SYNTHESIS AND TESTING OF NITROGEN-CONTAINING SUGAR ANALOGUES AS INHIBITORS OF THERMOANAEROBACTERIUM SACCHAROLYTICUM (3-XYLOSIDASE By TIMOTHY J. HffiBERT B.Sc, The University ofWinnipeg, 1995 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required j^ tandard THE UNIVERSITY OF BRITISH COLUMBIA May 1999 © Timothy J. Hiebert, 1999 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e head of my department o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . The U n i v e r s i t y o f B r i t i s h C o l umbia Vancouver, Canada 11 Abstract l-Deoxyxylonojirimycin, several -^substituted 1-deoxyxylonojirimycin derivatives, D-xylonolactam, and D-xylonojirimycin tetrazole were synthesised and tested as inhibitors of Thermoanaerobacteriwn saccharolyticum (3-xylosidase to investigate the influence of substitution at the endocyclic nitrogen, the importance of basicity at the ring nitrogen to inhibitory potency, and the influence of sp2 hybridisation at the anomeric centre on inhibitor binding in this model system. l-Deoxyxylonojirimycin is a competitive inhibitor of T. saccharolyticum (3-xylosidase with Ki values ranging from 1.5 mM at pH 4.5 to 13.5 pM at pH 6.5. This pH dependence is consistent with the inhibitor binding in its cationic form to the enzyme acid/base catalyst in its deprotonated form. N-Acetyl and JV-formyl- 1-deoxyxylonojirimycin bind poorly to T. saccharolyticum (3-xylosidase, as would be expected if the endocyclic nitrogen must be protonated for good inhibition of this enzyme by 1-deoxynojirimycin-type inhibitors. However, N-benzyloxycarbonyl- 1-deoxyxylonojirimycin also binds relatively tightly to the enzyme (K, = 110 pM, pH 5.5) while Af-benzyl- 1-deoxyxylonojirimycin is a poor inhibitor (Kj = 7.2 mM, pH 5.5). The inhibitory potency of Af-benzyloxycarbonyl-l-deoxyxylonojirimycin is therefore surprising, but may be the result of fortuitous noncovalent interactions between the pendant aromatic group and a hydrophobic site on the enzyme. Apparently these interactions are not accessible to the N-benzyl group. D-Xylonolactam is a weak inhibitor of T. saccharolyticum (3-xylosidase (K; = 3 mM, pH 5.5), while D-xylonojirimycin tetrazole binds relatively tightly (Kj = 100 pM, pH 5.5) consistent with the notion that proton transfer by T. saccharolyticum (3-xylosidase may take place in the plane of I l l the pyranose ring. Surprisingly, several of these inhibitors bind well to the Glul60Ala mutant of T. saccharolyticum p-xylosidase in which the putative acid/base catalyst has been replaced, suggesting that a third carboxyl group in the active site may assume a role in binding and possibly also catalysis in the mutant enzyme. N-Benzyloxycarbonyl- 1-deoxyxylonojirimycin was used as a glycosyl acceptor in transglycosylation reactions catalysed by the Glu358Ala mutant of Agrobacterium sp. [3-glucosidase. Both (3-1,3-linked and (3-1,4-linked products were formed in a total yield of 50.4%. While the enzyme initially formed p-l,4-linkages to the inhibitor only slowly, the 'disaccharide' product was a superior glycosyl acceptor and was quickly converted to tri- and tetrasaccharide products. Hexa-O-acetyl xylobiose was isolated on a gram scale by treatment of birchwood xylan with T. viride xylanase, followed by acetylation and column chromatography. This intermediate provided a convenient route to the xylanase substrates 2,5-dinitrophenyl (3-D-xylobioside and 2-nitrophenyl P-D-xylobioside in 41 and 48% overall yields respectively. iv Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures viii Abbreviations And Symbols xi Acknowledgements xiii Dedication xiv Chapter One: Introduction 1 1.1 General 2 1.2 Glycosidase Mechanisms 3 1.2.1 Inverting Glycosidases 3 1.2.2 Retaining Glycosidases 4 1.3 Evidence for the Double Displacement Mechanism 6 1.3.1 The Nucleophile and Acid/Base Catalysts 6 1.3.2 The Nature of the Glycosyl-Enzyme Intermediate 7 1.3.3 Oxocarbenium ion-like Transition states 9 1.3.4 Glycosidase Inhibitors 10 1.4 Nitrogen-containing Glycosidase Inhibitors 12 1.4.1 Aldonolactams 12 1.4.2 Glycosyl Amines 12 1.4.3 Nojirimycin and 1-Deoxynojirimycin 13 1.4.4 Derivatives of 1-Deoxynojirimycin 16 1.4.5 TAza Sugars 16 1.4.6 New Glycosidase Inhibitors with sp2-Hybridised 'Anomeric Centres' 17 1.5 Glycosidases in Enzymatic Synthesis 20 1.6 (5-Xylosidases 22 1.7 Aims of Thesis 23 1.7.1 Synthesis of and Testing of Potential Xylosidase Inhibitors 23 1.7.2 Preparation of Oligosaccharide Iminosugar Analogues by Enzymatic Transglycosylation 25 1.7.3 Preparation of Chromophoric Disaccharide Substrates for Xylanases 26 Chapter Two: Results and Discussion 27 2.1 Synthesis 28 2.1.1 Synthesis of Nitrogen-containing D-Xylopyranose Analogues 28 2.1.1.1 l-Deoxyxylonojirimycin 28 2.1.1.2 TV-Substituted Derivatives of l-Deoxyxylonojirimycin 30 2.1.1.3 Attempted Synthesis of D-Xylonolactam oxime (2.2) 31 2.1.1.4 Synthesis of 2,3,4-Tri-O-benzyl-D-xylonolactam (2.26) 34 2.1.1.5 Synthesis of 1,1-Dideutero-A -^benzyloxycarbonyl-1-deoxyxylonojirimycin 42 2.1.1.6 Preparation of D-Xylonojirmycin tetrazole (2.1) 44 2.7.2 Xylo-configured Imino Sugar Derivatives as Glycosyl Acceptors in Enzymatic Transglycosylation Reactions 45 2.1.2.1 Transglycosylation Experiments with AbgGlu358Ala 47 2.1.2.2 Transglycosylation Experiments with T. saccharolyticum Glu277Ala (3-xylosidase 51 2.1.3 Preparation ofAryl Xylobiosides 57 2.1.4 Conclusions 54 2.2 Enzyme Kinetics 55 2.2.7 The Significance of pH Dependence in Enzyme kinetics 55 2.2.2 pH Dependence of Wild-Type and Glul60Ala T. saccharolyticum /3-xylosidase Activity 57 2.2.3 Inhibition of Wild-Type T. saccharolyticum fi-xylosidase by l-Deoxyxylonojirimycin 58 2.2.3.1 pH Dependence of T. saccharolyticum (3-xylosidase Inhibition by l-Deoxyxylonojirimycin 59 2.2.4 The Influence of Inhibitor Structure on Inhibition of T. saccharolyticum (3-Xylosidase 62 2.2A. 1 Inhibition of T. saccharolyticum (3-Xylosidase by compounds with Substituents on the Ring Nitrogen 62 2.2.4.2 Inhibition of T. saccharolyticum (3-xylosidase by compounds with a sp2 hybridised anomeric centre 66 2.2.4.3 The Influence of the Position of the Endocyclic Nitrogen on Imino Sugar Inhibition of T. saccharolyticum [3-xylosidase 67 2.2.5 Inhibition Studies of Glul60Ala T. saccharolyticum fi-Xylosidase 67 2.2.6 Conclusions 68 Chapter Three: Materials and Methods 70 3.1 Synthesis .71 3.7.7 General 77 3.1.2 Transglycosylation Experiments 91 3.1.2.1 Transglycosylation Experiments with ZV-Benzyloxycarbony 1-1-deoxyxylonojirimycin (2.8) 91 3.1.2.2 Transglycosylation Experiments with l,l-Dideutero-7V-benzyloxocarbonyl-1-deoxyxylonojirimycin (2.9) 93 3.2 Enzyme Kinetics 95 3.2.1 General 95 3.2.2 Determination of Kinetic Parameters 95 3.2.3 Inhibition Studies on T. saccharolyticum (3-Xylosidase 96 References . Appendix A: Transition State Analogy A.l Transition state Theory A.2 Modified Substrates and Inhibitors to Probe Transition state Mimicry A.3 Probing Transition state Mimicry Using Enzyme Mutants Appendix B: Graphical Representation of Kinetic Data List of Tables Table 2.1 Summary of transglycosylation products from TV-Benzyloxycarbonyl-1-deoxyxylonojirimycin and AbgGlu358Ala 48 Table 2.2 'H-NMR data from p-l,3-linked disaccharides 49 Table 2.3 pH Dependence of T. Saccharolyticum P-Xylosidase Inhibition by l-Deoxyxylonojirimycin 61 Table 2.4 Inhibition of wild-type and Glul60Ala T. saccharolyticum p-xylosidases by various imino sugars 65 List of Figures Figure 1.1 The general reaction catalysed by glycosidases 3 Figure 1.2 Proposed mechanism of an inverting P-glucosidase 4 Figure 1.3 The double displacement mechanism, shown for a P-xylosidase 5 Figure 1.4 Affinity labels for retaining glycosidases; conduritol C cis-epoxide (1.1), 2,4-dinitrophenyl-2-deoxy-2-fluoro-P-D-glucopyranoside (1.2) 8 Figure 1.5 The glucosyl cation 9 Figure 1.6 Proposed transition state contributors in the retaining mechanism of a p-glucosidase 9 Figure 1.7 Inactivators of retaining glycosidases 10 Figure 1.8 Contributing resonance structures of D-gluconolactone 11 Figure 1.9 D-Gluconolactam 12 Figure 1.10 p-Glucosylamine 13 Figure 1.11 Nojirimycin (1.13) and 1-deoxynojirimycin (1.14) 13 Figure 1.12 Spontaneous dehydration of nojirimycin 14 Figure 1.13 A^V-Dimethyl-1 -deoxynojirimycin 15 Figure 1.14 Af-Alkyl derivatives of 1-deoxynojirimycin with biological activity: miglitol (1.16), and Af-butyl- 1-deoxynojirimycin (1.17) 16 Figure 1.15 Isofagomine (1.18) and fagomine (1.19) 17 Figure 1.16 Glycosidases inhibitors with sp2 hybridisation at the anomeric carbon 18 Figure 1.17 D-Gluconojirimycin tetrazole (1.31), imidazole (1.32), and triazole (1.33).. 19 Figure 1.18 A) Proposed direction of hydrogen bonding with lactam derivatives. B) 'Side on' protonation of substrate 19 Figure 1.19 Interception of glycosyl-enzyme intermediate resulting in transglycosylation by Abg P-glucosidase 21 Figure 1.20 Proposed mechanism of transglycosylation by AbgGlu358Ala 22 Figure 1.21 D-Xylonolactam (1.34) and 1-deoxyxylonojrimycin (1.35) 25 Figure 1.22 1-Deoxyxylobionojirimycn (1.36) 26 Figure 2.1 D-Xylo-configured sugar analogues containing an endocyclic nitrogen 28 Figure 2.2 Synthetic route towards 1-deoxyxylonojirimycin 29 Figure 2.3 Preparation of 15N-D-xylonolactam 30 Figure 2.4 Synthesis of Af-benzyloxycarbonyl- 1-deoxyxylonojirimycin (2.8), and Af-benzyl- 1-deoxyxylonojirimycin (2.3) 30 Figure 2.5 Synthesis of N-acyl derivatives of 1-deoxyxylonojirimycin 31 Figure 2.6 Synthesis of D-gluconolactam oxime from 2,3,4,6-tetra-t9-benzyl-D-gluconolactam 32 ix Figure 2.7 A) Attempted synthesis of D-xylonolactam oxime from D-xylonolactam. B) Proposed side reaction upon treatment of 2.8 with hydroxylamine 33 Figure 2.8 Synthesis of xylobionolactam oxime 33 Figure 2.9 Synthesis of D-gluconolactam oxime by Ganem et al 34 Figure 2.10 Compounds accessible from tri-O-benzyl-D-xylonolactam (2.26) 35 Figure 2.11 Conversion of 5-bromo-D-arabinolactone to D-arabinolactam 36 Figure 2.12 Retrosynthetic strategy towards tri-O-benzyl-D-xylonolactam 37 Figure 2.13 Unsuccessful route to tri-O-benzyl-D-xylonolactam 38 Figure 2.14 Retrosynthetic strategy to tri-O-benzyl-D-xylonolactam from methyl (3-D-xylopyranoside 38 Figure 2.15 A) Preparation of 2,3,4-tri-O-benzyl-D-xylonolactone. B) Proposed side reaction upon treatment of 2,3,4-tri-O-benzyl-D-xylonolactone (2.44) with base 39 Figure 2.16 Synthesis of 2,3,4-tri-0-benzyl-5-0-methanesulfonyl-D-xylonamide 40 Figure 2.17 Lactone formation upon treatment of 2,3,4-tri-0-benzyl-5-methanesulfonyl-D-xylononamide 41 Figure 2.18 Successful synthesis of tri-O-benzyl-D-xylonolactam 42 Figure 2.19 Introduction of deuterium at C-l affording benzyl carbamates 43 Figure 2.20 A) Preparation of 2,3,4,6-tetra-O-benzyl-D-mannonojirimycin tetrazole via intramolecular 1,3-dipoar cycloaddition. B) Preparation of D-xylonojirimycin tetrazole from 2,3,4-tri-O-benzyl-D-xylonolactam 44 Figure 2.21 A -^Benzyloxycarbonyl-l-deoxyxylonojirimycin (2.8), potential transglycosylation products 2.59 and 2.60 and l,l-dideutero-/V-benzyloxycarbonyl-l-deoxyxylonojirimycin (2.9) 46 Figure 2.22 Enzymatic transglycosylation mediated by AbgGlu358Ala (3-glucosidase using a-glucosyl fluoride as a glycosyl donor and 7V-benzyloxycarbonyl-1-deoxyxylonojirimycin (2.8) as a glycosyl acceptor 47 Figure 2.23 Transglycosylation products obtained by reaction of 2.8 and a-glucosyl fluoride mediated by AbgGlu358Ala 48 Figure 2.24 Chemo-enzymatic synthesis of aryl xylobiosides 53 Figure 2.25 k c a t/Km as a function of pH for wild-type T. Saccharolyticum |3-Xylosidase and Glul60Ala T. Saccharolyticum 58 Figure 2.26 Kj determination of T. saccharolyticum (3-Xylosidase inhibition by 1-Deoxyxylonojirimycin 59 Figure 2.27 pKa determination of 15N-l-deoxyxylonojirimycin by 15N-NMR 60 Figure 2.28 pK, as a function of pH for inhibition of T. saccharolyticum (3-xylosidase by 1-deoxyxylonojirimycin 62 Figure 2.29 TV-substituted imino sugars tested as inhibitors of T. saccharolyticum p-xylosidase 63 Figure 2.30 D-Xylonojirimycin triazole (2.79) 66 X Figure 2.31 D-Xyloisofagomine 67 Figure A.l The thermodynamic cycle 105 Figure A.2 a) log Kj versus log (Kra/kcat) for a series of substrates and phosphonic acid inhibitors of pepsin, b) log Kj versus log K m for the same series 107 Figure A.3 Linear free energy relationships for binding of a series of correspondingly substituted DNP glycosides and inhibitors with Abg p-glucosidase 108 Figure A.4 Acarbose 109 Figure A.5 Linear free energy relationship between kinetic parameters for the inhibitor acarbose and a-glucosyl fluoride as substrate with a series of mutants of CGTase 110 Figure B. 1 A) Determination of K m and V m a x for T. saccharolyticum (3-xylosidase using PX as substrate at pH 5.5 B-D) Kj plots for T. saccharolyticum (3-xylosidase inhibition by 1-deoxyxylonojirimycin at various pH values. .112 Figure B. 2 K; plots for the inhibition of T. saccharolyticum P-xylosidase by various inhibitors 113 Abbreviations And Symbols 2,4-DNP 2,4-dinitrophenol 2,4-DNPX 2,4-dinitrophenyl (3-D-xylopyranoside 2,5-DNP 2,4-dinitrophenyl 2,5-DNPX2 2,5-dinitrophenyl [3-xylobioside 3,4-DNPX 3,4-Dinitropheny 1 [3-D-xylopyranoside Abg Agrobacterium sp. (3-glucosidase Ac Acetyl aDKIE a-Deuterium kinetic isotope effect aGF a-Glucosyl fluoride BCX Bacillus circulans xylanase Bn Benzyl BSA Bovine Serum Albumin Cbz Benzyloxy carbonyl Cex Cellulomonas fimi exoglycanase CGTase Bacillus circulans cyclodextrin glucanotransferase DCI-MS Desorption chemical ionisation mass spectrometry DMF N-Dimethylformamide DMPU 1,3-Dimethyl-3,4,5,6-tetrahydro-2( l/7)-pyrimidinone DMSO A^N-Dimethyl sulfoxide E. coli Escherichia coli ES-MS Electropray ionisation mass spectrometry HMDS 1,1,1,3,3,3-Hexamethyldisilazane LSIMS-MS Secondary ion mass spectrometry MeOH Methanol Ms Methanesulfonyl MS Mass spectrometry NMR Nuclear magnetic resonance ONP 2-nitrophenol ONPX 2-nitrophenyl (3-D-xylopyranoside ONPX2 2-nitrophenyl (3-D-xylobioside Ph Phenyl PNPX p-nitropheny 1- (3 -D-xy lopy ranoside PX Phenyl (3-D-xylopyranoside py Pyridine T. saccharolyticum Thermoanaerobacterium saccharolyticum Tf Trifluororriethane sulfonyl TFA Trifluoroacetic acid THF Tetrahydrofuran TLC Thin layer chromatography TMS Trimethyl Silyl Tr Triphenylmethyl Ts p-toluenesulfonyl Kinetic Constants Catalytic rate constant Dissociation rate constant for the enzyme-inhibitor compli Michaelis-Menten constant Maximal rate of an enzyme-catalyzed reaction Amino Acid Abbreviations Ala (A) Alanine Arg (R) Arginine Asn (N) Asparagine Asp (D) Aspartate Cys (C) Cysteine Glu (E) Glutamate Gin (Q) Glutamine Gly (G) Glycine His (H) Histidine He (I) Isoleucine Leu (L) Leucine Lys (K) Lysine Met (M) Methionine Phe (F) Phenylalanine Pro (P) Proline Ser (S) Serine Thr (T) Threonine Typ (W) Tryptophan Tyr (Y) Tyrosine Val (V) Valine kcat Ki K m V m a X l l l Acknowledgements I wish to express my thanks to my supervisor, Dr. Steve Withers, for guidance, patience, and encouragement. Special thanks to Dr. R. V. Stick and Dr. Roland Hoos for many helpful discussions, to Ms. Karen Rupitz for advice and technical assistance, and Mr. David Vocadlo for providing the (3-xylosidase enzymes. I would also like to thank Mr. Manish Joshi for assistance with 15N-NMR spectroscopy and the Piers lab for the use of many chemicals. I am grateful to everyone in the Withers group for advice, helpful discussions, but especially for their friendships which were a constant source of encouragement. Thank you to my family and friends for their unwavering love and support. Most of all, thank you to my wife and best friend, Kendiss, for sticking with me through it all. xiv For Kendiss Chapter One: Introduction 2 1.1 General The importance of proteins and nucleic acids in fundamental life processes on this planet is widely recognised within the scientific community.and beyond, but we are only beginning to understand the significance of carbohydrates. Yet in one form or another, carbohydrates comprise more than two thirds of all carbon in the biosphere and perform a wide variety of vital roles and functions. In addition to their traditional roles in energy storage and structural support, saccharides, either as free molecules, or as a part of larger constructs, are involved in many biological events such as protein folding, cell-cell recognition, cell-cell adhesion, inflammation, and metastasis. The enzymatic biosynthesis and degradation of oligo- and polysaccharides are central to carbohydrate research. These processes are catalysed by two important classes of enzymes; glycosyl transferases, which are involved in biosynthesis, and glycosidases (glycosyl hydrolases), which catalyse the hydrolysis of glycosidic bonds. Glycosidases in particular have warranted increasing attention, largely because of their potential industrial applications and medical significance. Xylanases for example are currently used in the pulp and paper industry as alternatives to chemical bleaching(i). Glycosidase inhibitors have received attention as potential chemotherapeutic agents for cancer(2) and HIV(3-5). Zanamivir is a potent inhibitor of influenza neuraminidase(5-5) and has been shown to be effective as a treatment for influenza A and B in phase III clinical trials(9). Glycosidases catalyse the hydrolysis of the C-0 bond between a sugar (glycone) and its anomeric substituent, the aglycone, as shown in Figure 1.1. The aglycone is often another sugar residue, but may be one of many different groups. Since glycosidases constitute a 3 large, diverse group of enzymes they are classified according to several criteria. A glycosidase is named after the glycone portion of the substrate for which it has the highest activity, and categorised as an a- or p-glycosidase according to the anomeric configuration of its substrate linkage. Further subdivision occurs according to the stereochemical outcome of the catalysed reaction. If the first formed product has the same anomeric configuration as the substrate, then the enzyme in question is classified as retaining. Similarly, if the product has the opposite stereochemistry at the anomeric centre then the glycosidase is referred to as an inverting glycosidase. Glycosidases have also been classified into families according to amino-acid sequence similarities(70-72). Figure 1.1 The general reaction catalysed by glycosidases. 