1 hour at room temperature, poured into ice-cold anhydrous ether and stirred for 1 hour. The ether was decanted and the residual gum crystallized from M e O H to yield the desired product as fine white needles (61 mg, 22%); mp 150-151°C, ! H N M R (400 M H z , C D 3 O D ) 8: 4.81 (1 H , d, Jh2 = 8.8 Hz , H - l ) , 3.87 (2 H , C# 2 Br ) , 3.83 (1 H , dd, J5A = 5.2 Hz , 75,5' = 11.2 Hz , H-5), 3.47 (1 H , ddd, / 4 , 3 = 8.8 Hz , 7 4, 5 ' = 10.3 Hz , H-4), 3.35 (1 H , dd, 7 3 )2 = 8.8 Hz , H-3), 3.26 (1 H , dd, H-5'), 3.25 (1 H , dd, H-2). Ana l Calcd for C 7 H i 2 B r N 0 5 : C , 31.13; H , 4.48; N , 5.19. Found: C, 31.52; H , 4.40; N , 5.12. General synthesis of aryl 3,4,6-tri-0-acetyl-f5-D-xylopyranosides (2.2, 2.5, 2.8, 2.10, 2.12,2.16,2.17) To a mixture of 2,3,4-tri-Oacetyl-a-D-xylopyranosyl b romide 1 9 4 (1 eq.), tetrabutylammonium hydrogensulfate (1 eq.) and the acceptor phenol (2 eq.) was added sufficient dichloromethane (1 volume) to yield a solution containing the donor bromide at a concentration of 200 m M . A n equal volume of 1 M N a O H (1 volume) was then added Chapter 2 Mechanism ofT. saccharolyticum 0-Xvlosidase 95 and this mixture was rapidly stirred at room temperature for 1 to 3 hours. After the reaction was judged complete by T L C analysis, ethyl acetate (5 volumes) was added. The solution was washed with 1 M NaOH (4x2 volumes), water ( 2 x 2 volumes) and saturated sodium chloride solution (2 volumes). The organic layer was dried (MgSCu) and filtered, and the solvent was removed in vacuo. The resulting crude product was crystallized from a mixture of ethyl acetate and hexanes to provide yields, after collecting one crop of crystals, of between 55 and 80% of the desired aryl glycoside (2.2, 2.5, 2.8, 2.10, 2.12, 2.16, 2.17, Figure 2.32). ! H NMR data for the per-O-acetylated aryl xylopyranosides can be found in Tables 2.7 and 2.8. 'H NMR data for the aryl xylosides prepared by Lloyd MacKenzie and Lothar Ziser can also be found in these tables. All of these compounds except for 2.3, 2.7, and 2.14 have been previously prepared, however, NMR spectral data for these compounds was not reported.1 9 5"1 9 7 General synthesis of aryl fi-D-xylopyranosides (2.23, 2.24, 2.26, 2.28, 2.30-2.33) To a stirred solution of the specific aryl 2,3,4-tri-O-acetyl-P-D-xylopyranoside (2.5, 2.8, 2.10, 2.12-2.14, 2.16, 2.17, Figure 2.32) in anhydrous methanol was added a catalytic amount (1-5 drops) of a 1 M solution of sodium methoxide in methanol. The reaction mixture was allowed to stir for 2 to 4 hours at room temperature. In cases where a precipitate formed this was filtered and washed with methanol. The mother liquor of this reaction, and all other reactions, was processed in the same manner. Excess Amberlite IR-120 resin (H+) was added and the reaction mixture was filtered. The filtrate was concentrated in vacuo to provide a solid. Recrystallization of all compounds was accomplished using a mixture of methanol - ether - hexanes. Yields of the desired product after one recrystallization ranged from 45 to 85 %. lH NMR data, melting points and elemental analyses for compounds 2.23, 2.24, 2.26, 2.28, 2.30-2.33 can be found in Tables 2.6, 2.9, 2.10. ^ NMR Data for the aryl xylosides prepared by Lloyd MacKenzie and Lothar Ziser can also be found in these tables. All of these compounds except for 2.23 and 2.31 have been previously prepared, however, NMR spectral data for these compounds was not reported. 1 9 5 - 1 9 7 Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 96 General synthesis ofl-{2HJ- and 2,5-dinitrophenyl-(3-D-xylopyranosides (2.18 and 2.19) To a cooled (0 °C) stirred solution of either 1-{2H}- or 2,5-dinitrophenyl tri-0-acetyl-3-D-xylopyranoside (2.2 or 2.3, Figure 2.32) in anhydrous methanol (10 mg/mL) was added acetyl chloride in sufficient quantity to generate a final HC1 concentration of 5% (w/v). The reaction mixture was warmed to 4 °C and allowed to stir for 16 hours. The reaction mixture was concentrated in vacuo to provide a gum. Dry diethyl ether was added and the solid triturated after which the solvent was removed in vacuo. This process was repeated 6 times in order to remove any traces of HC1. The resulting gum was crystallized and then recrystallized using a mixture of methanol - ether - hexanes. Yields of the desired products after recrystallization were 45 and 61 % for compounds 2.18 and 2.19 respectively. *H N M R data, melting points and elemental analyses for compounds 2.18 and 2.19 can be found in Tables 2.6, 2.9, and 2.10. General synthesis of 1-{2HJ- and ortho-acetamidophenyl tri-0-acetyl-fi-D-xylopyranosides (2.13,2.14) The atmosphere of a solution of orr/io-nitrophenyl tri-0-acetyl-J3-D-xylopyranoside (140 mg, 0.35 mmol) (2.6 or 2.7) suspended in methanol (10 mL) was replaced with nitrogen and a catalytic amount of palladium on carbon was added. The atmosphere above the reaction was replaced with hydrogen and the mixture was stirred for 14 hours at which time the reaction was judged by T L C analysis to be complete. The reaction mixture was filtered through glass fiber filter paper and concentrated in vacuo to yield a white powder. To the resulting crude amine was added pyridine (5 mL) and acetic anhydride (1 mL). The mixture was stirred for 2 hours after which time the solvent was evaporated in vacuo to yield pale yellow crystals. Recrystallization of this material from ethyl acetate - hexanes afforded the title compounds 2.13 (110 mg, 0.27 mmol, 77 %) and 2.14 (96 mg, 0.24 mmol, 69 %). *H N M R data for these compounds can be found in Tables 2.7 and 2.8. These compounds, with the exception of the C - l deuterated analogues, have been previously prepared, however, their N M R spectral data was not reported. 1 9 5" 1 9 7 Chapter 2 Mechanism ofT. saccharolyticum B-Xvlosidase 97 Figure 2.32 Generalized structure of the series of aryl P-D-xylosides used in this study. The identities of the R-groups can be found on the subsequent page in Table 2.6 Chapter 2 Mechanism ofT. saccharolyticum (3-Xvlosidase 98 3 o PH bo e CJ U HH 3 O on Pi 0 ? a; T3 C >-. 3 CJ 5 -2-6 I £ T3 C 3 O a, £ o U N N T3 T3 C S a a I ) cj 'N 'N c c <o cj N N § § _o cj 'N 'N c c a) o M M u CJ X I X> CJ -a '> o T3 CJ •a '> o T 3 cj •O o N T O § CJ ' N c u ^ ! CJ s x> T3 CJ "H "> O N • CJ o C3 ;>-. X I -a CJ -a '> o N T3 s a e CJ cj c3 X I T3 CJ •a '> o oo Z Z oo cj r ( cfl r- N oo -o z § . .. cj St 'S o u U .2 -Z N CJ 12 N PH PH PH PH U o d CN (_)" >/"! , — 1 o & "c3 6 U O O 2 ^ ffi ffi ffi ffi X 33 u i n 3 - 5 u £ u * & ^ W 6 U o z c-i CJ X s C N S CJ X ! u N 2 Q Z Z S 8 £ ^ T3 u - 8 & s W PH CJ c o •SP -5 o3 2 >. >. x> x> -a -o C O CJ CJ 3 " O — C H > & = 2 E PH CJ 3 u u 1 — 1 o 2- "=3 O O XX'Z'ZXXXXX x x z x x x w z u x & x x x u x x o o o o Z Z E K Z Z K K K K S u o c j o o o c j c j o c j o o c j c j c j o < < < < < < < < < < < < < < < < x o z ON CN X Q C N 5 5C ffi W ffi tN (N o o Z Z, X X o ^ S ffi o o ffi K Z Z X ffi ffi X © - H N n CN CN CN CN tN CN CN CN X ^ • - u ->r in 2 co 0 \ ^ •* U U T t CN Q .^ z z CO C H O X o X X Q Z ^ K X u 5 ro" Chapter 2 Mechanism ofT. saccharolyticum 6-Xylosidase 99 o o\ VO Z N -o c a .« 'S3 c 5 T 3 |U •§ o a r i in c> 7,, >—1 >—1 u E E u N — 'N . c m o X £ CN 2 ^ >, O X l U -S u u <D S 00 w U S '§ iri o X & CN in U m m Z cb" IT) CN vd ^ xi r-^  >n IS r^ - s u6 £ U u u o O 0 M Ov VD EG r - m ^ o + w 6 E u T3 u-8 &1 OO I/-) v o <N oo m T-I o VO CN ro in U d ac in o + in vd in I - H X © w 6 u vo o"* I i q cs 3 K S 3 c i . 1 - 1 i i W 6 U Ov CN vd X u <K N N ^ W 6 U VO in U & x X X X X X X i - X ! « OH u X u o o _ z u X OH X u «' X X Z X X X X V) vo CN fN CN r i CN fN fN fN OV © CN ff> r i r i I-I cn r i fN r i r i 1/5 « 0 cn X t-l m x z X OH X z o X ca X o . ID Chapter 2 Mechanism ofT. saccharolyticum B-Xylbsidase 100 e, I S3 O. 3 i < 3 CN O •9 Tt 3 2 CN ca to 2 3 m o oi CN o < o X CN r -o CJ < O o x < CN O t - - v o o o CN CN CJ VO ^ P * CN co cj < o O < X O CN Ov OO o o < o X c o o o o 3 T t 3 2 CN « a GO o o i n a v o o o i n o o p i n i r -3 o o < o r o <H i n "-"T X o Hz Hz Hz Hz Ov Ov CN CN PP PP PP CN PP CN PP CN CN .o CN CN o CN -O 4 O CN Xi in Ov Tt Tt Tt Tt T 3 c3 C3 T3 0 0 13 v o O T t o o i n - a T3 T t r o CC CN 0 0 cc1 m Ov X o MH MH T3 0 0 Tt " a od i n VO to Xi m I - , --j N _ CC • § CN T t ° ° c n T 3 T 3 3 3 3 t~- o o m Tf TT Tt >-> e VO i n o CO i n * d co ov 7 3 Tt VO 0 0 X ) Ov n 3 T 3 • O Tt CN m Ov B v o o CO T 3 • a t o CN ^> CN ? o VO ^ 1—H m 3 T 3 ~ o v r-Ov <H T t X CJ 6 v o o CO 3 X PH T3 c o CN CO <H i n X OH CN CN 233 3 C l V\ V\ Tt Tt ' Tt" T t S S "S o c o CN O c o CN o c o CN X u 5 i n r o " < O x r o O CN T3 CC T3 CC • o p T3 p T3 i n i n • o -a - a • a T 3 Tt i n ^ - i—i o o r - VD v q xi v q Xi i n i n i n v q i n v q r o r o r o CO CO CO CO* CO cc CN T3 CN q £ CN a Tt N N X CC " O Ov Ov 7 3 ro ' i n ' Ov "1 Tt "I "~J CC CN ^ CN m ^ N CC o o vd N CC v o Tt X Q T t r o " Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 101 O o o O O < < < < < < < < o o o u o u o o o o X X X < X < X X X X CO CO r o o CN o CN r o r o r o c/T •A vT <A (A O r~ Os vo eo vo r o r - as r-o O i— i *™™• H « o o o CN CN CN CN CN CN CN CN CN CN 3 N •a « 2 as CN CN • . <n N SC o od Os s o C u -o T3 oo T t co 3 N ? vo ov r-Os Tt OO O vo o Tt 00 ra X o in m VO CN 00 o VO CN oo O «/•> vo CN X Q Tt cn X Tt o >—1 PP CN PP CN PP Os x 4 o CN X VO CN Tt Tt Tt oo ON Tt oo o Tt C O T i -r o Tt r o X T3 •a U O -o >/0 r --o -a i n r o T3 -a o vo 3 N N N is T 3 SC sc •o SC SC -o T 3 •—1 l-H -o •a •a -o T 3 O a x o X OO Ov Vi Vi ON Ov Tt Tf' T t £ £ £ -OSS •° S -* C  a x — - in in •a T 3 •a o r -o <*i N SC vo --H Tt r--Tt I T ) ^ T3 — T3 ° N r- ^ O I T ) T3 T3 Tt ^ uo CN r o CN U O L - > 3 o\ Tt vo CN r o CN vo CO V"-> <H V0 < X < z X Q V O CN vo CN • O <N 00 ^ CN r j vi "-T Q vo CN X I-I PQ T t .-a -o r o N SC oo oo CN -a CN T3 CN • T3 CN •a CN T 3 CN '—i -a *—* T3 T3 T 3 ^ - i X vo UO XI in oo X in CN X o X rf CN CN '—1 CN CN Tt Tt Tt Tt Tt N N - 3 ^ SC SC T3 i n r o S - ^ t^x in in CN vo CN B o I T ) CN X u r o Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 102 C3 fa 1 = " lx* f N O O I <D X J o 1 CO. "">, t3 o •4 ro" of o DO c I ) X i O- > 1) * C3 ai X J C 3 I o U 3 o-r o co — ' 2 vo OS z Sd u o u oi o oo oi X J X J oo O) CN Ov vd i Ov o Sd oi OI Ov Ov O Sd Sd XX r o r o X J ^ . 2 S os X ! 5 3 T t -d-' ~. % % r~ rN Ov vd o \ o ^ H r-; ^ H C-; oo oo ~H od r—* v d PP AR4 AR5 PP AR4 AR5 PP AR4 0 0 AR3, m oo r*i Ov m AR3, AR r-; AR OS< r o OS < t ~ --5 0 0 < 2 Z o U o u oi o 0 0 OI OI Ov vd - I Ov o O) Ov v d i Ov O OI Ov v d I Ov O f j ^ j N N N N N N H H H H H H H H H H H M MH * H 0 0 CN vd oi 2 os os X J < < Ov <*5 t^ '. oo < Z X CJ z^  33 p p 0 0 v—< X ) X ) 3 CO xi 2 0 0 io 01 Oj X ) m o ir-" 3 oo xi S X J § 0 0 fi" 01 g 3 O N 0 0 X cn r r oS as < < CN* C N OS OS < < V O >-> X ) X J r» Ov X PH < OS OS < < r— vo" vo" <*. % %. • vo >•-» r-x) vo o < X ) OV V O o N X o Ov c-l OS < as X OH < Z o oi Ov vo vo 3 o Ov OS X ) < ^H as od ^ Z no o < Z o oi - 1 O N ( N °°. V O X U 5 ro" z Sd U o < Z iz z X ) o oo vd vo oi o-r o Ov od cn •a Sg r o PJ" V O --5 X Q T t ro" X tu S o OH u u ov oi <—i X J X ) X ) X ) X J V O CO OI O j Ov r— c~ v d v d N X p Ov ^ H Ov od od vo vO X J OS < X J as < X J OS < o wi" o m" OI vn p OS < oo as < p OS < od v d < 3 3 3 3 oo vd ^ X u s Q T t ro" Chapter 2 Mechanism ofT. saccharolyticum B-Xvlosidase 103 < Z\ X ro vo ro as T t X ro as X Z < z < z: < z < Z < Z o CN T 3 O O ' O ON < T3 O „ Os g 2 X Q vo CN 2 o o i T 3 T3 O O O ON T-TD % as % Z X Q vo CN S OO oo OO vn ro g Z oo r o OO SO 00 VO X m N O od ro T t -O od CN -a m vo < < T3 CN as t-^  >-? < Z SS IT) C«1 CN CN • a 5 5 00 « * r- "-. < z sss PP PP T3 • a - a -o -o r - o\ T t VO vo VO H-< ^ OO o i *-H# oo r o r o -> od od vo SS o vo •-• CN a. Qi < < vo" vo" SS OO T t t~ od •O -o a. a.: < <i • < < ^ "-> as as vo o < z: Ov o\ vb I vo o X U ro Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 104 1 3 , 1— lo x Q I CO. " fe S3 fe O-I i3 3 VO X J K W •O m o \ T 3 in ob l/^  vn vn T3 O^ — i od Tt • . cn C O >~> XJ °° . oo ° ° Tt n • ^ n) co X J T3 v© 3 VO T3 X ! X J r~ in i CN i O 3 *-c ° d Tt •*. 3 Ov Tt O ° i >n "-T X o X ) X J co Tt XJ ; XJ N x> XJ XJ X XI r~ vn in VO rt CN vn o • Tt Tt Tt N N X X ) ^ Tt 2 d ^ in ^ i — i «j x i t— <n w-i r o ^ ^ 3 m 0 0 • ^ v n co "-> X o TD XJ cn Tt co 3 VO T 3 — X J v—i O Tt 3 XJ ^ Tt XJ 3 O m V-H r-; CO XJ X) m m 3 vq od JS N X X PP oo PP oo PP oo oo od o od o od Tt cn ^ CN vq cn CN vq cn CN cn ~1 CO CO X u Ov CN Tt Tt 3 S Tt X ) CN — i CN rf ca Tf CN in co oo in Ov CN Tt Tt Ov CN Tt Tt Ov CO Ov Tt 3 3 CN X J ^ CN CN rt" rt" Tt ^ ^ Tt in co o-. in ov CO Ov Tt Ov co Ov Tt X P H X OH CO CO Tt CO js 3 X Tt CN " " i 1 Tt ^ m co m CO CO Tt co co co Tt CO J3 JS N X X X X X) co XJ o X ! 00 XJ CN r-" CO i > ^ H vd VO o CN oo CN o CN 00 CN v i Tt in Tt "7 X u a in co" Tt to 3 CN CN XJ X) o £ CN rt vn Tt - l 33 vn N X J X XJ ^. o vo T3 TT • -u m ov oo 0 0 rt j in CO r— in co Tt co •n Tt CO X) ^ oo ^> O <H in '~r X Q Tt co" Chapter 2 Mechanism ofT. saccharolyticum d-Xylosidd.se 105 q x CN c o r -r o Os r o CN Os CN CN T t i n I o i n i n CN T f i n r o CN T t i n r o 00 CN c o T t r-r o T f r o r o T t c o 53 t -a •o oo oo -a -a r~ oo r o i n T3 -a so CN T t CN — > n ""1 T3 - H T3 <-i > n £ CN T t ^ T 3 >-H 7 3 — . i n £ CN rf T t -5 X) T3 00 OO T t -o T3 CN CN T t 53 TI-TS -o o CN T f CO T t i n o wo so m VD m CN so "O - t fj\ ON "° o\ OO I— « ca x> <n in un ^ 5 ^ VI " I "1. Tt" Tf Tt c o *-> ^ OO l /^ m m m Tf Tt Tf" c o ^ ^ , T t i n Os i n CN i n i n CO Os r o r-CN Os CN CN T t i n o i n i n CN T t r o i n r o CN T t i n c o oo CN C O T t f-T t c o c o c o T f c o r o c o i Os C O r -CN Os CN CN T f O i n CN T t i n c o 6 CN TT c o u o CO 00 CN c o T t r o T t r o r o T t c o js 53 53 53 53 T3 T t T3 T t T3 o T3 so Os SO O i > so SO r~ r l q CN CN CN T t T t i n u o X u o CM X tu Q T t co" X z X Q m CN" < z X Q < n CN 53 00 T f X PQ T t T3 00 m Os c i T t T3 CO T t oo n T t X u C O N 53 8 T f |-o OJ o CU . =3 Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 106 cd *4 I S 1 6 6 oo T f T-H" TD TD *-; T" TD TD TD TD r~ TD TD • o < TD TD TS 00 oo T - H T - H r - oo f - ,—i co T t CN CN co co o\ T - H o OS T t ON r-i> i> VO i> l> VO vb ro co CN o CN tu I TDI X i Q c i LH cd Si >s c cu X i & CU J 5 cd I Pi CN cd 3 I O U il uo od in vo TS ^ ~ ~ < < T t -tf* •»' ~. % % C - •-> •-> CN VO od - H ' r- *-> --> o CN 8 u o CN CN < z < z: < z z uo •o 3 T - H vn* uo Sil in vo •o ^ ^ TD 5 3 T t -VT ^* c -asas CN VO od H TD TD r-oi < 0. 0-< < f- *-> < z O T t C O 6 T t C O SS O co od - H t - *»-X o 1 6 T t C O TD TD in CN C O l> TS TD X X o C O od - - H ^ vn % 3 2 . < TS TD in o N X C O UO tn CN OJ 2 X o z o X PH s s TD ? < z od < Z TS C-q < T Z TD Ov < < z X C H X u 5 in co" X Q T t . co" X u o CN Ov od TD O oo < Z O 00 vb X <u s o X X u u Cfl" C/5* O vo CN v—i CN CN s co od vo: OO uj] VO ' Z Z 0 Hz s f—1 s oo s s O) s oo cK r-H CN od CN cn m Tf vo t n vO TS TS a: < TS AR TD AR TD TD AR co (N* 00 tN T - H tN* r - tN* oo CN ci q AR AR o AR uo AR as OC < oo a. < l> l> vb vb X 5 T t C O " Chapter 2 Mechanism of T. saccharolyticum 0-Xylosidase 107 < Z < < z < z < z < z: VO VO 3 3 3 - o XJ XJ oo Ov vd in * ^ t OO oo VO Ov fl CN OO CN fl fl Ov OO o l-" N-> 00 od VO X) fl N N N T? K ffi 3 m"0 q co m co c<i o t~ --> < Z < z 3 3 OV od 00 "5 co ?s i < - ? CN T3 od •a 3 33 CN CN od od fl TJ- vo . <~i 3" % r~~ >-> CO CN 3 3 I S 1 00 -~, 3 XJ XJ CN O od < z 3 3 3 CN fl fl 00 fl fl CN vo os os < < XJ XJ XJ r- •-> vo ov vd o r-' ^3 CN q Tf 00 N N X a Ov ON 00 CO Tt X J QS < X J AR X J OO cn oo m" CO Ov OS Ov co t~ < z z i 3 3 i n oo O r~ od ov < Z < z: N X Ov od VO CN Ov « vd < VO oo Ov OO 33 fl o CN CN XJ XJ m CN °. % < t - ^ 5 % 2 x z X Q m CN X Q i n CN X X PQ Tf X u co x) '1 o u XJ Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 108 2.7 A Molecular Biology Techniques 2.7.4.1 General Proceedures Growth media components were obtained from Difco. Plasmid-containing strains were grown in Luria Broth containing 50 p,g/mL kanamycin (LBk a n) or in TYP (16 g/L tryptone, 16 g/L yeast extract, 5 g/L NaCl, 2.5 g/L K2HPO4) containing 50 p,g/mL kanamycin (TYPkan). Pwo DNA polymerase and deoxynucleoside triphosphates were obtained from Boehringer Mannheim. Restriction endonucleases and T4 DNA ligase were from New England BioLabs. Escherichia coli TopplO cells and the pZeroBlunt cloning kit were from Invitrogen. The pET-29b(+) expression vector, E. coli BL21(DE3) cells, and His-Bind metal chelation resin were obtained from Novagen. The E. coli JM110 cells (rpsl, (Strr), thr, leu, thi-1, lacY, galK, galT, ara, tonA, tsx, dam, dcm, supE44, A(lac-proAB), [F' traD36 proAB lacqZAM15]) were from Stratagene. PCR DNA fragment purification and plasmid purification kits were from Qiagen and Promega. Preparation of oligonucleotide primers and DNA sequencing was performed at the Nucleic Acids and Peptide Service facility (NAPS), University of British Columbia. Complete sequencing of the gene within the final vector construct confirmed the desired sequence in all cases. 2.7.4.2 Amplification and Subcloning of XynB The gene encoding for the His6-tagged P-xylosidase gene from T. saccharolyticum (xynB), was amplified via polymerase chain reaction (PCR). The PCR mixture contained 10 uM oligonucleotide primers (shown below), 1 mM concentrations of the four deoxynucleoside triphosphates in 100 pi of DNA polymerase buffer and 50 ng of plasmid pXPH3, a generous gift from Dr. J. Gregory Zeikus (Department of Biochemistry, Michigan State University, East Lansing, MI). Plasmid pXPH3 contains a 2-kb Pstl - Hindlll fragment of T. saccharolyticum DNA, carrying a 1500 bp open Chapter 2 Mechanism ofT. saccharolyticum 6-Xylosidase 709 reading frame that contains the entire xynB gene. After heating the reaction mixture to 95 °C, the PCR reaction was started by adding 5 units of Pwo DNA polymerase (Boehringer Mannheim). Thirty PCR cycles (45 s at 94 °C, 45 s at 56 °C, and 70 s at 72 °C) were performed in a thermal cycler (Perkin Elmer, GeneAmp PCR System2400). Agarose gel electrophoresis of the PCR product revealed a single DNA fragment of approximately 1500 bp as estimated by gel electrophoresis. The forward primer was as follows: 5'-TAA CAT ATG ATT AAA GTA AGA GTG CCA GAT TTT-3' Ndel The reverse primer was as follows: 5'-TAA CTC GAG ATA TCC ATT TAT CTT GCT ATC-3' Xhol After purification of the PCR product using the Qiaquick PCR purification kit according to the manufacturer's protocol (Qiagen), a blunt-end ligation into plasmid pZero2.0 was performed according to the manufacturer's protocol (Invitrogen). Electrocompetent TopplO cells (Invitrogen) were subsequently transformed with the ligation mixture using a BioRad GenePulser II. Single colonies were selected and grown overnight in LBzeocin+kan, and DNA was isolated via minipreparation technique (Promega Wizard Plus kit). Restriction endonuclease mapping revealed positive clones, which were subsequently sequenced to verify the published sequence of xynB. The cloned xynB in pZero2.0 was cut out using the unique sites engineered into the oligonucleotide primers (see above). The His6-fusion protein expression vector, pET-29b(+) (Novagen), was also digested with Ndel and Xhol. A ligation reaction was performed at a ratio of 10:1 (insert to vector) using gel-purified DNA fragments and T4 DNA ligase (1 unit/10 ng of DNA) at 25 °C. The cloned product, called pET29xynBH6, was subsequently transformed into electrocompetent TopplO cells, selected by the Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 110 kanamycin resistance conferred by pET-29b(+). Single colonies were selected and grown overnight in LBkan, and DNA was isolated via the minipreparation technique. Restriction endonuclease mapping revealed positive clones. TopplO transformed cells were used for preparation of large amounts of plasmid pET29xynBH6 (Qiagen Plasmid Maxi kit) and long term storage of the vector as glycerol stock. 2.7.4.3 Site-Directed Mutagenesis by PCR 2.7.4.3.1 Mutagenesis of the Catalytic Nucleophile Glu277 The gene encoding for the T. saccharolyticum P-xylosidase cloned into pXHP3 (provided by Prof. Zeikus, University of Michigan) was used as a template for the mutagenic polymerase chain reaction. A four primer mutagenic strategy was devised to introduce the desired Glu277Ala mutation and a silent mutation that would introduce a unique Spel restriction site. Janine Foisy, a technician working in the laboratory of Dr. Warren, first carried out this approach. Unfortunately the recombinant pXHP3 bearing the mutant XynB gene had a single base mutation and consequently the author repeated the mutagenesis approach. Two separate polymerase chain reactions were carried out to first generate two larger fragments of 500 base pairs each. For the first reaction the mutagenic primer (forward primer shown below) contained a unique restriction enzyme cleavage site, the silent mutation (in bold below), and the mutation (in bold and italics). The reverse flanking primer (shown below) contained a unique restriction enzyme cleavage site (shown below). The forward primer (BXYEAF) was as follows: 5'-CAT ATA ACT GCG TAC AAT ACT A G T TAT AGT CCT CAA AAT CCT GTA CAC GA-3' Spel The flanking primer (P2) was as follows: 5'-GCC ATG TTC TTC ATC AAT CAA TTG TCT-3' Muni Chapter 2 Mechanism ofT. saccharolyticum O-Xylosidase 111 For the second reaction the mutagenic primer (BXYEAR, reverse primer shown below) contained a unique restriction enzyme cleavage site, the silent mutation (in bold below), and the mutation (in bold and italics). The forward flanking primer (PI, shown below) contained a unique restriction enzyme cleavage site (shown below). The reverse primer (BXYEAR) was as follows: 5'-AAG ACT ATA ACT AGT ATT GTA CGC AGT TAT ATG AAA CGG AAG GTT CGG AAA-3' Spel The forward flanking primer (PI) was as follows: 5'-GAT TTT GTA TCG CGA CAC GCT ACC AC-3' Nrul Two separate mixtures containing plasmid pXHP3 as a template and either set of the two primers just described was heated to 95 °C, after which the PCR reaction was started by adding 5 units of Pwo DNA polymerase (Boehringer Mannheim). Thirty PCR cycles (45 s at 94 °C, 45 s at 56 °C, and 70 s at 72 °C) were performed in a thermal cycler (Perkin Elmer, GeneAmp PCR System2400). Following agarose gel electrophoresis of the PCR reactions the two products were observed, one in each reaction. Both fragments were approximately 500 bp dsDNA when compared to a 100 bp dsDNA ladder. These two fragments were isolated using the Qiaquick PCR purification kit according to the manufacturer's protocol (Qiagen). The product from the reaction containing primers PI and BXYEAR was termed BX1 and the product of the second reaction was termed BX2. The BX1 and BX2 gel-purified PCR fragments were blunt-end ligated into EcoRl precut pZero-2.0 using T4 DNA ligase (1 unit/10 ng DNA) at a ratio of 10:1 (insert to vector) at 25 °C. The pZero-2.0 plasmids bearing each one of the fragments were transformed into electrocompetent TopplO cells and plasmids containing the BX1 or BX2 PCR fragments were selected by overnight culture on LBzeocin+Kan. plates Colonies were isolated and grown overnight in L B Z e 0cin+Kan and the DNA was isolated via the minipreparation Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 112 technique. This DNA was digested sequentially with Spel and either Nrul (in the case of BX1) or Muni (in the case of BX2) endonucleases and purified in the same manner as described. A third PCR reaction using pXHP3 as template containing the endonuclease-cut BX1 and BX2 fragments, along with flanking primers PI and P2, was carried out using the same conditions as described above. The product of this reaction was purified and treated in the same manner as the BX1 and BX2 fragments, to yield gel-purified pZero-2 containing the PCR fragment approximately 1000 bp in length. This gel-purified pZero-2.0 was digested sequentially with Nrul and Muni endonucleases. The digested fragment was purified as described above. Ligation of Nrul and Muni endonuclease digested pXHP3 and the digested PCR fragment was accomplished using T4 DNA ligase (1 unit/10 ng.DNA) at a ratio of 10:1 (insert to vector) at 25 °C. The cloned product, called pXHP3(E277A), was subsequently transformed into electrocompetent TopplO cells, selected by the ampicillin resistance conferred by pXHP3. Single colonies were selected and grown overnight in LBkan, and DNA was isolated via the minipreparation technique. Restriction endonuclease digest using Mscl and BstBl yielded a fragment that was gel-purified as described above. Ligation of this Mscl and BstBl endonuclease cut fragment with Mscl and BstBl endonuclease cut pET29xynfi/7,5 using T4 DNA ligase (1 unit/10 ng DNA) at a ratio of 10:1 (insert to vector) at 25 °C. TopplO transformed cells were used for preparation of large amounts of plasmid pET29xy«fi7Y6(E277A) (Qiagen Plasmid Maxi kit) and long term storage of the vector as glycerol stock. 