Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mechanism of glycoside hydrolase family 31 : mechanistic plasticity of glycosidic bond cleavage Lee, Seung Seo 2004

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

Item Metadata

Download

Media
831-ubc_2004-902188.pdf [ 11.68MB ]
Metadata
JSON: 831-1.0061201.json
JSON-LD: 831-1.0061201-ld.json
RDF/XML (Pretty): 831-1.0061201-rdf.xml
RDF/JSON: 831-1.0061201-rdf.json
Turtle: 831-1.0061201-turtle.txt
N-Triples: 831-1.0061201-rdf-ntriples.txt
Original Record: 831-1.0061201-source.json
Full Text
831-1.0061201-fulltext.txt
Citation
831-1.0061201.ris

Full Text

MECHANISM OF GLYCOSIDE HYDROLASE FAMILY 31 MECHANISTIC PLASTICITY OF GLYCOSIDIC BOND CLEAVAGE By SEUNG SEO LEE B.Sc, Seoul National University, 1991 M.Sc, Seoul National University, 1993 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 April 2004 © Seung Seo Lee, 2004 Abstract Glycoside hydrolase. (GH) family 31 contains enzymes that catalyze several different reactions on glycosides. These include a nucleophilic substitution reactions with net retention of stereochemistry, common to retaining glycosidases, as well as an unusual ^-elimination reaction. Since structures and mechanism are expected to be conserved in the same gene family, the two mechanisms were expected to feature common aspects. Three different G H family 31 enzymes were therefore studied. A double displacement mechanism for the retaining cc-glucosidase from Aspergillus niger was shown via trapping of a covalent glycosyl-enzyme intermediate with the mechanism based inactivator 5-fluoro-a-D-glucopyranosyl fluoride. The amino acid residue involved, Asp 224, was identified by L C M S / M S analysis of proteolytic digests. This residue is fully conserved in G H family 31 and has been suggested to be the catalytic nucleophile. A n unusual G H family 31 enzyme, the oc-l,4-glucan lyase from Gracilariopsis sp. (GLase) that cleaves the glycosidic bond of a-l,4-glucans via a net (3-elimination reaction was also studied. The trapping of a covalent glycosyl-enzyme intermediate using 5-fluoro-(3-L-idopyranosyl fluoride, another mechanism based inactivator of oc-glucosidases, strongly suggests that the mechanism also involves the formation of a covalent intermediate like that of a-glucosidases. The labeled amino acid residue was confirmed to be the highly conserved Asp 553, equivalent to Asp 224, the catalytic nucleophile in a-glucosidase from A. niger. A detailed mechanistic evaluation was also carried out. A classical bell shaped p H dependence of kcJKm indicates two ionizable ii groups (pKai =3 .1 , pKa2 ~ 6.7). Bronsted relationships of log & c a t versus pZa and log (kcat/Km) versus pKa for a series o f aryl glucosides both show a linear monotonic dependence on leaving group pKa with low p i g values of -0.32 and -0.33, respectively. The combination of these low p i g values with large a-secondary deuterium kinetic isotope effects (ku/ko - 1-16 ~ 1.19) on the first step indicate a transition state for the glycosylation step with substantial glycosidic bond cleavage and proton donation to the leaving group oxygen. Substantial oxocarbenium ion character at the transition state is also suggested by the potent inhibition afforded by acarbose and 1-deoxynojirimycin and by the substantial rate reduction afforded by adjacent fluorine substitution. For only one substrate, 5-fluoro-oc-D-glucopyranosyl fluoride, was the second, elimination, step shown to be rate-limiting. The large a-secondary deuterium kinetic isotope effect (kulko - 1.23) at C I and the small primary deuterium kinetic isotope effect (knlko - 1.92) at C2 confirm an E2 mechanism with considerable E l character for this second step. This considerable structural and mechanistic similarity with retaining a-glucosidases is a clear example of the mechanistic plasticity of glycosidic bond cleavage through evolution. Finally, an unknown protein (yicl) whose sequence has high similarity with G H family 31 was cloned from E. coli. and shown to be an a-xylosidase. Two new mechanism-based inactivators for a-xylosidases, (5S)- and (5R)-5-fluoro-a-D-xylopyranosyl fluorides were synthesized and shown to inactivate this enzyme. The amino acid residue labeled by these inactivators was identified as the invariant catalytic aspartate residue Asp 416, demonstrating the integrity of the mechanism within this gene family. i i i T A B L E O F C O N T E N T S A B S T R A C T ii 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 x L I S T O F S C H E M E S xv L I S T O F T A B L E S xvi i L I S T O F A B B R E V I A T I O N S xix A C K N O W L E D G E M E N T S xxi C H A P T E R I. Introduction 1 1.1. Enzymatic Cleavage of the Linkage between Two Sugars 2 1.1.1. Nucleophilic Substitution: Glycosyl Transfer 3 1.1.2. p-Elimination 4 1.2. Reaction Mechanisms of the Enzymatic Cleavage of the Linkage between Sugars.... 5 1.2.1. Cleavage of a Glycosidic Bond by Nucleophilic Substitution: Glycosidases 5 1.2.2. Eliminative Cleavage: Polysaccharide Lyases 12 1.3. Bioinformatics Applied to Enzymatic Glycosidic Bond Cleavage 14 1.4. Glycoside Hydrolase Family 31 16 1.4.1. a-Glucosidases 17 1.4.2. cc-Xylosidases 20 1.4.3. oc-l,4-Glucan Lyases 22 iv 1.5. Mechanism Based Inactivators in Mechanistic Studies of Enzymatic Glycosidic Bond Cleavage 27 1.5.1. Irreversible Inhibitors 27 1.5.2. Fluorosugars 29 1.6. A ims of Study 32 C H A P T E R II. Identification of the Catalytic Nucleophile of Glycoside Hydrolase Family 31 a-Glucosidase from Aspergillus niger 33 2.1. Background 34 2.2. Kinetic Evaluation of Aspergillus niger a-Glucosidase 35 2.3. Kinetic Evaluation of the Interaction of Aspergillus niger a-Glucosidase with 5-Fluoro-a-D-Glucopyranosyl Fluoride 38 2.4. Identification of the Catalytic Nucleophile of Aspergillus niger a-Glucosidase 42 2.5. Conclusion 47 C H A P T E R III. Detailed Mechanistic Study on an a- l ,4-Glucan Lyase from Gracilariopsis sp 48 3.1 Background 49 3.2. Previous Studies on the Mechanism of a- l ,4-Glucan Lyase 50 v 3.3. Probable Mechanisms 52 3.3.1. Enzymatic Elimination Reactions 52 3.3.2. Possible Mechanisms 53 3.4. The Reaction of the oc-l,4-Glucan Lyase from Gracilariopsis sp. with Various Substrates 56 3.4.1. Commercially Available Substrates 56 3.4.2. Synthetic A r y l Glycoside Substrates 59 3.5. Active Site Environments - p H Dependent Activity 64 3.6. A Covalent Intermediate and Two-Step Mechanism 65 3.6.1. Background 65 3.6.2. Reaction of a- l ,4-Glucan Lyase from Gracilariopsis sp. with 5 F a G l c F 67 3.6.3. The Reaction of a- l ,4-Glucan Lyase from Gracilariopsis sp. with 5FpTdoF ... 69 3.6.4. The Reaction of a- l ,4-Glucan Lyase from Gracilariopsis sp. with 2FocGlcF and l F G l c F 76 3.7. Transition State Structure 77 3.7.1. The First Step - Glycosylation, Part 1: Bronsted Relationship 77 3.7.2. The First Step - Glycosylation, Part 2: Fluorosugars and Inhibitors 80 3.7.3. The First Step - Glycosylation, Part 3: Kinetic Isotope Effects 85 3.7.4. Formulating the Transition State Structure of the First, Glycosylation, Step 88 3.7.5. Speculations on the Transition State Structure of the Second, Elimination, Step from the Measurement of Kinetic Isotope Effects 90 3.8. Speculations on the General Base Acting in the Second Step 93 3.9. Conclusion 94 v i C H A P T E R IV. Cloning and Kinetic Study of a G H Family 31 a-Xylosidase from Escherichia coli 97 4.1. Background 98 4.2. Sequence Alignment Analysis of yicl Gene Product from E. coli 99 4.3. Cloning and Overexpression of the yicl Gene 101 4.3.1. The Amplification and Cloning of yicl Gene 101 4.3.2. Overexpression of the jyz'cZ Gene 103 4.4. Characterization of the Protein Encoded by yicl 105 4.4.1. Molecular Weight and Analysis of Tryptic Fragments 105 4.4.2. Substrate Specificity 105 4.4.3. Factors Affecting the Activi ty of a-Xylosidase from E. coli (yicl): p H , Temperature and Metal Ions 107 4.5. Kinetic Evaluation of New Mechanism-Based Inactivators for a-Xylosidase from E. coli 110 4.5.1. a-Xylosidases in Nature 110 4.5.2. Synthesis o f (5S)- and (5R)-5-Fluoro-a-D-Xylopyranosyl Fluorides 111 4.5.3. The Inactivation o f a-Xylosidase from E. coli with 5-Fluoro-a-D-Xylopyranosyl Fluorides, 4.3 and 4.4 113 4.6. The Identification of the Catalytic Nucleophile of the a-Xylosidase from E. coli.. 126 4.7 Conclusion 131 Concluding Remarks 132 vi i C H A P T E R V . Experimental Methods 133 5.1. Synthesis 134 5.1.1. General Methods 134 5.1.2. General Materials 135 5.1.3. General Synthesis 136 5.1.4. Synthesis of A r y l a-D-Glucopyranosides 138 5.1.5. Synthesis of (5S)- and (5R)-5-Fluoro-a-D-Xylopyranosyl Fluorides 149 5.2. Enzymes 154 5.2.1. Materials and Methods 154 5.2.2. Cloning and Overexpression of the Gene (yicl) Encoding a-Xylosidase from Escherichia coli 155 5.2.3. Enzyme Kinetics 159 5.2.4. Trapping of the Covalent Glycosyl-Enzyme Intermediate and Identification of the Labeling Site by Tandem Mass Spectrometry 167 A P P E N D I X A . Crystal Structure of 5-Fluoro-Xylosyl Fluorides: Conformational Analysis 171 A . l . Anomeric Effect 172 A . 2 . The Anomeric Effect in Xylosy l Halides 175 A . 3 . Analysis of Crystal Structures of 5-Fluoro-a-D-Xylopyranosyl Fluorides 177 vi i i A P P E N D I X B . Kinetic Isotope Effects 188 B . 1. Primary Kinetic Isotope Effect 189 B.2 . Secondary Kinetic Isotope Effects 192 A P P E N D I X C. Graphical Representation of Data 195 R E F E R E N C E S 204 ix LIST O F F I G U R E S Figure 1.1. Ion-pair intermediate proposed for the lysozyme 8 Figure 1.2. Protonation trajectories of an oc-glycosidase 9 Figure 1.3. Illustration of some inhibitors for glycosidases 10 Figure 1.4. Substrate key polar group mapping of barley high p i a-glucosidase 17 Figure 1.5. Partial multiple sequence alignment of a-glucosidases 18 Figure 1.6. Small molecule substrates for oc-xylosidases 21 Figure 1.7. Scissile bonds of substrates for lyases 24 Figure 1.8. Degradation of a branched pentasaccharide by oc-l,4-glucan lyase 26 Figure 1.9. Affinity labels 27 Figure 1.10. Mechanism-based inactivators 28 Figure 1.11. Fluorosugar inactivators 31 Figure 2.1. 5-Fluoro-a-D-glucopyranosyl fluoride 38 Figure 2.2. Lineweaver-Burke plot for the inhibition kinetics of 5-fluoro-a-D-glucopyranosyl fluoride on A. niger a-glucosidase 39 Figure 2.3. Hydrolysis of 5-fluoro-a-D-glucopyranosyl fluoride catalyzed by A. niger a-glucosidase 41 Figure 2.4. Total ion chromatogram of the peptic digest of labeled A. niger a-glucosidase 43 Figure 2.5. Mass spectrum of the unlabeled (A) and labeled (B) A. niger a-glucosidase peptic digests taken at 27.63 min and 26.76 min, respectively 44 x Figure 2.6. The daughter ion spectrum ( M S / M S ) of m/z 1011 isolated from the labeled digest along with an interpretation of the spectrum 45 Figure 3.1. Enzymatic elimination 52 Figure 3.2. A coupled continuous assay using/>-nitrophenyl p-maltoside as a substrate. 57 Figure 3.3. pH-dependent hydrolysis of 2,4DNPccGlc by GLase 64 Figure 3.4. Difluorosugars designed as mechanisnvbased reagents for glycosidase trapping 66 Figure 3.5. Dixon plot showing the apparent reversible competitive inhibition of a-1,4-glucan lyase by 5FocGlcF 68 Figure 3.6. Inactivation of GLase by 5FpTdoF 70 Figure 3.7. Protection from inactivation in the presence of 0.2 p M acarbose (K\ - 0.02 p M ) at 30 m M 5FidoF and reactivation of inactivated a-l,4-glucan lyase upon removal of excess 5FIdoF 71 Figure 3.8. Detection of 5Fido-labeled peptide by comparative mapping 73 Figure 3.9. ESI M S / M S daughter ion spectrum along with interpretation 74 Figure 3.10. Bronsted plot constructed from data of Table 3.2 showing the relationship of the rate of cleavage of a series of aryl glucosides with the pKa o f the corresponding phenol 78 Figure 3.11. Schematic Diagram of the Transition State of the First Step of the Reaction Catalyzed by GLase 89 Figure 3.12. Schematic diagram of the transition state of the second step of the reaction (elimination) catalyzed by GLase 92 xi Figure 3.13. Schematic diagram of possible hydrogen abstraction by the catalytic residue which acts as a catalytic nucleophile in the first step of the reaction catalyzed by GLase 93 Figure 4.1. Partial sequence alignment of bacterial enzymes and hypothetical proteins of G H family 31 100 Figure 4.2. Designed primers for amplification of yicl gene by polymerase chain reaction and constructed plasmid containing yicl Gene 102 Figure 4.3. Agarose gel electrophoresis of amplified yicl gene 103 Figure 4.4. Electrophoretic analysis of expressed yicl protein 104 Figure 4.5. pH-dependent activity of the hydrolysis o f ocXylF by a-xylosidase from E. coli , 109 Figure 4.6. Temperature Dependence of the Hydrolysis of P N P a X y l by a-xylosidase from E. coli 110 Figure 4.7. Double reciprocal plots showing the apparent reversible competitive inhibition of a-xylosidase from E . coli by ax-5FaXF 4.3 (A) and eq-5FaXF 4.4 (B) 115 Figure 4.8. Labeling of Ful l Length a-Xylosidase from E . coli with ax -5FaXF 4.3 and eq-5FaXF 4.4 116 Figure 4.9. Inactivation of a-Xylosidase from E. coli with 4.3 and 4.4 at 10 °C 118 Figure 4.10. Reactivation at 37 ° C of inactivated (at 10 °C) a-xylosidase from E. coli.. 119 Figure 4.11. Hydrolysis of (5S)- and (5R)-5-fluoro-a-D-xylopyranosyl fluorides catalyzed by E. coli a-xylosidase 120 xn Figure 4.12. Hypothetical conformation change of sugar moiety of the covalent glycosyl-enzyme intermediate 125 Figure 4.13. Comparative Mapping of Peptic Digests of Control Sample taken at 30.2 minutes of total ion chromatogram (A), Labeled Peptide Samples with ax-5FocXF 4.3 taken at 30.4 min (B) and eq-5FaXF 4.4 taken at 30.5 min (C) 127 Figure 4.14. ESI M S / M S daughter ion spectrum of the peptide fragment m/z 902 (doubly charged) labeled by eq-5FocXF 4.4 129 Figure A . l . Schematic diagram of anomeric effect 172 Figure A . 2 . Schematic diagram of the origin of the anomeric effect 173 Figure A . 3 . Conformational equilibrium between 4 C , and XCA of 2,3,4-tri-0-acetyl-(3-D-xylopyranosyl chloride and fluoride in solution 175 Figure A .4 . X-ray crystal structure of per-O-acetyl diftuoro derivatives of xylopyranose 180 Figure A . 5 . Schematic diagram showing the conformations and selected bond lengths of various per-O-acetyl xylosyl fluorides 182 Figure A .6 . Schematic diagram showing selected bond angles ( 0 5 - 0 - C 2 , C1-05-C5 , and 05-C5-C4) excerpted from X-ray crystal structures of various per-O-acetyl xylosyl fluorides 183 Figure A . 7. Unfavorable dipole-dipole interaction between three parallel dipoles in the hypothetical 4 C i conformation of eq -5FpXF-OAc 184 Figure A . 8. Unfavorable dipole-dipole interaction between two parallel C -F bond dipoles that would form in ax-5FocXF-OAc 185 x in Figure B . l . Schematic Morse potential energy diagram showing the zero-point vibrational energy of C - H and C - D bonds 190 Figure B.2 . Schematic energy diagram corresponding to the reaction X - H + Y 192 Figure B.3 . Hybridization change causing secondary kinetic isotope effect 193 Figure C . l . Dixon plots showing the inhibition of a-glucosidase from Aspergillus niger by the reversible competitive inhibitors , 196 Figure C.2. Michaelis-Menten plots for the reaction a-l,4-glucan lyase with a series of aryl a-glucosides 197 Figure C.3. Michaelis-Menten Plots for the reaction of a-l,4-glucan lyase with fluorosugars 199 Figure C.4. Dixon plots showing the inhibition of a-l,4-glucan lyase by the reversible competitive inhibitors 201 Figure C.5. Reactivation at 10 ° C of inactivated (at 10 °C) a-xylosidase from E. coli. 203 xiv LIST O F S C H E M E S Scheme 1.1. Action of glycosidases 3 Scheme 1.2. Action of polysaccharide lyases 5 Scheme 1.3. The mechanism of a retaining oc-glycosidase 7 Scheme 1.4. The mechanism of an inverting oc-glycosidase 11 Scheme 1.5. The mechanism of a polysaccharide lyase 13 Scheme 1.6. Degradation of oc-l,4-glucans by the action of oc-l,4-glucan lyase 22 Scheme 1.7. In vivo transformation of 1,5-anhydro-D-fructose 23 Scheme 1.8. Inactivation of an cc-glycosidase by the reaction with conduritol B epoxide 29 Scheme 1.9. Inactivation of an oc-glycosidase by the action of a fluorosugar 30 Scheme 3.1. A direct an '/-elimination mechanism 54 Scheme 3.2. A proposed GLase mechanism involving a covalent glycosyl-enzyme intermediate 55 Scheme 3.3. The synthesis of 2,4-dinitrophenyl a-D-glucopyranoside 60 Scheme 3.4. BF3-diethyl etherate catalyzed synthesis of aryl glycosides 61 Scheme 3.5. SN2 type synthesis of aryl glycosides 62 Scheme 3.6. Possible inactivation scheme of GLase by the accumulation of a 2-keto glucosyl-enzyme intermediate from the reaction of GLase with l F G l c F 76 Scheme 4.1. Synthesis of axial (4.3) and equatorial (4.4) 5-fluoro-a-D-xylopyranosyl fluorides 113 xv Scheme 4.2. Kinetic scheme of the reaction of ax-5FccXF (or eq-5FaXF) with a-xylosidase (E) 123 xvi LIST O F T A B L E S Table 2.1. Michaelis-Menten Parameters for the Hydrolysis o f Substrates by A. niger a-Glucosidase 36 Table 2.2. The Structure of Inhibitors and^Ti values for A nigers a-Glucosidase 37 Table 3.1. Michaelis-Menten Parameters for the Hydrolysis of Maltooligosaccharides and Polysaccharides by Gracilariopsis a-l ,4-Glucan Lyase 51 Table 3.2. Michaelis-Menten Parameters for the Hydrolysis of a Series of A r y l a-Glucosides by Gracilariopsis a-l ,4-Glucan Lyase 63 Table 3.3. Michaelis-Menten Parameters for the Cleavage of Fluorosugars by Gracilariopsis a-l ,4-Glucan Lyase 80 Table 3.4. The Structure of Inhibitors and K{ values for Gracilariopsis a-l ,4-Glucan Lyase 84 Table 3.5. Kinetic Isotope Effects Measured for Deuterated Substrates with Gracilariopsis a-l ,4-Glucan Lyase 86 Table 4.1. Purification of Y i c l Protein 105 Table 4.2. The Hydrolysis of Various Substrates by yicl protein from E. coli 106 Table 4.3. Kinetic Parameters for the Hydrolysis o f Fluorosugars by the a-Xylosidase from E. coli Measured at 37 °C 122 Table A . l . Solvent Dependence of the Conformational Equilibrium of 2-Methoxytetrahydropyran 174 Table A . 2 . Experimental Details of X-Ray Crystallography 178 xvn Table A . 3 . Selected Torsion Angles of 2,3,4-Tri-0-Acetyl-5-Fluoro-D-Xylosyl Fluorides Excerpted from Crystallographic Data 184 xvm LIST O F A B B R E V I A T I O N S l F G l c F : 1 -fluoro-D-glucopyranosyl fluoride 2 ,4DNPaGlc : 2,4-dinitrophenyl a-D-glucopyranoside 2 ,5DNPaGlc : 2,5 -dinitrophenyl a-D-glucopyranoside 3 ,4DNPaGlc : 3,4-dinitrophenyl a-D-glucopyranoside 3 ,5DCPaGlc : 3,5 -dichlorophenyl a-D-glucopyranoside 5 , 5 F 2 p X F - O A c : 2,3,4-tri-0-acetyl-5,5-difluoro-P-D-xylopyranosyl fluoride 2FaGlcF: 2-deoxy-2-fluoro-a-D-glucopyranosyl fluoride 4C2NPocGlc: 4-chloro-2-nitrophenyl a-D-glucopyranoside 5FaGlcF: 5 -fluoro-a-D-glucopyranosyl fluoride 5FpTdoF: 5-fluoro-P-L-idopyranosyl fluoride Abg : Agrobacterium sp. (3-glucosidase aGlcF : a-D-glucopyranosyl fluoride ocKIE: a-secondary deuterium kinetic isotope effects a X y l F : a-D-xylopyranosyl fluoride a X y l F - O A c : 2,3,4-tri-O-acetyl-a-D-xylopyranosyl fluoride AnFru: 1,5-anhydro-D-ffuctose ax-5FaXF: (5S)-5-fluoro-a-D-xylopyranosyl fluoride ax -5FaXF-OAc: (5S)-2,3,4-tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride p X y l F - O A c 2,3,4-tri-O-acetyl-P-D-xylopyranosyl fluoride p X y l F - O B z : 2,3,4-tri-O-benzoyl-p-D-xylopyranosyl fluoride B S A : bovine serum albumin C B E : conduritol B epoxide D M F : dimethylformamide D M P U : 1,3-dimethyl-3,4,5,6-tetrahydro-2( 1 H)-pyrimidinone D N A : deoxyribose nucleic acid D N S : dinitrosalicylate D P : degree of polymerization E C : Enzyme Commission E D T A : ethylenediamine tetraacetic acid ESI M S : electrospray ionization mass spectrometry eq-5FaXF: (5R)-5-fluoro-a-D-xylopyranosyl fluoride eq-5FaXF-OAc: (5R)-2,3,4-tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride eq-5F(3XF-OAc: (5R)-2,3,4-tri-0-acetyl-5-fluoro-P-D-xylopyranosyl fluoride F D N B : fluorodinitrobenzene G A G : glycosaminoglycan G H : glycoside hydrolase GLase: a-l,4-glucan lyase H M P A : hexamethyl phosphoramide I U B M B : International Union of Biochemistry and Molecular Biology K I E : kinetic isotope effect L C : liquid chromatography xix MNPccGlc: 3-nitrophneyl a-D-glucopyranoside M S : mass spectrometry M S / M S : tandem mass spectrometry N A D : nicotinamide adenine dinucleotide N A D P : nicotinamide adenine dinucleotide phosphate N B S : N-bromosuccinimde N M R : nuclear magnetic resonance O N P a G l c : 2-nitrophenyl a-D-glucopyranoside O R P : open reading frame P C P a G l c : 4-chlorophenyl a-D-glucopyranoside P C R : polymerase chain reaction P N P a G l c : />-nitrophenyl a-D-glucopyranoside P N P p G l c : /7-nitrophenyl (3-D-glucopyranoside P N P a X y l : p-nitrophenyl a-D-xylopyranoside P N P p M a l : /7-nitrophenyl P-maltoside R P - H P L C : reverse phase-high performance liquid chromatography SDS P A G E : sodium dodecyl sulfate polyacrylamide gel electrophoresis TCPocGlc: 2,4,6-trichlorophenyl a-D-glucopyranoside T E M E D : N , N , N ' , N ' -tetramethylethylenediamine TIC: total ion chromatograms T L C : thin layer chromatography U V / V I S : ultraviolet/visible XX A C K N O W L E D G E M E N T S I would like to thank, first and most of all, my research supervisor, Professor Stephen G . Withers. His guidance and inspiration, not only in science but also in life, make possible what I am here and now. Professors David Dolphin, Lawrence Mcintosh and David Perrin are members of the guidance committee. They have been great inspiration on the way here. I would also like to thank former and present members of Withers Lab. Their friendship and help make life in the laboratory enjoyable. Special thanks to the following people: Dr. Shouming He is an expert in mass spectrometry and provided excellent spectra. Dr. Young Wan K i m is a highly talented molecular biologist who helped cloning E. coli gene. Dr. David Vocadlo, now Professor at Simon Fraser University, and Dr. Tanja M . Wrodnigg kindly provided their results and compounds. Dr. Shukun Y u is our faithful collaborator in a-l,4-glucan lyase project. I have really enjoyed the collaboration with him. Dr. Brian Patrick has provided nice X-ray crystal structures of xylosyl fluorides, which is greatly appreciated. M y thanks to University of British Columbia for providing me with University Graduate Fellowship and to the Department of Chemistry for providing me with Gladys Estella Laird Research Fellowship. Also, thanks to Arthur S. Hawkes Foundation and Boehringer Ingelheim (Canada) Ltd. for providing me with scholarships. To my parents and parents-in-law in Korea, I 'd like to express big thanks for all the love and support. For the biggest part of life here in Vancouver, I owed much to my family. M y lovely wife, Eun Joo, and son, Ryan (Dongchan), I would like to express gratitude from the deepest part of my heart for being the biggest reason to live for. xxi 1 Chapter I. Introduction 2 1.1. Enzymatic Cleavage of the Linkage between Two Sugars Carbohydrates, also known as sugars, are not only the most abundant carbon source in biosphere [1] but also versatile and vital elements in biological systems. Carbohydrates provide a carbon source for other compounds and chemical energy for immediate release or storage in the form of polysaccharides such as starch or glycogen. Many carbohydrates, including cellulose and chitin, play a role as structural units for support and protection of organisms. These structural polysaccharides are also found to be associated with proteins in the case of glycoproteins, or peptides in the case of peptidoglycans, both of which play important roles in living processes such as in recognition or in the structure of cell walls [2]. Reflecting this wide variety of roles in nature, carbohydrates have numerous structural features. Monosaccharides, the smallest unit of carbohydrates, can be linked to each other in numerous ways. Thus, while monomers can be linked linearly in a simple fashion, each linkage can be of a different anomeric configuration. Furthermore, there could be branches through each hydroxyl group of a monosaccharide unit, making an enormous number of possible structures. A glycosidic bond plays a central role in these carbohydrate structures. It links the acetal center (anomeric center) of a monosaccharide to a hydroxyl of another monosaccharide. This is a very stable chemical bond, with a half-life for cleavage under physiological conditions on the order of five mil l ion years. The rate acceleration achieved by enzymes performing glycosidic bond cleavage is amongst the highest of any known enzymes reaching 10 1 7 fold [3,4]. There are two major classes of reaction by which the linkage between two sugars can be broken, a nucleophilic substitution and a p-elimination. 3 1.1.1. Nucleophilic Substitution: Glycosyl Transfer Glycosyl transfer is formally a nucleophilic substitution at the saturated carbon of the anomeric center. Glycosyl transfer can occur between two oxygen nucleophiles or an oxygen and an atom other than oxygen, such as nitrogen, a halogen, or even carbon [5]. If a glycosyl group is transferred to a water molecule in the enzymatic reaction, this enzyme is a glycosidase (EC 3.2.1.-), also known as a glycoside hydrolase since it catalyzes hydrolysis of glycosides. The glycosidase is the most commonly occurring of all the glycosidic bond cleaving enzymes. A HO' B o OH Scheme 1.1. Action of glycosidases: A. a-glycosidases; B. p-glycosidases If another sugar is the glycosyl acceptor, the enzyme wi l l be called a transglycosidase. However, the reactions that occur in the two cases are the same. Many glycosidases also 4 catalyze transglycosidation reactions and vice versa. Thus, later discussions about nucleophilic substitution wi l l be limited to glycosidases or glycoside hydrolases that act on O-glycosides. The glycosyl donor could be a five-membered ring sugar (fiiranoses) or a six-membered ring sugar (pyranoses) and the glycosidic bond at the anomeric center of the donor could be either axial or equatorial. If substrates are D-series pyranoside sugars, enzymes acting on the axial glycosidic bond (cc-glycosides) at the anomeric center are called a-glycosidases while those acting on the equatorial glycosidic bond (p-glycosides) are called P-glycosidases. Glycosyl transfer can take place with either retention or inversion of the anomeric configuration. In the case where the reaction proceeds with overall retention of anomeric stereochemistry, the enzymes are termed retaining glycosidases, while inverting glycosidases generate products of inverted stereochemistry at the anomeric center. Actions of glycosidases on D-pyranoses are summarized in Scheme 1.1. 1.1.2. P-Elimination Whereas most enzymatic cleavages of the linkage between sugars occur through nucleophilic substitution reactions involving a glycosidic bond, the eliminative enzymes also play an important role in carbohydrate processing. Most of the elimination reactions involved in polysaccharide degradation can be found within the class of enzymes termed polysaccharide lyases [6,7]. These enzymes are responsible for the degradation of 5 glycosaminoglycans (GAGs) in prokaryotes or pectins in plants. GAGs are long, unbranched, highly sulfated, acidic polysaccharides that are markedly heterogeneous along their chain. GAGs are composed of an alternating sequence of uronic acid and hexosamine residues, sulfated at various positions. Polysaccharide lyases (EC 4.2.2.-) differ from glycosidases in that they cleave the linkage between C4 of the non-reducing end of uronic acids and the glycosidic oxygen by abstracting the C5 hydrogen of uronic acids and eliminating the sugar unit on C4 (Scheme 1.2). The reaction results in the formation of a C4-C5 double bond on the uronic acid residue. R: H , S 0 3 -R': uronic aicd R": hexosamine Scheme 1.2. Action of polysaccharide lyases 1.2. Reaction Mechanisms of the Enzymatic Cleavage of the Linkage between Sugars 1.2.1. Cleavage of a Glycosidic Bond by Nucleophilic Substitution: Glycosidases The mechanism of glycosidases has been the subject of many studies [5,8,9,10] and the mode of action has been studied in great detail. As described earlier, glycosidic 6 bond cleavage proceeds with overall retention or inversion of the anomeric center. The first proposal for the mechanism of glycosidases came from Koshland [11]. He pointed out that it was likely that glycoside hydrolases which yielded retention of anomeric configuration worked via a double displacement mechanism involving a covalent glycosyl-enzyme intermediate, whereas inverting glycosidases probably worked by a single chemical step involving the displacement of the aglycon leaving group by a water nucleophile (Scheme 1.3 and 1.4). Numerous mechanistic studies support this early proposal. 1) Retaining Mechanism. The double displacement mechanism of retaining glycosidases involves two successive nucleophilic substitutions at the anomeric center [5,8,9,10]. The first step (glycosylation) produces a covalent glycosyl-enzyme intermediate with inversion at the anomeric center, followed by the second step (deglycosylation) generating the final product with another inversion, thereby resulting in the net retention of stereochemistry (Scheme 1.3). Retaining enzymes typically have two catalytic carboxylic acids in their active site [5,10]. One carboxylic acid in the active site acts as the catalytic nucleophile which attacks the anomeric center from the opposite face to the aglycon leaving group, leading to the formation of the covalent glycosyl-enzyme intermediate, a glycosyl ester. The other plays the dual roles of the general acid in the first step and base catalyst in the second step. The carboxylic groups in retaining glycosidases are only 5.5 A apart, providing just enough space to accommodate only the substrates [8,9,10,12]. 7 H O H O O H Q J W V A A A A A A A A A A H O ^ R r; A, Glycosylation Deglycosylation J \ A A A / W W W V \ A H O &-6rR ° Y ° H 2 Q s ^ - ROH O H x / V A / W W W W W X H O , H O ' 1 °- C o - ^ o - H '"3 H O ^ H \ °Y°' *AAAAAAATV\AAAA H O H O O H Q H ° 6 - H H vAAAAAAAAAAAAA. Scheme 1.3. The mechanism of a retaining oc-glycosidase The first enzyme for which an X-ray crystal structure became available was hen's egg white lysozyme [13] which is a retaining p-glycosidase. Based on this early structure, an ion-pair intermediate was proposed instead of a covalent intermediate (Figure 1.1), suggesting that the role of the carboxyl group is to stabilize a glycosyl cation. 8 ^OH Figure 1.1. Ion-pair intermediate proposed for hen's egg white lysozyme However, accumulated experimental evidence strongly supports the existence o f the covalent glycosyl-enzyme intermediate. Several linear free energy relationship studies on glycosidases show that a plot of log kca, values for the hydrolysis of aryl glycosides versus aglycone pKa values produce biphasic, concave-downward Bronsted relationships, indicating a change in rate-limiting step [14,15,16]. The glycosylation step is rate-limiting step for substrates with poor leaving groups (high pKa values) while the rates for substrates with good leaving groups are essentially identical, indicating that the deglycosylation step is rate limiting. a-Secondary deuterium kinetic isotope effects (ocKIE) on glycosidase reactions with substrates having a rate limiting second step have values of greater than one, indicating that the hybridization of the anomeric carbon has changed from sp 3 to sp 2, consistent with a covalent intermediate [10,14]. Further, in several cases, the covalent intermediates have been trapped with mechanism-based inhibitors [17-23], and the labeled amino acid residues identified as aspartic acid or glutamic acid, with sialidases being the only exception. Several trapped complexes have been subjected to X-ray crystallographic analysis, providing valuable insight into the nature of the intermediate [24-27]. These studies have been carried out on a number of glycosidases, including hen's egg white lysozyme [27], and have clearly shown the involvement of a covalent intermediate. 9 A. B. HO' C H - O A o Figure 1.2. Protonation trajectories of an a-glycosidase: A . anti-protonation; B . syn-protonation While one carboxyl group attacks the anomeric center, the other donates a proton to the leaving group oxygen to facilitate the departure of the aglycone. It has been proposed recently that the protonation occurs laterally, and that two protonation trajectories are possible (Figure 1.2) [28]. The proton can come from the same side as the endocyclic oxygen or the opposite, thereby defining the syn or and protonation designation, respectively. The spontaneous hydrolysis of glycosides in aqueous medium has been shown to proceed through an oxocarbenium ion-like transition state [29,30], and a similar transition state has been assumed for glycosidase reactions [5,10,31]. The existence of oxocarbenium ion-like transition states in the glycosidase reaction has been illustrated by the measurement of large a-secondary deuterium kinetic isotope effects (a-KIE) on a range of glycosidases [5,10,14,32,33]. Oxocarbenium ion character is also suggested by the potent inhibition afforded by competitive inhibitors that mimic this positively charged transition state [28,31]. 10 B C D E Figure 1.3. Illustration of some inhibitors for glycosidases: A. acarbose; B. gluconolactone; C. 1-deoxynojirimycin; D. gluconohydroximolactam; E. isofagomine Positive charge at the endocyclic oxygen as well as the anomeric carbon, sp hybridization at the anomeric carbon and a half chair or envelope conformation enforcing planar geometry at the anomeric carbon are frequently utilized to generate a powerful inhibitor. Compounds such as nojirimycin, 1-deoxynqjirimycin, acarbose, isofagomine, glyconolactones, glyconolactams, and glyconohydroximolactams containing any of these features have shown improved affinity for glycosidases compared to substrates (Figure 1.3) [28,31]. 11 2) Inverting Mechanism. Scheme 1.4. The mechanism of an inverting a-glycosidase The mechanism of the inverting glycosidases has been discussed in terms of a single displacement by a water nucleophile as first proposed by Koshland [11] (Scheme 1.4). Inverting glycosidases also have two carboxyl groups in the active sites but these two carboxyl groups are 7 - 10.5 A apart on average, allowing the substrate and a water nucleophile to bind simultaneously, affording a single displacement reaction [8,12]. Thus, one carboxyl plays the role of a general acid, donating a proton to the glycosidic bond to facilitate the departure of the leaving group while the other acts as a general base deprotonating the incoming water nucleophile to assist a nucleophilic attack. The transglycosidation activity has not been detected in inverting glycosidases and this is quite reasonable since the product of the inverting glycosidases w i l l have the wrong configuration at the anomeric center and this cannot be the substrate for the reverse reaction to fulfill the requirements of microscopic reversibility [5]. The transition state structure of this enzymatic reaction also has substantial oxocarbenium ion character. In several cases, moderate to large a-secondary deuterium kinetic isotope effect values ( k H / k D = 1.09 - 1.17) have been observed [34,35], indicating substantial positive charge development at the transition state. Based on a-secondary tritium and primary 1 4 C kinetic 12 isotope effects, and computational modeling, the transition state for the hydrolysis of a -glucosyl fluoride catalyzed by an inverting glycosidase was modeled. A t the transition state, only 5% of the glycosidic bond is retained and the new bond being created is barely forming. The modeling also reveals that the anomeric center assumes a trigonal geometry and that the bond between the anomeric center and the endocyclic oxygen has a bond order of 1.92, which is fully consistent with an oxocarbenium ion like transition state [36]. 1.2.2. Eliminative Cleavage: Polysaccharide Lyases Whereas studies on glycosidases are facilitated by a number of available and easily monitored substrates and analogues, such is not the case for polysaccharide lyases. Natural substrates for polysaccharide lyases are highly heterogeneous mixtures of high molecular weight polymers which themselves generate other shorter substrates upon degradation by the enzyme. Thus, the lack of defined and reliable substrates has made it very difficult to perform mechanistic studies. Nonetheless, the mechanistic features of the polysaccharide lyase reaction have been proposed [37]. According to this, neutralization of the negative charge on the carboxylate anion of the uronic acid moiety is achieved either by metal chelation or protonation, followed by general base-catalyzed C5 proton abstraction. The ^-elimination of the 0 4 linked sugar then takes place (Scheme 1.5). A key element of this mechanism is the uronic acid moiety of the reducing sugar. The substrate carboxylic acid therefore facilitates the abstraction of the C5 proton by stabilizing the negative charge generated upon abstraction of the C5 proton. However, 13 whether this overall ^^-el iminat ion takes place in a stepwise or concerted fashion was determined only recently [38,39]. £>R HA o-HX R'O-Charge Neutralization N H A c H 0 ~ l H B: -OR" OH R: H, S03-R': uronic aicd R": hexosamine Proton Abstraction R'O + OH O r V O R H A ^ . 