1.2 Glycosidase Mechanisms 1.2.1 Inverting Glycosidases The inverting mechanism is shown in Figure 1.2. It is characterised by the direct attack of water at the anomeric centre assisted both by general acid and general base catalysis. The reaction is believed to proceed via an oxocarbenium ion-like transition state(7J). The active site of an inverting glycosidase contains two carboxyl residues, one positioned to act as the general acid assisting in departure of the aglycone, and the other as general base which activates the nucleophilic water molecule. 4 Figure 1.2 Proposed mechanism of an inverting [3-glucosidase. 1.2.2 Retaining Glycosidases Retaining glycosidases follow a two step mechanism which was first proposed by Koshland in 1953(74). The double displacement mechanism involves the formation of a covalent glycosyl-enzyme intermediate which, is subsequently attacked by a water molecule (Figure 1.3). The formation of this intermediate in the glycosylation step is achieved by attack of an active site carboxylate which acts as a catalytic nucleophile. At the same time, another carefully positioned carboxylic acid behaves as a general acid catalyst, thereby assisting in aglycone departure(73, 75). Subsequent hydrolysis (deglycosylation) results from direct attack of water on the covalent intermediate, this time through general base catalysis by the same residue which acted earlier as a general acid catalyst. The transition state for these steps is believed to have substantial oxocarbenium ion-like character, as shown in Figure 1.3. Noncovalent binding interactions between the enzyme and substrate are proposed to provide further rate enhancement by stabilisation of the transition state(7f5, 77). Figure 1.3 The double displacement mechanism, shown for a (3-xylosidase. 6 1.3 Evidence for the Double Displacement Mechanism. The double displacement mechanism has received general acceptance on the basis of a substantial body of experimental evidence. The following sections will highlight some of the more compelling evidence available to date in support of this mechanism's essential features. 1.3.1 The Nucleophile and Acid/Base Catalysts Koshland's proposal requires the correct spatial arrangement of two residues within the active site of a retaining glycosidase, one behaving as a catalytic nucleophile, the other as an acid/base catalyst. X-ray crystallography, affinity labels, site directed mutagenesis and detailed kinetic studies have been employed to demonstrate the validity of this assertion. Crystallography has revealed suitably positioned carboxyl groups in hen egg white lysozyme(iS, 19), (3-galactosidase from Escherichia coli(20), xylanase A from Streptomyces lividans(21), and Cellulomonas fimi exoglycanase(22) amongst many others(25). Glycosidases co-crystallised with substrate analogues(70, 24) or inhibitors provide still better structural evidence(25), but one must keep in mind that an inhibitor or substrate analogue may bind in a mode significantly different from that of the substrate. Recently, structures of Cel5A from Bacillus agaradhaerens in which the enzyme, either in its native form or complexed with bound substrate, a covalent intermediate, and product have provided strong evidence in support of the double displacement mechanism(26). Invaluable insights into the identity and function of catalytic residues may be obtained through site directed mutagenesis and detailed kinetic analysis. Potentially significant residues can be identified by sequence alignment with related enzymes. Replacement of 7 highly conserved residues by site directed mutagenesis and kinetic analysis of the enzyme mutants with a variety of substrates can demonstrate the importance of a specific group. For example, MacLeod and co-workers successfully identified the acid/base catalyst of Cellulomonas fimi exoglycanase (Cex) in this way(27). 1.3.2 The Nature of the Glycosyl-Enzyme Intermediate While mechanistic proposals include a stabilised ion-pair(79), prevailing wisdom supports a covalent intermediate for the double displacement mechanism. Positive a-secondary deuterium kinetic isotope effects (aDKIE) for Escherichia coli (lac Z) (3-galactosidase and Agrobacterium sp. (Abg) (3-glucosidase(26>) indicate that the intermediate has significant sp character at C-l and the transition state significant sp character. This is consistent with a covalent intermediate being cleaved via an oxocarbenium ion-like transition state. If the intermediate were an ion pair instead, one would expect an inverse effect. Other techniques have made the isolation of a covalent intermediate possible. Tao and co-workers employed low temperature 13C-NMR studies to demonstrate a covalent intermediate in pancreatic a-amylase using maltotetraose as substrate(29). Attempts have also been made to trap intermediate analogues through the use of specifically designed reagents. Conduritol epoxides for example, have been used to inactivate a number of enzymes, presumably through covalent modification of the active site nucleophile, reacting in a fashion similar to the natural substrate(75). These reagents have not proven sufficiently selective however, and have on occasion resulted in mislabelling. For example, the putative acid/base catalyst (Glu461) of E. coli (3-galactosidase was incorrectly identified as the catalytic nucleophile by labelling with conduritol C c/s-epoxide (\.\)(30, 31). 8 A novel class of mechanism based glycosidase inhibitors, the 2-deoxy-2-fluoro glycosides such as 1.2, provide a means to accumulate and identify covalent glycosyl-enzyme intermediates(3i-34). These inactivators are very similar to the enzyme's normal substrate, and are much more likely to interact with the enzyme by its true mechanism rather than by some other pathway. The fluorine substituent at C-2 slows both the glycosylation and deglycosylation steps by removing important stabilising noncovalent interactions and destabilising the positive charge in the transition state. Good leaving groups such as 2,4-dinitrophenolate increase the rate of glycosylation. The overall result is a greater rate decrease in the deglycosylation step than in the glycosylation step, and hence accumulation of the covalent intermediate. The catalytic competence of the intermediate has been demonstrated by reactivation of labelled Abg P-glucosidase(35). The presence of the glycosidic label, even after enzyme digestion, has been confirmed by 1 9 F NMR(J5) and more recently by X-ray crystallography(2(5). In combination with peptic digestion and a novel application of electrospray mass spectrometry (ES-MS) analysis in the neutral loss experiment these have allowed the identification of the catalytic nucleophile in a number of cases and demonstrated the formation of a covalent intermediate with the glycone(34). 1.1 1.2 Figure 1.4 Affinity labels for retaining glycosidases; conduritol C cis-epoxide (1.1), 2,4-dinitrophenyl-2-deoxy-2-fluoro-(3-D-glucopyranoside (1.2). 1.3.3 Oxocarbenium ion-like Transition states 1.3 Figure 1.5 The glucosyl cation. There is some debate surrounding the exact nature of the transition state in the retaining glycosidase mechanism. There are three structures which may contribute to a greater or lesser extent to the transition state (Figure 1.6)(57). Secondary deuterium kinetic isotope effects (otDKE) support an oxocarbenium ion-like transition state. An idealised glucosyl cation (1.3), as shown in Figure 1.5, distributes a full positive charge between 0-5 and C - l . It is also trigonal planar about these centres, requiring that C-5, 0-5, C - l , and C-2 be coplanar, distorting the ring into either a half-chair or a boat(2<5). ^OH .OH HO"" Figure 1.6 Proposed transition state contributors in the retaining mechanism of a (3-glucosidase. Positive ocDKIE measurements for both the glycosylation and deglycosylation steps indicate transition states with significant sp2 character at the anomeric centre(2S). Evidence from kinetic investigations on Abg p-glucosidase using a variety of substrates reveals a 10 change in the rate determining step as the pKa of the leaving group increases(2S). For substrates with good leaving groups (pKa < 8), the rate of reaction is independent of the leaving group, indicating that the deglycosylation step was rate limiting. For substrates with poorer leaving groups, glycosylation of the enzyme is rate limiting, and kcat is heavily dependent on leaving group ability. The magnitude of this dependence indicates a transition state with a large degree of bond cleavage, consistent with an oxocarbenium ion-like transition state. Synthetic substrates such as 1.2, 2-deoxy-2-fluoroglucopyranosyl fluoride (1.7)(3S), 2-deoxy-2,2-difluoroglucopyranosides (1.8)(J9) and 5-fluoro-D-glucopyranosyl fluorides (1.9)(40) contain substituents which inductively destabilise the developing positive charge in the transition state, and as a result are hydrolysed very slowly, often inactivating the enzyme by forming a covalent intermediate as described earlier in Section 1.3.2. This provides evidence for an oxocarbenium ion-like transition state with a substantial positive charge. The oxocarbenium ion-like model of the transition state is also supported by a large collection of potent inhibitors which resemble an oxocarbenium ion (13, 15, 41). Figure 1.7 Inactivators of retaining glycosidases. 1.3.4 Glycosidase Inhibitors Enzymes are believed to catalyse chemical reactions through preferential binding of the transition state relative to the ground state, so compounds which mimic the transition state have the potential for very tight binding as competitive inhibitors(42). Putative 1.7 1.8 1.9 11 transition state analogue inhibitors have proven to be valuable tools in the study of enzymatic mechanisms by shedding light on the nature of the transition state(75). The concept of transition state ana\ogy(43-46) will be discussed briefly in Appendix A. The most important characteristics for transition state analogy in glycosidases are believed to be a flattened half-chair conformation, trigonal planar geometry about the anomeric centre, and a positive charge distributed between C - l and the ring oxygen(75, 47, 48). The strong inhibition by compounds that resemble glycosyl cations supports the idea that an oxocarbenium ion-like transition state exists. Among the first known competitive inhibitors of glycosidases were the glyconic acid lactones (Figure 1.8)(49, 50). Aldonolactones; 1,5-lactones in particular, have long been known to inhibit (3-glycosidases and their inhibitory potency is explained by similarity to a cyclic oxocarbenium ion(75). X-ray(57) and NMR data(52) indicate that six-membered lactones exist primarily in a distorted half-chair conformation and are trigonal planar about the anomeric carbon as a result of sp2 hybridisation. Charge delocalisation also contributes to their inhibitory potency since the lactone bears a partial positive charge at the endocyclic oxygen, and may therefore interact favourably with negatively charged groups within the active site. Inhibition of glycosidases by lactones has proven difficult to quantify however as 1,5-lactones will quickly hydrolyse to aldonic acids and interconvert to 1,4-lactones in aqueous solution(75). OH OH 1.10 Figure 1.8 Contributing resonance structures of D-gluconolactone. 12 1.4 Nitrogen-Containing Glycosidase Inhibitors 1.4.1 Aldonolactams Aldonolactams, the nitrogen-containing analogues of aldonolactones, have similar properties to aldonolactones, but are stable to hydrolysis in aqueous solution (Figure \.9)(15). They are also trigonal planar about C-l, and have an even larger charge separation than their lactone analogues. Studies have shown that in many cases aldonolactams have inhibition constants of the same order of magnitude as their lactone analogues(i5). 1.11 Figure 1.9 D-Gluconolactam. 1.4.2 Glycosyl Amines Glycosyl amines have an amino group in place of the glycosidic oxygen. These basic sugar analogues such as 1.12 (Figure 1.10) are believed by some to emulate the protonated substrate (1.4, Figure 1.6) (37) and are bound by both a- and (3-glycosidases(/5). Inhibition by glycosyl amines can be most easily explained by ion pair formation between the protonated form of the inhibitor and a negatively charged group in the active site(75). Substituents on the nitrogen that weaken or abolish its ability to accept a proton greatly decrease inhibition. Kinetic studies with glycosyl amines are complicated once again, however, by spontaneous hydrolysis and rapid anomerisation in aqueous solution. 13 HO HO •NH 2 1.12 Figure 1.10 (3-Glucosylamine. 1.4.3 Nojirimycin and 1-Deoxynojirimycin In 1965 Nishikawa and Ishida discovered a naturally occurring antibiotic designated R-468(53"). This compound, later named nojirimycin (1.13), was the first sugar analogue identified with a basic nitrogen within the sugar ring. Both nojirimycin and 1-deoxynojirimycin (1.14), strongly inhibit a and P-D-glucosidases(i5). Figure 1.11 Nojirimycin (1.13) and 1-deoxynojirimycin (1.14). Nojirimycin (1.13), 1-deoxynojirimycin (1.14) and their homologues are often presented as transition state analogues(54). The positive charge on the ring nitrogen in its protonated form is proposed to mimic the oxocarbenium ion-like transition state in charge distribution. It is not possible however, to conclude that an inhibitor is a transition state analogue simply because it binds very tightly to an enzyme, even if the inhibitor is carefully designed to incorporate features of the transition state(55, 56). Rigorous techniques to confirm transition state mimicry do exist however. Corresponding changes to the structure of a substrate and a transition state analogue should OH OH 1.13 1.14 14 affect the rate constant for hydrolysis (kcat/Km) and inhibitor binding (Kj) in a similar manner(44). As a result, there should be a linear free energy relationship between log Kj and log(Km/k c a t) if the inhibitor does mimic the transition state (See Appendix A). Studies of this type were undertaken for a series of 1-deoxynojirimycins with Abg P-glucosidase(57, 58). No significant correlation was found between these parameters, indicating that 1-deoxynojirimycin is not a transition state analogue for this P-glucosidase by this criterion. The effect of pH on inhibition by 1.14 differs from the pH dependence of enzyme activity for almond (3-glucosidase providing further evidence against transition state analogy by 1.14 since changes in pH must have the same effect on transition state binding and enzyme activity(59). Instead, compounds of this type may be viewed as fortuitous binders(4i). As shown in Figure 1.12, nojirimycin exists in equilibrium with its dehydrated form in aqueous solution. The suggestion has been made that since the dehydrated form is planar about the anomeric centre, then inhibition may arise solely from binding of this species, even though it exists only as a minor component(<5<9). Many a and (3-glycosidases are also strongly inhibited by 1-deoxynojirimycin (1.14) however, which can not dehydrate(iJ). Figure 1.12 Spontaneous dehydration of nojirimycin. The inhibitory potency of nojirimycin and its analogues is generally believed to arise from ion pair formation between a carboxylate in the active site and the inhibitor in its protonated form(75). The exact mode through which this interaction comes about is not entirely clear in every case. There are two ways in which the interaction between these 15 iminosugars and a glycosidase may arise. It is possible for an inhibitor in its protonated form to diffuse into the active site and interact with an anionic residue. On the other hand, the neutral form of the inhibitor may diffuse into the active site where it is protonated by an acidic residue. The ionised inhibitor can now form an ion pair with a negatively charged residue, probably the very group which gave up its proton in the first place(7J, 61). These two pathways are difficult to distinguish experimentally. Which of these pathways applies to a specific system may in theory be determined by careful pH studies, but this often proves difficult in practice since the pKa values for iminosugar inhibitors and carboxylic acid residues are quite similar(75, 59). It is sometimes possible to determine the mode of binding if both a basic inhibitor and an analogous cationic compound are available. For example, if both 1-deoxynojirimycin (1.14) and its permanently cationic derivative 1.15 were both able to bind tightly to an enzyme, and their binding shows the same pH dependence, this would support binding of 1-deoxynojirimycin directly through its ionised form(75). This is exactly what is observed for glucosidase I, an a-glucosidase from the endoplasmic reticulum(52). In cases where the cationic derivative does not bind however, one may not make any conclusions since steric factors may be the cause of diminished binding, rather than the ionisation state of the inhibitor(75). OH 1.15 Figure 1.13 MiV-Dimethyl- 1-deoxynojirimycin. 16 1.4.4 Derivatives of 1 -Deoxynojirimycin In the search for effective glycosidase inhibitors with improved selectivity, specificity and biological activity, many derivatives of known glycosidase inhibitors have been synthesised(54). The 1-deoxynojirimycins have proven an attractive starting point due to their inhibitory potency and the relative ease with which they may be modified through TV-alkylation. One such derivative, Miglitol (TV-hydroxyethyl-1-deoxynojirimycin, 1.16) is commercially available as a treatment for diabetes(<5J, 64). TV-butyl-1-deoxynojirimycin (1.17) appears to have antiretroviral activity(3, 65). The effects of TV-substitution, which may influence the pK a of the 1-deoxynojrimycin derivatives and thereby influence their potency and specificity have also been studied(<5(5, 67). Derivatisation at the anomeric centre has also been investigated(68). OH 1.16 R = (CH 2) 2OH 1.17 R = rt-Butyl Figure 1.14 TV-Alkyl derivatives of 1-deoxynojirimycin with biological activity: miglitol (1.16), and TV-buty 1-1-deoxynojirimycin (1.17) 1.4.5 1-Aza Sugars A new group of hydroxypiperidines, the 1-aza sugars, have a nitrogen in place of the anomeric carbon rather than the endocyclic oxygen. The basic atom in place of C- l is believed to mimic the carbocationic form of a glycosyl cation (Figure 1.6)(57). Isofagomine (1.18) and its relatives are effective inhibitors of (3-glycosidases(37, 69-73). Indeed, 17 isofagomine is one of the strongest inhibitors of P-glucosidase from almonds (Kj = 0.1 pM). This is surprising since it lacks a hydroxyl group at C-2, the loss of which causes fagomine (1.19) to bind 280 times worse than 1.14 to P-glucosidase from almonds(74). Isofagomines are much less potent inhibitors of a-glycosidases however(57, 73). Superior binding of isofagomines to P-glycosidases relative to deoxynojirimycins suggests interaction with a carboxylate which is closer to the anomeric centre than to the endocyclic oxygen(75). Orientation, not just proximity, must also play a role, however(<57). It has also been proposed that the preferential inhibition of p-glycosidases by these inhibitors is an indication that the glycosyl carbocation resonance form of the oxocarbenium ion (1.3), believed to exist in an undistorted 4Ci chair (Figure 1.6) is a more important contributor to the transition state of p-glycosidases than for a-glycosidases, and that cleavage in p-glycosidases therefore occurs through the ground state conformation of the substrate(57). However, there is no particular support for this conclusion since recent X-ray studies have indicated that distortion of substrates from a 4 Q chair does occur in hen egg white lysozyme(75), endoglucanase 1(70) and Cel5A from Bacillus agaradhaerens(26). 