2.7.4.3.2 Mutagenesis of the Catalytic Acid/Base Glul60 The gene coding for the His6-tagged T. saccharolyticum (3-xylosidase cloned into pET 29b(+) (pET29xynBH<s) was used as a template for mutagenic polymerase chain reaction. A two primer mutagenic strategy was devised to introduce the desired Glul60Ala mutation and a silent mutation that would introduce a unique fig/II restriction site. The mutagenic primer (forward primer shown below) contained a unique restriction enzyme cleavage site, the silent mutation (in bold below), and the Glu to Ala mutation (in Chapter 2 Mechanism ofT. saccharolyticum B-Xylosidase 113 bold and italics). The flanking primer (shown below) contained a unique restriction enzyme cleavage site (shown below). The forward primer was as follows: 5'-GTC TTG AAG TGG CAA TTT GAG ATC TGG AAT GCA CCA AAC TTA AAA GAG-3' Mscl Bglll The reverse primer was as follows: 5'-ATC TCT TTC TTC GAA AAC GTC GCT GCT-3' BstBl A mixture containing plasmid pET29xynBH6 as a template and the two primers just described was heated to 95 °C, after which the PCR reaction was started by adding 5 units of Pwo DNA polymerase (Boehringer Mannheim). Thirty PCR cycles (45 s at 94 °C, 45 s at 56 °C, and 70 s at 72 °C) were performed in a thermal cycler (Perkin Elmer, GeneAmp PCR System2400). Following agarose gel electrophoresis of the PCR reaction the sole product, an approximately 500 bp dsDNA fragment, was isolated using the Qiaquick PCR purification kit according to the manufacturer's protocol (Qiagen). The restriction endonucleases Mscl and BstBl were used in separate digests to generate a double sticky end fragment that was purified using the Qiaquick PCR purification kit. In order to prepare non-dcm methylated plasmid DNA pET29xynfiiYl5 was transformed into electrocompetent JM110 cells and selected by the kanamycin resistance conferred by the plasmid. Single colonies were selected and grown overnight in L B k a n , and DNA was isolated via the minipreparation technique. This non-dcm methylated DNA was digested sequentially with Mscl and BstBl endonucleases and purified in the same manner as described. Ligation of the digested, gel-purified pET29bxynBH6 and the digested PCR fragment was accomplished using T4 DNA ligase (1 unit/10 ng DNA) at a ratio of 10:1 (insert to vector) at 25 °C. The cloned product, called pET29xynBH6(E160A), was subsequently transformed into electrocompetent TopplO cells, selected by the kanamycin resistance conferred by pET-29b(+). Single colonies were selected and grown overnight in LBkan, and DNA was isolated via the minipreparation technique. Restriction Chapter 2 Mechanism ofT. saccharolyticum 0-Xvlosidase 114 endonuclease digest using Bglll revealed positive clones. TopplO transformed cells were used for preparation of large amounts of plasmid pET29xynBH6 (Qiagen Plasmid Maxi kit) and long term storage of the vector as glycerol stock. 2.7.4.4 Overexpression and purification of His6-tagged XynB, XynB(E277A), and XynB(E160A) The constructed pET29xynBH6 expression vector was used to transform electrocompetent BL21(DE3) cells. The E. coli transformants were selected on LBkan (50 p.g/mL) agar plates. A single colony was picked and grown overnight in 3 mL of LBkan, and this culture was subsequently used to inoculate 1 liter of TYPk a n - After the culture grew to an A e o o of 2-3 at 30 °C, 0.4 mM isopropyl-|3-D-thiogalactoside (IPTG) was added to induce protein expression from the lac promoter and grown for an additional 4 hours at 25 °C. Overexpression of the enzyme was monitored by sampling of both induced and non-induced cells using SDS-PAGE. Induced cells were then harvested and suspended in 30 mL of binding buffer (5 mM imidazole, 500 mM NaCl, 20 mM Tris-HCl, pH 7.9). The cell suspension was passed two times through a French press at 0-5°C and centrifuged at 10 000 x g for 30 min at 4 °C to yield soluble cell extract. A 20-mL slurry of His-Bind resin (nitriloacetic acid-agarose, Novagen) was placed into a 50-mL column, yielding a bed volume of 10 mL. The column was washed with 10 bed volumes of sterile deionized water and then charged with nickel by adding 5 bed volumes of 50 mM N i S 0 4 . Unbound Ni 2 + was washed away with 5 bed volumes of binding buffer (see above). The soluble cell extract from a 1 liter cell culture was applied to the column. The column was washed with 5 bed volumes of binding buffer. XynBH6 protein was eluted from the Ni2+-column with a linear imidazole gradient from 30 to 200 mM in a buffer containing 500 mM NaCl and 20 mM Tris-HCl (pH 7.9). Fractions (3 mL) were collected (1.0 mL/min flow rate) and assayed for xylosidase activity using pNPX. Active fractions were further analyzed using SDS-polyacrylamide gel electrophoresis. Fractions containing pure enzyme were pooled and stored at 4°C. An identical protocol was used for the preparation of the mutants XynBH6(E277A) and Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 115 XynBH6(E160A) enzyme except that in the initial step the plasmids pET29xynBH6(E277A) and pET29xynBH6(E160A) were used to transform the electrocompetent BL21(DE3) cells. Great care was taken during the purification of XynBH6, XynBH6(E277A), and XynBH6(E160A) to avoid wild-type contamination of the mutant enzyme. The purification of the XynBHe(E160A) was conducted with equipment that had never come in contact with any XynBH6. The enzymes were concentrated to 5 mg/mL using a Centriprep concentrator (30 kDa) and dialyzed prior to use using a Slide-a-Lyzer (10 kDa cutoff) from Pierce against 50 mM phosphate buffer pH 7.00. 2.7.5 Mass Spectrometry Techniques Mass spectra were recorded on a PE-Sciex API 300 triple quadrupole mass spectrometer (Sciex, Thornhill, Ontario, Canada) equipped with an Ionspray ion source. Peptides were separated by reverse phase HPLC on an Ultrafast Microprotein Analyzer (Michrom BioResouces Inc., Pleasanton, CA) directly interfaced with the mass spectrometer. In each of the MS experiments, the proteolytic digest was loaded onto a C18 column (Reliasil, 1 x 150 mm), then eluted with a gradient of 0-60% solvent B over 60 minutes followed by 100% B over 2 minutes at a flow rate of 50 p,L/min (solvent A: 0.05% trifluoroacetic acid, 2% acetonitrile in water; solvent B: 0.045% trifluoroacetic acid, 80% acetonitrile in water). Spectra were either obtained in the single quadrupole scan mode (LC/MS), the tandem MS neutral loss scan mode, or the tandem MS product ion scan mode (MS/MS). In the single quadrupole mode (LC/MS) the quadrupole mass analyser was scanned over a mass to charge ratio (m/z) range of 400-1800 Da with a step size of 0.5 Da and a dwell time of 1.5 ms per step. The ion source voltage (ISV) was set at 5.5 kV and the orifice energy (OR) was 45 V. In the neutral loss scanning mode, MS/MS spectra were obtained by searching for the mass loss of m/z = 67.5, corresponding to the loss of the 2F-X label from a peptide Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 116 ion in the doubly charged state. To maximise the sensitivity of neutral loss detection, normally the resolution is compromised without generating artifact neutral loss peaks. 2.7.5.1 Labeling with 2F-DNPX and Proteolysis of T. saccharolyticum XynB (3-Xylosidase (15 uL, 3.4 mg/mL) was incubated with 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-xyloside (3 uL, 2.13 mM) at 37 °C for one hour. Complete inactivation ( > 99%) was confirmed by enzyme assay as above. This mixture was immediately subjected to peptic digestion as follows. (3-Xylosidase (10 u\L native or 12 uX 2F-Xyl-labeled, 3.4 mg/mL) was mixed with 150 mM phosphate buffer pH 2.00 (20 uL or 18 pL) and pepsin (10 uL, 0.4 mg/mL in 150 mM, pH 2.00 phosphate buffer). The mixture was incubated at room temperature for 20 minutes, frozen and analysed immediately by LC/MS upon thawing. 2.7.5.2 Labeling with NBX and Proteolysis of T. saccharolyticum XynBH6 Labeling of T. saccharolyticum P-xylosidase (1.3 mg/mL) was accomplished by incubating the enzyme with NBX (6.0 mM) for 20 minutes at 30 °C in 50 mM sodium citrate, pH 5.50 in a total volume of 30 uL. This sample was then analysed immediately by injecting the mixture onto a reverse-phase column (PLRP-S, 1 x 50 mm) equilibrated with solvent A [solvent A: 0.05% trifluoroacetic acid (TFA) / 2% acetonitrile in water] on an Ultrafast Microprotein Analyser (Michrom BioResources Inc., Pleasanton, CA, U.S.A.). Elution of the enzyme was accomplished using solvent A at a flow rate of 50 uL/min. Proteolytic digestion of the enzyme was performed by mixing the labeled enzyme (30 pL of 1.3 mg/mL) with 4 pL of 2.1 M sodium phosphate pH 1.7, and 4 pL of 1 mg/mL pepsin in 200 mM sodium phosphate pH 2.0. This sample and a control in which the enzyme was not exposed to the inhibitor were incubated at 25 °C for 30 minutes. After this time the sample was frozen at -78 °C. Analysis of these samples by Chapter 2 Mechanism ofT. saccharolyticum 3-Xylosidase 117 ESMS revealed that the digest time chosen ensured complete digestion of the enzyme and generated peptides of a size suitable for sequencing by MS/MS. 2.7.5.3 ESMS Analysis of the Proteolytic Digests Mass spectra were recorded on a PE-Sciex API 300 triple-quadrupole mass spectrometer (Sciex, Thornhill, Ontario, Canada) equipped with an Ionspray ion source. Peptides were separated by reverse-phase HPLC on an Ultrafast Microprotein Analyser (Michrom BioResources Inc., Pleasanton, CA, U.S.A.) directly interfaced with the mass spectrometer. In each of the MS experiments, the proteolytic digest was loaded onto a C18 column (Reliasil, 1 x 150 mm) equilibrated with solvent A [solvent A: 0.05% trifluoroacetic acid (TFA) / 2% acetonitrile in water]. Elution of the peptides was accomplished using a gradient (0-60%) of solvent B over 60 minutes followed by 100% solvent B over 2 minutes (solvent B: 0.045% TFA / 80% acetonitrile in water). Solvents were pumped at a constant flow rate of 50 uL / min. Spectra were obtained in the single-quadrupole scan mode (LC/MS) or the tandem MS product-ion scan mode (MS/MS). In the single-quadrupole mode (LC/MS), the quadrupole mass analyzer was scanned over a mass to charge ratio (m/z) range of 300-2200 Da with a step size of 0.5 Da and a dwell time of 1.5 ms per step. The ion source voltage (ISV) was set at 5.5 kV and the orifice energy (OR) was 45 V. In the tandem MS daughter-ion scan mode, the spectra were obtained in separate experiments by selectively introducing the labeled (m/z = 1075) or unlabeled (m/z = 885) parent ion from the first quadrupole (QI) into the collision cell (Q2) and observing the product ions in the third quadrupole (Q3). Thus, QI was locked on either m/z 1075 or m/z 885; Q3 scan range was 50-1100; the step size was 0.5; dwell time was 1 ms; ISV was 5 kV; OR was 45 V; Q0 = -10; IQ2 = -48. Chapter 2 Mechanism ofT. saccharolyticum 0-Xvlosidase 2.7.5.4 Aminolysis of the 2F-Xylosyl Labeled Enzyme 118 To a sample of the inactivated 2F-xylosyl enzyme (20 pL, 0.85 mg/mL) was added concentrated ammonium hydroxide (5 p:L). The mixture was incubated for 15 minutes at 50 °C, acidified with 50% TFA, and analysed by ESMS as described above. 2.7.6 Enzyme kinetics Michaelis-Menten parameters for aryl xylosides were determined by continuous measurement, at the appropriate wavelength, of the release of the substituted phenol product. Extinction coefficients and wavelength used are listed in Appendix 1 of this thesis. All spectrophotometric measurements were made using a Pye-Unicam PU8700 equipped with a circulating water bath, as described previously.43-44 Phenol pKa values used were taken from Barlin and Perrin,176 Kortum et al,111 and Robinson et a/. 1 7 8 Unless indicated otherwise reaction mixtures, in 25 mM citrate/25 mM phosphate buffer (pH 5.5 for XynBH6 and pH 6.5 for XynBH6(E160A) containing 0.1% BSA (buffer A) at 37 °C, were preincubated in the cell-holder at the appropriate temperature for 10 min prior to addition of enzyme. Exogenous nucleophiles were included in mixtures as indicated and care was taken to ensure the desired pH of the assay mixture was obtained and that the rate of spontaneous hydrolysis was less than 5%. Only freshly prepared solutions containing nucleophiles were used in kinetic studies. XynB-catalyzed hydrolysis for each substrate was measured at 8 to 10 different substrate concentrations ranging from about 0.14 Km to 7 Km, where practical. Values for Km and kcat were determined from the initial rates of hydrolysis (Vo) vs. substrate concentration, by nonlinear regression analysis using the computer program GraFit 4.0.198 In cases where significant transglycosylation occurs, a nonlinear regression was carried out on the data with substrate concentrations ranging from 0.14 to approximately 2 times Km and the resulting kinetic parameters (kcal and Km) are reported. The values of fccat and Km so obtained were then compared to those determined from linear regression of the reciprocal data as plotted according to Lineweaver-Burke and good agreement within the low Chapter 2 Mechanism ofT. saccharolyticum 6-Xvlosidase 119 concentration range was found in all cases. Michaelis-Menten parameters for pNPX, previously undetermined with this enzyme, were Km = 26 uM and kcat = 242 min"1 at 37 ° C in 50 mM sodium phosphate buffer pH 6.5, containing 0.1% bovine serum albumin. 2.7.6.1 Active-site Titration of XynBH6 using 2F-DNPX The concentration of XynBH6 active sites was determined by titrating samples of the enzyme with 2FDNPX at a concentration (0.57 mM) in large excess of enzyme (0.0025 mM - 0.01 mM). Solutions containing the inactivator in 50 mM sodium citrate buffer, pH 5.5, were equilibrated in a 1 cm path length quartz cell at 37 °C in the cell-holder for 10 minutes prior to addition of the enzyme. After addition of the enzyme, the absorbance of the solution at 400 nm was monitored over time until it was evident that the reaction had reached a slow steady state. Corrections were made for the initial absorbance of the xyloside sample. 2.7.6.2 pH Dependence of kcJKr The kcJKm values for the hydrolysis of chromogenic phenyl xylosides at each pH were determined from progress curves at low substrate concentrations as follows. For the XynBH6(E160A) mutant a solution of 3,4DNPX (800 uL, 2.4 uM, 0.01-0.2 x Km), 0.1% BSA, and the appropriate buffer was warmed to 37 °C. Identical conditions were used for the XynBH6 enzyme except that pNPX was used as a substrate (8 pM, 0.02 x Km, for native xylosidase). The reaction was initiated by the addition of a 10 uX aliquot of enzyme (0.4-5 mg/mL) and the release of phenolate monitored at either 400 nm (3,4DNPX) or 360 nm (pNPX) for 5 to 10 minutes at which time it was apparent that 5-7 half lives had passed. The pH of the reaction mixture was then checked and it was thus established that no significant change in pH had occurred during the course of the assay. The change in absorbance with time was fitted to a first-order rate equation using the program GraFit 3.0 (Leatherbarrow, 1990), yielding values for the pseudo-first-order rate Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 120 constant at each pH value. Since at low substrate concentrations ([S] « Km), the reaction rates are given by the equation: v = kM[E\[SVKm the fcobs values correspond to [E]0&Cat/^m. Thus, kcat/Km values can be extracted by division of these obtained rate constants by the enzyme concentration. The buffers used each contained 0.1% BSA (w/v) and were as follows: pH 4.5-6.5, 50 mM citrate and 150 mM sodium chloride; pH 6.5-8.0, 50 mM phosphate and 150 mM sodium chloride; pH 8.0-9.5 50 mM AMPSO and 150 mM sodium chloride. By analyzing the bell-shaped kcJKm versus pH plots using GraFit 3.0 two apparent pKa values of ionizable groups were assigned. Enzyme stability over the pH range, assay time, and at the temperature of the study was examined by adding enzyme at the same concentration as examined in the pH study to a preincubated cell containing, 0.1% BSA and the appropriate buffer at 37 °C. After 10 minutes an aliquot of the mixture was removed and injected into another preincubated solution containing either 3,4DNPX (600 \xM) or pNPX (600 pJVI) in 25 mM citrate/25 mM phosphate buffer pH 6.5. Data were retained for those pH values at which the enzyme was stable during the assay period. Data were discarded if more than 11 % enzyme death had occurred over a 10 minute period. 2.7.6.3 Measurement of Kinetic Isotope Effects Isotope effects were determined in two different ways due to limitations arising from the Km values of each substrate investigated. For oNPX (Km = 56 p,M) and 2,5DNPX (Km =12 pM) saturation conditions were readily obtained and therefore a D (V) isotope effects were determined. Accurate determinations of the a D (V/K) isotope effects for 2,5DNPX (Km = 12 yM) and oNPX (Km = 56 \xM) were precluded by the low concentration of substrate required (0.1 to 0.2 x Km) for these experiments and the consequent poor signal to noise in the progress curves. Conversely, for pNHAcX the relatively high Km (8.1 mM) prohibited measurement of the a D (V) isotope effect due to Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 121 the limited solubility of the substrate. Thus for this substrate the a D (V/K) isotope effect was determined by continuously monitoring the depletion of a low concentration of substrate (1/18 x Km) in the reaction at 287 nm. Quartz cells (1 cm path length) were filled with a solution containing buffer A (pH 5.5) equilibrated at 37 °C and either the enzyme (0.0037 mg/mL with 2,5DNPX; 0.0084 mg/mL with pNPX; 0.125 mg/mL with oNHAcX) or substrate (4.1 mM oNPX; 0.25 mM 2,5DNPX; 0.46 mM oNHAcX). The reaction was initiated either by the addition of an aliquot (20 pL) of thermally equilibrated substrate (oNPX or 2,5DNPX) or of enzyme (in the case of oNHAcX). Initial rates (for oNPX and 2,5DNPX) or second order rate constants (for oNHAcX) were measured alternately for protio and deuterio samples until at least 8 rates for each had been determined. Average rates or rate constants were then calculated for the protio and deutero substrates and the ratio was taken to yield the isotope effect. NMR analysis of the substrates used showed that the extent of isotopic incorporation was > 95%. 2.7.6.4 Inactivation of XynB by 2F-DNPX The inactivation of P-xylosidase by 2,4-dinitrophenyl-2-deoxy-2-fluoro-p-D-xyloside was monitored by incubation of the enzyme (0.68 mg/mL) under the above conditions in the presence of various concentrations of the inactivator (35.5-532.3 p,M). Residual enzyme activity was determined at several time intervals by addition of an aliquot (5-10 pL) of the inactivation mixture to a solution of pNPX (267 MJVI, 750 pJL) in the above mentioned buffer, and measurement of the rate of p-nitrophenolate release. Pseudo-first order rate constants at each inactivator concentration ( £ 0 b s ) were determined by fitting each inactivation curve to a first order rate equation (Leatherbarrow, 1992). Values for the inactivation rate constant (k\) and the dissociation constant for the inactivator (K\) were determined by fitting to the equation: Chapter 2 Mechanism ofT. saccharolyticum 3-Xylosidase 2.7.6.5 Protection from Inactivation by 2F-DNPX 122 Inactivation mixtures (60 uL) containing 0.68 mg/mL enzyme and 21.3 \iM 2F-DNPX were incubated in the presence and absence of thiobenzyl xyloside (11.33 mM, K\ = 5.3 mM). At several time intervals an aliquot (5 pL) was assayed for residual activity as above. 2.7.6.6 Reactivation of 2F-Xylosyl Enzyme Intermediate Complete formation of the covalent xylosyl enzyme intermediate was ensured by assaying for residual activity in the manner described above. The trapped intermediate (450 uL, 1.2 mg/mL) was freed of excess 2F-DNPX inactivator by concentration of the inactivated enzyme using 30 kDa nominal molecular weight cut-off centrifugal concentrators (Millipore Ultrafree-MC) followed by dilution of the concentrated enzyme stock (10-20 uL) to a final volume of 130 uL. Threefold repetition of the process ensured at least a 99% reduction in the free 2F-DNPX concentration. Aliquots (30 uX) of the inactivated enzyme freed from 2F-DNPX were then incubated at 37 °C in the presence of 50 mM phosphate buffer, pH 6.5, alone or with 45 mM xylobiose, 45 mM benzyl 1-thio-P-xyloside or 45 mM xylose in a final total reaction volume of 110 pL. Reactivation was monitored by removal of aliquots (5 or 10 pL) at several time intervals and assaying for activity as described above. Recovery of approximately 75% of activity relative to a control of native enzyme treated in an identical manner was observed. 2.7.6.7 Inactivation of XynBH6 by NBX The inactivation of P-xylosidase by NBX was monitored by incubation of the enzyme (0.61 mg/mL) in buffer A at 37 °C in the presence of various concentrations of the inactivator (1.49-17.91 mM). Residual enzyme activity was determined at several time intervals by addition of an aliquot (5 uL) of the inactivation mixture to a solution of Chapter 2 Mechanism ofT. saccharolyticum 3-Xvlosidase 123 pNPX (135 p,M, 800 pL) in buffer A. Pseudo-first-order rate constants at each inactivator concentration ( & 0 b s ) were determined by fitting each inactivation curve to a first-order rate equation. The second order rate constant for the inactivation process was determined by fitting to the equation: ^ o b , = [EUk/K-d 2.7.6.8 Protection from Inactivation by NBX The active-site directed nature of the inactivation was proven by demonstrating protection against inactivation by a competitive inhibitor. Inactivation mixtures (40 pL) containing 0.61 mg/mL enzyme and 11.9 mM NBX were incubated in the absence and presence of either benzyl P-D-thio-xylobiopyranoside (10.6 mM, K, = 5.3 mM) or xylose (80.8 mM, K{ = 20 mM). At various time intervals aliquots (5 pL) were removed and assayed for residual activity as described above. 2.7.6.9 Analysis of the Products of Hydrolysis by XynBH6 and XynBH6(E160A) Mixtures containing 5 mM 3,4DNPX, 2 M azide, and 0.5 mg/mL of XynBH6(E160A) or 0.1 mg/mL of XynBH6 in 50 mM phosphate buffer (pH 7.5) were incubated at 37 °C overnight. An aliquot of the reaction mixtures and standard (P-xylosyl azide reference) were applied to a 0.2 mm silica gel aluminum TLC plate and allowed to air-dry. The developing solvent was a mixture of ethyl acetate/methanol/water (7:2:1 v/v/v). After development, the chromatograms were air-dried for 5 min, then dipped in a solution of 10% H2SO4 in methanol, and heated until the charred reaction products were visible. Chapter 2 Mechanism ofT. saccharolyticum 0-Xylosidase 2.7.6.9.1 "H NMR Spectrometry of the Products of Enzymatic Hydrolysis 124 Mixtures containing 5 mM 3,4DNPX, 2 M azide, and 0.5 mg/mL of XynBH6(E160A) or 0.1 mg/mL of XynBH6 in 50 mM phosphate buffer (pH 7.5) were incubated at 37 °C overnight. Enzyme was removed by passing the solution through a 10 kDa cutoff polysulfone membrane (ultrafree-MC; Millipore). Samples of the filtrate were prepared for lH NMR analysis by repeated freeze-drying and dissolving in DMSO-D6. Spectra were recorded with a Bruker 400 MHz spectrometer and compared against those of a standard of (3-xylosyl azide 2.7.6.10 Determination of the Delta Absorption Coefficient (Asn) for Aryl Glycoside Substrates A series of three or four solutions containing varying concentrations of the intact glycoside of interest were made up in quartz cells in the assay buffer for XynBH6 (vide supra). The absorption of these solutions at the appropriate wavelength was measured and the data fitted by linear regression. In all cases the data revealed excellent linear correlations (r > 0.99) providing the extinction coefficient of the intact glycoside, for XynBH6 (10 \iL, 2.4 mg/mL) was added and the reaction allowed to proceed for six hours after which time the absorbance of each reaction mixture was again measured at the appropriate wavelength. The reaction was then allowed to proceed overnight and the absorbance of each solution was again recorded. These absorbance readings were in close agreement with the readings taken the previous day indicating the reactions had been complete after 6 hours. These data was fitted by linear regression using the program Grafit198 and in all cases the data revealed excellent linear correlations (r = 0.97) furnishing the extinction coefficient of the liberated phenol. The absorbace of the enzyme at each wavelength was also recorded. The extinction coefficient of the intact aryl glycoside of interest was subtracted from the extinction coefficient of the corresponding phenol and of the enzyme to provide the delta absorption coefficient (Aen). Chapter 3 Mechanism of ^ -Hexosaminidases 125 3 Mechanism of Retaining ^-Hexosaminidases 3.1 Background on Retaining ^ -Hexosaminidases (3-Hexosaminidases are a broad class of enzymes that catalyse the hydrolysis of the glycosidic linkages of 2-acetamido-2-deoxy-P-D-hexosaminyl units (Figure 3.1).11 These enzymes may be endo-glycosidases, which cleave an oligosaccharide or glycoconjugate at a hexosamine unit in the middle of the polysaccharide chain, or they may be exo-acting enzymes, cleaving hexosamine units from the non-reducing terminus of the polysaccharide chain. The best-known p-hexosaminidase is the grade-acting hen egg white lysozyme (HEWL) that catalyzes the hydrolysis of the peptidoglycan of Gram positive bacteria.13 The subject of this chapter will be the retaining exo-$-hexosaminidases. ROH Figure 3.1 Reaction catalyzed by retaining exo-p-hexosaminidases In the classification system developed by Henrissat the exo-acting P-hexosaminidases are found in family 20 of glycoside hydrolases.199 These enzymes are found in a wide range of organisms ranging from animals200,201 to plants202, and on to microbes.203-204 Many of these enzymes have high levels of activity on both 2-acetamido-2-deoxy-p-D-galactosides and 2-acetamido-2-deoxy-P-D-glucosides. Some of these enzymes, however, are highly specific for 2-acetamido-2-deoxyglucosides and have little activity on the 2-acetamido-2-deoxygalactosides. A number of hexosaminidases from various sources have been characterized with regard to their specificity for the Chapter 3 Mechanism of /^-Hexosaminidases 126 gluco- and ga/acfo-configured P-D-pyranosides and an examination of the data reveals no apparent correlation between organism type and P-hexosaminidase specificity (Table 3.1). Indeed, some organisms contain more than one P-hexosaminidase having different specificities (e.g., Vibrio furnisii). Table 3.1: Ratio of hydrolytic activity of selected p-hexosaminidases toward 2-acetamido-2-deoxy-P-D-gluco- and galactopyranosides. Organism Protein Relative Activity and Reference (GlcNAc/GalNAc) Homo sapiens HexB 2 0 5 8.1 Sus scrofa HexB 2 0 6 1.8 Entamoeba histolytica Hex 1 2 0 7 3.7 