0 - ' H X -Q R'O-NHAc -OR" JVVA/IAAA/* Elimination O R ^ ^ ^ R' HX Scheme 1.5. The mechanism of a polysaccharide lyase BH* O H NHAc The introduction of defined and easily measurable substrates has provided useful insights into the mechanism of polysaccharide lyases and has allowed studies that were not previously possible [38,39]. Thus, a primary kinetic isotope effect on kcaX/Km was measured using the 5-[2H]-substrate, indicating that the proton abstraction with such substrates occurs in the rate-limiting step. B y contrast, no dependence of the kcJKm values measured for breakdown of a series of 4-O-aryl glucuronic acid substrates upon the pKa values of the leaving groups was observed, showing that the C4-04 bond is probably not broken in the rate limiting step. These two results, coupled with the very low secondary deuterium kinetic isotope effect upon kcat/Km measured for the 4-[ FfJ-substrate, are fully consistent with the proposed stepwise E l c B mechanism involving an enolate intermediate. 14 1.3. Bioinformatics Applied to Enzymatic Glycosidic Bond Cleavage A s a complement to ITJBMB's Enzyme Commission (EC) classification, a new classification system was proposed for glycosidases or glycoside hydrolases in 1991 [40]. The significant number of available amino acid sequences, which is still increasing quickly, allowed the establishment of this new classification system for glycosidases based upon amino acid sequence similarities. Amino acid sequences of glycosidases or glycoside hydrolases were compared and enzymes which have significant similarity were grouped into the same family. The classification started with 35 families and has been constantly updated as the number of available sequences grows [41-44]. Currently, the number of families has reached 91. A similar classification system has been developed for the glycosyl transferases, enzymes forming glycosidic bonds using activated sugars such as nucleoside diphosphosugars and sugar-1-phosphates [45]. Polysaccharide lyases also have been classified by the same methods [45], the entry now containing 13 families. The families of these "Carbohydrate Active enZYmes" can be accessed through a web site ( C A Z Y web site), http://afmb.cnrs-mrs.fr/CAZY/index.html. This web site not only contains the above three enzyme groups but also families of carbohydrate esterases and carbohydrate-binding modules. Entries have been steadily updated. This classification system has obvious advantages over the traditional ITJBMB system. Since sequence similarity frequently reflects structural similarity, the members of the same family most likely have the same folding characteristics [46]. Indeed, with the growth of the number of available X-ray crystal structures of glycosidases, it was found 15 out that this is the case and even some of the different families share the same fold [9,47]. The families with related folding characteristics have been further grouped into a higher hierarchical level as clans [9,42,48]. Structural and sequence similarity usually leads to mechanistic similarity and it is notable that each glycoside hydrolase family presents only a single mechanism, either retaining or inverting. Also, enzymes in a given family display the same stereoselectivity, meaning that, with a very rare exception, each family does not contain both a- and (3-glycosidases. These findings indicate that the structure, and consequently reaction mechanism, are conserved through members of a given family. Thus, mechanistic information on one family member can be readily extended to all members of a given family. In the case of glycosidases, there are two crucial carboxylate groups in the active site. These active site residues can often be predicted by inspecting the invariant residues through sequence comparison within the family [41,42]. However, further structural or labeling information is needed to confirm this. It is also remarkable that, through this sequence alignment, an unknown protein or even a putative gene product can often be classified, and the reaction mechanism and amino acid residues involved in catalysis could be deduced before even performing enzymatic characterization. Many families of glycosidases are polyspecific, containing enzymes of different substrate specificity (different E C numbers). For example, glycoside hydrolase (GH) family 1 contains no less than 15 E C numbers and G H family 13 no less than 20 E C numbers. Interestingly, some families also contain enzymes normally not classified as O-glycoside hydrolases, such as myrosinase [43]. These could be indications of divergent evolution in the family. It is easy to deduce from this that an ancestor enzyme may 16 acquire a new substrate specificity through evolutionary events and this might be reflected in the polyspecificity of sequence-based families. On the other hand, enzymes of the same substrate specificity can be found in different families. For example, oc-glucosidases are found in G H family 13 and 31, and cellulases in 11 different families. 1.4. Glycoside Hydrolase Family 31 Glycoside hydrolase family 31 is one of two families that contain mostly oc-glucosidases. The other is G H family 13, the famous a-amylase superfamily. G H family 31 also contains cc-xylosidases and recently discovered a-l,4-glucan lyases. Enzymes of G H family 31 are formally retaining glycosidases [49,50]. Although G H families 13 and 31 share substrate specificity in large part, there is only extremely low amino acid sequence similarity between the two families, around the active site, leading to the concept of a clan [43,51]. However, no X-ray crystal structure of a G H family 31 enzyme has yet been determined. Only after this is done, w i l l it be verified that these two families indeed form a clan. Recently, a new class of starch-degrading enzyme that catalyzes a P-elimination reaction on a-l,4-glucans was discovered [52]: the a-l,4-glucan lyases. On the basis of amino acid sequence similarity, cc-l,4-glucan lyases belong to G H family 31 despite the difference in reaction catalyzed. 17 1.4.1. a-Glucosidases The primary role of a-glucosidases (EC 3.2.1.20) is to hydrolyze oligosaccharides produced enzymatically from starch, yielding glucose. These enzymes are exo-acting enzymes that release glucose from the non-reducing end of the oligosaccharide. In many cases, a p H optimum was observed in the acidic region (pH 4-5), with a few operating in the neutral region [49]. There are three types of a-glucosidases. While type I has a higher activity toward heterogeneous substrates such as aryl glucosides and sucrose than to maltooligosaccharides, type II has the opposite substrate preference. Type III has the same specificity as type II but can additionally hydrolyze polysaccharides such as starch and amylose. The majority of G H family 31 a-glucosidases have a preference for maltose and longer maltooligosaccharides (a-l ,4-l inked glucose units) and, therefore, belong to type II or III while a-glucosidases of G H family 13 belong to type I enzymes [49]. Figure 1.4. Substrate key polar group mapping of barley high pi a-glucosidase. Numbers are AAG* values for monodeoxygenated maltose derivatives, which were calculated to reflect the change on the free energy of activation for the first irreversible step (kcJKm) compared to the parent sugar, maltose, n.d. = not determined. Excerpted from references 49, 53 - 56. 18.7 kJ/mol 7kJ/mol -0.3kJ/mol -0.3 kJ/mol 9.7 kJ/mol n.d. 18 Molecular recognition analysis (Figure 1.4) reveals large contributions from the hydroxyl groups of the non-reducing end sugar and also significant contributions from the reducing end sugar to transition state stabilization [53-56]. Combined with the finding that subsites 1 - 3 are important for the binding of substrates, this supports the type II or III specificity of G H family 31 oc-glucosidases [57]. Family 31 a-glucosidases are also known for their transglucosidation activity [58]. A. S i g n a t u r e R e g i o n 1 * Barley High p i a-glucosidase 432 DGLWIDMNEISNF 444 Sugar beet a-glucosidase 464 DGIWIDMNEASNF 476 Human isomaltase 500 DGLWIDMNEVSSF 512 Human sucrase 1389 DGLWIDMNEPSSF 1401 Human lysosomal a-glucosidase 513 DGMWIDMNEPSNF 525 Schizosaccharomyces pombe a-glucosidase 47 6 SGIWTDMNEPSSF 489 Aspergillus niger a-glucosidase, P2 subunit 219 DGVWYDMSEVSSF 231 Garcilariopsis sp. a-l,4-glucan lyase 548 DFVWQDMTVPAMM 560 Morchella costata a-l,4-glucan lyase 543 EFVWQDMTTPAIH 555 Arabidopsis thaliana a-xylosidase (XYL1) 435 DGLWIDMNEVSNF 447 Tropaeolum majus a-xylosidase 454 DGLWIDMNEDLEF 466 Pinus pinaster a-xylosidase 433 DGLWIDMNEISNF 445 Lactobacillus pentosus a-xylosidase (XylQ) 409 DSFKTDFGER 418 Escherichia coli K12 ORF ( y i c l ) 411 DCFKTDFGER 420 Sulfolobus solfataricus a-xylosidase (XylS) 348 DAYWLDASEP 357 B. S i g n a t u r e R e g i o n 2 Barley High p i a-glucosidase 561 GADICGFNG Sugar beet a-glucosidase 595 GADICGFAE Human isomaltase 631 GADICGFVA Human sucrase 1527 GADICGFFN Human lysosomal a-glucosidase 643 GADVCGFLG Schizosaccharomyces pombe a-glucosidase 674 GADVCGFLG Aspergillus niger a-glucosidase, P2 subunit 421 GADTCGFNG Garcilariopsis sp. a-l,4-glucan lyase 692 GSDIGGFTS Morchella costata a-l,4-glucan lyase 662 GSDTGGFEP Arabidopsis thaliana a-xylosidase (XYL1) 590 GSDICGFYP Tropaeolum majus a-xylosidase 609 GSDICGFYP Pinus pinaster a-xylosidase 597 GADICGFYP Lactobacillus pentosus a-xylosidase (XylQ) 508 SHDIGGFED Escherichia coli K12 ORF, y i c l 509 SHDIGGFEN Sulfolobus solfataricus a-xylosidase (XylS) 455 TTDTGGFFS NTTEELCGRWIQLGAFYPFSR 590 STTEELCCRWIQLGAFYPFSR 624 ETTEELCRRWMQLGAFYPFSR 660 NSEYHLCTRWMQLGAFYPYSR 1556 NTSEELCVRWTQLGAFYPFMR 672 DSDEELCSRWMANGAFSPFYR 703 NSDEELCNRWMQLSAFFPFYR 450 ( 8)PCTGDLMVRYVQAGCLLPWFR 729 (10)YCSPELLIRWYTGSFLLPWLR 701 QPTEELCNRWIEVGAFYPFSR 619 GPTEELCNRWIEVGAFYPFSR 638 DTTEELCGRWIQLGAFYPFSR 626 ( 3)TPTADLYKRWSQFGLLSSHSR 540 TAPAHVYKRWC AFGLL SSHSR 538 ( 5)KAYAEIFVRWFQWSTFCPILR 489 Figure 1.5. Partial multiple sequence alignment of a-glucosidases, oc-l,4-glucan lyases and a-xylosidases of GH family 31: A. Signature Region 1; B. Signature Region 2: Proposed catalytic nucleophile is in bold character and indicated by an asterisk (*). Accession numbers: Barley high pi a-glucosidase (GenBank AF118226), sugar beet a-glucosidase (Swiss Prot O04931), human sucrase-isomaltase (Swiss Prot PI4410), human lysosomal a-glucosidase (Swiss Prot PI0235), Schizosaccharomyces pombe a-glucosidase (GenBank AB045751), Aspergillus niger a-glucosidase, P2 subunit (Swiss Prot P56526), Gracilariopsis sp. a-l,4-glucan lyase (Swiss Prot P81676), Morchella costata a-l,4-glucan lyase (Swiss Prot P81696), Arabidopsis thaliana a-xylosidase (GenBank AAD37363), Tropaeolum majus a-xylosidase (GenBank CAA10362), Pinus pinaster a-xylosidase (GenBank AF44821), Lactobacillus pentosus a-xylosidase (GenBank AAC62251), Escherichia coli K12 ORF yicl (Swiss Prot P31434), Sulfolobus solfataricus a-xylosidase (Swiss Prot Q9P999) 19 a-Glucosidases of G H family 31 are retaining enzymes which utilize a double displacement mechanism involving a covalent glycosyl enzyme intermediate (Scheme 1.3). Therefore, two carboxyl groups were expected to participate in catalysis by a -glucosidases. From the analysis of sequence alignments, seven well-conserved regions have been found and two of those highly conserved regions were designated as signature regions 1 and 2 (Figure 1.5) [49]. Indeed, there are highly conserved carboxylic acid residues found in signature regions 1 and 2, and it is highly likely that catalytic residues are among these. Based on labeling studies with conduritol B epoxide (CBE) , a highly conserved aspartic acid was labeled and suggested to be the catalytic nucleophile [59-63]. The site-directed mutagenesis of this residue and accompanying kinetic analysis provided equivocal results that were inconsistent with an essential role for this residue [60,63]. Indeed, C B E has often lead to the incorrect labeling of active site residues (discussed later), indicating that the identity of the catalytic nucleophile still remains in doubt. The identity of the acid/base catalytic residue is also controversial. Mutation of the highly conserved aspartate residue in signature region 2 in human lysosomal and Schizosaccharomyces pombe a-glucosidase caused only a 50 - 1000 fold reduction [49,63], a much smaller change than expected for mutants of the acid/base residue. Mutation of another aspartate residue in the a-glucosidase from Schizosaccharomyces pombe results in a substantial 10 4 - 10 5 fold reduction in activity [63]. However, this residue is not invariant through all species. In neither case, have other supporting kinetic studies such as rescue kinetics been performed. a-Glucosidases of G H family 31 have been shown to be inhibited by several glycosidase inhibitors including nojirimycin, deoxynojirimycin, gluconolactone and 20 acarbose, providing some insights into the structure of the positively charged transition state and anionic charge in the active site [5,31,64,65]. Also, substantial a K I E values upon kcat measured for the hydrolysis of p-chlorophenyl a-glucopyranoside and isomaltose indicate a substantially oxocarbenium ion-like transition state [32,66]. Computational modeling of the transition state structure for the hydrolysis of a -glucopyranosyl fluoride based on the a-tritium secondary and 1 4 C primary kinetic isotope effects suggests 70% cleavage of the bond between the anomeric center and the leaving group, and a half chair conformation of the substrate, clearly showing the significant oxocarbenium ion character [36]. In summary, a-glucosidases of G H family 31 have long been studied and significant mechanistic features have been revealed. However, the catalytic residue which is the key nucleophile in the double displacement mechanism adopted by retaining glycosidases is still to be clearly identified. 1.4.2. a-Xylosidases a-Xylosidases catalyze the hydrolysis of xyloglucan oligosaccharides, products resulting from the degradation o f xyloglucan. Xyloglucan is the widely distributed hemicellulose found in the primary cell wall o f plants. It contains a p-(l,4)-D-glucan backbone with a-( 1,6)-D-xylopyranosyl residues linked to about 75% of the backbone glucosyl residues [67-72]. Additional a-L-fbcosyl-(l,2)-P-D-galactosyl-(l,2)- branches linked to the 2-position of xylosyl units are also found, dependent on the species of plants. 21 Thus, the cooperative action on xyloglucan by xyloglucan endoglucanase, exo-p-glucosidase, a-fucosidase, and P-galactosidase is required to yield xyloglucan oligosaccharides [68-72]. In addition to the structural role, xyloglucan-derived oligosaccharides play the role of regulating auxin and acid pH-induced growth [73,74]. The a-xylosidase cleaves the a-xylosyl residue attached to the glucose residue farthest from the reducing end of the xyloglucan oligosaccharides. In plants, this action contributes to the maintenance of the appropriate level of xyloglucan in the organism since the removal of the xylosyl residues from the xyloglucan oligosaccharides by the a-xylosidase seems to be a necessary prerequisite for the oligosaccharide degradation [70-74]. In contrast to plant enzymes, the exact role and function of a-xylosidase in microorganisms are yet to be studied. A. B. Figure 1.6. Small molecule substrates for a-xylosidases: A. isoprimeverose; B. p-nitrophenyl a-D-xylopyranoside Two different types of a-xylosidases have been observed. Those from fungi and bacteria have high activity for simple xylosides such as />-nitrophenyl a-xylopyranoside and isoprimeverose (Figure 1.6), but little or relatively low activity for terminal xylosyl units from xyloglucan-derived oligosaccharides and a-glucosides [67,68]. The others, mostly plant a-xylosidases, hydrolyze xyloglucan oligosaccharides efficiently while the 22 activity for /?-nitrophenyl a-xylopyranoside and the disaccharide, isoprimeverose, is very low [70-72]. These plant enzymes also hydrolyze maltooligosaccharides as efficiently as xyloglucan oligosaccharides while the first type of enzymes do not. A recently cloned a-xylosidase from an archaeon, Sulfolobus solfataricus was shown to have similar substrate specificity to the first type [68]. Remarkably, all the a-xylosidases that have been sequenced belong to G H family 31 [67,68,70-72]. Two signature regions are well conserved through all these a-xylosidases (Figure 1.5), indicating that the double displacement mechanism would be operating. However, there has been no mechanistic study on this class of enzymes, reflecting its still nascent stage of research. 1.4.3. oc-l,4-Glucan Lyases 1,5 -anhydro-D- fructose Scheme 1.6. Degradation of o>l,4-glucans (starch or glycogen) by the action of a-l,4-glucan lyase a- l ,4-Glucan lyase (EC 4.2.2.13, GLase) was first discovered in Gracilariopsis sp. during the course of study on the starch degrading enzymes in the red algae [75]. GLase was one of two enzymes purified and initially it was thought to be an a-glucosidase. Upon the identification of the product, 1,5-anhydro-D-fructose, this enzyme 23 was renamed as oc-l,4-glucan lyase. GLase was first purified from Gracilariopsis sp. in 1993 [52] and, since then, two isozymes were purified from the same source and also three from fungi, Morchella costata, Morchella vulgaris and Peziza ostracoderma [75]. A l l these have been sequenced and cloned [75,76,77]. GLase acts on a-l,4-glucans such as starch and glycogen, as well as short maltooligosaccharides including maltose and maltotriose [75]. GLase is an exo-acting enzyme which specifically cleaves the oc-l,4-glycosidic bond in oc-l,4-glucans to yield a (3-elimination product from the non-reducing end sugar, leaving the reducing end sugar as free glucose (Scheme 1.6). The product, 1,5-anhydro-D-fructose (AnFru), has been detected in several other species, while GLases have been isolated and cloned only from red algae and fungi. Thus, AnFru has also been identified in Escherichia coli, rat liver and human leukemic cells [78,79,80], increasing the possibility o f the existence o f GLase in these organisms. raicrothecin Scheme 1.7. In vivo transformation of 1,5-anhydro-D-fructose: A . in bacteria and mammals, : 1. 1,5-anhydro-D-fructose reductase, 2. hexose kinase; B . in algae and fungi, 1. dehydratase However, its in vivo role is not clearly defined yet. In E. coli and rat liver, it has been found that 1,5-anhydro-D-fructose is reduced to 1,5-anhydro-D-glucitol by a specific 24 NADH-dependent reductase (Scheme 1.7). 1,5-Anhydroglucitol is possibly phosphorylated to 1,5-anhydroglucitol phosphate, which is the final product of this glycogenolysis in E. coli. Based on these findings, a third metabolic pathway of glycogen involving GLase and reductase has been suggested [78,79]. In red algae and fungi, a different pathway that GLase participates in has been suggested [75]. It was observed that AnFru is converted to microthecin which is a precursor of antibiotics by a 1,5-anhydro-D-fructose dehydratase [81], suggesting a protective mechanism. GLase, though catalyzing a p-elimination reaction, is different from polysaccharide lyase in many aspects. First of all, while substrates for polysaccharide lyases are uronic acid-containing polysaccharides, those for GLases are simple a-1,4-glucans. Secondly, GLase cleaves the glycosidic bond between the anomeric carbon of the non-reducing sugar and the glycosidic oxygen while polysaccharide lyase cleaves the bond between the glycosidic oxygen and C4 of the reducing sugar (Figure 1.7). Figure 1.7. Scissile bonds of substrates for lyases. Scissile bonds are designated by arrows: A. a-l,4-glucan lyase; B. polysaccharide lyase The location of the double bond generated by the reaction is also different. GLase produces a double bond between C I and C2 of the non-reducing sugar whereas polysaccharide lyase creates unsaturation between C4 and C5 of the reducing sugar. Amino acid sequence alignments also reveal striking differences [75]. The two A. B. 25 subfamilies of sequenced GLases, algal and fungal enzymes, have no amino acid sequence similarity with any known polysaccharide lyases. Instead, GLases display substantial amino acid sequence similarity with G H family 31 [75]. Remarkably, 23 -28% sequence similarity exists between them and a-glucosidases of G H family 31 including two signature regions (Figure 1.5). Therefore, the two subfamilies (algal and fungal) of GLases are assigned to G H family 31 and posted as such on the C A Z Y web site. Initial characterization has shown that GLase is a single, non-glycosylated polypeptide chain with molecular mass of 117 - 122 kDa and an optimum p H for activity of 3.8 - 4.1 for algal lyases and 6.5 for fungal lyase [52,75]. These p H optima are consistent with the acidic optimum of G H family 31. GLases have high affinity for starch, allowing use of a starch affinity column for purification, although no starch-binding domain homologous to any currently known carbohydrate binding domains has yet been identified [52,75]. GLase from algae has high specificity for the a- l ,4-glucan linkage. While linear maltooligosaccharides were completely degraded to 1,5-anhydro-D-fructose and glucose, a branched pentasaccharide was degraded to AnFru and branched tetrasaccharide, for example (Figure 1.8). GLase has virtually no activity toward other a - l , 3 - or a - l , 6 -linkages [82]. Likewise, GLase does not act on other sugars such as galactosides and mannosides. This high specificity for the a- l ,4-glucan linkage is consistent with the existence of several subsites, leading to higher activity for glucan polymers and maltooligosaccharides such as starch, glycogen, maltose and maltotriose than for synthetic substrates such as/>-nitrophenyl a-D-glucopyranoside [82,83]. 26 Figure 1.8. Degradation of a branched pentasaccharide by a-l,4-glucan lyase: the reaction stops after removing a terminal glucosyl unit, leaving 1,5-anhydrofructose and a tetrasaccharide Besides the initial characterization, there have been very few kinetic studies on this enzyme although the mechanistic features should be very interesting [83,84]. GLase is mechanistically unusual. It catalyzes formally a p-elimination as do polysaccharide lyses and, thus, GLase itself is a polysaccharide lyase. However, while other polysaccharide lyases act on substrates containing a uronic acid moiety, which activates the proton abstracted during the reaction, GLase acts on simple glucans containing no functional group to activate the proton on C2 which is abstracted. In addition, it seems that GLase activity does not depend on redox chemistry which might activate the proton in situ since no N A D , N A D P or metal dependence have been found. It is therefore intriguing to see how GLase can overcome the enormous activation barrier involved in abstracting a non-acidic, non-activated proton. It is also interesting that the amino acid sequence of GLase has significant similarity not with polysaccharide lyases which catalyze similar reactions, but with a-glucosidases of G H family 31 which catalyze 27 hydrolysis. This might have substantial implications in predicting a reaction mechanism of GLase. 1.5. Mechanism Based Inactivators in Mechanistic Studies of Enzymatic Glycosidic Bond Cleavage 1.5.1. Irreversible Inhibitors Reversible competitive inhibitors, usually transition state analogues, have been useful tools to show features of transition state structure, active site structure and even the protonation pattern o f the glycosidase reaction [28]. However, they provide little insight into active site structure, thus other direct methods have been devised for this purpose including affinity labels (Figure 1.9). Affinity labels for glycosidases consist o f two parts, a sugar moiety that provides recognition for the active site o f glycosidases and a reactive group that can form an adduct with reactive active site groups. Br A. B. C. Figure 1.9. Affinity labels: A. epoxy-alkyl (3-D-glucoside; B. N-bromoacetyl p-D-glucosylamine; C. a-C-bromo-ketone-mannoside 28 However, due to the inherent reactivity of this functional group, affinity labels have often labeled not only the desired catalytic residue but also other residues close to the active site and beyond. Therefore, affinity labels have been best used in combination with site directed mutagenesis where the importance of the residue tagged can be further evaluated [85-93]. A n improved version of label is the mechanism-based inactivator. These reagents have the same kind of affinity for enzymes, but are unreactive until they react with the enzyme, which unleashes reactive grouping [31]. The best known mechanism-based inactivators for glycosidases are triazenes and conduritol epoxides (Figure 1.10 ). B. A. Figure 1.10. Mechanism-based inctivators: A. a glycosylmethyltriazene; B. conduritol B epoxide Glycosylmethyltriazenes decompose upon enzymatic protonation to generate the highly reactive carbenium ion which reacts with active site groups [94]. Conduritol B epoxide would be arguably one of the most versatile-mechanism based reagents [5,31]. The inactivation occurs by formation of a stable covalent intermediate upon opening of the epoxide ring via simultaneous action of the general acid catalyst and nucleophilic attack (Scheme 1.8). This reagent is a symmetrical molecule along the axis bisecting the epoxide ring and, therefore, i f it is flipped over the symmetry axis, only the epoxide oxygen wi l l change its orientation (Figure 1.10). Consequently, 29 conduritol B epoxide is able to mimic both retaining a- and (3-glycosides depending on the orientation of the epoxide oxygen and, therefore, it acts on both a- and p-glycosidases [59-63,95]. HCI X H O r H O ^ V ^ ^ y f° ° H O - V ^ V A ° HO-A^-^^-Ss! a-glycosidase HO-X*—•^*V)H O h ^ OH Scheme 1.8. Inactivation of an a-glycosidase by the reaction with conduritol B epoxide 1.5.2. Fluorosugars Some highly sophisticated mechanism-based inactivators for retaining glycosidases are the activated fluorosugars [8,10,19,22,23]. The central strategy is to destabilize the highly positively charged transition states of the glycosidase reaction. The reaction mechanism involves a covalent glycosyl enzyme intermediate which is formed and hydrolyzed through oxocarbenium ion like transition states. The incorporation of a highly electron-withdrawing fluorine adjacent to the site of positive charge development should slow down both steps, i.e. the formation and hydrolysis of the intermediate, by destabilizing both transition states. However, a good leaving group such as fluoride or a 2,4-dinitrophenyl group at the anomeric center ensures that the formation of the intermediate (glycosylation) is faster than its hydrolysis (deglycosylation). As a result, the glycosyl-enzyme intermediate will accumulate (Scheme 1.9). 30 H f l " ^ o ^ - A ' H O i K 2 . F Y M V ^ H 3 I H O I pHO . t ^ , I" O H ^ " F M U O H V H \ C y O O ^ O - ^ H O ^ O H jwvJv/wv/w- J W W A T W U W J W V \ A / \ / W W » k 2 » k 3 Scheme 1.9. Inactivation of an a-glycosidase by the action of a fluorosugar: An example containing a 5-fluorosugar is illustrated. The first such fluorosugars to be employed were the 2-deoxy-2-fluoro-p-D-glycopyranosyl fluorides and 2,4-dinitrophenyl 2-deoxy-2-fluoro-P-D-glycopyranosides (Figure 1.11 A ) [96,97]. Numerous studies have now been performed utilizing 2-fluorosugars on (3-glycosidases [22] and the intermediates shown to be genuine a-linked glycosyl-enzyme intermediates by N M R spectrometry and X-ray crystallography [22,98]. The intermediate is kinetically competent since it slowly turns over, especially in the presence of suitable acceptors. This reasonably stable intermediate has allowed the characterization of the intermediate, often leading to the identification of the catalytic nucleophile [8,10,19,22]. 2-Fluorosugars have rarely proved useful in studying oc-glycosidases. Instead, they behaved like slow substrates with a rate limiting glycosylation step, thus, no intermediate accumulated. A n alternative to 2-fluorosugars was developed by incorporating an extra fluorine at C2 to further slow deglycosylation (Figure 1.1 IB) [99]. Although the additional fluorine further slows both steps, the incorporation of a far better leaving group such as a 2,4,6-trinitrophenyl group facilitates the first step, allowing the accumulation of the intermediate. 31 F Figure 1.11. Fluorosugar inactivators: A. 2-fluorosugars; B . 2,2-difluorosugar; C. 5-fluorosugar The other strategy was to incorporate fluorine in place of hydrogen at the 5-position instead (Figure 1.11C) [18]. A s the endocyclic oxygen is supposed to bear the majority of the positive charge at the oxocarbenium ion-like transition state, fluorine at C5 is also expected to substantially destabilize the transition states. Modeling studies and mechanistic studies have indicated that the greatest difference in the amount of charge build-up between the ground state and the oxocarbenium ion-like transition state occurs at 0 5 , the endocyclic oxygen, rather than at C I , the anomeric center [100-102]. Therefore, destabilizing effects by fluorine substitution at C5 might be expected to be greater. Indeed, 5-fluoro-a-D-glycosyl fluorides has been successfully employed in inactivating a-glycosidases and trapping the intermediate, often leading to the identification of catalytic nucleophiles [18,23,51]. 5-Fluoro-p-glycosyl fluorides were also synthesized and shown to work well for inactivating p-glycosidases [18], thus, proving the generality of the 5-fluorosugar approach in contrast to what was found for 2-deoxy-2-fluorosugars. 32 1.6. Aims of Study Glycoside family 31 contains both glycosidases and lyases cleaving the glycosidic bond in two different ways, nucleophilic substitution and p-elimination, respectively. While the two reactions are significantly different, enzymes in the same sequence family are expected to utilize the same mechanistic machinery. Thus, in this thesis, the detailed reaction mechanism of enzymes in G H family 31 w i l l be thoroughly scrutinized including representative enzymes for each reaction, a-glucosidase and a-l ,4-glucan lyase. In addition, an unknown homologous gene product from Escherichia coli w i l l be studied to provide further insights into mechanisms in this sequence family. 33 Chapter II. Identification of the Catalytic Nucleophile of Glycoside Hydrolase Family 31 a-Glucosidase from Aspergillus niger * * A version of this chapter has been published: Lee, S. S., He, S. and Withers, S. G. (2001) Identification of the Catalytic Nucleophile of the Family 31 a-Glucosidase from Aspergillus niger via Trapping of a 5-Fluoroglucosyl Enzyme Intermediate. Biochemical Journal 359: 381 - 386 34 2.1. Background Aspergillus niger a-glucosidase ( E C 3.2.1.20) is an exo-acting retaining glycosidase that releases a-glucose from the non-reducing end of starch and malto-oligosaccharides [50,57]. It is a glycoprotein containing 25.5 - 27.6% carbohydrates which contain mostly mannosyl units and show a novel structure containing a -D-galactofuranosyl linkages [50,58,103]. The enzyme contains two subunits, PI and P2 of combined molecular mass 125,000 D a [57]. The full amino acid sequence is available [104, accession number Swiss-Prot P56526], and from this, it has been assigned to glycoside hydrolase (GH) family 31. a-Glucosidases in G H family 31 are assigned to type II or III a-glucosidases according to substrate specificities and A. niger a -glucosidase has shown the substrate specificity corresponding to type III [49]. Thus, it has higher activity for maltooligosaccharides and soluble starch than for synthetic substrates such as aryl glucosides [57,58]. Mechanistic features of this enzyme have been relatively well characterized. A s a retaining glycosidase, this enzyme utilizes a double displacement mechanism in which a p-glucosyl-enzyme intermediate is formed and hydrolyzed. A water soluble carbodiimide inactivated this enzyme, suggesting the essential role of carboxylic acid containing residues and two apparent pKa (3.2 and 6.4) values have been measured and indicated protonated and deprotonated carboxylic acids as have been found in other retaining glycosidases [105]. Oxocarbenium ion character at the transition states has also been implicated by the measurement of a large a-deuterium secondary kinetic isotope effect (1.16) on kcat for hydrolysis of isomaltose [66]. The identity of the catalytic nucleophile 35 was also suggested on the basis of the use of conduritol B epoxide (CBE) [62]. However, C B E is a reagent that does not form a kinetically competent intermediate and has been shown in several cases to label other residues than the catalytic nucleophile [8,19,106,107]. Thus, the identity of the catalytic nucleophile is still ambiguous. Specific Aims of the Chapter In addition to a-glucosidases, a-xylosidases and an unusual class of enzyme, the a-l,4-glucan lyases are found in G H family 31. In order to understand and compare these enzymes, it is therefore important to unequivocally identify the catalytic nucleophile of one of the family 31 a-glucosidases with a more reliable probe. 2.2. Kinetic Evaluation oi Aspergillus niger a-Glucosidase Since A. niger a-glucosidase is a type III enzyme, its activity for the synthetic substrate, /?-nitrophenyl a-D-glucopyranoside (PNPaGlc) is reported to be very low. The bell-shaped p H dependent activity profile shows that the optimum p H of A. niger a-glucosidase is p H 4 - 4.5 [105]. Thus, hydrolysis of P N P a G l c cannot be monitored at 400 nm at an optimal p H since the phenolate product (p^Ca = 7.15) is almost entirely protonated and therefore colorless. I f the p H is raised to improve the sensitivity of the assays, the activity of the enzyme is too low even at p H 6. In combination with the intrinsic low activity of this enzyme for P N P a G l c , it becomes impossible to perform a continuous assay with this substrate. However, at a wavelength o f 360 nm, there is a 36 substantial difference in molar extinction coefficient (1.88 ± 0.01 IvT'cm"1) between substrate and product at p H 4.5, allowing a continuous assay. Kinetic parameters for the hydrolysis of P N P a G l c of Km = 0.31 (± 0.02) m M and kcal= 2.3 (± 0.06) s"1, respectively, were determined in this way. These are consistent with values in the literature (Table 2.1). Table 2.1. Michaelis-Menten Parameters for the Hydrolysis of Substrates by A niger a-Glucosidase substrate km (s"1) (mM) kcat/Km ( s 'mM" 1 ) maltose 144 0.75 192 maltotriose 181 0.69 262 maltotetraose 193 1.1 175 maltopentaose 147 1.9 77.4 maltodextrin (DP = 17) 127 11 11.5 soluble starch 163 4.3 37.9 phenyl a-maltoside 181 0.87 208 phenyl oc-D-glucoside 1.83 0.34 5.38 p-nitrophenyl a-D-glucoside 4.3; 0.7; 6.14; 2 . 3 ± 0 . 0 6 a 0.31 ± 0 . 0 2 a 7.4 a 2,4-dinitrophenyl 178 ± 6 a 1.7 + 0 . 1 a 105 a a-D-glucoside a-D-glucosyl fluoride 3 6 4 ± 1 2 a 3.0 ± 0 .2 a 121 a a Measured in this study; others are excerpted from references 57 and 58. However, the highly activated substrates 2,4-dinitrophenyl a-gtucopyranoside and a-glucopyranosyl fluoride both provide more convenient assay conditions. Since the pKa 37 of the leaving group (2,4-dinitrophenol) of the former is 3.96, the reaction is readily monitored at the optimal p H of 4.5 at a wavelength of 400 nm. The latter substrate generates fluoride which can be monitored using a fluoride-selective electrode. Both substrates display fairly high kcat (178 ± 6 s"1 and 364 ± 12 s"1, respectively) and moderate Km (1.7 ± 0.1 m M and 3.0 ± 0.2 m M ) values, reaching the same level as those for natural substrates (Table 2.1). Some of the glycosidase inhibitors that have shown potency for a-glucosidases before have been tested with A. niger a-glucosidase (Table 2.2). Acarbose and 1-deoxynojirimycin have been shown previously to be highly potent inhibitors of a-glucosidases of G H family 31 [64,65], with K; values in the micromolar range (6.6 ± 0.2 p M and 1.5 ± 0.1 p M , respectively). However, of interest is the poor inhibition by hydroximinogluconolactam with a K value of only 1.7 ± 0.1 m M . In contrast, this compound is a tight binding inhibitor of the G H family 13 yeast a-glucosidase with a K of 2.9 uM[108] . Table 2.2. The Structure of Inhibitors and K\ values for A niger a-Glucosidase Inhibitors K\ (pM) Acarbose 6.6 ± 0.2 1 -Deoxynojirimycin 1.5 ± 0 . 1 Hydroximinogluconolactam 1.7 (± 0.1) x 10 3 38 Compounds of this class are thought to only inhibit arcft'-protonators with their acid residue residing in an anft'-relationship to the substrate endocyclic oxygen based on studies on (3-glycosidases [28,109]. Since G H family 13 enzymes are known to be anti-protonators, this raises the possibility that G H family 31 enzymes are 5y«-protonators. 