1.18 1.19 Figure 1.15 Isofagomine (1.18) and fagomine (1.19). 1.4.6 New Glycosidase Inhibitors with sp2-Hybridised 'Anomeric Centres' While the presence of a positive charge is an important feature in many glycosidase inhibitors, a flattened anomeric conformation may be even more important in mimicry of the 18 oxocarbenium ion-like transition state(47, 55). In the last several years a large number of potential inhibitors derived from lactones and lactams have been synthesised and tested. Figure 1.16 depicts examples of these new inhibitor classes. O H 1.27 R = H 1.30 1.28 R = n-butyl 1.29 R = Dodecyl Figure 1.16 Glycosidases inhibitors with sp2 hybridisation at the anomeric carbon. Lactone oximes (1.20 - 1.22) are moderate inhibitors of P-glucosidase from sweet almonds(76). Glycohydroximolactams such as 1.23 strongly inhibit a-glycosidases(77) and P-glycosidases(7S). Substituted lactone and lactam oximes have also shown inhibitory potency(79, 80). Lactam O-alkyl oximes (1.25, 1.26) are strong inhibitors of p-glucosidase from Caldocellum saccharolyticum, especially when the substituent is a sugar moiety(6>7). Another interesting inhibitor type is based on gluconamidine (1.27), a hydrolytically unstable yet potent inhibitor of glucosidases(6>2). Amidrazones (1.30) and N-alkyl amidine derivatives such as 1.28 and 1.29 are relatively stable and have shown effective inhibition of several P-glucosidases(6>2, 83), almond P-glucosidase(S2), and a number of a and p-19 mannosidases(6>2, 84) are inhibited by amidrazones. . O H . O H HO v N ~ N HO v ^ X ~ N O H O H 1.31 1.32 1.33 Figure 1.17 D-Gluconojirimycin tetrazole (1.31), imidazole (1.32), and triazole (1.33). Bicyclic inhibitors such as D-gluconojirimycin tetrazole (1.3l)(48, 85) and imidazole (1.32)(S6, 87) are moderate to strong inhibitors of a number of glycosidases (Figure 1.17). These compounds do not contain a basic group, and binding is believed to arise from hydrogen bonding of the exocyclic heteroatom closest to C-l with the acid/base catalyst(S7, 88). This is consistent with stronger inhibition by imidazoles than tetrazoles, as imidazoles are more basic. As further evidence, inhibition is abolished when this crucial heteroatom is removed as in the triazole (1.33)(S9). This mode of binding implies that the acid/base catalyst may be located to the side of, rather than above, the anomeric centre as depicted in Figure 1.18. A ) B) , O H O H H O - V ^ N J N o n - ^ Q O^r-O 0 ^ 0 ^ B Figure 1.18 A) Proposed direction of hydrogen bonding with lactam derivatives. B) 'Side on' protonation of substrate. 20 These groups of compounds are often described as transition state analogues for the enzymes they act on, a conclusion usually based only on their inhibitory potency, even though tight binding provides no guarantee of transition state mimicry. Partial, though imperfect transition state character has been attributed to amidines on the basis of 'H-NMR studies and molecular modelling(90). Studies where the ratios (Km/Kj) and (k c at/k u n c a t) are compared have also been used as evidence for partial transition state character by amidines(5'5). However, these criteria are not sufficient to conclusively prove or disprove transition state analogy. The nojirimycin tetrazoles were the first glycosidase inhibitors for which a rigorous treatment of transition state mimicry was conducted(4S). Detailed kinetic studies have demonstrated a linear free energy relationship between log K, for D-mannonojirimycin tetrazole and D-gluconojirimycin tetrazole and log (Km/k c a t) for corresponding manno- and gluco-configured substrates for a number of a- and P-glycosidases. They also show no correlation between log K; and log K m , indicating the absence of inhibitor binding through analogy to the substrate. In addition, D-gluconojirimycin tetrazole binds 200 times less tightly to a mutant of Abg P-glucosidase in which the catalytic nucleophile is replaced (Glu358Asp) than the wild-type. This residue is very important to transition state binding but is not believed to play a role in ground state binding(97). This is also consistent with transition state mimicry by nojirimycin tetrazoles. 1.5 Glycosidases in Enzymatic Synthesis Oligosaccharide synthesis presents major obstacles to synthetic chemists and enzymatic synthesis of oligosaccharides is an attractive alternative to lengthy and expensive 21 chemical synthesis. Glycosyl transferases can and have been used in the synthesis of new compounds, but these enzymes have limited availability and require expensive substrates(92, 93). Retaining glycosidases have the potential for transglycosylation rather than hydrolysis when it is favoured by high concentrations of a glycosyl acceptor(92, 94). As shown in Figure 1.19 for Abg P-glucosidase, the glycosyl enzyme intermediate can react with another sugar rather than with water, thereby yielding a transglycosylation product(95). This approach has been demonstrated with Abg using a variety of glycosyl acceptors(95). A cellulase from Trichoderma sp. was recently used to prepare 'deoxycellobionojirimycin'(96). Unfortunately, the transglycosylation products are themselves subject to cleavage by the enzyme, resulting in diminished yields(95). Glul70 Glul70 Glu358 Glu358 Figure 1.19 Interception of glycosyl-enzyme intermediate resulting in transglycosylation by Abg p-glucosidase. Recently, a mutated version of Abg p-glucosidase has shown potential as a "glycosynthase"(97). The Glu358Ala mutant of Abg (AbgGlu358Ala) in which the catalytic nucleophile has been replaced by alanine is able to utilise activated glycosyl donors with opposite configuration at the anomeric centre in the synthesis of oligosaccharides as shown in 22 Figure 1.20. Since the mutant lacks a catalytic nucleophile, it is not able to hydrolyse the product at an appreciable rate(9S) giving improved yields. This technique has the potential for widespread application to many other p-glycosidases, creating a library of 'glycosynthases'. Glul70 Glul70 Ala358 Ala358 Figure 1.20 Proposed mechanism of transglycosylation by AbgGlu358Ala. 1.6 P-Xylosidases The most abundant component of plant cell wall hemicelluloses is xylan, a polymer composed primarily of D-xylopyranosyl units connected through P-l,4-linkages(99). Enzymatic degradation of xylan is an attractive alternative to the environmentally damaging process of chemical bleaching in the pulp and paper industry(7OO). The complete breakdown of xylan requires the concerted action of several enzymes including P-xylosidases, glucuronidases and P-xylanases. P-Xylosidases, which cleave xylose residues from the non-reducing end of the polymer are classified within families 39 and 43 of the glycosyl hydrolases(707, 702). A 55 kDa P-xylosidase has been isolated from Thermoanaerobacterium saccharolyticum strain B6A-RI, an organism isolated from Yellowstone National Park, 23 which subsists on xylan as its primary carbon source. T. saccharolyticum (3-xylosidase is a member of family 39(702). Thus far, there are no crystallographic data available on this enzyme and it has not yet been studied to any great extent. It cleaves xylo-oligosaccharides and nitrophenyl-P-D-xylopyranosides but has no activity against oat spelt xylan(702). T. saccharolyticum P-xylosidase is stable to high temperatures and displays optimum activity at pH 5.5, and 7O°C(702). T. saccharolyticum P-xylosidase has been shown to operate with overall retention of anomeric configuration(707) and is believed to follow the canonical double displacement mechanism. The catalytic nucleophile has been identified as Glu-277 through derivatisation with 2,4-dinitrophenyl-2-deoxy-2-fluoro-P-D-xyloside and detection of the glycosylated peptide in a pepsin digest thereof by electrospray MS in the neutral-loss mode(705J). The acid/base catalyst has also been tentatively assigned to Glul60 by sequence alignment with other family 39 enzymes and kinetic studies of the Glul60Ala mutant of T. saccharolyticum p-xylosidase(704). T. saccharolyticum p-xylosidase exhibits significant transglycosylation activity when p-nitrophenyl-p-D-xylopyranoside (PNPX) is used as substrate(707). The transglycosylation activity is observed even under conditions which strongly favour hydrolysis over transglycosylation: aqueous solution without organic solvent, and low substrate concentrations. The transglycosylation reaction lacks regioselectivity however, yielding a mixture of nitrophenyl oligosaccharides containing p-1,2, P-1,3, and P-1,4 linkages(707). 1.7 Aims of Thesis 7.7.7 Synthesis of and Testing of Potential Xylosidase Inhibitors Xylosidases have not been studied as extensively as other glycosidases and the 24 decision was made to investigate the effects of changes in inhibitor structure on T. saccharolyticum p-xylosidase, an enzyme which has not been greatly studied to date. Studies with xylo-configured inhibitors may also provide a means to probe mechanistic differences between retaining and inverting p-xylosidases. The synthesis of pentose iminosugars is also advantageous because it presents fewer synthetic obstacles than the synthesis of hexose analogues. Aldohexoses have a stereocentre at C -5, and maintaining the correct stereochemistry at this centre can be difficult when replacing the endocyclic oxygen. In aldopentoses however, C -5 is not a stereogenic centre, greatly simplifying the introduction of nitrogen in the synthesis of pentose iminosugar analogues. In addition, the synthetic routes and techniques used in the synthesis of these sugar analogues will parallel those needed in the preparation of xylanase inhibitors in the future. There are several features which may contribute to the transition state character and binding of a glycosidase inhibitor including a positive or basic centre, sp2 hybridisation at the anomeric centre, flattened half-chair conformation, and correct stereochemistry at all centres(75, 55, 61). A goal of this work is to explore how changes to certain structural features influence the binding of nitrogen-containing sugar analogues to T. saccharolyticum p-xylosidase. l-Deoxyxylonojirimycin will be tested as an inhibitor of T. saccharolyticum p-xylosidase. The pH dependence of T. saccharolyticum p-xylosidase inhibition by 1-deoxyxylonojirimycin will also be investigated to gain insight into the mode of inhibitor binding. The effect of substitution at the ring nitrogen and the importance basicity of the ring nitrogen to inhibitor binding will be investigated by measuring the inhibition of T. saccharolyticum P-xylosidase by derivatives of 1-deoxyxylonojirimycin (1.35) which vary in 25 these respects. Inhibitors such as D-xylonolactam (1.34), which exhibit sp2 hybridisation at the anomeric centre will also be tested as inhibitors of T. saccharolyticum [3-xylosidase. Synthetic routes to a collection of nitrogen-containing sugar analogues in the xylo-configuration which vary in these respects will be pursued, and the inhibitory potency of these potential inhibitors will be measured against wild-type T. saccharolyticum p1-xylosidase, and its Glul60Ala mutant in which the acid base catalyst has been removed. Figure 1.21 D-Xylonolactam (1.34) and 1-deoxyxylonojrimycin (1.35). 1.7.2 Preparation of Oligosaccharide Iminosugar Analogues by Enzymatic Transglycosylation. Oligosaccharides containing an iminosugar residue are of great potential interest as inhibitors of endoglycosidases such as Bacillus circulans xylanase (BCX). Enzymes of this type do not bind monosaccharides and inhibition studies require oligosaccharide inhibitors. The preparation of disaccharides or trisaccharides containing an imino sugar by chemical synthesis is a daunting task however, and chemo-enzymatic strategies such as the use of a glycosynthase to form the glycosidic linkage are desirable alternatives. The AbgGlu358Ala glycosynthase, and the Glu277Ala mutant of T. saccharolyticum p-xylosidase in which the catalytic nucleophile has been replaced, will be tested as catalysts for the preparation of imino sugar-containing oligosaccharides using xylo-configured iminosugars as glycosyl acceptors. Oligosaccharide imino sugars such as OH 1.34 1.35 26 1-deoxyxylobionojirimycin (1.36) may provide valuable insights into the binding interactions of enzymes such as Cellulomonas fimi exoglycanase and Bacillus circulans xylanase. Figure 1.22 1-Deoxyxylobionojirimycn (1.36). 1.7.3 Preparation of Chromophoric Disaccharide Substrates for Xylanases Kinetic studies of Bacillus circulans xylanase are complicated by the lack of useful substrates. Conventional assays of xylanase activity employ substrates of inconsistent and uncharacterised composition. Aryl xylobiosides are the most useful substrates for kinetic analysis of xylanases, but these disaccharide substrates are not readily available. Chemical syntheses of these compounds from monosaccharide starting materials are laborious and time consuming(7G>5, 106). The most logical route to aryl xylobiosides requires a source of xylobiose, which is a known compound, but it is not readily available commercially and a convenient method for its preparation has not yet been presented. A chemo-enzymatic route to aryl xylobiosides in which Trichoderma viride xylanase is used to hydrolyse xylan to xylobiose will be investigated as a more practical alternative. 1.36 Chapter Two: Results and Discussion 28 2.1 Synthesis 2.1.1 Synthesis of Nitrogen-Containing D-Xylopyranose Analogues The syntheses of several nitrogen-containing sugar analogues based on D-xylonolactam (1.34) and 1-deoxyxylonojirimycin (1.35) were attempted. The target compounds are shown in Figure 2.1. 1.35 R = H 2.6 2.7 2.8 R - H 2.3 R = CH 2Ph 2 9 R = D 2.4 R = COCH 3 2.5 R = CHO Figure 2.1 D-Xylo-configured sugar analogues containing an endocyclic nitrogen. 2.1.1.1 l-Deoxyxylonojirimycin A relatively straightforward route to 1-deoxyxylonojirimycin from D-xylose was developed by Godskesen et al(107). The critical ring closure was achieved in two steps by treatment of mesylate (2.10) with aqueous ammonia followed by acetonide deprotection as shown in Figure 2.2. The five-membered acetonide ring prevented concomitant lactamisation and the amino amide (2.11) spontaneously cyclised once this constraint was removed. D-Xylonolactam (1.34) was converted to 1-deoxyxylonojirimycin (1.35) in 95% yield through 29 the stepwise protection with TMS protecting groups, borane reduction and subsequent hydrolysis of the silyl ether protecting groups. N H 3 , H 2 0 • 1% H C l , MeOH, : reflux, 38% O H 1.34 1) HMDS, TMSC1 2) B H 3 . S M e 2 3) H C l , H 2 0 95% HO HO NH.HC1 [ O - ^ ^ OH 1.35 Figure 2.2 Synthetic route towards l-deoxyxylonojirimycin(7(97). A 15N-labelled sample of 1-deoxyxylonojirimycin was desired as a precursor to 1 5N-labelled disaccharide inhibitors to be used in 15N-NMR studies with BCX. The procedure shown above was modified slightly in order to prepare an 15N-labelled sample of 1-deoxyxylonojirimycin. In place of aqueous ammonia, 2.10 was treated with a mixture of excess 15N-enriched ammonium chloride and sodium hydroxide in DMF affording the ^ re-labelled 2.12 in 25% yield (Figure 2.3). The lactam (2.12) was converted to 15N-1-deoxyxylonojirimycin (2.6) by a route identical to that described above for 1.35. 30 1) 1 5NH 4C1, NaOH, DMF H Q • HO 2) 1% HC1, MeOH 25% O OH 2.10 2.12 Figure 2.3 Preparation of 15N-D-xylonolactam 2.1.1.2 ^-Substituted Derivatives of 1 -Deoxyxylonojirimycin The benzyl carbamate (2.8) and Af-benzylamine (2.3) derivatives of 1-deoxyxylonojirimycin were prepared directly from 1.35 in moderate yield without the need for protecting groups. Exposure of 1.35 to benzyl chloroformate and 2 equivalents of triethylamine in aqueous solution furnished carbamate 2.8 in 56% yield without reaction of the hydroxyl groups. The /V-benzyl derivative (2.3) was prepared in 84% yield by treatment of 1.35 with excess benzyl bromide and triethylamine in DMF. o 2.3 Figure 2.4 Synthesis of /V-benzyloxycarbonyl- 1-deoxyxylonojirimycin (2.8), and ZV-benzyl-1-deoxyxylonojirimycin (2.3). 31 /V-Acylated derivatives of 1-deoxyxylonojirimycin were prepared by the use of pentafluorophenyl ester reagents. These reagents allow for rapid and selective acylation of unhindered amines under very mild conditions(70#, 709). TV-Acetyl-1-deoxyxylonojirimycin (2.4) was prepared in 70% yield by reaction of 1.35 with pentafluorophenyl acetate and triethylamine in DMF. To simplify purification, TV-formyl- 1-deoxyxylonojirimycin (2.5) was prepared in two steps in 90% overall yield by treatment of 2,3,4-tri-t9-benzyl-l-deoxyxylonojirimycin (2.13) with pentafluorophenyl formate in dichloromethane followed by catalytic hydrogenolysis. The preparation of 2.13 will be discussed in Section 2.1.1.4. O H O ^ - \ ' N H H C l Q Et 3N, DMF H O - " " \ H O - W - ^ + If • H O - V - - * * A OH C 6 F s O ^ 70% OH 1.33 2.4 O B n O ^ ' N « O DCH2C12 HO^X B n O - * ^ \ ^ + J l • H O - V - - ^ A OBn C 6 F 5 0 ^ 2)H 2 Pd/C,AcOH OH. 2.13 MeOH, 92% 2 5 Figure 2.5 Synthesis of TV-acyl derivatives of 1-deoxyxylonojirimycin. 2.1.1.3 Attempted Synthesis of D-Xylonolactam oxime (2.2) Vasella et al. have reported the synthesis of D-gluconolactam oxime (1.20) from benzyl protected D-gluconolactam (2.14) as shown in Figure 2.6 (78, 87). Significant epimerisation can occur at C-2 upon treatment of the lactam with Lawesson's reagent however(7S). The route of Vasella et al. also required a number of deprotection and reprotection steps, adding significantly to the complexity of the synthetic route. 