Bombyx mori HexC 2 0 8 1.0 Candida albicans H e x l 2 0 9 1.3 Trichoderma harzianum H e x l 2 1 0 1.3 Turbatrix aceti Hexl 2 1 1 1.2 Dictyostelium discoideum HexA 2 1 2 5.6 Serratia marcescens C h b 2 1 3 3.3 Vibrio furnisii E x o l 2 1 4 2.1 Vibrio furnisii Exo2 2 1 5 36 Alteromonas sp. HexA 2 1 6 55 Streptomyces thermoviolaceus N A G A 2 1 7 38.5 Bacillus subtilis Y C F O 2 1 8 co Escherichia coli Y C F O 2 1 9 OO 3.1.1 Human p-Hexosaminidase The best-known member of the exo-acting P-hexosaminidases is the human P-hexosaminidase (phex). Like all known mammalian P-hexosaminidases human Phex is compartmentalized within the lysosome where it acts to degrade gangliosides and other glycoconjugates. Gangliosides are lipid glycoconjugates and are found extensively Chapter 3 Mechanism of B-Hexosaminidases 127 within the cellular membranes of the central nervous system where they are involved in cell adhesion and migration.220 These gangliosides are constantly being broken down and recycled within the lysosome. This cellular compartment contains catabolic enzymes that might be dangerous to the viability of the cell if they were not contained.221 Within the lysosome the sequential action of a series of glycosidases systematically degrades the ganglioside, removing one sugar unit at a time. The deficiency of (3-hexosaminidase has a terrible effect, resulting from the accumulation of ganglioside G M 2 (Figure 3.2) within the cell.2 2 2 Figure 3.2 Structure of ganglioside G M 2 indicating the site of action of human P-hexosaminidase The physiological effects of the accumulation of this ganglioside within the brain are severe and in many cases include retardation and death within early childhood. The clinical disorders known as Tay-Sachs and Sandhoff diseases stem from a defect in one of the two subunits (a or (3) that form the dimeric enzyme. Depending on the combination of these subunits different isozymes of Phex result and these have different substrate specificities. The Phex isozyme A (HEXA) is a heterodimer composed of one a and one P subunit and has optimal activity on negatively charged substrates such as GM2 ganglioside. The B isozyme (HEXB) is a homodimer of P subunits and cleaves only neutral substrates including glycoproteins, proteoglycans, and glycosaminoglycans. A third isozyme, PHexS, is a highly unstable homodimer of a subunits that has no known physiological function and perhaps exists only under in vitro conditions. Chapter 3 Mechanism of ^ -Hexosaminidases 128 3.1.2 Mechanism of Retaining p-Hexosaminidases As mentioned earlier hen egg white lysozyme (HEWL) is a p-hexosaminidase, and in the early days following the elucidation of the structure of HEWL several mechanistic alternatives were put forward to explain its catalytic action. These hypotheses included the Koshland mechanism (Path A, Scheme 3.1), the Phillips mechanism (Path B, Scheme 3.1), and a mechanism involving anchimeric assistance from the 2-acetamido group of the substrate (Path C, Scheme 3.1). The anchimeric assistance mechanism has significant precedence from the work of Piszkiewicz and Bruice, which revealed that the acid catalyzed hydrolysis of aryl 2-acetamido-2-deoxy-P-D-glucosides proceeds by a mechanism involving participation of the 2-acetamido group.25 Indeed, Ballardie et al have observed a fleeting oxazoline intermediate during the methanolysis of 2-acetamido-2-deoxy-P-D-glucopyranosyl fluoride.26 E35 R' = NAG or H R" = peptidyl group or H Scheme 3.1 Three possible mechanisms proposed for hen egg white lysozyme (HEWL). A, Koshland double displacement mechanism; B, Phillips ion-pair mechanism; C, anchimeric assistance mechanism. Chapter 3 Mechanism of B-Hexosaminidases 129 Soon after these initial proposals, however, the anchimeric assistance mechanism was ruled out for lysozyme by elegant experiments in which it was found that substrates lacking the 2-acetamido group were cleaved efficiently by HEWL. 1 7 0 The result was that, to some extent, the Phillips mechanism was vindicated and the anchimeric assistance mechanism was temporarily banished to occupy a place as a historical footnote. In 1972 the stereochemical outcome of the reaction catalyzed by boar epididymis P-hexosaminidase (henceforth in this chapter P-hexosaminidase will refer to exo-acting enzymes only) was found to be retention, making an anchimeric assistance mechanism with these enzymes a possibility. However, the failure of the anchimeric assistance mechanism in HEWL, somewhat unjustifiably, carried over to these enzymes and the sway of the Phillips mechanism was so strong that it was readily generalized. Even in the absence of supporting experimentation all retaining glycosidases including, of course, the p-hexosaminidases were presumed to operate by this mechanism. In defense of this position it must be asserted that studies with competitive inhibitors, considered transition state analogues, revealed that the inhibitory response of these p-hexosaminidases was similar to that of other retaining glycosidases (For a partial list see: 223-231) No mention of the anchimeric assistance mechanism can be found in the literature until 1973 when the N-acetyl-P-glucosaminidase from Aspergillus oryzae was examined. 232-235 A s e r j e s G f studies involving the substitution of the 2-acetamido group with other N-acyl groups having various steric and electronic properties (Figure 3.3) prompted these workers to propose a mechanism involving anchimeric assistance from the 2-acetamido group. 236 NH R = CH 3, H, CF 3 , CH 2C1, CH 2Br,etc R'=/7NP R Figure 3.3 Structure of some N-acyl substituted glucosaminides. Chapter 3 Mechanism of ft-Hexosaminidases 130 While insightful, these studies suffered from poor experimental design making the interpretations and conclusions suspect. Yamamoto had not evaluated the Michaelis-Menten parameters but only examined relative rates of hydrolysis at a fixed concentration of substrate. A more rigorous study published in 1980 by Jones and Kosman27 found that varying the electronic properties of the acetamido group had a pronounced effect on the rate of the enzyme catalyzed reaction. This study did not, however, make any attempt to distinguish whether the effect was truly electronic in nature or the consequence of increasing the steric bulk of the acetamido group. Additionally, the Aspergillus niger (3-hexosaminidase they studied had significant activity on aryl-p-glucosides that lack the 2-acetamido substituent. Regardless of these minor criticisms, these works can be considered the first positive evidence pointing to an anchimeric assistance mechanism. Lastly, in 1992 Bollhagen and Legler, using (±)-6-acetamido-l,2-anhydro-6-deoxy-myo-inositol (3.1, Scheme 3.2) found unusual inhibition kinetics.237 As discussed in the introduction, these cyclohexane epoxides are mechanism-based inactivators that typically result in the time-dependent inactivation of the enzyme being studied. When they subjected the P-hexosaminidases from jack beans and bovine kidney to epoxide 3.1 in the presence of a chromogenic substrate they observed a most unusual result. The rate of substrate hydrolysis remained constant for a short time, then decreased, and then later increased. On the basis of these kinetic results they suggested that inositol 3.1 was a psewc-io-substrate for these enzymes. They proposed that these two enzymes utilized anchimeric assistance from the 2-acetamido group to generate, in situ, an oxazoline intermediate (3.2) that is a tight binding inhibitor of the enzyme. This intermediate was, however, slowly turned over by the enzyme to generate a hypothetical N-acetyhnosarnine product (3.3). Curiously, the Aspergillus niger enzyme studied by Jones and Kosman was not affected at all by the inhibitor and the P-hexosaminidase from Helix pomatia was irreversibly inactivated in a time dependent fashion. Chapter 3 Mechanism of 0-Hexosaminidases 131 ~ ^ * N H P-hexosaminidase N N O P-hexosaminidase NH H 3 C ^ O Y C H 3 3.1 3.2 3.3 Scheme 3.2 Proposed reaction mechanism explaining the pseudo substrate behaviour of (±)-6-acetamido-l,2-anhydro-6-deoxy-m>>o-inositol with certain P-hexosminidases. These few scattered reports form the only support for a mechanism involving anchimeric assistance in p-hexosaminidases and were entirely ignored in the literature prior to 1995. However, it is important to remember that all other contemporary studies on the mechanism of P-hexosaminidases prior to 1995 were consistent with the Phillips mechanism.13 The power of the HEWL paradigm was such that the 1969 review by Capon is the last significant mention of the possibility of an anchimeric assistance mechanism made in any major review.29 Lai and Withers uncovered another unexpected result in the P-hexosaminidase-catalyzed hydration of the 2-acetamido glucal (3.4, Scheme 3.3).238 Glycals are slow substrates for retaining P-glycosidases and are hydrated by the enzyme to produce a hemiacetal. Typically, protonation of the 2 position is found to occur from the bottom or a face of the pyranose ring (Path A, Scheme 3.3). With three different p-hexosaminidases (jack bean, bovine kidney, and human placenta), however, the protonation was found to occur exclusively from the top or P face of the saccharide (Path B, Scheme 3.3). Scheme 3.3 Two possible routes for the hydration of 2-acetamido glucal. Path A, Hydration from the bottom (a) face. Path B, Hydration from the top (P) face. Chapter 3 Mechanism of B-Hexosaminidases 132 Although these authors point out that an anchimeric assistance mechanism would enforce such a stereochemical outcome for the hydration reaction a mechanism involving an enzymic nucleophile was still favoured. With the increasing involvement of structural biologists in elucidating the mechanisms of glycosidases it is unsurprising that the idea of anchimeric assistance would be rehypothesized from this quarter. In 1995, Thunnissen et al solved the first three-dimensional structure of a cloned family 18 endo-transglycosylase in complex with the inhibitor bulgecin ( 3 . 5 ) . 2 3 9 Although the logic used is unclear, these authors, based on their 2.8A data and the modeling of a MurNAc residue in place of the proline moiety, hypothesized that the carbonyl oxygen of the acetamido group could participate in catalysis. Presumably, the modeled MurNAc residue can be overlaid such that the acetamido oxygen of the MurNAc moiety overlaps with the carbonyl of the proline group. By then making the further assumption that the ring nitrogen of the proline occupies the same position as the anomeric center of the MurNAc the authors were able to model a surprising conformation for the MurNAc residue. In this imaginative model the acetamido group was suitably oriented toward the anomeric center and presumably would be able to participate in catalysis. Figure 3.4 Structural resemblance of bulgecin (3.5), allosamidin (3.6) and a putative oxazoline intermediate (3.7). Another, related, X-ray crystallographic study involved the elucidation of the structure of the plant defense enzyme hevamine. This retaining enzyme, also a member Chapter 3 Mechanism of 0-Hexosaminidases 133 of family 18, cleaves bacterial cell walls and chitin, which is a component of insect exoskeletons and fungal cell walls. The X-ray structure of hevamine in complex with the competitive inhibitor allosamidin (3.6) led Terwisscha van Scheltinga et al to propose a catalytic mechanism involving anchimeric assistance from the 2-acetamido group of the substrate.240 These workers pointed out that the five-membered carbocylic 'reducing end' of allosamidin resembled the putative oxazoline intermediate (3.7) that might be involved in an anchimeric assistance mechanism. 3.2 Objectives of this Work Despite the interest in these enzymes, none of the studies discussed above have been conducted on p-hexosaminidases of known primary sequence. The catalytic mechanism of this class of enzyme remains unclear with most presupposing participation of an enzymic nucleophile while scattered reports hint that an anchimeric assistance mechanism may be operating for at least some of these enzymes. The goals of this chapter of this thesis include the elucidation of the detailed catalytic mechanism of representative members of retaining exo-P-hexosaminidases. Specifically, we wish to conclusively determine whether any of these enzymes use anchimeric assistance and if so what differences there are in the nature of the transition state. We have selected three such P-hexosaminidases as representative enzymes from this class of enzyme. One enzyme, from jack bean, is known to use a retaining mechanism and is thought to belong to family 20 of glycoside hydrolases,238 another from Streptomyces plicatus, is known to belong to family 20 and the last, from Vibrio furnisii,215 has not been classified into a specific family but displays some sequence similarity with P-glucosidases. Chapter 3 Mechanism of ^ -Hexosaminidases 134 3.3 NAG-thiazoline; An N-Acetyl-p-Hexosaminidase Inhibitor That Implicates Acetamido Participation 3.3.1 Introduction The primary objective of this chapter is to establish the mechanism of (3-hexosaminidases. By analogy to a widely cited mechanism for retaining (3-glycosidases, retaining P-hexosaminidases have commonly been thought to operate by stabilizing a transition state leading to a covalent glycosyl enzyme complex (Scheme 3.4, Path A). As discussed above there are a few tantalising suggestions that the catalytic mechanism of certain P-hexosaminidases may involve participation of the neighbouring C-2 acetamido group, leading initially to a cyclized oxazoline intermediate (Scheme 3.4, Path B). ENZ h° ENZ Scheme 3.4 Two possible catalytic mechanisms for retaining N-acetyl-(3-hexosaminidases. Path A, Double displacement mechanism involving a covalent glycosyl enzyme intermediate. Path B, Anchimeric assistance mechanism involving an oxazoline ion intermediate (3.8). There are several possible methods to discern whether an anchimeric assistance mechanism may operate for any given P-hexosaminidase. The most direct would be to prepare the putative oxazoline intermediate (Scheme 3.4, 3.8) and then examine whether the enzyme in question can turn over this species. Khorlin and coworkers241 report to Chapter 3 Mechanism of B-Hexosaminidases 135 have prepared oxazoline 3.8 although efforts to duplicate these results by the capable group of Capon have failed.26 Indeed, these workers have transiently observed the NAG-oxazoline (3.8) and insist that it is too unstable to be isolated. Some early solvolysis studies have been carried out on the more stable and simple analogue, 2-methyl oxazoline (3.9). Martin and Parcell show that this compound undergoes acid catalyzed hydrolysis to generate O-acetylethanolamine (kobs = 0.024 min"1) as the first formed product 2 4 2 The analogous 2-methyl thiazoline (3.10), however, undergoes ring opening 25 times more slowly (kobs = 0.001 min"1)243 and, according to this team, in both of these cases the attack of water occurs only at the acyl center. 3.9 3.10 Figure 3.5 Chemical structures of 2-methyl oxazoline (3.9) and 2-methyl thiazoline (3.10) The hydrolysis of NAG-oxazoline is likely to be more complicated owing to the additional possibility for the attack of water at the anomeric centre (Scheme 3.5). In all known cases, however, the acid catalysed hydrolysis of the per-O-acetyl derivative of oxazoline 3.8 leads first to the per-O-acetyl-2-amino sugar (3.11) and this kinetic product can only arise from attack of water at the acyl center.244 The base catalysed attack of 1 80 enriched water on the same compound reveals complete incorporation of the isotope into the resulting 2-acetamido sugar (Scheme 3.5, 3.12).245 Chapter 3 Mechanism of ^ -Hexosaminidases 136 HN OH 3 . 1 2 >=o CH 3 Scheme 3.5 Possible routes leading to the hydrolysis of oxazoline 3.8. Path A, attack of water at the anomeric center. Path B, attack of water at the acyl center yielding an orthoimidate intermediate. Path C, acid catalyzed decomposition of the orthoimidate to yield a 2-amino sugar. Path D, pathway favoured for the decomposition of the orthoimidate in the presence of base. This site specific isotope incorporation indicates that base catalyzed hydrolysis also occurs exclusively at the acyl center. Interestingly, in no case has hydrolysis by attack of water at the anomeric center (Path A, Scheme 3.5) been observed although this is the pathway that retaining -^hexosaminidases must catalyze if anchimeric assistance is involved. 3.3.2 NAG-Thiazbl ine is a Potent Inhibitor of Jack Bean (3-Hexosaminidase The corollary of the studies of Martin and Parcell is that l,2-dideoxy-2'-methyl-oc-D-glucopyranoso-[2,l-ti]-A2'-thiazoline (systematic name (3ai?,5/?,65,7/?,7a/?)-5-(acetoxymethyl-6,7-diacetoxy-2-methyl-5,6,7,7a-tetrahydro-3a//-pyrano[3,2-ci]thiazole), NAG-thiazoline, 3 . 1 4 ) may be expected to be considerably more stable to hydrolysis than the NAG-oxazoline. As NAG-oxazoline has been reportedly too unstable to isolate we prepared NAG-thiazoline (Scheme 3.06) as a stable analogue with the expectation that it Chapter 3 Mechanism of B-Hexosaminidases 137 would be sufficiently stable under the acidic (pH 4.25 - 5.00) assay conditions used for the analysis of several p-hexosaminidases 2 4 6 Scheme 3.6 Synthesis of NAG-thiazoline (3.14) from 2-acetamido-l,3,4,6-tetra-0-acetyl-2-deoxy-p-D-glucopyranose. i) a) Lawesson's reagent, tol, A, 94 %; ii) NH 3 , MeOH, 0 °C—•RT, 86 %. Prof. Knapp and coworkers first prepared NAG-thiazoline in their laboratory and found it to be both stable and readily isolated. It is stable to base, as evidenced by the near quantitative conversion of 3.13 to 3.14 in the presence of sodium methoxide and methanol and is stable overnight in water. Thiazoline 3.14 is also unaffected by protonation with trifluoromethanesulfonic acid in chloroform solution. Curiously, the per-O-acetylated NAG-thiazoline had been prepared earlier but had never been deprotected.247 NAG-thiazoline (3.14) was found to be a potent competitive inhibitor of jack bean p-hexosaminidase using p-nitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside (Km = 0.62 mM) as substrate. A reciprocal replot of apparent Km values versus inhibitor concentration, as shown in Figure 3.6, reveals a K for 3.14 of 280 nM. The thiazoline therefore binds almost 20 000 times more tightly than the parent sugar N-acetyl-D-glucosamine (Ki = 5 mM), comparable to the best inhibitors of this enzyme, such as 2-acetamido-l,2-dideoxynojirimycin (Ki = 140-230 nM), and 2-acetamido-2-deoxynojirimycin (1.2 nM). 2 2 3 Chapter 3 Mechanism of B-Hexosaminidases 138 l/[pNP-GlcNAc] mM 1 [Thiazoline] uM Figure 3.6 Inhibition of jack bean NAGase catalyzed hydrolysis of PNPGlcNAc by NAG-thiazoline 3.14. (a) Double reciprocal plot showing competitive inhibition kinetics. The concentrations of 3.14 used were 0.000 (•), 0.087 (•), 0.203 (•), 0.434 (•), 0.868 (A), and 2.026 (A ) uM. (b) Graphical determination of K, by plotting Km apparent against [NAG-thiazoline]. Presumably, the thiazoline is a potent inhibitor because it geometrically resembles the putative oxazoline intermediate or a derived transition state. Therefore, the powerful inhibitory properties of NAG-thiazoline with jack bean P-hexosaminidase suggest that this enzyme uses a catalytic mechanism involving anchimeric assistance. The inhibition data alone, however, cannot be used to rule out a mechanism involving a covalent glycosyl enzyme intermediate. Indeed, it is certainly possible that thiazoline 3.14 is merely a serendipitous inhibitor and its similarity to the putative oxazoline intermediate is simply coincidental. 3.3.3 MuTAG as a Pseudo-Substrate for Jack Bean p-Hexosaminidase We speculated that if jack bean P-hexosaminidase uses the acetamido group for anchimeric assistance in the cleavage of the p-glycosidic linkage, then a glycoside of 2-deoxy-2-thioacetamido-P-D-glucose might be converted by the enzyme to the stable, inhibitory, thiazoline 3.14, resulting in a time-dependent loss of enzyme activity. The 4-methylumbelliferyl 2-deoxy-2-thioacetamido-P-D-glucopyranoside 3.17 (MuTAG) was prepared from 4-methylumbelliferyl 2-acetamido-2-deoxy-3,4,6-tri-0-acetyl-p-D-Chapter 3 Mechanism of /^-Hexosaminidases ; 139 glucopyranoside248 3.15 by treatment with Lawesson's reagent and then sodium methoxide (Scheme 3.13). This compound was first prepared in the laboratory of Prof. Knapp. Scheme 3.7 Synthesis of 4-methylumbelliferyl 2-deoxy-2-thioacetamido-p-D-glucopyranoside 3.17. i) Lawesson's reagent, THF, 74%; ii) NaOMe, MeOH, 80%. The thioacetamido sugar 3.17 was isolated after crystallization as a white solid, and then was further purified by HPLC to remove traces of thiazoline 3.14. Incubation of jack bean p-hexosaminidase with 3.17, while monitoring the release of 4-methylumbelliferone fluorometrically,249 resulted in slow, time-dependent loss of activity, as shown in Figure 3.7. Reaction over a 300 min period resulted in a 21-fold reduction in rate with release of 2.2 mol of 4-methylumbelliferone, consistent with near stoichiometric conversion of 3.14 to 3.17. No loss of enzyme activity occurred in the absence of 3.17, and the putative precursor 3.17 is itself a poor inhibitor of jack bean p-hexosaminidase when measured over time periods too short for formation of significant quantities of 3.17. The resulting time-dependent decrease in activity was shown to be due to the build up of a reversible inhibitor rather than covalent inactivation by repeating the experiment at a higher enzyme concentration and then diluting the sample prior to assay. Under these conditions essentially no time-dependent loss of enzyme activity was observed. These observations suggest that MuTAG is a pseudo-substrate for jack bean P-hexosaminidase and furthermore that the enzyme catalyzes the nucleophilic attack of the thioacetamido group to form the potent inhibitor NAG-thiazoline. Chapter 3 Mechanism of ^ -Hexosaminidases 140 0 100 200 300 400 Time (min) Figure 3.7 Time course of release of 4-methylumbelliferone from 0.65 mM MuTag ( 3 . 1 7 ) upon reaction with jack bean NAGase ( O ) ; Spontaneous decomposition of 3 . 1 7 ( • ) . Two earlier reports describe the inhibition and hydrolysis of the closely related para-nitrophenyl 2-deoxy-2-thioacetamido-(3-D-glucopyranoside (pNPTAG, 3.18, Figure 3.8) by P-hexosaminidases from Turbatrix aceti and Aspergillus niger. As discussed in the introduction, Jones and Kosman suggested that the A. niger P-hexosaminidase catalyzes hydrolysis of glycosides by an anchimeric assistance mechanism.27 One caveat to consider before drawing parallels is that the homogenous enzyme Jones and Kosman studied also had a low level of P-glucosidase activity, making it difficult to discern whether these two different activities result from two different active sites on the enzyme or whether this is the consequence of a mechanism involving an enzymic nucleophile. Regardless, these workers found that the A. niger enzyme hydrolyzed pNPTAG with a second order rate constant, kcJKm, of 130 mM"1 min"1. The parent compound, para-nitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranoside (pNP-GlcNAc), was cleaved some 250 fold faster (kcJKm = 33000 mM"1 min"1) than pNPTAG. One serious complication that renders meaningful comparisons difficult is that Kosman and Jones performed only a stopped assay. As a result it is unlikely that they would have observed any time dependent inhibition as was seen in the studies described in this chapter. In an alternative approach Bedi et al. examined pNPTAG for its inhibitory effect on the P-hexosaminidase from T. aceti250 and did not determine whether the enzyme actually hydrolyzed the compound.211 They estimated a Kx of 5.0 u\M which, remarkably, is lower than that determined for the putative transition state analogue, N-acetylglucosamino-1,5-Chapter 3 Mechanism of ^ -Hexosaminidases 141 lactone (3.19, Figure 3.8) K\ of 5.8 |iM. This is in disagreement with our results and the results of Kosman and Jones who determined a Km for pNPTAG of 1.3 mM. Indeed, such potent inhibition from a ground state analogue is surprising and to the author it seems likely that a small amount of contaminating NAG-thiazoline was present in the assay mixture. We found that even after crystallization of MuTAG (3.17) traces of NAG-thiazoline (3.14) could be detected by HPLC. Consequently, the two studies by Bedi et al. and Kosman and Jones provide little evidence to either corroborate or contradict the results obtained here. Figure 3.8 Chemical structures of para-nitrophenyl 2-deoxy-2-thioacetamido-P-D-glucopyranoside (3.18, pNPTAG) and N-acetylglucosamino-l,5-lactone (3.19) The results of the studies of MuTAG and NAG-thiazoline with jack bean (3-hexosaminidase forward strong evidence for a mechanism involving acetamido participation and an oxazoline intermediate (Scheme 3.4, Path B). Furthermore, NAG-thiazoline represents a novel class of potent P-hexosaminidase inhibitor that may prove to be of use in further mechanistic and clinical studies. Indeed, several such studies are already underway. 