2.3. Kinetic Evaluation of the Interaction of Aspergillus niger a-Glucosidase with 5-Fluoro-a-D-Glucopyranosyl Fluoride 5-Fluoro-a-D-glucopyranosyl fluoride (5FaGlcF, Figure 2.1) has proved to be a useful reagent for trapping the intermediate formed by retaining a-glucosidases [18,23,51]. The highly electronegative fluorine at C5 destabilizes the oxocarbenium ion-like transition states, slowing both the formation and the hydrolysis of the intermediate. However, the presence of a good leaving group, fluoride, at the anomeric center ensures that the formation of the intermediate (glycosylation) is faster than its hydrolysis (deglycosylation). A s a result, the glucosyl-enzyme intermediate w i l l accumulate, in some cases resulting in inactivation of the enzyme i f turnover is very slow, as described earlier [23]. Figure 2.1. 5-Fluoro-a-D-glucopyranosyl fluoride 39 This approach involving 5-fluoroglycosides has been employed previously to identify the catalytic nucleophiles o f a-glucosidases in family 13 [51] as well as a-galactosidases in family 27 [110] and a-mannosidases in family 38 [111,112]. When 5 F a G l c F was incubated with Aspergillus niger a-glucosidase and aliquots were removed at time intervals, no time-dependent inactivation was observed. Instead, a substantially reduced activity was measured at time zero, and this did not change with time. This phenomenon indicates that turnover occurs at rates that are comparable to those of intermediate formation and, thus, the outcome is steady state accumulation of the intermediate. The compound therefore appears to be acting as a tight binding, reversible inhibitor. 9 11 13 1/[S] - 0 . 5 Figure 2.2. Lineweaver-Burke plot for the inhibition kinetics of 5-fluoro-a-D-glucopyranosyl fluoride on A. niger a-glucosidase: • , [5FaGlcF] = 0 uM; • , [5FaGlcF] = 5 uM; • , [5FaGlcF] = 10 uM; • , [5FaGlcF] = 20 uM 40 5FocGlcF was therefore analyzed as a competitive inhibitor in a separate set of experiments, yielding an apparent K, or KC of 2.5 (± 0.13) u M (Figure 2.2). However, considering that the Km value of the parent compound, aGlcF , is 3 m M and that the incorporation of an axial fluorine on C5 should not create a favorable situation for binding, this tight binding would be quite remarkable i f it reflected simple, reversible association. Similar behavior has been seen previously with several other 5-fluoro glycosyl fluorides and glycosidases, and shown to be due to steady-state accumulation of a 5-fluoroglycosyl-enzyme intermediate, due to the relatively slow turnover, yet rapid formation, of the intermediate [51,110,112]. The apparent K, does not measure the true inhibitor-enzyme affinity. Instead, it includes contributions from both the binding event (Ki) and the chemical reactions which are, in this case, the steady state formation (ki) and turnover (k^) of the covalent glycosyl enzyme intermediate according to the following expression [31,113], k ' , r r k 2 k 3 E + I - " EI » E-I • E + P k-i EI: Michaelis complex E-I: covalent intermediate l + fo/fo Thus, the apparent K, represents a minimum value of the true Kt. 41 0 1000 2000 3000 4000 5000 6000 7000 8000 Time ( s e c o n d ) Figure 2.3. Hydrolysis of 5-fluoro-a-D-glucopyranosyl fluoride catalyzed by A. niger a-glucosidase. Concentrations of 5FaGlcF are as designated on each profile. The inhibition behavior was further probed by direct monitoring of turnover using a fluoride ion-specific electrode. Thus, 5FocGlcF was tested as a substrate at various concentrations. The enzymatic reaction continued at a constant rate until almost all the substrate was consumed when the rate suddenly dropped, this same rate being seen at a range of concentrations of 5-fluoro-a-D-glucopyranosyl fluoride (Figure 2.3). Both of these observations indicate a very low Km value and are consistent with the low KC value measured since the Km value for this compound as a substrate must equal its Kt' value. The turnover number, kcat, was determined from the slope of these plots of fluoride release versus time, and found to be 0.055 (± 0.001) sec"1. Unfortunately, a Km value could not be directly determined from this experiment due to the lack of sensitivity of the fluoride electrode towards the very low substrate concentrations (1-5 pM) needed for its 42 determination. However, the previous K\ value of 2.5 u M is a good estimate of the Km value. The kcat value is comparable to those found for 5-fiuoroglycosyl fluoride with other enzymes. Thus the kcaX for 5FaGlcF with yeast a-glucosidase was 0.11 ( ± 0 . 0 1 ) s"1 [51] and that for 5-fluoro-ri-L-gulopyranosyl fluoride (epimer of the corresponding mannosyl compound at C5) with bovine kidney lysosomal a-mannosidase was 0.022 s"1 [112]. 2.4. Identification of the Catalytic Nucleophile oi Aspergillus niger a-Glucosidase Steady state accumulation of the covalent glycosyl enzyme intermediate indicates that this mechanism-based reagent does not operate as an inactivator since it turns over slowly. Interestingly, 5-fluoroglycosyl fluorides have often shown this behavior on both a - and P-glycosidases [110-112,114]. Although turnover occurs, it is usually slow enough for the intermediate to be isolated and analyzed. In many cases, this intermediate could be digested with a protease and the resultant peptide mixture analyzed by a combination of liquid chromatography/mass spectrometry ( L C / M S ) [51,112,114]. Comparative mapping of labeled digests along with a digest of a control sample to look for differences in peptide masses between the two spectra (control and label) often lead to the identification of a labeled peptide which has a higher mass than a control peptide, followed by tandem mass spectrometry ( M S / M S ) to identify the labeled residue [19]. A. niger a-glucosidase was thus incubated with 5 F a G l c F to form an accumulated intermediate, then immediately subjected to peptic digestion at p H 2. A sample of 43 unlabeled a-glucosidase was also digested by pepsin under the same conditions. The peptic digests from the labeled and unlabeled enzymes were loaded onto the microbore reverse phase-high performance liquid chromatography (RP-HPLC) connected to the electrospray ionization mass spectrometer (ESI-MS), and total ion chromatograms (TIC) for each sample were obtained. £ 0.5 10 20 30 40 T i m e (m in) 50 60 Figure 2.4. Total ion chromatogram of the peptic digest of labeled A. niger a-glucosidase. Fig 2.4 shows the TIC of the labeled enzyme: that of the unlabeled enzyme is practically indistinguishable from this, and is not shown. In order to identify the labeled peptide, the masses of the peptides under each peak in the labeled sample were compared with those of the peptides from the unlabeled sample in the corresponding region of the TIC. The masses of the peptides from the two samples were identical with one exception. A peptide of mass 1011 Da eluting at 27.76 minutes was observed only in the TIC of the labeled sample (Fig 2.5). If this is the labeled peptide of interest, then a peptide of mass 44 -830 might be expected in the TIC of the unlabeled sample, this being the mass difference between the peptide of mass 1011 and the 5-fluoro-glucosyl label of mass 181. - 0.5 o > o 443.5 A 843.5 705.5 573.0 hjj«iln ni n I 400 600 800 1000 1200 1400 • 0.5 > m 5 443.5 B 1011.0 843.5 705.5 573.0 l i k i j l u i u J i 400 600 800 1000 1200 1400 m/z, amu Figure 2.5. Mass spectrum of the unlabeled (A) and labeled (B) A. niger a-glucosidase peptic digests taken at 27.63 min and 26.76 min, respectively. Arrow indicates the peptide unique to the peptic digest of labeled enzyme. Singly charged fragments (z = 1). Unfortunately, no such peptide was observed (Fig 2.5), possibly indicating that the unlabeled peptide is susceptible to further peptic digestion and has been converted to smaller fragments. Such differences in proteolytic cleavage as a consequence of the presence of a sugar residue have been demonstrated previously [115]. 45 830. H 2N— W' C-N—Y—C II H » O O 464. U N — M - c H " O 235.Q9 f a -si-•C-N—E— c II H " O O OH 159.09 350.15 465.18 655.25 812.5 WYDMSE + 5FGIu 1011.0 (ft c = 0.5 > or 100 200 300 400 500 600 700 800 900 1000 m/z, amu Figure 2.6. The daughter ion spectrum (MS/MS, z = 1) of m/z 1011 isolated from the labeled digest along with an interpretation of the spectrum. To further investigate the possibility of this being the labeled peptide and assign the labeled amino acid residue, the fraction containing the putative labeled peptide was purified by RP-HPLC and subjected to electrospray ionization tandem mass spectrometry (ESI MS/MS) fragmentation analysis. The daughter ion spectrum arising from the peptide of m/z 1011 is shown in Fig 2.6, along with an interpretation of the spectrum. This spectrum shows the fragment of m/z 830, which was absent in the TIC of the unlabeled enzyme and corresponds to the mass of the parent ion (1011 minus the mass of label 181), confirming that this peptide of mass 1011 Da is the labeled peptide of interest. 46 Analysis of the fragmentation pattern revealed that the peptide sequence was W Y D M S E as shown in the above spectrum. Two prominent peaks are observed that arise from fragments containing the sugar moiety. One of these, 776.5 Da, corresponds to W Y D M + 5 F G l c . The other, 646 Da, corresponds to either W Y D + 5 F G l c or D M S E + 5 F G l c . Within the peptides W Y D M and W Y D either the tyrosine residue or the aspartate could be the catalytic nucleophile. However, the tyrosine is not conserved while the aspartate is invariant. Should the labeled fragments be W Y D M and D M S E only the aspartic acid is a feasible candidate. Both analyses point to the catalytic nucleophile being an aspartic acid corresponding to Asp 224 of the P2 subunit of the enzyme. Previous labeling studies with other family 31 a-glucosidases using conduritol B epoxide have involved labeling of human intestinal sucrase-isomaltase [20], human lysosomal a-glucosidase [60], sugar beet a-glucosidase [61], A. niger a-glucosidase [62] and Schizosaccharomyces pombe a-glucosidase [63]. A l l these studies indicated that the nucleophile is an invariant aspartate residue corresponding to Asp 224 of A. niger a-glucosidase. However, the use of conduritol epoxides could well lead to the misassignment of the labeled residue because of the relatively high reactivity of its epoxide, coupled with the absence of a C6 hydroxymethyl group. A s described earlier, the 6 -OH contributes substantially to the recognition of the substrate and the stabilization of the transition state (Figure 1.6). Thus, when this group is missing, binding can occur in alternate modes, leading to nonspecific labeling. Indeed the use of conduritol B epoxide has previously resulted in misassignments of the catalytic nucleophiles of E. coli (3-galactosidase [106] and human lysosomal (3-glucosidase [107], amongst others. B y comparison, labeling with the mechanism based reagent 5FaGlcF , which forms a 47 catalytically relevant and kinetically competent intermediate evidenced by its turnover, is more reliable and unambiguously confirms the identity of the catalytic nucleophile. 2.5. Conclusion A long inconclusive catalytic nucleophile of an a-glucosidase of glycoside hydrolase family 31 has been identified by use of the mechanism based reagent, 5FaGlcF . The 5-fluorosugar approach has proved itself to be a powerful tool for probing the active sites of retaining a-glucosidases, especially when combined with the analytical technology of tandem mass spectrometry. Upon "reaction with A. niger a-glucosidase, 5FaGlcF formed a catalytically competent glycosyl-enzyme intermediate, allowing the identification of the catalytic nucleophile as Asp-224 within the sequence W Y D M S E . This study is an essential step for setting the stage for the next step for the mechanistic study of new enzymes of G H family 31, an a-l,4-glucan lyase, catalyzing a (3-elimination, and a product of an open reading frame of Escherichia coli. 48 Chapter III. Detailed Mechanistic Study on an a-l,4-Glucan Lyase from Gracilariopsis sp.* * Versions of this chapter have been published: Lee, S. S., Yu, S. and Withers, S. G. (2002) oc-Glucan Lyase Performs a Trans-elimination Reaction via a Nucleophilic Displacement Followed by a Syn-elimination, Journal of the American Chemical Society, 124, 4948 - 4949; Lee, S. S., Yu, S. and Withers, S. G. (2003) Detailed Dissection of a New Mechanism for Glycoside Cleavage: the cc-l,4-Glucan Lyase, Biochemistry, 42, 13081 - 13090 49 3.1. Background a-l ,4-Glucan lyase (EC 4.2.2.13, GLase) degrades a-l,4-glucans and maltooligosaccharides via a nonhydrolytic, ^-elimination process to release 1,5-D-anhydrofructose from the non-reducing end. GLase is a polysaccharide lyase, substrates of which are mainly a-l,4-glucan polysaccharides. However, this enzyme is different from other polysaccharide lyases in many respects. Firstly, GLases work on simple a-l,4-glucans and cleave the bond between C I ' and O I ' , abstracting the C 2 ' proton and generating a double bond between C I ' and C 2 ' of the non-reducing end sugar (Fig 1.7) while other polysaccharide lyases cleave the bond between C4 and O I ' of uronic acid-containing polymers to generate a double bond between C4 and C5 of the reducing end sugar by abstracting the C5 proton and facilitating elimination (Figure 1.7). The C6 carboxyl group of the uronic acid moiety plays a key role in the reaction mechanism of polysaccharide lyases by increasing the acidity o f H5 . In contrast, substrates of GLase are simple glucans and catalyze the abstraction of a proton without the aid of an adjacent electron-withdrawing carbonyl. Since GLase does not depend on redox chemistry [75], it is intriguing to speculate on the strategy employed for abstraction of a non-activated proton. Secondly, full amino acid sequence alignment reveals significant similarity (23 ~ 28%) with retaining oc-glycosidases from glycoside hydrolase family 31, with complete conservation of the residue previously identified as the nucleophile in these glycosidases (Figure 1.5), leading to the assignment of GLase to G H family 31. A l l of these above 50 point to an unusual mechanism for GLases, having more in common with that of retaining glycoside hydrolases than of regular polysaccharide lyases. Specific Aims of This Chapter This chapter focuses on the investigation of the reaction mechanism of an a-1,4-glucan lyase in detail to provide clear and direct evidence that the reaction mechanism of this enzyme is highly relevant to the general mechanistic picture of G H family 31. 3.2. Previous Studies on the Mechanism of a-l,4-Glucan Lyase Previous studies on the mechanism of a-l,4-glucan lyases have been very limited, consisting essentially of initial kinetic evaluations [83,84]. However, these studies still give some insights into the action of GLase. GLase was, first of all, shown to be inhibited by acarbose and 1-deoxynojirimycin [82,84]. Inhibition of glycosidases by these compounds has been considered as evidence for oxocarbenium ion character at the transition state and of anionic groups in the active site. Carboxylic acid groups in the active site are suggested by the finding that a water soluble carbodiimide, l-ethyl-3-(3-dimethylaminopropyl)-carbodiimide, inactivates GLase [83,84]. Since carbodiimides are known to be specific for carboxylic acid functionalities, this indicates the involvement of carboxylic acid groups in the reaction mechanism. 51 Table 3.1. Michaelis-Menten Parameters for the Hydrolysis of Maltooligosaccharides and Polysaccharides by Gracilariopsis a-1,4-Glucan Lyase a Average Chain „ , A „ , , _K U . + T +U K™ (mM) &cat (s ) substrate Length Maltose 2.3 + 0.1 56 + 0.3 Maltotriose 0.4 + 0.1 60 + 4 Maltotetraose 0.1+0.01 45 + 2 Maltopentaose 0.1+0.03 42 + 5 Maltohexaose 0.1+0.01 53 + 3 Maltoheptaose 0.1+0.4 50 + 0.1 Rabbit Liver Glycogen 16 0.03 ± 0.05 60 + 3 Barley Amylopectin 20 0.03 ± 0.08 89+13 Potato Amylopectin 24 0.03 ± 0.03 100+14 a Reference 83 Values of kcat and Km have been measured for a series of maltooligosaccharides [83]. It was shown that kcat values were virtually identical for all substrates whereas Km values decreased as the number of glucose units increased up to four (maltotetraose) and beyond that, remained identical (Table 3.1). This indicates the existence of four subsites which have favorable binding interactions with substrates. It is therefore likely that the subsite substrate specificity of GLases is very similar to that of type III a-glucosidases. These findings suggest a similarity of GLases and G H family 31 a-glucosidases, but the difference in mechanism remains to be explained. 52 3.3. Probable Mechanisms 3.3.1. Enzymatic Elimination Reactions A B C Figure 3.1. Enzymatic Elimination: A. enolase superfamily; B. N-acetylneuraminate lyase superfamily; C. crotonase A s described earlier, polysaccharide lyases that carry out eliminative glycosidic bond cleavage exploit the presence of an activated proton that is acidified by the adjacent carboxylic acid group. The general base residue that is responsible for abstracting the proton is often an oxygen atom in polysaccharide lyases. Since the proton is transferred from a less electronegative carbon atom to a more electronegative oxygen atom, the aid of some other acidifying group is absolutely necessary for the catalysis. The formation of this oxyanion intermediate is followed by (3-elimination. This strategy is adopted by most enzymes catalyzing elimination reactions. The enolase superfamily is among the largest 53 of the groups of enzymes performing a p-elimination reaction. Abstraction of a proton activated by an adjacent carboxylic acid group, generates an anionic enolate intermediate which subsequently undergoes either anti- or svrc-elimination (Figure 3.1. A ) [116]. Dehydratases in the N-acetylneuraminate lyase superfamily utilize a slightly different, but identical in principle, strategy in activating the proton. The P-elimination is initiated by abstracting a proton that is activated by a protonated Schiff base (Figure 3.1 B) [116]. Crotonase also catalyzes a p-elimination involving an oxyanion intermediate (Figure 3.1 C), though in this case the a-proton of a thioester is abstracted [117]. These examples, without exception, utilize an anionic mechanism ( E l c B ) with the aid of an electron withdrawing functionality in the substrate. There have been only very few examples of mechanisms of enzymatic elimination reactions other than the E l c B type [118,119]. However, only superficial descriptions of them including the product analysis have been made and these have failed to draw the attention of the research community. 3.3.2. Possible Mechanisms To define the reaction mechanism of a new enzyme, a couple of things should be performed first. One should look into existing mechanisms for similar reactions and also search databases for sequence-based relatives to see what mechanisms they adopt. Based on these, predictions about the mechanism could be made. The reaction catalyzed by a-1,4-glucan lyase is an overall ^^-e l imina t ion . Whether syn- or arcri-elimination is involved, the mechanism ordinarily utilized for an elimination reaction in nature involves 54 anionic transition states and intermediates. This requires the presence of an electron withdrawing group to activate the proton and to stabilize anionic transition states and intermediates. However, substrates of GLase have no electron-withdrawing groups and the active site environment of GLase is likely to be anionic. Therefore, a direct E l c B type reaction is unlikely. Further the sequence similarity with glycosidases also argues for a cationic transition state rather than an E l c B type mechanism [75]. Two possible mechanisms are proposed. A s the overall reaction is an an^'-elimination, the reaction could proceed via direct a«ri-elimination reaction (Scheme 3.1). This is likely to be stereoelectronically favored since the cleaved bond and the newly generated electron pair are antiperiplanar to each other thereby allowing favorable interaction between the electron pair and the anti-bonding orbital of the cleaving bond. Scheme 3.1. A direct a«//-elimination mechanism Such as mechanism could be either concerted or stepwise. However, a completely synchronous concerted mechanism (E2) is unlikely since it does not involve any cationic species. A fully stepwise mechanism ( E l ) is also unlikely since this would require a highly unstable glycosyl cation intermediate. Thus, i f a direct a«ri-elimination operates, 55 the departure of the leaving group should precede the abstraction of the proton, generating an asymmetric transition state in which glycosyl cationic character prevails. Scheme 3.2. A proposed GLase mechanism involving a covalent glycosyl-enzyme intermediate The other possible mechanism by analogy with retaining a-glucosidases would involve the formation of a covalent glycosyl-enzyme intermediate via an oxocarbenium 56 ion like transition state, followed by a syn-elimination (Scheme 3.2). This two-step mechanism is analogous to the double displacement mechanism of retaining glycosidases. In the first step, a combination of nucleophilic attack and general acid catalysis would operate to form a covalent glycosyl-enzyme intermediate, providing a better leaving group for an upcoming (3-elimination reaction. The elimination step would follow, involving the same kind of asymmetric cationic transition state as described earlier for the direct elimination. This mechanism would be overall energetically more favorable than the direct elimination reaction since the poor leaving group is expelled by nucleophilic substitution in the first step and, thus, the second step benefits from the presence of a good leaving group. 3.4. The Reaction of the a-l,4-Glucan Lyase from Gracilariopsis sp. with Various Substrates 3.4.1. Commercially Available Substrates Throughout this study, a-l,4-glucan lyase with a molecular mass of 117,030 Da from Gracilariopsis sp. has been used. This enzyme, the first GLase to be separated and cloned [75] is a generous gift from Dr. Shukun Y u of Danisco Innovation, Denmark. Reliable and reproducible substrates are essential for performing mechanistic studies. Due to its similar substrate specificity to type III a-glucosidases, GLase has very low activity for a synthetic substrate such as the convenient substrate, /?-nitrophenyl a -D-57 glucopyranoside (PNPaGlc). While this substrate had reasonable kca, and Km values with A. niger a-glucosidase (2.3 s"1 and 0.31 mM, respectively), it turned out to be a relatively poor substrate for GLase, with kcat = 0.40 s"1 and Km-2A mM at pH 6.0. Thus, PNPaGlc was not suitable for routine kinetic experiments. A previous study suggested that a modification of the traditional dinitrosalicylate (DNS) reagent method was useful [120]. In the DNS method, a mixture of the DNS reagent and a reaction mixture containing reducing sugar released by the enzyme are boiled to generate the color that is then to be detected spectrophotometrically. In contrast, the modified method for monitoring GLase does not require boiling of the mixture. Upon standing at room temperature, the solution develops a color spontaneously without heating. It is not certain why this happens but is presumably related to the structure of the unsaturated sugar product, 1,5-anhydro-D-fructose. However, this method requires a stopped assay, which is cumbersome and the sensitivity of the assay is unsatisfactory. Agrobacterium p-Glucosidase detected by the UV spectrometer Figure 3.2. A coupled continuous assay usingp-nitrophenyl (i-maltoside as a substrate 58 Thus, another continuous, coupled assay was devised. It has been reported that the activity of GLase for a- and p-maltosides was almost identical [84]. The new assay method was inspired by this observation. If p-nitrophenyl (3-maltoside (PNPpMal) underwent reaction by GLase, /?-nitrophenyl p-D-glucopyranoside (PNPpGlc) would be produced along with 1,5-anhydro-D-fructose. This P N P P G l c is not a substrate for GLase but is readily cleaved by p-glucosidases while the original P N P p M a l would not serve as an initial substrate for the latter. Therefore, a (3-glucosidase could be used as a secondary enzyme in the coupled assay of GLase (Figure 3.2). Agrobacterium sp. p-glucosidase was chosen as the secondary enzyme and kcat and Km values were measured to be 10.5 (± 0.2) s"1 and 0.048 (± 0.002) m M at p H 6.0, respectively. This kcat value is 4 - 5 fold lower than that of maltooligosaccharides (Table 3.1) while the Km value is forty times lower than that of maltose and six fold lower than that of maltotriose. The reduced turnover may be attributed to the P-linked p-nitrophenyl group while the improved binding of P N P P M a l may be attributed to the hydrophobic interactions with the phenyl ring, as observed often in glycosidases. A potential alternative substrate was oc-D-glucopyranosyl fluoride (aGlcF) . This substrate has an activated aglycone leaving group but is, nevertheless, reasonably stable. Kinetic parameters have been measured and kcaX and Km values shown to be 505 s"1 and 27.9 m M , respectively. The kcat value is the largest ever measured for GLases but the Km is also quite high. Although this is promising, a G l c F was not used as a routine substrate for several reasons. Firstly, the fluoride selective electrode used to monitor reaction is not sufficiently sensitive at low concentration. Secondly, the Km value is too high for routine analysis where saturation is required. 59 Consequently, the coupled assay involving P N P p M a l was used for kinetic evaluations in the initial phases of the work. 3.4.2. Synthetic A r y l Glycoside Substrates 3.4.2.1. Synthesis. A) Background. A r y l glycosides have proved to be useful substrates for the mechanistic study of glycosidases. Upon hydrolysis by glycosidases, they generate phenols which can be easily detected spectrophotometrically. Such substrates can also be used for studies of linear free energy relationships by varying the substituents on the phenyl ring thereby providing minimal structural alterations with maximal electronic effects. Such substrates have been especially useful for (3-glycosidases. On the other hand, a relatively limited number of aryl a-glycosides has been synthesized and used for a-glycosidase studies. In particular, only a few aryl a-glucosides have been introduced. Currently, only two aryl a-glucosides, P N P a G l c (3.6) and phenyl a-glucosides (PaGlc , 3.11) are commercially available. The syntheses of aryl glucosides whose substituted aglycone groups have relatively high pKa values (> 7), including 2-nitrophenyl a-D-glucopyranoside ( O N P a G l c , 3.7) and 4-chlorophenyl a-D-glucopyranoside (PCPaGlc , 3.10) have been described [121,122]. However, these are all poor substrates for GLase. Thus, during the course of the mechanistic study of GLase, aryl a-glucosides containing phenol groups of lower pKA were needed, thus were synthesized. Syntheses generally started from either 2,3,4,6-tetra-60 O-acetyl-a-D-glucopyranosyl fluoride or 2,3,4,6-tetra-O-acetyl-P-D-glucopyranosyl chloride synthesized from 1,2,3,4,6-penta-O-acetyl-D-glucopyranose. B) 2,4-Dinitrophenyl a-D-Glucopyranoside (2,4DNPaGlc, 3.1). 2,4-Dinitrophenyl 2,3,4,6-tetra-O-acetyl a-D-glucopyranoside was synthesized according to the published method [123]. This method utilizes a base-catalyzed in situ anomerization from the corresponding p-anomer, itself synthesized from 2,3,4,6-tetra-O-acetyl-D-glucopyranose and fluorodinitrobenzene ( F D N B ) . Reaction for 24 hours in dimethylformamide ( D M F ) in the presence of anhydrous K2CO3 afforded a tetra-O-acetyl derivative of 3.1. Deprotection of this glycoside was challenging and had not been achieved previously, due to the lability of the product. Scheme 3.3. The synthesis of 2,4-dinitrophenyl a-D-glucopyranoside: i . anhydrous Na 2 C0 3 , DMF; ii. acetyl chloride (10%) in MeOH at 4 °C The traditional Zemplen method or even the milder ammonia-promoted deprotection resulted in the decomposition of the desired product 3.1. A milder deprotection method was needed and the acid catalyzed methanolysis, achieved by the in situ generation of H C l from a mixture of acetyl chloride in dry methyl alcohol was employed (Scheme 3.3). Usually, this deprotection method takes 24 - 36 hours [124]. However, 3.1 decomposed extensively after 5 hours of reaction, so reaction was limited to less than 5 hours, which still gave a reasonably good yield (74%). 61 C) 3,4-Dinitrophenyl a-D-Glucopyranoside (3,4DNPaGlc, 3.3), 3,5-Dichlorophenyl a-D-Glucopyranoside (3,5DCPaGlc, 3.8) and 3-Nitrophneyl a-D-Glucopyranoside (MNPaGlc, 3.9). Tetra-O-acetyl derivatives of these substrates were synthesized from 2,3,4,6-tetra-O-acetyl oc-D-glucopyranosyl fluoride by a modification of the published method [122]. Boron trifluoride-diethyl etherate was used to afford protected aryl glucosides. Thus, boron trifluoride was added in excess to maintain the acidic condition to confirm the thermodynamic control (Scheme 3.4). It is notable that, in this procedure, the substituted phenol, not the sugar, was the limiting reagent. Under these conditions the product formed initially is the kinetically favored p-anomer. However, under the acidic conditions the product anomerizes to the more stable a4inkage [122]. ^OAc . O A c .OH Scheme 3.4. BF 3-diethyl etherate catalyzed synthesis of aryl glycosides: i . a desired phenol, BF 3-diethyl etherate, acetonitrile; i i . acetyl chloride (10%) in MeOH at 4 °C for 3.3; i i i . anhydrous N H 3 (saturated) in MeOH for 3.8 and 3.9 A limitation of this procedure is that it is not applicable to syntheses involving o-nitro phenols, which hinder the action of the Lewis acid, BF3. Deprotection of the 3,4-dinitrophenyl derivative to give 3.3 was carried out via in situ HCl-promoted deprotection, as described for 3.1. 3.8 was afforded through ammonia-promoted deacetylation. X: 3,4-dinitrophenyl (3.3), 3,5-dichlorophenyl (3.8) 3-nitrophenyl (3.9) 62 D) 2,5-Dinitrophenyl a-D-Glucopyranoside (2,5DNPaGlc, 3.2), 2,4,6-Trichlorophenyl a-D-Glucopyranoside (TCPaGlc, 3.4), 4-Chloro-2-nitrophenyl a-D-glucopyranoside (4C2NPaGlc, 3.5). Since the BF3-diethyl etherate catalyzed reaction could not be used for o-nitro phenols, these were prepared by the modification of another procedure devised for synthesizing ONPocGlc [121]. The direct displacement of the chlorine on a p-glycosyl chloride by a substituted phenolate in a highly polar solvent such as hexamethyl phosphoramide (HMPA) or l,3-dimethyl-3,4,5,6-tetrahydro-2(lH)-pyrimidinone (DMPU) afforded the desired tetra-O-acetyl aryl a-glycosides (Scheme 3.5), presumably via a SN2 reaction. This approach also proved useful for 2,4,6-trichlorophenyl derivative which was not accessible by the B F 3 catalyzed reaction, despite the absence of an o-nitro group on the phenyl ring. 3.2 and 3.5 were deacetylated through the HCl/MeOH method and 3.4 by NH 3 /MeOH. ^OAc ^,OAc ^OH Scheme 3.5. S N 2 type synthesis of aryl glycosides: i . desired phenol, NaH, DMPU; ii . acetyl chloride (10%) in MeOH at 4 °C for 3.2 and 3.5; ii . anhydrous N H 3 (saturated) in MeOH for 3.4 3.4.2.2. Measurement ofkcat and Km for Aryl Glycoside Substrates The series of aryl cc-D-glucosides synthesized were analyzed as substrates for GLase by kinetic studies. These substrates have aglycone leaving groups with values X: 2,5-dinitrophenyl (3.2), 2,4,6-trichlorophenyl (3.4), 4-chloro-2-nitro (3.5) 63 in the range between 3.96 and 9.99. Kinetic parameters determined (kcaX, Km, and kcaX/Km) are presented in Table 3.2 along with pKa values of phenols and extinction coefficient differences (Ae) between each aryl glucoside and its phenol at the relevant wavelength at p H 6.0. In general, the more activated substrates such as those containing dinitrophenyl leaving groups are the best substrates for glycosidases [10]. This trend was also found with GLase. Table 3.2. Michaelis-Menten Parameters for the Hydrolysis of a Series of A r y l a-Glucosides by Gracilariopsis a- l ,4-Glucan Lyase substrate kcal (s-') a Km ( m M ) a (s^M" 1) A8 b ( m M - W ) 2 , 4 D N P a G 3.96 27 3.8 7100 10.3 2 , 5 D N P a G 5.15 11 9.9 1100 3.59 (at 440 nm) 3,4DNPocG 5.36 1.7 0.59 2900 10.6 T C P a G 6.39 1.0 0.77 1300 . 1.60 (at312nm) 4 C 2 N P a G 6.45 6.3 2.7 2300 0.710 (at 425 nm) P N P a G 7.18 0.40 2.1 190 1.66 O N P a G 7.22 5.3 11 480 1.07 3 , 5 D C P a G 8.19 1.5 3.1 480 0.726 (at 280 nm) M N P a G 8.39 0.66 2.2 300 0.307 (at 380 nm) P C P a G 9.38 0.44 2.1 210 0.574 (at 278 nm) P a G 9.99 0.11 4.4 25 0.760 (at 277 nm) a Errors in kinetic parameters are less than 10 %. B A E values were measured at wavelength of 400 nm unless otherwise noted. However, the dependence is relatively moderate. Both & C a t and kcJKm are only moderately sensitive to changes in pKa value, with kca, values changing only 250 fold from the most to the least reactive substrates ( A p ^ a = 6). Activated substrates such as 3.1, 3.2 and 3.3 were shown to be only moderately good substrates, none of which have & c a t values comparable to that of ocGlcF (kcai = 505 s"1). 64 3.5. Active Site Environments - pH Dependent Activity Retaining glycosidases contain two carboxylic acids in the active site, acting as acid/base catalyst and as nucleophile/leaving group. Reflecting this, a bell-shaped p H dependent activity has been observed for most retaining glycosidases, indicating a deprotonated carboxylate acting as a nucleophile and a protonated carboxylic acid as a general acid. Since GLase has similar structural features as suggested by the invariant aspartic acid and glutamic acid of G H family 31, a similar dependence of the activity on p H might be expected. 2 4 ; 6 8 p H Figure 3.3.pH-Dependent Hydrolysis of 2,4DNftxGlc by GLase 65 Values of kcat/Km were measured at a series of p H values by the substrate depletion method using 2,4-dinitrophenyl a-D-glucopyranoside (3.1) as substrate. The low pKa value (= 3.96) of the product phenol group simplified detection at the lower p H values. The resulting pH-activity profile proved to be a classical bell-shaped curve suggesting that there are at least two essential ionizable groups (Figure 3.3). Consistent with the initial expectation, the p H optimum (= 4.8) and the pH-dependence of kcat/Km (pKai =3.1, pKa2 - 6.7) for GLase are very similar to those of the family 31 Aspergillus niger a-glucosidase which has a p H optimum of 4.5, and pKa\ = 3.2 and pKa2 - 6.4 [105]. Thus the free enzymes (upon which kcJKm reports) have very similar active site environments. 3.6. A Covalent Intermediate and Two-Step Mechanism 3.6.1. Background The similarity of the active site environment of GLase to that of retaining a-glucosidases, in combination with other previously described evidence indicates the close kinship of GLase with G H family 31 but did not distinguish between possible mechanisms. The most direct approach to distinguish between the two possible mechanisms (a direct a«#-elimination and an indirect mechanism involving a covalent intermediate) was to search for the intermediate. The obvious reagents for this are the fluorosugars described earlier [22,23]. The mechanism-based inactivators (Figure 3.4), 5-66 fluoro-a-D-glucopyranosyl fluoride (5FaGlcF) and 5-fluoro-P-L-idopyranosyl fluoride (5FpidoF), have been used to trap covalent glycosyl-enzyme intermediates on cc-glucosidases [18,51], the trapping of the covalent glycosyl-enzyme intermediate on a G H family 31 a-glucosidase by the former agent having been described in the previous chapter. In many cases, epimers of the parent fluorosugars inverted at C5 have turned out to be more effective trapping reagents [111,112] and indeed 5F|3IdoF is a more potent a-glucosidase inactivator [51]. Thus, these two reagents were synthesized and used for trapping studies with GLase. Figure 3.4. Difluorosugars designed as mechanism-based reagents for glycosidase trapping : A. 5-fluoro-a-D-glucopyranosyl fluoride; B . 5-fluoro-P-L-idopyranosyl fluoride; C. 1-fluoro-D-glucopyranosyl fluoride; D. 2-deoxy-2-fluoro-oc-D-glucopyranosyl fluoride In addition, 2-deoxy-2-fluoro-a-D-glucopyranosyl fluoride (2FaGlcF) and 1-fluoro-D-glucopyranosyl fluoride (lFGlcF) were prepared and tested with GLase (Figure 3.4). 2FaGlcF has been shown to act as a slow substrate rather than an inactivator for a-glucosidases [23]. lFGlcF has been developed as a probe for the development of positive charge during glycoside hydrolysis and acted as a slow substrate for both a- and p-glycosidases [125]. 