32 BnO Lawesson's reagent BnO NH 2 OH.HCl NaHC0 3 , BnO BnO' BnO BnO :o 99% S MeOH, 92% 1) Na, N H 3 2) Ac 2 0, Py 66-78% HO AcO NaOMe HO A c O —OAc N - - O H 88% N Figure 2.6 Synthesis of D-gluconolactam oxime from 2,3,4,6-tetra-(3-benzyl-D-gluconolactam(76>, 87). To reduce the risk of epimerisation and shorten the synthesis, a new strategy towards the synthesis of D-xylonolactam oxime (2.2) was proposed from tri-O-acetyl-D-xylonolactam (2.18). Protected lactam 2.18 was prepared in 55% yield by treatment of D-xylonolactam with acetic anhydride and pyridine. Exposure of lactam 2.18 to Lawesson's reagent provided the thionolactam (2.19) in 71% yield. The thionolactam (2.19) was exposed to hydroxylamine and the crude reaction mixture was treated with acetic anhydride in pyridine. Unfortunately, the expected product was not isolated. In fact, tic analysis did not reveal any significant organic material after workup using either molybdate stain or sulfuric acid charring. This result was unexpected as a parallel route was recently used to prepare the xylobiosyl lactam oxime (2.22) as shown in Figure 2.8(770). Hydroxylamine may have acted as a base rather than a nucleophile, thereby leading to elimination products such as those shown in Figure 2.7B. Presumably, these side products were not detected as they were extracted into the aqueous phase on workup. Formation of an aromatic, substituted pyridine, can be seen as the driving force for this side reaction. 33 A) Ac 2 0 , Py AcO ^ AcO 55% 1.34 Lawesson's reagent ^ A c G ^ AcO OAc 7 1 % 2.18 OAc 2.19 B ) AcO"' • ft-AcO OAc 2.19 - N H S NH 2OH_ A c 0 , , s NH 2 OH^ / / " N SH OAc OH 2.20 2.21 Figure 2.7 A) Attempted synthesis of D-xylonolactam oxime from D-xylonolactam. B) Proposed side reaction upon treatment of 2.8 with hydroxylamine. AcO : 0 ^ ^ ^ ^ * \ ^ 0 - ^ V . AcO NH OAc 2.22 OAc 1) NH 2OH, NaHC0 3, MeOH 2) Ac 20, Py Figure 2.8 Synthesis of xylobionolactam oxime(770). In order to avoid these difficulties, it would be necessary to employ protecting groups which are less prone to eliminate under basic conditions. One such approach would be to use the benzyl protected lactam (2.26) and proceed as shown above in the D-gluconolactam oxime case. Care would be required to minimise epimerisation at C-2 in formation of the thionolactam however, and additional protection steps would also be required for isolation of 34 the product. Another approach to lactam oximes was earlier applied by Ganem et al(47). In this approach the trimethylsilyl protected lactam was converted to the thionolactam without epimerisation. Treatment of the thionolactam with hydroxylamine afforded the lactam oxime directly in good yield (Figure 2.9). TMSO TMSO TMSO NH 1) Lawesson's reagent HO >- HO O 2) HC1, MeOH HO OTMS 63% NH NH 2.24 2.25 HO-NH 2 OH • HO-S 75% H O — V ^ ^ - ^ - ^ N -OH OH 1.20 -OH Figure 2.9 Synthesis of D-gluconolactam oxime by Ganem et a\(47). 2.1.1.4 Synthesis of 2,3,4-Tri-O-benzyl-D-xylonolactam (2.26) Attention now shifted to the synthesis of the benzyl protected D-xylonolactam (2.26) which would provide a useful starting material in routes toward a number of interesting species (Figure 2.10). Notably, D-gluconojrimycin tetrazole (1.31), imidazole (1.32), and D-gluconolactam oxime (1.20) have all been prepared from a similarly protected D-gluconolactam(55, 77, 78, 111, 112). In addition, 2.26 is a useful starting material for the preparation of 1,1-dideutero- 1-deoxyxylonojirimycin (2.7) and its benzyl carbamate derivative (2.9) which proved necessary for transglycosylation studies (see 2.1.2). Protected lactam 2.6 would also provide potential avenues towards D-xylonojirimycin tetrazole 2.1, imidazole 2.27, triazole 2.28, and lactam oximes 2.2, and 2.29. 35 2.2 R=H 2.29 R=C(0)NHAr Figure 2.10 Compounds accessible from tri-O-benzyl-D-xylonolactam (2.26). The synthesis of 2.26 proved to be more challenging than was anticipated. Initially, the preparation of 2.26 was attempted by treatment of unprotected lactam (1.34) with benzyl bromide and sodium hydride. Unfortunately, these conditions produced a complex mixture and the desired product was not isolated, presumably as a consequence of /V-alkylation and elimination reactions in the presence of the strong base. Benzyl trichloroacetimidate, in combination with an acid catalyst, provides a mild system for the preparation of benzyl ethers in cases where strong bases can not be used. Benzyl trichloroacetimidate is easily prepared from benzyl alcohol and trichloroacetonitrile upon exposure to catalytic sodium hydride(7i5). A number of monosaccharide derivatives 36 containing from one to three hydroxy groups have been alkylated with this reagent(774). 0-alkylation in the presence of acetal, ester, amide and lactone functional groups proceeds without complication^ 73, 775). These reaction conditions were unsuccessful when applied to 1.34 however, due to the poor solubility of the lactam in dichloromethane, 1,4-dioxane and DMF, the solvents most commonly used in this reaction. Since direct benzylation of D-xylonolactam (1.34) proved difficult, a new strategy was attempted in which benzyl groups would be introduced prior to formation of the lactam ring. Cyclisation of an amino amide could then furnish the desired product in a similar fashion as used for the preparation of the unprotected aminoamide (2.33) in Figure 2.11(707). HO OH 2.30 N H , HO O HO 2.34 O ^ N H 2 H O — — O H — O H ' Br 2.31 HO-O ^ N H 2 — O H :o 2.32 < W N H 2 HO--OH I—OH - N H 2 2.33 Figure 2.11 Conversion of 5-bromo-D-arabinolactone to D-arabinolactam(707). The retrosynthetic strategy outlined in Figure 2.12 proposed methyl 2,3,4-tri-O-benzyl-5-O-methanesulfonyl-D-xylonate (2.36) as a crucial precursor to the protected lactam 37 (2.26) via a route similar to that described above. This intermediate (2.36) would be obtained from the protected aldehyde 2.39. 2.39 2.38 2.37 Figure 2.12 Retrosynthetic strategy towards tri-O-benzyl-D-xylonolactam. The route to the triphenylmethyl ether (2.39) from D-xylose is outlined in Figure 2.13. The anomeric centre was masked as a diethyl dithioacetal giving 2.41 in excellent yield, leaving the remaining hydroxyl groups free. The polyol in 2.41 was selectively converted to a triphenylmethyl ether giving 2.42 and the remaining hydroxyl groups were benzylated with benzyl bromide and sodium hydride in 72% overall yield to give 2.43. Treatment of the dithioacetal 2.43 with mercuric chloride and calcium carbonate afforded the aldehyde 2.39 in 72% yield. Oxidation of 2.39 to the corresponding acid was unsuccessful upon treatment of the aldehyde with sodium chlorite and hydrogen peroxide in sodium dihydrogen phosphate buffer, as only starting material and decomposition products were observed upon workup. 38 HO HO [ 0 ^ ^ _ 2.40 EtSH.HCl HO-""V " 0 H SEt T r C 1 ' p y 0 H O H * * " H 0 ^ " ^ S E i 2.41 HO HO 2.42 OH^SEt -OTr OBn 2.39 BnBr, NaH B n 0 9 0 % NaC102, NaH 2PQ 4 H 2 0 2 - ^ ^ O T V S E t H g C l 2 , C a C 0 3 , ^ B n ° ~ ^ ^ O ^ S E t C H 3 C N , H 2 0 7 0 % 2.43 Starting material + Assorted products Figure 2.13 Unsuccessful route to tri-O-benzyl-D-xylonolactam. Rather than search for alternative means to prepare acid 2.38 from 2.39, this approach was abandoned for a more direct strategy towards lactam 2.26 which involved 2,3,4-tri-O-benzyl-D-xylonolactone (2.44) as a key intermediate (Figure 2.14). Methanolysis of this lactone would give an appropriately protected 5-hydroxy methyl ester, a direct precursor to mesylate 2.36. The lactone itself could be readily obtained from methyl (i-D-xylopyranoside (2.45), an inexpensive starting material. OMs OBn OCH 3 2.25 BnO o 2.36 H ( H O Q OCH3 OH 2.45 Figure 2.14 Retrosynthetic strategy to tri-(9-benzyl-D-xylonolactam from methyl p-D-xylopyranoside. 39 As shown in Figure 2.15A, benzylation of methyl (3-D-xylopyranoside (2.45) with benzyl bromide and sodium hydride provided 2.46 in 90% yield. Deprotection of the anomeric centre was performed by treatment of the methyl xyloside with sulfuric acid in acetic acid at reflux to give 2.47 as a mixture of anomers in 77% yield. Hemiacetal 2.47 was oxidised to the lactone 2.44 in 93% yield using DMSO and acetic anhydride. Attempts to open the lactone ring of 2.44 to afford a methyl ester were unsuccessful under both basic conditions with sodium methoxide in methanol and acidic conditions with HCl in methanol. Upon treatment of the lactone with base it is probable that a competing elimination reaction may have occurred, affording the cc,P-unsaturated lactone (2.48) as shown in Figure 2.15B(ii5). Treatment of the lactone with HCl in methanol was unsuccessful presumably due to the stability of the 6-membered ring. Heating of the lactone under acidic conditions lead to decomposition rather than acyclic ester formation. A) H Q — \ " ° \ BnBr, NaH B n P - ^ \ ° \ H O - X - ^ - ^ O C H j 1 • ^ n O - A ^ ^ ^ - O C H 3 X0 H DMF, 90% NOBn 2.45 2.46 I AcOH, I H 2 S0 4 , 77% ^DMSO,Ac 2Q B r £ ^ \ ~°> 93% OH OBn 2.47 -Q NaOMe^ BnO { )=0 OBn 2.48 Figure 2.15 A) Preparation of 2,3,4-tri-O-benzyl-D-xylonolactone. B) Proposed side reaction upon treatment of lactone (2.44) with base. 40 While 2.44 could not be converted to an acyclic methyl ester in our hands, 2,3,4-tri-O-benzyl-D-xylononamide (2.49) was formed in nearly quantitative yield upon treatment of lactone 2.44 with ammonia in methanol (Figure 2.16). Treatment of amide 2.49 with methanesulfonyl chloride in pyridine provided the mesylate 2.50 in 79% yield. When 2,3,4-tri-O-benzyl-5-methanesulfonyl-D-xylonamide was treated with ammonia in an attempted cyclisation only decomposition was observed, however. This outcome was not expected as the mesylate 2.10 reacted under identical circumstances. In this case, displacement of a mesylate by ammonia may proceed by way of an intermediate epoxide as discussed for the related bromide (Figure 2.11). Thus, the reaction may have failed for want of a vicinal hydroxyl group. NH r ] n O - _^-*—Q » - U I I V J ^ — x ^ - j j — - ^ n 2 » - U I » J N V ^ - ^ i > n 2 OBn 95% B n 0 0 79% B n O Q 2.44 2.49 2.50 Figure 2.16 Synthesis of 2,3,4-tri-0-benzyl-5-t9-methanesulfonyl-D-xylonamide (2.50). The literature contains a number of instances of /V-alkylation of amides(776). If the amide is first converted to its conjugate base, treatment of the amide anion with alkyl halides or sulfonates may afford the ZV-alkylated product in good yield(777, 775). Thus, Zwierzak and co-workers have reported the use of sodium hydroxide in combination with t-butylammonium hydrogen sulfate, a phase transfer reagent, in the N-alkylation of amides and sulfonamides with simple alkyl halides and sulfonic acid esters(777, 118). Application of this method did not afford the desired cyclisation product, however. Exposure of the mesylate 2.50 to phase transfer conditions led only to slow hydrolysis of the methanesulfonate ester. 41 Treatment of the mesylate 2.50 with n-butyl lithium or sodium hydride did not result in lactamisation. Instead, the major reaction product was the benzylated lactone arising from O-alkylation and subsequent hydrolysis on workup. Conversion of the methanesulfonate group to a trifluoromethanesulfonate group was intended to increase the vulnerability of this centre to nucleophilic attack by the amide. Treatment of 2.49 with trifluoromethanesulfonic anhydride again led to the formation of the lactone 2.44, presumably by way of immediate O-alkylation of the expected triflate derivative (2.52) which was not isolated (Figure 2.17). L 2-52 J Figure 2.17 Lactone formation upon treatment of 2,3,4-tri-<9-benzyl-5-methanesulfonyl-D-xylononamide. Conversion of the mesylate 2.50 to a primary amine by introduction of an azide group was the next strategy attempted (Figure 2.18). Displacement of the methanesulfonate group by sodium azide in DMPU provided 5-azido-2,3,4-tri-(3-benzyl-D-xylonamide 2.53 in 95% yield. Subsequent exposure of the azide to triphenylphosphine in aqueous THF in a modified Staudinger reaction resulted in formation of the lactam 2.26 in 91% yield. This reaction is believed to involve a short-lived phosphinimine intermediate 2.54, the hydrolysis of which forms an amino amide which spontaneously cyclises. The azide group was also reduced in a 42 separate experiment by treatment of 2.53 with hydrogen sulfide and pyridine and once again the lactam was obtained. This offers support for the mechanism shown in Figure 2.18 where the intermediate phosphinimine is hydrolysed prior to reaction with the amide. NaN 3 > DMPU n _ „ ^ N 3 B v- OMs h O - X - ^ ^ T r - - - N H 2 BnOfj 2.50 BT8 nO BnO BnO NH, NH, B n 0 0 2.35 91% NH OBn 2.26 95% nO nO NH, BnOQ 2.53 Ph3P, H 2 0 - N = P P h 3 -NH 2 BnOQ 2.54 Figure 2.18 Successful synthesis of tri-O-benzyl-D-xylonolactam. 2.1.1.5 Synthesis of 1,1 -Dideutero-N-benzyloxycarbonyl-1 -deoxyxylonojirimycin. iV-Benzyloxycarbonyl- 1-deoxyxylonojirimycin (2.8) was intended to be used as a glycosyl acceptor in enzymatically catalysed transglycosylation reactions. Its role as a glycosyl acceptor is complicated by the symmetry of this molecule (see Section 2.1.2) which would make it difficult to identify the glycosidic linkages in the resulting transglycosylation products. In order to facilitate the identification of any transglycosylation products, a means of eliminating the symmetry of this molecule was needed. As a solution to this problem, 43 chirality was introduced by the specific introduction of deuterium at C-l in N-benzyloxycarbonyl-l,l-dideutero-l-deoxyxylonojrimycin. The synthesis of a selectively deuterated sample of 2.8 is not readily accomplished using the original route to 1.35 presented by Godskesen et al(107). Their synthesis utilised BH3 .Me2S in the reduction of TMS protected lactam (2.24). Here, this approach could not be adopted as BD3.Me2S is not easily accessible. Reduction of lactam 2.26 by lithium aluminum deuteride was proposed as the most convenient route to 1,1-dideutero-1-deoxyxylonojirimycin as benzyl groups are stable to hydride reductions and bestow solubility in relatively nonpolar aprotic solvents. Treatment of lactam 2.26 with lithium aluminum hydride or lithium aluminum deuteride in anhydrous ether and THF provided 2.13 and 2.55 as in 95% and 93% yield respectively. The benzyl ether groups were removed by catalytic hydrogenolysis with hydrogen catalysed by palladium on carbon in acetic acid and the benzyloxycarbonyl substitu'ent was introduced as described in section 2.1.1.2. BnO BnO o LiAlH 4 or LiAlD 4 Et 20, THF • R H 2 , Pd/C, •NH R R 2.13 R=H, 95% 2.55 R=D, 93% AcOH, 80% HO I R 2.26 1.35 R=H 2.7 R=D O CBzCl, Et 3N HO | R 2.8 R=H, 57% 2.9 R=D, 64% Figure 2.19 Introduction of deuterium at C-1 affording benzyl carbamates 2.8 and 2.9. 44 2.1.1.6 Preparation ofD-Xylonojirmycin tetrazole (2.1) Previous syntheses of D-glycononojirimycin tetrazoles have involved an intramolecular 1,3-dipolar cycloaddition as shown below. The reaction efficiently provides the desired product but synthesis of the precursor requires many steps and affords low overall yields(&5, 88). Vasella and Vonhoff have recently shown that perbenzylated glyconolactams are convenient starting materials for the synthesis of glycononojirimycin tetrazoles(779). Protected D-xylononojirmycin tetrazole 2.58 was formed in 54% yield upon treatment of D-xylonolactam 2.26 with sodium azide and trifluoromethanesulfonic anhydride. Deprotection of 2.58 by catalytic hydrogenolysis with H 2 and palladium on carbon, followed by crystallisation provided the tetrazole (2.1) in 69% yield. A) BnO 'OBn NaN 3, DMSO • 110°C, 85% OBn N=N^ BnO v OBn OBn OBn 2.56 2.57 B) OBn 2.26 NaN 3, Tf 20 * CH 3 CN, 54% OBn 2.58 H 2 , Pd/C AcOH, 69% 2.1 Figure 2.20 A) Preparation of 2,3,4,6-tetra-O-benzyl-D-mannonojirimycin tetrazole (2.57) via intramolecular 1,3-dipoar cycloaddition. B) Preparation of D-xylonojirimycin tetrazole from 2,3,4-tri-C7-benzyl-D-xylonolactam. 45 2.7.2 Xylo-configured Imino Sugar Derivatives as Glycosyl Acceptors in Enzymatic Transglycosylation Reactions. Glycosidases have proven to be useful catalysts in oligosaccharide synthesis(92). Studies using Abg p-glucosidase and a p-glucanase from Bacillus licheniformis have also shown that replacement of the catalytic nucleophile in a retaining P-glycosidase may provide a mutant enzyme capable of performing transglycosylation reactions(97, 720). One goal of this work was to attempt the chemo-enzymatic synthesis of oligosaccharide glycosidase inhibitors containing a xylo-configured iminosugar residue by use of two different enzymes, a mutant of T. saccharolyticum p-xylosidase in which the catalytic nucleophile has been replaced (Glu277Ala), and the AbgGlu358Ala glycosynthase mutant of Abg P-glucosidase. TV-Benzyloxycarbonyl-1-deoxyxylonojirimycin (2.8) was chosen as a potential glycosyl acceptor to be used in reactions catalysed by the Glu277Ala mutant of T. saccharolyticum P-xylosidase or by AbgGlu358Ala. This compound was chosen for two main reasons. Firstly, previous reports of glycosidases as catalysts for oligosaccharide synthesis reveal that compounds with pendant aryl groups near the "reducing end" are often useful as glycosyl acceptors in enzymatic transglycosylation reactions(95, 97). /?-Nitrophenyl p-D-xylopyranoside (ONPX), for example, acted as both a glycosyl donor and acceptor in transglycosylation reactions with wild-type T. saccharolyticum P-xylosidase (see Section 1.6). Nonspecific, noncovalent interactions of the aryl ring with the protein are presumably responsible for an increased affinity of this substrate for the +1 subsite, and the benzyl carbamate substituent was chosen with the hope that it would enhance binding at the +1 subsite through the interactions mentioned above. Secondly, carbamoylation eliminates the 46 basicity of the endocyclic nitrogen which was expected to greatly decrease its inhibitory potency by disrupting specific binding at the -1 subsite of the enzyme active site. In summary, N-benzyloxycarbonyl-l-deoxyxylonojrimycin (2.8) was hoped to behave poorly as a competitive inhibitor while retaining the potential to act as a glycosyl acceptor. Since the carbamate 2.8 is a symmetrical molecule its use as a glycosyl acceptor in transglycosylation experiments presents difficulties in the identification of the oligosaccharide products. As shown in Figure 2.21, substitution at C-2 or C-4 of 2.8 can not be easily distinguished by conventional analytical techniques. A means was needed to destroy the symmetry of the acceptor molecule without significantly altering its properties. The specific introduction of deuterium at C - l was seen as a means to desymmetrise 2.8 without altering its reactivity. 