2 5 1 C H 3 3.18 3.19 Chapter 3 Mechanism of B-Hexosaminidases 142 3.4Mechanism of Action and Identification of Asp242 as the Catalytic Nucleophile of Vibrio furnisii ^-Hexosaminidase Using 2-Acetamido-2-deoxy-5-fluoro-a-L-idopyranosyl Fluoride In the same year that the results of the preceding section of this chapter appeared in print, two other papers were published that also provided strong evidence for a catalytic mechanism involving anchimeric assistance. The first paper, by Tews et al., involved the elucidation of the three dimensional structure of the first P-hexosaminidase from family 20 of glycoside hydrolases.76 The structure revealed an intact molecule of chitobiose bound across the active site (subsite -1 and +1) with the acetamido group of the -1 sugar oriented in a position where it could act as a nucleophile. Moreover, this X-ray structure revealed that there was no suitably disposed enzymatic nucleophile, further implicating the acetamido group in catalysis. In the second paper, a group of Japanese workers, who apparently prepared the elusive and controversial NAG-oxazoline, obtained more compelling evidence.2 5 2 Although their report is brief and details are lacking they demonstrate that a P-hexosaminidase from Bacillus circulans can use NAG-oxazoline as a substrate in a transglycosylation reaction to generate chitobiose. These three papers, each from an independent laboratory, resulted in the abandonment of the idea that all retaining P-hexosaminidases use a catalytic mechanism involving an intimate ion pair or covalent glycosyl enzyme intermediate. Indeed, many researchers turned to adopt the position that all p-hexosaminidases utilize a mechanism involving anchimeric assistance. Some even went so far as to suggest that the catalytic mechanism of hen egg white lysozyme may involve anchimeric assistance from the 2-acetamido group. 1 0 ' 2 5 3 As a consequence of this change of opinion we were considerably interested when Kehani and Roseman published a paper on the cloning of a P-hexosaminidase having some limited sequence similarity to some members of family 3 of glycoside hydrolases.215 The mechanism of family 3 P-glucosidases254 and P-glucan exohydrolases255 has been the subject of some investigation. Stereochemical outcome studies with both Hordeum vulgare P-glucan exohydrolase256 and Aspergillus wentii P-glucosidase257 have shown that this family operates via a retaining mechanism. Thus, Chapter 3 Mechanism of B-Hexosaminidases 143 owing to the slight sequence similarity with the family 3 enzymes, we hypothesized that the P-N-acetylglucosaminidase from Vibrio furnisii might also belong to this family of glycosidases. Furthermore, we expect that if it is a member of family 3 of glycoside hydrolases it is likely that it catalyzes hydrolysis of its substrates using a double displacement mechanism in which two active site carboxyl groups assist in the formation and breakdown of a glycosyl enzyme intermediate. We therefore closely examined the biochemical literature in order to establish if other p-hexosaminidases with biochemical properties similar to that of the V. furnisii enzyme had been previously isolated. Upon inspection of the literature we found that there were essentially two classes of p-hexosaminidase. The first class has fairly relaxed specificity for the orientation of the 4-hydroxyl inasmuch as both 2-acetamido-2-deoxy-P-gluco- and galactopyranosides are cleaved with similar catalytic efficiency. The second class has much more stringent specificity for 2-acetamido-2-deoxy-P-D-glucosides over 2-acetamido-2-deoxy-P-D-galactosides. This marked difference in specificity is outlined in table 3.1 (Section 3.1 of this chapter). A further distinction between these two classes of p-hexosaminidase is the pH optimum of each group. The first class of P-hexosaminidase is most commonly active at acidic pH while the second class of enzyme typically has a pH optimum in the neutral range. The enzyme described by Kehani and Roseman is a member of this second class of 'neutral' p-hexosaminidases. Using the basic local alignment search tool ( B L A S T ) 1 5 3 we searched all known protein databases for proteins similar to the V. furnisii P-hexosaminidase (ExoII). This family comprises primarily p-glucosidases, P-xylosidases, cellodextrinases, and exo-P-l,3-l,4-glucanases However, fifteen proteins were identified (Figure 3.9) as being more closely related (identical amino acids >30 %, similar amino acids > 45%, and gaps in amino acid sequence < 5%) and these form a separate branch within family 3 of glycosyl hydrolases. Three of these proteins have had their substrate specificity investigated, revealing that they are N-acetyl-P-glucosaminidases. On the basis of the strong similarity of these fifteen closely related proteins we propose that they are all N-acetyl-P-glucosaminidases. Chapter 3 Mechanism of B-Hexosaminidases 144 NAG_Vibfu NAG_Ecoli NAG_Haein NAG_Myctu NAG_Mycle NAG_Ricpr NAG_Zymmo NAG_Bacsu NAG_Altso NAG_Strepth NAG_Strepco NAG_Synpc NAG_Thema NAG_Borbu NAG_BorbuA consensus EXO_Horvu 139 142 142 157 157 151 147 158 164 182 153 162 139 169 167 169 jgJLTYSSAYM: ALAIASRFI AVNLATAFI STAYAGAYAi RTAYAGAYA: WVPLFLSAI SAALGRAVL: TSRLGLYTM: JTKLGLAQ ^^RMVAAQV \RHTTAWI SGTLVREFI E^HGARAC BSLLSLAFY: SGLMVEAFI LETUYDER-KETBCDPRP LETJYDDRT GGVTTPP-TGGVTTPP-IELllIDTS KALHWDS-YGLKLVSHG TGLBRVDHD TGFBVITHT HDLBAYTAG LHLHVIPHP LTLHWDAP INl|lINSN JKYLAFLPYS KHfPGhG DsH p IVQSMTELIPGLQGDVPKDFTSGMPFVAGKNKVAACAKHFVGDGGTVDGINENNTIINRE H NAG_Vibfu NAG_Ecoli NAG_Haein NAG_Myctu NAG_Mycle NAG_Ricpr NAG_Zymmo NAG_Bacsu NAG_Altso NAG_Strepth NAG_Strepco NAG_Synpc NAG_Thema NAG_Borbu NAG_BorbuA NAG_consensus EXO_Horvu consensus NAG_Vibfu NAG_Ecoli NAG_Haein NAG_Myctu NAG_Mycle NAG_Ricpr NAG_Zymmo NAG_Bacsu NAG_Altso NAG_Strepth NAG_Strepco NAG_Synpc NAG_Thema NAG_Borbu NAG_BorbuA NAG_consensus EXO_Horvu consensus 229 227 233 233 246 246 242 235 261 267 274 244 253 230 262 254 EAGILD; JRENKLD; QQLHSQNKLD; PTI JSTQAPVG\ RTLflADAPVG\ KELAKYDYIKLJ IQHLRN AW "IDAG-ADK EI0KASPPGK HKAG-IDSI QG-VDSj JEREKKVTJ |KILlQEN-IP\ FGRAAK F| jIYPHYDAQ BIYSDVDPR SIYSQCDSQ EGHLQVPGLTGSE IGHLDVPGLTGDD IYTALDPDN IYQAWDKES QFPAFDDTTYKSKLDGSDIL' QYPALDNSKWNS-QGESMI QFPALDPSGD LVPAHDPEL LVPAWDDSN RYSSIDSL \YPKLTNGEN 1 /PKISKD i dm pf i -GLMNIHMPAYKNAMDKG-mm ahv pas vlr giv -VSTVMISYSSWN -GVKMHANQDLVT EGAAIMGGPAERAQQSffl EGAAIMGSYAERGQASI KGAGVMGNFVERSK: ISDRFGVSEAVLRTi i SDRYGLADAVLRTI QALSGSMADITKGAJi: ALDGDPVTRALRVF -JKAIADHFGQEEAWMAB GISDFFNPVDATIETF: EGVRTKYGDDRVPVL. AVTRRYGIDGATVK JGGITDIASPREVAVR. -BSAVSNNFSVEEIVSL: (L-MNAVNYNNESIYNTIERI |SYD-MGAITRSFSNIENAIKKS| sD 1 m 27 0 GYLKDTLK FKGFVISDWEGIDRITTPAGSDYSYSVKASILAGLDMIMVPNNYQQF Figure 3.9 Partial multiple sequence alignment of the cloned and putative family 3 p-N-acetylglucosaminidases. Consensus of P-N-acetylglucosaminidases is shown below (NAG_consensus) with upper case letters indicating entirely conserved residues and lower case letters indicating similar residues. P-N-Acetylglucosaminidases were aligned using ClustalW.2 5 8 Dark shading indicates highly conserved residues, while light shading indicates conserved similar residues. Shading was performed using Version 3.21 of BoxShade. The sequence of Hordeum vulgare p-glucan exohydrolase was manually aligned in conjunction with use of ClustalW. 2 5 8 The consensus of all sequences is shown (consensus) at bottom: (*) indicates entirely conserved Chapter 3 Mechanism of B-Hexosaminidases 145 residues, and (:) indicates similar residues. Numbers to the left denote residue positions. The abbreviations used, references to the published sequence, and data bank accession numbers are as follows: NAG_Vibfu, N-acetyl-p-D-glucosaminidase from V. furnisii2^5 (GenBank identifier U52818); NAG_Ecoli, putative N-acetyl-p-D-glucosaminidase from Escherichia coli259 (GenBank identifier AE000302); NAG_Haein, putative N-acetyl-p-D-glucosaminidase from Haemophilus influenzae260 (GenBank identifier U32777); NAG_Myctu, putative N-acetyl-P-D-glucosaminidase from Mycobacterium tuberculosis2^ (GenBank identifier AL021929); NAG_Mycle, putative N-acetyl-p-D-glucosaminidase from Mycobacterium leprae2 6 2 (GenBank identifier CAA18563); NAG_Ricpr, putative N-acetyl-p-D-glucosaminidase from Rickettsia prowazekii2^ (GenBank identifier CAA15141); NAG_Zymmo, putative N-acetyl-p-D-glucosaminidase from Zymomonas mobiliz264 (GenBank identifier AF124757.1); NAG_Bacsu, putative N-acetyl-P-D-glucosaminidase from Bacillus subtilis265 (GenBank identifier AB002150); NAG_Altso, N-acetyl-p-D-glucosaminidase from Alteromonas sp. 2 1 6 (GenBank identifier D17399); NAG_Strepth, N-acetyl-P-D-glucosaminidase from Streptomyces thermoviolaceus2^ (GenBank identifier AB008771); NAG_Strepco, putative N-acetyl-p-D-glucosaminidase from Streptomyces coelicolor266 (GenBank identifier AL023702); NAG_Synpc, putative N-acetyl-p-D-glucosaminidase from Synechocystis sp. 2 6 7 (Genbank identifier D90914); NAG_Thema, putative N-acetyl-P-D-glucosaminidase from Thermotoga maritima2^ (Genbank identifier AE001748); NAG_Borbu, putative N-acetyl-p-D-glucosaminidase from Borrelia burgdorferi2^ (Genbank identifier AE001163); NAG_BorbuA, putative N-acetyl-p-D-glucosaminidase from Borrelia burgdorferi269 (Genbank identifier AE001115); EXO_Horvu, eA;o-p-l,3-l,4-glucanase from Hordeum vulgare256 (Genbank identifier AF102868). The nucleophile identified in this work and the corresponding residue in Hordeum vulgare exo-p-l,3-l,4-glucanase are indicated by (•). The only three dimensional structural information available of any family 3 enzyme is a recently published X-ray structure of the family 3 H. vulgare p-glucan exohydrolase by Varghese et a l . 2 7 0 Analysis of the sequence alignments benefits from consideration of the global protein fold of this glucanase despite the poor similarity of the H. vulgare P-glucan exohydrolase to the putative class of P-N-acetylglucosaminidases. One interesting difference that can be noted in the multiple sequence alignments between the P-N-acetylglucosaminidases and other members of this family is a region that lies at the end of P-strand e. The highly conserved residues Lys-His-Phe, found after the P-Chapter 3 Mechanism of 0-Hexosaminidases 146 strand, have been shown to contain two c/s-peptide bonds in the Hordeum vulgare P-glucan exohydrolase structure, and come in close contact with the C-3 and C-4 hydroxyls of the substrate. Immediately following this unusual feature is a stretch of 11 amino acids within a loop region. Close inspection of this region using sequence databases revealed a short conserved sequence that we assign as a putative N-acetyl binding site in all 15 sequences: 3 cloned family 3 p-N-acetylglucosaminidases (Vibrio furnisii, Streptomyces thermoviolaceus, and Alteromonas sp.), 5 open reading frames classified on the basis of global alignments as P-N-acetylglucosaminidases (Escherichia coli, Haemophilus influenza, Mycobacterium tuberculosis, Streptomyces coelicolor, and Borrelia burgdorferi), and 7 other open reading frames (Zymomonas mobiliz, Rickettsia prowazekii, Mycobacterium leprae, Bacillus subtilus, Thermotoga maritima, Borrelia burgdorferi, and Synechocystic sp.). (Figure 3.10). N A G . N A G . N A G . N A G . N A G . NAG. NAG. NAG. NAG. NAG. B G L . B G L . EXO. B G L . B G L . V i b f u _ H a e i n . E c o l i _ A l t s o .Bacsu . S t r e p t h _ S t r e p c o . R i c p r _Zymmo _ B o r b u A _ A g r t u .Clotm _ H o r v u _ E c o l i _ A s p a c 160 163 163 185 179 203 174 172 168 188 143 152 206 208 189 K H F P G H G K H F P G H G K H F P G H G K H F P G H G K H F P G H G K H F P G H G K H F P G H G KHIJPGHG K H g P G H G K H F P G J l G K H F J p l K H F g i K H F K H F K H ¥ &VIA riVLA fWTA DTHV DTDV DTAV DVAV RATV QAKV 3TTT 1ESEIERQTMS NNQEHRRMTVD IGTVDGINEN .VEGGKEYN ILNEQEHFRQVAE Figure 3.10 Partial multiple sequence alignment of 15 selected family 3 members including 10 p-N-acetylglucosaminidases. The region aligned extends from the end of |}-strand e into a loop region containing the putative N-acetyl binding region of the p-N-acetylglucosaminidases. Alignments were performed using ClustalW.2 5 8 Dark shading indicates highly conserved residues, while light shading indicates conserved similar residues. Shading was performed using Version 3.21 of BoxShade. Abbreviations used, references to the published sequences, and data bank accession numbers can be found in the legend of Figure 3.9 or below: BGL_Agrtu, P-glucosidase from Agrobacterium tumefaciens21 '(Genbank identifier M59852); BGL_Clotm, P-glucosidase from Clostridium thermocellum212 (Genbank identifier X15644); BGL_Ecoli, p-glucosidase from E. coli259 (Genbank identifier AE000302); BGL_Aspac, P-glucosidase from Aspergillus aculeatus2^3 (Genbank identifier D64088). Chapter 3 Mechanism of B-Hexosaminidases 147 The function of this region can only be guessed at, although from examination of the Hordeum vulgare P-glucan exohydrolase structure it appears that in P-N-acetylglucosaminidases these residues may come into contact with the 2-acetamido group of the substrate. We would like to propose this stretch of primary sequence [(KH-(F/I)-PGH-(G/L)-X-X-X-X-D(S/T)H] as a unique identifier for family 3 p-N-acetylglucosaminidases which may prove to be of utility given the large amount of information becoming available from genome sequencing work. 3.4.1 Identity of the Catalytic Nucleophile of Family 3 On the basis of the H. vulgare X-ray structure, Varghese et al. tentatively proposed Asp285 as the nucleophile and Glu491 as the acid/base catalyst. These authors, however, indicated some uncertainty in their assignments of these catalytic carboxyl groups. 2 7 0 Legler and coworkers conducted some elegant early work, using the active site directed inactivator conduritol B epoxide2 3 7 and the slow substrate D-glucal with the Aspergillus wentii p-glucosidase257, proposing an aspartic acid within a 63 residue peptide as the nucleophile. Sequence classifications had not been described at the time of this work and indeed the sequence of this enzyme has not yet been determined. The size of the enzyme, however, suggests it to be a member of family 3 and subsequent studies on the closely related Aspergillus niger enzyme showed this be highly probable.2 7 4 As discussed in the introduction (Section 1.1), however, in a number of cases conduritol epoxides have been found to label residues other than the catalytic nucleophile. For example, in the case of E. coli P-galactosidase, the D-galacto derivative labeled the acid/base residue2 7 5 while in bitter almond P-glucosidase a carboxylate in close proximity to the active site was tagged by conduritol B epoxide.2 7 6 Working with the A. wentii P-glucosidase, Legler also conducted a series of kinetic experiments including kinetic isotope effects, Hammett correlations, and inhibition studies.254 Relative to the glucosidases in this family, the detailed mechanism of family 3 p-N-acetylglucosaminidases remains essentially uncharacterized. While the weak similarity of the putative N-acetyl-p-glucosaminidases with other family 3 enzymes is suggestive of Chapter 3 Mechanism of B-Hexosaminidases 148 a common mechanism, one cannot conclude in advance that these enzymes use a classical double displacement mechanism. Indeed, as outlined above (Section 3.11), strong evidence for catalysis involving neighbouring group participation from the 2-acetamido group has been obtained for the functionally related family 20 P-N-acetylglucosaminidases. It is reasonable to conjecture that the family 20 enzymes may have evolved independently of the family 3 enzymes to use an anchimeric assistance mechanism, using the global fold of the family 3 enzymes merely as a scaffold. Thus whether these N-acetyl-P-glucosaminidases operate with an anchimeric assistance mechanism in which the conserved aspartate is no longer of critical importance for catalysis, or whether this residue functions as a critically important nucleophile as in most retaining glycosidases is unclear and merits investigation. Should the enzyme utilise an enzymic nucleophile, we hoped to identify the residue using a mechanism-based inhibitor, thereby establishing the mechanism of action and also removing any ambiguity regarding the identity of this residue. Since the 2-acetamido group of the substrate is reportedly important for catalysis by ExoII, 2 1 5 it was necessary to prepare an inhibitor bearing this functional group. However, compounds in which the 2-acetamido and 2-fluoro groups are both present are known to be unstable.277 Consequently, we would like to prepare the novel substrate analogues 2-acetamido-2-deoxy-5-fluoro-p-D-glucopyranosyl fluoride (5FGlcNAcF) and 2-acetamido-2-deoxy-5-fluoro-a-L-idopyranosyl fluoride (5FIdNAcF) (Figure 3.11). Figure 3.11 Chemical structures of 2-acetamido-2-deoxy-5-fluoro-p-D-glucopyranosyl fluoride (3.18) and 2-acetamido-2-deoxy-5-fluoro-a-L-idopyranosyl fluoride (3.19). Previous work by McCarter and Withers has shown that the unnatural C-5 epimeric 5-fluoro glycosyl fluorides function as mechanism based inhibitor at least as well as the 'normally' configured inhibitors.183 Therefore, while it would be most 3.18 3.19 Chapter 3 Mechanism of' ^ -Hexosaminidases 149 desirable to obtain the gluco configured inhibitor the ido analogue is a legitimate target as it may function adequately for our purposes. We envisioned that 5FGlcNAcF and 5FIdNAcF (3.18 and 3.19) could be synthesized from the known 2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-a-D-glucosyl fluoride (3 .20) 2 6 by slightly different three-step procedures (Scheme 3.19). Radical bromination278 of 3.20 in carbon tetrachloride generated the 2-acetamido-3,4,6-tri-<9-acetyl-5-bromo-2-deoxy-P-D-glucosyl fluoride (3.21) although attempts to isolate this material were unsuccessful. Monitoring of a crude sample of the 5-bromo derivative (3.21) in CDCI3 by 1 9 F NMR suggested that the product underwent significant decomposition overnight. The bromo compound (3.21) was therefore used immediately after minimal purification on a plug of silica. Displacement of the bromine at C-5 using silver fluoride in acetonitrile yielded the product with the inverted stereochemistry at C-5, namely, 2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-5-fluoro-a-L-idosyl fluoride (3.22), in a modest but acceptable yield. Attempts to obtain the 2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-P-D-glucopyranosyl fluoride (3.23) precursor to 5FGlcNAcF by treatment of the 5-bromo compound using silver tetrafluoroborate in ether resulted in a complex mixture of products. 1 9 F NMR analysis of the crude reaction mixture suggested that per-O-acetylated 5FIdNAcF (3.22) was still the predominant epimer under these conditions. We consequently opted to pursue the synthesis and testing of 5FIdNAcF as a suitable mechanism-based probe before making more efforts toward the synthesis of 5FGlcNAcF. To this end, deprotection of the acetylated 5-fluoro-idosyl fluoride compound (3.22) was accomplished by treatment with ammonia in methanol to yield, after purification, one of the target compounds 5FIdNAcF (3.19). Chapter 3 Mechanism of B-Hexosaminidases 150 AcO AcO AcO 2 HN 3.20 C H 3 AcO-, A c O — ^ - \ - - - - Q A c O - X ^ - ^ ^ F Rr ^ H N ^ o C H 3 3.21 AcO-AcO-AcO-HN O / C H 3 3.23 R O ^ H N ^ Q r C H 3 3.22,R = OAc 3.19, R = OH Scheme 3.8 Routes to the synthesis of 5FGlcNAcF and 5FIdNAcF. i) NBS, A, CC14; ii) AgF, MeCN, 8% over two steps; iii) AgBF4OEt2, OEt2, 0%; iv) NH 3, MeOH, 88%. We opted to conduct the reaction series on the desired 2-acetamido compound and accept the loss of yield arising from partial N-halogenation of the alkyl amide rather than employ a transient protecting group on a 2-amino-sugar. This decision was reached on the basis of avoiding potential difficulties in deprotection and reacetylation during subsequent steps, since compounds with two anomeric fluorine atoms are relatively reactive. The source of enzyme for the studies on ExoII was the cloned hexahistidine tagged enzyme. Dr. Christoph Mayer, a postdoctoral fellow in the laboratory of Dr. Withers, constructed the expression plasmid. The original clone of the exoll gene was generously provided by Dr. Saul Roseman, a professor in the department of Biology at John Hopkins University. Because of the addition of the hexahistidine tag to the C-terminus of the enzyme we decided it would be prudent to compare the kinetic behaviour of the polyhistidine-tagged enzyme to the kinetics previously obtained by Kehani and Roseman for the untagged cloned enzyme. Comparison of the Michaelis-Menten parameters for the hydrolysis of pNPGlcNAc by the His6-tagged ExoII (23.5 °C, Km = 0.73 mM, Vmax = 2.37 p,mol mg"1 min"1) used in this study and the wild-type ExoII as reported by Kehani and Roseman (22 °C, Km = 0.44 mM, Vmax =1.1 u.mol mg"1 min"1) indicates that modification of the wild-type ExoII by addition of the hexa-histidine tag Chapter 3 Mechanism of ^ -Hexosaminidases 151 does not alter its catalytic function significantly. The greater specific activity of the His6-tagged ExoII may arise from the expedient purification of the His6-tagged ExoII as compared to the wild-type ExoII that is known to be thermally unstable. Returning to the testing of the 5FIdNAcF mechanism based inhibitor, we found that incubation of V. furnisii p-N-acetylglucosaminidase with 5FIdNAcF did not result in time-dependent inactivation of the enzyme relative to a control. The absence of any observable inactivation of the enzyme by 5FIdNAcF therefore prompted us to investigate whether the enzyme was capable of hydrolysing this analogue. Rates of release of fluoride upon incubation of V. furnisii P-N-acetylglucosaminidase at a number of different inhibitor concentrations [(0.15-5) x Km] as measured by a fluoride ion selective electrode indicated that the enzyme hydrolyzed 5FIdNAcF very slowly as a Michaelian substrate (Km = 0.23 ± 0.02 mM, kcal = 0.0091 ± 0.0004 s"\ Figure 3.12). The hydrolysis of 5FIdNAcF by the enzyme follows the scheme shown below where E = enzyme, F = fluoride, 5FIdNAc = 2-acetamido-2-deoxy-5-fluoro-L-idosyl, E:S = Michaelis complex, and EI = enzyme intermediate. The 5FIdNAc product rapidly decomposes with the release of a second equivalent of fluoride; thus, all kinetic parameters have been calculated on the basis of two equivalents of fluoride released per molecule of inhibitor (see below in methods, section 3.4.2.7.2). 0 0.2 0.4 0.6 0.8 1 1.2 1.4 [5FIdNAcF] m M Figure 3.12 Michaelis-Menten plot of initial rates of hydrolysis of 5FTdNAcF by V. furnisii ExoII. Chapter 3 Mechanism of B-Hexosaminidases 152 The observed Km is somewhat lower than that observed for pNPGlcNAc (Km = 0.44 mM) despite the absence of an aromatic aglycon, which has generally been found to enhance binding. This relatively low Km is even more surprising given the structurally significant epimerization of the 5-position from the configuration found in GlcNAc to that found in IdNAc. This modification would be expected to result in decreased inherent binding affinity of the enzyme for the inhibitor. It therefore seems likely that this relatively low Km value has its origin in a kinetic phenomenon rather than reflecting enhanced binding interactions. The Km value for a two-step enzyme having a ping-pong-type mechanism (see chapter 2, page 63) in which one of the substrates is water is expressed as: Km=[(k.l+k2)/ki][k3/(k2+h)] As can be seen from the expression, a low Km can result from an increase in k\ relative to k.\ or k2, or a decrease in k^ relative to k2, which in either case results in an increased steady-state concentration of the enzyme intermediate (EI). However, in this case the low observed kcat suggests that &3 has decreased significantly relative to k2 such that the breakdown of the enzyme intermediate is now rate-determining. Such kinetic behavior is consistent with that observed for 5-fluoroglycosyl fluorides studied previously, as discussed in the introduction, 183,279,280^ a n ( j 0 p e n s the possibility of investigating the nature of what must be a fairly well-populated enzyme intermediate. The mass of native V. furnisii P-N-acetyl-glucosaminidase was found by ESMS to be 37 258 Da (Figure 3.13a), which is in agreement, within error, with the theoretical mass of the cloned enzyme (37 247 Da). After incubation with 5FIdNAcF, two species are observed: the native, unlabeled enzyme; and another species with a mass of 37 479 Da (Figure 3.13b). The mass difference observed between the native and inhibited enzymes is 221 Da, a value which is consistent, within error, with the addition of a single 5FIdNAc label (222 Da). The partial labeling of the enzyme is consistent with the expectation of a high steady-state population of the enzyme intermediate rather than a stoichiometricalry inactivated species. Chapter 3 ; Mechanism of 8-Hexosaminidases ' 153 36500 37000 37500 mass (Da) Figure 3.13 Transform of the electrospray mass spectrum of (a) native ExoII and (b) ExoII incubated with 8.2 mM 5FTdNAcF. The observation of a covalent glycosyl enzyme intermediate provides strong evidence for a mechanism involving an enzymic nucleophile rather than the mechanism involving substrate assistance from the 2-acetamido group followed by the family 20 p-N-acetylglucosaminidases. This latter mechanism involves the formation of an oxazoline or oxazolinium ion intermediate, noncovalently bound to the enzyme, and it is very unlikely that such a complex would survive the LC/ESMS conditions. Indeed, subsequent peptide analysis (vide infra) confirms the covalent nature of the intermediate. This enzyme therefore operates by a double displacement mechanism involving an enzymic nucleophile as do other family 3 members. The observation of a transient covalent glycosyl enzyme intermediate with this substrate analogue also suggests that the catalytic mechanism of ExoII with its natural substrate proceeds through a covalent intermediate. Of course, an ion pair mechanism as suggested by Phillips for HEWL could also operate for ExoII although there is no evidence to support such a view. Having clarified the mechanism of action of ExoII we can now proceed to identify the enzymic nucleophile. Chapter 3 Mechanism of ^ -Hexosaminidases 154 The relatively large steady-state 5FIdNAc-enzyme population observed by ESMS should permit the determination of the residue to which the label is attached. 