67 3.6.2. Reaction of a-l,4-Glucan Lyase from Gracilariopsis sp. with 5FaGlcF 3.6.2.1. Inactivation Kinetics. When 5FaGlcF was incubated with GLase from Gracilariopsis sp. and small aliquots were removed and assayed, only a small extent of inactivation was observed, much as seen with A. niger a-glucosidase. The inactivation was neither time-dependent nor did it go to completion at high concentrations of 5FaGlcF . Reduced activity was measured at the shortest time interval possible (30 seconds) upon mixing of GLase and 5FaGlcF , and did not change with time. The lower rates observed presumably arise from 'reversible' inhibition afforded by the 5 F a G l c F carried over into the assay mixture, suggesting that 5 F a G l c F might be acting as a reversible inhibitor. Therefore, 5 F a G l c F was tested as a reversible inhibitor by measuring its effects on reaction rate. A Dixon plot was constructed by measuring rates with several concentrations of substrate (2,4-dinitrophenyl a-D-glucopyranoside) with varying concentrations of 5FaGlcF . The resulting Dixon plot shows that 5 F a G l c F acts as an apparent competitive inhibitor with an apparent dissociation constant, K{\ of 6.48 m M (Figure 3.5). As seen with A. niger a-glucosidase, 5 F a G l c F apparently binds more tightly than the parent sugar, a -D-glucopyranosyl fluoride (Km - 27.9 m M ) . This improved binding affinity is unlikely to come from an increase of intrinsic affinity since the addition of a second fluorine is more likely to invoke unfavorable interactions than to improve binding. Thus, the lower K\ 68 value may indicate that this is inhibiting by virtue of its reaction as a substrate, possibly via the accumulation of a covalent glycosyl-enzyme intermediate since K\ can combine contributions from the binding and chemical events as described in chapter 2. 200 i 180 A -10 -5 0 5 10 [SFccGcF] mM Figure 3.5. Dixon plot showing the apparent reversible competitive inhibition of a-l,4-glucan lyase by 5FaGlcF. 2,4-DNPaGlc was used as the substrate: • , 0.5 mM 2,4DNPccGlc; 1 mM 2,4DNPaGlc; • , 2 mM 2,4DNPaGlc. The horizontal line corresponds to the line, y = l / V ^ . 3.6.2.2. Detection of a Covalent Glycosyl-Enzyme Intermediate Since there was a possibility that the observed reduction of activity came from the accumulation of the intermediate and thus it might be detected by mass spectrometry, a number of efforts to detect such a covalent species were made. Unfortunately, a mass for the intact protein could not be measured, as no distinct peaks were observed. This rendered direct observation of labeling by mass spectrometry impossible. Thus, labeled and unlabelled enzymes were digested by pepsin followed by HPLC/ESI MS analysis in 69 the hope of finding a labeled peptide. However, a labeled peptide could also not be detected under those conditions. 3.6.3. The Reaction of oc-l,4-Glucan Lyase from Gracilariopsis sp. with 5FfjIdoF 3.6.3.1. Inactivation Kinetics When oc-l,4-glucan lyase from Gracilariopsis sp. was incubated with more potent inactivator, 5FIdoF, time-dependent inactivation was observed. When the residual activities from each time interval were measured, it was clearly shown that the inactivation followed pseudo first-order kinetics (Figure 3.6). Values of pseudo first order kinetic constants, k0bS, at various concentrations of 5F(3ldoF were obtained by fitting measured residual activities to the first order rate equation. Subsequently, the £ 0 b s values obtained were fitted to the equation, k, = k0bJ(Ki + [I]) to yield an inactivation rate constant kt of 0.61 ± 0.02 min" 1 (or 0.010 s"1) and an inactivator dissociation constant, Ku o f 2 8 . 3 ± 2 . 1 m M . Figure 3.6. Inactivation of GLase by 5F(iIdoF: A . Residual activity at each time interval; B . Replot of apparent rate constants from A . 71 0.03 Q= T B 0.02 h -©— w/oacarbose — w/acarbose 10 20 t i m e [min ] 30 0.01 h T 1 1 1 r 20 40 60 time (min) Figure 3.7. A . Protection from inactivation in the presence of 0.2 u M acarbose (AT; = 0.02 | iM) at 30 m M 5FidoF, O without acarbose; V with acarbose. B . Reactivation of inactivated oc-l,4-glucan lyase upon removal of excess 5FIdoF. To confirm that the inactivation occurs at the active site, protection experiments were performed. In this experiment, a competitive inhibitor is expected to decrease k0bs by competitively binding at the active site. Thus, when 5FpIdoF was incubated with GLase in presence of the competitive inhibitor acarbose (0.2 p M , Kt = 0.02 p M ) , the pseudo first-order rate constant for inactivation by 30 m M 5FIdoF was reduced from 0.31 min" 1 in the absence of acarbose to 0.055 min" 1 in its presence (Figure 3.7). This reduced 72 rate constant is consistent with the value predicted on the basis of competitive binding (k = 0.053 min"1). The catalytic competence of the trapped intermediate and the relevance of the inactivation process were assessed by monitoring reactivation. In most cases with retaining glycosidases, accumulated glycosyl-enzyme intermediates have been shown to be kinetically competent and to turn over upon removal of excess inactivator. Thus, GLase was pre-inactivated through incubation with 5F(3ldoF and then excess inactivator was removed by repeated centrifugal filtration. The resulting GLase was then incubated and aliquots were removed for activity assay. Slow turnover was observed according to first order kinetics with a rate constant of A^eact. — 0.036 min" 1 or 6.0 x 10^ s"1 (Figure 3.7). These results indicate that the mechanism of GLase is closely related to that of retaining glycosidases of G H family 31. Further, the protection experiments proved that all events occurred at the active site of GLase and the reactivation of inactivated GLase through turnover of the covalent intermediate is clear evidence for the competence of the intermediate. 3.6.3.2. Identification of the Catalytic Nucleophile A s the inactivation of GLase is highly likely to be due to the accumulation of a covalent glycosyl-enzyme intermediate, the identification of the amino acid residue bearing the mechanism based inactivator was attempted. Since the full-length protein was not detected by electrospray ionization mass spectrometry (ESI M S ) , fully inactivated GLase, along with a control sample, was subjected to peptic digestion followed by L C / E S I M S comparative mapping. 73 A. 1 i isity 0.8 inter 0.6 -> 0.4 -relat 0.2 -0 B 1 >. 0.8 H jj 0.6 6 1 0.4 J O 0.2 0 jLiUhnHtmuL tin i m/z 1025 300 800 1300 1800 m/z Figure 3.8. Detection of 5FIdo-labeled peptide by comparative mapping. Mass spectrum (z = 1) of the peptic digest from the labeled enzyme (A) taken at 40.48 min and from the unlabeled enzyme (B) taken at 41.55 min. The peptic digests from the labeled and unlabeled enzymes were loaded onto the microbore reverse phase-high performance liquid chromatography (RP-HPLC) connected to the electrospray ionization mass spectrometer (ESI M S ) , and total ion chromatograms (TIC) for each sample were obtained. TICs of the labeled and unlabeled enzyme are practically indistinguishable, as indicated in Chapter 2. In order to identify the labeled peptide, the masses of the peptides under each peak in the labeled sample were compared with those of the peptides from the unlabeled sample in the corresponding region of the TIC. The masses of the peptides from the two samples were identical, with one exception 74 in each sample. A peptide fragment corresponding to m/z 1206 was detected only in the inactivated sample while a fragment of m/z 1025, lower in mass by the amount expected for the 5-fluon>|3-L-idosyl moiety (m/z = 181), was detected in the control sample (Figure 3.8). This strongly suggests that the fragment of m/z 1206 is the active site peptide labeled by 5FpidoF and therefore contains the catalytic nucleophile. This fragment was isolated by liquid chromatography (LC) monitored by the mass spectrometer and sequenced by tandem mass spectrometry (Fig. 3.9). rrVz 1025 779 594 204 m/z 433 561 676.5 778 908 1025 100 300 500 m / z 700 900 1100 1300 w c 0) 0.5 > jo m/z sequence 645 DMTV+5Fldo 989 FVWQDM+5Fldo 1089.5 FVWQDMT+ 5Fldo 1206 FVWQDMTV+5Fldo 658 645 433 \ 543.5 4,1 1206 789.5 8 8 9 5 1007.5 989 1025 1089.3 100 300 500 m / z 7 0 0 900 1100 1300 Figure 3.9. ESI MS/MS daughter ion spectrum along with interpretation; A. daughter ion spectrum (z -of m/z 1025 peak from spectrum B; B. daughter ion spectrum of m/z 1206 peak from labeled peptide 0 75 The daughter ion spectrum (Fig 3.9.B) reveals a fragment corresponding to mass 1025 (mass difference by m/z =181 from the parent ion) arising from the peptide without label. However, interpretation of the rest of the fragmentation pattern was not fruitful. Fortunately, an excellent fragmentation pattern was observed in the daughter ion spectrum of the species of m/z 1025 generated by orifice fragmentation (Fig 3.9 A ) . This readily yielded a sequence of F V W Q D M T V and revealed that the uninterpretable peaks in Figure 3.9 B arose from the loss of water or a hydroxyl from the peptide. Inspection of the daughter ion spectrum further reveals fragments of m/z 645, 989 and 1089 which are consistent with peptides D M T V , F V W Q D M and F V W Q D M T , each bearing the 5FpIdo moiety. Fully conserved amino acid residues within the smallest peptide correspond to Asp 553 and Met 554 in the sequence [75]. Asp 553 is absolutely conserved in all family 31 enzymes including GLases and this equivalent residue has been proposed previously as the catalytic nucleophile in family 31 a-glucosidases [49,75]. These results indicate that Asp 553 is the catalytic nucleophile of GLases, which strongly suggests that GLases adopt a mechanism very similar to that of retaining a-glucosidases, involving formation, of a covalent intermediate followed by a syn-elimination rather than a direct anti-elimination. The inactivation by fluorosugars implicates not only a mechanism involving a covalent intermediate but also positively charged transition states. Therefore, the first step of the GLase mechanism presumably resembles that of retaining glycosidases involving covalent intermediate formation via a cationic transition state. The second step 76 of the GLase reaction must be a P-elimination that also occurs via a cationic transition state in order for the intermediate to be trapped. 3.6.4. The Reaction of a-l,4-Glucan Lyase from Gracilariopsis sp. with 2FocGlcF and lFGlcF 2 F a G l c F and l F G l c F did not function as time-dependent inactivators of GLase since their incubation with GLase did not result in decreased activity in either case. This result with 2 F a G l c F was not unexpected given that 2FccGlcF had not successfully labeled other a-glucosidases [126]. However, it had been hoped that l F G l c F might inactivate, either via similar inductive effects on the intermediate, or v ia the alternative mechanism shown in Scheme 3.6. involving accumulation of a 2-keto glucosyl-enzyme intermediate. Scheme 3.6. Possible inactivation scheme of GLase by the accumulation of a 2-keto glucosyl-enzyme intermediate from the reaction of GLase with lFGlcF This could have arisen i f the initially formed 1-fluoro-p-glucosyl enzyme had undergone a l,2-<2M#-elimination of H F to yield an enolic glycosyl enzyme which would then ketonize (Scheme 3.6). The fact that this otherwise chemically preferred ^^ '-el imination did not occur may be suggestive of a finely tuned and optimized enzymatic syn-77 elimination process. It could also imply that the nucleophilic residue may act as the base in the second step (discussed later). However, both compounds do indeed function as slow substrates, as described in Section 3.7.2. 3.7. Transition State Structure 3.7.1. The First Step - Glycosylation, Part 1: Bronsted Relationship A n understanding of the enzymatic transition state structure requires, first, a definition of which step is rate-limiting for any particular substrate. In order to probe this, the kinetic parameters for a series o f aryl glucosides (Table 3.2) were plotted in the form of Bronsted relationships as shown in Figure 3.10. These plots of log kcat and log hcJKm vs. pKa exhibit a disappointing level of scattering (R = 0.83 and 0.90, respectively), but clearly reveal a dependence, albeit shallow (Pig = -0.32 and -0.33, respectively), of rate on leaving group ability. The similarity in these slopes suggests that the same step is rate-limiting in the two cases and the clear dependency on pKa indicates that this step involves breakage of the aryl glycoside bond. This is reinforced by the fact that kcJKm reports on the first irreversible step, which has generally proved to be the formation of the glycosyl-enzyme for retaining glycosidases. 78 CD O o 0 B 4 h E •4—' CO o 0 1 i 1 < 1 " 1 o o -1 o 1 o 1 1 1 i i 4 6 „ 8 PK 3 10 1 • i i i I I I 1 i i i i i 6 8 pK a 10 Figure 3.10. Bransted plot constructed from data of Table 3.2 showing the relationship of the rate of cleavage of a series of aryl glucosides with the pKa of the corresponding phenol. A. log kca< versus pKa; B. log kaJKm versus pA:a. The absence of any 'break' in this plot implies that no change in rate-limiting step occurs as leaving group ability increases, as might be expected given that oc-glucosyl fluoride has a kcat value of 505 s"1, almost 20 times greater than that of the best aryl glycoside substrate. Interestingly, in no case yet has a biphasic plot been seen for any a-glycosidase, whereas this behavior is quite common for (3-glycosidases [10,14-16]. The low pig value indicates that relatively little negative charge is present on the glycosidic 79 oxygen at the transition state. Thus either there is very little glycosidic bond cleavage at the transition state or considerable proton donation has occurred. The former interpretation is rendered unlikely by the inactivation afforded by a fluorosugar, 5FpidoF, and by the measurement of large a-deuterium kinetic isotope effects (discussed later) both of which indicate substantial oxocarbenium ion character thus, substantial bond cleavage. Early proton donation, thereby reducing net negative charge, therefore seems likely as suggested for the first step of the sucrase-isomaltase complex of the same gene family, G H family 31 [32]. Similarly low, but clear, dependence of rates on the pKa values of leaving groups was observed in that case along with a large a-deuterium kinetic isotope effect. Hammet-Hansch analysis for both enzymes (sucrase and isomaltase) yielded a low p value, which was very close to that measured for acid-catalyzed hydrolysis of aryl glucosides in which the glycosidic oxygen is almost fully protonated with the glycosidic bond being largely broken at the transition state. This led to the suggestion that the transition state would mimic the features of that of acid-catalyzed hydrolysis. Early protonation preceding bond breakage is not a rare suggestion for the first step of retaining glycosidases: indeed similarly low p\ g values have been seen with the Cellulomonas fimi exoglycanase and interpreted likewise [33]. 80 3.7.2. The First Step - Glycosylation, Part 2: Fluorosugars and Inhibitors 3.7.2.1. Kinetic Evaluation of Fluorosugars Three of the fluorosugars, l F G l c F , 2FccGlcF and 5FccGlcF acted as slow substrates, thus were subjected to kinetic analysis using the fluoride ion electrode. l F G l c F and 2 F a G l c F were found to bind poorly with high Km values that could not be accurately measured. Only values of kcJKm were measured for these two compounds, though values of kcat and Km could be measured for 5 F a G l c F (Table 3.3). Also given in Table 3.3 are values of kca, and Km for cc-glucopyranosyl fluoride (aGlcF) . A A G 1 values for all three difluorides ( l F G l c F , 2 F a G l c F and 5FaGlcF) were calculated to reflect the change in the free energy of activation for the first irreversible step (kcJKm) compared to the parent sugar, a G l c F using the equation A A G * = R T In [ ( ^ a t / ^ m ) F / ( & c a t / £ m ) F 2 ] where the subscript F indicates the value for the monofluoride and the F2 for the difluorides. Table 3.3: Michaelis-Menten Parameters for the Cleavage of Fluorosugars by Gracilariopsis q- l ,4-Glucan Lyase substrate kcat Km kcat/Km K ' A A G * (s' ') a ( m M ) a (s- 'M' 1 ) (mM) (kJ/mol) l F G l c F 13.8 18.1 2FccGlcF 3.81 21.1 5FaGlcF 0.131 10.7 12.2 6.48 18.4 a G l c F 5 0 5 27.9 1.81 x 10 4 3 Errors in kinetic parameters are less than 10%. A) Probing Transition State Destabilization. One additional approach to probe charge development involves measurement of effects of fluorine substitution adjacent to the reaction center upon reaction rates. A s can be seen from Table 3.3, the effects of 81 substitution of a second fluorine at C - l , C-2 and C-5 on rates of cleavage of a G l c F have been explored. Substitution at C - l and C-5 involves replacement of H by F, which may cause very minor steric penalties, but does not remove any potentially important hydrogen bonding interactions, as is possibly the case at C-2. Further, substitution at C-2 and C-5 is adjacent to the developing oxocarbenium ion, whereas substitution at C - l is directly on the carbocation. Previous non-enzymatic solvolysis studies with 2-deoxy-2-fluoro sugar derivatives have revealed an approximately 60-fold rate reduction as a consequence of substitution at the 2-position [127]. Effects of substitution right at the cationic center are less predictable. Indeed a stabilizing effect by lone pairs on fluorine substituted directly on carbocations has been seen [128], though studies of spontaneous solvolysis of l F G l c F have revealed substantial rate reductions corresponding to an increase of free energy of activation ( A A G J ) of 18.5 kJmol" 1 [125]. The difference in behavior in these two cases presumably resides in the oxocarbenium ion character in this case: lone pair donation by the oxygen likely dominates any possible donation by fluorine, leaving only the destabilizing inductive effect. Interestingly, A A G + values for all three substitutions on the enzymatic reaction were approximately 20 kJmol" 1, thus are very similar to that seen for spontaneous solvolysis. Since the non-enzymatic process has been shown to involve substantial carbocationic character [29,30], a similar transition state is predicted for the enzymatic process. Therefore, since kcJKm reflects the first irreversible step, the changes in the free energy of activation reflect the destabilization of an oxocarbenium ion-like transition state of the first, glycosylation, step. B) Kinetic Analysis of the Reaction of GLase with SFaGlcF. While only hcJKm values could be measured for l F G l c F and 2FccGlcF, values of kcaX (0.13 s"1) and Km (10.7 m M ) 82 were readily measured for 5FccGlcF (Table 3.3). The kca, value was 4000 fold lower than that of a G l c F , as expected for a 5-fluorosugar. Remarkably, the Km value is almost 3-fold lower than that of the parent oc-glucosyl fluoride (Km - 27.9 m M ) . This reduction in Km is quite significant given that an increase in true affinity as a consequence of the substitution of the C-5 hydrogen by fluorine is unlikely. A more likely explanation is that the second step has been slowed substantially more than the first by the fluorine substitution, with an accumulation of intermediate resulting. This phenomenon leads to a lowering of the Km value when the second step is clearly rate-limiting. The expression for Km for a two-step enzymatic reaction is as follows: k L _ k 2 k 3 E + S - * ES » E-S • E + P k i ES: Michaelis complex E-S: covalent intermediate where is the rate constant for the first step and ks that for the second step. If the second step is rate limiting, then » k3 and the second factor w i l l be reduced to k3/k2 ( « 1) which would yield a lower Km value. Similar cases have been observed in several glycosidases. Km values for 5 F a G l c F were too small to measure with A. niger a-glucosidase in the previous chapter. Similarly low Km values were proposed for yeast a-glucosidase (family 13) and Coffea arabica a-galactosidase (family 27) [18,110]. 83 3.7.2.2. Reversible Inhibitors Some indications of positively charged transition states are also found in the inhibition studies performed. In the case of A. niger a-glucosidase in the previous chapter and other enzymes of G H family 31, 1-deoxynojirimycin and acarbose were shown to be potent inhibitors [64,65]. Likewise, 1-deoxynojirimycin and acarbose were previously known to inhibit GLase [75,84], but no inhibition parameters had been determined. Thus, along with a few other inhibitors, these were subjected to analysis. Table 3.4 shows the K\ values of the inhibitors tested, along with their structures. 1-Deoxynojirimycin and acarbose both proved to be both potent and competitive inhibitors with K\ values of 0.13 and 0.020 p M , respectively. These values are substantially lower than those measured for A. niger a-glucosidase (1.5 and 6.6 uM) . The potent, competitive inhibition by 1-deoxynojirimycin suggests that a cationic species has high affinity for the active site of GLase while the highly potent nanomolar inhibition exhibited by acarbose (K\ = 20 nM) , itself a proven transition state analogue inhibitor of G H family 13 a-glucosidases [129], strongly supports reaction via an oxocarbenium ion-like transition state. Hydroximinogluconolactam has been shown previously to be a potent inhibitor of G H family 13 a-glucosidase (K[ = 2.9 u M with yeast a-glucosidase) [108]. In the case of GLase, this compound was shown to be a poor inhibitor (K\ = 1.28 ± 0.06 m M ) . This poor binding affinity of hydroximinogluconolactam is not completely surprising in light of the same level of poor inhibition for a-glucosidase from A. niger of G H family 31 (K[ = 1.7 m M ) described in Chapter 2. This strongly implies that G H family 31 enzymes including GLase, for which no 3-dimensional structures have yet been reported, are 'syn' 84 protonators, with their acid catalysts sitting in a syrc-relationship to the substrate endocyclic CI-05 bond thus unable to interact with the lone pair on the exocyclic nitrogen. On the other hand, G H Family 13 glycosidases are known to be a«fr'-protonators and only tfraft'-protonators are significantly inhibited by compounds of this class [28,109]. Table 3.4. The Structure of Inhibitors and K\ values for Gracilariopsis oc-1,4-Glucan Lyase Inhibitors (uM) 1-Deoxynojirimycin (1) 0.13 ±0.001 Acarbose (2) 0.020 ± 0.0003 Hydroximinogluconolactam (3) 1.3 (± 0.064) x 103 D-Glucono-1,5-lactone (4) N D a a not detected HO HO 3 4 D-Glucono-l,5-lactone also did not inhibit GLase as previously reported [75]. Poor binding of gluconolactone is less surprising since this is known to be a poor inhibitor of oc-glycosidases [31]. 85 3.7.3. The First Step - Glycosylation, Part 3: Kinetic Isotope Effects Further evidence for this mechanism is derived from kinetic isotope effect (KIE) analysis. Kinetic isotope effects have proved to be some of the best probes of transition state structure for mechanistic studies of retaining glycosidases [5,10]. B y introducing deuterium either at C I or C2 , K I E analysis can directly reveal the structure of the transition state. Primary deuterium kinetic isotope effects are measured in the reaction during which a hydrogen atom is directly involved in bond making or breaking [130,131]. Values of the primary K I E are often large, within the range of ku/k-o = 2 - 7 , dependent on the position of the proton/deuteron at the transition state. When the transition state structure is symmetric with the proton half transferred, the K I E value w i l l be close to the theoretical maximum of 7. If the hydrogen is closer to either of its new or old bonding partner, there wi l l be significant decrease in the K I E value. The presence of a primary K I E (magnitude of 2 - 7) is often used as evidence for a mechanism in which proton transfer occurs in the rate-limiting step. The secondary deuterium kinetic isotope effects arise when there is substitution adjacent to a hydrogen which is not directly involved in the bond making or breaking [130]. The magnitude of the secondary deuterium K I E is between 0.7 - 1.5. This often occurs when there is a change in the hybridization on carbon and the carbon-hydrogen bond is affected by this change during the reaction. I f the hydrogen/deuterium is attached to the carbon that changes hybridization, the K I E is an a-secondary K I E and i f the hydrogen is on the adjacent carbon, it is a P-secondary K I E . In the case of an a-secondary deuterium K I E , i f the hybridization changes from sp 3 to sp 2 and the transition state assumes sp 2 character, the K I E value w i l l have a value 86 greater than unity. More sp 2 character at the transition state, in this case, w i l l give a larger K I E value. Therefore, the presence and the magnitude of the a-secondary deuterium K I E 3 2 can serve as a probe of the development of carbocationic character (sp to sp ) at the transition state. Thus, the K I E measured for deuterium substitution on C - l is an a-secondary deuterium K I E and would reveal the existence and extent o f oxocarbenium ion-like character in the transition states of either the first or second step. The K I E measured for the 2-deutero substrates, on the other hand, would reveal the features of the transition state of the elimination reaction. Depending on whether or not the rate-limiting step involves the proton transfer from C2, either a primary K I E or a p-secondary K I E would be expected from the 2-position. Combined with the K I E from the 1-position, the K I E from 2-position could provide insight into the transition state for the elimination step as has been done in the study of non-enzymatic elimination reactions [132]. Table 3.5. Kinetic Isotope Effects Measured for Deuterated Substrates with Gracilariopsis q-l ,4-Glucan Lyase substrate K I E upon kcat K I E upon kcat/K„, l - [ 2 H]-PNPaGlc 1.19 ± 0 . 0 2 l - [ 2 H]-aGlcF 1.14 ± 0 . 0 5 1.16 ± 0 . 0 1 l - [ 2 H]-5FaGlcF 1.23 ± 0 . 1 0 2-[ 2 H]-PNPaGlc 1.06 ± 0 . 0 1 2-[ 2 H]-aGlcF 1.06 ± 0 . 0 4 1.07 ± 0 . 0 1 2-[ 2 H]-5FaGlcF 1.92 ± 0 . 1 5 Kinetic isotope effects were measured for /?-nitrophenyl a-D-glucopyranoside, a -D-glucopyranosyl fluoride and 5-fluoro-a-D-glucopyranosyl fluoride, substituted with 87 deuterium, separately at the 1- and 2-positions. (Table 3.5). Values of K I E upon kcat were measured with P N P a G l c and KIEs upon both kcat and kcat/Km were measured with a G l c F . For 5FaGlcF , only the K I E upon kcax was measured. Isotope effects arising from deuterium on C - l and C-2 were measured separately. A s noted earlier, the first, glycosylation, step is rate-limiting for P N P a G l c and the second, elimination, step is rate-limiting for 5FaGlcF while the rate-limiting step is unknown for aGlcF . Values of KIEs upon kcaX at C I and C2 are of the expected magnitude for P N P a G l c , for which the first step (glycosylation) is rate-limiting. The value of K I E upon kcat measured for l - [ 2 H ] - P N P a G l c corresponds to an a-secondary K I E and the large value measured indicates that there is a large degree of oxocarbenium ion character at the transition state. Similarly large a-secondary KIEs (1.12 - 1.21) have been reported previously for the hydrolysis of /7-chlorophenyl a-glucopyranoside by sucrase-isomaltase and isomaltose by a-glucosidases from sugar beet and A. niger, all belonging to G H family 31 [32,66], indicating that closely related mechanisms are followed between two different types of enzyme in the same gene family. More diagnostically useful are the K I E values measured for the 2-deutero substrate. If GLases adopted a directly operating a«//-elimination reaction mechanism, a significant primary kinetic isotope effect would be expected. However, the observed value with 2 - [ 2 H]-PNPaGlc (1.06) is not big enough for a primary K I E and presumably represents a (3-secondary K I E . This is fully consistent with the result provided by the trapping of the covalent intermediate. The similarity of the KIEs measured upon kcat for P N P a G l c and kcJKm for a G l c F implies that the same step is rate-limiting in the two cases. Since kcat/Km monitors the first irreversible step, this finding is fully consistent with the observation from the Bronsted 88 relationship that the formation of the glycosyl-enzyme is rate-limiting for P N P a G l c . Interestingly, the equivalence of the K I E values on & c a t and kcJKm for a G l c F implies that the same step is rate-limiting for both since kcJKm reflects the glycosylation step, thus implies that despite the & c a l value of 505 s"1, glycosylation is still rate-limiting for kcm. In summary, in all three cases where the first step, the formation of the covalent glycosyl-enzyme intermediate, is rate-limiting, large a-secondary kinetic isotope effects have been measured from 1 -deutero substrates, indicating very substantial oxocarbenium ion character at the transition state of the displacement reaction. These are fully consistent with those measured on other G H family 31 glycosidases. The absence of a primary K I E but the presence of a P-secondary K I E on 2-deutero substrates confirms that a direct awrZ-elimination mechanism is not in operation, but that the rate-limiting step involves formation of the covalent glycosyl-enzyme intermediate. In contrast to these results, KIEs upon kca, at each center measured with 5 F a G l c F are quite different since the second step is suggested to be rate-limiting for 5 F a G l c F unlike P N P a G l c and GlcF. These results w i l l be discussed later in Section 3.7.5. 3.7.4. Formulating the Transition State Structure of the First, Glycosylation, Step When all of the experimental results concerning the transition state structure of the first step are combined, the following features emerge: 89 A ) Low Pig observed in the Bransted relationship of pKa vs. log kcat and log kcat/Km: early protonation on the glycosidic oxygen. B) Large A A G * calculated from kcJKm of fluorosugars: cationic transition states. C) Highly potent inhibition by transition state analogues, acarbose and 1-deoxynojirimycin: oxocarbenium ion-like transition state. D) Large a-deuterium kinetic isotope effects: substantial oxocarbenium character at the transition state. Low Pig values could mean a small degree of glycosidic bond cleavage rather than early protonation. However, all other evidence points to a large degree of bond breaking at the transition state. A s described in Chapter 1, in a modeling study on sugar beet a -glucosidase of G H family 31, 70% of the glycosidic bond was cleaved at the transition state for the hydrolysis of ocGlcF and almost no bond formed from the incoming enzymatic nucleophile, indicating highly substantial oxocarbenium ion character [36]. A similar transition state is predicted for GLase. J W \ Z W W V W V V W V V V V O H H Figure 3.11. Schematic diagram of the transition state of the first step of the reaction catalyzed by GLase: The glycosidic bond cleaves substantially to give a highly positively charged oxocarbenium ion character while the negative charge on the glycosidic oxygen is masked by the proton donation. 90 Thus, it is likely that the glycosidic oxygen is protonated extensively, almost masking the developed negative charge, while the glycosidic bond between the anomeric center and the glycosidic oxygen is broken considerably to give substantial oxocarbenium ion character at the transition state of the first step in the degradation of a-l,4-glucans by GLase (Figure 3.11). 3.7.5. Speculations on the Transition State Structure of the Second, Elimination, Step from the Measurement of Kinetic Isotope Effects In order to probe the transition state structure of the second, elimination, step, it was necessary to identify substrates for which this step is rate-limiting. Unfortunately, none of the aryl glycosides meet this challenge, as shown by the linear Bronsted plot, and even a-glucosyl fluoride, with a £ c a t value of 505 s"1, appears to have rate-limiting glycosylation as evidenced by the similarity o f the deuterium kinetic isotope effects on £ C a t measured at each center with that measured for P N P a G l c . In an attempt to generate a substrate for which deglycosylation was rate-limiting, attention was turned to modified glycosyl fluorides containing a second fluorine substituent. A s noted earlier, the best candidate for such a study was 5-fluoro a-glucosyl fluoride (5FaGlcF) . The shift of rate-limiting step from the first to the second step by fluorine substitution, thereby resulting in accumulation of the covalent intermediate, leads to a lowering of the Km value (Table 3.3). This lower Km value allowed the measurement of rates at high substrate concentration (8.x Km) and, thus, made measurement of isotope effects upon kcm possible. 91 Accordingly, 1-[2H]- and 2-[ 2 H]-5FaGlcF were synthesized and values of K I E upon kcm at each center were measured. A s seen in Table 3.5, a very large a-deuterium kinetic isotope effect of knlko = 1.23 was observed for l - [ 2 H]-5FaGlcF , consistent with fully developed oxocarbenium ion character at the transition state. This is accompanied by a substantial isotope effect of knlko = 1.92 for the 2-deutero substrate (2-[ 2 H]-5FaGlcF). This value is clearly outside the range of a p-secondary isotope effect, and within the range of a primary kinetic isotope effect, thereby providing the first direct kinetic proof of a two step mechanism for which the second step involves at least partially rate-limiting C - H bond cleavage. This primary K I E is relatively small for a fully concerted, synchronous E2 mechanism while the K I E measured with the 1-deutero derivative is quite large for an a-secondary deuterium K I E . Based on this pair of values, some distinction can be made between stepwise and concerted mechanisms for the elimination step and a picture of the transition state structure in the second step can be proposed. The large secondary a - K I E might, on its own, indicate an E l mechanism. However, the presence of a primary K I E at C2 eliminates a fully stepwise E l mechanism. A n E l mechanism involving anchimeric assistance by the P-hydrogen and nucleophilic attack of solvent on the P-hydrogen has been previously shown to yield small primary K I E values of 1.72 ~ 1.85 [133]. However, such a mechanism is not possible in this case since the p-hydrogen is located on the same face of the pyranose ring as the scissile bond. A concerted mechanism with an asymmetric transition state involving extensive cleavage of the glycosidic bond and substantial proton transfer is more probable and consistent with oxocarbenium ion chemistry. 92 Figure 3.12. Schematic diagram of the transition state of the second step of the reaction (elimination) catalyzed by GLase: The glycosidic bond cleaves substantially to give a highly positively charged oxocarbenium ion character, making (3-hydrogen acidic enough to be abstracted. A small primary K I E may mean that the proton is either less than half or more than half transferred at the transition state. The latter case is inconsistent with substantial positive charge at the anomeric center, thus delayed proton transfer seems more probable. Indeed a mechanism in which substantial heterolysis of the glycosidic bond occurs w i l l generate a species with substantial oxocarbenium ion character, thereby acidifying the P-hydrogen to be transferred (Figure 3.12). The E2 mechanism is not necessarily completely synchronous and there can be a broad spectrum from carbocation-like to completely synchronous to carbanion-like in the E2 mechanism, as excellently depicted by the potential energy surface diagrams of More O'Ferrall [134]. Such an E2 mechanism leaning towards E l was previously invoked to explain low K I E values (2.4 ~ 2.6) of some non enzymatic E2 reactions [135]. Further, a similar kind of E - l type E-2 enzymatic ^ - e l i m i n a t i o n was found in UDP-N-acetylglucosamine 2-epimerase where a primary K I E of 1.8 was measured [119]. 93 3.8. Speculations on the General Base Acting in the Second Step The identity of the base responsible for H-2 abstraction is not clear at this stage, but the following possibilities should be taken into consideration. A proton relay system is one candidate. The general acid catalyst deprotonated in the first step would be the starting point of the proton relay and could generate a deprotonated active-site water that would play the role of the base as in the example of 2-deoxy-D-ribose-5-phosphate aldolase [136]. However, GLase is likely to be a .syn-protonator and consequently the general acid catalyst would be placed on the same side as the endocyclic oxygen of the glycone, placing the position of the general acid catalyst well away from the 2-proton and rendering this unlikely. Figure 3.13. Schematic diagram of possible hydrogen abstraction by the catalytic residue which acts as a catalytic nucleophile in the first step of the reaction catalyzed by GLase. A stronger candidate for this role must be the departing carboxylate oxygen of the catalytic nucleophile itself (Asp 553). This group is correctly positioned on the p-face of the sugar and, since the transition state is late, will be available to act as a base (Figure 3.13). Indeed three-dimensional structures of intermediates trapped on other a-glycosidases and transglucosidases [137,138] reveal that the carbonyl oxygen of the t-HO HO 94 nucleophile is situated in close proximity to the endocyclic oxygen of the sugar ring. A small rotation around the C - 0 bond would suffice to place the oxygen correctly to act in this role. The lack of inactivation by l F G l c F through the alternative mechanism discussed in Section 3.6.4 may also be indicative of this possibility since the formation of a 2-keto-glucosyl intermediate cannot occur i f the leaving group, carboxylate, is the base residue (Scheme 3.6). Additional support for this notion comes from the observation that retaining glycosidases catalyze the hydration of glycal substrates via a covalent glycosyl-enzyme intermediate and do so via the ^ - a d d i t i o n of the proton at C-2 and the enzymatic nucleophile. It has been generally assumed that the nucleophile itself donates this proton in a somewhat concerted process highly analogous to that proposed for GLase [31]. In addition, Wolfenden earlier noted that the ribosylation reaction catalyzed by nucleoside 2-deoxyribosyltransferase is accompanied by an occasional elimination reaction, generating D-ribal [118]. This also points to the existence of a mechanistic continuum. 