7V-Benzyloxycarbonyl-l,l-dideutero-1-deoxyxylonojirimycin (2.9) was designed as a glycosyl acceptor so that the transglycosylation products could be more easily identified. Figure 2.21 ZV-Benzyloxycarbonyl-1 -deoxyxylonojirimycin (2.8) possesses a plane of symmetry. Potential transglycosylation products 2.59 and 2.60 are difficult to distinguish. 1,1 -Dideutero-TV-benzyloxycarbonyl-1 -deoxyxylonojirimycin (2.9) is asymmetric. 47 2.1.2.1 Transglycosylation Experiments with AbgGlu358Ala Experiments have shown that the AbgGlu358Ala glycosynthase has somewhat relaxed specificity and is thus able to use a variety of different compounds as glycosyl acceptors including D-xylo-configured monosaccharides(97). D-Xylo-configured iminosugars, D-xylonolactam (1.34), 1-deoxyxylonojirimycin 1.35, N-benzyl-l-deoxyxylonojirimycin 2.3, and carbamate derivative 2.8, were screened as potential glycosyl acceptors for this mutated enzyme using a-glucosyl fluoride as the glycosyl donor as depicted in Figure 2.22 for 2.8. 2.8 AbgGlu358Ala ((j-glucosidase Transglycosylation Products Figure 2.22 Enzymatic transglycosylation mediated by AbgGlu358Ala (3-glucosidase using a-glucosyl fluoride as a glycosyl donor and AT-benzyloxycarbonyl-1-deoxyxylonojirimycin (2.8) as a glycosyl acceptor. While D-xylonolactam (1.34), 1-deoxyxylonojirimycin (1.35) and N-benzyl-l-deoxyxylonojirimycin (2.3) proved incompetent as glycosyl acceptors, a number of transglycosylation products were observed when 2.8 was combined with a-glucosyl fluoride in the presence of AbgGlu358Ala. TLC analysis indicated several new products were which absorbed UV-light, confirming that 2.8 was indeed a substrate in the transglycosylation reactions. The crude transglycosylation mixture was then treated with acetic anhydride and 48 pyridine. The mixture, after workup was then resolved by column chromatography. To aid in the identification of the oligosaccharide products, the reaction was repeated on a larger scale, using a higher enzyme concentration, and deuterated benzyl carbamate 2.9 as a glycosyl acceptor. The identities of products from both reactions are listed in Table 2.1. Acceptor Glycosidic Chain Structure Yield (%)t linkage Length 2.8 P-1,3 2 2.62 10.0 P-1,3 3 2.63 -P-1,4 2 2.66* 5.0 (3-1,4 3 2.67* 3.2 P-1,4 4 2.68 -18.2 2.9 P-1,3 2 2.64 18.0 P-1,3 3 2.65 4.4 P-1,4 2 2.69 -P-1,4 3 2.70* 14.8 P-1,4 4 2.71* 13.2 50.4 Table 2.1 Summary of transglycosylation products from 2.8 and 2.9 with AbgGlu358Ala. t yields based on amount of glycosyl donor used (* tentative assignments). R' R 2.62 H H 2.63 Glu H 2.64 H D 2.65 Glu D n R 2.66 2.67 2.68 2.69 2.70 2.71 H H H D D D Figure 2.23 Transglycosylation products obtained by reaction of 2.8 and a-glucosyl fluoride mediated by AbgGlu358Ala. Mass spectral analysis of the transglycosylation products was consistent with 49 oligosaccharide structures containing between one and three glucose residues and a single iminosugar residue. The major product in both cases were the (3-1,3-linked disaccharides 2.62, and 2.64 in 10 and 18% yields respectively. These structures were confirmed by mass spectral analysis and 'H-NMR spectroscopy (Table 2.2). The 'H-NMR spectrum of the nondeuterated disaccharide displayed symmetry in the iminosugar ring which could only arise if the glycosidic linkage is located at C-3. As additional evidence, the intensity of the signals assigned to the protons on C-5 and C-l of the nondeuterated sample are reduced by half in the corresponding deuterated sample. Compound Assignment 8 Integral 2.62 H-la, H-5a H-lb, H-5b H-2, H-4 H-3 4.31 2H 2.75 2 H 4.79 2 H 4.93 1 H 2.64 H-2, H-4 H-3 H-5a H-5b 4.79 2 H 4.93 1 H 4.31 1 H 2.75 1 H Table 2.2 1 H-NMR data from (3-1,3-linked disaccharides 2.62 and 2.64. In the first transglycosylation reaction using the nondeuterated carbamate 2.8 and a lower concentration of enzyme a second disaccharide product (2.66) was obtained as a minor product. A trisaccharide product was also observed. The linkage in the second disaccharide and trisaccharide samples could not be determined since they did not contain deuterium labels at C-l however. It is probable that these products contain (3-1,4-linkages however, and the structure shown for 2.66 and 2.67 have been tentatively assigned. When the experiments were repeated with the deuterated carbamate 2.9 using a higher enzyme concentration, a much higher yield was observed, but the (3-1,3-linked disaccharide 50 2.64 was the only disaccharide isolated. Two trisaccharide products and one tetrasaccharide were also isolated in moderate yields however. As a result of the higher enzyme concentration, the disaccharide products were elongated much more quickly than in the earlier experiment. One of the trisaccharide products was identified as (3-1,3-linked 2.65 by comparison of the chemical shift of its H-5 protons with the spectrum of (3-1,3-linked disaccharide 2.64. This compound was isolated in only 4.4% yield however, while the other trisaccharide was obtained in 14.8% yield. The 'H-NMR spectra of the major trisaccharide (14.8%) and tetrasaccharide (13.2%) products show that these two compounds were not (3-1,3-linked at the iminosugar residue however, as evidenced by the significant differences between the 'H-NMR spectra of these samples with spectra of 2.62 and 2.64. Despite the introduction of deuterium labels at C-l, the 'H-NMR spectra of these compounds could not be assigned due to overlapping resonances. Comparison of the spectra of the deuterated and nondeuterated samples was not helpful since the only differences fell within a complex region of the spectra. Because wild-type Abg P-glucosidase normally cleaves P-l,4-linked substrates, and p-l,2-Linked products have never been reported in transglycosylation reactions using either wild-type Abg P-glucosidase or the Glu358Ala glycosynthase, it is likely that the observed tri- and tetrasaccharide products contain P-l,4-linkages to the iminosugar rather than P-l,2-linkages. The structures 2.68, 2.70, 2.71 have been cautiously assigned to the major oligosaccharide products. It is interesting to note that the major transglycosylation product is a P-l,3-linked disaccharide but the major tri- and tetrasaccharide products do not contain a P-l,3-linkage. This differs from previous work using />nitrophenyl p-D-xyloside (PNPX) as the glycosyl 51 acceptor with AbgGlu358Ala where all of the observed products were (3-l,3-linked(97). These results suggest that although AbgGlu358Ala forms |3-l,3-linkages to 2.8 and 2.9 more quickly than (3-1,4-linkages, the resulting P-l,3-linked product is a poor glycosyl acceptor and is extended very slowly. The P-l,4-linked disaccharide is apparently formed more slowly, but once formed it is a good glycosyl acceptor by the enzyme, leading completely to oligosaccharide products. 2.1.2.2 Transglycosylation Experiments with T. saccharolyticum Glu277Ala fi-xylosidase l-Deoxyxylonojirimycin (1.36) is a compound of particular interest as a potential inhibitor of xylanases, and a chemo-enzymatic route to this compound was sought as an alternative to chemical synthesis. A 15N-labelled sample of 1.36 could be especially useful for 1 5 N-NMR studies with Bacillus circulans xylanase. Since T. saccharolyticum P-xylosidase has inherent transglycosylation activity(iOi), it was hoped that a mutant in which the catalytic nucleophile (Glu277) is removed would provide a "glycosynthase" with specificity for D-xylo-configured substrates. A route to 1-deoxyxylobionojirimycin (1.36) was therefore proposed via transglycosylation onto the carbamate derivative (2.8) catalysed by the Glu277Ala mutant of T. saccharolyticum P-xylosidase using a-xylosyl fluoride as a glycosyl donor. Unfortunately, the mutant enzyme did not display any significant transglycosylation activity thus its use as a glycosynthase was not possible. 2.1.3 Preparation of Aryl Xylobiosides The Bacillus circulans xylanase (BCX) will only cleave oligomers of D-xylose containing three or more sugar residues. This strict substrate requirement presents significant 52 hurdles to the study of xylanases since xylan, the natural substrate, does not have a fixed composition, and other 'natural' substrates of known composition are not readily available. Activated aryl glycosides are convenient substrates for kinetic studies of many glycosidases since their hydrolysis can be monitored spectrophotometrically. While aryl (3-D-xylosides are not hydrolysed by xylanases, BCX does cleave activated aryl xylobiosides such as p-nitrophenyl p-xylobiosides, and compounds of this type are consequently useful xylanase substrates. Unfortunately, these disaccharide substrates are not readily available due to the difficulty in obtaining xylobiose, the logical precursor for their synthesis. Before this work was begun, two approaches to aryl xylobiosides had been presented. The first approach involved chemical synthesis from monosaccharide precursors(i05). This route involves a large number of synthetic steps and provides aryl xylobiosides in low yields. The second approach utilised xylobiose as the starting material. This disaccharide has been isolated on a small scale by the acidic digestion of oat spelt xylan(i2i-i24), but isolation of unprotected xylobiose from the resulting mixture is difficult. Xylobiose prepared in this way was utilitised in the synthesis of 2,5-dinitrophenyl P-D-xylobioside(70<5). Since xylanases such as Trichoderma viride xylanase and BCX digest xylan affording xylobiose as the major product, the use of a xylanase as a catalyst for the direct preparation of xylobiose was investigated. The soluble fraction of birchwood xylan, dissolved in sodium acetate buffer, was digested with T. viride xylanase. In order to avoid the rigorous purification steps required to separate xylo-oligosaccharides, the resulting crude mixture containing xylotriose, xylobiose, and D-xylose was acetylated and hexa-0-acetyl xylobiose (2.72) in pure form was isolated by column chromatography on silica gel. This procedure allowed for the isolation of several grams of protected xylobiose in one batch. Penta-O-53 acetyl-a-xylobiosyl bromide (2.73) was prepared in high yield by treatment of the hexa-O-acetate (2.72) with hydrobromic acid in acetic acid. The bromide was readily converted to acetylated 2,5-dinitrophenyl (3-xylobioside and acetylated 2-nitrophenyl P-xylobioside in good yields using the Koenigs-Knorr reaction(706) and the free sugars 2.74 and 2.75 were obtained by treatment of the acetylated glycosides with ammonia in anhydrous methanol. The aryl xylobiosides were thus obtained in 41 and 48% overall yields from the hexa-O-acetate (2.72) in three steps. While this work was in progress, Mechaly et al. used xylanase T-6 to degrade xylan to xylobiose and employed a very similar route to 4-nitrophenyl (3-xylobioside(725). This method allows for convenient access to chromophoric xylanase substrates in relatively large quantities. Birchwood Xylan •OAc HBr, AcOH CH2C12 2.72 AcO' ROH, 2,6-lutidine OR Br 2.73 2.74 R = Ac, R'=2,5DNP 2.76 R = H, R'=2,5DNP 2.77 R = H, R'=ONP Figure 2.24 Chemo-enzymatic synthesis of aryl xylobiosides. 54 2.1.4 Conclusions A series of interesting D-xylo-configured iminosugars were prepared as potential (3-xylosidase inhibitors. An improved synthesis of D-xylonolactam and 1-deoxyxylonojirimycin was developed which also provided access to 1,1-dideutero-l-deoxyxylonojirimycin and 1,1-dideutero-N-benzyloxycarbonyl-l-deoxyxylonojirimycin. An intermediate in the synthetic route, 2,3,4-tri-O-benzyl-D-xylonolactam (2.26), is a useful starting material which may be converted to a number of other interesting species. This intermediate was converted to D-xylonojirimycin tetrazole, a new inhibitor of xylosidases. N-acyl, N-benzyl, and N-benzyloxycarbonyl derivatives of 1-deoxyxylonojirimycin were also prepared in good yields. The synthetic strategies employed here may also be used in the synthesis of disaccharide iminosugars. N-Benzyloxycarbonyl- 1-deoxyxylonojirimycin was used as a glycosyl acceptor in the enzymatic synthesis of di-, tri- and tetrasaccharide products using the Abg Glu358Ala glycosynthase. The major reaction product was the (3-1,3-linked disaccharide (2.62) which proved to be a poor glycosyl acceptor and was not extended. The enzyme formed (3-1,4-linkages to the iminosugar more slowly, but the product was a superior glycosyl acceptor, and was quickly converted to tri- and tetrasaccharide products. This work shows that iminosugars can be used in transglycosylation reactions with the AbgGlu358Ala glycosynthase to make oligosaccharide imino sugars. An improved chemo-enzymatic strategy was developed for the convenient synthesis of aryl xylobiosides. Hexa-0-acetyl xylobiose, prepared through the degradation of xylan by T. viride xylanase followed by acetylation and straightforward column chromatography was used in the synthesis of 2-nitrophenyl (3-D-xylobioside and 2,5-dinitrophenyl (3-D-55 xylobioside. This route provided aryl xylobiosides in greater yield and in fewer steps than the previously published syntheses(7(25, 106). 2.2 Enzyme Kinetics T. saccharolyticum P-xylosidase has not yet been extensively studied. This enzyme has been shown to be retaining with transglycosylation activity(707). The catalytic nucleophile(7(93) and acid/base catalyst have been assigned(704), but thus far, no competitive inhibitors have been tested with this enzyme. Since 1-deoxynojirimycin and its analogues are powerful inhibitors of many glycosidases, 1-deoxyxylonojirimycin was tested as an inhibitor of T. saccharolyticum (3-xylosidase. The pH dependence of T. saccharolyticum p-xylosidase inhibition by 1-deoxyxylonojirimycin was explored to gain insights into the binding of 1-deoxynojirimycin type inhibitors. The inhibition constants of a number of 1-deoxyxylonojirimycin derivatives with various nitrogen substituents were also measured to investigate the influence of substitution at the ring nitrogen, the importance of basicity at the ring nitrogen to inhibitor binding, and the effect of pendant aromatic groups. D-xylonolactam and D-xylonojirimycin tetrazole, compounds which exhibit sp2 hybridisation at the anomeric centre were tested as inhibitors of T. saccharolyticum P-xylosidase as well. A number of these compounds were also tested as inhibitors of the Glu 160Ala mutant of T. saccharolyticum P-xylosidase in which the acid/base catalyst has been replaced by alanine to investigate the influence of this residue on inhibitor binding. 2.2.1 The Significance of pH dependence in Enzyme kinetics Any enzyme is active only within a limited range of pH values as a result of changes in the ionisation state of the enzyme itself, the substrate(s), and cofactor(s) involved in the 56 reaction as the pH is altered(72c5). Usually only one ionic form of an enzyme active site is catalytically active, and one may treat such as enzyme as a dibasic acid. In order for the enzyme to be active, one residue must be protonated and the other deprotonated(72r5, 127). When this model is applied to an enzyme system, measurements of kinetic parameters can provide information about the important ionisable groups in the enzyme. Measuring kcat /K m or K m over an appropriate pH range can provide pK a values for the free enzyme while measurements of k c a t can furnish pK a values for the enzyme-substrate complex of the rate determining step(126). Measurements of K; as a function of pH can provide pK a values for both the enzyme and inhibitor. The importance of pH in enzyme kinetics has been described in detail by Dixon and Webb(72(5). In order for the model briefly described above to be valid for a specific enzyme system, a number of conditions must be met(128). First of all, the observed changes in enzyme activity must truly be changes in activity and not simply the result of enzyme instability. Secondly, as has been already intimated, there must be only one ionisation state of the enzyme that is catalytically active. Otherwise, there may be multiple parallel or diversionary pathways which the enzyme-catalysed reaction may follow. This model also requires that there be no change in the rate determining step with pH, and that all protonation and deprotonation steps are fast relative to substrate-product conversion(72S). Since the inhibition constant (Kj) can be considered a true dissociation constant, it may seem that the interpretation of its dependence on pH is not subject to such difficulties. It is important to keep in mind however, that Kj is equal to [E][I]/[EI] and is influenced by the pH functions for all three of these factors(728). It is also important that the pK a values of inhibitor and the free enzyme be sufficiently separated so that their influences may be 57 distinguished(59). 2.2.2 pH Dependence of Wild-type and Glul60Ala T. saccharolyticum {5-xylosidase Activity The pH dependence of wild-type T. saccharolyticum P-xylosidase activity using p-nitrophenyl P-D-xylopyranoside as a substrate and the corresponding Glu 160Ala mutant using 3,4-dinitrophenyl p-D-xylopyranoside (3,4-DNPX) as a substrate are reported here courtesy of J. Wicki(729). Values of k c a t /K m for the wild-type p-xylosidase are seen to be dependent on two ionisations (Figure 2.25). The activity of the native enzyme is defined by a pK a of 4.1 for the acidic limb and a pK a of 6.8 in the basic limb which presumably represent the pK a values of two important ionisable groups in the enzyme active site (126). Our understanding of the double-displacement mechanism leads to a tentative assignment of the observed pK a values of 4.1 and 6.8 in the wild-type P-xylosidase to the catalytic nucleophile (Glu277) and acid/base catalyst (Glu 160) respectively. For the enzyme to be active, the acid/base catalyst must be protonated and the catalytic nucleophile unprotonated. The Glul60Ala T. saccharolyticum P-xylosidase mutant has significant catalytic activity against aryl xylosides(i(34). The pH dependence of the Glul60Ala mutant also presents as a bell shaped curve (Figure 2.25). The curve is defined by pK a values of 6.5 in the acidic limb and 9.0 for the basic limb. The shift in pH dependence of the Glul60Ala mutant relative to the wild-type P-xylosidase is not fully understood, but one explanation may involve a third residue, presumably a carboxyl, in the active site which assumes the role of acid/base catalyst in the mutant enzyme. When Glu 160 is removed, the pK a of the catalytic nucleophile would be expected to shift due to the change in its environment, and the pK a of the other carboxyl residue would be reflected in the basic limb of the pH dependence curve. 