3.4.1.1 Identification of the Labeled Active Site Peptide. Peptic hydrolysis of the 5FIdNAc-enzyme resulted in a mixture of peptides which were separated by reverse-phase HPLC using the ESMS as detector. When scanned in the normal LC/MS mode, the total ion chromatogram (TIC) showed a large number of peaks, each corresponding to one or more peptides in the digest mixture (Figure 3.14a). The 5FIdNAc-peptide was then located in a second experiment using the. tandem mass spectrometer set up in the neutral-loss mode. This technique involves the limited fragmentation of the ions by collision with an inert gas (N2) in a collision chamber (Q2) between two quadrupoles (QI and Q3). Since the ester linkage between label and enzyme is one of the most labile linkages present, homolytic cleavage of this bond is expected to occur in the collision chamber. The resulting neutral loss of this label leaves the peptide with an unchanged charge state but a mass decrease of 222 Da. The two quadrupoles were therefore scanned in a linked mode to permit only passage of those ions which lose a mass of 222 Da in the collision cell. A peak at 23.0 min was observed (Figure 3.14b), and no such peak was observed in a control in the neutral-loss spectrum of the unlabeled -N-acetylglucosaminidase (Figure 3.14c). These results indicate that a singly charged peptide bearing the 5FIdNAc label is preferentially detected with a mass of 1248 Da (Figure 3.14d); thus, the uncharged peptide has a mass of 1247 Da (1248 - 1 H). Since the mass of the label is 222 Da, the unlabeled, protonated peptide must have a mass of 1026 Da (1247 - 222 + 1 H). Aminolysis of the labeled peptide resulted in disappearance of the peak in the ion chromatogram of m/z 1247 (±1) (data not shown) and an increase in the intensity of the peak at 1027. The mass loss of 221 Da (1248 - 1027) resulting from aminolysis of the labeled peptide to yield a lower molecular weight species (m/z = 1027) is consistent, within error, with the expected mass loss from cleavage of the ester-linked 5FIdNAc label. Chapter 3 Mechanism of B-Hexosaminidases 155 § > 2 100 75-1 SO 25 a 100 75 50 25 10 1001 75 50 25 10 30 30 4) Time(min) 23.0 20 33 40 Time (min) 20 30 40 Time (min) 50 [ d 1248 lr-«-l r — " r l , - * n a _ ™ „ , 400 600 800 1000 1200 1400 1600 1800 m/z (amu) Figure 3.14 ESMS experiments on peptic digests of V. furnisii ExoII. (a) Enzyme incubated with 5FIdNAcF, TIC in normal MS mode, (b) Enzyme incubated with 5FIdNAcF, TIC in the neutral-loss mode, (c) Unlabeled enzyme, TIC in the neutral-loss mode, (d) Mass spectrum of the peptide eluting at 23.0 min. The susceptibility of the label to aminolysis provides evidence for an ester linkage between the sugar moiety and the enzyme and is entirely consistent with the formation of a glycosyl-enzyme intermediate on an enzymic carboxylate. Candidate peptides of this mass (1026 Da) were identified by inspection of the enzyme amino acid sequence and searching for all possible peptides of this mass. Three possible peptides with a mass of Chapter 3 Mechanism of B-Hexosaminidases 156 1026 ± 1 D a were identi f ied: i 3 5 D V Q T V L T Y S 1 4 3 , 2 3 8 l V F S D D L S M 2 4 6 , and 287PISVVPQAQS296 The peptide bearing the label was then identif ied unambiguously by peptide sequencing using M S / M S . 3.4.1.2 Peptide Sequencing In fo rmat ion on the sequence was obtained by addit ional f ragmentat ion o f the peptide o f interest (m/z 1248) in the daughter ion scan mode (Figure 3.15). B o t h the labeled parent ion and the unlabeled peptide (m/z = 1026) arising f r o m loss o f the 5FIdNAc label appear as singly charged species. Peaks resulting f r o m B ions correspond to I V F (m/z 360) , I V F S (m/z 447) , I V F S D (m/z 562) , I V F S D D (m/z 677) , I V F S D D L (m/z 790) , I V F S D D L S (m/z Sll), and I V F S D D L S M (m/z 1008). Several peaks w i t h m/z values 18 units lower than the B ions are observed. These ions are believed to arise f r o m dehydrat ion o f a serine residue and are marked ( • ) on the spectrum. This sequence corresponds exactly to that o f 238IVFSDDLSM246 as deduced f r o m the gene sequence. W i t h i n this peptide, there are four amino acids that cou ld potential ly funct ion as nucleophiles: t w o serines and t w o aspartic acids. Precedent alone w o u l d make the t w o serine residues unl ikely candidates, and both the susceptibil ity o f the linkage to aminolysis (vide supra) and the lack o f sequence conservation (vide infra) con f i rm this. This leaves t w o aspartic acid residues that could funct ion as the catalytic nucleophile, and, unfortunately, the M S / M S sequencing does not al low the assignment o f which one was labeled. However , sequence conservation and precedent f r o m other enzymes provides a clear answer, as described below. Chapter 3 Mechanism of B-Hexosaminidases 157 -NH.Hle-Val--Phe/Ser/Asp/Asp/xu^eryMet-Ca/)H * 360 447 562 677 790 877 1008 Bions 100 -i 80 H £ 60 '55 c 4> 1) > 40 4-. 3> 20 0 I UuiL * [M+HJ+ [M+H] + 1248 200 400 600 800 1000 1200 w/z (amu) Figure 3.15 ESMS/MS daughter-ion spectrum of the 5FIdNAc-peptide (m/z 1248, in the singly charged state). Observed B series fragments are shown below the peptide sequence. An asterisk indicates the unlabeled peptide arising from neutral loss of the label. B ions arising from dehydration of serine are indicated by (•). 3.4.1.3 Multiple Sequence Alignments Multiple sequence alignments of the members of family 3 glycosyl hydrolases revealed that Asp242 is fully conserved among all known family members whereas Asp243 is not, as shown in the partial multiple sequence alignment of Figure 3.9 and 3.10. This observation, in conjunction with the results from the tandem MS sequencing of the labeled peptide, permits the identification of Asp242 as the catalytic nucleophile. This assignment is in agreement with the results of both Legler, 2 8 1 Fincher, 2 7 0 and very recently Siegel et al214 and thereby both unambiguously establishes the identity of the Chapter 3 Mechanism of B-Hexosaminidases 158 nucleophile within this family and demonstrates that the P-hexosaminidases within this family operate via the same mechanism as that of the glucosidases. This is in contrast to the family 20 P-hexosaminidases that are known to use a catalytic mechanism involving anchimeric assistance from the 2-acetamido group. 3.5 Comparative Analysis of exo-p-hexosaminidases from Families 3 and 20. To this point in chapter 3 we have provided evidence that there are two distinct classes of ew-P-hexosaminidase catalyzing the same overall reaction yet operating by different catalytic mechanisms. On first consideration this realization appears contrary to the hypothesis that all retaining P-glycosidases use a fundamentally similar mechanism. The apparent distinction in the catalytic mechanisms of these two classes of P-hexosaminidase rests in the identity of the nucleophile. For one class, the family 3 P-hexosaminidases, an anionic enzymic carboxylate group acts as the nucleophile while for the other class of p-hexosaminidase, comprising family 20 of glycoside hydrolases, the neutral 2-acetamido group of the substrate acts as nucleophile. The consequences, however, of differing nucleophiles on the structure of the transition-state for each reaction are unknown. Indeed, little mechanistic work has been carried out with P-hexosaminidases from families 3 or 20. Unfotunately, the works of Jones and Kosman27 and of Yamamoto236 have been carried out on isolated enzymes for which the family is not known. The work contained in this section of chapter 3 will focus on elucidating the detailed catalytic mechanisms of the P-hexosaminidase from families 3 and 20. Several key questions will be addressed including the extent of involvement of the 2-acetamido group, the extent of charge development in the transition state and whether the oxazoline intermediate is covalent or a stabilized cation. To this end we will undertake a comparative study of the family 20 S. plicatus, and family 3 V. furnisii P-hexosaminidases. Chapter 3 Mechanism of B-Hexosaminidases 159 3.5.1 NAG-Thiazoline as a Probe of Mechanism As discussed earlier, NAG-thiazoline may be a specific inhibitor of family 20 (3-hexosaminidases owing to its similarity to the putative NAG-oxazoline intermediate or a derived transition state involved in the anchimeric assistance mechanism. The bicylic structure of NAG-thiazoline, however, has little similarity to any putative intermediate involved in the classical double displacement mechanism. Consequently, we expected that it would be a very poor inhibitor of the family 3 P-hexosaminidases. Gratifyingly, when we tested the NAG-thiazoline as an inhibitor of the activity of ExoII on pNPGlcNAc we found that it was a very poor inhibitor with a K, estimated at greater than 30 mM. When SpHex was assayed in the same manner the dissociation constant (K{) for NAG-thiazoline was determined to be 20 pM (Table 3.2). NAG-thiazoline therefore binds some 70 times weaker to SpHex than to jack bean P-hexosaminidase (Ki = 280 nM) and approximately 2000 times more weakly than to human HEXB (Ki ~ 10 nM 2 8 2). The basis for the poorer binding of NAG-thiazoline to the bacterial SpHex as compared to the plant or mammalian enzyme is unclear, however, the compound is still a reasonably potent inhibitor of SpHex. GlcNAc alone binds with similar efficiency to both SpHex and ExoII (Table 3.2) and so NAG-thiazoline discriminates between ExoII and SpHex by at least five orders of magnitude. Given the great disparity in substrate specificity of the family 3 and family 20 P-hexosaminidases toward 2-acetamido-2-deoxy-p-D-gluco and -galactosides (Table 3.1,Section 3.1) we were interested in whether the inhibitory profile of a galacto derivative of NAG-thiazoline would parallel the substrate specificity. Using a synthetic approach analogous to that used for the preparation of NAG-thiazoline (Scheme 3.6, Section 3.1.1) we prepared NAGal-thiazoline. As expected NAGal-thiazoline binds at least as poorly to ExoII as does NAG-thiazoline. The inhibition constants of NAGal and NAG-thiazolines do follow the same trend as the glucoside and galactoside specificity of SpHex. NAG-thiazoline binds to SpHex five-fold more tightly that does NAGal-thiazoline and this selectivity parallels the specificity [(&Cat/#'m)pNPGicNAc/ (fcCat/£m)pNPGaiNAc] of SpHex for pNPGlcNAc (kcJKm = 3700 ± 300 m M ' W 1 , kcat = 185 Chapter 3 Mechanism of B-Hexosaminidases 160 ± 3 sec-1, Km = 0.050 ± 0.004 mM) over pNPGalNAc (kcJKm = 670 ± 50 mJVJ/W1, fccat = 39 ± 1 sec-1, Km = 0.056 ± 0.004 mM). Table 3.2: Inhibitor! of 3-hexosaminidase catalyzed hydrolysis of pNPGlcNAc by NAG-thiazoline, NAGal-Thiazoline, GlcNAc, and GalNAc. Enzyme Ki (NAG-Thiazoline) mM Ki GlcNAc mM Ratio Ki (GlcNAc/ Thiazoline) Ki (NAGal-Thiazoline) mM GalNAc mM Ratio Ki (GalNAc/ Thiazoline) SpHex 0.02 3.0b 150 0.10" 1.0" 10 ExoII >30b 0.4C <0.01 >30a >10a <0.3 a The Ki was estimated to be greater than the noted value as incubation of the enzyme with 30 mM NAGal-thiazoline or 10 mM GalNAc resulted in less than 50% inhibition of enzymatic activity toward pNPGlcNAc. b The K was estimated by range finder Kr c Data from Chitlaru and Roseman.215 From this limited data it appears as though these thiazoline inhibitors are specific for family 20 enzymes, having little affinity for the family 3 P-hexosaminidases. These data further support the view that these two enzymes catalyze hydrolysis of N-acetyl-glucosaminides by different mechanisms and that NAG-thiazoline is a mimic of a putative oxazoline intermediate or a similarly structured transition state. These results also establish the pair of thiazoline inhibitors as a useful tool for rapidly distinguishing whether a P-hexosaminidase belongs to family 3 or 20 of glycoside hydrolases. This may be useful for experimentalists desiring to isolate or clone P-hexosaminidases with specific physicochemical properties. 3.5.2 Requirement for the Substrate 2-Acetamido Group. From the studies with NAG-thiazoline and NAGal-thiazoline it certainly appears that the 2-acetamido group is involved in catalysis by the family 20 P-hexosaminidases. Both ExoII and SpHex, however, have been reported to have an absolute requirement for the substrate to bear a 2-acetamido group.215 For the family 3 enzymes this requirement Chapter 3 Mechanism of 0-Hexosaminidases 161 may simply be for recognition of the substrate rather than participating directly in catalysis. What is not known is the extent of involvement of the acetamido group in the transition state for catalysis by each enzyme. Firstly, we would like to unambiguously establish the requirement for the acetamido group. Secondly, we would like to establish the electronic demands on the acetamido group in the transition state for each enzyme. To examine the requirement for the acetamido group we decided to examine whether 2,4-dinitrophenyl (3-D-glucopyranoside (2,4DNPG, 3.24) is a substrate for ExoII and SpHex. If, as is believed, the family 20 enzymes use the acetamido group in catalysis we expect that even a glucoside bearing an excellent leaving group such as 2,4-dinitrophenyl should not be cleaved by SpHex. Conversely, for ExoII, which we believe uses an enzymic nucleophile, we expect that the enzyme may well have detectable activity on 2,4DNPG if the acetamido group does not fulfill a critical catalytic role. No significant activity above that arising from spontaneous hydrolysis could be detected when SpHex (0.2 mg/mL) was incubated overnight with 20 mM 2,4DNPG. When ExoII was incubated with 2,4DNPG, however, significant activity was observed. This observation is in disagreement with the results of Chitlaru and Roseman who could not detect glucosidase activity215 possibly because they were not using a highly activated substrate. Regardless, the Michaelis-Menten parameters for the hydrolysis of 2,4DNPG could be measured conveniently by continuous assay over five minutes (Figure 3.16, kcJKm = 0.030 ± 0.003 m M ' W 1 , kcat = 0.009 ± 0.001 sec"1, Km = 0.31 ± 0.02 mM). The &cat value is only approximately 350 fold lower than that found for pNPGlcNAc {kcJKm = 4.4 ± 0.2 m M ' W 1 , £ c a t = 3.20 ± 0.04 sec"1, Km = 0.73 ± 0.03 mM). This suggests that the acetamido group does not play a critical role in the catalytic mechanism (such as nucleophile where a 105 to 107 fold rate reduction has been found) but is somehow involved in stabilizing the transition state. Chapter 3 Mechanism of 3-Hexosaminidases 162 0 10 20 30 40 [2,4DNPGlc] mM Figure 3.16 Michaelis-Menten plot of the hydrolysis of 2,4DNPG by ExoII. Inset: Region of low 2,4DNPG concentration showing apparent Michaelian saturation kinetics. The chemical structure of 2,4DNPG (3.24) is shown. Plots with the shape found in figure 3.16 commonly indicate for retaining glycosidases that the second step of the reaction is rate determining and that by increasing the substrate concentration a second molecule of substrate intercepts the glycosyl enzyme intermediate thereby increasing its rate of breakdown. This sequence of events is known as transglycosylation and occurs at elevated substrate concentrations. Thus, the shape of the plot in figure 3.16 suggests that the rate determining step in the ExoII catalyzed hydrolysis of 2,4DNPG is deglycosylation. Recall that very similar behaviour was described for XynB in Chapter 2. Presumably, as was the case for XynB, the very good leaving group makes the first step, glycosylation, faster than the second, deglycosylation, step. The low Km value for 2,4DNPG is consistent with such an interpretation as it is unlikely that a glucoside would have the same affinity as the glucosaminide substrate. The low Km can most likely be ascribed to the consequence of accumulation of the glycosyl enzyme intermediate as described in Chapter 2 and above (Chapter 3, Section 3.1.1). These results using 2,4DNPG further support the critical role of the acetamido Chapter 3 Mechanism of B-Hexosaminidases 163 group in catalysis by family 20 P-hexosaminidases and indicate a non-critical role for the acetamido moiety in catalysis by the family 3 P-hexosaminidases. 3.5.3 Electronic Requirement for the 2-Acetamido Group. To study the electronic requirements of ExoII and SpHex for the acetamido substituent we prepared a series of para-nitrophenyl 2-deoxy-2-fluoroacetamido-P-D-glucoside substrates (See Chapter 3, Section 3.1.2 or 3.7.2.1 for chemical structures of these compounds) using the same approach as Yamamoto233. Increasing levels of substitution of fluorine for hydrogen on the acetamido group decrease the basicity of the carbonyl oxygen nucleophile, making this group a poorer nucleophile. Analysis of pNPGlcNAc, pNPGlcNAcF, pNPGlcNAcF2, and pNPGlcNAcF3 as substrates of both enzymes reveals significant differences in the processing of these substrates by ExoII and SpHex (Table 3.3). Indeed, increasing fluorine substitution has little effect on the Michaelis parameters for ExoII while for SpHex a dramatic effect is seen. The data in Table 3.3 can be represented graphically in a Taft-like free energy relationship. Table 3.3: P-Hexosaminidase-catalyzed hydrolysis of fluoro-substituted para-nitrophenyl 2-acetamido-2-deoxy-P-D-glucosides. Substrate a* Enzyme ^cat (mM) (sec1) (sec^mM"1) pNPGlcNAc 0.0 ExoII 0.73 ± 0.03 3.20 ± 0.04 4.4 ± 0.2 SpHex 0.050 ± 0.004 185 ± 3 3700 ± 3 5 0 pNPGlcNAcF 0.8 ExoII 0.68 ± 0.04 3.4 ± 0 . 2 5.0 ± 0.3 SpHex 0.55 ± 0.02 390 ± 40 720 ± 80 pNPGlcNAcF2 2.0 ExoII 0.33 ±0.01 2.04 ± 0.02 6.2 ±0 .1 SpHex 0.66 ± 0.02 39 ± 1 59 ± 2 pNPGlcNAcF3 2.8 ExoII 0.138 ±0.007 0.9 ±0 .1 7 ± 1 SpHex 1.7 ± 0.2 0.95 ± 0.04 0.6 ±0 .1 a* Is the Taft-parameter, values of which were obtained from reference Chapter 3 Mechanism of 3-Hexosaminidases 164 - 1 0 1 2 3 - 1 0 1 2 3 o* & Figure 3.17 Substituent analysis of the ExoII and SpHex P-hexosaminidase catalyzed hydrolysis of a series of fluoroacetamido derivatives, a) Plot of log (kcJKm) against a*, b) Plot of log (&cat) against a*. In each plot the open circles (O) are the data for ExoII and closed circles (•) for SpHex. The data and o* values used are tabulated in Table 3.3. By plotting the log of the first (&cat) and second order (kcJKm) rate constants against the Taft parameters for methyl, monofluoromethyl, difluoromethyl, and trifluoromethyl we obtain the linear relationships shown in Figure 3.17. The function governing a Taft relationship is given by the following relationship where p* is the apparent sensitivity of the reaction to the substitution, CT* is the Taft-parameter for the methyl group, 8 is the sensitivity of the reaction to steric effects, and E s is the steric parameter for the methyl group. p * = a * log(rate constant) + 8ES Given the series of substrates studied here it is impossible to dissect the steric and electronic effects of the substitutions. An additional complication is that the data are limited to a set of four points and that there appears to be considerable scatter in the plot of log £ c a t against CT* for the SpHex catalyzed reaction. Scatter in plots of £Cat can be attributed to differences in ground state binding as &cat is governed by the largest energy Chapter 3 Mechanism of B-Hexosaminidases 165 barrier from any stable enzyme-substrate or enzyme intermediate complex to the susequent transition state. Alternatively, the data may not be scattered, but actually represent a biphasic plot indicating that the rate determining step varies for SpHex depending on the extent of fluorine substitution of the substrate. Increasing the extent of fluorine substitution will make the acetamido group a poorer nucleophile and a better leaving group and this should have the two-fold effect of slowing down the cyclization step and accelerating the ring-opening step. As the ring-opening step is known to be rate limiting for pNPGlcNAc (vide infra) it is quite possible that the cyclization step becomes rate determining for pNPGlcNAcF2 and/or pNPGlcNAcF3. Regardless of the source of these complications we will avoid vagaries and avoid the interpretation of such data. Despite these limitations several important points can be made on the basis of the data outlined in Table 3.3 and Figure 3.17. Firstly, as expected the two enzymes have greatly differing requirements for the acetamido group and enzymes from each family can be readily and rapidly distinguished using these substrates. Indeed, on this basis we can infer that the enzymes that Yamamoto and Kosman isolated from natural sources are both family 20 P-hexosaminidases. Secondly, the substituted methyl group exerts neither an appreciable steric or electronic effect on the transition state for the ExoII catalyzed hydrolysis of these substrates (p*kcat = -0.17 and p*kcat/Km = 0.07) and therefore is not a critical recognition element for this enzyme. Indeed, there is little effect on kcat or on Km as can be seen from Table 3.3. In conjunction with the results of the experiments using 2,4DNPG we can conclude that it is only the amide functionality that is the important component of the acetamido moiety for catalysis by ExoII. Thirdly, the methyl group is strictly recognized in the transition state of SpHex catalyzed hydrolysis of these substrates (p* + 5 = -1.29). Although we cannot distinguish the extent of steric contributions (8) from the electronic contributions (p*) it is likely that both have some effect. Jones and Kosman, using pNPGlcNAc, pNPGlcNAcF2, and pNPGlcNAcF3 as substrates for Aspergillus niger P-hexosaminidase determined from their three point correlation a slope of -1.41 (slope = p* + 8 = -1.41) but in the absence of additional information ascribed it entirely to electronic effects (p*). It it primarily on this rather slender evidence that they hypothesized a mechanism involving anchimeric assistance for Aspergillus niger P-hexosaminidase. Yamamoto and Chapter 3 Mechanism of B-Hexosaminidases 166 coworkers made attempts to separate the electronic and steric effects by preparing a wider range of para-nitrophenyl glucoaminides modified at the acetamido substituent. They were able to partition the contributions arising from sterics (8 = -0.46), electronics (p* = -0.46), and hyrophobicity and increase the linear correlation of their data. Although their studies were compromised by the fact that they only performed their kinetic analysis at one fixed concentration of substrate (1 mM) their studies suggest that both steric and electronic factors play some role in catalysis. Regardless, the pronounced negative slope for the substrates used in our study is consistent with those observed by the independent groups of Kosman and Yamamoto and suggests that the active site architecture surrounding the acetamido substituent is tightly held by the enzyme. This is consistent with the observation that the acid catalyzed solvolysis of sugar oxazolines typically proceeds by attack of water at the acyl center yet the enzyme-catalyzed reaction proceeds exclusively by attack at the anomeric center. The enzyme presumably forms a tight fitting pocket wherein the acetamido group is carefully positioned for catalysis and the acyl center of the oxazolinium ion intermediate is shielded from water. Fourthly, we can hypothesize that the acetamido group is interacting with a positively charged center in the transition state of the SpHex catalyzed reaction. A steep negative correlation in Taft analyses of nucleophilic participation, such as that found for SpHex, may be considered to indicate that the nucleophile interacts with a cationic center at the transition state. For SpHex it is quite reasonable to speculate, on the basis of considerable precedent with other glycosidases, that this cationic center is the anomeric carbon. 