3.9. Conclusion Based on findings in this chapter, the mechanism of GLase is suggested to be as depicted in Scheme 3.2 with some fine tuning, thus highly similar in many aspects to that of a-glucosidases of glycoside hydrolase family 31. Such a mechanism involves the formation and base-catalyzed elimination of a covalent glycosyl-enzyme intermediate. The glycosyl enzyme intermediate is formed through a transition state with significant 95 oxocarbenium ion character in which the carbon-oxygen bond is broken substantially but the negative charge on the leaving group oxygen is largely masked by advanced proton donation by the general acid/base catalyst. The covalent intermediate so formed then undergoes an elimination reaction which follows an E l - l i k e E2 mechanism via a transition state in which the C-O bond at the anomeric center is largely broken and the anomeric center assumes substantial oxocarbenium ion character. The oxocarbenium-ion character of this species leads to substantial acidification of the C-2 proton allowing its relatively facile removal. This mechanism is an example of a finely tuned enzymatic approach to the glycosidic bond cleavage. Simple glucans do not have a functional group that can activate the proton and the aglycone at the reaction center is a poor leaving group. Consequently, chemically preferred direct onri-elimination is difficult due to the high energy barrier. Nature's solution to this problem is that the poor leaving group is expelled by nucleophilic substitution such that the resulting covalent intermediate has a good enough leaving group to provide the oxocarbenium ion character capable of acidifying the proton. Thus, GLase selects an indirect path involving a covalent intermediate and a less favored syrc-elimination as a smart strategy to overcome the high barrier of the reaction due to the lack of assisting functional groups. This is a novel mechanism for glycosidic bond cleavage, but one in which all processes except the second step are common aspects of a-glucosidases of family 31, leading to the idea that GLase utilizes mechanistic machinery that is common to family 31. It therefore represents an example of mechanistic plasticity by which enzymatic mechanisms can evolve through subtle changes in active site constitution. There are several cases in which the different reactions catalyzed by enzymes in the same gene 96 family proceed through similar mechanisms. One such example is found in the enolase superfamily [116]. Enzymes in the enolase superfamily catalyze a wide variety of reactions including racemization, epimerization and (3-elimination. However, all these reactions proceed through a common step, which is the formation of a stable enolate intermediate by metal-assisted, general base-catalyzed abstraction of a proton. The fate of this intermediate determines the outcome of the overall reaction. Other examples include the N-acetylneuraminate lyase superfamily, in which a common enamine intermediate leads to a variety of reactions, and the crotonase superfamily where the common species is the oxyanion intermediate of a thioester [116]. These commonalities highlight the advantages of the sequence-based classification of glycosidases in considering reaction mechanisms and provide further caution about the annotation of genomes purely on the basis of sequence similarity. 97 Chapter IV. Cloning and Kinetic Study of a GH Family 31 oc-Xylosidase from Escherichia coli. 98 4.1. Background Glycoside hydrolase family 31 is a unique glycoside hydrolase gene family which contains both hydrolases (a-glucosidases and a-xylosidases) and lyases, unlike other G H families. These enzymes are widely distributed being found in Eukaryota, Bacteria and Archaea [49]. While GLases were isolated from only algae and fungi, this activity has been detected in a wide variety of species including mammals, plants and bacteria [75-80]. G H family 31 a-xylosidases are also found in plants and some microorganisms [67-72]. However, no G H family 31 enzymes from E. coli have been purified and characterized, although two open reading frames have been shown to have substantial sequence similarity with other G H family 31 members. Two genes, yicl and yihQ, encode these open reading frames and were reported as a part of the E. coli genome sequencing project, located at 81.5 - 84.5 min and 87.2 - 89.2 min regions on the chromosome, respectively [139,140]. In E. coli, the product of GLase, 1,5-anhydro-D-fructose has been detected and the compound shown to be further metabolized to 1,5-anhydro-D-glucitol and its 6-phosphate. This observation resulted in the suggestion of the existence of a third glycogenolysis pathway in E. coli and, later, in mammals [78,79], thereby necessitating the existence of GLase in E. coli. Since known a- l ,4-glucan lyases belong to G H family 31, these ORFs (yicl and yihQ) are strong candidates as their genes. However, the possibility that these are a-glucosidases or a-xylosidases cannot be ruled out. 99 Specific Aims of Chapter A s a continuation of the mechanistic study on enzymes of G H family 31, this chapter describes the cloning and characterization of a protein encoded by the yicl gene, one of the two ORFs . The gene product w i l l be subjected to kinetic analysis to ascertain its function and probe its catalytic mechanism. 4.2. Sequence Alignment Analysis of yicl Gene Product from E. coli The hypothetical protein encoded by the yicl gene from E. coli shows considerable sequence similarity (more than 50%) with hypothetical proteins from Salmonella typhimurium, Salmonella enterica, Bacillus halodurans, Streptomyces coelicolor and Clostridium acetobutylicum, possibly the result of sharing an evolutionary origin. Among the known G H family 31 enzymes, sequence alignment analysis yields the highest score with a-xylosidase from Lactobacillus pentosus (46%) (Figure 4.1). However, this E. coli protein also shows 20 - 29% similarity with mammalian, plant and fungal a-glucosidases, and plant a-xylosidases. When aligned with a-l,4-glucan lyase, disappointingly, only 20% similarity was observed. The protein encoded by the other gene yihQ, however, shows lower similarity in general, thus, presented no better a target. Two signature regions can be found throughout all members of G H family 31 (Figure 1.5 in Chapter 1), though there are some variations between proteins from bacteria and other higher organisms, especially in the signature region 1. 100 Y I C I _ E C O L I 3 00 LHVFHFDCFWMKAFQWCDFEWDPLTFPDPEGMIRRLKAKGLKICVWINPYIGQKSPVFKE AAL2260 8 3 00 LHVFHFDCFWMKAFQWCDFEWDPVTFPDPKGMIRRLKAKGLKVCVWINPYIGQKSPVFQE CAD0324 5 3 00 LHVFHFDCFWMKAFQWCDFEWDPVTFPDPKGMIHRLKAKGLKVCVWINPYIGQKSPVFQE Q9K3K5 1 QWCDFEWDPDVFPDPDGMLARLKAKGLRVSAWINPYIAQKSPLFDE Q9WYE4 294 LHVFHFDCFWMKEFHWVDLEWNRENFPDPEGLLKRLKEKGLKVCVWINPYVSQFSSLFDE Q9KBM1 2 99 VHVFHFDCFWMKEFEWCNFEWYRRVFPEPEKMLQRLKEKGLKLSVWINPYIAQRSPLFQE Q97K36 297 LDVFHFDCFWMKEFEWCNFKWDDRMFKNPEKMLKTIKSKGIKTCVWINPYIAQKSPLFNE P96793 2 98 LDVFHFDCFWQKGFEWCTLEWDKEQFPDPEGLLKKIHDRGIKVCVWLNPYIAQKSPLFKE Y I C I _ E C O L I 3 60 LQEKGYLLKRPDGSLWQWDKWQPGLAIYDFTNPDACKWYADKLKGLVAMGVDCFKTDFGE AAL22608 360 LKEKGYLLKRPDGSLWQWDKWQPGLAIYDFTNPQACEWYADKLKGLVEMGVDCFKTDFGE CAD03245 360 LKEKGYLLKRPDGSLWQWDKWQPGLAIYDFTNPQACEWYADKLKGLVEMGVDCFKTDFGE Q9K3K5 47 AAALGHLVRRPDGDIWQWDLWQAGMGLVDFTSPAARDWYAGKLKPLLDQGVDCFKTDFGE Q9WYE4 3 54 GKEKGYFLKKPDGDVWQTDDWQPGMAIIDFTNPEVRKWFASKLERLIDMGVDCFKTDFGE Q9KBM1 3 59 AAANGYLLKKENGDVWQWDLWQPGMGVVDFTNPDARIWYQDHLRRLLEMGVDCFKTDFGE Q97K36 3 57 AVENGYLLKRANGEVWQWDMWQAGMGVVDFTNPKATVWFQNKLEELVDMGVDAFKTDFGE P96793 358 AKDKGYLLTRENGDIWQWDLWQAGNGFVDFTNPAAVKWYQDKLKVLLDMGVDSFKTDFGE WiDMnE Y I C I _ E C O L I 420 RIP-TDVQWFDGSDPQKMHNHYAYIYNELVWISTVLKDTVGEEEAVLFARSASVGAQKFPVH AAL22608 420 RIP-TDVQWFDGSDPQKMHNHYAYIYNELVWNVLKETVGVEEAVLFARSASVGAQQFPVH CAD03245 420 RIP-TDVQWFDGSDPQKMHNHYAYIYNELVWNVLKETVGVEEAVLFARSASVGAQQFPVH Q9K3K5 107 RIP-TDWWHDGADPERMHNYYTHLFNRTVFELLEKERGQGEAVLFARSATAGGQQYPVH Q9WYE4 414 K I P - T D W Y Y D G S D P E K M H N Y Y T Y L Y N K W F E T I E R K L G K R N A W F A R S A T A G S Q K F P V H Q9KBM1 419 RIP-TDWYHDGSDPEKMHNYYTFLYNQTVFDVLKQVKGNHEAVLFARSATAGSQQFPVH Q97K36 417 RIP-TDWYYDGSNAKKMHNYYTYLYNKTVFDLLKRKKGEKEAVLFARSATVGGQKFPVH P96793 418 RIPAEDVKFFDGSNPQQEHNYYTLQYNRAVYEVIQQEKGADEAVLFARSQRLVHNPIQYT Figure 4.1. Partial sequence alignment of bacterial enzymes and hypothetical proteins of G H family 31: underlined sequences are the active site sequence and the arrow indicates the proposed catalytic nucleophile. WiDMnE is the corresponding sequence in enzymes of higher organisms in G H family 31: These are hypothetical proteins unless otherwise noted, AAL22608, Salmonella typhimurium; CAD03245, Salmonella enterica; Q9K3K5, Streptomyces coelicolor (fragment); Q9WYE4, Thermotoga maritima; Q 9 K B M 1 , Bacillus halodurans; Q97K36, Clostridium acetobutylicum; P96793, a-xylosidase from Lactobacillus pentosus {XylQ). The consensus sequence surrounding the catalytic aspartate nucleophile in this region of plant, mammalian and fungal enzymes, including a-glucosidases and a-xylosidases, is W i D M n E . GLase also retains this sequence except that E is replaced by V or T, giving W i D M n X ( V or T) as the concensus. However, corresponding sequences in bacterial proteins are shown to have the sequence K T D F G E and are absolutely invariant. The sequence, W i D M n E , even with some gaps, is not found among bacterial proteins. Small but clear variations in amino acid sequence between bacterial and higher organisms are also found in signature region 2 (Figure 1.5). Therefore, there are clear distinctions between amino acid sequences of higher organism proteins and bacterial proteins of G H 101 family 31, suggesting that these bacterial proteins took a different path early in the evolution of G H family 31 [70,141]. Though none of the bacterial proteins presented in Figure 4.1 except for the a-xylosidase from Lactobacillus pentosus, have been characterized, the high amino acid sequence similarity suggests that these proteins, including that encoded by the yicl gene of E. coli, may be a-xylosidases. 4.3. Cloning and Overexpression of the yicl Gene 4.3.1. The Amplification and Cloning of yicl Gene Sequences covering a number of base pairs of the 5'- and 3'-end sequences of genomic D N A of the target gene, yicl from E. coli K-12 , were used to design the oligosaccharides for polymerase chain reaction (PCR) primers. Ndel and Xhol restriction enzyme sites were appended at the 5' and 3' ends of D N A , respectively. Designed primers are presented in Figure 4.2. 102 P r i m e r s Yicl-TOP : 5'-CA G A A C TA A G G A A C G CAT AT G A AAATTA G C -3' Y ic l -END : 5 ' -ATCAAG CT CG A G C A A C G T A A T T G T C A G C G C-3' Sph I Sma I Figure 4.2. Designed primers for amplification of yicl gene by polymerase chain reaction and constructed plasmid containing yicl Gene; Sequences in bold character are restriction enzyme sites, C A T A T G , AWel site and C T C G A G , Xhol site. Synthesized primers were subjected to P C R using E. coli genomic D N A as a template. The amplified yicl gene was confirmed by agarose gel electrophoresis (Figure 4.3). The D N A molecule consists o f 2319 base pairs and a band of this size was clearly visible as a single band on the gel. The prepared gene was purified using the QIAquick® gel extraction kit ( Q I A G E N Inc.) and digested with Ndel and Xhol restriction enzymes to generate sticky ends. After digestion with restriction enzymes, the yicl gene was purified through agarose gel electrophoresis and subsequent gel extraction. The expression vector for the yicl protein was constructed by inserting the purified gene into the plasmid, pET29a. The constructed vector, designated as 103 pET29EcoXyl31A (Figure 4.2) was then transformed into E. coli Top 10. The transformants were screened by incubation in the presence of an antibiotic, kanamycin and the resulting clones were purified from the cell lysates using the QIAprep® Spin Miniprep kit ( Q I A G E N Inc). Figure 4.3. Agarose gel electrophoresis of Amplified yicl Gene: The size of yicl is 2319 bps. 4.3.2. Overexpression of the yicl Gene Plasmids containing the yicl gene (pET29EcYICI) were transformed into E. coli BL21 (DE3) and the resulting cells cultured in the presence of 1 m M IPTG to induce the overexpression of protein. The overexpressed protein was identifiable on the SDS polyacrylamide gel electrophoresis as a band at 89 kDa (Figure 4.4) and was purified by N i - N T A affinity column chromatography. During pooling and concentration of the target protein fractions, it was observed that the protein was susceptible to aggregation when the 104 concentration of imidazole was lowered. This may be the consequence of the presence of a membrane associated hydrophobic region [67], which could lead to aggregation in the absence of imidazole. Alternatively it would be due to the formation of a complex between several 6 x His tags and co-eluted N i 2 + ion. Indeed, when EDTA was added, Y i c l protein aggregation no longer occurred and the Y i c l protein remained soluble even after EDTA was removed by centrifugal filtration and buffer was exchanged from elution buffer (pH 7.9, 0.02 M Tris buffer, 0.5 M imidazole, 0.5 M NaCl) to pH 7.0, 0.05 M phosphate buffer. This strongly suggests that aggregation was a consequence of the presence of co-eluted N i 2 + ion. Once metal ion was removed, it was shown that the Y i c l protein was soluble in a range of buffers at pH 7.0 including phosphate buffer (0.05 M), HEPES buffer (0.05 M) and MOPS buffer (0.05M). Purification is summarized in Table 4.1. kDa 113 Figure 4 .4. Electrophoretic analysis of expressed Yicl protein. Lanes: 1, molecular weight marker, 2. Soluble fraction of crude cell lysate; 3. After Ni -NTA affinity column 105 Table 4.1. Purification of Y i c l protein a Purification step Protein Activity b Specific activity Purification Yie ld mg units 0 units/mg fold % Cell Supernatant 1460 21.9 0.015 1.0 100 N i - N T A 160 14.0 0.088 5.9 64 a From 2 L culture. b Assays were performed in p H 7.0, 0.05 M phosphate buffer at 37 °C using 4-nitrophenyl a-xylopyranoside as substrate. c One unit is defined as the amount of enzyme causing release of one micromole of 4-nitrophenol per minute. 4.4. Characterization of the Protein Encoded by yicl A A A . Molecular Weight and Analysis of Tryptic Fragments The molecular weight of the protein encoded by the recombinant gene yicl is calculated to be 89,030 D a based on the primary structure including 6 x His tags. The SDS P A G E shows that the molecular weight of the expressed protein is indeed about 89 k D a (Figure 4.4). The Y i c l protein was digested with trypsin and the resulting peptide fragments were analyzed by liquid chromatography/electrospray ionization mass spectrometry (LC/ESI M S ) . A total of 141 peptide fragments were expected from tryptic digests of the Y i c l protein and 102 peptide fragments matched with the predicted masses of peptides that would be generated by trypsin from the Y i c l protein were identified by M S . These peptides cover 83% of the entire Y i c l protein, thereby confirming the cloning of the correct protein. 106 4.4.2. Substrate Specificity To investigate the enzymatic activity of the Y i c l protein, 12 different glycoside substrates were incubated with the protein in p H 7.0, 0.05 M phosphate buffer at 37 °C. A s expected from the high sequence similarity with the G H family 31 a-xylosidase from L. pentosus, the Y i c l protein has substantial activity with /?-nitrophenyl a - D -xylopyranoside ( P N P a X y l ) and a-D-xylopyranosyl fluoride ( a X y l F ) as substrates. Lower, but significant activity was observed with /?-nitrophenyl a-D-glucopyranoside, while other substrates such as a (3-glucoside, galactosides and mannosides were not hydrolyzed at all (Table 4.2). Table 4.2. The Hydrolysis of Various Substrates by Y i c l protein from E. coli k IK Substrate ^ ( s " 1 ) Km(mM) j j i ^ PNPp-D-mannopyranoside a N D b - -PNPp-D-galactopyranoside N D PNP|3-D-glucopyranoside N D PNP|3-D-xylopyranoside N D PNPa-L-arabinopyranoside N D PNPa-D-mannopyranoside N D PNPa-D-galactopyranoside N D PNPa-D-glucopyranoside 0.046 ± 0.001 c 7.7 ± 0.4 6.0 PNPa-D-xylopyranoside 0.13 ± 0.01 c 0.71 ± 0 . 0 4 180 a-D-xylopyranosyl fluoride 75 ± 3 0.97 ± 0.13 7 . 7 x 1 0 Maltose N D Maltotriose N D a P N P : p-nitrophenyl; b Not Detected; c Ae of 4-nitrophenol at p H 7.0 was measured to be 9.13 (± 0.02) mM"'cm"' . Rates were calculated using this value. ,4 107 Thus, this protein seems to cleave a-xylosides and a-glucosides. However, when maltose and maltotriose were incubated with the Y i c l protein, no activity was observed at all, even after an extended period of time. These results suggest that gluco-oligosaccharides are not good substrates for the Y i c l protein. A s noted earlier, two types of substrate specificity have been observed for a-xylosidases. The a-xylosidases of the first type are mostly plant enzymes and the best substrates for them are xyloglucan oligosaccharides [69-72,141]. a-Glucosides are also good substrates. However, small a-xylosides such as P N P a X y l and isoprimeverose are poor substrates. On the other hand, a-xylosidases from microorganisms such as S. solfataricus, L. pentosus and Bacillus sp. have a stringent specificity for smaller a-xylosides [67,68,142]. These have relatively high activity for P N P a X y l and isoprimeverose and little or no activity for xyloglucan oligosaccharides, a-glycosides and maltooligosaccharides. The current findings for this new E. coli enzyme are therefore consistent with those found for a-xylosidases from microorganisms. Values of &Cat and Km for the hydrolysis o f P N P a X y l by this protein are in the same range as those for two other bacterial a-xylosidases (kca, = 0.21 s"1 and Km = 1 m M for the enzyme from Bacillus sp. and kcat = 0.1 s"1 and Km — 1.3 m M for that from L. pentosus). When the Y i c l protein was incubated with either P N P a X y l or P N P a G l c and products were checked by T L C , only hydrolyzed products, xylose or glucose appeared, indicating the absence of lyase activity. A s a result, the Y i c l protein has been named an a-xylosidase. Interestingly, including the present enzyme, all a-xylosidases whose sequences are known belong to G H family 31. 108 4.4.3. Factors Affecting the Activity of a-Xylosidase fromis. coli (yicl): pH, Temperature and Metal Ions 4.4.3.1. pH-Dependent Activity The measurement of the pH-dependent activity of the newly identified a-xylosidase from E. coli (yicl) was performed using a-D-xylopyranosyl fluoride (ocXylF) as a substrate. This choice is based on the fact that a X y l F is the best substrate to date and that fluoride ion measurement at lower p H is easier than the measurement, at lower p H , of/>-nitrophenol released from P N P a X y l . A classical bell-shaped dependence of kcJKm upon p H was observed as expected, indicating at least two ionizable groups involved in the activity of the free enzyme (Figure 4.5). The two apparent pKa values are pKai - 4.9 ± 0.2 and pKa2 = 7.9 ± 0.1 with an estimated p H optimum of p H 6.4. This value is somewhat higher than those (pH = 4-5) for a-glucosidases of G H family 31, usually known as acid a-glucosidases [49]. Plant a-xylosidases are also acidic enzymes, showing similar p H optima to those of a-glucosidases [69-72]. However, a-xylosidases from microorganisms show somewhat higher p H optima. The optimum p H values for the activity o f a-xylosidases from S. solfataricus and Bacillus sp. are 5.5 and 7.5, respectively [68,143] while the p H optimum of L. pentosus enzyme has not been measured. 109 4.4.3.2. Effects of Temperature and Metal Ions The apparent enzyme activity on PNPocXyl was measured at various temperatures (Figure 4.6). The activity increased until 50 °C, then dropped and was lost at 70 °C. At 60 °C, a high activity could still be measured but was lost in a few minutes. The increase in rate with temperature is simply that expected from any reaction, while the drop in activity at high temperature is a consequence of time-dependent thermal denaturation. The shape of such curves is therefore highly dependent on how the enzyme is assayed and should not be interpreted too seriously. 110 0 20 40 60 80 Temperature (°C) Figure 4.6. Temperature dependence of the hydrolysis of PNPaXyl by a-xylosidase from E. coli At lower temperatures, the hydrolysis of P N P a X y l was detected down to 10 °C though only 2% of maximal activity was observed at this point. The a-xylosidase stored in p H 7.0, 0.05 M phosphate buffer retains full activity for 48 hours at 37 °C and for several months at 4 °C. The effects of various metal ions including C a 2 + , M g 2 + , M n 2 + , Z n 2 + , N i 2 + , C u 2 + and C o 2 + on the activity of a-xylosidase were examined and none of these affected the apparent rate of hydrolysis of P N P a X y l . This result was expected since none of the enzymes of G H family 31 has shown any metal ion dependence [49]. I l l 4.5. Kinetic Evaluation of New Mechanism-Based Inactivators for a-Xylosidase from E. coli 4.5.1. a-Xylosidases in Nature a-Xylosidases are involved in the metabolism of oligosaccharides derived from xyloglucan and isoprimeverose (Figure 1.6), a small unit of xyloglucan. Plant cell walls are composed of networks built by cellulose microfibers cross-linked by xyloglucan chains, and pectins [144]. A s described in Chapter 1, xyloglucan consists of a P-(l,4)-D-glucan backbone that mainly carries a-D -xylosyl residues and a-L-fucosyl-(l,2)-|3-D-galactosyl-(l,2)-a-xylosyl residues as side chains to the 6 -OH of P-glucosyl unit of the backbone. Thus, the cooperative action on xyloglucan by xyloglucan endoglucanase, exo-p-glucosidase, a-fucosidase, and P-galactosidase is required to yield oligosaccharides [68-72]. In addition to their structural role, xyloglucan-derived oligosaccharides play a role in regulating auxin- and acid pH-induced growth [73,74]. a-Xylosidases cleave the a -xylosyl residue attached to the glucose residue farthest from the reducing end of the xyloglucan derived oligosaccharides. This removal of the terminal xylosyl residue makes possible the action of p-glucosidase since P-glucosidase only hydrolyzes P-glucans with a free non-reducing end 6-OH. Thus, its removal serves as a control point for the regulation of the appropriate level of xyloglucan oligosaccharide [69-74]. In contrast to the processing of xyloglucan-derived oligosaccharides in plants, that of isoprimeverose is relatively little known. Currently, the best substrate for microbial a-xylosidases is isoprimeverose and the xylPQ regulon of Lactobacillus pentosus, the only known 112 microbial genetic system encoding a trnasporter protein and a-xylosidase, seems to be involved in the metabolism of isoprimeverose [67]. However, details of a-xylosidases found in this xylPQ regulon and other microorganisms are yet to be studied [67,68]. 4.5.2. Synthesis of (5S)- and (5R)-5-Fluoro-a-D-Xylopyranosyl Fluorides While there have been very few studies on a-xylosidases, considerable work has been done on enzymes acting on p-xylosidic linkages [16,109,145-148]. This is probably due to the higher frequency of occurrence of P-xylosidic linkages in nature as found in, for example, xylan and xylobiose. Accordingly, a number of mechanistic probes such as mechanism-based inactivators and various transition state analogues for xylanases and p-xylosidases have been synthesized [16, 109, 145]. In contrast to the abundant repertoire of chemical probes for the mechanistic analysis of p-xylosidases, virtually no effort has been made to study a-xylosidases, largely, due to the lack of demand. Thus, during the course of this study on the new a-xylosidase from E. coli, the synthesis of new mechanism-based inactivators specific for a-xylosidases became necessary. Consequently, 5-fluoro-a-D-xylopyranosyl fluorides were designed and synthesized. Since the xylopyranosyl unit does not have a hydroxymethyl group at C5 , two forms of 5-fluoro-a-D-xylopyranosyl fluoride are possible with either an axial or an equatorial fluorine at C5 . 113 4.3 4.4 Scheme 4.1. Synthesis of axial (4.3) and equatorial (4.4) 5-fluoro-a-D-xylopyranosyl fluorides The synthesis started from the per-O-acetylated derivative of D-xylose (Scheme 4.1). 1,2,3,4-Tetra-O-acetyl-p-D-xylopyranose underwent anomeric fluorination in the presence of HF/pyridine at -20 °C [149] to yield 2,3,4-tri-O-acetyl-a-D-xylopyranosyl fluoride. Subsequent photobromination [150] using N-bromosuccinimide under light afforded the axial 5-bromide, (5S)-5-bromo-a-D-xylopyranosyl fluoride in low yield (20%). Competing bromination of the acetyl groups at extended reaction times resulted in low yields and difficult purification. Therefore, the reaction time was limited to 24 hours. Fluorination [151] of the bromo-derivative using A g B F 4 in toluene on 4 A molecular sieves afforded a mixture of axial and equatorial epimers, (5S)- and (5R)-2,3,4-tri-0-acetyl-5-fluoro-oc-D-xylopyranosyl fluorides, 4.1 and 4.2, respectively after less than an hour of reaction. The mixture of epimers was easily separable by column chromatography and both were obtained in crystalline forms. NFfj-promoted deacetylation of each epimer of the per-O-acetylated compounds afforded the desired 114 axial and equatorial epimers, (5S)-5-fluoro-a-D-xylopyranosyl fluoride (ax-5FaXF 4.3) and (5R)-5-fluoro-a-D-xylopyranosyl fluoride (eq-5FaXF 4.4), respectively. 4.5.3. The Inactivation of a-Xylosidase from E. coli with 5-Fluoro-a-D-Xylopyranosyl Fluorides, 4.3 and 4.4. 4.5.3.1. Inactivation Kinetics In the post genomic era, the catalytic nucleophile o f a-xylosidase from E. coli could probably be predicted through sequence alignment with other members of G H family 31 as noted earlier. However, it was shown that bacterial enzymes including a -xylosidases from E. coli and L. pentosus differ substantially from the a-glycosidases of higher organisms in their amino acid sequence around the predicted catalytic residue (Figure 4.1). While this variation in amino acid residues may be a minor factor in the mechanism, it could lead to mechanistic alterations. Thus, it is still necessary to confirm the prediction experimentally. Further, it is also interesting to observe the effect o f axial and equatorial fluorine substitutions at the 5-position of substituted xylosyl fluorides on the rate of inactivation. When axial 5-fluoro-a-D-xylopyranosyl fluoride 4.3 was incubated with E. coli a-xylosidase at 37 °C, apparent inactivation was observed by measuring activities of aliquots. However, this substantially reduced activity was observed at the shortest time interval measurable and did not change with time. Further, using low concentrations of ax -5FaXF 4.3, inactivation did not go to completion. This behavior is reminiscent of that 115 observed in the reaction of 5-fluoro-oc-D-glucopyranosyl fluoride (5FocGlcF) with a -glucosidase from A. niger and with the Gracilariopsis a-l,4-glucan lyase in which apparent tight binding was also observed (Chapter 2 & 3). Thus, ax-5FocXF 4.3 was tested as a competitive reversible inhibitor and the apparent K, value was determined to be 9.8 p M . The equatorial epimer eq-5FaXF 4.4 behaved likewise with an apparent Kt value of 0.45 p M revealing even tighter apparent binding (Figure 4.7). 1 2 3 1/[PNPaXyl] mM' 1 1 2 3 1/[PNPaXyl] mM"1 Figure 4.7. Double reciprocal plots showing the apparent reversible competitive inhibition of a-xylosidase from E. coli by ax-5FaXF 4.3 (A) and eq-5FocXF 4.4 (B). The concentration of each inhibitor is as follows: A. [ax-5FoXF 4.3], • , 0 uM; • , 2.25 uM; • , 4.5 uM; O , 9 uM; B. [eq-5FaXF 4.4], O , 0 uM; A, 0.1 uM; 0,0.25 uM; • , 0.5 uM. A s noted earlier, when turnover of the intermediate is marginally slower than the rate of formation of the intermediate, or when the turnover rate is substantial even though the rate difference between the two events is considerable, such behavior would be observed [112,114]. To confirm that this rapid inactivation is due to the accumulation of the covalent intermediate, reaction mixtures of both samples were subjected to analysis by electrospray ionization mass spectrometry, which revealed that the mass of the protein 116 increased by a mass of 157 (reaction with 4.3) and 158 (reaction with 4.4). This is in general agreement with the expected difference for addition of the mass of a 5-fluoro-xylosyl unit (mass of 151) from ax -5FaXF 4.3 and eq-5FaXF 4.4 (Figure 4.8). The labeling of the a-xylosidase by these two reagents indicates that the reduced rate is indeed due to the accumulation of the intermediate. Figure 4.8. Labeling of full length a-xylosidase from E. coli with ax-5FaXF 4.3 and eq-5FaXF 4.4: A. Control sample; B. Labeled protein with 43; C. Labeled protein with 4.4 117 Even though time-dependent inactivation could not be measured under the conditions employed, it might be possible to observe time dependent inactivation under other conditions, such as lower temperature, which could reduce the turnover number. Thus, inactivation was re-examined at lower temperature. When ax-5FaXF 4.3 and. eq-5FaXF 4.4 were incubated with a-xylosidase and the activity at appropriate time intervals was measured at 10 °C, time-dependent inactivation was indeed observed for both reagents (Figure 4.9). Even though the inactivation was not complete at a low concentration of inactivator and a steady state phase was observed, the initial phase of inactivation showed a time-dependent rate reduction according to pseudo first-order kinetics in both cases (Figure 4.9). The apparent first order rate constants, kQbs at each concentration of 4.3 and 4.4 were fitted to k0bS = &i[I]/([I] + Kt) as before, yielding kt and Ki values for both compounds. For ax-5FaXF 4.3, k-, value was 0.17 + 0.01 s"1 and Kt value of 2.4 ± 0.3 mM while k,• = 0.16 ± 0.02 s"1 and Kt = 0.31 ± 0.06 mM for eq-5FaXF 4.4. 118 0 2 4 6 8 10 12 14 16 T i m e ( m i n u t e s ) B 6 I i i i i | i i i i | i i i i | i i i i | i i i i | i I Figure 4.9. Inactivation of oc-Xylosidase from E. coli with 4.3 and 4.4 at 10 °C. The enzyme was incubated with the following concentrations of ax-5FaXF 4.3 and eq-5FoXF 4.4 at 10 °C and assayed with PNPocXyl at 10 °C. A . Inactivation by 4.3, O 0.1 mM, • 0.2 mM, • 0.3 mM, • 0.5 mM, A 0.8 mM, • 2.5 m M ; B. Replot of Pseudo first order kinetic constants ( k o b s ) from A vs concentration of 4.3; C . Inactivation by 4.4, O 0.05 mM, • 0.1 mM, • 0.12 m M , • 0.15 m M , A 0.25 mM, • 0.3 m M , V 0.45 m M ; D. Replot of kobs values from C vs concentration of 4.4. .a o 2 h 119 ' I ' I I 1 I 1 I 1 O 0 20 40 60 80 100 120 140 160 time (min) B 0.04 i — i — | — i — j — i — | — i — \ — i — | — i — | — i — | — r c E < "> o < 0.02 0 0 20 4 0 60 80 100 120 140 160 time (min) Figure 4.10. Reactivation at 37 °C of inactivated (at 10 °C) a-xylosidase from E. coli: A . enzyme inactivated with ax-5FaXF 4.3; B . enzyme inactivated with eq-5FaXF 4 . 4 The kinetic competence of the intermediates formed was also assessed. Excess inactivator was removed from each enzyme sample that had been fully inactivated by the corresponding inactivator at 10 °C, then the reactivation of each inactivated enzyme was monitored by measuring activity at a series of time intervals at 10 °C and also at 37 °C. The reactivation for both samples followed pseudo first-order kinetics (Figure 4.10) and yielded rate constants of = 2.8 (± 0.1) x 10"4 s"1 for ax-5FocXF 4.3 and 4.5 (± 1.0) x 10"5 s"1 for eq-5FaXF 4.4 at 10 °C and £ r e a c t = 3.2 (± 0.4) x 10~3 s"1 for ax-5FocXF 4.3 and 120 3.2 (± 0.1) x 10"4 s"1 for eq-5FccXF 4.4 at 37 °C, proving that inactivation occurred through the normal enzymatic reaction mechanism. 0 2000 4000 6000 8000 10000 Time (second) time (second) Figure 4.11. Hydrolysis of (5S)- and (5R)-5-fluoro-a-D-xylopyranosyl fluorides catalyzed by E. coli a-xylosidase. Concentrations of ax-5FocXF 4.3 (A) and eq-5FaXF 4.4 (B) are as designated on each profile. Finally, 4.3 and 4.4 were analyzed as substrates at 37 °C by measuring the released fluoride anion using a series of substrate concentrations. In each case, the enzymatic reaction continued at a constant rate until almost all the substrate was consumed, and, further, the same rate was observed at a range of concentrations of ax-121 5 F a X F 4.3 and eq-5FaXF 4.4, respectively (Figure 4.11). These observations for both compounds are also reminiscent of what was seen during the reaction of a-glucosidase from A. niger with 5 F a G l c F (Chapter 2) and consistent with the low Ki' values measured since the Km value of each compound as substrate should be equal to its K' values. The turnover number, £ c a t , was determined from the slope of these plots of fluoride release versus time, yielding 2.16 (± 0.02) x 10"3 sec"1 for ax -5FaXF 4.3 and 1.22 (± 0.08) x 10"4 sec"1 for eq-5FaXF4.4 . 4.5.3.2. Analysis of the Kinetic Parameters Obtained from the Inactivation Reactions The kinetic behavior of the axial and equatorial 5-fluoro-a-D-xylopyranosyl fluorides with the a-xylosidase in which slow on-set of inhibition and apparent tight binding was observed is reminiscent o f that seen for the hydration of glycals by glycosidases [113]. The only kinetic parameters that could be measured were those for the overall enzymatic reaction. However, at lower temperatures where turnover of the intermediate is significantly slowed, time-dependent inactivation was observed, allowing independent measurement of the rates of formation and hydrolysis of the covalent 5-fluoro-P-D-xylosyl enzyme intermediates. These kinetic parameters are presented in Table 4.3 along with those for hydrolysis of a-D-xylopyranosyl fluoride ( a X y l F ) , and clearly reveal the effect of the highly electronegative fluorine substituent on catalysis by the a-xylosidase. Values of A A G * for hydrolysis of 4.3 and 4.4 (difluorides) were calculated to assess the change in free energy of activation for the first irreversible step 122 (kcat/Km) compared to the parent sugar, a X y l F (monofluoride) using the equation A A G * = R T In [(kcJKm^/tycJKm)?!] where the subscript F indicates the value for the monofluoride and F2 refers to the difluorides. In this calculation, apparent K\ (K\) values for difluorides were used instead of Km values since Km values were too low to be measured and since the Kf values should be equal to the Km values. Table 4.3. Kinetic Parameters for the Hydrolysis of Fluorosugars by the oc-Xylosidase from E. coli Measured at 37 °C. Substrate h (s"')a ^turnover (S ) K{ (mM) ax-5FoXF 4.3 - 0.17 ± 0 . 0 1 3.2 (± 0.4) x 10"3 2.4 ± 0 . 3 eq-5FaXF 4.4 0.16 ± 0 . 0 2 3.2 ( ± 0 . 1 ) x 10"4 0.31 ± 0 . 0 6 ocXylF N A N A N A a measured at 10 ° C ; b reactivation rate constant & r eact Substrate kcat (S _ 1) KmoxK? (pM) kczilKm (sAUAf A A G * (kJ/mol) ax-5FoXF 4.3 2.2 (± 0.1) x 10' 3 9.8 ± 0.6 b 224 (± 17) 15.1 ( ± 2 . 2 ) eq-5FoXF 4.4 1.2 ( ± 0 . 1 ) x 10"4 0.45 ± 0.05 b 267 ( ± 3 5 ) 14.6 ( ± 2 . 7 ) o X y l F 75 (± 3) 970 ± 130 7.73 (± 1.08) x 10 4 0 K\ values; d KC values were used for 4.3 and 4.4. It is obvious from these data that hydrolysis of these substrates proceeds through highly oxocarbenium ion-like transition states since a 3.5 x 10 4 fold reduction in & c a t value for ax-5FocXF 4.3 and a 6.0 x 10 5 fold reduction for eq-5FocXF 4.4 from that for a X y l F were observed. The destabilization of the oxocarbenium ion by the fluorine at C5 is reflected in the large A A G * values. Since fluorine substitution is not expected to create a significant steric penalty, a substantial portion of the destabilization must be caused by the inductive effect of the highly electronegative fluorine adjacent to the center of positive charge at the transition state. Previous measurement of the A A G * value for 123 spontaneous hydrolysis o f fluorosugars, where oxocarbenium ion-character at the transition states is fully developed, yielded values of 18 - 21 kJmol" 1 (Chapter 3) [125]. These are very close to the values obtained here for xylosyl fluorides. Interestingly the similarity of the A A G * values for ax-5FocXF 4.3 and eq-5FocXF 4.4 implies very similar destabilization of the first step afforded by the axial and the equatorial fluorine substituents at C5 . In contrast, the value of kcaX for ax-5FccXF 4.3 is 15 fold higher than that of eq-5FaXF 4.4, indicating that the second step is strongly influenced by the orientation of the substituents. This result is consistent with the higher kcaX values measured for enzymatic reactions with other reagents bearing an axial fluorine at C5 compared to those with an equatorial fluorine. Thus for yeast a-glucosidase [114] and a-glucan lyase (Chapter 3), kcaX values for 5-fluoro-a-D-glucopyranosyl fluoride were at least 3140 and 220 fold higher than the turnover numbers of the covalent glycosyl-enzyme intermediate (kTeacX) with 5-ffuoro-p-L-idopyranosyl fluoride. Since kcaX values are smaller than the individual rate constants for each step, these comparisons represent a minimal ratio, thus the destabilizing effect may be greater. However, D-glucopyranose has a hydroxymethyl group at C5 and the wrong location of this group may well be a big factor in the destabilization. k •] k j ^turnover E + 5FXF E-5FXF • E-5FX • E + Product Scheme 4.2. Kinetic scheme of the reaction of ax-5FocXF (or eq-5FaXF) with a-xylosidase (E). E»5FXF is the non-covalent Michaelis complex between the enzyme and the inactivator. E-5FXF is the covalent intermediate. k\ and k.