58 Figure 2.25 kCat/Km as a function of pH for wild-type T. Saccharolyticum P-Xylosidase using p-nitrophenyl (3-D-xylopyranoside as substrateO) and Glul60Ala T. Saccharolyticum (3-Xylosidase using 3,4-dinitrophenyl P-D-xylopyranoside as substrate(O). 2.2.3 Inhibition of Wild-type T. saccharolyticum j3-xylosidase by l-Deoxyxylonojirimycin Inhibition of wild-type T. saccharolyticum P-xylosidase by 1-deoxyxylonojirimycin was measured at pH 5.5 revealing a Kj of 118 pM. Presentation of the experimental results as a double-reciprocal plot in Figure 2.26 reveals that the lines for each concentration of enzyme inhibitor intersect near the y-intercept, as expected for a competitive inhibitor. 59 [I] = 30.2 uM [I]=45.3 uM [I] = 75.5 uM [TJ = 151 uM rrj = 226 uM [I] = 302 uM Figure 2.26 Kj determination of T. saccharolyticum P-Xylosidase inhibition by l-Deoxyxylonojirimycin. 2.2.3.1 pH Dependence of T. saccharolyticum (3-xylosidase Inhibition by 1-Deoxyxylonojirimycin. As mentioned earlier, 1-deoxynojirimycins may bind to glycosidases in two different ways: the neutral inhibitor may enter the active site where it interacts with a protonated residue, or the inhibitor may be protonated before it enters the active site and then interact with an anionic residue. In theory, the mode of binding can be dissected by careful studies of the pH dependence of substrate and inhibitor binding(59). The pH dependence of wild-type T. saccharolyticum P-xylosidase inhibition by 1-deoxyxylonojirimycin was investigated to gain insight into the mode of inhibitor binding. An accurate value for the pK a of 1-deoxyxylonojirimycin was required in order to dissect the effect of pH on inhibitor binding. The pK a of 15N-l-deoxyxylonojirimcyn was determined using 1 5 N-NMR by observing the chemical shift of the endocyclic nitrogen over a 60 range of pH values with the assistance of Mr. Manish Joshi, University of British Columbia, Department of Biochemistry. The resulting pH titration curve from which a pK a of 7.30 was determined is shown below (Figure 2.27). 36 4 6 8 10 PH Figure 2.27 pK a determination of 15N-1-deoxyxylonojirimycin by 1 5 N-NMR. pK a = 7.30. Inhibition constants for 1-deoxyxylonojirimycin with wild-type T. saccharolyticum (3-xylosidase at a series of pH values are listed in Table 2.3. Inhibition improved with increasing pH from pH 4.5 to 6.5 and remained essentially constant from pH 6.5 to 7.5. Plotting pK; as a function of pH gave a plot with a linear region from pH 4.5 - 6.0 with a slope of 1.1, and a flat portion from pH 6.5 to 7.5 (Figure 2.28). The linear and horizontal portions of the graph intersect at pH 6.3 indicating a pK a of 6.3. Since 1-deoxyxylonojirimycin has a pK a of 7.30, these results show that the pH dependence of enzyme inhibition by 1-deoxyxylonojirimycin is not simply due to the ionisation state of the inhibitor. The observed pK a therefore must also reflect a change in ionisation state of the enzyme. This would imply that the residue in the enzyme active site of the enzyme which interacts with 1-deoxyxylonojirimycin must be deprotonated for optimal interaction. It is 61 likely that 1-deoxyxylonojirimycin is interacting with the acid/base catalyst (Glu 160) of T. saccharolyticum P-xylosidase. The observed pK a value of 6.3 may be tentatively assigned to the acid/base catalyst, but earlier, a pK a of 6.8 was assigned to this residue according to the pH dependence of k c a t / K m (129). The apparent pK a of 6.3 must therefore be regarded with caution. PH Ki (pM) 4.5 1500 ± 100 5.0 400 ± 30 5.5 118 ± 10 6.0 34 ± 3 6.5 14 ± 1 7.0 14 ± 3 7.5 17 ± 1 Table 2.3 pH Dependence of T. Saccharolyticum P-xylosidase inhibition by 1-deoxyxylonojirimycin. These observations are consistent with binding of the inhibitor in its cationic form through interactions with a deprotonated, negatively charged group in the enzyme active site, presumably the acid/base catalyst. If the inhibitor were binding in its neutral form to a protonated carboxyl in the active site instead, inhibition would not remain constant over any range of pH values. While the pK a observed in the pH dependence study does not match perfectly with the pK a from pH activity studies, these results are nevertheless consistent with the 1-deoxyxylonojirimycin binding to T. saccharolyticum P-xylosidase in its cationic form. The pH dependence of T. saccharolyticum P-xylosidase by 1-deoxyxylonojirimycin also shows that this inhibitor is not a true mimic of the transition state for this enzyme. The binding of a transition state analogue must show the same dependence on pH as the activity of the enzyme. This is not the case for 1-deoxyxylonojirimycin. 62 -• n u l l 11111 I I11II1 • • l l l l l l l 4 4.5 5 5.5 6 6.5 7 7.5 8 pH Figure 2.28 pKj as a function of pH for inhibition of T. saccharolyticum P-xylosidase by 1-deoxy xy lonoj irimycin. 2.2.4 The Influence of Inhibitor Structure on Inhibition ofT. saccharolyticum fi-Xylosidase 2.2.4.1 Inhibition ofT. saccharolyticum $-Xylosidase by Imino Sugars with Substituents on the Endocyclic Nitrogen N-Benzyloxycarbonyl- 1-deoxyxylonojirimycin (2.8) was initially designed as a potential glycosyl acceptor for transglycosylation reactions mediated by the nucleophile mutant of T. saccharolyticum P-xylosidase. The benzyloxycarbonyl moiety was employed in an attempt to weaken the binding of 2.8 as a competitive inhibitor without affecting its ability to act as a glycosyl acceptor. Inhibition kinetics on wild-type T. saccharolyticum P-xylosidase reveal that 2.8 is bound as tightly as 1-deoxyxylonojirimycin at pH 5.5 with a Kj of 110 pM and only 4.5 fold more weakly at pH 6.5. The carbamate derivative (2.8) was also bound tightly by Glul60Ala T. saccharolyticum P-xylosidase (Table 2.4). The unexpected inhibitory potency of 2.8 prompted an exploration of the important features in the binding of 63 deoxynojirimycin type inhibitors to this enzyme. Iminosugars 2.3, 2.4, and 2.5 were prepared, not as potential glycosyl acceptors, but to probe the importance of removing the basicity of the endocyclic nitrogen, and the influence of an aromatic substituent on the endocyclic nitrogen. 2.3 R = CH 2Ph 2.78 2.4 R = COCH 3 2.5 R = CHO 2.8 R = COOCH2Ph Figure 2.29 TV-substituted iminosugars tested as inhibitors of T. saccharolyticum P-xylosidase. Studies of 1-deoxyxylonojirimycin derivatives with TV-benzyl (2.3), TV-acetyl (2.4) and TV-formyl (2.5) substituents as inhibitors of T. saccharolyticum P-xylosidase are summarised in Table 2.4. 2.3, 2.4 and 2.5 showed weak inhibition of T. saccharolyticum P-xylosidase, much weaker than 1-deoxyxylonojirimycin (1.35) or its carbamate derivative (2.8). Since the acyl and formyl groups are substantially smaller than the benzyloxycarbonyl group, the decreased inhibition can not be due to steric effects. Rather, it seems the TV-acetyl (2.4) and TV-formyl (2.5) derivatives of 1-deoxyxylonojirimycin are poor inhibitors of T. saccharolyticum P-xylosidase because the endocyclic nitrogen in these compounds is unable to accept a proton or form a hydrogen bond. Curiously, the TV-benzyl derivative (2.3) was among the poorest inhibitors of T. saccharolyticum P-xylosidase in this group, even though the nitrogen is still able to accept a proton. The benzyl group in 2.3 does not extend as far from the sugar ring as the benzyl carbamate in 2.8 however, and it is presumably unable to interact with the enzyme at the same site. Poor inhibition by benzyl derivative 2.3 is most 64 likely the result of unfavourable steric interactions. The position of the benzyl carbamate group in 2.8 is clearly very important, as a xyloisofagomine derivative with a benzyl carbamate substituent (2.78) is a very weak inhibitor of the enzyme despite presenting the same pendant group to the enzyme(770). Although the carbamate substituent on the endocyclic nitrogen in 2.8 is believed to disrupt interactions in the +1 subsite, this compound is still able to bind relatively tightly to T. saccharolyticum (3-xylosidase presumably through additional interactions between the benzyloxycarbonyl group and the enzyme. Because the basicity of the endocyclic nitrogen is removed, Af-benzyloxycarbonyl- 1-deoxyxylonojirimycin 2.8 is not believed to exploit any specific interactions at the active site, and its inhibition of T. saccharolyticum P-xylosidase is the result of fortuitous nonspecific interactions between the benzyl carbamate group with a hydrophobic site on the enzyme. These results show that the basicity of the endocyclic nitrogen is very important to iminosugar binding to T. saccharolyticum P-xylosidase. Aromatic substituents are only able to enhance binding if they are perfectly positioned so as to allow interactions with a hydrophobic site in the enzyme. 65 Inhibitor PH Wild-type Ki(uM) Glul60Ala Ki (uM) H O ^ ' N H 1.35 O H 5.5 6.5 7.7 118 + 6 13.5 ± 1 210 ± 10 30 ± 2 68 ± 7 o H < H ^ ^ 2.8 H O \ ^ 5.5 6.5 7.7 110 ± 3 0 70 ± 7 130 ± 5 65 ± 1 40 ± 2 H 0 A ^ \ [ I I 2.3 O H 5.5 6.5 7200 ± 700 3500 ± 300 O H O ^ ^ N x f ^ H O — V - ^ - A 2.4 O H 5.5 5700 ± 100 O H O ^ - \ H Q _ V - ^ A 2.5 O H 5.5 3100 ± 4 0 0 2.78 6.5 >1000* >1000* H O ^ C - N « H O - X ^ ^ ^ O 1.34 O H 5.5 6.5 3000 ± 300 2300 ± 200 170 ± 5 H O - X — « ^ > — N 2.1 O H 5.5 6.5 7.7 102 ± 6 135 ± 7 23 ± 1 25 ± 5 64 ± 6 2.80 H < ^ ~ ^ * * ^ N H 6.5 1.3* 200* Table 2.4 Inhibition of wild-type and Glu 160Ala T. saccharolyticum (3-xylosidases by various imino sugars. (*Datafrom (7/0)) 66 2.2.4.2 Inhibition ofT. saccharolyticum (3-xylosidase by compounds with a sp2 hybridised anomeric centre. K; values for the inhibition of wild-type T. saccharolyticum P-xylosidase by D-xylonolactam (1.34) and D-xylonojirimycin tetrazole (2.1) are listed in Table 2.4. The lactam is a very weak inhibitor of the wild-type P-xylosidase (K; = 3 mM), indicating that while compounds with sp2 hybridisation at C-l may be structurally and to an oxocarbenium ion, this provides no guarantee of strong inhibition. Tetrazole 2.1 was bound much more tightly by T. saccharolyticum P-xylosidase with a Kj of 102 pM at pH 5.5. Glyconojirimycin tetrazoles are believed to bind through hydrogen bonding of the exocyclic nitrogen closest to C-l with the acid/base catalyst(6>7, 88). Inhibition by tetrazoles suggests that the acid/base catalyst is positioned for 'side on' protonation of the substrate (Figure 1.18) and inhibition by D-xylonojirimycin tetrazole in the micromolar range suggests that the acid/base catalyst in T. saccharolyticum P-xylosidase is positioned to protonate the substrate from the side. The tetrazole (2.1) inhibits the acid/base mutant more than the wild-type however, and these results must therefore be interpreted with caution. Further inhibition studies with triazole (2.79) could provide additional information. 2.79 Figure 2.30 D-Xylonojirimycin triazole (2.79) 67 2.2.4.3 The Influence of the Position of the Endocyclic Nitrogen on Imino Sugar Inhibition In general, 1-aza sugars such as isofagomine (1.18) tend to bind more tightly to retaining (3-glycosidases than do the l-deoxynojirimycins(37). In agreement with this trend, xyloisofagomine was recently found to inhibit wild-type T. saccharolyticum (3-xylosidase very tightly with a Kj of 1.3 pM at pH 5.5(110), approximately 100 times more strongly than 1-deoxyxylonojirimycin at the same pH, even though the lack of a hydroxyl group at C-2 would be expected to be detrimental to binding. Although studies of the pH dependence of 2.80 would allow for a more meaningful comparison between these two inhibitors, it is clear that orientation and position of the nitrogen centre have a major influence on binding, and the nitrogen in 2.80 is better oriented to interact at the active site of this enzyme. Figure 2.31 D-Xyloisofagomine. 2.2.5 Inhibition Studies of Glul60Ala T. saccharolyticum [5-xylosidase The observed inhibition of Glul60Ala T. saccharolyticum P-xylosidase mutant by several iminosugars is summarised in Table 2.4. l-Deoxyxylonojirimycin is believed to bind in large part through interactions with the acid/base catalyst yet it inhibits the Glul60Ala mutant almost as strongly as the wild-type P-xylosidase with a Kj of 30 p M at pH 6.5. D-Xylonojirimycin tetrazole binds even more tightly to the mutant enzyme than to the wild-type and D-xylonolactam (1.34) binds ten times more tightly to the mutant enzyme. Since inhibition by 2.8 is not specifically dependent on interactions with the acid/base catalyst, it ofT. saccharolyticum P-xylosidase 2.80 68 was able to bind tightly to the mutant Glu 160Ala T. saccharolyticum P-xylosidase as well. These results are puzzling as these inhibitors are believed to bind in large part through interactions with the acid/base catalyst. 1-Deoxynojirimycins are believed to interact ionically with this residue(75) while binding of nojirimycin tetrazoles has been explained by the formation of a hydrogen bond with the acid/base catalyst(S9). Superficially, these results seem to suggest that Glu 160 is in fact not the acid/base catalyst, but this mutant enzyme displays a significant activity increase in the presence of azide, and the identity of the azide reaction product has been confirmed, providing strong evidence that Glu 160 is indeed the acid/base catalyst(704). One explanation for these observations may be the presence of a third carboxyl residue in the active site of T. saccharolyticum P-xylosidase in close proximity to Glu 160. When Glu 160 is replaced with an alanine, this carboxyl may assume an important role in binding and catalysis. This would explain the bell shaped pH dependence of the enzyme activity, and compounds which are expected to interact with the acid/base catalyst in the wild-type enzyme may instead interact with the additional carboxyl residue. 2.2.6 Conclusions 1-Deoxyxylonojirimycin is a competitive inhibitor of T. saccharolyticum P-xylosidase with a Kj of 118 uM at pH 5.5 and a Kj of 13.5 pM at pH 6.5. The pH-dependence of T. saccharolyticum P-xylosidase inhibition by 1-deoxyxylonojirimycin was measured from pH 4.5 to 7.5. The results suggest that 1-deoxyxylonojirimycin binds to this enzyme as a cation and interacts with a deprotonated, negatively charged residue within the active site. 1-Deoxyxylonojirimycin is not a transition state analogue for T. saccharolyticum P-xylosidase since the pH dependence of inhibitor binding does not correspond with the pH activity profile 69 of this enzyme. The effects of N-substitution on the inhibitory potency of 1-deoxynojirimycin type compounds were investigated by testing JV-benzyloxycarbonyl, JV-benzyl, JV-acetyl, and N-formyl derivatives of 1-deoxyxylonojirimycin as inhibitors of T. saccharolyticum p--xylosidase. 7V-Benzyloxycarbonyl- 1-deoxyxylonojirimycin (2.8) was a moderately strong inhibitor of both wild-type and Glul60Ala T. saccharolyticum P-xylosidase. The N-acyl derivatives bound poorly to the enzyme however, demonstrating the importance of basicity at the endocyclic nitrogen to binding. AM3enzyl-1-deoxyxylonojirimycin (2.3) and N-benzyloxycarbonyl xyloisofagomine (2.78) also bound poorly to this enzyme, an indication that the location of the phenyl ring in 2.8 is crucial. N-Benzyloxycarbonyl-1-deoxyxylonojirimycin is not believed to exploit any specific interactions at the active site, but rather its binding is believed to arise through fortuitous interactions between the aryl ring on the inhibitor with a hydrophobic site on the enzyme. Two inhibitors with sp2 hybridisation at the anomeric centre, D-xylonolactam (1.34) and D-xylonojirimycin tetrazole (2.1) were tested as inhibitors of T. saccharolyticum P-xylosidase. While 1.34 was a poor inhibitor, the tetrazole showed moderate inhibition of this enzyme. Inhibition studies on the Glul60Ala mutant of T. saccharolyticum P-xylosidase support the possibility of another residue in the active site assuming the role of acid/base catalyst in this enzyme. Binding of 1-deoxyxylonojirimycin decreased only slightly while binding of D-xylonojirimycin tetrazole and D-xylonolactam to the mutant enzyme was improved by the absence of Glul60, suggesting the possibility that these inhibitors are interacting with another residue in the mutant enzyme. Chapter Three: Materials and Methods 71 3.1 Synthesis 3.1.1 General ' i i nuclear magnetic resonance (NMR) spectra were recorded on the following instruments with frequencies for protons as indicated: a Briiker AC-200 at 200 MHz, a Varian XL-300 at 300 MHz, or a Briiker WH-400 at 400 MHz. 1 3 C - N M R spectroscopy was performed on a Briiker AC-200 at 50 MHz or on a Varian XL-300 at 75 MHz for 1 3 C . ! H samples dissolved in CDCI3, acetone-D6 or MeOD were internally referenced to the solvent and for samples dissolved in D2O, 2,2-dimethyl-2-silapentane-5-sulfonate (5 = 0.015 ppm) was used as an external reference. C spectra are proton decoupled and the chemical shifts are internally referenced to CDCI3 or MeOD or for samples dissolved in D 2 0 , Methanol was used as an external reference. 1 5 N-NMR spectra were acquired at 30°C using a Varian Unity Spectrometer operating at 500 MHz for protons with the assistance of Mannish Joshi, University of British Columbia, Department of Biochemistry. Micro-analyses were performed in the Microanalytical Lab, University of British Columbia, by Mr. Peter Borda. Low- and high-resolution desorption chemical ionisation mass spectrometry (DCI-LRMS and DCI-HRMS) were performed on a Delsi Nermag RIO-IOC single quadrupole mass spectrometer using ammonia as the reagent gas. Low and high-resolution secondary ion mass spectrometry (LSIMS-LRMS and LSfMS-HRMS) were performed on a Kratos-Concept II H mass spectrometer equipped with a cesium-ion gun using glycerol and methanol as the matrix. Melting points were obtained using a Mel Temp II melting point apparatus and are uncorrected. Thin layer chromatography (tic) was performed on Merck silica gel 60 F 2 5 4 72 aluminum backed plates and visualised under U V light and by charring with 10% sulfuric acid in methanol or 10% ammonium molybdate in 1 M sulfuric acid. Column chromatography was performed under elevated pressure on Kieselgel 60 (230-400 mesh) silica. All solvents and reagents used were of either reagent grade, certified or spectral grade. Dry solvents were prepared as follows: Acetonitrile, pyridine and dichloromethane were distilled from calcium hydride. Diethyl ether, tetrahydrofuran (THF), and 1,4-dioxane were distilled from sodium benzophenone ketyl. N,N'-dimethyl formamide (DMF) was held at reflux over calcium sulfate for several hours then distilled from 4A molecular sieves under reduced pressure. Unless otherwise stated, all other chemicals were purchased from Sigma Chemical Company and used without further purification. D-Xylonolactam (1.34) Lactam (1.34) was prepared from D-xylose exactly as described by Godskesen et al(707). mp: 170-171°C. Lit: 169-171°C. 