3.5.4 Exocyclic Oxonium Ion or Oxocarbenium Ion-like Transit ion States In an early publication Kosman and Jones27 advanced a different mechanistic proposal involving the formation of an exocyclic oxonium ion-like transition state (Similar to (a), Figure 3.18) in which they stated: "The glycosidic oxygen is protonated very early in the reaction, perhaps even in the Michaelis complex". Chapter 3 Mechanism of ^ -Hexosaminidases 167 a A' •ENZ b A' •ENZ HO HO R HO R NHAc Figure 3.18 Structures of a) exocylic oxonium ion and b) oxocarbenium ion. In such a transition state there would be very little fission of the glycosidic bond but nearly complete protonation of the glycosidic oxygen. At the time of this proposal such a view was in keeping with the equilibrium protonation advocated by Phillips and coworkers for HEWL (See Chapter 1). Such a view, however, is quite at odds with the current view of acid catalysis by glycosidases. Kosman and Jones rationalize their conclusion on the basis of the small solvent isotope effect (&H/&D = 1.27) observed during the hydrolysis of pNPGlcNAc, which indicates that the proton is either almost completely transferred or hardly transferred to the glycosidic oxygen. They also indicate that the small, slightly negative Pig(kcat/Km) value (Pig = - 0.1), determined from the A. niger (3-hexosaminidase catalyzed hydrolysis of a series of four para-substituted 2-acetamido-2-deoxy-P-D-glucopyranosides suggests little charge development on the glycosidic oxygen, supporting their view. For spontaneous and alkaline hydrolysis of glucosides the Pig value is typically large and negative (Pig = - 1) while for specific acid catalysis where an oxonium ion is formed the Pig value is typically small and positive (Pig = 0.2 to 0.02).284 An alternative interpretation of the data of Jones and Kosman would be that the transition state is more oxocarbenium ion-like with significantly advanced fission of the glycosidic bond ((b), Figure 3.18) and significant proton donation. Indeed, such a late transition state is more consistent with the Taft analysis that indicates formation of a positively charged center. The limited Br0nsted analysis carried out by Kosman and Jones is the only such study carried out with any P-hexosaminidase. Although we suspect that the A. niger enzyme they were working with is a family 20 glycosidase we felt it would be prudent to conduct a more rigorous study comparing the family 3 ExoII and family 20 SpHex enzymes. To this end we prepared a series of aryl N-acetyl-glucosaminides with varying Chapter 3 Mechanism of ^-Hexosaminidases ; 168 leaving groups. The kinetic parameters found with each enzyme are summarized in Table 3.4. The kinetic parameters can be plotted in a series of Br0nsted linear free energy relationships (Figure 3.19) where several striking trends emerge. Firstly, with respect to ExoII we see behavior similar to that described in Chapter 2 for XynB. The plot of \og{kcJKm) against pKa of the phenol leaving group shows a strong correlation for all substrates. The plot of log(fccat) against the pKa of the phenol leaving groupalso reveals a significant correlation but only for substrates bearing poor leaving groups with pKa > 7. The values of pig(kcat) = -0.78 (n = 6, r = -0.99) obtained from a linear regression of the data for all substrates other than 2,3DNPGlcNAc and the pig(kcat/Km) = -0.79 (n = 7, r = -0.99) obtained from a linear regression of all the data are very similar. Indeed, identical slopes are expected if the rate determining step, governed by kcaU is also the first irreversible chemical step of the reaction, as governed by kcJKm. Thus the similarity of the pig values suggests that the rate determining step for these substrates (pKa > 7) with ExoII must be the glycosylation step. The deviation of 2,3DNPGlcNAc from the correlation of log(fccat) presumably arises because the rate-determining step for this substrate with ExoII is the deglycosylation step. The relatively low Km value that is found for 2,3DNPGlcNAc as compared to poorer substrates (pKa > 7) supports this view. This low Km value may simply be a consequence of accumulation of the glycosyl enzyme intermediate as discussed earlier in this chapter (Section 3.1.2). Although studies to substantiate this claim have not been carried out for ExoII similar behavior has been observed with several retaining exo-p-glycosidases including p-glucosidases from sweet almonds,56 Agrobacterium sp.,44 and Pyrococcus furiosus,5 and the p-xylosidase from Thermoanaerobacterium saccharolyticum (Chapter 2). The magnitudes of the pig values observed in this work are also consistent with the studies just listed where Pig values commonly lie between -0.8 and -0.95. Secondly, linear regression of the data from SpHex for both \og(kcJKm) and log(£ c at) versus pKa of the leaving group phenol yields good correlations throughout the entire range of substrates. This indicates that the rate-determining step remains the same for all substrates. Chapter 3 Mechanism of 0-Hexosaminidases 169 Table 3.4 Kinetic parameters for the hydrolysis of a series of aryl glucosaminides catalyzed by SpHex and ExoII. Substrate Phenol pKa Chemical Structure Enzyme Km ^cat/Km (sec1) (mM) (sec" mM" ) ExoII 10.8 + 0.2 0.086 ±0.01 130 ± 10 SpHex 235 ± 9 0.019 ± 0.002 12000±1000 ExoII 4.8 + 0.4 1.2 + 0.1 4.0 ± 0 . 4 SpHex 222 + 6 0.048± 0.005 4600 ± 5 0 0 ExoII 3.20 + 0.04 0.73 ± 0.03 4.4 ± 0 . 2 SpHex 193 ± 3 0.049 ± 0.004 3900 + 400 ExoII 2.08 + 0.08 1.2 + 0.1 1.8 ± 0.1 SpHex 180 ± 7 0.054 ± 0.03 3300 ± 300 ExoII 0.33 ±0.01 1.5 ± 0.1 0.22 ± 0.02 SpHex 197 ± 7 0.15 + 0.03 1300±100 ExoII N.D. N.D. N.D. SpHex 206 + 8 0.22 ±0 .01 940 ± 90 ExoII 0.054 + 1.1 ±0 .1 0.05 ± 0.06 0.005 SpHex 187 + 8 0.27 ± 0.04 700 ± 1 0 0 ExoII 0.012 ± 1.0 ±0 .1 0.011 ±0.002 0.005 SpHex 172 + 3 0.48 ±0.01 360 ± 70 2,3DNPGlcNAc 4.96 3,5DNPGlcNAc 6.69 pNPGlcNAc 7.18 MuGlcNAc 7.50 mNPGlcNAc 8.39 pNHAcGlcNAc 9.50 mNHAcGlcNAc 9.60 PGlcNAc 9.99 n OaN HO—V-J °\ sN°i AcHN \ ^ HO,. HO— .NO; H o \ ^ » ^ 0 ^ / = ^ AcHN N02 AcHN \^~~ND.. " ° \ a o~S° HO—^-\--0. ,NO HO-X««-^ «\-- 0-"-£*\ AcHN HO~ HO—^ A.—°. H0A^^-X-0^/^==v AcHN \^~~NH HO-. HO-V«"^ »\^ 0~-^ =\ AcHN ^ ^  HO—^-X—O. AcHN pKa values used for phenols were taken from references '76-178 The markedly different pig(kcat) (-0.02) and Pi^ cat/Km) (-0.29) values further suggest that the rate determining step is not the first irreversible chemical step. Indeed &cat, which reflects the rate determining step, shows essentially no dependence on leaving group ability (Pig(kcat) ~ 0) indicating that the rate determining step is the deglycosylation step. The trend in Km also supports such a hypothesis. As the leaving group ability improves, Km decreases (Table 3.4) and this indicates that the steady state concentration of an enzyme sequestered intermediate is increasing (see Chapter 2, page 90 for a discussion). An alternative explanation for the Br0nsted data that would invoke a non-chemical rate Chapter 3 Mechanism of 0-Hexosaminidases 170 alternative explanation for the Br0nsted data that would invoke a non-chemical rate determining step is ruled out by the cc-deuterium kinetic isotope effect of significant magnitude (vide infra) measured on &cat using pNPGlcNAc. The magnitude of the Pig(kcat/Km) value is greater that that estimated by Kosman and Jones although the sign is the same in both studies. The small negative (3ig(kCat/Km) (-0.29) reflects the accumulation of a relatively small amount of negative charge on the glycosidic oxygen at the transition state of the glycosylation step and suggests that cleavage of the glycosidic bond is slightly, yet significantly, more advanced than is proton donation to the glycosidic oxygen. Such a small negative (3ig value has been observed for several retaining P-glycosidases including the xylanases from Cellulomonas fimfi3 and Bacillus circulans.2^5 6 5 4 3 5 -> ££> OD 1 O ~" 0 -1 -2 -3 4 5 6 7 8 9 10 11 pKa 6 5 4 "s 3 2 a £ 1 M £ 0 -1 -2 -3 4 5 6 7 8 9 10 11 pKa Figure 3.19 Br0nsted plots of the log of the first (&cat) and second (kcJKm) order rate constants for the ExoII and SpHex hydrolysis of a series of aryl glucosaminide substrates, (a) Plot of log(^cat^m) against the pKa of the leaving group phenol, and (b) plot of log(£ c a t) against the pKa of the leaving group phenol. Filled circles (•) represent data for ExoII and open circles (O) represent data for SpHex. Chapter 3 Mechanism of 0-Hexosaminidases 171 The negative Pig values measured for both enzymes indicate that the glycosidic oxygen in the transition state accumulates net negative charge, presumably as a consequence of the cleavage of the glycosidic bond. The difference in the magnitude of the Pig(kcat/Km) values between SpHex and ExoII hints at a different requirement for acid catalysis for each enzyme with proton donation playing a more significant role for the family 20 SpHex P-hexosaminidase. 3.5.5 a-Deuterium Kinetic Isotope Effects . One direct and rapid test to gain insight into the nature of the transition state would be to carry out a-deuterium kinetic isotope effect studies. As described in the introduction and in chapter 2, a-deuterium kinetic isotope effects (aD-KIE's) can be of considerable use in the study of glycosidase mechanism and here they would serve a two-fold purpose. Firstly, they would clarify whether the transition state is more oxonium ion-like ((a), Figure 3.18) as suggested by Kosman and Jones or more oxocarbenium ion-like ((b), Figure 3.18) as we advocate here. Secondly, they may serve to ascertain whether the intermediates of the ExoII and SpHex catalyzed reactions are covalent or stabilized oxocarbenium ions. If the transition state is oxonium ion-like then we expect that the aD-KIE will be approximately zero as the anomeric center is essentially sp hybridized in both the starting material and the transition state. Conversely, if the transition state is oxocarbenium ion-like, then the hybridization of the anomeric center 3 2 changes from sp in the ground state to sp in the transition state, thus a normal aD-KIE should be observed. In every case where such studies have been carried out they have revealed that the transition states leading to the formation and breakdown of the covalent glycosyl enzyme have considerable oxocarbenium ion-like character.11 However, the only P-hexosaminidase to have been investigated by kinetic isotope effect studies is HEWL where a normal aD-KIE of &H/&D =1.14 was found. No such studies have been carried out on any exo-P-hexosaminidase. If a significant and normal aD-KIE value is measured for kcal with SpHex, as we expect, we can infer that the intermediate must be a covalent oxazoline. Some investigators have suggested that the intermediate is an Chapter 3 Mechanism of B-Hexosaminidases 172 oxocarbenium ion intermediate merely stabilized by the acetamido group286 while others have voiced their uncertainty.27-240 The ion pair intermediate also has considerable support from those who maintain the Phillips ion-pair mechanism for HEWL. 1 3 Lastly, in addition to clarifying these aspects of the mechanism of the family 20 (3-hexosaminidases these isotope effects would provide further insight into the similarities and differences of the anchimeric assistance mechanism and the double displacement mechanism. To this end we prepared pNPGlcNAc isotopically enriched with deuterium at the anomeric center. Scheme 3.9 Synthesis of 1-{2H)-pNPGlcNAc. i) D 20, NaBD4, THF, 82%; ii) Pd-C, H 2 , EtOAc, MeOH, 88%; iii) HC1, AcCl, 0 °C -> RT; iv) pNP, Bu 4NHS0 4, CH2C12, 1 M NaOH, 73% over two steps; v) a) NaOMe, MeOH, b) Amberlite IR-120 (H+) resin, 78%. The known lactone 3.25 was prepared in four steps from GlcNAc according to the procedure of Granier and Vasella.287 Reduction of the lactone using NaBD4 followed by hydrogenolysis of the benzyl ether protecting groups provided C-l deuterated GlcNAc (3.26). Treatment of the hemiacetal with hydrogen chloride-saturated acetyl chloride afforded the crude glycosyl chloride donor (3.28), which was used without further purification. Glycosylation of donor 3.28 using the phase transfer method of Roy 2 4 8 followed by global deprotection provided the substrate para-nitrophenyl 2-acetamido-2-deoxy-1 -deutero-P-D-glucopyranoside (3.30). With the deuterium labeled substrate in hand we sought to determine the aD-KIE values for both enzymes. Unfortunately, owing to the limited solubility of the substrate Chapter 3 Mechanism of B-Hexosaminidases 173 and the relatively high Km value of this substrate for ExoII (Km = 0.73 mM) complete substrate saturation could not be reached and an a~D(V)iciE could therefore not be measured. The a~D(V/K)KiE value, however, could be measured for this enzyme using the substrate depletion method (Described in Chapter 2, section 2.4.2). Conversely, for SpHex, the Michaelis constant is so small (^m = 0.05 mM) as to make accurate spectrophotometric measurements by the substrate depletion method difficult. Therefore the isotope effect was measured under conditions of saturating substrate yielding an a_D(V)iciE value for SpHex. As mentioned above the kcat value for all aryl glucosaminides, including pNPGlcNAc, reflects the deglycosylation step for SpHex. However, as the cyclization and ring-opening steps are the near microscopic reverse of each other, these values provide some evidence into the natures of the transition states for both steps. The isotope effect measured for SpHex (&H/£D = 1.07 ± 0.01) reveals a transition state having significant oxocarbenium ion character that is consistent with the Taft-like analysis, which suggested that the acetamido group interacts with a positively charged center in the transition state. This isotope effect also indicates that the oxazoline intermediate is a covalent species and not an ion-pair as has been posited.27-239-286 The isotope effect measured for ExoII (k^lko = 1.10 ± 0.02) is very close to that measured for SpHex and suggests that the charge build up at C-l in the transition states of both enzymes is similar. These values are also within the range found for retaining P-glycosidases to date (1.05-1.11 for the glycosylation step and 1.08-1.25 for the deglycosylation step)11 and therefore support the idea of a common catalytic mechanism involving oxocarbenium ion-like transition states bracketing a covalent intermediate for all of these enzymes regardless of the nature of the nucleophile. On the basis of these results we speculate that the intermediate in the catalytic pathway of any given retaining P-glycosidases may be either a covalent glycosyl enzyme or a covalent bicyclic oxazoline. Chapter 3 Mechanism of B-Hexosaminidases 174 3.6 Summary The studies on the retaining exo-P-hexosaminidases outlined in this chapter reveal that there are two distinct classes of these enzymes. One group comprise family 20 of glycoside hydrolases and the second group are part of family 3. A series of studies on the Streptomyces plicatus P-hexosaminidase using the novel thiazoline inhibitors in conjunction with substrate reactivity studies using N-fluoroacetyl glycosides and X-ray crystallographic analysis of the NAG-thiazoline-enzyme complex reveals that this family 20 enzyme uses the 2-acetamido group of the substrate as a nucleophile. Indeed, Brian Mark, a doctoral candidate in the research laboratory of Prof. M.N.G. James, recently solved the three dimensional structure of SpHex in complex with NAG-thiazoline (Figure 3.20). This structure reveals that the NAG-thiazoline adopts a relaxed AC\ chair conformation with the acyl center of the thiazoline ring shielded by tryptophan residues, much as predicted. Conversely, mass spectrometric evidence using the mechanism based inhibitor, 5FIdNAcF, suggests that the family 3 Vibrio furnisii P-hexosaminidase uses an enzymic nucleophile to form a covalent glycosyl enzyme intermediate. Figure 3.20 NAG-thiazoline (NGT) and glycerol (Gol) bound to sugar binding subsites -1 and +1 of SpHEX, respectively. The catalytic triad (Glu314, His250, and Asp 191) has been drawn along with its hydrogen-bonding network. The glycerol hydroxyl group hydrogen bonding to the carboxylate of Glu314 is believed to occupy the position that an incoming water molecule would take to nucleophilically attack C- l . WAT indicates the conserved incoming water molecule proposed by Reference.76 Figure provided courtesy of Brian Mark. Chapter 3 Mechanism of 0-Hexosaminidases 175 Despite this difference in the identity and nature of the nucleophile the transition states for both enzymes appear remarkably similar. Cleavage of the glycosidic linkage is significantly advanced, with considerable charge development on the anomeric center as evidenced by the normal and significant a-deuterium kinetic isotope effects observed for both enzymes. Br0nsted analysis of both enzymes also reveals that proton donation from the acid catalyst lags behind cleavage of the glycosidic bond to a varying extent. The consequence of these interpretations is that the transition state can be viewed as highly polar, with accumulation of net negative charge on the glycosidic oxygen and the -1 saccharide unit having significant oxocarbenium ion-like character. These results are consistent with the very large majority of studies on retaining (3-glycosidases and advance the central hypothesis of this thesis that all retaining [3-glycosidases use a fundamentally similar mechanism. The remaining known exception to this theory is HEWL, which has long been considered to proceed by a reaction mechanism involving an ion-pair intermediate. The subsequent chapter will center on attempts to resolve this apparent conflict. 3.7 Experimental Procedures 3.7.1 General Procedures All buffer chemicals and other reagents were obtained from the Sigma/Aldrich/Fluka Chemical Co. unless otherwise noted. Solvents and reagents used were either reagent, certified, or spectral grade. Anhydrous solvents were prepared as follows. Methanol was distilled from magnesium turnings in the presence of iodine, tetrahydrofuran was distilled from sodium in the presence of benzophenone. Dichloroethane, toluene, pyridine, triethylamine, acetonitrile, and dichloromethane were prepared by distillation from calcium hydride. Solvents were distilled immediately prior to use. Dimethylformamide was dried sequentially over 4 A molecular sieves. Chapter 3 Mechanism of 0-Hexosaminidases 176 Synthetic reactions were monitored by TLC using Merck Kieselgel 60 F254 aluminum-backed sheets (thickness 0.2 mm). Compounds were detected by ultraviolet light (254 nm) and/or by charring with 10% ammonium molybdate in 2 M H2SO4 and heating. Flash chromatography under a positive pressure was performed with Merck Kieselgel 60 (230-400 mesh) using the specified eluants. [ H NMR spectra were recorded on a Bruker WH-400 spectrometer at 400 MHz, a Bruker AV-300 at 300 MHz, or a Bruker AC-200 at 200 MHz. 1 9 F NMR spectra were recorded on a Bruker AC-200 at 188 MHz or a Bruker AV-300 at 282 MHz and are proton-coupled with CF 3 C0 2 H as a reference.Chemical shifts are reported on the 8 scale in parts per million from tetramethylsilane (TMS) and were measured relative to CDCI3, CD3OD, or to DSS when taken in D2O. The abbreviations used in describing multiplicity are: s-singlet, d-doublet, 1 t-triplet, m-multiplet, and br-broad. Carbon nuclear magnetic resonance ( C-NMR) spectra were obtained on a Varian XL-300 spectrometer at 75.5 MHz. Signal positions are given in parts per million (ppm) from tetramethylsilane and were measured relative to the signal of CDCI3, CD3OD, or D2O. The Mass Spectrometry Laboratory, University of British Columbia performed both high and low resolution mass spectra. Mr. Peter Borda of the Microanalytical Laboratory, University of British Columbia performed elemental analyses. Melting points were recored using a Laboratory Devices Mel-Temp II melting point apparatus and are uncorrected. 3.7.2 Generous Gifts Prof. Saul Roseman graciously provided the plasmid bearing the ExoII gene. The ExoII gene was subcloned into a high expression vector by Dr. Christoph Mayer, a postdoctoral fellow in the laboratory of Prof. Withers. Purified Vibrio furnisii ExoII was kindly provided by Dr. Christoph Mayer and Melanie Mah, a technician working in the laboratory of Prof. Warren. The plasmid bearing the Streptomyces plicatus SpHex gene was provided by Brian Mark a doctoral candidate in the laboratory of Prof. James at the University of Alberta and the protein was purified essentially as outlined in reference 2 8 8 . pNPGlcNAcF and reference samples of pNPGlcNAcF2 and pNPGlcNAcF3 were Chapter 3 Mechanism of B-Hexosaminidases 177 generously provided by Dr. Tom Harvey, a postdoctoral fellow working in the laboratories of Profs. Weiler and Withers. 3.7.2.1 Syntheses The synthesis of pNPGlcNAcF2 and 2-deoxy-3,4,6-tri-0-acetyl-2-trifluoroacetamido-p-D-glucopyranosyl bromide289 was performed as described in the literature. 2-Acetamido-3,4,6-tri-0-acetyl-2-deoxy-5-fluoro-a-L-idosylfluoride (3.22). 2-Acetamido-3,4,6-tri-0-acetyl-2-deoxy-P-D-glucopyranosyl fluoride (3.20)26 (0.922 g, 4.1 mmol) was suspended in carbon tetrachloride (220 mL), and N-bromosuccinimide (2.36 g, 13.3 mmol) was added. This suspension was heated to reflux between two 200 W light bulbs. After 2 h, the reaction mixture was allowed to cool to room temperature, and dichloromethane (160 mL) was added. The solution was washed with saturated sodium hydrogen carbonate (150 mL) and water (2 x 150 mL). The organic layer was dried (MgSCU) and filtered, and the solvent was removed in vacuo. The resulting yellow gum was immediately purified on a short column of silica gel (ethyl acetate/hexanes/triethylamine, 2:1:0.015) to yield the crude intermediate 2-acetamido-3,4,6-tri-0-acetyl-5-bromo-2-deoxy-p-D-glucosyl fluoride (3.21): 1 9 F NMR (188 MHz, CDC13, CF 3 C0 2 H reference) -44.7 (1 F, dd, 7Fi,Hi = 57.0 Hz). This golden syrup (223 mg) was dried under vacuum for 3 h, and, under an atmosphere of nitrogen, silver fluoride (0.94 g, 7.4 mmol) and activated 4 A molecular sieves (50 mg) were added. Dry acetonitrile (50 mL) was added to the flask, and the reaction mixture was stirred under nitrogen for 18 h. The reaction mixture was then filtered through Celite/siiica to remove silver salts and the solvent removed in vacuo to yield a translucent pale yellow syrup. This residue was purified by careful gradient flash chromatography (ethyl acetate/hexanes, 1:1 to 2:1) to yield the product (3.22) as a clear gum (120 mg, 0.33 mmol, 8%): 1 9 F NMR (188 MHz, CDC13, CF 3 C0 2 H reference) -44.3 (1 F, dd, 7FI.HI = Chapter 3 Mechanism of 0-Hexosaminidases 178 52.4 Hz, 7 F 1 , H 2 = 10.1 Hz, F-l), -30.8 (1 F, dddd, 7 F 5 , H 6 = 17.0 Hz, 7 F 5 , H 6 ' = 15.3 Hz, 7 F 5 , H 4 = 8.5 Hz, 7F5,HI = 1.5 Hz, F-5); ! H NMR (400 MHz, CDC13) 5.99 (1 H, d, 7 N H , H 2 = 8.6 Hz, NH), 5.63 (1 H, ddd, 7 H i , H 2 = 3.9 Hz, H-l), 5.47 (1 H, dd, 7 H 4 , H 3 = 5.6 Hz, H-4), 5.08 (1 H, dd, 7 H 3 , H 2 = 6.0 Hz, H-3), 4.47 (1 H, ddd, H-2), 4.31 (1 H, d, H-6), 4.31 (1 H, d, H-6'), 2.1 (9 H, s, 3 x OAc), 2.0 (3 H, s, NAc); +CIMS (NH3): m/z 385 (M + NH 4) + (22%), 368 (M + H)+ (100%), 348 (M - F)+ (20%); high-resolution +CIMS (M + NH 4)+: calculated, 368.1158; found, 368.1150. Synthesis of' 2-acetamido-2-deoxy-5-fluoro-oc-L-idosyl fluoride (5FIdNAcF, 3.19). The protected 5-fluoro-sugar (23.5 mg, 0.064 mmol) (3.22) was dissolved in dry methanol (3 mL). After being cooled to 4 °C, the stirred solution was saturated with ammonia. This mixture was warmed to room temperature and stirred for 25 min, after which time the solvent was evaporated in vacuo to give a clear gum. The desired product was then purified by flash chromatography on silica gel (ethyl acetate/methanol/water, 17:2:1) to yield the desired product (3.19) (13.6 mg, 0.056 mmol, 88%) as a transparent gum: 1 9 F NMR (188 MHz, CD 3OD, CF3CO2H reference) -44.1 (1 F, dd, 7 F i , H 1 = 54.7 Hz, 7 F 1 , H 2 = 13.6 Hz, F-l), -33.2 (1 F, dddd, 7 F 5 , H 6 = 14.6 Hz, 7 F 5 , H 6 ' = 12.2 Hz, 7 F 5 , H 4 = 12.2 Hz, 7 f 5,HI = 1.7 Hz, F-5); 'H NMR (400 MHz, CD3OD) 5.49 (1 H, dd, 7 H i , H 2 = 3.9 Hz, H-1), 4.19 (1 H, ddd, 7 H 2 , H 3 = 8.4 Hz, H-2), 3.98 (1 H, dd, 7 H 4 ,H 3 = 7.1 Hz, H-4), 3.80 (1 H, dd, 7 H 6 , H 6 ' = 12.2 Hz, H-6), 3.72 (1 H, dd, H-6'), 3.71 (1 H, dd, H-3), 2.01 (3 H, s, NAc); +CIMS (NH4)+: m/z 242 (M + H)+ (29%), 222 (M - F)+ (100%); high-resolution +CIMS (M + H)+: calculated, 242.0840; found, 242.0838. Synthesis of l,2-dideoxy-3,4,6-tri-0-acetyl-2 '-methyl-a-D-glucopyranoso-[2,l-d]-A2 '-thiazoline. (3.13) The protected 2-acetamido-2-deoxy-l,3,4,6-tri-0-acetyl-p-D-glucopyranose (1.69 g, 4.3 mmol) in anhydrous toluene (20 mL) was treated with Lawesson's reagent (1.14 g, 2.84 mmol). This mixture was heated to 85 °C for 105 minutes after which time the reaction was judged complete by TLC analysis. The reaction mixture was cooled to Chapter 3 Mechanism of B-Hexosaminidases 179 room temperature and concentrated in vacuo to yield a yellow syrup. The desired product was then purified by flash chromatography on silica gel (ethyl acetate / hexanes, 1:1) to yield the desired product (920 mg, 4.0 mmol, 94 %) as a transparent gum: *H NMR (200 MHz, CDC13) 6.22 (1 H, d, 7Hi,H2 = 7.0 Hz, H-1), 5.55 (1 H, dd, 7H3,H2 = 3.2 Hz, 7H 3,H4 = 1-8 Hz, H-3), 4.94 (1 H, ddd, 7H 4,H5 = 9.4 Hz, 7H 4,H2 = 1.8 Hz, H-4), 4.46 (1 H, m, H-2), 4.11 (2 H, m, H-6, H-6a), 3.54 (1 H, ddd, 7H 5 >H6 = 4.5 Hz, 7H 5,H6a = 4.5 Hz, H-5), 2.31 (3 H, d, 7M e,H2 = 2.1 Hz, Thiazoline Me), 2.12, 2.06, 2.02 (3 H, s, OAc). Synthesis of l,2-dideoxy-2'-methyl-a-D-glucopyranoso-[2,l-d]-A2'-thiazoline. (NAG-thiazoline, 3.14) A cooled (0 °C) solution of l,2-dideoxy-3,4,6-tri-0-acetyl-2'-methyl-a-D-glucopyranoso-[2,l-d]-A2'-thiazoline (588 mg, 1.7 mmol) in anhydrous methanol (25 mL) was saturated with anhydrous ammonia gas. The reaction mixture was allowed to slowly warm to room temperature and after 3 hours the reaction was judged by TLC analysis to be complete. The mixture was concentrated in vacuo to provide a pale yellow gum. Flash column chromatography on silica gel (ethyl acetate / methanol / water, 14:2:1) yielded the desired product. (316 mg, 1.45 mmol, 85 %) as a clear gum. 