\ are the rate constants for rapid and reversible formation of the initial Michaelis complex. Other rate constants are as explained in text. 124 While & c a t values only provide information on the overall enzymatic reaction, each individual step (formation, h and turnover, turnover o f the covalent intermediate) can be assessed through measured parameters of inactivation and reactivation (Scheme 4.2). Values o f the reactivation rate constant & r e a c t reflect turnover o f the trapped intermediates, thus, ^react values are used as values for ^turnover- K\ values represent the true binding affinity of inactivators ax-5FocXF 4.3 and eq-5FaXF 4.4. The value of h of ax-5FocXF 4.3 is essentially identical to that for eq-5FaXF 4.4, indicating that there is no difference in destabilizing effects from an axial or an equatorial fluorine at C5 for the first step, consistent with similar A A G * values calculated for both reagents. On the other hand, the turnover value of ax -5FaXF 4.3 is 10 fold higher than that of eq-5FccXF 4.4, indicating that the equatorial fluorine substituent has a greater effect on the second, deglycosylation, step of the enzymatic reaction than does the axial. This might arise from different substrate reactivities for the two steps. The leaving group in the first step of the enzymatic reaction, fluoride, is a much better leaving group than the enzymatic carboxylate o f the second step. This greater reactivity could decrease the selectivity o f the reaction. However, it may be hard to imagine that the inductive through-bond effects exerted by the axial and equatorial fluorine on the adjacent oxocarbenium ion are significantly different. Differences may be attributed to the change in the conformation of the sugar moiety of the covalent intermediate. The xylopyranosyl unit is well known for its conformational flexibility (discussed in Appendix A ) . Previously, the solution conformation of 2,3,4-tri-O-acetyl-p-D-xylopyranosyl fluoride was shown to be a ' Q chair with fluorine and acetyl groups in the axial orientation [152]. This was attributed to the powerful anomeric effect o f a fluorine atom overcoming unfavorable 1,3-diaxial 125 interactions of acetyl groups. The covalent intermediate formed from eq-5FocXF 4.4 w i l l have only the equatorial fluorine, while the intermediate from ax-5FocXF 4.3 w i l l have the axial fluorine. Therefore, the sugar moiety of the intermediate formed from eq-5FccXF 4.4 may change its conformation to a 'C4 chair or a B i i 4 boat conformation to place the fluorine in the axial orientation, thereby breaking the important stabilizing effects on the transition state by the interactions between the sugar moiety and enzymatic groups (Figure 4.12). second step Figure 4.12. Hypothetical conformation change of sugar moiety of the covalent glycosyl-enzyme intermediate: A. the intermediate formed from the reaction of a-xylosidase with eq-5FaXF 4.4 may experience conformation change from 4 C i to ' C 4 or B | ? 4 ; B . the intermediate formed from the reaction of a-xylosidase with eq-5FaXF 4.3 will not experience conformation change The K, value of the equatorial epimer 4.4 is 'almost 8 fold lower than that of the axial epimer 4.3, indicating a higher binding affinity for 4.4. The lower affinity for the compound with an axial fluorine substituent might indicate that there are unfavorable interactions with enzymatic groups around the position occupied by the axial fluorine. Alternatively, the distortion of the sugar ring observed in the small molecule crystal structure (Appendix A ) may well result in poorer binding. 126 The apparent K, (Ki) is almost 20 fold lower for the equatorial epimer 4.4 than for the axial epimer 4.3. A s noted earlier, KC expresses contributions from both the binding event (K,) and the chemical reactions including the formation (Jc,) and turnover (kturn0ver) of the covalent glycosyl-enzyme intermediate, through the relation, K? = Kt/(l + &i/&tumover) [31,113]. For both reagents, the ki/ktum0ver term is large, resulting in a low value of KC The lower Ki and turnover values resulted in a yet lower Ki for eq-5FaXF 4.4. 4.6. The Identification of the Catalytic Nucleophile of the oc-Xylosidase from E. coli Since inactivation of a-xylosidase was a consequence of the accumulation of the covalent glycosyl-enzyme intermediate, as evidenced by the labeling of fiill-length protein (Figure 4.8), the identification of the amino acid labeled was performed. Fully inactivated enzyme was prepared with both inactivators, separately, along with a control sample containing no inactivator. These were subjected to peptic digestion, followed by L C / E S I M S comparative mapping. A comparison of all masses of all the peptides in the inactivated and control samples revealed that the only significant difference was a peptide fragment corresponding to m/z 601 (triply charged) and 902 (doubly charged) which was detected in the two inactivated samples while no such peptide was detected in the control sample (Figure 4.13). 127 350 450 550 650 750 850 950 1050 ml: Figure 4.13. Comparative Mapping of Peptic Digests of Control Sample taken at 30.2 minutes of total ion chromatogram (A), Labeled Peptide Samples with ax-5FaXF 4.3 taken at 30.4 min (B) and eq-5FctXF 4.4 taken at 30.5 min (C). If these are the labeled peptides of interest, then a peptide of mass ~550 (triply charged) or -827 (doubly charged) might be expected in the TIC of the unlabeled sample, this being the mass difference between the peptide of mass 1011 and the 5-fluoro-xylosyl 128 label of mass 51 (triply charged) or 75 (doubly charged). Unfortunately, no such peptide was observed (Fig 4.13), possibly indicating that the unlabeled peptides are susceptible to further peptic digestion and have been converted to smaller fragments. Differences in proteolytic cleavage as a consequence o f the presence of a sugar residue are not rare [Chapter 2 and Reference 115]. Thus, this fragment was isolated from both inactivated samples by H P L C and sequenced by ESI tandem mass spectrometry (Fig. 4.14). Daughter ion spectra of peptide samples derived from inactivation with axial and equatorial 5-fluoro substituents are identical and only the M S / M S analysis of the peptide labeled by eq-5FocXF 4.4 is shown. The daughter ion spectrum reveals a doubly charged fragments corresponding to m/z 902 and 827 (mass difference of m/z = 75; i f singly charged, Am/z =150) arising from the peptide with and without a label (5-fluoro43-D-xylosyl moiety). The pattern of other singly charged fragments shown in Figure 4.12 readily yielded a sequence of F K T D F G E R I P T D Q V . 129 A . Sequence m/z FKTDFGERIPTDVQ 827 FKTDFGERI 1094 FKTDFGE 780 FKTD 492 FKT 377 FK 276 TDFGERIPTDVQ 1377 DFGERIPTDVQ 1276 FGERIPTDVQ 1161 GERIPTDVQ 1014 RIPTDVQ 828 IPTDVQ 672 PTDVQ 559 DVQ 361 VQ 246 FKTDFGERIPTDVQ + 5 F 0 t X y l 902 FKTDFGERIPTDV + 5 F 0 C X y l 1657 FKTDFGERI + 5 F a X y l 1245 FKTDFGER + 5 F a X y l 1131 FKTDFG + 5 F 0 C X y l 847 TDFGERIPTDVQ + 5 F 0 C X y l 1528 ( d o u b l y charged) ( d o u b l y charged) 0 500 1000 1500 m/z Figure 4.14. ESI MS/MS daughter ion spectrum of the peptide fragment m/z 902 (doubly charged) labeled by eq-5FotXF 4.4: A. Interpretation; B. MS/MS Spectrum 130 Further inspection of the daughter ion spectrum reveals singly charged fragments of m/z 1528, 1245 and 827, which are consistent with peptides T D F G E R I P T D Q V , F K T D F G E R I and F K T D F G , each bearing the 5-fluoro-p-D-xylosyl moiety. Consensus amino acid residues correspond to Thr 415, Asp 416, Phe 417 and Gly 418 in the sequence (Figure 4.1) [67]. Only Asp 416 among these four residues is absolutely conserved through all enzymes of G H family 31, and this also corresponds to the invariant residue proposed previously as the catalytic nucleophile in G H family 3 1 a -glucosidases and a-l,4-glucan lyases [Chapter 2 and 3, and Reference 49]. These results strongly suggest that Asp 416 is indeed the catalytic nucleophile of this a-xylosidase, despite the irregularities found in the amino acid sequence around the catalytic nucleophile. The amino acid sequence around Asp 416 o f bacterial enzymes, which is different from those of higher organisms, could then be considered a minor variation that is specific for the class, but does not affect the mechanistic process. Therefore, the catalytic nucleophile of this a-xylosidase has been identified experimentally for the first time. The identification o f the catalytic residue through labeling with mechanism-based inactivators reveals clearly that O R F yicl of E. coli is indeed an a-xylosidase of G H family 31 and follows the normal double displacement mechanism proceeding through oxocarbenium ion-like transition states. 131 4.7 Conclusion a-Xylosidic linkages are relatively rare in nature systems, thus the enzymatic cleavage of cc-xylosidic linkages has been studied very little. Recently, the importance of a-xylosides of xyloglucans of the plant cell wall has been confirmed and an additional role of xyloglucan as a signal regulator in cell wall elongation was discovered [73,74,153,154]. Thus, several plant a-xylosidases have been cloned and shown to catalyze the hydrolysis of xyloglucan-derived oligosaccharides resulting from the degradation of xyloglucans by glucanases [69-72,141]. On the other hand, only two a-xylosidases from microorganisms have been cloned and characterized [67,68]. Neither enzyme has been studied beyond analysis of sequence. Thus, this study adds another example with a bacterial a-xylosidase from E. coli and, for the first time, provides the unequivocal assignment of the catalytic nucleophile of an a-xylosidase, clearly showing the integrity of mechanistic features within the same gene family. During the course of this study, two new mechanism-based inactivators for a-xylosidase have been synthesized and X-ray crystal structures of per-O-acetyl derivatives were determined, providing interesting insight into the influence of anomeric substutuents at both C I and C5 , and of both configurations, on ring conformations, and bond lengths (discussed in Appendix). 132 Concluding Remarks Glycoside hydrolase (GH) family 31 is by far the most interesting family o f the glycosidases in terms of its mechanistic plasticity. It is usual for enzymes with different substrate specificities to belong to the same family. This polyspecificity is probably acquired through divergent evolution from the one ancestral enzyme of a single substrate specificity. However, G H family 31 shows another type of divergent evolution. Whereas a-glucosidases and a-xylosidases of G H family 31 catalyze hydrolysis, formally a nucleophilic substitution, a-l,4-glucan lyases, also included in G H family 31, catalyze an overall anti p-elimination. Examples of members of the same gene family catalyzing different reactions are not uncommon outside the glycosidase world and studies have shown that major elements of the mechanism are largely conserved among them, often involving one or more common steps. In this case, it has been shown that p-elimination of a-l,4-glucan lyase proceeds through the formation of the same covalent glycosyl-enzyme intermediate that is formed by retaining a-glycosidases of this family. It is not hard to imagine that the acquisition of such a new mechanism (P-elimination) from a nucleophilic substitution or vice versa, could be achieved through alterations in amino acid residues during evolution. The key step is conserved, along, presumably, with much of the active site structure. Thus, it is quite impressive that active sites of similar structure could lead to quite different overall reactions through slight changes in active site microenvironment. This mechanistic plasticity of enzyme active sites is another strategy of nature leading to the balance between diversity and efficiency for cleavage of glycosidic bonds. 133 Chapter V. Experimental Methods 5.1. Synthesis 134 5.1.1. General Methods A l l ' H nuclear magnetic resonance ( N M R ) spectra were recorded at either 200 or 300 M H z and 1 9 F N M R spectra at 188 M H z or 282 M H z using a Bruker AC-200 or A V -300 spectrometer. 1 9 F N M R spectra were referenced to C F 3 C O O H . Mass spectrometry was performed by Dr. Shouming He using a PE-Sciex A P I 300 triple quadrupole mass spectrometer (Sciex, Thornhill, Ont, Canada) equipped with an electrospray ionization ion source or performed by the mass spectrometry laboratory at the University of British Columbia. X-Ray crystal structure determination was performed by Dr. Brian Patrick of the X - R a y Crystallographic Services at the University of British Columbia. Elemental analysis was performed by the Microanalysis Laboratory at the University of British Columbia. Melting points were measured with a M e l Temp II melting point apparatus and are uncorrected. Thin layer chromatography (TLC) was performed on aluminum-backed sheets of silica gel 6OF254 (Merck) of thickness 0.2 mm and visualized using U V light (254 nm) and/or by applying a solution of 10% ammonium molybdate in 2 M H2SO4 followed by heating. Column chromatography was performed on 230 - 400 mesh silica gel (Silicycle, Quebec, Canada). Solvents used were freshly distilled over CaH2 except methanol which was distilled over magnesium. Dimethylformamide ( D M F ) and chloroform were dried extensively over 4 A molecular sieves. 135 5.1.2. General Materials Acarbose was a generous gift of Miles Inc. (now Bayer). 1-Deoxynojirimycin was purchased from Toronto Research Chemicals Inc. (North York, O N , Canada). Hydroximinogluconolactam was a generous gift from Dr. Tanja M . Wrodnigg. 1-Fluoro-D-glucopyranosyl fluoride ( l F G l c F ) and 2-deoxy-2-fluoro-a-D-glucopyranosyl fluoride (2FocGlcF) were synthesized by M r . L loyd McKenzie . 5-Fluoro-a-D-glucopyranosyl fluoride (5FaGlcF) and 5-fluoro-P-L-idopyranosyl fluoride (5FpIdoF) were prepared as previously described [18]. (5R)-2,3,4-friacetyl-5-fluoro-P-D-xylopyranosyl fluoride and 2,3,4-triacetyl-5,5-difluoro-p-D-xylopyranosyl fluoride was synthesized by Dr. David J. Vocadlo. 2,4-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside and 4-chlorophenyl a-D-glucopyranoside (PCPaGlc) were synthesized by Dr Carl S. Rye and M s Viv ian Y i p . 4-Nitrophenyl a-D-glucopyranoside (PNPaGlc) , phenyl a-D-glucopyranoside (PaGlc) , 4-nitrophenyl a-D-xylopyranoside ( P N P a X y l ) and D-glucono-l,5-lactone were purchased from Aldr ich Chemical Company. 2-Nitrophenyl a-D-glucopyranoside (ONPaGlc , 3.7) was prepared as previously described [121]. l-[ 2H]-D-glucopyranose and 2-[ 2H]-D-glucopyranose were purchased from Cambridge Isotope Laboratories, Inc. (Andover, M A , U S A ) . Deutero substrates, p-nitrophenyl 1-[2H]-a-D-glucopyranoside and p-nitrophenyl 2-[ 2H]-a-D-glucopyranoside [121], l-[ 2H]-a-D-glucopyranosyl fluoride and 2-[ 2H]-a-D-glucopyranosyl fluoride [149], and l-[ 2H]-5-fluoro-a-D-glucopyranosyl fluoride and 2-[ 2H]-5-fluoro-a-D-glucopyranosyl fluoride [18] were synthesized as previously described except that the 136 2 2 corresponding l -[ H]-D-glucopyranose or 2-[ H]-D-glucopyranose was used as a starting material. A l l other chemicals and reagents were purchased from Sigma/Aldrich Chemical Co. unless otherwise noted and used without further purification. 5.1.3. General Synthesis 5.1.3.1. General Acetylation General acetylation procedures were taken from known methods [155]. There are two general methods, one of which yields a mixture of a- and (3-anomers and the other only the fJ-anomer. Anomeric Mixture. The free sugar (10 g) was dissolved in a solution of acetic anhydride (50 g) and pyridine (65 g) and the solution was stirred at room temperature until T L C indicated a complete reaction. The solution was poured into an ice-water and crude crystals were formed. After filtration, the crude crystals were redissolved in ethyl acetate and washed with water and saturated N a C l . After drying over MgSG^ , the product was concentrated under reduced pressure. Pure product was obtained as white needles through recrystallization from ethyl acetate and petroleum ether. /3-Anomer. A solution of sodium acetate (25 g) and acetic anhydride (350 mL) was brought to a boil and the sugar (50 g) was then slowly added without heating. The acetylation reaction was exothermic, thus the reaction mixture continued boiling once reaction had been initiated. After all the sugar had been added and the reaction had 137 subsided, the reaction mixture was heated to a full boil . The solution was immediately cooled and poured into ice-water. After 3 hours with occasional stirring, crystalline material was isolated by suction filtration. Pure product was obtained through recrystallization from ethanol. 5.1.3.2. Ammonia-Promoted Deacetylation The protected sugar (typically 1 mmol) was dissolved in dry methanol (30 mL) , cooled on ice and anhydrous ammonia bubbled through until the solution was saturated with ammonia (typically 5 minutes). The reaction mixture was warmed to room temperature, stirred until the reaction was completed, then concentrated under vacuum. 5.1.3.3. Acetyl Chloride-Promoted Deacetylation Dried acetyl chloride (10% w/w) was added to a solution of the protected sugar (typically 1 mmol) in dry methanol (90 mL) and the reaction mixture was stirred at 4 ° C until T L C analysis indicated a complete reaction. The mixture was then concentrated under vacuum and the resulting oi l was either purified by column chromatography or directly crystallized. 5.1.3.4. a-D-Glycosyl Fluoride Synthesis This method is taken from reference 149. Per-O-acetylated sugar (typically 15 mmol) was added to HF/pyridine (25 g, Aldr ich Chemical Co.). The anomeric configuration of the starting material did not affect the results of this reaction. The reaction mixture was stirred under a N2 atmosphere at room temperature for 2 hours 138 (glucosyl fluoride) or at -20 °C for 1 hour (xylosyl fluoride). The reaction mixture was poured into cold saturated NaHCG"3 and the organic material was extracted with chloroform (3 times). The organic layer was washed with saturated N a H C 0 3 , water and saturated N a C l and dried over MgSCM. After the solvent was evaporated under reduced pressure, the product was crystallized from ethyl acetate/petroleum ether. 5.1.3.5. J3-D-Glycosyl Chloride Synthesis This method was taken from reference 156. Per-O-acetylated P-glycopyranose was dissolved in dried chloroform (15% w/v). Into this solution was added crushed anhydrous AICI3 (2 equivalents). The reaction mixture was stirred in the dark at room temperature for 30 minutes. Then, dry benzene was poured into the reaction mixture, followed by silicic acid, and the precipitate was filtered off. The filtrate was added to ice-water in a separatory funnel and the organic layer was separated after shaking. The solvent was evaporated under reduced pressure and the resulting oi l was crystallized from diethyl ether and petroleum ether. 5.1.4. Synthesis of Aryl a-D-Glucopyranosides 5.1.4.1. Aryl 2,3,4,6-Tetra-O-Acetyl-a-D-Glucopyranosides a) BF3-Diethyl Etherate-Catalyzed Synthesis. This method was a modification of the known procedure [122] and was used for synthesizing aryl glucosides formed from phenols without an ortho-mtio substituent. 139 Under a nitrogen atmosphere, boron trifluoride-diethyl etherate (BF3-OEt2, 7.63 g, 0.054 mol = 6 equiv.) was added to a mixture of 2,3,4,6-tetra-O-acetyl-a-D-glucopyranosyl fluoride (3.2 g, 0.009 mol) and the desired phenol (0.4 equivalent). The mixture was stirred for 3 hours at room temperature and then saturated N a H C C h (60 mL) was added. The organic layer was extracted with ethyl acetate (60 m L x 2), washed with water (150 mL) and saturated N a C l (150 mL) , dried over M g S 0 4 , concentrated under vacuum and the product purified by column chromatography. 3' ,4-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-cc-D-glucopyranoside. Purification by column chromatography (97/3 chloroform/acetone, Rf 0.24) and crystallization from diethyl ether yielded white yellowish plates (52%): m.p. 167 °C, ! H N M R (CDC1 3 , 300 M H z ) 5 8.03 (d, 1 H , J5..6> 9.1 Hz , H5 ' ) , 7.52 (d, 1 H , J2>,6> 2.5 Hz, H2 ' ) , 7.38 (dd, 1 H , h\e 2.5, J 5 - , 6 ' 9.1 Hz , H6 ' ) , 5.85 (d, 1 H , J 1 > 2 3.7 Hz, HI ) , 5.63 (dd, 1 H , J 2 i 3 10.4, J 3 , 4 9.8 Hz, H3), 5.14 (dd, 1 H , J 3 , 4 9.8, J 4 , 5 10.3 Hz , H4), 5.07 (dd, 1 H , J 2 , 3 10.4, J i , 2 3.7 Hz , H2), 4.20 (dd, 1 H , J 5 , 6 a 5.1, J 6 a ,6b 12.4 Hz, H6a), 4.02 (dd, 1 H , J5,6b 2.2, J 6 a , 6 b 12.4 Hz, H6b), 3.97 (ddd, 1 H , J 4 , 5 10.3, J 5 i 6 a 5.11, J 5 , 6 b 2.2 Hz, H5), 2.02 - 2.07 (4s, 1 2 H , 4 x C H 3 C O ) Anal . Calcd. for C 2 0 H 2 2 N 2 O 1 4 : C, 46.70; H , 4.31; N , 5.45. Found: C, 47.05; H , 4.39; N , 5.83 AcO 140 3,5-Dichlorophenyl 2,3,4,6-tetra-O-acetyl- a-glucopyranoside. CI Purification by column chromatography (97/3 chloroform/acetone, Rf 0.48) yielded a white solid upon the evaporation of solvent (76%): m.p. 113.5 - 114.5 °C, ' H N M R (CDC1 3 , 300 M H z ) 5 7.06 (dd, 1 H , J 1.8 Hz , H4 ' ) , 7.02 (d, 2 H , J 1.8 Hz, H 2 ' , H6 ' ) , 5.67 (d, 1 H , J 1 > 2 3.6 Hz , H I ) , 5.60 (dd, 1 H , J 2 , 3 10.0, J 3 , 4 9.6 Hz , H3), 5.10 (dd, 1 H , J 3 , 4 9.6, J 4 , 5 9.9 Hz, H4) 5.00 (dd, 1 H , J , , 2 3.64, J 2 , 3 10.0 Hz , H2), 4.24 (dd, 1 H , J 5 , 6 a 6.0, J 6 a ,6b 12.8 Hz, H6a), 4.00 - 4.06 (m, 2 H , H5 , H6b), 2.02 - 2.06 (4s, 12 H , 4 x C H 3 C O ) Anal . Calcd. for C 2 oH 2 20,oCl 2 : C , 48.69; H , 4.49. Found: C, 48.62; H , 4.42 3-Nitrophenyl 2,3,4,6-tetra-O-acetyl- a-glucopyranoside. Purification by column chromatography (96/4 chloroform/acetone, Rf 0.44) and crystallization from ethyl acetate/hexanes yielded short white needles (45%): m.p. 101 -102 °C, ' H N M R (CDC1 3 , 300 M H z ) 5 7.92 - 7.97 (m, 2 H , H 2 ' , H4 ' ) , 7.40 - 7.49 (m, 2 H , H 5 \ H6 ' ) , 5.79 (d, 1 H , J 1 > 2 3.6 Hz, H I ) , 5.67 (dd, 1 H , J 2 ; 3 9.9, J 3 , 4 9.8 Hz , H3), 5.15 (dd, 1 H , J 3 , 4 9.8, J 4 > 5 9.7 Hz , H4), 5.05 (dd, 1 H , J , , 2 3.6, J2,3 9.9 Hz , H2), 4.23 (dd, 1 H , 141 •/5,6a 4.9, J6a,6b 12.2 Hz, H6a), 4.07 (ddd, 1 H , J4,5 9.7, J 5 , 6 a 4.9, J 5 , 6 b 2.1 Hz , H5), 4.03 (dd, 1 H , J5,6b 2.1, J6 a,6b 12.2 Hz, H6b), 2.02 - 2.06 (4s, 12 H , 4 x C H 3 C O ) Anal . Calcd. for C20H23NO12: C, 51.18; H , 4.94; N , 2.98. Found: C, 51.21; H , 4.77; N , 3.16 b) Direct Displacement Reaction at the Anomeric Center. A modification of the published procedure [121] involving direct displacement of chloride at the anomeric center by the desired phenolate was employed for all aglycones with an ortho nitro substituent. Under a nitrogen atmosphere, 2,3,4,6-tetra-O-acetyl-f}-D-glucopyranosyl chloride (2.9 g, 0.008 mol) was added to a solution of 1,3-dimethyl-3,4,5,6-tetrahydro-2(lH)-pyrimidinone ( D M P U , 30 mL) containing the sodium salt of the desired phenol (3 equivalents) prepared in situ from the phenol and sodium hydride. The solution was stirred for 6 hours at 40 0 C, then overnight at room temperature. A n ice-water mixture (100 mL) was added and the precipitate filtered with Celite, washed with water, and then dissolved in methylene chloride (100 mL). The organic layer was washed with N a H C 0 3 (150 mL), water (150 mL) and saturated N a C l (150 mL) , dried and then the solvent was evaporated under vacuum. 2,5-Dinitrophenyl 2,3,4,6-tetra-O-acetyl- a-D-glucopyranoside. 0 2 N 142 Product was purified by column chromatography (97/3 chloroform/acetone, Rf 0.4) and triturated in 1/1 ethyl acetate/hexanes followed by crystallization from diethyl ether as yellowish short needles (11%): m.p. 134 - 136 °C, ' H N M R (CDC1 3 , 300 M H z ) 5 8.25 (d, 1 H , J4-,6' 2.2 Hz, H6 ' ) , 8.06 (dd, 1 H , Jy,4> 8.8, J 4 , j 6 , 2.2 Hz, H4 ' ) , 7.92 (d, 1 H , J3'.4' 8.8 Hz , H3 ' ) , 5.92 (d, 1 H , J , , 2 3.7 Hz , HI ) , 5.58 (dd, 1 H , J 2 3 10.3, J 3 , 4 9.8 Hz , H3), 5.14 (dd, 1 H , J 3 4 9.8, J 4 , 5 9.7 Hz, H4), 5.02 (dd, 1 H , J U 2 3.7, J 2 > 3 10.3 Hz , H2), 4.27 (dd, 1 H , J 5 , 6 a 5.5, J6a,6b 12.1 Hz , H6a), 4.18 (ddd, 1 H , J 4 , 5 9.7, J 5 , 6 a 5.5, J 5 j 6 b 1.5 Hz , H5), 4.07 (dd, 1 H , J 5 , 6 b 1.5, J 6 a , 6 b .12.1 Hz, H6b), 2.03 - 2.09 (4s, 12 H , 4 x C H 3 C O ) Anal . Calcd. for C 2 0 H 2 2 N 2 O i 4 : C, 46.70; H , 4.31; N , 5.45. Found: C, 46.79; H , 4.30; N , 5.80 2,4,6- Trichlorophenyl 2,3,4,6-tetra-O-acetyl- a-D-glucopyranoside. The direct displacement method was employed nicely in this case. Column chromatography was performed (97/3 chloroform/acetone, Rf 0.35) and the product was N M R (CDCI3, 300 M H z ) 8 7.32 (s, 2 H , H 3 \ H5 ' ) , 5.85 (d, 1 H , J l i 2 4.0 Hz, HI ) , 5.70 (dd, 1 H , J 3 4 9.7, J 2 , 3 10.5 Hz , H3), 5.18 (dd, 1 H , J i , 2 4.0, J 2 , 3 10.5 Hz , H2), 5.15 (dd, 1 H , J 3 i 4 9.7, J 4 , 5 10.3 Hz , H4), 4.79 (ddd, 1 H , J 4 , 5 10.3, J 5 6 a 4.0, J5,6b 2.4 Hz , H5), 4.25 (dd, 1 H , J6a,6b 12.5, J5,6a 4.0 Hz, H6a), 4.11 (dd, 1 H , J 6 a, 6b 12.5, J 5 ; 6b 2.4 Hz, H6b), 2.02 ~ 2.07 (4s, 1 2 H , 4 x C H 3 C O ) AcO crystallized from ethyl acetate/hexanes as a white solid (42%): m.p. 108 - 109 °C, H 143 Anal . Calcd. for: C20H21O10CI3: C, 45.52; H , 4.01. Found: C, 44.56; H , 3.99 4-Chloro-2-nitrophenyl 2,3,4,6-tetra-O-acetyl- a-D-glucopyranoside. 0 2 N Column chromatography was performed (97/3 chloroform/acetone, Rf 0.29) and the product was crystallized from ethyl acetate/hexanes as yellowish white needles (10%): m.p. 141 °C, ' H N M R (CDC1 3 , 300 M H z ) 8 7.80 (d, 1 H , J3-,5> 2.5 Hz , H3 ' ) , 7.47 (dd, 1 H , J 3 - > 5 - 2.5, J5>,6- 8.9 Hz , H5 ' ) , 7.28 (d, 1 H , J5-,6> 8.9 Hz , H6 ' ) , 5.80 (d, 1 H , J 1 2 3.7, H I ) , 5.60 (dd, 1 H , J 2 > 3 9.8, J 3 , 4 9.8 Hz, H3), 5.15 (dd, 1 H , J 3 , 4 9.8, J 4 , 5 9.8 Hz , H4), 4.98 (dd, 1 H , J 2 , 3 9.8, J, 2 3.7 Hz, H2), 4.23 (dd, 1 H , J 5 , 6 a 4.4, J 6 a , 6 b 12.0 Hz , H6a), 4.16 (ddd, 1 H , J 4 , 5 9.8, J 5 , 6 a 4.43, J 5,6b 1.8 Hz, H5), 4.08 (dd, 1 H , J5,6b 1.8, J 6 a , 6 b 12.0 Hz, H6b), 2.03 -2.07 (4s, 1 2 H , 4 x C H 3 C O ) Anal . Calcd. for C 2oH2 2NOi 2CI: C , 47.68; H , 4.40; N , 2.78. Found: C, 47.96; H , 4.42; N , 3.00 144 5.1.4.2. Aryl a-D-Glucopyranosides 2,4-Dinitrophenyl a-D-glucopyranoside (2 ,4DNPaGlc , 3.1). H O 3.1 2,4-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the acetyl chloride-promoted deacetylation procedure. Reaction time was limited to 5 hours to minimize the formation of a side product (Rf 0.17). The product was purified by column chromatography (25/3/2 ethyl acetate/methanol/acetic acid, Rf 0.36), then crystallized from methanol/chloroform/hexanes as a pale yellowish solid (74%). m.p. 100 - 112 °C (decomp.), ' H N M R ( D 2 0 , 300 M H z ) 5 8.80 (d, 1 H , JV ; 5> 2.8 Hz , H3 ' ) , 8.42 (dd, 1 H , J 5-, 6- 9.4, J3>,5- 2.8 Hz , H5 ' ) , 7.55 (d, 1 H , J 5 - , 6 . 9.4 Hz , H6 ' ) , 5.96 (d, 1 H , J u 3.4 Hz, H I ) , 3.85 (dd, 1 H , J 2 , 3 9.8, J 3 > 4 9.6 Hz , H3), 3.72 (dd, 1 H , J 2 > 3 9.8, J , , 2 3.4 Hz, H2), 3.55 - 3.63 (m, 3 H , H5, H6a, H6b), 3.44 (dd, 1 H , J 3 > 4 9.6, J 4 , 5 9.1 Hz , H4), ESI M S [ M + Na] + 369.1 Anal . Calcd. for C , 2 H i 4 N 2 O i 0 + 1/2H 2 0: C, 40.57; H , 4.26; N , 7.89. Found: C, 40.92; H , 4.50; N , 7.44 145 2,5-Dinitrophenyl a-D-glucopyranoside (2,5DNPccGlc, 3.2). o 2 N 3.2 2,5-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the acetyl chloride-promoted deacetylation procedure. The reaction was performed for 5 hours. The product was purified by column chromatography (27/3 ethyl acetate/methanol, Rf 0.37) and crystallized from methanol/chloroform/hexanes as white yellowish needles (67%). m.p. 149 - 150 °C, ! H N M R ( D 2 0 , 300 M H z ) 5 8.21 (d, 1 H , J4'i6> 2.2 Hz , H6 ' ) , 8.03 (d, 1 H , J3>,4< 8.9 Hz, H3 ' ) , 7.96 (dd, 1 H , J 3 > 4 , 8.9, J 4 . , 6 . 2.2 Hz , H4 ' ) , 5.93 (d, 1 H , J , , 2 3.4 Hz, H I ) , 3.83 (dd, 1 H , J 2 , 3 9.8, J 3 , 4 9.8 Hz , H3), 3.72 (dd, 1 H , h,i 3.4, J 2 , 3 9.8 Hz, H2), 3.60 - 3.65 (m, 3 H , H5 , H6a, H6b), 3.43 (dd, 1 H , J 3 ; 4 9.8, J 4 , 5 9.7 Hz, H4), ESI M S [ M + N a ] + 369.1 Anal . Calcd. for C i 2 H i 4 N 2 O 1 0 + 1/2H 2 0: C , 40.57; H , 4.26; N , 7.89. Found: C, 40.73; H , 4.07; N , 7.83 146 3,4-Dinitrophenyl a-D-glucopyranoside (3,4DNPccGlc, 3.3). 3,4-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the acetyl chloride-promoted deacetylation procedure. The reaction was performed for 5 hours. The product was purified by column chromatography (27/3 ethyl acetate/methanol, R f 0.26), and crystallized from methanol/chloroform/hexanes as white plates (65%). m.p. 87 - 89 °C, ' H N M R ( D 2 0 , 300 M H z ) 5 8.06 (d, 1 H , J5>,6> 9.1 Hz, H5 ' ) , 7.62 (d, 1 H , h',6' 2.5 Hz , H2 ' ) , 7.42 (dd, 1 H , J2>,6- 2.5 Hz , J 5 . , 6 . 9.1 Hz , H6 ' ) , 5.74 (d, 1 H , J i , 2 3.5 Hz , H I ) , 3.81 (dd, 1 H , J 3 , 4 9.3, J 2 , 3 9.9 Hz, H3), 3.68 (dd, 1 H , J 2 , 3 9.9, J u 2 3.5 Hz , H2), 3.50 - 3.60 (m, 3 H , H5, H6a, H6b), 3.42 (dd, 1 H , J 3 , 4 9.3, J 4 , 5 9.5 Hz , H4), ESI M S [ M + N a f 369.1 Anal . Calcd. for C i 2 H i 4 N 2 O i o + 1/2H 2 0: C , 40.57; H , 4.26; N , 7.89. Found: C, 40.70; H , 4.33; N , 7.79 2,4,6-Trichlorophenyl a-D-glucopyranoside (TCPocGlc, 3.4). 3.4 2,4,6-Trichlorophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the ammonia-promoted deacetylation procedure. The reaction 147 time was 1.5 hours, T L C was carried out in 27/3 ethyl acetate/methanol (Rf 0.24) and the product directly crystallized from methanol as white needles (85%). m.p. 168 °C, *H N M R ( C D 3 O D , 300 M H z ) 8 7.40 (s, 2 H , H 3 ' , H5 ' ) , 5.90 (d, 1 H , J i , 2 3.9 Hz , HI ) , 4.2 (ddd, 1 H , J 4 , 5 10.0, J 5 , 6 a 3.2, J5,6b 3.0 Hz, H5), 3.90 (dd, 1 H , J 2 , 3 9.8, J 3 , 4 9.2 Hz, H3), 3.65 - 3.75 (m, 2 H , H6a, H6b), 3.60 (dd, 1 H , J , , 2 3.9, J 2 , 3 9.8 Hz , H2), 3.47 (dd, 1 H , J 3 , 4 9.2, J 4 > 5 10.0 Hz, H4) Anal . Calcd. for C ! 2 H i 3 0 6 C l 3 : C , 40.08; H , 3.64. Found: C, 40.09; H , 3.70 4-Chloro-2-nitrophenyl a-D-glucopyranoside (4C2NPaGlc , 3.5). 0 2 N 3.5 4-Chloro-2-nitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the acetyl chloride-promoted deacetylation procedure. Reaction time was 24 hours. Column chromatography was performed with 26/4 ethyl acetate/methanol (Rf 0.20) and the product was crystallized from methanol/chloroform/hexanes as white yellowish plates (46%). m.p. 96.5 - 97 °C, ' H N M R ( D 2 0 , 300 M H z ) 8 7.90 (d, 1 H , J 3 , 5 , 2.6 Hz, H3 ' ) , 7.55 (dd, 1 H , J 3 ' , 5 ' 2.6, J5>,6- 9.1 Hz, H5 ' ) , 7.33 (d, 1 H , J 5 , 6 - 9.1 Hz , H6 ' ) , 5.78 (d, 1 H , J , , 2 3.5 Hz , H I ) , 3.82 (dd, 1 H , J 2 , 3 9.6, J 3 > 4 9.2 Hz, H3), 3.55 - 3.67 (m, 4 H , H2, H5, H6a, H6b), 3.40 (dd, 1 H , J 3 > 4 9.2, J 4 , 5 9.2 Hz , H4). 148 Anal . Calcd. for d 2 H I 4 N 0 8 C l : C , 42.94; H , 4.20; N , 4.17. Found: C, 42.54; H , 4.59; N , 4.18 3,5-Dichlorophenyl a-D-glucopyranoside ( D C P a G l c , 3.8). 3,5-Dichlorophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the ammonia-promoted deacetylation procedure. The reaction was performed for 2 hours. The final product was purified by column chromatography (27/3 ethyl acetate/methanol, R f 0.21) and was obtained as a white solid upon the evaporation of solvent (71%). m.p. 168 °C, ' H N M R ( D 2 0 , 300 M H z ) 5 7.08 (dd, 1 H , J 1.6 Hz, H4 ' ) , 7.05 (d, 2 H , J 1.6 Hz , H 2 ' , H6 ' ) , 5.52 (d, 1 H , J, 2 3.6 Hz, HI ) , 3.78 (dd, 1 H , J 2 , 3 9.5, J 3 , 4 9.3 Hz , H3), 3.55 - 3.64 (m, 4 H , H2, H5, H6a, H6b), 3.39 (dd, 1 H , J 3 , 4 9.3, J 4 , 5 9.6 Hz , H4). Anal . Calcd. for d 2 H 1 4 0 6 C i 2 : C , 44.33; H , 4.34. Found: C, 44.55; H , 4.42 149 3-Nitrophenyl a-D-glucopyranoside ( M N P a G l c , 3.9). 3.9 3-Nitrophenyl 2,3,4,6-tetra-O-acetyl-a-D-glucopyranoside was deacetylated according to the ammonia-promoted deacetylation procedure. The reaction was performed for 2 hours. The product was purified by crystallization from methanol as short yellowish white needles (78%). m.p. 183 - 184 °C, ' H N M R ( D 2 G \ 300 M H z ) 5 7.87 - 7.94 (m, 2 H , H 2 \ H4 ' ) , 7.48 - 7.51 (m, 2 H , H 5 \ H6 ' ) , 5.70 (d, 1 H , Jh2 3.7 Hz , HI ) , 3.88 (dd, 1 H , J2.3 9.7, J3A 9.2 Hz, H3), 3.44 - 3.70 (m, 5 H , H2, H4, H5, H6a, H6b) Anal . Calcd. for C i 2 H 1 5 N 0 8 : C , 47.84; H , 5.02; N , 4.65. Found: C , 48.09; H , 4.95; N , 4.60 5.1.5. Synthesis of (5S)- and (5R)-5-Fluoro-oc-D-Xylopyranosyl Fluorides 5.1.5.1. (5S)- and (5R)-2,3,4-Tri-0-Acetyl-5-Fluoro-a-D-Xylopyranosyl Fluorides (5S)-2,3,4-Tri-0-Acetyl-5-Bromo-a-D-Xylopyranosyl Fluoride. 150 The photobromination procedure was taken from the reference 150. 2,3,4-Tri-O-acetyl-a-D-xylopyranosyl fluoride was dissolved in C C U in the Ace photochemical reaction assembly (Aldrich Chemical Company), followed by the addition of N -bromosuccinimde (NBS) . The resulting mixture was irradiated with a 200 W halogen bulb for 24 hours under a N 2 atmosphere. After 24 hours, the reaction mixture was diluted with CHCI3, filtered and washed with saturated N a H C 0 3 , water and saturated N a C l . The organic layer was dried over MgSCM and concentrated under reduced pressure. The product was purified by column chromatography (50/1 methylene chloride/diethyl ether, Rf 0.36) and crystallized from diethyl ether as yellowish needles (20 %). m.p. 93 -93.5 °C, ' H N M R (CDC1 3 , 300 M H z ) 5 6.45 (d, 1 H , J 4 j 5 4.5 Hz , H5), 5.87 (dd, 1 H , J 2 > 3 10.3, J 3 , 4 10.3 Hz , H3), 5.86 (dd, 1 H , J l i F i 52.6, J U 2 3.2 Hz , H I ) , 4.96 (ddd, 1 H , J 2 ,FI 23.5, J 2 , 3 10.3, J i , 2 3.2 Hz , H2), 4.82 (dd, 1 H , J 3 , 4 10.3, J 4 , 5 4.49 Hz, H4), 2.10 (3s, 9 H , 3 x CH3CO), l 9 F N M R (CDCI3, 282 M H z ) 5 -68.86 (dd, J U F i 52.6, J 2 j F i 23.5 Hz , F l ) Anal . Calcd. for C n H i 4 B r F 0 7 : C, 36.99; H , 3.95; Br , 22.37; F, 5.32; O, 31.36. Found: C, 37.14; H 3.94 (5S)- and (5R)-2,3,4-Tri-0-Acetyl-5-Fluoro-a-D-XylopyranosylFluorides (4.1 and 4.2) F F 4.1 4.2 (5S)-2,3,4-Tri-0-acetyl-5-bromo-a-D-xylopyranosyl fluoride was dissolved in toluene under a N 2 atmosphere. To the solution was added AgBF4 and 4 A molecular 151 sieves, and the resulting mixture was stirred for 50 minutes at room temperature, when T L C (3/1 petroleum ether/ethyl acetate) showed that the starting material had disappeared and two new compounds appeared. The compound corresponding to Rf 0.2 was (5S)-2,3,4-tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluorides (4.1) and the one corresponding to Rf 0.32 was (5R)-2,3,4-tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride (4.2). After adding ethyl acetate the reaction mixture was filtered with Celite 545 (Fisher Chemicals, N J , U S A ) and the filtrate was washed with saturated N a H C 0 3 , water and saturated N a C l . The solvent was evaporated under reduced pressure, and 4.1 and 4.2 were separated by column chromatography (3/1 petroleum ether/ethyl acetate). Each compound was obtained as a white solid. 