'H-NMR (200 MHz, D 20): 5 4.0 (d, 1 H, J 2 , 3 9.25 Hz, H-2), 3.93 (m, 1 H, H-4), 3.68 (t, 1 H, J 3 , 4 9.1 Hz, H-3), 3.48 (dd, 1 H, J 4 , 5 b 5.7 Hz, J 5 a , 5 b 12.4 Hz, H-5b), 3.11 (dd, 1 H, J 4 , 5 a 8.9 Hz, H-5a). Anal. Calc for C5H9NO4: C, 40.82; H, 6.17; N, 9.52; 0,43.50. Found: C, 40.85; H, 6.22; N, 9.47. 15N-D-Xylonolactam (2.12) Methanesulfonate 2.10 (1.0 g, 3.35 mmol) was combined with 1 5NH 4C1 (1.0 g, 18.3 mmol) and NaOH (700 mg, 17.5 mmol) in deionised water (5 ml) in a sealed flask. The reaction was allowed to stir at room temperature for seven days then concentrated by rotary evaporation, redissolved in methanol and acidified with cone. HC1 then reconcentrated by 73 rotary evaporation. The residue was dissolved in 1% aqueous HCl (15 ml) and heated to reflux for 18 hours. The solution was allowed to cool to room temperature then eluted through anion exchange resin (Amberlite IR-45, H C 0 3 , 10 ml) and concentrated by rotary evaporation^ 07). The crude product was purified by flash chromatography (5:2 ethyl acetate/methanol) furnishing 2.12 as a crystalline white solid (118 mg, 24%). mp: 170-171°C. 'H-NMR (200 MHz, D 20): 8 4.0 (d, 1 H, J 2 , 3 9.25 Hz, H-2), 3.93 (m, 1 H, H-4), 3.68 (t, 1 H, J 3 , 4 9.1 Hz, H-3), 3.48 (dd, 1 H, J 4 , 5 b 5.7 Hz, J 5 a ; 5 b 12.4 Hz, H-5b), 3.11 (dd, 1 H, J 4 , 5 a 8.9 Hz, H-5a). LSIMS-MS: m/z 149 (M + H +). 15N-1 -Deoxyxylonoj irimycin (2.6). The lactam (2.12) (100 mg, .675 mmol) was combined with hexamethyldisilizane (0.5 ml, 2.2 mmol) and trimethylsilylchloride (0.05 ml, 0.3 mmol) in anhydrous acetonitrile (1 ml) and stirred for one hour at 82°C. The solution was filtered, concentrated by rotary evaporation and dissolved in 1,4-dioxane (5 ml) under an inert atmosphere. BH 3 .Me2S (10 M , 0.4 ml, 4 mmol) was added and the solution was stirred for 5 hours at 100°C then allowed to stand for 16 hours at room temperature. 1 M HCl (3 ml) was added slowly and the solution was heated to reflux for 1 hour, concentrated with 1% HCl in methanol (3x10 ml). Crystallisation from water and ethanol provided 2.6 (70 mg, 80%). mp: 125-126°C. ] H-NMR (200 MHz, D 2 0): 8 3.76 (ddd, 2 H, J 2 , 3 = J 3 , 4 8.4 Hz, J , , 2 = J 4 , 5 4.5 Hz, J r , 2 = U,s- 10.2 Hz, H-2, H-4), 3.50 (t, 1 H, H-3), 3.43 (dd, 2 H, J 1 > r = J 5 , 5 ' 12.7 Hz, H-5, H-l), 2.91 (dd, 2 H, H - l \ H-5'). LSIM-HRMS: calculated for C 5 H , , 1 5 N 0 3 + H + : 135.0739. Found: 135.07900. The pK a of 15N-1-deoxyxylonojirimycin (2.6) was determined by 1 5 N-NMR by measuring the chemical shift of the endocyclic nitrogen over a range of pH values . The pK a 74 was determined to be 7.30. D-Xylose diethyl dithioacetal (2.41) D-Xylose (10 g, 66.7 mmol) and ethanethiol (2 ml, 160 mmol) were dissolved in concentrated HC1 (10 ml) at 0°C. The reaction was warmed to room temperature and allowed to stir for five hours. The solution was then diluted with methanol (20 ml), neutralised with lead (IT) carbonate then filtered through Celite. The Celite bed was washed with hot methanol (5 x 20 ml) and the pooled filtrate was concentrated in vacuo giving 2.41 as a thick yellow syrup which was pure enough for to be used in the next reaction (18.0 g, >95%). 'H-NMR (400 MHz, CDC1 3, 1 drop D 20): 5 4.04 (d, 1 H, J 1 > 2 8.5 Hz, H-l), 4.02 (dd, 1 H, J 2 3 2.2 Hz, J 3 , 4 3.7 Hz, H-3), 3.83 (m, 1 H, H-4), 3.71 (d, 2 H, J 4 , 5 5.0 Hz, H5, H5), 3.67 (dd, 1 H, H-2), 2.75-2.59 (m, 4 H, 2 CH 2 ) , 1.25 (t, 6 H, J 7.4 Hz, 2 CH 3 ) . 5-0-Triphenylmethyl-D-Xylose diethyl dithioacetal (2.42) 1,1-Dithioethyl-D-xylose (2.41) (6.0 g, 23.3 mmol) and chlorotriphenylmethane (8.4 g, 30 mmol) were dissolved in pyridine (100 ml) and stirred for 20 hours at 35°C. Water (20 ml) was added and the solution was concentrated by rotary evaporation to a thick syrup which was redissolved in CH 2 C1 2 (200 ml) and washed with water (2 x 75 ml), dried over MgS0 4 , filtered and concentrated in vacuo. Chromatography (5:1 - 2:1 petroleum ether / ethyl acetate) gave 2.42 as a thick yellow syrup (9.7 g, 82%). 'H-NMR (200 MHz, CDC13): 8 7.40-7.20 (m, 15 H, Ph), 4.2 (s, 1 H, H-3), 4.05 (d, 1 H, J, i 2 9.0 Hz, H-l), 3.9 (m, 1 H, H-4), 3.55 (dd, 1 H, J 2 i 3 10.2 Hz, H-2), 3.4-3.2 (m, 2 H, H 5 a , H 5 b ) , 2.7 (m, 4 H, CH 2 ) , 1.2 (m, 6 H, CH 3 ) . 1 3 C - N M R (50 MHz, CDC13): 8 148 (C, CPh3), 146 (C, Ph), 129 (CH, Ph), 128 (CH, Ph), 127 75 (CH, Ph), 75 (CH), 73 (CH), 68 (CH), 66 (CH 2 , C-5), 28 (CH 2), 25 (CH 2), 17 (CH 3). 2,3,4-Tri-O-benzyl-1,1 -dithioethyl-5-O-triphenylmethyl-D-xylose (2.43) Sodium hydride (2.18 g, 55-65% suspension in oil, 55 mmol) was washed with hexanes (2 x 10 ml) and suspended in anhydrous DMF (30 ml) under an inert atmosphere. The solution was cooled to 0°C and 2.42 (6.0 g, 12.1 mmol) was added in D M F (20 ml) over 20 minutes. The solution was allowed to stir an additional 10 minutes before benzyl bromide (6.5 ml, 36 mmol) was added over 20 minutes. The reaction mixturewas then allowed to stir at room temperature overnight. Methanol (5 ml) was added and the solution was concentrated by rotary evaporation. The resulting slurry was dissolved in CH 2 C1 2 (100 ml) and was washed with 1 M HC1 (50 ml), water (50 ml), dried over MgSCU, and concentrated by rotary evaporation. Flash chromatography (30:1 - 20:1 petroleum ether / ethyl acetate) provided 2.43 as a white solid (8.05 g, 86.9 %). 'H-NMR (200 MHz, CDC13): 5 7.5-7.1 (m, 30 H, Ph), 4.7-4.6 (m, 4 H, PhCH 2), 4.59 (d, 1 H, J 11.7Hz, PhCH2), 4.53 (d, 1 H, J 11.7Hz, PhCH2), 4.25 (dd, 1 H, J 1 > 2 6.4 Hz, J 2 , 3 4.2 Hz, H-2), 3.83 (m 1 H, H-3), 3.78 (m, 2 H, H - l , H-4), 3.4 (m, 2 H, H 5 , H 5 ), 2.62 (q, 4 H, J 7.4 Hz, 2 CH 2 ) , 1.19 (t, 3 H , J 7.4 Hz, CH 3 ) , 1.16 (t, 3 H, J 7.4 Hz, CH 3 ) . 2,3,4-Tri-0-benzyl-5-0-triphenylmethyl-D-xylose (2.39) Dithioacetal 2.43 (1.93 g, 2.51 mmol) was dissolved in a mixture of acetonitrile and water (45 ml, 10:1) and to this was added HgCl 2 (2.04 g, 7.53 mmol), C a C 0 3 (0.753 g, 7.53 mmol) and the reaction was allowed to stir at room temperature for 70 minutes. The mixture was filtered through Celite, and the Celite bed was washed with CH 2 C1 2 (3 x 10 ml) and the pooled filtrate was concentrated by rotary evaporation. Flash chromatography (20:1 76 petroleum ether / ethyl acetate) provided 2.39 as a clear oil (1.2 g, 73%). 'H-NMR (400 MHz, CDCI3): 8 9.7 (s, 0.7 H, H-l), 7.7-7.2 (m, 30 H, Ph), 4.7-4.53 (m, 5 H, PhCH2), 4.35 (d, 1 H, J 11.7 Hz, PhCH2), 4.18 (t, 1 H, J 3 , 4 5 Hz, H-3), 3.93 (dd apparent, 1 H, H-4), 3.88 (d, 1 H, J2,3 4.9 Hz, H-2), 3,47 (dd, 1 H, J 4 , 5 a 5.4, J 5 , 5 b 10.0 Hz, H-5a), 3.35 (dd, 1 H, J 4 , 5 b 5.3, H-5b). Methyl-2,3,4-tri-0-benzyl-P-D-xylopyranoside (2.46) Methyl P-D-xylopyranoside (4.0 g, 24.36 mmol) was dissolved in anhydrous DMF (100 ml) under an inert atmosphere and cooled on ice before sodium hydride (4.4 g, 55-65% suspension in oil, 110 mmol), benzyl bromide (13.2 ml, 109.6 mmol) and ammonium iodide (300 mg, 2.1 mmol) were added. The solution was allowed to stir for 12 hours at room temperature and quenched by the addition of 1 M HCl (20 ml) and saturated (10 ml) sodium bicarbonate. The resulting slurry was concentrated to a small volume by rotary evaporation and diluted with methylene chloride (200 ml), and washed with deionised water (100 ml). The organic phase was dried with MgS0 4 , filtered and concentrated to a thick yellow oil which was purified by flash chromatography (30:1 - 20:1 petroleum ether / ethyl acetate) to yield 2.46 as a white solid (9.51 g, 89.8%). 'H-NMR (400 MHz, CDC13): 8 7.40-7.18 (m, 15 H, Ph), 4.90-4.80 (m, 3 H, PhCH 2), 4.72 (d, 1 H, J 11.6 Hz, PhCH 2), 4.69 (d, 1 H, J 11.6 Hz, PhCH 2), 4.61 (d, 1 H, J 11.6 Hz, PhCH 2), 4.25 (d, 1 H, J>,2 7.53 Hz, H-l), 3.93 (dd, 1 H, J 5 a , 5 b 11.6 Hz, J 4 , 5 a 4.9 Hz, H-5a), 3.63-3.52 (m, 2 H, H-2, H-4), 3.52 (s, 3 H, OCH 3 ) , 3.34 (t, 1 H, J 2 , 3 ~ J3,4 8 Hz, H-3), 3.21 (dd, 1 H, J 4 , 5 b 9.6 Hz, H-5b). 1 3 C - N M R (50 MHz, CDC13):8138.7 (C, Ph), 138.6 (C, Ph), 138.2 (C, Ph), 128.5 (CH, Ph), 128.4 (CH, Ph), 128.04 (CH, Ph), 128.0 (CH, Ph), 127.9 (CH, Ph), 127.6 (CH, Ph), 83.7 (CH), 82.0 (CH), 77.9 (CH), 75.6 77 (PhCH2), 74.9 (PhCH2), 73.4 (PhCH2), 63.9 (CH 2 , C-5), 57.0 (OCH 3). Anal. Calc. for C27H30O5: C, 74.63; H, 6.96; Found: C, 74.58; H, 6.76. DCI-MS: m/z 452 (M + NH 4 + ) . 2,3,4-Tri-O-benzyl-D-xylopyranose (2.47) The benzylated methyl glycoside 2 .46 (3.0 g, 6.9 mmol) was dissolved in glacial acetic acid (60 ml) and heated to reflux then 1 M sulfuric acid (30 ml) was added over 10 minutes. The reaction mixture was allowed to reflux for an additional 2 hours and cooled to room temperature before the addition of cold deionised water (600 ml). The resulting suspension was left at 4°C overnight and filtered. The precipitate was triturated with cold water (50 ml), redissolved in acetone (100 ml), dried with magnesium sulfate, filtered and concentrated in vacuo to a slightly yellow solid. Crystallisation (ethyl acetate / petroleum ether) afforded a pure sample of 2 .47 as a mixture (30% a, 70% (3) of anomers (2.218 g, 76.5%). 'H-NMR (400 MHz, CDC13): selected data only, 8 5.1 (d, 0.3 H, J i a , 2 3.6 Hz, H-la), 4.68 (d, 0.7 H, J i P ; 2 6.9 Hz, H-lp). 1 3 C (50 MHz, CDC13): 8 138.7 (C, Ph), 138.6 (C, Ph), 138.4 (C, Ph), 138.3 (C, Ph), 138.1 (C, Ph), 137.9 (C, Ph), 128.54 (CH, Ph), 128.48 (CH, Ph), 128.42 (CH, Ph), 128.1 (CH, Ph), 128.04 (CH), 127.9 (CH, Ph), 127.84 (CH, Ph), 127.81 (CH, Ph), 127.7 (CH, Ph), 97.81 (CH), 91.44 (CH), 83.32 (CH), 82.45 (CH), 80.55 (CH), 79.52 (CH), 77.74 (PhCH2), 77.56 (PhCH2), 74.81 (PhCH2),73.42 (PhCH2), 73.25 (PhCH2), 63.77 (CH 2 , C-l), 60.33 (CH 2 , C-l). Anal. Calc. for C 2 6 H 2 8 0 5 : C; 74.26, H; 6.71, Found: C; 74.24, H; 6.54. DCI-MS: m/z 438 (M + NH 4 + ) . 2,3,4-Tri-C>-benzyl-D-xylono-l,5-lactone (2.44) 2,3,4-Tri-<9-benzyl-D-xylopyranose (1.52 g, 3.62 mmol) was dissolved in DMSO (15 78 ml) and acetic anhydride (10 ml). The reaction mixture was stirred for 16 hours at room temperature and quenched by the addition of deionised water (10 ml). A further volume of water (20 ml) was added and the resulting white precipitate was collected by filtration and washed with deionised water (10 ml), collected and dried in vacuo over phosphorus pentoxide to yield 2.44 as a white crystalline solid (1.40 g, 93%). mp: 114-116°C. 'H-NMR (400 MHz, CDC13): 5 7.5-7.2 (m, 15 H, Ph), 5.01 (d, 1 H, J 11.6 Hz, PhCH 2), 4.64 (d, 2 H, J 11.1 Hz, PhCH 2), 4.60-4.46 (m, 3 H, PhCH 2), 4.38 (ddd, 1 H, J 5 a , 5 b 12.3 Hz, J 4 , 5 a 3.4 Hz, J 3 , 5 a 1.5 Hz, H-5a), 4.27 (dd, 1 H, J 4 , 5 b 2.05 Hz, H-5b), 4.12 (d, 1 H, J 2 , 3 6.6 Hz, H-2), 3.88 (dt, 1 H, J 3 > 4 2.0 Hz, H-3), 3.75 (m, 1 H, H-4). Anal. Calc. for C 2 6 H 2 6 0 5 : C, 74.62; H, 6.26; Found: C, 74.31; H , 6.22. DCI-MS: m/z 419 (M+H+), 436 (M + NH 4 + ) . 2,3,4-Tri-0-benzyl-D-xylonamide(2.49) Lactone 2.44 (1.23 g, 2.94 mmol) was suspended in 40 ml anhydrous methanol at 0°C and N H 3 was bubbled through for 10 minutes. The flask was sealed and allowed to stir for a further 60 minutes. Concentration yielded 2.49 as a white solid (1.28 g, 95%). mp: 137-138°C. 1 H-NMR (400 MHz, CDC13): 5,7.40-7.20 (m, 15 H, Ph), 6.59 (br s, 1 H, NH), 5.60 (br s, 1 H, NH), 4.72-4.46 (m, 6 H, PhCH 2), 4.08 (d, 1 H, J 2 , 3 2.6 Hz, H-2), 4.04 (dd, 1 H, J 3 , 4 6.9 Hz, H-3), 3.75 (m, 1 H, H-4), 3.65 (m, 1 H, H-5a), 3.46 (m, 1 H, H-5b), 2.1 (br s, 1 H, OH). I 3 C - N M R (50 MHz, CDC13): 5 173.8 (CO), 137.5 (C, Ph), 137.4 (C, Ph), 136.4 (C, Ph), 128.7 (CH, Ph), 128.5 (CH, Ph), 128.39 (CH, Ph), 128.36 (CH, Ph), 127.9 (CH, Ph), 79.7 (CH), 79.6 (CH), 79.2 (CH), 75.3 (PhCH2), 73.6 (PhCH2), 73.3 (PhCH2), 61.7 (CH 2 , C-5). IR (CDCI3): cm"1 3516m, 3400m, 3034w, 2879w, 1690vs, 1568m, 1497w, 1454w, 1404w, 1342w, 1209w, 1123m, 1052s, 827w. Anal. Calc. for C 2 6 H 2 9 N 0 5 : C; 71.70; H, 6.71; N, 3.22; 79 Found: C; 71.71, H; 6.60, N, 3.13. DCI-MS: m/z 436 (M + FT). 2,3,4-Tri-(9-benzyl-5-(9-methanesulfonyl-D-xylonamide (2.50) The amido alcohol 2.49 (1.00 g, 2.3 mmol) was combined with pyridine (1 ml) in CH 2 C1 2 (25 ml) and to this methanesulfonyl chloride (0.3 ml, 3.5 mmol) was added dropwise over 5 minutes. The reaction mixture was allowed to stir for 18 hours at room temperature, diluted with CH 2 C1 2 (40 ml), and washed with water (50 ml), 1 M HCl (30 ml), saturated NaHC03(aq) (30 ml), dried over MgS04 and concentrated to a yellow oil. Crystallisation from ethyl acetate and petroleum ether afforded a white crystalline sample (979 mg, 82%). mp: 109-110°C. 'H-NMR (400 MHz, CDC13): 5 7.40 - 7.20 (m, 15 H, Ph), 6.58 (br s, 1 H, NH), 5.53 (br s, 1 H, NH), 4.66 - 4.47 (m, 7 H, 6 PhCH 2 , H-5a), 4.20 (dd, 1 H, J 5 a , 5 b 11.4 Hz, J4,5b 7.0 Hz, H-5b), 4.17 (d, 1 H , J 2 , 3 2.5 Hz, H-2), 4.03 (dd, 1 H, J 3 , 4 6.3 Hz, H-3), 3.88 (dt, 1 H, J 4 , 5 a 2.2 Hz, H-4). 1 3 C - N M R (50 MHz, CDC13): 5 173.8 (CO), 137.5 (C, Ph), 137.3 (C, Ph), 136.2 (C, Ph), 128.8 (CH, Ph), 128.62 (CH, Ph), 128.56 (CH, Ph), 128.48 (CH, Ph), 128.40 (CH, Ph), 128.32 (CH, Ph), 128.15 (CH, Ph), 128.0 (CH, PH), 79.0 (CH), 78.32 (CH), 77.09 (CH), 74.9 (PhCH2), 73.9 (PhCH2), 73.4 (PhCH2), 70.63 (CH 2 , C-5), 37.2 (CH 3). Anal. Calc. For C 2 7 H 3 i N 0 7 S : C; 63.14, H; 6.08, N; 2.73, Found: C; 63.08, H; 6.02, N; 2.72. DCI-MS: m/z 514 (M + H + ). 5-Azido-2,3,4-tri-0-benzyl-5-deoxy-D-xylonamide (2.53) The methanesulfonate (2.50) (500 mg, 0.97 mmol) and sodium azide (630 mg, 9.7 mmol) were combined in DMPU (7 ml) and heated to 80 °C for 5 hours. The solution was then allowed to cool and diluted with deionised water (20 ml) The resulting precipitate was 80 triturated with deionised water, and collected (440 mg, >95%). mp: 98-100°C. 'H-NMR (400 MHz, CDC13): 6 7.4-7.2 (m, 15 H, Ph), 6.57 (br s, 1 H, NH), 5.54 (br s, 1 H, NH), 4.67-4.53 (m, 5 H, PhCH2), 4.47 (d, 1 H, J 11.4 Hz, PhCH2), 4.05 (d, 1 H, J 2, 3 2.5 Hz, H-2), 3.97 (dd, 1 H, J 3 > 4 6.7 Hz, H-3), 3.72 (td, 1 H, J 4 , 5 a , 2.9 Hz, H-4), 3.30 (dd, 1 H, J 5 a, 5b 13.3 Hz, H-5a), 3.12 (dd, 1 H, J 4 l 5 b 7.2 Hz, H-5b). IR (CDC13): cm"1 3517m, 3400m, 3068w, 3034w, 21025, 16925, 1567m, 1497w, 1454w, 1321w, 1212w, 1081m. DCI-HRMS: calculated for C 2 6 H 2 8 0 4 N 4 + H +: 461.21887. Found: 461.21926. 2,3,4-Tri-0-benzyl-D-xylono-l,5-lactam (2.26) The azide 2.53 (440 mg, 0.97 mmol) was combined with triphenylphosphine (350 mg, I. 15 mmol) in a 9:1 mixture of THF and deionised water. After stirring for 12 hours at room temperature, the solution was concentrated by rotary evaporation and the residue was purified by flash chromatography (petroleum ether / ethyl acetate 2:3) giving a semicrystalline white solid (370 mg, 91%). mp: 106-108°C. 1 H-NMR (CDCI3): 8 7.45-7.2 (m, 15 H, Ph), 6.5 (br s, 1 H, NH), 5.05 (d, 1 H, J 11.5 Hz, PhCH2), 4.69 (d, 2 H, J 10.1 Hz, PhCH2), 4.64 (d, 1 H, J I I . 3 Hz, PhCH2), 4.59 (d, 1 H, J 11.8 Hz, PhCH2), 4.52 (d, 1 H, J 11.8 Hz, PhCH2), 3.94 (d, 1 H, J 2, 3 7.1 Hz, H-2), 3.80 (t, 1 H, J 3 , 4 7.1 Hz, H-3), 3.69 (m, 1 H, H-4), 3.30 (dt, 1 H, J4,5a~J5a,NH 3.9 Hz, J 5 a , 5 b 12.7 Hz, H-5a), 3.18 (ddd, 1 H J 4 , 5 b 6.9 Hz, J 5 b , N H 2.7 Hz). I 3 C-NMR (50 MHz, CDC13): 8 171.6 (CO), 137.96 (C, Ph), 137.85 (C, Ph), 137.67 (C, Ph), 128.44 (CH, Ph), 128.35 (CH, Ph), 128.30 (CH, Ph), 128.18 (CH, Ph), 127.91 (CH, Ph), 127.87 (CH, Ph), 127.69 (CH, Ph), 82.07 (CH), 78.83 (CH), 76.00 (CH), 74.14 (PhCH2), 73.80 (PhCH2), 71.96 (PhCH2), 42.12 (CH2, C-5). Anal. Calc. For C 2 6H 2 7N0 4: C; 74.80, H; 6.52, N; 3.35, Found: C; 74.70, H; 6.40, N; 3.25. 81 1 -Amino-1,5-anhydro-2,3,4-tri-O-benzy 1-1 -deoxy-D-Xylitol (2,3,4-Tri-O-benzyl-1 -deoxyxylonojirimycin) (2.13) Under an inert atmosphere, lactam 2.26 (1.51 g, 3.62 mmol) was added to anhydrous diethyl ether (75 ml) and heated to reflux. Anhydrous THF (20 ml) was added to dissolved remaining starting material and the solution was cooled to room temperature before the addition of lithium aluminum hydride (1.5 g, 35 mmol). The solution was heated at reflux for 2.5 hours and cooled to room temperature. Ethyl acetate (150 ml) was added slowly and the reaction was quenched by the addition of 1 M NaOH (60 ml) The solution was extracted with ethyl acetate (100 ml) washed with saturated aqueous NaHC03 (50 ml), dried over MgS0 4 , filtered and concentrated by rotary evaporation affording crude 2.13 as a yellow oil. Crystallisation (ethyl acetate / petroleum ether) and drying in vacuo over P2O5 provided 2.13 as an analytically pure sample (1.412 g, 95%). mp: 49-50°C. 'H-NMR (200 MHz, CDC13): 8 7.5-7.2 (m, 15 H, Ph), 4.80 (s, 2 H, PhCH2), 4.63 (d, 2 H, J 11.8, PhCH2), 4.56 (d, 2 H, J 11.8, PhCH2), 3.55-3.39 (m, 3 H, H-2, H-4, H-3), 3.21 (dd, 2 H, J 1 2=J 5 2 4 Hz, J U .=J„.12 Hz, H - l , H-5), 2.50 (dd, 2 H, J1 > 2=J5,4 9.3 Hz, H - l ' , H-5'), 2.1 (br s, 2 H, NH + H 20). 1 3 C (50 MHz, CDCI3): 5 139.0 (C, Ph), 138.7 (C, Ph), 128.4 (CH, Ph), 128.0 (CH, Ph), 127.8 (CH, Ph), 127.7 (CH, Ph), 127.5 (CH, Ph), 85.7 (CH, C-3), 79.7 (CH, C-2, C-4), 75.5 (CH 2 , PhCH2), 72.9 (CH 2 , PhCH2), 49.1 (CH 2 , C - l , C-5). Anal. Calc. For C 2 6 H 2 9 N0 3 : C; 77.39, H; 7.24, N; 3.47, Found: C; 77.10, H; 7.18, N; 3.42. 1 -Amino-1,5-anhydro-2,3,4-tri-0-benzyl-1,1 -dideutero-1 -deoxy-D-xylitol (2,3,4-Tri-O-benzyl-1,1 -dideutero-1 -deoxyxylonoj irimycin) (2.55) The protected lactam 2.26 (500 mg, 1.21 mmol) was reduced to 2.55 by treatment 82 with L i A l D 4 (450 mg, 10.9 mmol) in diethyl ether (25 ml) and THF (5 ml) and chromatographed as described above for 2.13 (450 mg, 92%). mp: 49-50°C. 'H-NMR (400 MHz, CDC13): 5 7.5-7.2 (m, 15 H, Ph), 4.92 (s, 2 H, PhCH2), 4.73 (d, 2 H, J 11.7 Hz, PhCH2), 4.67 (d, 2 H, J 11.7 Hz, PhCH2), 3.55 (t, 1 H, J 2 , 3 ~ J 3 , 4 8.3 Hz, H-3), 3.5-3.4 (m, 2 H, H-2, H-4), 3.22 (dd, 1 H, J 4 > 5 a 4.8 Hz, J 5 a , 5 b 12.3 Hz, H-5a), 2.51 (dd, 1 H, J 4 ; 5 b 10.0 Hz, H-5b), 1.45 (s, 1 H, NH). 