'H NMR (200 MHz, CDCI3) 6.36 (1 H, d, 7Hi,H2 = 9.0 Hz, H-1), 4.38-4.25 (1 H, m, H-2), 4.12 (1 H, dd, 7H 3,H2 = 4.7 Hz, 7H 3,H4 = 3.9 Hz, H-3), 3.73 (1 H, dd, 7H 6,H6a = 12.1 Hz, 7H 6,H5 = 2.6 Hz, H-6), 3.57 (1 H, dd, 7H6a,H5 = 6.1 Hz, H-4), 3.51 (1 H, m, H-4), 3.31 (1 H, ddd, 7H 4,H 5 = 9.4 Hz, H-5), 2.23 (3 H, d, 7M e,H2 = 2.0 Hz, Thiazoline Me); Anal. Calcd for C 8 H 1 3 N0 4 S; C, 43.82; H, 5.98; N, 6.39. Found: C, 43.64; H, 6.03; N 6.12. Synthesis of l,2-dideoxy-tri-0-acetyl-2 '-methyl-a-D-galactopyranoso-[2,l-d]-A2 '-oxazoline (3.31). To a solution of the protected 2-acetamido-2-deoxy-1,3,4,6-tri-0-acetyl-(3-D-galactopyranose (155 mg, 0.4 mmol) in anhydrous toluene (3 mL) was treated with Lawesson's reagent (98 mg, 0.24 mmol). This mixture was heated to reflux for 90 minutes after which time the reaction was judged complete by TLC analysis. The Chapter 3 Mechanism of 0-Hexosaminidases 180 reaction mixture was cooled to room temperature and concentrated in vacuo to yield a yellow syrup. The residue was subjected to flash chromatography on silica gel (ethyl acetate / hexanes, 1:1) to yield the desired product (920 mg, 4.0 mmol, 94 %) as a transparent gum: *H NMR (400 MHz, CDC13) 6.22 (1 H, d, 7 H I ,H2 = 6.30 Hz, H-l), 5.49 (1 H, dd, 7H4,H3 = 3.3 Hz, 7H4,H5 = 3.3 Hz, H-4), 5.18 (1 H, dd, 7 H 3 , H 2 = 7.8 Hz, H-3), 4.36 (1 H, ddd, 7 M e , H 2 = 1-4 Hz, H-2), 4.29-4.24 (1 H, m, H-5), 4.19 (1 H, dd, 7 H 6 ,H6a = H-3 Hz, 7H6,H5 = 7.4 Hz, H-6), ), 4.07 (1 H, dd, 7H6a.H5 = 5.6 Hz, H-6a), 2.23 (3 H, d, Thiazoline Me), 2.10, 2.06, 2.02 (3 H, s, OAc); mp 155-156 °C; Anal. Calcd for C 8Hi 3N0 4S; C, 43.82; H, 5.98; N, 6.39. Found: C, 43.73; H, 5.85; N 6.32. Synthesis of l,2-dideoxy-2 '-methyl-a-D-galactopyranoso-[2,1 -d]-A2 '-thiazoline (NAGal-thiazoline, 3.32) To a solution of l,2-dideoxy-tri-0-acetyl-2'-methyl-a-D-galactopyranoso-[2,l-d]-A2'-oxazoline (100 mg, 0.29 mmol) in anhydrous methanol (20 mL) was added 5 drops of a 1 M solution sodium methoxide in methanol. After 4 hours the reaction was judged to be complete by TLC analysis. Amberlite IR-120 resin (H+) was added until the reaction mixture was neutral (pH 7). The resin was removed by filtration and the solution was concentrated in vacuo to yield a crystalline solid. Recrystallization of this residue from methanol - diethylether provided the desired product (55 mg, 0.25 mmol, 86 %) as a white needles: 'H NMR (300 MHz, CD3OD) 6.36 (1 H, d, 7 H i , H 2 = 6.5 Hz, H-l), 4.26 (1 H, dd, 7 H 2 , H 3 = 6.8 Hz, 7 H 2 .Thiazoline Me — 1.5 Hz, H-2), 3.93-3.82 (3 H, m, H-3, H-4, H-5), 3.76 (1 H, dd, 7H6,H6a = 11-6 Hz, 7 H 6 > H 5 = 6.7 Hz, H-6), 3.69 (1 H, dd, 7 H 6 a,H5 = 5.0 Hz, H-6a), 2.23 (3 H, d, Thiazoline Me); mp 155-156 °C; Anal. Calcd for C 8Hi 3N0 4S; C, 43.82; H, 5.98; N, 6.39. Found: C, 43.73; H, 5.85; N 6.32. Synthesis of para-nitrophenyl 2-deoxy-3,4,6-tri-0-acetyl-2-trifluoroacetamido-fi-D-glucopyranoside (3.37) An alternative to the approach described in the literature233 was chosen to prepare large quantities of this compound. Dr. Tom Harvey a postdoctoral fellow in the Chapter 3 Mechanism of B-Hexosaminidases 181 laboratory of Prof. Withers first carried out this synthesis. Under an atmosphere of nitrogen, to a solution of 2-deoxy-3,4,6-tri-0-acetyl-2-trifluoroacetamido-P-D-glucopyranosyl bromide (3.48 g, 7.5 mmol)289 in anhydrous acetonitrile (30 mL) was added CaSC>4 (2 g). To a different solution, also under an atmosphere of nitrogen, containing 2,6-lutidine (1.86 mL, 18.8 mmol, 2.5 equiv.) and /?ara-nitrophenol (dried in vacuo overnight, 2.1 g, 15.1 mmol, 2 equiv.) in anhydrous acetonitrile (30 mL) was added CaSC>4 (2 g). These two separate slurries were then stirred at room temperature for 30 minutes after which time the mixture containing the phenol was cannulated into the flask containing the glycosyl bromide. Silver carbonate was added (1.32 g, 0.75 eq) to the suspension and the resulting mixture was stirred vigorously in the dark for 3 hours after which time TLC analysis revealed that the glycosyl bromide had entirely reacted. The reaction mixture was diluted with acetonitrile and the solids removed by filtration through a bed of silica and Celite. The filtrate was concentrated in vacuo to provide a gray gum. The desired product was then purified by flash chromatography on silica gel (ethyl acetate/hexanes;l:l) to yield the desired product (3.37, Figure 3.21) (3.05 g, 5.9 mmol, 78%) as a crystalline solid. 'H and 1 9 F NMR data for compound 3.37 can be found in tables 3.6 and 3.7. General synthesis of aryl 2-acetamido-2-deoxy-3,4,6-tri-0-acetyl-f3-D-glucopyranosides (3.33-3.36 and 3.38-3.41) To a mixture of 2-acetamido-2-deoxy-3,4,6-tri-0-acetyl-a-D-glucopyranosyl chloride248 (1 eq.), tetrabutylammonium hydrogensulfate (1 eq.) and the acceptor phenol (2 eq.) was added sufficient dichloromethane (1 volume) to yield a solution containing the donor chloride at a concentration of 200 mM. An equal volume of 1 M NaOH (1 volume) was then added and this mixture was rapidly stirred at room temperature for 1 to 3 hours. After the reaction was judged complete by TLC analysis ethyl acetate (5 volumes) was added. The solution was washed with 1 M NaOH (4x2 volumes), water (2x2 volumes) and saturated sodium chloride solution (2 volumes). The organic layer was dried (MgS04) and filtered, and the solvent was removed in vacuo. The resulting crude product was crystallized from a mixture of ethyl acetate and hexanes to provide Chapter 3 Mechanism of 0-Hexosaminidases 182 yields after 1 crop of between 45 and 70% of the desired aryl glycoside (3.33-3.36 and 3.38-3.41, Figure 3.21). *H NMR data for the per-acetylated aryl N-acetyl-glucosaminides can be found in Tables 3.6 and 3.7. Compounds 3.35, 2 9 0 3.38,291 3.39,248 and 3.412 4 8'2 9 1 are known. Spectral data reported in this thesis are consistent with the available data. General synthesis of aryl 2-acetamido-2-deoxy-f3-D-glucopyranosides (3.42-3.52) To a stirred solution of the specific aryl 2-acetamido-2-deoxy-3,4,6-tri-(3-acetyl-p-D-glucopyranoside (3.33 to 3.41, Figure 3.21) in anhydrous methanol was added a catalytic amount (1-5 drops) of a 1 M solution of sodium methoxide in methanol. The reaction mixture was allowed to stir for 2 to 4 hours at room temperature. In cases where a precipitate formed this was filtered and washed with methanol. The mother liquor of these reaction and all other reactions were processed in the same manner. Excess Amberlite IR-120 resin (H+) was added and the reaction mixture was filtered. The filtrate was concentrated in vacuo to provide a solid. Recrystallization of all compounds was accomplished using a mixture of methanol - ether - hexanes. Yields of the desired product after one recrystallization ranged from 30 to 75 %. 'H NMR data, melting points and elemental analyses for compounds 3.42 to 3.52 can be found in Tables 3.5, 3.8, and 3.9. Compounds 3.44,290 3.49,291, 3.50248 and 3.52248-291 are known. Spectral data reported in this thesis are consistent with the available data Chapter 3 Mechanism of B-Hexosaminidases 183 T3 C ca Oi <^ o u JC 6J) 'S c o T3 O c o 3 X o o T3 u ca i CM ca a> cj « 3 -S — u ca J S o J S e o -o c 3 <2 JO t> cn 3 3 .5? ° Chapter 3 Mechanism of 3-Hexosaminidases 184 o VO W TD C n3 o Q Z~ Q Q Q Q Z Q Z~ Q Z~ Q Z i n i n <N Z T t . -CN ° T t ^ u O N O O SC °°. S> «-> 2d* & + u 6 U W i n oo u 6 ^ u CN * z ^ i n TH BC x - 3 3 2 m IT) £ u u rt ij a, & ^ * ^ ~ . CN CN Z~ Z 2 o CN" CN ^ CN . . «-> 2 BC a u £»" °° ^  ^ o 2 a CN ^ 5 <-> u 6 « * B U W cn °° od ^ 8 21 S SS CN ^ Tt CN J J ^ O O cn O O O O cn cn o cn cn cn cn c •8 O BC BC BC X X Z BC o Z a) o e cd o 3 tf o z o z <N r-i r-J O O O Z Z Z BC BC BC X Z BC BC BC o z o z o z o z o tf o z •BC BC BC BC BC BC BC 0 0 0 0 0 0 0 < < < < < < < o z BC BC BC BC BC BC E 3 01 cd I o X i 6 CS o a. B o U cn T t i n Cn VO 00 ©\ © * H f ; T t T t <n CN T t cn T t ci T t T t T t ci co o 3 cd H I z T3 C 3 O a £ o U o o I o < z o o Q CN o < o t o < z o 5 Q o < o I o < z o a o < o 1 o < z o 5 5 o o PH O t o a o < o o 5 o $ Cu o < O o < z o 5 o o < o o z o 5 OH o < z o o Q o $ o 5 Q o < z o 0 o < Z o 0 Chapter 3 Mechanism of B-Hexosaminidases 185 o cj 3 u o 60 CN CN « u CH ° T t o CN CJ <D 3 u o m CN i CN CO CN O c n o 4) T3 ^ S « ^ T f m CN CN Q 6 ^ a & B1 « x 6 U W -a U o 3 ^ £ Z J z z ~* CN ^ °° £ 8 "1 ^ W 3 ffi 33 ¥ <3 .« ^ TC £ <-> u 6 u w I CN CN VO CN CN vo VO Z" 00 . .. T t - m „°X u o •-° IT? ^ 2 >o jo SS + u i r i a * 6 U W 5? PJ VO o CN cj ^ vo <* 3 ^ a a a i u u B 13 *f Sv s z z VO vo 2 3 s u u E U PJ T t O N ^ T t T t VO ffi 53" U vo AT ^ ^ § £ W O CN o £ &13 x o z o z o z o z ex-ec X X X X X X X X X X X vc oo Ti-en as Tt o CJ < z o PH u CJ I CJ $ O H CJ o CH Chapter 3 Mechanism of B-Hexosaminidases 186 I i 3 OAc OAc OAc OAc OAc OAc 2.06 s, NAc 2.04 s, NAc 2.05 s, NAc x OAc UAc NAc 2.00 s, NAc 2.03 s, NAc 2.04 s, NAc x OAc OAc NAc c/T Cfl" to" CO CO <N «? to" CO CO CO CO* (A CO <N Cfl Cfl" o o VO Os i n O O . O SO O T t OS aT i^l i n r - ° ° ) CN O CN °) Tt o Tt OS i n O i n OS cfl i n m t— O Os CN rt CN CN CN — ' O N H CN rt CN rt CN rt O CN rt Cfl" to" co CN Cfl" CO c/T CN 00 o OS O oo O CN o i n O vo O CN CN CN CN CN CN SI X C N C N K S 73 C N T t VO C N vo O « ^ 3 S D \£> Tt v> in X C N O S 2 ^ i n os wS S 2 . sc "O i n S3 o\ o „ Tt < z o Q m Tt C N Tt C N C N Tt C N Tt S S s S x t-; r-; c n c n i n OS -o o o OS T t T t T t T t rt vO OS T3 T3 N rt W 2 C N 00 <N C N " Tt S. g rt rt T3 i n T t c n i n o i OS X3 N 2 W rt Os T3 -a T3 T 3 •a -o Tt T t •a -o C N Tt SO <""< C N ^ Tt « rt OS T t T t C N Tt i n -o -a 00 « N C N ' - ' Tt S. rt.rN.r-; i n i n C N 00 ta X) - 1 - > 2 . £ O Os 2 ^ Os i n ° \ rt ^ Tt OS rt T3 •° rt rt T t O o 00 rt _' d c n _f -o -o -o N vo C N „ T3 7 3 T t T t OS N 2 SC 7 3 rt rt T f 00 O o — rt T 3 C N OS -o T3 oo c n os 3 Tt OS T t „ T J T3 •a t-Tt O S -a •a X o o OS T ? O S N 2 SC T3 T t rt o ±3 N N T3 X •a X X T J X T J X T J X T J X OS vq i n rt c n i n VO rt 00 c n i n T t i n c n i n o d i n o d Tt o d Tt 00 rt o d rt o d C N o d i n C N in" C N i n C N i n i n cs i n C N i n C N N N N X X X X T J vq T J o o T3 vq T3 i n C N l > o d CN VO t- ; l > cs T t o o o d fN c n sq 00 CN i n I i n a" z i n a i n X z S m ^ X S o o M " X ' ° s o T3 ° ° ' c ^ t T3 ° ° C N t m 0d a r- < s r- < «n x -i n " z o Q m o < § O-, vb CJ I o o £ o I C H T3 T 3 rt C N T3 T3 o o £ j C N ' _ ' rf ^ sS X rt i n T t c n m" C N Os « Xi N 2 SC 2 os Sci Tt T3 T3 T t OS T3 Tt o 9 cS T3 m c n • so 0 0 CN i n x -CJ 1 = Chapter 3 Mechanism of 3-Hexosaminidases 187 < O >/-> o C N oo o TD T3 N 3 ^ <=> - H CO 1 — 1 -a -a T3 r -ON N N vo r~; i r i C N 3 w 7 3 VO ^ H OV >n 5 TD T3 O T3 i n 3, o cn m cn • ^ T t N TD S VO ^ cn oo >n <H N X vq oo n a" z V O 00 vo PH u a o o s Q TD o C CJ C 3 N X S o o C N TD CJ o cj cd ' td Q >1 TD 3 Tt cn o H J m m rn CJ TD o CJ _3 "3>| Q cat VO T t cn o cj Tp C N 6 TD CJ CJ cd i CN cd cj J3 cj Hz cn od TD TD < < VO r -2 s as < Z S S as as T3 T 3 ai • ai cj < a < z TD TD O VO S S cn T t 00 H H °5 < < r- '-s 'x. Z < o Q cn CN Z TD TD TD O cn S TD T 3 cn T t N N ac HI cn in T3 od ^ r -< Z CJ oo OV TD as so SO N p od 3 1 N HI q od in •» O at $ CN CN 2 < C N Z OV vd >n cn < z m cn Z TD TD O O in o CN TD 00 S CN 3 ^ CN TD 00 cj cn 2 u o o T t cn U P s u a 0H CJ I O H z in o TD oo z TD oo Hz S S Hz Hz CN CN 00 H H cn cn CN CN od H-I Ov as PP H/ \D ai ai < < AR3 PP T-a: a: < < TD AR3 TD 2 < VO CN O cs r i m CN r i oo ai ai < • < cn Ov ai < q ai ai < < q ai < q < t ~ vd CJ I Chapter 3 Mechanism of 3-Hexosaminidases 188 \< 2" o TD OO < Z TD oo cn ON VO cs" t t , o < z o TD tu o c a o o T t TD -a u o o o tu _u ct) cd cd cd a Q TD <rt ci o n Tt ci t/3 t U -O CJ 3 "5b| a t l o tu •a C N 6 TD CJ u cd i CN cd tu 60| C o c tu 4= cd PP CN vo X Ov 3^ cn VD CN vd o CN 'Tt T D ffi T t O V cn _-3 S VO 0 0 cn „-N TD K TD cn O ^ H T t "1 T D ^ OV CN oo i n <~i cj i Q >n cn o < z ov OV cn O m 3 w cn cn ov cn _-oo od TD S TD o d 00 cn m . N TD K VO vO — ; od i n <H o < Q cn CN CJ i Ov cn -°. x>" v£> oo oo T t Ov oo T t cn CN Ov 8 2 oo oo oo Tt OV T3 S Tt ^ H H oo i n <H < z o VO ov TD TD oo T D ^ X cn O 1 £ tN OV vO cn >^o ov T t Tt 0< cn' 5 vo TD TD >n ? o TD • O T t cn o OV H H cn •a Ov H H OO i n « cj •< z o < z VO Ov TD TD VO VO VO i n cn i n oo cn ov T t «n ov cn m T t Ov cn m T t i n oo cn ov N TD W ^ H T t p od i n <*i i o CJ < Chapter 3 Mechanism of 3-Hexosaminidases 189 Z Ov < z VO OV $ VO OV PH u CN 00 T t OH u T3" T3. C OH U o T3 T t T3 <ri T t <h O vd T3 oo IT) I c OH PH u T3 T t f-T c PH -o T3 O T3 T3 oo CN Ov ^ H iri T3 T3 S c n xT cn vo T3 tS cn x T oo CN OV IT) 00 OV cn T3 T3 oo CN CN CN •«3 5 3 OV vq^  w CN CN OV vq -O N uv "2 uv ^ H N X —' S —' •= DC •H OV VO ^ H cn IT) 6 VO cn cn i VO T t 3. c c 1 a 1 T3 i n xf cn iri CN Ov « Xl VD vo cn in w iri cn <r> cn IT) cn cn CN VO cn vo T t •o " ° T t 2 OV T3 •O cn T t oo Ov N cn -T t Ov cn 2 3 T3 T t OV CN 3 K a S ^ w £ 2 oo cn' -H 5 3 T3 7 3 T t T t OV T J 7 3 vo VO • VD 0 0 c n 5 3 T3 S3 06 Tt cn _,-vS oo cn 5 00 CN Ov IT) oo Ov cn • N DC T t d N T3 ^ T3 T t vo d o ~ H N T3 T t r - , O O ^ H T3 ^ T3 cn 8 2 T3 B o e N r-ov oo T t M . TD a CN p OO iri i-i Z OV CN 00 iri <H T3 ^ IT) CN 00 >ri <H in CN 00 ir! <"i c 3 IS o o T t < Z O o < Z C H o OH o CJ < < z o o X OH OH PH CJ < z o OH OH O I T3 CJ O o CS Chapter 3 Mechanism of B-Hexosaminidases 190 ca h loo •a •8 CN LTj O <N Tt u 'co o O _3 "oo Q CO. I >s X o <u o -o & B "5 o S3 i ca CO >s ca co J3 i ca ON _o S i ca H c 3 I o U £ X as as n-} T- vo OS OS ~ < < C N CN CN ^ % % oo >-> •»> < z o Q i n cn < Z •a cn as E S S cn T f 06 rt as as < < < z Q cn C N < z pp B o o ON m i > N N X X cn rt od- od PP AR6 PP AR6 C N m IT) r- OS < i n AR T t oo oo 00 < z T t 00 00 00 < z oo O C N cn C N 7 3 r— C N 3 % % IT) \6 NO f -as < Z C N C N C N C N T3 "O CO r-' < cj z o •< z r-o C N < Z NO as NO T t o\ NO T t as < Z z rt o OS Z Z „ 3 NO ON NO Z o CJ z rt CJ < z a rt CJ < z a CJ < z a < N T3 T 3 T 3 i n i n NO rt rt rt i > t-» "O "O -o T3 OS O rt oo oo 00 rt T t rt rt . rt rt od od od < 7. T3 PP Hz Hz •a "O -a O T t rt OO 00 00 T t C N od rt rt od od od rt fc: CO co cn T t ON Os C<) OS Cl ON NO "O AR -a as < •a AR rt i n CN i n NO CN rt rt as rt rt % l > r-" o Z o rt Chapter 3 Mechanism of 3-Hexosaminidases 191 ON od PH o < z o < Z TD o o CN < Z o CN N N X X CN CN ON ON co TD CC < TD CC < VO ri O CC <* CN < r~ PH O < z o OH 3 s o o T t TD CD o o 3 TD Chapter 3 Mechanism of 0-Hexosaminidases 192 Synthesis of l-(2H}-2-acetamido-2-deoxy-3,4,6-tri-0-benzyl-a/fi-D-glucopyranose ( 3 . 2 6 ) To a stirred solution of 3,4,6-tri-0-benzyl-D-glucono-l,5-lactone ( 3 . 2 5 ) 2 8 7 (9.7 g, 28 mmol) in anhydrous THF (130 mL) was added a solution of sodium borodeuteride (0.22 g, 5.2 mmol) in 0.99 mL of D2O. The reaction mixture was stirred at room temperature for 2 hours, after which time the reaction was judged complete by TLC analysis. The resulting solution was concentrated in vacuo to a pale yellow syrup and dichloromethane (200 mL) was added. After the syrup had dissolved the organic layer was washed with water (2 x 100 mL) and saturated sodium chloride solution (1 x 100 mL). The organic layer was dried over MgS04, concentrated in vacuo to a syrup, and crystallized from ethyl acetate - hexanes to provide a mixture of the title compounds as a white crystalline solid (82% yield, 8.0 g, 23 mmol). ! H NMR (400 MHz, CDCI3) 7.35-7.15 (15 H, m, Ph), 5.37 (1 H, d, 7 N H,H2 = 8.8 Hz, NH), 4.86-4.76 (3 H, m), 4.68-4.48 (5 H, m), 4.11 (1 H, dd, / H 2 , H 3 = 10.3, H-2), 4.03 (1 H, ddd, 7 H 5 ,H4 = 9.7 Hz, 7 H 5 , H 6 = 4.8 Hz, 7H5,H6a = 2.8 Hz, H-5), 3.80 (2 H, m), 3.68-3.60 (3 H, m), 1.81 (3 H, s, NAc) Synthesis of l-2H-2-acetamido-2-deoxy-a/fi-D-glucopyranose (3.27) The atmosphere of a solution of l-2H-2-acetamido-2-deoxy-3,4,6-tri-0-benzyl-P-D-glucopyranose (1.2 g, 2.43 mmol) in a mixture of ethyl acetate and methanol (10 mL each) was replaced with nitrogen and a catalytic amount of palladium on carbon was added. The atmosphere above the reaction was replaced with hydrogen and the mixture was stirred for 16 hours at which time the reaction was judged by TLC analysis to be complete. The reaction mixture was filtered through glass fiber filter paper and concentrated in vacuo to yield a white powder (475 mg, 2.14 mmol, 88 %). *H NMR (400 MHz, D20) 3.80-3.52 (4.7 H), 3.46-3.32 (1.3 H, m), 1.93 (3 H, s, NAc). Chapter 3 Mechanism of 0-Hexosaminidases 193 3.7.2.2 Enzyme Kinetics 3.7.2.2.1 Enzyme kinetics with ExoII All kinetic studies involving ExoII were performed in 50 mM sodium phosphate, 200 mM sodium chloride, pH 7.00, containing 0.1% bovine serum albumin. For each substrate a continuous spectrophotometric assay based on the rate of release of the phenolic leaving group upon hydrolysis of the chromogenic substrate was used. The resulting absorbance change was measured using a Pye-Unicam 8700 UV/VIS spectrophotometer equipped with a circulating water bath set at 23.5 C. The wavelengths monitored varied for each substrate and are listed in Table 3.10. Michaelis-Menten parameters for the hydrolysis of all substrates by ExoII were determined by directly fitting the intial rate data to the Michaelis-Menten equation using GraFit version 3.0.198 Km and &cat values for the hydrolysis of 5FIdNAcF by ExoII were determined by monitoring the release of fluoride using a 9606BN Ionplus Orion fluoride ion electrode interfaced to a pH/ion selective meter (Fischer Scientific) at 25 °C. Stock enzyme (10 pL) was added to glass cells containing various concentrations of 5FIdNAcF in the above-mentioned buffer to give a final volume of 250 p,L. Reaction cells were preequilibrated in a water bath to 25 °C. It is known that 5-fluoroglycosyl fluorides, after being hydrolyzed to give the hemiacetal, rapidly lose another fluoride ion, resulting in a net release of two fluoride ions for each catalytic event. Initial rates, which were measured over 5 min, were therefore corrected by dividing by a factor of 2. Calculation of kinetic parameters for this substrate was carried out by directly fitting the intial rate data to the Michaelis-Menten equation using GraFit version 3.0.198 The concentrations of enzyme and substrate used in each assay can be found in Appendix 1. 3.7.2.2.2 Enzyme Kinetics with SpHex All kinetic studies involving SpHex were performed in 25/25 mM sodium citrate/sodium phosphate, 150 mM sodium chloride, pH 4.50. For each substrate a continuous spectrophotometric assay based on the rate of release of the phenolic leaving Chapter 3 Mechanism of ^ -Hexosaminidases 194 group upon hydrolysis of the chromogenic substrate was used. The resulting absorbance change was measured using a Pye-Unicam 8700 UV/VIS spectrophotometer equipped with a circulating water bath at 37.0 °C. The wavelengths monitored varied for each substrate and are listed in Appendix 1. Michaelis-Menten parameters for the hydrolysis of all substrates by SpHex were determined by directly fitting the intial rate data to the Michaelis-Menten equation using GraFit version 3.0.198 The concentration of enzyme and substrate used in each assay can be found in Appendix 1. 3.7.2.2.3 Kinetic Isotope Effects Isotope effects using l-{2H}-pNPGlcNAc were determined in two different ways due to limitations arising from the Km values with each enzyme and the limited solubility of the substrate. For ExoII the relatively high Km value (Km = 0.73 mM) prohibited measurement of the a D (V) isotope effect due to the limited solubility of the substrate. Thus for this ExoII the a D (V/K) isotope effect was determined by continuously monitoring the depletion of a low concentration of substrate (1/18 x Km) in the reaction at 400 nm. For SpHex saturation conditions were readily obtained (Km = 50 uM) and therefore a D (V) isotope effects were determined. An accurate determination of the a" D(V/K) isotope effect for SpHex was precluded by the low concentration of substrate required (0.1 to 0.2 x Km) for these experiments and the consequent poor signal to noise in the progress curves. The standard assay conditions were used as described above except that for the experiments involving ExoII the concentration of l-{2H}-pNPGlcNAc was 70 pM and for SpHex a concentration of 1.2 mM was used. The reaction was initiated by the addition of an aliquot (20 U.L) of thermally equilibrated substrate. Initial rates or second order rate constants were measured alternately for protio and deuterio samples until at least 8 rates for each had been determined. Average rates or rate constants were then calculated for the protio and deutero substrates and the ratio was taken to yield the isotope effect. 'H NMR analysis of the sample of l-{2H}-pNPGlcNAc used showed that the extent of isotopic incorporation was > 95%. Chapter 3 Mechanism of 0-Hexosaminidases 3.7.2.2.4 Fluorescence Measurements with Jack Bean ^-Hexosaminidase 195 The time course of release of 4-methylumbelliferone from MuTAG ( 3 . 1 8 ) by jack bean NAGase was followed by establishing a series of reaction mixtures, each containing NAGase (2.4 |ig/mL) and 3 . 1 8 (0.65 mM) in 420 pL of 50 mM citrate buffer containing 100 mM NaCl and 0.1% BSA at pH 5.0. These mixtures were quenched at various incubation times by adding 1.26 mL of 0.2 M glycine buffer, pH 10.65. The fluorescence due to 4-methylumbelliferone was then measured at 450 nm. A Perkin Elmer LS-5B fluorometer was used for all measurements with the excitation (368 nm) slit width set at 6 nm and the emmision (450 nm) slit width set at 7 nm. A standard curve (8 points, r2 = 0.999) was constructed to determine the concentration of 4-methylumbelliferone. 3.7.2.2.5 Determination of the Delta Absorption Coefficient (Aen) for Aryl Glycoside Substrates A series of three or four solutions containing varying concentrations of the intact glycoside of interest were made up in quartz cells in the assay buffer for SpHex (vide supra). The absorption of these solutions at the appropriate wavelength was measured and the data fitted by linear regression. In all cases the data revealed excellent linear correlations (r > 0.99) providing the extinction coefficient of the intact glycoside. SpHex (10 pL, 1.2 mg/mL) was added and the reaction allowed to proceed for six hours after which time the absorbance of each reaction mixture was again measured at the appropriate wavelength. The reaction was then allowed to proceed overnight and the absorbance of each solution was again recorded. These absorbance readings were in close agreement with the readings taken the previous day indicating the reactions had been complete after 6 hours. These data was fitted by linear regression using the program Grafit198 and in all cases the data revealed excellent linear correlations (r = 0.97) furnishing the extinction coefficient of the liberated phenol. The absorbace of the enzyme at each wavelength was also recorded. The extinction coefficient of the intact Chapter 3 Mechanism of 0-Hexosaminidases 196 aryl glycoside of interest was subtracted from the extinction coefficient of the corresponding phenol and of the enzyme to provide the delta absorption coefficient (Aen). 3.7.2.2.6 Labeling and Proteolysis. Labeling of V. furnisii p-N-acetylglucosaminidase (8.2 mg/mL) was accomplished by incubating the enzyme with 5FIdNAcF (8.2 mM) for 5 min in 50 mM sodium phosphate, 200 mM sodium chloride, pH 7.00. This sample was then analyzed immediately by injecting the mixture onto a reverse-phase column (PLRP-S, 1 x 50 mm) equilibrated with solvent A [solvent A: 0.05% trifluoroacetic acid (TFA)/2% acetonitrile in water] on an Ultrafast Microprotein Analyzer (Michrom BioResources Inc., Pleasanton, CA). Elution of the enzyme was accomplished using solvent A at a flow rate of 50 u\L/min. Proteolytic digestion of the enzyme was performed by mixing the labeled enzyme (25 pL of 8.2 mg/mL) with 20 pL of 2.1 M sodium phosphate, pH 1.7, 135 pL of H2O, and 20 pL of 1 mg/mL pepsin in 200 mM sodium phosphate, pH 2.0. This sample and a control in which the enzyme was not exposed to the inhibitor were incubated at 25 °C for 2 h. At 0.25, 0.5, 1.0, and 2.0 h, aliquots (50 pL) of the sample were frozen over CCVacetone. Analysis of these samples by ESMS revealed that the 2 h digest ensured complete digestion of the enzyme and generated peptides of a size suitable for sequencing by MS/MS. 3.7.2.2.7 ESMS Analysis of the Proteolytic Digest. Mass spectra were recorded on a PE-Sciex API 300 triple-quadrupole mass spectrometer (Sciex, Thornhill, Ontario, Canada) equipped with an Ionspray ion source. Peptides were separated by reverse-phase HPLC on an Ultrafast Microprotein Analyzer (Michrom BioResources Inc.) directly interfaced with the mass spectrometer. In each of the MS experiments, the proteolytic digest was loaded onto a Qg column (Reliasil, 1 x 150 mm) equilibrated with solvent A [solvent A: 0.05% trifluoroacetic acid (TFA)/2% Chapter 3 Mechanism of 0-Hexosaminidases 197 acetonitrile in water]. Elution of the peptides was accomplished using a gradient (0-60%) of solvent B over 60 min followed by 100% solvent B over 2 min (solvent B: 0.045% TFA/80% acetonitrile in water). Solvents were pumped at a constant flow rate of 50 pL/min. Spectra were obtained in the single-quadrupole scan mode (LC/MS), the tandem MS neutral-loss scan mode, or the tandem MS product-ion scan mode (MS/MS). In the single-quadrupole mode (LC/MS), the quadrupole mass analyzer was scanned over a mass-to-charge ratio (m/z) range of 300-2200 Da with a step size of 0.5 Da and a dwell time of 1.5 ms per step. The ion source voltage (ISV) was set at 5.5 kV, and the orifice energy (OR) was 45 V. In the neutral-loss scanning mode, MS/MS spectra were obtained by searching for the mass loss of m/z = 222, corresponding to the loss of the 5FIdNAc label from a peptide in the singly charged state. In the tandem MS daughter-ion scan mode, the spectrum was obtained by selectively introducing the parent ion (m/z = 1248) from the first quadrupole (QI) into the collision cell (Q2) and observing the product ions in the third quadrupole (Q3). Thus, QI was locked on m/z 1248; the Q3 scan range was 50-1260; the step size was 0.5; the dwell time was 1 ms; ISV was 5 kV; OR was 45 V; Q0 = -10; IQ2 = -48. 