4.1 and 4.2 were recrystallized from diethyl ether as white plates and white needles, respectively. (5S)-2,3,4-Tri-0-Acetyl-5-Fluoro-a-D-Xylopyranosyl Fluoride (ax-5FccXF-OAc, 4.1) m.p. 103.5 - 104 °C, ' H N M R (CDC1 3 , 300 M H z ) 5 5.79 (dd, 1 H , J 2 , 3 10.4, J 3 , 4 10.4 Hz , H3), 5.78 (m, 2 H , J 1 > F 1 , J 5 , F 5 53.3, J i , 2 , J 4 , 5 3.1 Hz , H I , H5), 4.96 (m, 2 H , J 2 , F 1 , J 4 , F 5 23.3, J 2 > 3 , J 3 > 4 10.4, J , , 2 , J 4 , 5 3.1 Hz, H2, H4), 2.10 (3s, 9 H , 3 x C H 3 C O ) , 1 9 F N M R (CDC1 3 , 282 M H z ) 6 -63.6 (m, JJ.PI, J 5 , F 5 53.3, J 4 , F 5 , J 2 ,FI 23.3 Hz , F l , F5) Anal . Calcd. for C i , H 1 4 F 2 0 7 : C, 44.60; H , 4.76; F, 12.83; O, 37.81. Found: C, 44.54; H 4.78 (5R)-2,3,4-Tri-0-Acetyl-5-Fluoro-a-D-XylopyranosylFluoride (eq-5FaXF-OAc, 4.2) m.p. 107 - 108 °C, ' H N M R (CDC1 3 , 300 M H z ) 8 5.80 (dd, 1 H , J , , F i 53.2, J i , 2 2.6 Hz , H I ) , 5.62 (dd, 1 H , J 5 , F 5 52.3, J 4 , 5 6.3 Hz, H5), 5.45 (dd, 1 H , J 2 , 3 9.2, J 3 , 4 8.8 Hz , H3), 5.17 (ddd, 1 H , J 4 , F 5 12.0, J 4 , 5 6.3 Hz , J 3 , 4 8.8 Hz , H4), 5.11 (ddd, 1 H , J 2 , F , 24.0, J 2 > 3 9.2, 152 Ji,2 2.6 Hz, H2), 2.10 (3s, 9 H , 3 x C H 3 C O ) , U F N M R (CDC1 3 , 282 M H z ) 8 -68.77 (ddd, JI.FI 53.2, J2.F1 24.0, J F i , F 5 8.5 Hz, F l ) , -69.12 (ddd, J5.F5 52.3, J 4 .F5 12.0, JFI.FS 8.5 Hz, F5) Anal . Calcd. for C1H14F2O7: C , 44.60; H , 4.76; F, 12.83; O, 37.81. Found: C, 44.41; H 4.75 (5S)-5-Fluoro-a-D-XylopyranosylFluoride (ax-5FaXF, 4.3) (5S)-2,3,4-Tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride 4.1 was dissolved in M e O H and deacetylated according to the ammonia-promoted deacetylation procedure. The product was purified by column chromatography (27/3 ethyl acetate/methanol, R f 0.21) to yield a white wax. [ H N M R ( D 2 0 , 300 M H z ) 8 5.75 (m, 2 H , J U F i , J 5 , F 5 54.1, J , , 2 , J 4 , 5 3.0 Hz, H I , H5), 3.96 (dd, 1 H , J 2 , 3 9.7, J 3 , 4 9.5 Hz , H3), 3.72 (m, 2 H , J 2 , F i , J 4 , F 5 25.5, J 3 , 4 9.5, J i , 2 , J 4 > 5 3.0 Hz, H2, H4), 1 9 F N M R ( D 2 0 , 282 M H z ) 8 -64.58 (m, J 1 > F i , J 5 , F 5 54.1, J 4 .F5, J2.F1 25.5 Hz , F1 ,F5) Anal . Calcd. for C 5 H 8 F 2 0 4 : C, 35.30; H , 4.74; F, 22.34; O, 37.62. Found: C , 35.51; H , 4.75 (5R)-5-Fluoro-a-D-Xylopyranosyl Fluoride (eq-5FaXF, 4.4) F 4.3 F 4.4 153 (5R)-2,3,4-Tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride 4.2 was dissolved in M e O H and deacetylated according to the ammonia-promoted deacetylation procedure. The product was purified by column chromatography (27/3 ethyl acetate/methanol, Rf 0.35) to yield a white wax. ' H N M R ( D 2 0 , 300 M H z ) 8 5.64 (dd, 1 H , J i , F 1 51.8, J U 2 2.6 Hz , H I ) , 5.40 (dd, 1 H , J 5 , F 5 53.8, J 4 , 5 7.3 Hz , H5), 3.55 - 3.70 (m, 2 H , H2, H3), 3.50 (ddd, 1 H , J 4 j F5 15.0, J 3 , 4 9.14, J 4 , 5 7.3 Hz , H4), 1 9 F N M R ( D 2 0 , 282 M H z ) 8 -71.19 (dd, JI,FI 51.8, J 2 ,FI 29.1 Hz, JFI,FS ~0 H Z , F l ) , -74.61 (dd, J 5 , F 5 53.8, J 4 , F 5 15.0 Hz , F5) Anal . Calcd. for C 5 H 8 F 2 0 4 : C, 35.30; H , 4.74; F, 22.34; O, 37.62. Found: C, 35.79; H , 4.77 154 5.2. Enzymes 5.2.1. Materials and Methods 5.2.1.1. Materials a-Glucosidase from Aspergillus niger was purchased from Megazyme International Ireland Ltd. (Bray, Ireland) and dialyzed against p H 4.5, 0.1 M sodium acetate buffer at 4°C for 48 hours before use. oc-l,4-Glucan lyase from Gracilariopsis sp. was a kindly gift from Dr. Shukun Y u of Danisco Innovation, Denmark. Pepsin (from porcine mucosa) was purchased from Boehringer Mannheim. p-Glucosidase from Agrobacterium sp (Abg) was cloned by Ms . Karen Rupitz. Yeast extracts and Tryptone were purchased from Difco Laboratory, M I , U S A . 40% acrylamide/N,N'-methylene bisacrylamide and N , N , N ' , N ' -tetramethylethylenediamine ( T E M E D ) were purchased from Bio-Rad, C A , U S A . A l l other chemicals and reagents were purchased from Sigma Chemical Co. unless otherwise noted. 5.2.1.2. General Methods The Bradford assay [157] was used for the determination of the concentration of proteins. 0.8% agarose (Bio-Rad, C A , U S A ) gel was used for all electrophoresis of D N A . Sodium dodecyl sulfate-polyacryl amide gel electrophoresis ( S D S - P A G E ) [158] was performed to analyze the homogeneity of proteins. 155 5.2.2. Cloning and Overexpression of the Gene (yicl) Encoding oc-Xylosidase from Escherichia coli. 5.2.2.1. Preparation of Genomic DNA ofE. coli E. coli (K-12) was grown on L B media at 37 °C for 24 hours. Genomic D N A was extracted from bacterial cells using the DNeasy® Tissue K i t ( Q I A G E N Sciences, Maryland, U S A ) , following the procedure provided by the manufacturer. Prepared genomic D N A dissolved in Buffer A E provided by the manufacturer was stored at 4 °C. 5.2.2.2. Polymerase Chain Reaction (PCR) Primers for the amplification of the yicl gene via the polymerase chain reaction (PCR) were designed as the following sequences: 5' - C A G A A T T A A G G A A C G C A T A T G A A A A T T A G C G A T - 3 ' and 5' - A T C A A T C T C G A G C A A C G T A A T T G T C A G C G C - 3 ' , which introduces Ndel and Xhol restriction enzyme cleavage sites at the 5'- and 3'-ends of the yicl gene (underlined). These oligonucleotides were synthesized by the Nucleic Acids Protein Services (NAPS) Unit at the University of British Columbia. P C R was performed using a GeneAmp 2400 P C R machine (Perkin Elmer). In 100 u L of final volume of P C R buffer (pH 8.85, 10 m M T r i - H C l , 25 m M KC1, 5 m M (NFL^SGu, 2 m M M g S 0 4 ) , genomic D N A (100 ng) was used as a template. Pwo D N A polymerase (2.5 units, Roche Applied Science) was used as the enzyme. 20 nmol of each dNTP and 50 pmol of each of the above primers were included. The program used was as 156 follows: denature template for 2 minutes at 94 °C; 10 cycles of 94 °C for 15 seconds, 50 °C for 30 seconds, and 72 °C for 1 minute; 15 cycles of 94 °C for 15 seconds, 50 °C for 30 seconds, and 72 °C for time which started from 50 seconds and increased by 5 seconds each cycle. The reaction mixture was extracted with phenol-chloroform. Amplified D N A was recovered by precipitation with ethanol and 200 m M N a C l . 5.2.2.3. Plasmid Preparation and Cloning of the yicl Gene Encoding a-Xylosidase The amplified yicl gene was digested with the Ndel and Xhol restriction enzymes and then purified by 0.8% agarose gel electrophoresis and subsequent gel extraction using QIAquick® gel extraction kit ( Q I A G E N Sciences, Maryland, U S A ) . Separately, plasmid pET29a was digested with Ndel and Xhol and purified. The resulting gene fragment was ligated into plasmid pET29a using T4 D N A ligase (Roche Applied Science). The ligation mixture was then transformed into E. coli Top 10. Transformation was performed as follows: Host competent cells were mixed with the plasmid and the mixture was kept on ice for 30 minutes, then transferred to a water bath at 42 °C for a heat shock for 2 minutes. The mixture was then quickly moved to ice and allowed to stand on ice for more than 1 minute. The transformants were incubated in L B media at 37 °C for 1 hour and subsequently cultured on L B media plates at 37 °C containing 20 pg/mL kanamycin, an antibiotic. Several colonies were inoculated on the L B media and cultured for 24 hours at 37 °C in the presence of kanamycin. Cultured cells were centrifuged at 10,000 x g and 4 °C for 10 minutes and plasmids were prepared from the cells, using the QIAprep® Spin Miniprep kit ( Q I A G E N Inc). The prepared plasmids 157 were stored at -20 °C and a small portion of these were digested with Ndel and Xhol to confirm that the yicl gene was inserted. D N A sequencing was performed by the N A P S unit at the University of British Columbia. In the constructed vector, pET29EcoXyl31 A , the yicl site is under control of the T7 promoter inducible by isopropyl l-thio-(3-D-galactopyranoside (IPTG) and 6 histidines are appended to the C-terminal of the final expressed protein. 5.2.2.4. Overexpression of yicl and Purification of a-Xylosidase (Yicl) The plasmid, pET29EcoXyl31A, was transformed into E. coli B L 21 (DE3) by electroporation using the Genepulser II (Bio-Rad Inc.). The resulting E. coli B L 21 (DE3) cells containing pET29EcoXyl31A were cultured at 37 °C in 2 L L B media containing 20 pg/mL kanamycin. Expression was induced by the addition of 1 m M IPTG when the culture reached an A6oo level of 0.5. Growth proceeded for 24 - 26 hours and cultured cells were harvested by centrifugation at 10,000 x g and 4 °C for 20 minutes. The resulting cell pellets were resuspended in 50 m L o f the binding buffer (pH 7.9, 0.02 M Tr i s -HCl , 0.005 M imidazole, 0.5 M NaCl) and treated with Benzonase (Novagen). Cells were then homogenized by French Ce l l treatment, followed by centrifugation (10,000 x g, 4 °C, 20 minutes). The supernatant was collected and subjected to purification. The supernatant was applied to a HiTrap™ Chelating H P Column (Pharmacia) which had been equilibrated with NiSGM and binding buffer, sequentially, and was attached to the Fast Protein Liquid Chromatography (FPLC) equipment (Pharmacia). The unbound portion was washed away with washing buffer (pH 7.9, 0.02 M Tr i s -HCl , 0.06 M imidazole, 0.5 M NaCl) and the bound protein was eluted with a linear gradient of 0 to 158 100% elution buffer (pH 7.9, 0.02 M Tr is -HCl , 0.5 M N a C l , 0.5 M imidazole). Fractions containing proteins were pooled and concentrated using a 10 K nominal molecular weight cut-off centrifugal filter, Amicon Ultra (Millipore), followed by buffer exchange into p H 7.0, 0.05 M phosphate buffer. Buffer exchange was performed as follows. Pooled fractions (15 mL) were transferred to a centrifugal filter and centrifuged at 4 °C and 4,000 x g. To the retentate (1 mL) was added p H 7.0, 0.05 M phosphate buffer containing 5 m M E D T A (15 mL) and the resulting solution was concentrated again by centrifugation. This was repeated three times. Then, the same process was repeated three times with p H 7.0, 0.05 M phosphate buffer without E D T A . The homogeneity of the protein was analyzed by 10% SDS P A G E [162], stained with Coomassie blue. Molecular weight markers, a pre-stained SDS P A G E standard (Bio-Rad), contained phosphorylase b 106 kDa, bovine serum albumin 80 kDa, ovalbumin 49.5 kDa, carbonic anhydrase 32.5 kDa, soybean trypsin inhibitor 27.5 kDa, and lysozyme 18.5 kDa. The concentration of the protein was determined by the Bradford assay [163]. The tryptic fragments of the protein were analyzed with electrospray ionization mass spectrometry to confirm the identity of the protein: The protein (1 mg/mL) was incubated with 10 unit of trypsin in p H 7.0 phosphate buffer (1 mL) for 1 hour. The resulting digest was applied to liquid chromatography (LC)/electrospray ionization mass spectrometry (ESI M S ) analysis and masses of fragments were analyzed by Dr. Shouming He. 159 5.2.3. Enzyme Kinetics 5.2.3.1. Michaelis-Menten Kinetics: Spectrophotometric Measurements A l l experiments were carried out at 37 °C in p H 4.5, 0.1 M acetate buffer containing 0.1% bovine serum albumin (BSA) for the a-glucosidase from Aspergillus niger, at 30 °C in p H 6.0, 0.1 M M E S buffer for the a-l,4-glucan lyase from Gracilariopsis sp., and at 37 °C in p H 7.0, 0.05 M phosphate buffer containing 0.1 % B S A for the a-xylosidase from Escherichia coli unless otherwise noted. Michaelis-Menten parameters for the aryl glycosides were measured by monitoring the release of phenol at 400 nm with either a U N I C A M UV7VIS spectrophotometer or a Varian Cary 300 Bio U V / V I S spectrometer equipped with a circulating water bath. Cuvettes of either 1 cm or 0.1 cm path length were used. The typical substrate concentration range employed was 0.3 x Km - 3 x Km. The difference in extinction coefficient, Ae, between each aryl glycoside and substituted phenol at the relevant wavelength at the appropriate p H and temperature was determined by measuring the difference in absorbance between fixed equal concentrations of each pair of phenols and aryl glucosides. Kinetic parameters were obtained by direct fit of the data to the Michaelis-Menten equation using the software GraFit 4.0 (Leatherbarrow, R J . , Erithacus Software Ltd., Staines, U K ) . Kinetic parameters for hydrolysis of 4-nitrophenyl a-D-glucopyranoside (PNPaGlc) at p H 4.5 were determined from the initial linear increase in absorbance at 360 nm upon the addition of A . niger a-glucosidase (final concentration 6.94 pg/ml) to a range of concentrations (0.1 - 1.2 m M ) of P N P a G l c . The difference of extinction 160 coefficients Ae between P N P a G l c and 4-nitrophenol at 360 nm at p H 4.5, 30 °C was determined to be 1.88 (± 0.01) mM^cm" 1 by measuring the difference in absorbance between fixed equal concentrations of 4-nitrophenol and P N P a G l c . Rates were calculated using the determined extinction coefficient and kinetic parameters were obtained by direct fit of the data to the Michaelis-Menten equation using GraFit 4.0. The activity of a-l,4-glucan lyase (GLase) measured with 4-nitrophenyl p - D -maltoside (PNPpMal) as a substrate by the coupled assay using A b g as an auxiliary enzyme was measured as follows. In this coupled assay, GLase produces 4-nitrophenyl p-D-glucopyranoside (PNPPGlc) and A b g hydrolyzes P N P P G l c to generate 4-nitrophenol which is detected spectrophotometrically. Kinetic parameters for cleavage of P N P p M a l were determined from the initial linear increase in absorbance at 400 nm upon the addition of enzyme (final concentration 3.63 pg/ml) to a range of concentrations (0.02 -1.0 m M ) of P N P p M a l containing 4.3 unit of Abg . Control runs in which the amount of A b g was doubled revealed no increase in rate, while doubling the GLase concentration doubled the rate. This confirms that the GLase, not Abg , was rate-limiting. The difference in extinction coefficients Ae between P N P p M a l and 4-nitrophenol at 400 nm at p H 6.0, 30 °C was determined to be 1.66 (± 0.01) mM^cm" 1 by measuring the difference in absorbances between fixed equal concentrations of 4-nitrophenol and P N P p M a l . Rates were calculated using the determined extinction coefficient and kinetic parameters were obtained by direct fit of the data to the Michaelis-Menten equation using GraFit 4.0. 161 5.2.3.2. Michaelis-Menten Kinetics: Simple Glycosyl Fluorides Kinetic parameters with simple glycosyl fluorides (a-D-glucopyranosyl fluoride or oc-D-xylopyranosyl fluoride) were determined as previously described [112]. A range of concentrations of the substrates were prepared and in the presence of the desired concentrations of enzyme, the release of fluoride was monitored using an Orion 96-09 combination fluoride ion electrode interfaced to a computer running the LoggerPro software (Vernier Software Ld.). Initial rates were used for the determination of kinetic parameters, which were obtained by direct fit o f the data to the Michaelis-Menten equation using GraFit 4.0. 5.2.3.3. pH-Dependence Studies Measurement of the pH-dependent activity of a-1,4-glucan lyase (GLase) was carried out spectrophotometrically at 400 nm using 2,4-dinitrophenyl a -D-glucopyranoside (2 ,4DNPaGlc) as substrate, using the following buffers: Glyc ine -HCl (pH 2.42 - 3.5), sodium acetate (pH 4.0 - 5.5), M E S (pH 6.0), M O P S (pH 6.5 - p H 7.5), G l y - G l y (pH 8). A l l buffers were 0.05 M and contained 0.05 M N a C l . The substrate-depletion method was employed as follows [142]. A solution of substrate (0.03 m M , 100 fold lower than the value of Km) was preincubated at p H 6.0 and 30 °C, then the release of 2,4-dinitrophenol after addition of GLase was monitored using a U N I C A M U V / V I S spectrophotometer equipped with a circulating water bath until at least 80% depletion of substrate. 1 cm or 0.1 cm path length cuvettes were used. Appropriate controls confirmed 162 that, at 30 °C, GLase is stable over the reaction time. Fitting of the data to a first order equation (GraFit 4.0), yielded an apparent rate constant for the reaction, from which the kcJKm value for each substrate was calculated by dividing that rate constant by the concentration of GLase. Obtained kcJKm values were then plotted versus p H and fitted to the appropriate curve using GraFit 4.0, thereby yielding apparent pKa values. Measurement of the pH-dependent activity of a-xylosidase was carried out likewise except that the first-order fluoride release was measured with an Orion 96-09 combination fluoride ion electrode in this case. 5.2.3.4. Bronsted Analysis Michaelis-Menten parameters for the hydrolysis of various aryl glucosides by GLase at 30 °C were determined by monitoring the release of phenols at the appropriate wavelength (Table 3.2) with a UV7VTS spectrophotometer equipped with a circulating water bath. The typical substrate concentration range employed was 0.3 x I m - 3 x Km. The difference in extinction coefficient, Ae, between each aryl glucoside and substituted phenol at the relevant wavelength at p H 6.0, 30 °C was determined by measuring the difference in absorbance between fixed equal concentrations of each pair of phenols and aryl glucosides. Data are presented in Table 3.2. Kinetic parameters were obtained by direct fit of the data to the Michaelis-Menten equation. Logarithms of obtained kinetic parameters (& c a t and kcJKm) were fitted linearly with leaving group pKa values. The Bronsted coefficient, p i g , was obtained from the slope of this plot. A l l data fitting was performed using GraFit 4.0. 163 5.2.3.5. Inhibitor Studies Inhibition studies were performed by measuring the activity of enzyme (A. niger a-glucosidase or a-1,4-glucan lyase) in the presence of various concentrations of each inhibitor, using 2 ,4DNPaGlc as substrate. A. niger a-glucosidase (3.4 pg/mL) or GLase (9.8 pg/mL), preincubated at 37 °C (a-glucosidase) or 30 °C (GLase), was added to 200 p L of buffer solution containing 2 ,4DNPaGlc and varying amounts of inhibitors, also preincubated at the corresponding temperature. The release of 2,4-dinitrophenol was monitored spectrophotometrically at 400 nm. The experiments were repeated at different concentrations of 2 ,4DNPaGlc . A Dixon plot of 1/v versus inhibitor concentration intersects a line given by 1 / V m a x at an inhibitor concentration equal to -Kj. Some K\ values were also obtained by fitting the data to the equation of competitive inhibition. 5.2.3.6. Fluorosugar Kinetics: Difluoro Sugars The kinetic parameters for the breakdown of difluoro sugars by A. niger a-glucosidase, a-l,4-glucan lyase and a-xylosidase were measured according to the procedure described in 5.2.3.2 except for the following: Since 2 equivalents of fluoride are produced upon hydrolysis of difluorosugars, rates were divided by 2 [114]. 1) Values of kcat for the Hydrolysis of 5FaGlcF by A. niger a-glucosidase . A s initial rates of fluoride release at all concentrations (0.35 - 2.67 m M ) assayed were identical, this rate was taken as V m a x and kC3t was calculated from V m a x . 164 2) Kinetic Parameters for the Reaction of a-l,4-Glucan Lyase with lFGlcF, 2FaGlcF and 5FaGlcF Measured initial rates at various substrate concentrations were used for the determination of kinetic parameters. For l F G l c F and 2FaGlcF , no saturation was observed up to a concentration of 100 m M . Therefore, only kcJKm values could be obtained from the slope of the initial low concentration part of the plot. A A G * was calculated according to the equation, A A G * = R T In [(kcJKm)f/(kcat/Km)f2] where (£ c a t / .rvm)F is determined for a G l c F and (kcJKm)v2 for either l F G l c F or 2FaGlcF . Kinetic parameters for 5 F a G l c F were determined by fitting the data to the Michaelis-Menten equation, using GraFit 4.0. 3) Values of kcat far the Hydrolysis of (5S)-5-Fluoro-a-D-Xylopyranosyl Fluoride (4.3) and (5R)-5-Fluoro-cc-D-XylopyranosylFluoride (4.4) by a-Xylosidase A s initial rates of fluoride release at all concentrations of 4.3 (0.1 - 2 m M ) and 4.4 ( 0 . 1 - 1 . 2 m M ) assayed were identical, this corresponding rate for each substrate was taken as the V m a x value and each kcat value was calculated from these V m a x values by dividing each V m a x by the corresponding enzyme concentration. A A G * was calculated according to the equation, A A G * = R T In [(kcJKm)F/(kcat/Km)F2] where (kcJKm)F is determined for ccXylF and (kcJKm)F2 for either 4.3 or 4.4. However, since Km values for 4.3 and 4.4 were not determined, KC values for these were used instead. 5.2.3.7. Inactivation Kinetics For analysis of inactivation kinetics, the enzyme (final concentration 0.2 - 3.6 mg/ml) was pre-incubated with a range of concentrations of an appropriate mechanism 165 based inactivator (5FccGlcF, 5F(3ldoF, l F G l c F , 2 F a G l c F , 4.3 or 4.4) at 37 °C (A. niger a-glucosidase or E. coli a-xylosidase), 30°C (a-1,4-glucan lyase) or 10 °C (E. coli a-xylosidase). 10 p i aliquots of the sample were withdrawn at time intervals and added to 500 - 800 p i of an appropriate substrate whose concentration at least equaled or was higher than the Km value of the substrate. Each substrate solution was pre-equilibrated at appropriate temperature in the U V / V i s i b l e spectrometer. The residual enzyme activity at each time interval at each concentration of inactivator was measured in this way. Pseudo-first order rate constants at each inactivator concentration (k0bS) were determined by fitting each curve to a first order rate equation. Values for the inactivation rate constant (k\) and the inactivator dissociation constant (K\) were determined by fitting k0bS values and inhibitor concentrations to the equation. Reactivation of the inactivated enzyme was carried out as follows. Enzyme (100 p L , 0.30 mg/mL) fully inactivated by the inactivator as above was concentrated using 10 kDa nominal cut-off centrifugal concentrators (Amicon Corp., Danvers, M D ) to a volume of approximately 40 p L and diluted with 1000 u L of buffer. This was repeated twice, and the retentate was diluted to a final volume of 100 p L . The inactivated enzyme was then incubated at 30 °C and reactivation was monitored by removal of aliquots (10 pL) at appropriate time intervals and assayed as described above. Measured activities were corrected for decreases in activity due to denaturation over this time course using data for 166 noninactivated control samples. The reactivation rate constant, Ar react, was determined by fitting the data to a first order rate equation, as described above. 5.2.3.8. Kinetic Isotope Effects (KIE) The KIEs upon kcat of P N P l-[ 2H]-ocGlc and P N P 2-[ 2H]-ocGlc were determined by measuring initial rates of GLase with each protio and deutero substrate by monitoring the release of P N P from each substrate at a concentration of 20 m M (Km = 2.1 m M ) as described above and dividing the rate for the protio-substrate by the rate for the deutero-substrate in each case. The KIEs upon kcat for l - [ 2 H]-aGlcF and 2-[ 2 H]-aGlcF was determined likewise, except at a concentration of 270 m M (Km = 27.9 m M ) , by monitoring the release of fluoride as described above. In addition, the KIEs upon kcat for l - [ 2 H]-5FaGlcF and 2-[ 2 H]-5FaGlcF were determined likewise, except at a concentration of 80 m M (Km = 10.7 mM) . Each measurement was repeated at least 6 times. The KIEs upon kcat/Km of l - [ 2 H ] - a G l c F and 2-[ 2 H]-aGlcF were determined as follows. Each protio- and deutero-substrate solution was prepared at a concentration of 1 m M , far below the value of Km (28 m M ) , and in the presence of GLase the release of fluoride was monitored using an Orion 96-09 combination fluoride ion electrode until 80% depletion of substrate. B y fitting data to a first order depletion equation (GraFit 4.0, Erithacus Software Ltd., Staines, U K ) , the kcat/Km value for each substrate was obtained. KIEs were determined by dividing the kcat/Km value for the protio-substrate by that for each deutero-substrate. 167 5.2.4. Trapping of the Covalent Glycosyl-Enzyme Intermediate and Identification of the Labeling Site by Tandem Mass Spectrometry 5.2.4.1. Labeling and Proteolysis for Electrospray Mass Spectrometry a. a-Glucosidase from A. niger A stock solution of the enzyme (20 u.L, 8.13 mg/ml) was incubated with 5FcxGlcF (20 pL, 20 m M ) at 37 °C for 5 min. The sample was diluted with 0.05 M phosphate buffer (pH 2, 90 uL) and incubated with pepsin (15 p L , 1 mg/ml) for 15 min at room temperature. The sample was then frozen quickly and analyzed immediately upon thawing. A control sample was prepared according to the same procedure, except that no 5FocGlcF was added. b. a-l ,4-Glucan Lyase from Gracilariopsis sp. A stock solution o f the enzyme (20 uL, 11.89 mg/ml) was incubated with 5FocGlcF (20 pL , 30 m M ) , 5FpIdoF (20 uL, 80 m M ) , l F G l c F (20 (iL, 100 mM) or 2FccGlcF (20 uL, 100 m M ) at 30 °C for 30 min. The sample was diluted with 0.05 M phosphate buffer (pH 2, 90 pi) and incubated with pepsin (15 | i l , 1 mg/ml) for 15 min at room temperature. The sample was then frozen quickly and analyzed immediately upon thawing. A control sample was prepared according to the same procedure, except that no inactivator was added. c. a-Xylosidase from E. coli A stock solution of the enzyme (20 pL, 8.8 mg/ml) was incubated with (5S)- or (5R)-5FocXylF (20 pL , 10 m M ) at 37 °C for 30 min. The sample was diluted with 0.05 168 M phosphate buffer (pH 2, 90 pi) and incubated with pepsin (15 u\L, 1 mg/ml) for 15 min at room temperature. The sample was then rapidly frozen and analyzed immediately upon thawing. A control sample was prepared according to the same procedure, except that no inactivator was added. 5.2.4.2. Mass Spectrometry Mass spectrometry was performed by Dr. Shouming He. For the study of the oc-glucosidase from A. niger and the a-l,4-glucan lyase from Gracilariopsis sp., mass spectra were recorded on a PE-Sciex A P I 300 triple quadrupole mass spectrometer (Sciex, Thornhill, O N , Canada) equipped with an electrospray ionization ion source. Peptides were separated by reverse phase high performance liquid chromatography ( H P L C ) on an Ultrafast Microprotein Analyzer (Michrom BioResource Inc., Pleasanton, C A , U S A ) directly interfaced with the mass spectrometer. In each M S experiment, the protein sample or the proteolytic digest was loaded onto a C I 8 column (Reliasil, 1x150 mm) and eluted with a gradient of 0 - 60 % eluting solvent (0.045% trifluoroacetic acid, 80% acetonitrile in water) over 60 min at a flow rate of 50 pL/min. A post-column splitter was used in all experiments, splitting off 85% of the sample into a fraction collector and sending 15% into the mass spectrometer. Spectra were obtained in either the single-quadrupole scan mode ( L C / M S ) or the tandem M S daughter ion scan mode ( M S / M S ) . In the single quadrupole mode ( L C / M S ) , the quadrupole mass analyzer was scanned over a m/z range of 400-2000 Da with a step size of 0.5 Da and a dwell time of 1.0 ms/step. The ion source voltage was set at 5 k V and the orifice energy was 50 V . 169 After the L C / M S experiment, total ion chromatograms of the labeled and unlabeled enzyme digests were compared to find the fraction containing the labeled peptide fragments. Samples of the labeled peptide were collected from the post-column flow splitter and lyophilized. The concentrated sample was then sequenced via tandem M S fragmentation analysis. In the tandem M S daughter ion scan mode, mass spectra were obtained by selectively introducing the m/z 1011 peptide (a-glucosidase from A. niger) or m/z 1206 or 1205 (a-l,4-glucan lyase from Gracilariopsis sp.) from the first quadrupole (QI) into the collision cell (Q2) and observing the daughter ions in the third quadrupole (Q3). The following settings were applied: QI was locked on m/z 1011 (a-glucosidase from A. niger) or m/z 1206 or 1205 (a-1,4-glucan lyase from Gracilariopsis sp.); Q3 scan range m/z 100-1020 (a-glucosidase from A. niger) or m/z 100-1300 (a-l,4-glucan lyase from Gracilariopsis sp.); step size, 0.5; dwell time 1 ms; ion source voltage was 5 k V ; orifice voltage was 50 V ; the focusing ring voltage was 200 V ; Q0 potential was -10 V ; Q2 potential was -42 V ; the collision gas was N 2 . For the study of the a-xylosidase from E. coli, mass spectra were recorded using an A B 1 M D S - S C I E X A P I Q S T A R Pulsar i mass spectrometer (Sciex, Thornhill, ON) . Peptides were separated on a reverse phase C I 8 column using an Ultimate Capillary H P L C system (LC Packings, Amsterdam, Netherlands) interfaced with the mass spectrometer. In L C / M S experiments, proteolytic digests of the protein were loaded onto a C I 8 column (300 | im x 150 mm) and eluted with a gradient of 2 to 40% eluting solvent (0.1% formic acid and 85% acetonitrile in water). The mass analyzer was scanned over a mass-170 to-charge ratio range of 300 - 2400 amu, with a step size of 0.1 amu and a scan time of 1 second. The ion source potential was set at 5 k V ; the orifice energy was 50 V . To determine the amino acid sequence, the mass spectrometer operated in an I D A (information dependent acquisition) M S / M S mode, where the precursor ion is selected "on the fly" from the previous scan. A n m/z ratio (902 in this case) for an ion that had been selected for fragmentation was placed in a list. The peptides including m/z 902 (doubly charged) previously fractionated by H P L C were introduced into the mass spectrometer via a nanospray ion source (Protana, Staermosegaardvej, Denmark). Following mass selection in the first quadrupole ( Q l ) , the peptide of interest was fragmented by collision with nitrogen gas in the second quadrupole (Q2) and the resulting product ions were analyzed in the T O F mass analyzer. The following settings were used: T O F scan range of m/z 100 - 1820 amu, step size of 0.1 amu, and the scan time of 1 second, Q2 potential of -42 V and source voltage of 1000 V . Appendix A. Crystal Structure of 5-Fluoro-Xylosyl Fluorides Conformational Analysis 172 A. 1. Anomeric Effect The anomeric effect is often called the Edward-Lemieux effect, following the names of the initial advocates [159,160], and concerns the phenomenon that an electronegative substitutent at C I in a pyranose ring prefers an axial orientation over an equatorial orientation (Figure A . l A ) . O O C H , axial O equatorial O C H 3 B R + Y gauche anti O: element X Figure A . l . Schematic diagram of the anomeric effect: A. anomeric effect in tetrahydropyran system; B. generalized anomeric effect (GAE) Thus for D-series hexopyranoses, the a-anomer should be more conformationally stable than the [3-anomer. This is opposite to predictions based solely on steric interactions, since bulky anomeric substituents should prefer the equatorial orientation due to unfavorable 1,3-diaxial interactions. Later, the anomeric effect was expanded beyond cyclic pyranoses and sugars to include other molecules whose components are R - X - A - Y , where A is an element of intermediate electronegativity, Y is more electronegative than A , 173 X possesses lone pairs and R stands for H or C (Figure A . I B ) . This effect was termed the generalized anomeric effect ( G A E ) . The G A E predicts that Y and R prefer a synclinal (gauche) position to the antiperiplanar (anti) position. The anomeric effect has been covered by some excellent reviews and monographs [160-164]. The origin of the anomeric effect has been the subject of debate. While many hypotheses have been proposed to explain it, the most accepted rationalizations have converged into two models, both of which relate to the importance of the participation of the lone pairs on the element X in Figure A . I B . Figure A.2. Schematic diagram of the origin of the anomeric effect: A. dipole-dipole interaction model; B. hyperconjugation ofno-cr*c-x model. In the first model (Figure A . 2 A ) , for example of a pyranose ring, an unfavorable, repulsive interaction between the ring dipole, generated by the lone pair electrons of the endocyclic oxygen, and the nearly parallel polar bond in the equatorial conformer at C I is suggested as an origin of the anomeric effect. This unfavorable interaction does not exist in the axial conformer, leading to higher stability of this conformer. This effect w i l l be B . 174 reduced in a solvent with a higher dielectric constant. Consistent with this explanation, axial preference decreases with increasing solvent dielectric constant (Table A . l ) . Table A . l . Solvent Dependence of the Conformational Equilibrium of 2-Methoxytetrahydropyran [165] OMe Solvent Dielectric constant (e) % axial conformer CCL, 2.2 83 Benzene 2.3 82 C S 2 2.6 80 CHC1 3 4.7 71 Acetone 20.7 72 Methanol 32.6 69 Acetonitrile 37.5 68 Water 78.5 52 This electrostatic model also explains the formation of stronger anomeric effects with more electronegative anomeric groups X . However, this model fails to explain the variations of bond lengths and angles accompanying the anomeric effect. The bond lengths and angles vary in a characteristic fashion accompanying the anomeric effect. In the axial form, the endocyclic C l - 0 bond tends to shorten significantly while the exocyclic C l - X bond lengthens in the presence of an electronegative substitutent, compared to the equatorial form. Also, larger bond angles of O - C l - X and C - O - C l are observed in the axial form than the equatorial form. The alternative hyperconjugation or 175 double-bond-no-bond resonance model emerged to explain this. According to this model, the stabilization of the axial conformer is attributed to the derealization of electrons in a lone-pair orbital on oxygen into the antibonding orbital of the C - X bond (Figure A . 2 B ) . This interaction (n-cj*) results in a shortening of the C l - O bond due to the increased bond order and the opposite effect on the C l - X bond. Further, O and C I assume partial sp character, thereby increasing bond angles. With the advent of these influential models, many experimental and theoretical results have been presented to advocate each school. However, as these results have accumulated, it has been obvious that no single factor is predominant in the anomeric effect and both contribute substantially. A.2. The Anomeric Effect in Xylosyl Halides x OAc I OAc X : CI or F Figure A.3. Conformational equilibrium between 4 C , and *C 4 of 2,3,4-tri-0-acetyl-(3-D-xylopyranosyl chloride and fluoride in solution. The strong tendency of highly electronegative halogens in glycosyl halides to adopt an axial orientation is a strong manifestation of the anomeric effect. D-Xylopyranosyl halides are molecules that particularly reflect the great influences of the anomeric effect on conformations due to their relatively high conformational flexibility. In the cases of 176 2,3,4-tri-O-acetyl-p-D-xylopyranosyl chloride and fluoride, the tendency towards an axial orientation of the halogen is so strong that the molecules adopt sterically unfavorable 'C4 conformations in chloroform and acetone, respectively, instead of the otherwise expected 4 C i conformation as determined by N M R [166,167] (Figure A.3) . This illustrates the power of the anomeric effect from a single halogen overcoming destabilizing 1,3-diaxial interactions resulting from bulky acetyl groups. Crystal structures of xylosyl fluorides have provided considerable insight into variations in bond lengths and angles as a consequence of anomeric effects. 3-Dimensional X-ray crystal structures o f 2,3,4-tri-O-acetyl-p-D-xylopyranosyl fluoride ( p X y l F - O A c ) and its cc-anomer (c tXylF-OAc) were determined [168,169]. In contrast to its solution conformation, the p-anomer adopts a 4 C i conformation in the crystal with fluorine taking the equatorial orientation. On the other hand, 2,3,4-tri-O-benzoyl-P-D-xylopyranosyl fluoride (pXylF-OBz) was shown to adopt a *C4 conformation in the crystal, placing fluorine in an axial orientation [168]. The axial C-F bond in 2,3,4-tri-O-acetyl-a-D-xylopyranosyl fluoride and 2,3,4-tri-O-benzoyl-p-D-xylopyranosyl fluoride are 1.397 and 1.396 A long, respectively. These values are significantly longer than that of the equatorial C-F bond in 2,3,4-tri-O-acetyl-P-D-xylopyranosyl fluoride, 1.367 A . The 05-C1-C2 and C 1 - 0 - C 5 bond angles are also informative. Two compounds, a X y l F - O A c and p X y l F - O B z have angles o f 111.9 and 114.7 degrees for C 1 - 0 - C 2 , and 113.7 and 112.6 (or 114.5) degrees for C 1 - 0 - C 5 , respectively. On the other hand, C 1 - 0 - C 2 is 108.6 degrees and C1-0 -C5 is 109.9 degrees in p X y l F - O A c . Difference in bond lengths between the axial and equatorial C -F bond at the anomeric center clearly show the importance of the double-bond-no-bond resonance in the anomeric effect. Larger bond 177 angles in a X y l F - O A c and p X y l F - O B z are also indicative of sp 2 character in the anomeric carbon and the endocyclic oxygen. A.3. Analysis of Crystal Structures of 5-Fluoro-a-D-Xylopyranosyl Fluorides A n electronegative heteroatom at C5 of a xylosyl unit converts that carbon center to an anomeric center. Thus such a molecule possesses two anomeric centers sharing a common oxygen. These therefore represent an intriguing class of molecules with which to probe anomeric effects and the effects of anomeric substituents thereon. Through these studies, it should be possible, in a self-consistent manner, to probe the angular dependence of these electronegative effects. Previously, Dr. David J. Vocadlo of this laboratory has synthesized and crystallized an equatorial 5-fluoro derivative, (5R)-2,3,4-tri-0-acetyl-5-fluoro-|3-D-xylopyranosyl fluoride (eq-5F(3XF-OAc) and a 5,5-difluoro derivative, 2,3,4-tri-O-acetyl-5,5-difluoro-(3-D-xylopyranosyl fluoride (5,5F2(3XF-OAc). The crystal structures of the two compounds were determined (Figure A . 4 C , D) . In the present work, the two epimers (axial or equatorial fluorine at C5), (5S)- and (5R)-2,3,4-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride were synthesized (ax-5FocXF-OAc and eq - 5 F a X F -OAc , respectively). Both epimers were nicely crystallized and X-ray structures were determined (Figure A . 4 A and A . 4 B , respectively). Experimental details of X-ray crystallography are summarized in Table A . 2 . Selected data including bond lengths and 178 bond angles are presented in Figures A.5 and A .6 , and torsion angles are tabulated in Table A . 3 . Also included in Figures A.5 and A .6 , and Table A.3 are corresponding structural data of 2,3,4-tri-O-acetyl-a-D-xylopyranosyl fluoride and its (3-anomer and 2,3,4-tri-0-benzoyl-(3-D-xylopyranosyl fluoride excerpted from references 168 and 169. Table A . 2 . Experimental Details of X-Ray Crystallography A . Crystal Data ax-5FccXF-OAc eq -5FaXF -OAc eq-5F(3XF-OAc 5,5F 2|3XF-OAc Empir ical formula C11H14F2O7 C i i H 1 4 F 2 0 7 C11H14F2O7 C n H 1 3 F 3 0 7 Formula weight 296.22 296.22 296.22 314.21 Crystal dimension 0.5 x 0.4 x 0.2 m m 0.5 x O . 