3.7.2.2.8 Aminolysis of the Labeled Enzyme Digest. Concentrated ammonium hydroxide (5 u\L) was added to an aliquot of the 2 h digest mixture (10 pL, 1.0 mg/mL). This sample was incubated for 15 min at 50 °C, acidified with 50% TFA, and analyzed by LC/MS as described above. Chapter 4 Mechanism of Hen Ess White Lysozyme 198 4 Mechanism of Hen Egg White Lysozyme 4.1 Background of HEWL Lysozyme was discovered pseudo-serendipitously by Alexander Fleming during his systematic search for antibacterial compounds.292 Apparently, while working one day in 1921, Fleming had a cold and intentionally let some nasal secretion drop from his nose onto a plate of bacteria. He then cultured the plate to see what the effect of the mucus on the growth of the bacteria would be. To his great surprise, the following morning an area around the drop of mucus was clear of bacteria, indicating that his nasal secretions were bactericidal. In this way Fleming had discovered the natural defense mechanism of animals against certain bacteria. All bacteria form a very durable and resilient coating known as the cell wall. Bacteria synthesize their own cell wall from oligosaccharides that are extensively cross-linked by short peptides (Figure 4.1). This peptidoglycan layer acts to counter the severe osmotic pressure that forms across the bacterial membrane as a result of the high ionic strength of the contents of the cell, thereby preventing the cell from rupturing.221 ETC—O->—ETC Glu-D L-Ala ETC—O-O—ETC Figure 4.1 Simplified structure of a fragment of the bacterial cell wall of E. coli. Chapter 4 Mechanism of Hen Ess White Lysozyme 199 The bacteriocidal activity of lysozyme stems from its ability to disrupt the structure of the cell wall by hydrolyzing the oligosaccharide chains.293 Once the cell wall has been compromised the membranes of the denuded cells burst open and the bacterium dies. Not all bacteria are susceptible to the action of lysozyme. The peptidoglycan of Gram positive bacteria is exposed to the environment making these bacteria vulnerable to lysozyme. Gram negative bacteria, however, have an additional membrane surrounding the peptidoglycan that protects it from lysozyme. It is likely that many bacteria that are usually harmless are only rendered thus by the action of lysozyme. After the discovery of lysozyme by Fleming, other similar enzymes were found in every other animal and plant investigated. Indeed, all animal secretions, from tears to mucus, contain lysozyme, which acts to protect the animal from colonization by Gram positive bacteria. Despite considerable experimentation all known lysozymes have been found to be unsuitable for the treatment of bacterial infections. From these rather inconspicuous beginnings, research into understanding the properties of hen egg white lysozyme flared brightly for a period of 10 or 15 years and is still the topic of a considerable number of papers. Owing to the relatively small size of the hen egg white lysozyme molecule and its ready availability many experimentalists, seeking to understand the structure of proteins and function of enzymes, have opted to investigate HEWL. The advancing technique of protein sequencing was applied to HEWL independently in the laboratories of Jolles294 and Canfield295 resulting in the elucidation of the primary sequence of its 129 amino acid peptide chain. Over time it has become apparent that there are three distinct classes of lysozyme that are only distantly related; phage lysozymes are (3-inverting glycosidases from family 24 of glycoside hydrolases, G-type lysozymes are P-inverting enzymes from family 23, and C-type lysozymes are P-retaining enzymes from family 22 that include as one of their members HEWL. The sequencing of several C-type (chicken type) lysozymes has also revealed that Asp and Glu residues are invariably found in a positions corresponding to the Asp52 and Glu35 of HEWL. The high resolution (2.0 A) tertiary structure of HEWL was solved by Blake et al not long after (1965) the primary sequence had been determined.16 At this Chapter 4 Mechanism of Hen Ess White Lysozyme 200 time little was known about the mechanism of action of HEWL or indeed, any glycosidase, but this was soon to change. Soon after the initial structure of HEWL was solved, Phillips and coworkers obtained a high resolution X-ray crystal structure of HEWL in complex with chitotriose (Figure 4.2).17 ' ON V . O H Y . O H . V O H A / - 4 ( A ) -3(B) -2(C) -1(D) + 1 ( E ) +2 (F) Figure 4.2 Schematic of chitotriose bound in the active site of HEWL. No effort has been made to accurately represent the geometry of the active site or the relative orientations of the sugar residues. Now, in more modern times, it is difficult to imagine the excitement that must have resulted upon the unveiling of this structure. Here was the first glimpse at the catalytic machinery of any enzyme. Given the chemical information that chitotriose was a very poor substrate for H E W L 2 9 6 and the absence of any change in the diffraction pattern over several weeks, Phillips believed that this complex must be a non-productive species incapable of leading to catalysis. In what must be considered prescient; Phillips modeled three more residues extending off of the reducing end of the bound chitotriose molecule (Figure 4.3). On the basis of this relatively crude model Blake et al suggested in their 1967 paper (here follows a direct quote):17 (1) that Asp52 carries a negative charge which promotes the formation of a carbonium ion at C(l) of residue D and stabilizes it when formed; (2) that distortion of residue D, from the chair conformation into the very half-chair conformation in which a carbonium ion at C(l) would also be stabilized by sharing Chapter 4 Mechanism of Hen Ess White Lysozyme 201 (3) its charge with the ring oxygen, also contributes to the formation of this carbenium ion and to the consequent weakening of the C(1)0 bond; and that Glu35 can act as a proton donor facilitating the formation of a hydroxyl group with the bridge oxygen atom and the release of residues E and F. H O • O H -4 (A) -3 (B) -2 (C) -1(D) +1 (E) +2(F) Figure 4.3 Schematic representation of the model based on the non-productive complex of chitotriose with HEWL. No effort has been made to accurately represent the geometry of the active site, only the relative conformations of the pyranose rings and the position of the active site residues. Asp52 is positioned below the sugar residue in the -1 (D) site and Glu35 is positioned above the pyranose ring of the same sugar. By the time of the Phillips proposal it was known that the bacterial cell wall was composed of alternating N-acetyl glucosamine (NAG) and N-acetyl muramic acid (NAM) residues. Furthermore, Sharon had determined that a cell wall tetrasaccharide composed of NAG-NAM-NAG-NAM was cleaved to yield two identical disaccharide products, namely NAG-NAM. 2 9 7 On the basis of these facts in conjunction with the X-ray data it was further possible to propose that the B, D, and F subsites could accommodate an N-acetyl muramic acid residue thus explaining the exclusive cleavage of the glycosidic bond between a NAM and NAG residue. This remarkably detailed hypothesis was forwarded in the absence of any experimental evidence pointing to the identity of the catalytic residues and any detailed solution studies. Immediately, numerous group began testing various aspects of the Phillips mechanism. Lin and Koshland found that chemical modification of Asp52 results in complete inactivation of the enzyme.298 As mentioned in the introduction, Asp52 was also labeled using an epoxypropyl glycoside.299 Beautiful chemical mutagenesis work exploiting the Chapter 4 Mechanism of Hen Egg White Lysozyme 202 complete specificity of the 2',3'-epoxypropyl chitobioside (1.13) affinity label for Asp52 also confirmed the great importance of this residue. In that study Eshdat et al were able to reduce the esterified Asp52 to an inactive enzyme containing homoserine at position 52 (Scheme 4.1). HO^ °Y° ASP52 DTT NaBH4 ^OH HOMOSER52 INaB 4 A. pH 8.5 [ Scheme 4.1 Strategy for the chemical mutagenesis of Asp52 to homeserine. An extremely careful and early application of site directed mutagenesis also found that Asp52 was important for catalysis. The activity of the Asp52Asn mutant enzyme on defined substrates such as chitopentaose or 4-methylumbelliferyl chitotrioside (Figure 4.4) was greatly reduced, to less than 0.1% and 2% respectively of the rate of the wild-type enzyme.163 Figure 4.4 Chitopentaose (at left) and 4-methylumbelliferyl chitotrioside (at right). Interestingly however, this mutant exhibited considerable (5.5 ± 2.5%) activity in the classical cell wall assay using Micrococcus luteus cell walls. The same study also revealed that mutagenic substitution of the proposed catalytic acid/base residue, Glu35, with Gin completely obliterated all activity. Other mutagenesis studies have also confirmed the importance of these residues in catalysis.63-300'301 Neutron diffraction studies provided an elegant means to ascertain which residue functioned as the catalytic acid/base. These coworkers found that in the free enzyme Glu35 was protonated, strongly suggesting that this group and not Asp52 was the catalytic acid/base residue.302 Chapter 4 Mechanism of Hen Ess White Lysozyme 203 Together, these studies have provided conclusive evidence supporting the original assignments of these residues by Phillips. Another important aspect of the Phillips proposal involved the prediction that the NAM-NAG glycosidic linkage was cleaved with retention of configuration. The first studies devised to determine the stereochemical outcome of the reaction stemmed from investigations into the transfer products generated by HEWL. After cleaving the glycosidic bond and forming the enzyme intermediate, the aglycone diffuses away from the active site and an acceptor molecule, most commonly water but occasionally another saccharide moiety, diffuses into the active site. The properly positioned hydroxyl of the acceptor sugar moiety may then intercept the intermediate, resulting in a transglycosylation reaction (Figure 4.5). ASP52 TRANSGLYCOSYLATION Figure 4.5 Transglycosylation reaction catalyzed by HEWL showing the Phillips mechanism. The donor is chitotetraose and the acceptor is cellobiose. Chapter 4 Mechanism of Hen Egg White Lysozyme 204 From product analysis of the HEWL catalyzed transglycosylation reaction using chitotetraose as a donor and (14C) N-acetyl-glucosamine as an acceptor Rupley and Gates found that the transglycosylation products comigrated during charcoal-Celite chromatography.87 A more detailed analysis using methanol as an acceptor confirmed that the stereochemistry at the anomeric center was retained and furthermore that the reaction proceeded with at least 99.7% stereoselectivity.20 The substrate specificity of HEWL has also been rigorously investigated. As discussed earlier, the groove shaped binding site observed by Blake et al is consistent with an endo-mode of action for HEWL. 1 6 The crystal structure of HEWL in complex with chitotriose reveals that the 3-Olactyl side chain of a NAM residue cannot be accommodated in the -4(A), -2(C), or +1(E) subsites. With this information and recalling that Sharon found that the NAG-NAM-NAG-NAM tetrasaccharide is split symmetrically between the NAM and NAG residues one can conclude that the cleavage site cannot be between sites -4(A) to -3(B), sites -2(C) to -1(D), or sites +1(E) to +2(F). Furthermore, it is known that the trisaccharide complex observed by Blake et al is non-productive and thus the cleavage site cannot be between sites -3(B) and -2(C). Only one alternative remains; cleavage must occur between the -1(D) and +1(E) subsites, exactly as predicted by Phillips. In order to examine the relative importance of the binding sites toward catalysis Rupley and Gates studied the relative rates of hydrolysis of a series of chitooligomers (NAGn, where n = 2 to 6).87 They found that the hexamer was cleaved rapidly and that shorter chain length oligomers were cleaved correspondingly slower. A summary of selected kinetic data is shown below in Table 4.1. They also found that the lactyl side chains were not essential for catalysis, contributing only a two-fold enhancement in the first order rate constant for cleavage of the hexasaccharide. Also interesting was the cleavage pattern of the hexamer; it was cleaved primarily at one site to generate a dimer and a tetramer, directly confirming that the site of cleavage is between the -1(D) and +1(E) subsites. These workers also confirmed that it is the C-l to 0-1 bond that is cleaved by HEWL. By conducting the HEWL catalyzed hydrolysis reaction of chitotriose in 180-water they found incorporation of the heavy oxygen only in the newly liberated hemi-acetal at the reducing terminus.221 Chapter 4 Mechanism of Hen Ess White Lysozyme 205 Table 4.1 Cleavage data for selected oligosaccharides." Saccharide ^cat ( s . ) (NAG)3 8.3 x 10 ,-6 (NAG)4 (NAG)5 (NAG)6 6.6 x 10' 1-5 0.033 0.25 (NAG-NAM)3 0.5 Table adapted from Reference111. Most of the features of the Phillips mechanism including the substrate binding sites, the stereochemical outcome of the reaction, the identity of the catalytic residues, and the site of bond cleavage have thus been validated. Two of the most contentious features of the Phillips mechanism that remain unresolved are substrate distortion and the nature of the intermediate. Good evidence for substrate distortion was obtained early on by Chipman and Sharon during their studies on the free energies of association of various oligosaccharides to HEWL. They found a positive binding term (3 to 6 kcal/mol) associated with binding of NAM to the -1(D) subsite of HEWL that they explained by proposing that this unfavorable binding term arose from the postulated ground state distortion of the pyranose ring.3 0 3 Although this interpretation is in complete agreement with the substrate distortion proposed by Phillips it does not constitute a proof. Indeed, others dispute the role of distortion in catalysis. Using energy minimization calculations on the binding of a hexasaccharide to the active site of HEWL Levitt and Warshel have argued that the enzyme is too flexible to physically distort the substrate.23 Similarly, Pincus and Scheraga have also concluded that an oligosaccharide can bind without distortion and further rule out a mechanism involving distortion.304-305 As pointed out in the introduction, Post and Karplus have suggested an alternative endocyclic ring opening mechanism on the basis of molecular dynamics calculations on the behaviour of chitohexaose and HEWL for 55 ps.14 The Post and Karplus mechanism has been ruled out and Strynadka and James have raised significant criticisms concerning the minimization calculations of the previous two groups.21 They point out that those calculations also resulted in several other conformations, some of which involve Chapter 4 Mechanism of Hen Egg White Lysozyme 206 distortion, that were similar in energy. They also point out that given the limitations in the calculations and the approximations used, accurately discerning the true conformation among the many conformers of similar energy is unlikely. Another criticism against substrate distortion has been leveled on the basis of binding free energies of oligosaccharides. Schindler et al suggest that it is the 3-0-lactyl group of the pyranose sugar bound in site -1(D) of HEWL that contributes 2.9 kcal/mol to an unfavorable binding energy term.306 If the distortion contributes to catalysis it is difficult to reconcile the result of Schindler et al with the apparent lack of importance of the lactyl groups in catalysis (vide supra, Table 4.1). More tangible evidence for distortion in the catalytic mechanism of HEWL can be found in the work of Strynadka and James. They observed, using X-ray diffraction techniques, a product complex of NAM-NAG-NAM bound in sites -3(B), -2(C), and -1(D) of HEWL (Figure 4.6).21 -4(A) -3(B) -2(C) -1(D) +1(E) +2 (F) Figure 4.6 NAM-NAG-NAM bound to the active site of HEWL. No attempt is made to illustrate the relative orientation of the sugar residues. Apparently, the NAM residue bound to the -1(D) subsite is distorted from the low energy 4 C i conformation into a sofa conformation. The high temperature factors associated with the ring atoms of the pyranose unit (39 A2) relative to the enzyme (15.4 A 2), however, make precise interpretation of the distorted conformation of this sugar moiety difficult.21 Indeed, an earlier interpretation of the same data, but unrefined and at lower resolution, concluded that there was no distortion at subsite -1(D).74 Another group have observed oligosaccharide product complexes with a NAG residue bound to subsite -1(D) in the 4 Ci conformation but oddly, as a mixture of a- and (3-hemiacetals.307 These complexes also Chapter 4 Mechanism of Hen Ess White Lysozyme 207 reveal that the sugar units are bound in a slightly different position than is NAM-NAG-NAM strongly suggesting that they may not represent the product complex initially formed after the last chemical step (deglycosylation). Despite these concerns these structures have also been used to argue against substrate distortion in HEWL. Clearly the detailed interpretation of crystallographic data is a delicate and complex matter. As discussed in the introduction, the role of substrate distortion in catalysis by |3-retaining glycosidases remains to be further validated and accepted as scientific fact although there is certainly strong evidence that such distortion does occur in all cases yet observed.61-76-77 Regarding the role of substrate distortion in HEWL, the best and most careful studies point to a role for distortion in the mechanism and are also supported by the compelling studies of other functionally related enzymes. Despite the continued debate this author tentatively submits that substrate distortion is involved in the mechanism of HEWL but that convincing evidence is still required. If, as we expect, this distortion is eventually proved, Phillips would posthumously deserve hearty congratulations for his predictions. The other flashpoint of the Phillips proposal centers on the nature of the reaction intermediate. Attempts to support the existence of an oxocarbenium ion intermediate as proposed by Phillips have focused on the measurement of aD-KIE's. Early studies have found that the aD-KIE associated with the HEWL catalyzed hydrolysis of defined synthetic substrates is quite large, with knlkD = 1.13.308 A lovely study, by Smith et al, using chitotriose labeled with tritium at C-l of every sugar unit (Figure 4.7) revealed a kinetic isotope effect (ku/kr) of 1.19, which is equivalent to a deuterium kinetic isotope effect (ku/kv) of 1.14.309 Rosenberg and Kirsch have also measured a large primary 1R isotope effect during HEWL catalyzed hydrolysis of an aryl glycoside with O incorporated in the glycosidic linkage.310 Figure 4.7 Tritium enriched chitotriose. Chapter 4 Mechanism of Hen Egg White Lysozyme 208 All of these studies, however, have used substrates for which the first chemical step, the cleavage of the glycosidic bond, is rate determining. Therefore, they only provide evidence pointing to the nature of the first transition state along the reaction coordinate. Regardless of this hmitation, these isotope effects have often been used to argue that the intermediate is a sp2 hybridized oxocarbenium ion. In fact, because these isotope effects report only on the nature of the first transition state these results say only that the transition state is oxocarbenium ion-like and say little about the intermediate. Indeed, Smith et al claim that:309 There still exists the possibility of post-rate-determining collapse of the carbonium (oxocarbenium) ion to a covalent intermediate, either via attack by the anion Asp52 or by the N-acetyl group to give the oxazoline intermediate. In fact, the observed half-life of the intermediate, which must be long enough to allow the observed transfer reaction to another saccharide, would suggest some such stabilization. Crystallographers, including Phillips, have argued that the distance (~3 A) between C-l and Asp52 is too great to bridge and that this residue is too tightly held to approach C-l more closely than 2.3 A . 1 7 - 2 1 The distance that is observed, however, is very close to the van der Waals limit (-2.8 A) between these two groups. Given our relatively limited understanding of catalysis and the nature of protein folds and stability, let alone the complex energetics involved in small rearrangements within proteins the use of this negative evidence to preclude the possibility of a covalent intermediate seems somewhat unfounded. This over aggressive interpretation of the crystallographic data is perhaps understandable in view of the nimbus of great respectability borne by the Phillips proposal.17 The great sway of the Phillips proposal is well illustrated by the understanding of the catalytic mechanism of retaining (3-glycosidases by the scientific community. Despite the absence of any positive evidence for an ion-pair intermediate and a few dissenting voices, all of these enzymes were presumed to catalyze the hydrolysis of glycosides by a two step mechanism involving an ion pair intermediate. Indeed, the scientific literature is full of papers, reviews, and textbooks depicting ion-pairs as the intermediate in the catalytic mechanism of these enzymes. A seminal paper by Sinnott and Souchard,45 Chapter 4 Mechanism of Hen Egg White Lysozyme 209 however, incontestably revealed that a covalent intermediate operated in the mechanism of the p-galactosidase from Escherichia coli. Due to the sway held by the ion-pair hypothesis the text was phrased very delicately:45 A scheme consisting of a conformation change, liberation of a galactopyranosyl cation in an intimate ion-pair, non-productive but preferential collapse of the ion-pair to a covalent species and reaction of the galactosyl enzyme through the ion-paired form is proposed. The key experiment involved determining the a-deuterium kinetic isotope effect using a substrate for which the breakdown of the intermediate was rate determining and the result was unequivocal; the kinetic isotope effect was significantly greater than one. Indeed, the prominence of the Phillips mechanism was such that Sinnott, many years later, claimed that it was not for several weeks after the first experiment that he was convinced that Souchard had not mixed up the samples of protiated and deuterated substrate.311 Regardless of the initial uncertainty, this convincing evidence pointing to the nature of any glycosyl-enzyme intermediate was soon followed by several other reports involving a-deuterium KIE's that confirmed a covalent intermediate operated in the mechanism of several other retaining glycosidases. Despite the mounting evidence suggesting that covalent intermediates operate in the mechanism of other retaining glycosidases this evidence was largely overlooked and it had little effect on the Phillips hypothesis as espoused in the textbooks. This is in part because of the laborious work required to prepare defined substrates, and also because the best efforts of many notable chemists failed to provide a substrate for which the rate determining step was the breakdown of the intermediate.24 Thus no evidence pointing to either a covalent or an ion-pair intermediate can be found and the Phillips mechanism perseveres. 4.2 Objectives of this Work As discussed above, based on the available literature the nature of the intermediate is unclear: does electrostatic catalysis occur to generate a stabilized ion-pair intermediate or does nucleophilic catalysis operate to yield a covalent glycosyl-enzyme intermediate? Chapter 4 Mechanism of Hen Egg White Lysozyme 210 A priori it is impossible to be certain but by inductive reasoning we can conjecture that the intermediate should be a covalent intermediate as has been found for other glycosidases. Indeed, it would make sense that enzymes catalyzing very similar reactions would obey similar physical principles and thus use similar catalytic mechanisms. Many eminent researchers, however, have argued against a covalent intermediate and their reasoned arguments can not simply be dismissed. Therefore the research contained within this chapter will center on addressing their objections by seeking evidence for a covalent glycosyl-enzyme intermediate in the catalytic cycle of HEWL. If such an intermediate can be detected this would bring the mechanism of H E W L in line with all other well investigated (3-retaining glycosidases and allow the formulation of a general theory for catalysis by all such enzymes. 4.3 Hydrolysis of a 2-fluoro-glycoside H E W L does not only catalyze the hydrolysis of oligosaccharides but is also a very efficient transglycosylase. As discussed above, the intermediate, whatever it may be, partitions favorably to oligosaccharide acceptors. A considerable number of researchers have used this transglycosylating ability of HEWL to generate, on a preparative scale, several complex oligosaccharides. Others have exploited this phenomenon to study catalysis by HEWL. In this manner Raftery and Rand-Meir studied the importance of the 2-acetamido group of the N A G residue bound in subsite -1(D). By incubating chitotetraose with para-nitrophenyl (3-D-glycopyranoside (pNPGlc), para-nitrophenyl 2-acetamido-2-deoxy-P-D-glycopyranoside (pNPGlcNAc), or /rara-nitrophenyl 2-deoxy-P-D-glycopyranoside (pNP2dGlc) they generated, in situ, oligosaccharide substrates bearing the aryl glycoside at the non-reducing terminus (Scheme 4.2). Chapter 4 Mechanism of Hen Egg White Lysozyme 211 Scheme 4.2 Transglycosylation reaction catalyzed by HEWL generates complex aryl oligosaccharides. Other oligosaccharides are generated having one less or one more NAG residue at the non-reducing terminus but are not shown here for clarity. These products are, in turn, hydrolyzed by H E W L to liberate the chromophore p-nitrophenolate. In this way they established that the products resulting from transfer to pNPGlc or pNP2dGlc function as substrates that hydrolyze with similar efficiency as do transfer products to pNPGlcNAc. The relative rates of cleavage defined in this way were found to be 1 to 16 to 2, respectively.312 This study is, of course, fraught with possible sources of error including selectivity in the transfer reaction affecting the concentration of the aryl oligosaccharide substrate. To strengthen this preliminary investigation Rand-Meir, Dahlquist, and Raftery prepared p-nitrophenyl 2-acetarnido-2-deoxy-P-D-glucopyranosyl-(l-4)-(3-D-glucopyranoside (NAG-Glc-pNP), p-nitrophenyl 2-acetamido-2-deoxy-P-D-glucopyranosyl- (l-4)-2-deoxy-P-D-glucopyranoside (NAG-2dGlc-pNP), and p-nitrophenyl 2-acetamido-2-deoxy-glucopyranosyl-P-D-( 1 -4)-2-acetamido-2-deoxy-p-D-glucopyranoside (NAG-NAG-pNP) (Table 4.2). Using these defined substrates they determined the Michaelis-Menten parameters for the hydrolysis of these substrates by H E W L (Table 4.2). 1 7 0 Interestingly, the kinetic parameters for all three of these substrates are comparable. The similar magnitude of the second order rate constants (Vmax/^ m) 1 S particularly important as this parameter is insensitive to the non-productive binding that often plagues detailed comparison of the kinetic data obtained in s