1 5 x O . l mm 0.15 x 0.3 x 0.35 m m 0 . 1 x 0 . 2 5 x 0 . 3 m m Crystal system Tricl inic Tr ic l inic Monocl in ic Monocl inic Lattice type Primitive Primitive Primitive Primitive Lattice parameters a = 8.6224 (8) A b = 13.421 (2) A a = 7.9989 (9) A b - 7.3754 (8) A a = 8.529 (2) A b = 16.795 (5) A a = 5.6002 (9) A b = 13.903 (5) A c = 18.027 (1) A c = 11 .074(1) A c = 9.3511 (4) A c = 9.2080 (8) A a =85.35 ( 1 ) ° o t = 9 0 ° P = 91.6948 ( 1 1 ) ° p= 1 0 3 . 0 4 0 ( 2 ) ° P = 75.23 ( 1 ) ° p = 9 1 . 6 5 ( 1 ) ° V = 1338.9(4) A3 V = 698.4 (2) A 3 7 = 8 3 . 4 2 ( 1 ) ° V = 2000.9 (4) A 3 7 = 9 0 ° V = 653.1 (1) A 3 Space group P i (#2) P i (#2) P2 , /c (#14) P2, (#4) Z value 6 2 4 2 D Calc 1.475 g /cm3 1.510 g /cm 3 1.469 g W 1.494 g /cm 3 Fooo 924.00 308 616 324 ^ ( M o K a ) 1.39 c m - ' 1.43 cm"' 1.39 cm"1 1.48 cm"' B. Data Acquisit ion Diffractometer R i g a k u / A D S C C C D Radiation M o K a (k = 0 .71069 A), graphite monochromated Detector aperture 94 m m x 94 m m Data images 460 exposure @ 12 seconds 460 exposure @ 59 seconds 460 exposure @ 30 seconds 462 exposure @ 8 seconds (> oscillation range (X = -90.0) 0 . 0 - 190 .0° 0 . 0 - 1 9 0 . 0 ° 0 . 0 - 1 9 0 . 0 ° 0 . 0 - 1 9 0 . 0 ° co oscillation range (X = -90.0) - 1 7 . 0 - 2 3 . 0 ° - 1 7 . 0 - 2 3 . 0 ° - 2 2 . 0 - 1 8 . 0 ° - 2 3 . 0 - 1 8 . 0 ° Detector position 38.79 mm 38.83 m m 39.15 m m 38.84 m m Detector swing angle - 5 . 5 2 ° - 5 . 6 0 ° - 5 ° - 1 0 . 0 ° 20max 5 5 . 7 ° 59.1 0 5 0 ° 6 1 . 2 ° N o . o f reflections Total: 17083 Total: 44912 Unique: 2330 Total: 6461 Unique: 8015 (R,„ , = 0.045) Unique: 2341 ( R ; „ , = 0.056) Unique: 1884 (R,„ , = 0.056) Corrections Lorentz polarization Lorentz polarization Lorentz polarization Lorentz polarization Absorption/scaling /decay (corr. factors: Absorption/scaling /decay (corr. factors: 0.6703 Absorption/scaling (trans. Factors: 0.684 - 1.005) Absorption/scaling /decay (corr. factors: 0 .8516 -0 . 6 7 7 4 - 1.0000) - 1 . 0 0 0 0 ) Secondary extinction (coeff.: 1.14 (8) x 10"°) 1.0000) 179 Table A .2 . (continued) Experimental Details of X-Ray Crystallography C . Refinement ax-5FoXF-OAc eq-5FaXF-OAc eq-5FpXF-OAc 5,5F2(3XF-OAc Structure solution Direct method (SI R97) Refinement Ful l matrix least squares F 2 Function minimized Zco(F02-Fc2)2 Least square weights co= l/cr^FoV (0.0732P) 2 where P = (Max (F 0 2 ,0 ) + 2Fc 2 ) /3 co= l / a W ) + (0.0716P) 2 + 0.0258P where P = (Max (F 0 2 ,0 ) + 2Fc 2 ) /3 co= l / o 2 ( F 0 2 ) = [ o c 2 ( F 0 2 ) + p 2 Fo 2 /4 ] - ' co- l / a W ) Anomalous dispersion A l l non-hydrogen atoms N o . observations (I > O.OOa(r)) 8015 2341 2330 3185 No . variables 548 183 182 189 Reflection/parameter Ratio 14.63 12.80 16.85 Residuals (refined on F 2 , all data): R l ; w R 2 0.077; 0.173 0.045; 0.099 0.080; 0.065 0.093; 0.062 Goodness o f fit indicator 1.10 1.02 1.76 1.32 M a x shift/error in final cycle 0.00 0.00 0.0006 0.0008 N o . observation (I>2o-(I)) 6089 2072 1261 Residuals (calculated on F, I>2o-(I)): R; R w 0.060; 0.157 0.039; 0.096 0.045; 0.031 0.045; 0.028 M a x i m u m peak in final difference map 0.32 e" / A 3 0.42 e" / A 3 0.57 e / A 3 0.51 e V A 3 M i n i m u m peak in final difference map -0.32 e ' / A 3 -0.15 e" /A 3 -0.59 e - / A 3 -0.60 e" / A 3 180 Figure A.4. X-ray crystal structure of per-O-acetyl difluoro derivatives of xylopyranose: A. (5S)-2,3,4-tri-O-acetyl-5-fluoro-a-D-xylopyranosyl fluoride, ax-5FaXF-OAc; B. (5R)-2,3,4-tri-0-acetyl-5-fluoro-a-D-xylopyranosyl fluoride, eq-5FocXF-OAc 181 Figure A.4. (continued) X-ray crystal structure of per-O-acetyl difluoro derivatives of xylopyranose: C. (5R)-2,3,4-tri-0-acetyl-5-fluoro-p-D-xylopyranosyl fluoride, eq-5F(3XF-0Ac; D. 2,3,4-tri-0-acetyl-5,5-difluoro-P-D-xylopyranosyl fluoride, 5,5F2pXF-0Ac. 182 1.362 (5) ax-5FocXF-OAc: 4 C , 5,5F2pXylF-OAc: 4 C , Figure A.5. Schematic diagram showing the conformations and selected bond lengths of various per-O-acetyl xylosyl fluorides. Data are excerpted from X-ray crystal structures of each compound. Data for compounds fSXylF-OAc and pXylF-OBz are quoted from reference 168, and those for aXylF-OAc from reference 169. Bond lengths are quoted in A. 183 OAc OAc eq-5F(3XF-OAc: ] H 3 5,5F2pXylF-OAc: 4 C , Figure A.6. Schematic diagram showing selected bond angles (05-C1-C2, C1-05-C5, and 05-C5-C4) excerpted from X-ray crystal structures of various per-O-acetyl xylosyl fluorides. Data for compounds pXylF-OAc and pXylF-OBz are quoted from reference 168, and those for ocXylF-OAc from reference 169. Angles are quoted in degrees. 184 Table A . 3 . Key Torsion A n g Excerpted from Crystallogra les of 2,3,4-Tri-0-Acetyl-5-Fluoro-D-Xylosyl Fluorides phic Data Compounds F l - C I - 0 5 - C 5 C I - 0 5 - C 5 - F 5 p X y l F - O A c ( 4 C,) N A o X y l F - O A c ( 4Ci) N A p X y l F - O B z C ' C ^ N A ax -5FaXF-OAc ( 4 C,) 71.6 (2) 70.2 (3) 70.9 (2) -71.1 (2) -70.9 (3) -74.0 (2) eq-5FocXF-OAc ( 4 C,) 62.3 (2) 173.87 (18) eq-5FpXF-OAc ( 'H 3 ) -61.9(3) 96.8 (3) 5 , 5 F 2 X F - O A c ( 4Ci) -175.5 (4) -177.0 (4) equatorial F -63.7 (5) axial F A l l compounds but two, p X y l F - O B z and eq-5FpXF-OAc, adopt a 4 Q conformation. A s noted earlier, P X y l F - O B z adopts a ' C 4 conformation and a substantially distorted ] H 3 conformation was observed for eq-5FpXF-OAc (Figure A.4) . OAc Figure A.7. Unfavorable dipole-dipole interaction between three parallel dipoles in the hypothetical 4C\ conformation of eq-5FbXF-OAc. This is quite remarkable since the corresponding mono fluoride p X y l F - O A c adopts a normal 4 C i chair. This is most likely a consequence of the powerful effects of unfavorable dipole-dipole interactions. If eq -5FpXF-OAc were to adopt a 4 C i conformation, the additional fluorine would generate a dipole (C5—>F) approximately parallel to the two existing dipoles (lone pairs of oxygen and CI—»F) which would be destabilizing (Figure A . 7). It appears that the molecule distorts to relieve these 185 interactions, resulting in a half chair conformation. It does not assume a *C4 chair conformation as in p X y l F - O B z since, even though the anomeric effects would be favorable, there would be two adjacent C—>F dipoles in an axial orientation along with three bulky acetyl groups. These unfavorable dipole-dipole interactions seem to be relieved by the addition of a dipole in the opposite direction since 5,5F2XF-OAc assumes a 4 C i chair conformation in which an additional fluorine at C5 in axial orientation makes a dipole in the opposite direction to other three dipoles. Two 5-fluoro-a-D-xylosyl fluorides, ax-5FaXF-OAc and eq-5FaXF-OAc, assume conformations close to 4 Q chair as expected. While both fluorines are in axial orientations in ax-5FocXF-OAc, the influence of dipole-dipole interactions on the conformation can also be discerned. Both C - F bonds are oriented somewhat outwards and displaced slightly from each other as seen in the torsion angles, F1-C1-05-C5 and C l -05-C5-F5, which are almost 10 degrees larger than the corresponding torsion angles involving axial C-F bonds found in eq-5FocXF-OAc, eq-5FpXF-OAc, and 5,5F2XF-OAc. This seems to be the result of repulsion between two adjacent parallel C—>F dipoles (Figure 4.8). In contrast to ax-5FocXF-OAc, eq-5FaXF-OAc does not show significant deviations from the standard 4 C i conformation, probably due to relatively well balanced dipoles with F l in the axial orientation and F5 in the equatorial orientation. Figure A.8. Unfavorable dipole-dipole interaction between two parallel C-F bond dipoles that would form in ax-5FaXF-OAc. 186 Bond lengths and bond angles are characteristic in general. The length of axial C -F bonds is between 1.388 - 1.398 A, significantly longer than that (1.332 - 1.367 A) of equatorial C -F bonds. Also, the bonds between the anomeric center and the endocyclic oxygen ( C I - 0 5 and/or C5-05) are shorter when the C-F bond is axial than when it is equatorial. This trend is consistent with the double-bond-no-bond resonance hypothesis of the anomeric effect. Even the pseudo-axial C l - F l bond in eq-5F(3XF-OAc (1.395 A) is in the range of the length of the axial C-F bonds, indicating the contribution of electron derealization from the endocyclic oxygen. Particularly interesting is the C5-F5 bond of this compound eq-5F(3XF-OAc. This bond is in neither an axial nor an equatorial orientation since C5 is located within the plane 05-C5-C4 . Interestingly, the C5-F5 bond length is intermediate between those of an axial and an equatorial C -F bond. In addition, the C l - 0 5 bond connected to the axial C l - F l bond is shorter than the C5-05 bond, indicating the difference in bond order. These results suggest that the n-o* interaction is frilly formed between a lone pair of 0 5 and the antibonding orbital of C l - F l while it is only partially formed between the non-bonding orbital of 0 5 and a* of C5-F5 in eq-5 F p X F - 0 A c . Distorted axial C-F bonds in ax-5FocXF-OAc also show some interesting features. Bond lengths of C l - F l and C5-F5 are shortened compared to that of the corresponding "undistorted" axial C -F bond in a X y l F - O A c , e q - 5 F a X F - O A c and eq-5F(3XF-OAc. The apparent shortening of these anomeric axial bonds could arise from the imperfect overlap of the non-bonding orbital (no) of the endocyclic oxygen and antibonding orbitals o* of C-F bonds. This imperfect overlap might be due to the distortion of two bonds away from a fully axial position, which would have resulted in deviation from the anti-periplanar positioning of a lone pair. Further, the competition 187 between two possible interactions, no-o~*ci-Fi and no-CJ*c5-F5 may contribute to the shortening since the overlap from the non-bonding orbital of oxygen could not be fully manifested at C l - F l and C5-F5 bonds simultaneously. Consistent with this, both C l - 0 5 and C5-05 bonds are also somewhat longer than that found in a X y l F - O A c . The axial C 5 -F5 bond in the trifluoro-derivative 5 , 5 F 2 X F - O A c is longer than the two equatorial C-F bonds. However, the two equatorial C-F bonds (1.349 A for C l - F l and 1.332 A for C 5 -F5) are significantly shorter than other equatorial C-F bonds (1.366 and 1.367 A). In particular, the C5-F5 bond is extremely short, indicating that the derealization of electrons from the lone pairs of equatorial fluorine (F5) may occur and compete with that from the endocyclic oxygen to the antibonding orbital of the axial C5-F5 bond. Also it is hard to explain the bond lengths of C l - 0 5 and 05 -C5 . The C l - 0 5 bond bearing the equatorial C l - F l bond is shorter than the 05 -C5 bond and also significantly short compared to the corresponding bond of other molecules ((3XylF-OAc, eq -5FaXF-OAc) . A plausible explanation may require other examples and efforts in computation. There is a general trend that the bond angles 05 -C1-C2 , C1-05-C5 , and 0 5 - C 5 -C4 are larger when the C-F bond is in an axial orientation than when it is in an equatorial orientation. While this trend is clear in monofluorides, there are some deviations in difluorides. For example, the C5-C1-C2 angle (axial C l - F l ) in eq-5FocXF-OAc is only 109.0 degrees, not much different from that of the equatorial monofluoride fSXylF-OAc while the 05 -C5-C4 angle in eq-5F(3XF-OAc is 113.3 degrees, larger than that of the axial monofluoride a X y l F - O A c even though the C-F bond at C5 is not in the axial orientation. Appendix B. Kinetic Isotope Effects 189 Kinetic Isotope Effects The effect of isotopic substitution on a rate constant is referred to as a kinetic isotope effect. For example, in the reaction: A + B -> C The effect of isotopic substitution in reactant A is expressed as the ratio of rate constants k[/kh where the subscripts 1 and h represents reactions in which molecules A contain the light and heavy isotopes, respectively. The magnitude of this kinetic isotope effect (KIE) depends on the location of the isotope with respect to the center of reaction. According to this, commonly encountered types of effects are primary K I E and secondary K I E . B . l . Primary Kinetic Isotope Effect A primary kinetic isotope effect arises when the bond to the isotopic substituent is broken at or before the transition state. A s an example, for the reaction in which a C - H (or C-D) bond cleavage occurs and the hydrogen is transferred to another atom at the rate-limiting step, the vibrational energy of the C - H bond is expressed as follows: Evib = h v(n + —) 2 where n is each vibrational level, h is Planck's constant and 190 2jC\fJx-r where k is the vibrational force constant and p is the reduced mass of the X - Y bond (in this case C - H and C - D bond). The reduced mass p is as follows: m\mi ju= m\ + mi At the ground state, the zero-point energy would be 0.5 hv. The Morse potential energy diagram (Figure B . l ) is unaltered on going from H to D and consequently, k remains constant. However, the reduced mass p is different. As a good approximation, reduced mass increases by a factor of 2 on going from C - H to C-D . Therefore, the frequency v of C - D w i l l be lower than that of C - H , leading to the lower zero-point energy for C - D than C - H . Figure B.l . Schematic Morse potential energy diagram showing the zero-point vibrational energy of C-H and C-D bonds: E0(H), zero-point energy for C-H bond; E0(D), zero-point energy for C-D bond. Since the energy is the same at the dissociation limit, the bond energy is effectively higher for the C - D bond. Therefore, any reaction in which the C - H bond is broken during 191 the rate-limiting step wi l l be slower i f the hydrogen is replaced by a deuterium. The relative rate ka/ko is in the range of 2 - 7. Theoretically, the maximal effect is ~7 when the C - H bond is fully broken at the transition state and therefore the zero-point energy difference between C - H and C - D is zero. It should be noted that primary deuterium KIEs much larger than 7 have been measured, which is a result of quantum mechanical tunneling. In reality, the zero-point energy difference between C - H and C - D at the transition state is not truly zero since the H (or D) is still bonded to other atoms at the transition state, generating different zero-point vibrational energies at the transition state. However, this difference at the transition state is smaller than that at the reactant ground state and this phenomenon is the origin of the primary kinetic isotope effect. In the following reactions: X - H + Y [ X — H - - Y X + H Y (1) X - H + Y ^ [ X - H Y l * — X + H Y (2) X - H + Y ^ | X - - H - Y f ^ X + H Y (3) In the reaction (1), the H (D) atom is half transferred to Y at the transition state and w i l l be bonded equally to both atoms. This wi l l lower the force constant, and consequently, reduce the vibrational frequency. This w i l l cause the zero-point energy difference at the transition state to be rather small. A s a result, the K I E w i l l be high. On the other hand, in reactions (2) and (3), the transition state is reactant-like or product-like where H (D) is hardly transferred or almost transferred to the Y atom from the X atom at the transition state. In these cases, the zero-point energy difference at the transition state w i l l be close to that at the ground state of the reactant or product, leading to a small K I E (Figure B.2). 192 AEr Transition State IX—H-Y? < AE D X-H X-D Reactant Ground State IX—D-Yl* B Transition State [ X-H-—Yr* or I X--H-Y1* A E L KE D X-H X-D Reactant Ground State [X-D—YF or [ X--D-Y1* Figure B.2. Schematic energy diagram corresponding to the reaction X-H + Y: A. equation (1) showing large primary KIE; B. equations (2) and (3) showing small primary KIE. Thus, the measurement of a primary KIE w i l l give the following information. The existence of a primary KIE indicates that the bond to the substituted atom (hydrogen in this example) is cleaved in a rate-limiting step. Also , the magnitude of the primary KIE provides insight into the structure of the transition state for the cleavage of the bond. B.2. Secondary Kinetic Isotope Effects Secondary KIE arises when there is an isotopic substitution to an atom which is not directly involved in the bond being formed or broken. This effect is much smaller than the primary KIE and usually occurs in the range of 0.7 - 1.4. When the hydrogen in C - H is replaced by deuterium and the bond forming or breaking does not involve this hydrogen, the KIE is called a secondary deuterium kinetic isotope effect. When the C - H 193 bond is on a center undergoing reaction, an a-secondary deuterium K I E wi l l be measured and when the C - H bond is adjacent to the reaction center, a p-secondary deuterium K I E wi l l be measured. Essentially, a secondary a-deuterium K I E is explained in the same way as the primary effect, in terms of changes in vibrational energy and thus zero-point level. The secondary K I E is usually found in reactions in which the hybridization of the reaction center at the transition state is different from that at the ground state. For example, in glycoside cleavages, where the transition state assumes an oxocarbenium ion character, the initial sp 3 hybridization of the reaction center changes to a certain degree of sp 2 hybridization (Figure B.3). S P 3 sp 2 character Figure B.3. Hybridization change causing secondary kinetic isotope effect. 3 2 A decrease in the bending frequency of the C - H bond occurs when the change sp —> sp takes place, thus there is a decrease with the formation of such a transition state. Therefore, energy levels between C - H and C - D at the transition state are closer than those at the reactant state and the energy diagram resembles Figure B . 2 A . If the reaction proceeds with the change in the opposite direction, sp 2 —> sp 3, K I E value w i l l be less than 1. A secondary K I E larger than 1 is called a normal K I E and that less than 1 is an inverse K I E . Often secondary a-deuterium K I E serves as a probe of hybridization change during 15+ c x \ 194 the reaction. The theoretical maximal normal secondary a-deuterium K I E is 1.41 for the process where full hybridization change from sp 3 to sp 2 occurs at the transition state. In reality, normal secondary a-deuterium KIEs (ka/ko) between 1.1 and 1.3 are regarded as 3 2 evidences for substantial hybridization change from sp to sp . Appendix C. Graphical Representation of Data 196 Figure C. 1. Dixon plots showing the inhibition of a-glucosidase from Aspergillus niger by the reversible competitive inhibitors. The substrate used was 2,4-dinitrophenyl a-D-glucopyranoside. (37 °C, 0.1 M sodium acetate buffer, p H 4.5) 197 Figure C.2. Michaelis-Menten plots for the reaction a-l,4-glucan lyase with a series of aryl a-glucosides. (30 °C, 0.05 M M E S buffer, p H 6.0) [24DNPocGlc] mM 2,4-dinitrophenyl a-glucopyranoside 0 0.4 0.8 1.2 1.6 2 [34DNPaGlc] mM 3,4-dinitrophenyl a-glucopyranoside [4C2NPaGlc] mM 4-chloro-2-nitrophenyl a-glucopyranoside [25DNPaGlc] mM 2,5-dinitrophenyl a-glucopyranoside T—i—i—i—i—i—i—i—i—i—i—r [246TCPaGlc] mM 2,4,6-trichlorophenyl a-glucopyranoside [PNPaGlc] mM 4-nitrophenyl a-glucopyranoside 13 n> 3 CfQ o o 13 O o 13 3* a> 3 CfQ 5* o o 13 PJ 3 o co & ho 1 3 •8 cr CfQ ?T 0 o o 13 o o cr ST I-I o 13 cr 3 era o o 13 B o y I cr. o' -cr o 13 cr a> 3 CfQ ST o o 13 s o CD rate s _ 1 o o o o o k> 4^. b ) c o ° | ^ L . I I I I I I I I I oo 199 Figure C.3. Michaelis-Menten Plots for the reaction of oc-l,4-gtucan lyase with fluorosugars (30 °C, 0.05 M M E S buffer, p H 6.0). 1. 1-Fluoro-D-glucopyranosyl fluoride [1 FGIcF] mM 2. 2-Deoxy-2-fluoro-a-D-glucopyranosyl fluoride 0 20 40 60 80 [2FaGlcF] mM 3. 5-Fluoro-a-D-glucopyranosyl fluoride 0 20 40 60 80 [5FcxGlcF] mM 201 Figure C.4. Dixon plots showing the inhibition of a-l,4-glucan lyase by the reversible competitive inhibitors. The substrate used was 2,4-dinitrophenyl a-D-glucopyranoside. (30 °C, 0.05 M MES buffer, p H 6.0). 1. Inhibition by 1-Deoxynojirimycin 450 i [I] u M 2. Inhibition by Hydroximinogluconolactam 1000 l 3. Inhibition by Acarbose. 203 Figure C.5. Reactivation at 10 ° C of inactivated (at 10 °C) a-xylosidase from E. coli: the substrate used was 3 m M P N P a X y l (pH 7.0, 0.05 M phosphate buffer) 1. Reactivation of the enzyme inactivated with ax-5FaXF 4.3 1 ' r 0 200 400 time (min) 2. Reactivation of the enzyme inactivated with eq-5FaXF 4.4 0 200 400 time (min) 204 REFERENCES 1. Gruber, E . (1976) Papier 30, 133 - 138 2. Zubay, G . Biochemistry, 2nd ed., M c M i l l a n Publishing Co. , New York, 1988, p i31 3. Wolfenden, R., L u , X . and Young, G . (1998) J. Am. Chem. Soc. 120, 6 8 1 4 - 6 8 1 5 4. Wolfenden, R., Snider, M . , Ridgway, C . and Mil ler , B . (1999) J. Am. Chem. Soc. 121, 7 4 1 9 - 7 4 2 0 5. Sinnott, M . L . (1990) Chem. Rev. 90, 1171 - 1202 6. Ernst, S, Langer, R., Cooney, C. L . and Sasisekharan, R. (1995) Crit. Rev. Biochem. Mol. Biol. 30, 3 8 7 - 4 4 4 7. Linker, A . , Meyer, K . and Hoffman, P. (1956) J . Biol. Chem. 219, 13 - 25 8. McCarter, J. D . and Withers, S. G . (1994) Curr. Opin. Struct. Biol. 4, 885 - 892 9. Davies, G . J. and Henrissat, B . (1995) Structure 3, 853 - 859 10. Zechel, D . L . and Withers, S. G . (2000) Acc. Chem. Res. 33, 11 - 18 11. Koshland, D . E . (1953) Biol. Rev. 28, 4 1 6 - 4 3 6 12. Wang, Q., Graham, R .W. , Trimbur, D. , Warren, R. A . J. and Withers, S. G . (1994) J. Am. Chem. Soc. 116, 11594- 11595 13. Blake, C. C . F., Mair , G . A . , North, A . C . T., Phillips, D . C . and Sharma, V . R. Proc. R. Soc. London B 1967,167, 365 - 377 14. Kempton, J. and Withers, S. G . (1992) Biochemistry 31, 9961 - 9969 15. L i , Y . - K . , Chir, J. and Chen, F . - Y . (2001) Biochem. J. 355, 835 - 840 16. Vocadlo, D . J., Wick i , J., Rupitz, K . and Withers, S. G . (2002) Biochemistry 41, 9727 - 9 7 3 5 205 17. McCarter, J. D . , Yeung, W. , Chow, J., Dolphin, D . and Withers, S. G . (1997) J. Am. Chem. Soc. 119, 5792 - 5797 18. McCarter, J. D . and Withers, S. G . (1996) J. Am. Chem. Soc. 118, 241 - 242 19. Withers, S. G . and Aebersold, R. (1995) Protein Sci. 4, 361 - 372 20. Quaroni, A . and Semenza, G . (1976) J. Biol. Chem. 251, 3250 - 3253 21. Iwanami, S., Matsui, H . , Kimura, A . , Ito, H . , M o r i , H . , Honma, M . and Chiba, S. (1995) Biosci. Biotech. Biochem. 59, 459 - 463 22. W i c k i , J., Rose, D . R. and Withers, S. G . (2002) Method Enzymol. 354, 8 4 - 105 23. Mosi , R. M . and Withers, S. G. (2002) Method Enzymol. 354, 64 - 84 24. Davies, G . , Mackenzie, L . F., Varrot, A . , Dauter, M . , Brzozowski, A . M . , Schulein, M . and Withers, S. G . (1998) Biochemistry 37, 11707 - 11713 25. Burmeister, W . P., Cottaz, S., Driguez, H . , Iori, R., Palmieri, S. and Henrissat, B . (1997) Structure 5, 663 - 675 26. White, A . , Tul l , D . , Johns, K , Gilkes, N . R., Withers, S. G . and Rose, D . R. (1996) Nat. Struct. Biol. 3, 149 - 154 27. Vocadlo, D. J., Davies, G. J., Laine, R. and Withers, S. G . (2001) Nature 412, 835 -839 28. Heightman, T. D . and Vasella, A . T. (1999) Angew. Chem. Int. Ed. 38, 750 - 770 29. Sinnott, M . L . and Jencks, W . P. (1980) J. Am. Chem. Soc. 102, 2026 - 2032 30. Banait, N . S. and Jencks, W . P. (1991) J. Am. Chem. Soc. 113, 7958 - 7963 31. Legler, G . (1990) Adv. Carbohydr. Chem. Biochem. 48, 3 1 9 - 3 8 5 32. Cogoli, A . and Semenza, G . (1975) J. Biol. Chem. 1975, 250, 7802 - 7809 33. Tull , D . and Withers, S. G . (1994) Biochemistry 33, 6363 - 6370 206 34. van Doorslaer, E. , van Opstal, O., Kersters-Hilderson, H . and De Bruyne, C. K . (1984) Bioorg. Chem. 12, 1 5 8 - 1 6 9 35. Matsui, H . , Blanchard, J. S., Brewer, C. F. and Hehre, E . J. (1989) J. Biol. Chem. 264, 8 7 1 4 - 8 7 1 6 36. Tanaka, Y . , Tao, W. , Blanchard, J. S. and Hehre, E . J. (1994) J. Biol. Chem. 269, 32306-32312 37. Gacesa, P. (1987) FEBS Lett. 212, 199 - 202 38. Rye, C. S. and Withers, S. G . (2002) J. A m . Chem. Soc. 124, 9756 - 9767 39. Rye, C. S. and Withers, S. G . (2002) Anal. Biochem. 308, 77 - 82 40. Henrissat, B . (1991) Biochem. J. 280, 309 - 316 41. Henrissat, B . and Bairoch, A . (1993) Biochem. J. 293, 781 - 788 42. Henrissat, B . and Bairoch, A . (1996) Biochem. J. 316, 695 - 696 43. Henrissat, B . (1998) Biochem. Soc. Trans. 26, 153 - 156 44. Campbell, J. A . , Davies, G . J., Bulone, V . and Henrissat, B . (1997) Biochem. J. 326, 929 45. Coutinho, P. M . and Henrissat, B . in Recent Advances in Carbohydrate Bioengineering (Gilbert, H . J., Davies, G . J., Henrissat, B . and Svensson, B . , eds.) The Royal Society of Chemistry, Cambridge, p3. 46. Chothia, C. and Lesk, A . M . (1986) EMBO J. 5, 823 - 826 47. Jenkins, J., Lo Leggio, L . , Harris, G . and Pickersgill, R. (1995) FEBS Lett. 362, 281 -285 48. Bourne, Y . and Henrissat, B . (2001) Curr. Opin. Struct. Biol. 11, 593 - 600 49. Frandsen, T. P. and Svensson, B . (1998) Plant Mol. Biol. 37, 1 - 13 207 50. Chiba, S., Hiromi, K . , Minamura, N . , Ohnishi, M . , Shimomura, T., Suga, K . , Suganiima, T., Tanaka, A . , Tomioka, S. and Yamamoto, T. (1979) J. Biochem. 85, 1135 - 1 1 4 1 51. McCarter, J. D . and Withers, S. G . (1996) J. Biol. Chem. 271, 6889 - 6894 52. Y u , S., Kenne, L . and Pedersen, M . (1993) Biochim. Biophys. Acta 1156, 313 - 320 53. Frandsen, T. P., Stoffer, B . B . , Palcic, M . M . , Hof, S. and Svensson, B . (1996) J. Mol. Biol. 263, 79 - 89 54. Lemieux, R. U . , Spohr, U . , Bach, M . , Cameron, D . R., Frandsen, T. P., Stoffer, B . B . , Svensson, B . and Palcic, M . M . (1996) Can J. Chem. 74, 319 - 335 55. Sierks, M . R. and Svensson, B . (1992) Protein Eng. 5, 185 - 188 56.Sierks, M . R., Bock, K . , Refin, S. and Svensson, B . (1992) Biochemistry 31, 8972 -8977 57. Ki ta , A . , Matsui, PL, Somoto, A . , Kimura, A . , Takata, M . and Chiba, S. (1991) Agric. Biol. Chem. 55, 2327 -2335 58. McCleary, B . , Gibson, T. S., Sheehan, H . , Casey, A . , Horgan, L . and O'Flaherty, J. (1989) Carbohydr. Res. 185, 147 - 162 59. Quaroni, A . and Semenza, G . (1976) J. Biol. Chem. 251, 3250 - 3253 60. Hermans, M . M . P., Kroos, M . A . , van Beeumen, J., Oostra, B . A . and Reuser, A . J. J. (1991) J. Biol. Chem. 266, 13507 - 13512 61. Iwanami, S., Matsui, H . , Kimura, A . , Ito, H . , M o r i , H . , Honma, M . and Chiba, S. (1995) Biosci. Biotech. Biochem. 59, 459 - 463 62. Kimura, A . , Takata, M . , Fukushi, Y . , Mor i , H . , Matsui, H . and Chiba, S. (1997) Biosci. Biotech. Biochem. 61, 1091 - 1098 208 63. Okuyama, M . , Okuno, A . , Shimizu, N . , Mor i , H . , Kimura, A . and Chiba, S. (2001) Eur. J. Biochem. 268, 2270 - 2280 64. Chambers, J. P., Elbein, A . D . and Will iams, J. C. (1982) Biochem. Biophys. Res. Commun. 107, 1 4 9 0 - 1496 65. Hanozet, G . , Percher, H . - P . , Vanni , P., Oesch, B . and Semenza, G . (1981) J. Biol. Chem. 256, 3703 - 3 7 1 1 66. Igaki, S. and Chiba, S. (1989) Abstracts of the Annual Meeting of the Japan Agricultural Society, Niigata, Japan, Agricultural Chemical Society of Japan, Tokyo, 1989; 3 H p l 6 , p 292 67. Chaillou, S., Lokman, B . C , Leer, R. J., Posthuma, C , Postma, P. W . and Pouwels, P. H . (1998) Bacteriol. 180, 2312 - 2320 68. Moracci, M . , Ponzano, B . C , Trincone, A . , Fusco, S., De Rosa, M . , van der Oost, J., Sensen, C. W. , Charlebois, R. L . and Rossi, M . (2000) J. Biol. Chem. 275, 22082 - 22089 69. Acebes, J. L . , Moral , R. and Zarra, I. (1993) Phytochemistry 33, 1343 - 1345 70. Sampedro, J., Sieiro, C , Revilla, G . , Gonzallez-Villa, T. and Zarra, I. (2001) Plant Physiol. 126 ,910-920 71. Crombie, H . J., Chenggappa, S., Jarman, C , Sidebottom, C. and Reid, J. S. G . (2002) PlantalU, 4 0 6 - 4 1 3 72. Sanchez, M . , Gianzo, C , Sampedro, J., Revilla, G . and Zarra, I. (2003) Plant Cell Physiol. 44, 1 3 2 - 138 73. McDougall , G . and Fry, S. C. (1990) Plant Physiol. 93, 1042 - 1048 74. Lorences, E . P., McDougal, G . and Fry, S. C. (1990) Physiol. Plant, 79, 563 - 569 209 75. Y u . S., Bojsen, K . , Svensson, B . and Marcussen, J. (1999) Biochim. Biophys. Acta 1433, 1 - 15 76. Bojsen, K . , Y u , S., Kragh, K . M . and Marcussen, J. (1999) Biochim. Biophys. Acta 1430, 396 - 402 77. Bojsen, K . , Y u , S. and Marcussen, J. (1999) Plant Mol. Biol. 40, 445 - 454 78. Shiga, Y . , Kametani, S., Kadokura, T. and Akanuma, H . (1999) J. Biochem. 125, 166 - 172 79. Kametani, S., Shiga, Y . and Akanuma, H . (1996) Eur. J. Biochem. 242, 832 - 838 80. Suzuki, M . , Kametani, S., Uchida, K . and Akanuma, H . (1996) Eur. J. Biochem. 240, 2 3 - 2 9 81. Broberg, A . , Kenne, L . and Pedersen, M . (1995) Phytochemistry 41, 1 5 1 - 1 5 4 82. Y u , S., Ahmad, T., Kenne, L . and Pedersen, M . (1995) Biochim. Biophys. Acta 1244, 1 - 9 83. Yoshinaga, K . , Fujisue, M . , Abe, J., Hanashiro, I., Takeda, Y . , Muroya, K . and Hizukuri , S. (1999) Biochim. Biophys. Acta 1472, 447 - 454 84. Nyval l , P., Pedersen, M . , Kenne, L . and Gacesa, P. (2000) Phytochemistry 54, 139 -145 85. Keitel, T., Simon, O., Borriss, R. and Heinemann, U . (1993) Proc. Natl. Acad. Sci. U.S.A. 90, 5287 - 5291 86. Bordier Hoj, P., Condor, R., Traeger, J., McAulif fe , J. and Stone, B . (1992) J. Biol. Chem. 267,25059 - 25066 87. Legler, G . and Bause, E . (1973) Carbohydr. Res. 28, 45 - 52 210 88. Tul l , D . , Burgoyne, D . L . , Chow, D . T., Withers, S. G . and Aebersold, R. (1996) Anal. Biochem. 234, 1 1 9 - 125 89. Keresztessy, Z. , Kiss, L . and Hughes, M . A . (1994) Arch. Biochem. Biophys. 315, 323 - 3 3 0 90. Yariv, J., Wilson, K . , Hildesheim, J. and Blumberg, S. (1971) FEBS Lett. 15, 24 - 26 91. Black, T. S., Kiss , L . , Tul l , D . and Withers, S. G . (1993) Carhohydr. Res. 250, 195 -202 92. Howard, S. and Withers, S. G . (1998) Biochemistry 37, 3858 - 3864 93. Bombard, S., Maillet, M . and Campau, M . L . (1995) Carbohydr. Res. 275, 4 3 3 - 4 4 0 94. Sinnott, M . L . and Smith, P. J. (1978) Biochem. J. 173, 1 - 1 4 95. Legler, G . (1968) Z. Physiol. Chem. Hoppe-Seyler 349, 767 - 774 96. Withers, S. G. , Rupitz, K . , and Street, I. P. (1988) J. Biol. Chem. 263, 7929 - 7932 97. Street, I. P., Kempton, J. B . and Withers, S. G . (1992) Biochemistry 31, 9970 - 9978 98. Withers, S. G . and Street, I. P. (1988) J . Am. Chem. Soc. 110, 8551 - 8553 99. Braun, C , Grayer, G . and Withers, S. G. (1995) J. Biol. Chem. 270, 26778-26781 100. Kajimoto, T., L iu , K . K . - C , Pederson, R. L . , Zhong, Z . , Ichikawa, Y . , Porco, J. A . J. and Wong, C . - H . (1991) J . Am. Chem. Soc. 113, 6187 - 6196 101. Winkler, D . A . and Holan, G . (1989) J. Med. Chem. 32, 2084 - 2089 102. Namchuk, M . N . , McCarter, J. D . , Becalski, A . , Andrews, T. and Withers, S. G . (2000) J. Am. Chem. Soc. 122, 1270-1277 103. Takayanagi, T., Kimura, A . , Chiba, S. and Ajisaka, K (1994) Carbohydr. Res., 256, 149-158 211 104. Kimura, A . , Takata, M . , Sakai, O. , Matsui, H . , Takai, N . , Takayanagi, T., Nishimura, I., Uozumi, T. and Chiba, S. (1992) Biosci. Biotech. Biochem., 56, 1368 -1370 105. Kimura, A . , Somoto, A . , Mor i , H . , Sakai, O., Matsui, H . and Chiba, S. (1997) Biosci. Biotech. Biochem. 61, 475 - 479 106. Gebler, J. C , Aebersold R. and Withers, S. G . (1992) J. Biol. Chem. 267, 11126 -11130 107. Miao S., McCarter J. D . , Grace, M . E . , Grabowski, G . A . , Aebersold R. and Withers, S. G. (1994) J. Biol. Chem. 269, 10975 - 10978 108. Hoos, R., Vasella, A . , Rupitz, K . and Withers, S. G . (1997) Carbohydr. Res. 298, 291 - 2 9 8 109. Will iams, S. J., Hoos, R. and Withers, S. G . (2000) J. Am. Chem. Soc. 122, 2223 -2235 110. L y , H . D . , Howard, S., Shum, K . , He, S., Zhu, A . and Withers, S. G . (2000) Carbohydr. Res. 329, 539 - 547 111. Howard, S., He, S. and Withers, S. G . (1998) J. Biol. Chem. 273, 2067 - 2072 ' 112. Numao, S., He S., Evjen G. , Howard S., Tollersrud, O. K . and Withers, S. G . (2000) FEBSLett. 484, 1 7 5 - 178 113. Wentworth, D . F. and Wolfenden, R. (1974) Biochemistry 13, 4715 - 4720 114. McCarter, J. D . (1995) Ph. D. Thesis, University of British Columbia 115. Stoll, D . , He, H . , Withers, S. G . and Warren, R. A . J. (2000) Biochem. J. 351, 833 -838 116. Babbit, P. C and Gerlt, J. A . (1997) J. Biol. Chem. 272, 30591 - 30594 212 117. Babbit, P. C , Renyon, G . L . , Martin, B . M . , Charest, H . , Slyvestre, M . , Scholten, J. D. , Chang, K . - H . , Liang, P. - H . and Dunaway-Mariano, D . (1992) Biochemistry 31, 5 5 9 4 - 5 6 0 4 118. Smar, M . , Short, S. A . and Wolfenden, R. (1991) Biochemistry 30, 7908 - 7912 119. Morgan, P. M . , Sala, R.F. and Tanner, M . E . (1997) J. Am. Chem. Soc. 119, 10269 -10277 120. Y u , S., Olsen, C. E . and Marcussen, J. (1998) Carbohydr. Res. 305, 73 - 82 121. Apparu, M . , Blanc-Muesser, M . , Defaye, J. and Driguez, H . (1981) Can. J. Chem. 59 ,314-320 122. Yamaguchi, M . , Horiguchi, A . , Fukuda, A . and Minami , T. (1990) J. Chem. Soc. Perkin Trans. 7, 1079 - 1082 123. Kooners, H . J., de Kok, A . J., Romers, C. and van Boom, J. H . (1980) Reel. Trav. Chim. Pays-Bas 99, 355 - 362 124. Ballardie, F., Capon, B . , Sutherland, J. D . G. , Cocker, D . and Sinnott, M . L . (1973) Chem. Soc. Perkin Trans. 1, 2418 - 2419 125. Konstantinidis, A . and Sinnott, M . L . (1991) Biochem. J. 279, 587 - 593 126. Braun, C. (1995) Ph.D. Thesis, University of British Columbia 127. Withers, S. G. , MacLennan, D . J. and Street, I. P. (1986) Carbohydr. Res. 154, 127 -144 128. Blint, R. J., McMahon, T. B . and Beauchamp, J. L . (1974) J. Am. Chem. Soc. 96, 1 2 6 9 - 1278 129. Mos i , R., Sham, H . , Uitehaag, J. C. M . , Ruiterkamp, R., Dijkstra, B . W . and Withers, S. G. (1998) Biochemistry 37, 1 7 1 9 2 - 17198 213 130. Cleland, W . W. , O'Leary, M . H . and Northrop, D . B . , Eds (1977) Isotope Effects on Enzyme-Catalyzed Reactions, University Park Press, Baltimore 131. Northrope, D . B . (1975) Biochemistry 14, 2644 - 2651 132. Fry, A . (1972) Chem. Soc. Rev. 1, 163 - 210 133. Cram, D . J. and Tadanier, J. (1959) J. Am. Chem. Soc. 81, 2737 - 2748 134. More O'Ferrall, R. A . (1970) J. Chem. Soc. B, 274 - 277 135. Bunnet, J. F., Davis, G . T. and Tanida, H . (1962) J. Am. Chem. Soc. 84, 1606 - 1614 136. Heine, A . , DeSantis, G. , Luz, J. G . , Mitchell, M . , Wong, C . - H . and Wilson, I. A . (2001) Science 294, 369 - 374 137. Uitdehaag, J. C. M . , Mos i , R., Kalk, K . H . , van der Veen, B . A . , Dijkhuizen, L . , Withers, S. G . , Dijkstra, B . W . (1999) Nat. Struct. Biol. 6, 432 - 436 138. Numao, S., Kuntz, D . A . , Withers, S. G . and Rose, D . R. (2003) J. Biol. Chem. 278, 48074 - 48083 139. Burland V . D . , Plunkett G . I l l , Daniels D . L . and Blattner F. R. (1993) Genomics 16, 551 - 5 6 1 140. Plunkett G . I l l , Burland V . D . , Daniels D . L . and Blattner F. R. (1993) Nucleic Acids Res. 21,3391 - 3 3 9 8 141. Monroe, J. D. , Garcia-Cazarin, M . L . , Poliquin, K . A . and Aivano, S. K . (2003) Plant Physiol. Biochem. 41, 877 - 885 142. Zong, N . , Kamiyama, Y . and Yasui , T. (1989) Agric. Biol. Chem. 53, 2129 - 2139 143. Zong, N . and Yasui , T. (1989) Agric. Biol. Chem. 53, 1 8 7 - 195 144. Roberts, K . (2001) Plant Physiol. 125, 127 - 130 214 145. Vocadlo, D . J., W i c k i , J., Rupitz, K . and Withers, S. G . (2002) Biochemistry 41, 9736 - 9746 146. Connelly, G . P., Withers, S. G . and Mcintosh, L . P. (2000) Protein Sci. 9, 512 - 524 147. Joshi, M . D . , Sidhu, G . , Nielsen, J. E. , Brayer, G . D . , Withers, S. G . and Mcintosh, L . P. (2001) Biochemistry 40, 10115-10139 148. Poon, D . K . , Webster, P., Withers, S. G . and Mcintosh, L . P. (2003) Carbohydr. Res. 338 ,415-421 149. Hayashi, M . , Hashimoto, S. and Noyori , R. (1984) Chem. Lett. MAI - 1750 150. Praly, J.-P. and Descotes, G . (1987) Tetrahedron Lett. 28, 1405 - 1408 151. Igarashi, K . , Honma, T. and Irisawa, J. (1969) Carbohydr. Res. 11, 577 - 578 152. Kothe, G . , Luger, P. and Paulsen, H . (1979) Acta Crystallogr. 35, 2079 - 2087 153. Fry, S. C. (1989) Physiol. Plant 75, 532 - 536 154. Cosgrove, D . (2001) Plant Physiol. 125, 131 - 134 155. Wolfrom, M . L . and Thompson, A . (1963) Methods Carbohydr. Chem. 2, 211 - 215 156. Lemieux, R. U . (1963) Methods Carbohydr. Chem. 2, 224 - 225 157. Bradford, M . (1976) Anal. Biochem. 72, 248 - 254 158. Laemmli, U . K . (1970) Nature 227, 680 - 685 159. Edward, J. T. (1955) Chem. Ind. (London) 1102 - 1104 160. Lemieux, R. U . (1971) Pure Appl. Chem. 25, 527 - 547 161. El ie l , E . L . (1972) Angew. Chem., Int. Ed. Engl. 11, 739 - 750 162. Juriasti, E . and Cuevas, G . (1992) Tetrahedron 48, 5019 - 5087 215 163. Szarek, W . A . and Horton, D . Eds. (1979) Anomeric Effect: Origin and Consequences; A C S Symposium Series No . 87, American Chemical Society, Washington D C 164. Thatcher, G . R. J. Ed . (1993) The Anomeric Effect and Associated Stereoelectronic Effects, A C S Symposium Series N o . 539, American Chemical Society, Washington D C 165. Lemieux, R. U . , Pavia, A . A . , Martin, J. C. and Watanabe, K . A . (1969) Can. J. Chem. 47, 4427 - 4439 166. Holland, C. V . , Horton, D . and Jewell, J. S. (1967) J. Org. Chem. 32, 1 8 1 8 - 1821 167. Hall , L . D . and Manville, J. F. (1969) Can. J. Chem. 47, 19 - 30 168. Kothe, G. , Luger, P. and Paulsen, H . (1979) Acta Crystallogr. 35, 2079 - 2087 169. Luger, P., Buschmann, J., Schmidt, H.-J . and Paulsen, H . (1982) Acta Crystallogr. 38 ,2732-2735 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061201/manifest

Comment

Related Items