1 3 C - N M R (50 MHz, CDC13): 8 139.0 (C, Ph), 138.7 (C, Ph), 128.4 (CH, Ph), 128.0 (CH, Ph), 127.8 (CH, Ph), 127.7 (CH, Ph), 127.5 (CH, Ph), 85.7 (CH, C-3), 79.8, 79.7 (CH, C-2, C-4), 75.5 (CH 2 , PhCH2), 72.9 (CH 2 , PhCH2), 49.1 (CH 2 , C-5) (C-l not apparent). Anal. Calc. For C 2 6 H 2 7 D 2 N 0 3 : C; 77.01, H; 7.19, N; 3.45, Found: C; 76.87, H; 7.20, N; 3.33. 1 -Amino-1,5-anhydro-1 -deoxy-D-xylitol hydrochloride (1 -Deoxyxylonojirimycin hydrochloride) (1.35) A sample of 1-deoxyxylonojirimycin (1.35) was prepared from D-xylonolactam (1.34) exactly as described by Godskesen et a\(107). Alternatively, benzylated 1-deoxyxylonojirimycin 2.13 (500 mg, 1.24 mmol) was dissolved in a 1:1 mixture of methanol and glacial acetic acid (10 ml) and treated with 10% Pd/C (50 mg) and exposed to H 2 ( g ) for 24 hours at 1 atmosphere. The suspension was filtered through Celite and the filtrate was concentrated and co-concentrated with 1 M HC1 (3 x 25 ml). Crystallisation (methanol / diethyl ether) gave a pure sample of 1.35 (191 mg, 80%). mp: 125-126°C. Lit: 126.5-127.5. 'H-NMR (200 MHz, D 20): 8 3.76 (m, 2 H, H-2, H-4), 3.50 (t, 1 H, J 2 , 3 = J 3 , 4 8.4 Hz, H-3), 3.43 (dd, H, J 5 , 5 - 12.7 Hz, H-5), 2.91 (dd, 1 H, H-5'). Anal. Calc. for C 5 H 1 2 N0 3 C1: C, 35.41; H,7.13;N, 8.26; 0,28.30; Cl, 20.91. Found: C, 35.60; H, 7.30; N, 8.09. 83 1,1 -Dideutuero-1 -deoxyxylonojirimycin (2.7) The deuterated iminosugar 2.55 (500mg, 1.24 mmol) was converted to 2.7 exactly as described above for 1.35. Crystallisation from methanol / diethyl ether gave a sample of 2.7 as the acetate salt (200 mg, 84%). mp: 125-126°C. 'H-NMR (200 MHz, D20): 8 3.76 (ddd, 2 H, J 2, 3 = J3,4 8.4 Hz, Ji,2 = J 4 l 5 4.5 Hz, Jl',2 = J4,5' 10.2 Hz, H-2, H-4), 3.50 (t, 1 H, H-3), 3.43 (dd, 1 H, J 5, 5' 12.7 Hz, H-5, H-l), 2.9 (dd, 1 H, H-5'), 2.05 (s, 3 H, Ac). LSLMS-HRMS: calc. for C 5Hi 0D 2NO 3 (M + 1): 136.0941. Found: 136.09433. Anal. Calc. for C7Hi2D2N05: C, 43.07; H + D, 7.74; N, 7.18. Found: C, 43.01; H, 7.45; N, 7.03. N-Benzyloxycarbonyl-1-deoxyxylonojirimycin (2.8) l-Deoxyxylonojirimycin 1.35 (120 mg, 0.71 mmol) was dissolved in a 1:1 mixture of 1,4-dioxane and deionised water (10 ml) and treated with triethylamine (250 pi) and benzyl chloroformate (400 pi). After 16 hours at room temperature the solution was concentrated by rotary evaporation and 2.8 was isolated by flash chromatography (37:2:1 ethyl acetate / methanol / water) (120 mg, 55.8%). mp: 151-152°C. 'H-NMR (400 MHz, CD3OD): 8 7.55-7.25 (m, 5 H, Ph), 5.12 (s, 2 H, PhCH2), 4.08 (dd, 2 H, J u - = J5,5- 12.5 Hz, J,, 2 = J 4, 5 4.5 Hz, H-l, H-5), 3.4-3.3 (m, 2 H, H-2, H-4), 3.24 (t, 1 H, J 2, 3 = J 3 , 4 8.4 Hz, H-3), 2.74 (br s, 2 H, H-1', H-5'). 1 3 C (75 MHz, D20): 8 156.8 (CO), 136.4 (C, Ph), 129.0 (CH, Ph), 128.7 (CH, Ph), 128.0 (CH, Ph), 78.0 (CH, C-3), 69.2 (CH, C-2, C-4), 68.2 (PhCH2), 47.6 (CH2, C-l, C-5). Anal. Calc. For Ci 3Hi 7N0 5: C; 58.42, H; 6.41, N: 5.24, Found: C; 58.10, H; 6.44, N; 5.09. DCI-HRMS: calc. for C 1 3 H n N0 5 (M + H+): 267.11069. Found: 267.11032. 84 TV-Benzyloxycarbonyl-1 -deoxyxylonojirimycin (2.9) The benzyl carbamate 2.9 was prepared from 1,1-dideutero-1-deoxyxylonojirimycin 2.7 (120 mg, 0.62 mmol) as described in the preparation of 2.8 (121 mg, 64%). mp: 97-98°C. 'H-NMR (200 MHz, CD3OD): 5 7.55-7.25 (m, 5 H, Ph), 5.12 (s, 2 H, PhCH2), 4.08 (dd, 1 H, J 4 > 5 4.5 Hz, J 5 a, 5b 12.5 Hz, H-5), 3.4-3.3 (m, 2 H, H-2, H-4), 2.74 (s, 1 H, H-5') 3.24 (t, 1 H, J 2, 3 ~ J3>4 8.4 Hz, H-3). 1 3 C (75 MHz, D20): 8 156.8 (OAc), 136.4 (C, Ph), 129.0 (CH, Ph), 128.7 (CH, Ph), 128.0 (CH, Ph), 78.0 (CH, C-3), 69.2 (br, CH, C-2, C-4), 68.2 (PhCH2), 47.6 (CH2, C-5) 47.6 (CH2, C-5) (C-l not apparent). Anal. Calc. For Ci 3Hi 5D 2N0 5 + 0.5 H20: C; 56.11, H; 6.87, N; 5.03, Found: C; 56.21, H; 6.19, N; 4.73. (6R,1S, 8S')-6,7,8-Tris(benzyloxy)-5,6,7,8-tetrahydropyrido/'i,2-li;tetrazole (2.58) The protected lactam 2.26 (300 mg, 0.72 mmol) was dissolved in anhydrous acetonitrile (10 ml) and cooled to -30 °C under N 2 ( g ) before sodium azide (56 mg, 0.86 mmol) was added. After stirring for 5 minutes, trifluoromethanesulfonic anhydride (0.18 ml, 1.1 mmol) was added and the reaction was allowed to stir at -10 °C for 12 hours. The solution was diluted with ethyl acetate (25 ml), washed with saturated aqueous NaHC03 (25 ml), dried over MgS04, and concentrated by rotary evaporation. Column chromatography (4:1 petroleum ether / ethyl acetate) provided 2.58 as a white solid (171 mg, 54%). 1 H-NMR (400 MHz, CDC13): 8 7.5-7.25 (m, 15 H, Ph), 5.08 (d, 1 H, J 11.9 Hz, PhCH2), 4.89 (d, 1 H, J 11.9 Hz, PhCH2), 4.77 (d, J 2 > 3 2.9 Hz, H-2), 4.68 (d, 1 H, J 11.5 Hz, PhCH2), 4.59 (d, 1 H, J 11.8 Hz, PhCH2), 4.55 (dd, 1 H, J 4 , 5 a 3.6 Hz, J 5 a, 5b 13.4 Hz, H-5a), 4.49 (d, 2 H, J 11.8 Hz, PhCH2), 4.42 (dd, 1 H, J 4 , 5 b 2.9 Hz, H-5b), 4.21 (dd, 1 H, J 3, 4 2.8 Hz, H-3), 4.00 (ddd, 1 H, H-4). 1 3 C-NMR (50 MHz, CDC13): 8 136.8 (C, Ph), 136.7 (C, Ph), 128.7 (CH, Ph), 128.43 (CH, Ph), 85 128.35 (CH, Ph), 128.1 (CH, Ph), 127.9 (CH, Ph), 75.9 (CH), 73.1 (PhCH2), 72.7 (PhCH2), 72.1 (CH), 71.9 (PhCH2), 68.6 (CH), 47.1 (CH 2 , C-5). DCI-HRMS: calc. for C 2 6H 2 7 N 4 03 (M + H +): 443.20700. Found: 443. 20831. (6R, IS, 8,S)-5,6,7,8-Tetraliydropyrido^,2-fif7tetrazole-6,7,8-triol (D-xylonojirimycin tetrazole) (2.1) The benzylated tetrazole 2.58 (146 mg, 0.33 mmol) was dissolved in a 1:1 mixture of methanol and glacial acetic acid (5 ml), treated with 10% Pd/C (10 mg) and exposed to H 2 ( g ) at 1 atmosphere for 3 days. Filtration through Celite, concentration, and crystallisation (methanol, diethyl ether) provided a crystalline sample of 2.1 (39 mg, 69%). mp: 151-152°C. 'H-NMR (400 MHz, D 20): 8 4.95 (d, 1 H, J 2 , 3 7.0, H-2), 4.80 (dt, 1 H, J 4 , 5 a ~ h,5b 9.0 Hz, H-4), (q, 2 H, H-5a, H-5b), 4.03 (t, 1 H, J 3 , 4 7.5 Hz H-3). 1 3 C - N M R (50 MHz, D 20): 8 154.5 (C, C-l), 74.4 (CH), 66.9 (CH), 65.9 (CH), 49.2 (CH 2 , C-5). Anal. Calc. For C 5 H 8 N 4 0 3 : C; 34.89, H; 4.68, N; 32.55, Found: C; 35.12, H; 4.67, N; 32.45. /V-Formyl- 1-deoxyxylonojirimycin (2.5) The benzylated imino sugar 2.13 (220 mg, 0.545 mmol) was dissolved in anhydrous CH 2 C1 2 (5 ml) and treated with pentafluorophenyl formate (1.0 mmol, 210 mg) and allowed to stir for 45 minutes at room temperature. Flash chromatography of the crude reaction mixture (1:1 petroleum ether / ethyl acetate) provided the protected amide as a clear oil which was dissolved in a one to one mixture of methanol and glacial acetic acid (10 ml), and treated with 10% Pd/C (20 mg) and H 2 ( g ) at 1 atmosphere for 12 hours. Filtration through Celite, concentration by rotary evaporation followed by crystallisation (methanol, diethyl ether) 86 provided 2.5 as a white crystalline solid (80 mg, 91.9%). mp: 140-142°C. 'H-NMR (400 MHz, D 20): 5 8.0 (2, 0.9 H, CHO), 4.25 (ddd, 1 H, J l a l a , 1.9 Hz, J , a 2 6.5 Hz, J l a l b 12.7 Hz, H-la), 3.92 (ddd, 1 H, J l a , 2 , 5.2 Hz, J, a„b. 13.1 Hz, H-la'), 3.53 (m, 1 H, H-2), 3.45 (m, 2 H, H-2', H-3), 3.06 (dd, 1 H, l l b 210.2 Hz, H-lb), 2.70 (dd, 1 H, l l b , 2 , 10.2 Hz, H-lb'). Anal. Calc. For C 6 H „ N 0 4 : C; 44.72, H; 6.88, N; 8.69, Found: C; 44.17, H; 7.66, N; 8.53. N-Acetyl-1 -deoxyxylonojirimycin (2.4) l-Deoxyxylonojirimycin 1.35 (100 mg, 0.57 mmol) was dissolved in anhydrous DMF (1 ml) and triethylamine (0.15 ml, 1.15 mmol) was added. After stirring at room temperature for 5 minutes, pentafluorophenyl acetate (260 mg, 1.14 mmol) was added and the reaction mixture was stirred for an additional 15 minutes. The reaction was neutralised by the addition of saturated aq NH4C1 (0.2 ml) and concentrated by rotary evaporation. Flash chromatography (7:2:1 ethyl acetate / methanol / H 20) afforded a white crystalline sample of 2.4 (70 mg, 70%). mp: 186-187°C. 'H-NMR (400 MHz, D 20): 5 4.40 (ddd, 1 H, J l a l a , 2.3 Hz, J l a 2 4.6 Hz, J l a l b 13.0 Hz, H-la), 3.94 (ddd, 1 H, J, a , 2 , 5.0 Hz, J l a , lb, 13.5 Hz, H-la'), 3.54 (m, 1 H, H-2), 3.43 (m, 1 H, H-2), 3.35 (t, 1 H, J 2 3 ~ J 2 , 3 8.4 Hz, H-3), 3.04 (dd, 1 H, J l b 2 10.7 Hz, H-lb), 2.63 (dd, 1 H, J l b , 2 . 10.2 Hz, H-lb'). Anal. Calc. for C 7 H 1 3 N0 4 : C; 47.99, H; 7.48, N; 8.00, Found: C; 47.80, H; 7.56, N; 7.86. Af-Benzyl-1 -deoxyxylonojirimycin (2.3) l-Deoxyxylonojirimycin 1.35 (200 mg, 1.14 mmol), triethylamine (1 ml) and benzyl bromide (1 ml) were combined in D M F (5 ml) and stirred for 2 hours at room temperature. The solution was then concentrated onto silica gel by rotary evaporation and purified by column chromatography (37:2:1 - 27:2:1 diethyl ether / methanol / H 20) to a clear oil. 87 Crystallisation (methanol, diethyl ether) provided an analytically pure sample of 2.3 (220 mg, 84%). mp: 151-152°C. 'H-NMR (400 MHz, CD3OD): 8 7.4-7.2 (m, 5 H, Ph), 3.58 (s, 2 H, PhCH2), 3.48 (m, 2 H, H-2, H-2'), 3.08 (t, 1 H, J„ ~ J 2, 3 8.9 Hz, H-3), 2.93 (dd, 2 H, J l a l b ~ Jla, lb, 10.8 Hz, H-la, H-la'), 1.95 (t, 2 H, H-lb, H-lb'). ,3C-NMR (50 MHz, CD3OD): 8 139.4 (C, Ph), 131.7 (CH, Ph), 130.1 (CH, Ph), 129.2 (CH, Ph), 81.8 (CH, C-3), 72.3 (CH, C-2, C-2'), 63.8, 60.0 (CH2, C-l, C-l', PhCH2). Anal. Calc for Ci2Hi7N03: C; 64.55, H; 7.67, N; 6.27, Found: C; 64.59, H; 7.57, N; 6.20. Chemoenzymatic Preparation of 2,3,4,2',3',4'-Hexa-t7-acetyl-xylobiose (2.72) Xylan from birchwood (Sigma, 10 g) was suspended in sodium acetate buffer (50 mM, pH 4.65, 1 L) and heated to 35 °C for 8 hours. The non-soluble portion was removed by suction filtration and the filtrate was incubated with xylanase from Trichoderma viride (1 mg, Sigma, 100-300 U/mg) for 3 days at 35 °C. The solution was then lyophilised and the residue extracted with methanol (10 x 50 ml). The combined extracts were concentrated in vacuo, redissolved in pyridine (50 ml) and treated with acetic anhydride (22 ml) at room temperature over 24 hours. The solution was concentrated in vacuo, redissolved in CH2C12 (150 ml), washed with 1 M HCl (3 x 50 ml), 50% saturated aqueous NaHC03 (2 x 50 ml), dried over MgS04, filtered and concentrated by rotary evaporation. Purification by column chromatography (2:1 - 1:1 petroleum ether / ethyl acetate) provided an anomeric mixture of 2.72 as a foamy white solid (2.89 g). 'H-NMR (400 MHz, CDC13): 8 6.19 (d, 0.25 H, J, a , 2 3.7 Hz, H-la), 5.61 (d, 0.75 H, J i p > 2 7.3 Hz, H-l(3), 5.14 (t, 1 H, J 2, 3 ~ J 3 , 4 8.6 Hz, H-3), 5.06 (t, 1 H, J3.,4- 7.8 Hz, H-3'), 4.94 (m, 1 H, H-2), 4.85 (m, 1 H, H-4'), 4.78 (dd, 1 H, h:r 7.8 Hz, H-2'), 4.54 (d, 1 H, J r,r 6.0 Hz, H-l'), 4.07 (dd, 1 H, J 4., 5 a. 4.7 Hz, J5a-,5b' 12.0 Hz, H-88 5a'), 3.99 (dd, 1 H, J 4 , 5 a 5.1 Hz, H-5a), 3.84 (m, 1 H, H-4), 3.44 (dd, 1 H, J4,5b 9.3 Hz, J 5 a, 5b 11.9 Hz, H-5b), 3.36 (dd, 1 H, J 4, 5 b- 7.8 Hz, H-5'), 2.05-1.98 (m, 18 H, Ac). Anal. Calc. For C22H30CM5: C; 49.44, H; 5.66. Found: C; 49.55, H; 5.75. 2,5-Dinitrophenyl-2,3,4,2',3',4'-penta-0-acetyl-p-xylobioside (2.74) Xylobiose per-0-acetate 2.72 (0.5 g, 0.91 mmol) was dissolved in anhydrous CH2CI2 (5 ml) under N2 3 8.7 Hz, H-2), 3.51 (m, 1 H, H-4'), 3.32 (t, 1 H, J2-,3- ~ J 3 > > 4. 9.0 Hz, H-3'), 3.23 (m, 2 H, H-2', H-5b'). Anal. Calc. for C 1 6H2oN 2Oi 3 + H20: C, 41.17; H, 4.75; N, 6.0. Found: C, 41.30; H, 4.64; N, 5.85. 2-Nitrophenyl-2,3,2',3',4'-penta-0-acetyl p-xylobioside (2.75) The bromide (2.73) was prepared from 2.72 (0.5 g, 0.91 mmol) as described above and combined with 2-nitrophenol (265 mg, 1.9 mmol), 2,6-lutidine (0.2 ml, 1.9 mmol), and 90 drierite in anhydrous acetonitrile (10 ml) as described for 2.74. Crystallisation from ethyl acetate and petroleum ether provided 2.75 in pure form (350 mg, 64%). mp: 186 - 187°C. Lit: 186-187°C(705). 'H-NMR (200 MHz, CD3OD): 8 7.76 (dd, 1 H, J2.,3„ 8.0 Hz, J2,.4„ 1.8 Hz, H-2"), 7.50 (dt, 1 H, J3„4.. 8.0 Hz, H-4"), 7.26 (dd, 1 H, J4..5.. 7.9 Hz, J3.,5„ 1.0 Hz, H-5"), 7.13 (dt, H-3"), 5.26 (d, 1 H, J 1 2 5.0 Hz, H-l), 5.17 (t, 1 H, J 2 3 ~ J 3 4 6.5 Hz, H-3), 5.11 (t, 1 H, J 3 4 . ~ J 2 3 . 7.5 Hz, H-3'), 5.05 (dd, 1 H, J 2 3 6.5 Hz, H-2), 4.90 (dt, 1 H, J4,5b. 8.1 Hz, J 4 5 a . 4.0 Hz, H-4'), 4.84 (dd, 1 H, J2,3. 7.9 Hz, H-2'), 4.59 (d, 1 H, J„2. 6.2 Hz, H-l'), 4.11 (dd, 1 H, H-5a), 4.09 (dt, 1 H, H-5a'), 3.86 (dt, 1 H, J 4 5 b 6.6 Hz, H-4), 3.54 (dd, 1 H, J 5 a 5 b 12.2 Hz, H-5b), 3.38 (dd, 1 H, J5 a,5 b, 11.9 Hz, H-5b'), 2.10 (s, 6 H, Ac), 2.03 (s, 6 H, Ac), 2.01 (s, 3 H, Ac). 13C (50 MHz, CDC13): 5 169.9, 169.7, 169.6, 169.1 (5 OAc), 149.1 (Ph), 141.1 (Ph), 133.7, 125.1, 122.9, 118.0 (Ph), 99.8, 98.6 (C-l, C-l'), 73.4, 70.79, 70.5, 69.8, 69.0, 68.4 (C-2, C-3, C-4, C-2', C-3', C-4'), 61.8, 61.5 (C-5, C-5'), 20.7, 20.6, 20.5 (5 OAc). Anal. Calc. for C26H3iNOi6: C, 50.90; H, 5.09; N, 2.28. Found: C, 50.97; H, 5.15; N, 2.20. 2-Nitrophenyl [3-xylobioside (2.77) 2.75 (200 mg, 0.326 mmol) was deprotected as described above for 2.76. The product was purified by column chromatography (27:2:1 ethyl acetate / methanol / H20) yielding 2.77 as a slightly yellow solid (99 mg, 75.2%). mp 112-114°C. Lit. 113-115°C (705). 'H-NMR (200 MHz, D20): 5 7.77 (dd, 1 H, h"A", 8.0 Hz, J r , 5 " 1.5 Hz, H-3"), 7.51 (dt, 1 H, J5",6" 7.5 Hz, H-5"), 7.25 (dd, 1 H, H-6"), 7.09 (dt, 1 H, J4-,5" 7.6 Hz, J4-,6" 1-0 Hz, H-4"), 5.1-5.04 (m, 1 H, H-l), 4.31 (d, 1 H, h;r 7.8 Hz, H-l'), 3.99 (dd, 1 H, H-5a), 3.81 (dd, 1 H, H-5a'), 3.72 (dt, 1 H, J 4 ) 5b 9.5 Hz, J 4 j 5 a 5.0 Hz, H-4), 3.59-3.40 (m, 2 H, H-2, H-3), 3.40 (dd, 1 H, J5a,5b 12.0 Hz, H-5b), 3.26 (t, 1 H, J3-,4- 9.0 Hz, H-3'), 3.48 (ddd, 1 H, J4-,5a- 5.3 Hz, J 4 . i 5 b . 91 11.5 Hz, H-4'), 3.14 (t, 1 H, J5a',5b- 11.5 Hz, H-5b'), 3.10 (dd, 1 H, h\y 9.0 Hz, H-2'). 3,1.2 Transglycosylation Experiments The Abg p-glucosidase mutant (AbgGlu358Ala) was provided by Karen Rupitz, Department of Chemistry, UBC. The enzyme was purified according to the procedure of Withers et al(9i). a-D-Glucosyl fluoride was kindly provided by David Zechel, UBC Department of Chemistry. The synthesis of all glycosyl acceptors is described above. All other chemical were purchased from Sigma Chemical Company and used without further purification. 3.1.2.1 Transglycosylation experiments with N-Benzyloxycarbonyl-1 -deoxyxylonojirimycin (2.8) The benzyl carbamate 2.8 (30.9 mg, 0.115 mmol) and a-glucosyl fluoride (22 mg, 0.12 mmol) and were combined in 150 mM phosphate buffer, pH 7.5 (2.0 ml) and treated with AbgGlu358Ala (0.17 mg/ml). The reaction mixture was allowed to incubate for 7 days at room temperature then diluted with methanol (20 ml) and concentrated in vacuo. The residue was redissolved in acetic anhydride and pyridine (2:3 v/v, 10 ml) and stirred for 24 hours at room temperature. The acetylated products were isolated by column chromatography (2:1 petroleum ether / ethyl acetate). Three products were isolated: p-1,3-linked disaccharide 2.62 (8.1 mg, 0.012 mmol, 10%), p-l,4-linked disaccharide 2.66 (4.3 mg, 0.006 mmol, 5%), p-l,4-linked trisaccharide 2.68 (1.9 mg, 0.002 mmol, 3.2%). Total yield: 18.2%. Yields are based on the amount of glycosyl donor. Structures of 2.66 and 2.67 are tentatively assigned. 92 23,4,6-Tetra-O-acetyl-p-D-gluco^ deoxyxylonojirimycin. (2.62) Rf 0.32 (3:2 petroleum ether - ethyl acetate). *H (400 MHz, CDC13): 5 7.35-7.25 (m, 5 H, Ph), 5.20-4.97 (m, 4 H, H-3, H-3', H-4', H-5'), 4.93 (t, 1 H, J2-,3' 9.0 Hz, H-3), 4.79 (br s, 1 H, H-2, H-4), 4.58 (d, 1 H, J,.,2. 9.0 Hz, H-l'), 4.31 (dd, 2 H, J l a , l b = J 5 a , 5 b 12.8 Hz, J l a , 2 = J 4 , 5 a 4.1 Hz, H-la, H-5a), 4.25-4.17 (br s, 1 H, PhCH2), 4.17-3.90 (br s, 1 H, PhCH2), 3.74-3.53 (m, 2 H, H-6a', H-6b'), 2.75 (dd, 2 H, J l b > 2 = J 4 , 5 b 10.1 Hz, H-lb, H-5b), 2.10-1.95 (m, 18 H, Ac). LSIMS-HRMS: calc for C3,H4oNOi6(M + H+): 682.2347. Found: 682.23394. 2,3,4,6-Tetra-(9-acetyl-P-D-glucopyranosyl-(l—»4)-2,4-di-0-acetyl-A7-benzyloxycarbonyl-l-deoxyxylonojirimycin. (2.66) Rf 0.28 (3:2 petroleum ether - ethyl acetate). ! H (400 MHz, CDC13): 6 7.35-7.20 (m, 5H, Ph), 5.3-5.1 (br s, 1 H), 5.16 (t, 1 H, J 9.5 Hz), 5.05 (br s, 1 H), 5.04 (t, 1 H, J 9.6 Hz), 4.94 (dd, 1 H, J, , 2 . 8.0 Hz, J 2 , 3 . 9.5 Hz, H-2'), 4.89 (br s, 1 H), 4.72 (br s, 1 H), 4.64 (d, 1 H, h\T 8.0 Hz, H-l'), 4.28 (dd, 1 H, J 4 , 5 a 4.8 Hz, J 5 a , 5 b 12.3 Hz, H-5a), 4.09 (dd, 1 H, J 4 , 5 b 2.3 Hz, J 5 a , 5 b 12.3 Hz, H-5b), 3.83 (t, 1 H, J 5.5 Hz), 3.75-3.50 (m, 4 H), 2.10-1.95 (m, 18 H, Ac). ES-LRMS: calc for C 3 1H 4 0NO, 6 (M + H+): 682.4. Found: 682.4. 2,3,4,6-Tetra-0-acetyl-p-D-glucopyranosyl-(1^ 4)-2,3,4,6,-tetra-0-acetyl-p-D-glucopyranosyl-(l->4)-2,4-di-(9-acetyl-A?-benzyloxycarbonyl-l-deoxyxylonojirimycin (2.67) Rf 0.14 (3:2 petroleum ether - ethyl acetate). 'H-NMR (400 MHz, CDC13) 8 7.35-7.20 (m, 5 H, Ph), 5.23-5.15 (br s, 1 H), 5.125, 5.12 (2 t, 2 H, J 5.9 Hz, J 5.9 Hz), 4.87 (m, 2 H), 4.68 (br s, 1 H), 4.61 (d, 1 H, J 8.0 Hz), 4.57 (d, 1 H, J 7.9 Hz), 4.56 (dd, 1 H, J 2.0 Hz, J 93 12.5 Hz), 4.33 (dd, 1 H, J 4.4 Hz, J 12.5 Hz), 4.09 (dd, 1 H, J 51 Hz, J 12.0 Hz), 4.02 (dd, 1 H, J 2.1 Hz, J 12.4 Hz), 3. 80 9t, 1 H, J 5.6 Hz), 3.77-348 (m, 7 H), 2.1-1.95 (m, 27 H, Ac). LSIMS-HRMS: calc for C43H56NO24 (M + H +): 970.31918. Found: 970.31955. 3.1.2.2 Transglycosylation Experiments with 1,1-Dideutero-N-benzyloxocarbonyl-l-deoxyxylonojirimycin (2.9) The deuterated benzyl carbamate 2.9 (48 mg, 0.26 mmol), a-glucosyl fluoride (43 mg, 0.28 mmol) were combined with AbgGlu358Ala (1.1 mg/ml) in 150 mM phosphate buffer, pH 7.5 (3.0 ml) and allowed to stand at room temperature for 3 days. The solution was diluted with methanol (20 ml), concentrated and the products were acetylated and isolated by column chromatography as described above: [3-1,3-linked disaccharide 2.64 (32.1 mg, 0.047 mmol, 18.0%), (3-1,3-linked trisaccharide 2.65 (6.0 mg, 0.006 mmol, 4.4%), (3-1,4-linked trisaccharide 2.70 (20.1 mg, 0.021 mmol, 14.8%), and (3-1,4-linked tetrasaccharide 2.71 (15.9 mg, 0.016 mmol, 13.2%). Total yield: 50.4% based on glycosyl donor. Structures of 2.65, 2.70 and 2.71 are tentatively assigned. 2,3,4,6-Tetra-(3-acetyl-P-D-glucopyranosyl-( l->3)-2,4-di-0-acetyl-/V-benzyloxycarbonyl-1,1-dideutero-1 -deoxyxylonoj irimycin. (2.64) R f 0.32 (1:1 petroleum ether - ethyl acetate). 'H (400 MHz, CDCI3): 5 7.35-7.25 (m, 5 H, Ph), 5.20-4.97 (m, 4 H, H-3, H-3', H-4', H-5'), 4.93 (t, 1 H, J 2 ' , 3 ' 9.0 Hz), 4.79 (br s, 1 H, H-2, H-4), 4.58 (d, 1 H, J , - , r 9.0 Hz, H-l'), 4.31 (dd, 1 H, J 5 a,5b 12.8 Hz, J 4 , 5 a 4.1 Hz, H-5a), 4.25-4.17 (br s, 1 H, PhCH2), 4.17-3.90 (br s, 1 H, PhCH2), 3.74-3.53 (m, 2 H, H-6a', H-6b'), 1.99 (dd, 1 H, J 4 , 5 b 10.1 Hz, H-5b), 2.10-1.95 (m, 18 H, Ac). LSIMS-HRMS: calc for 94 C 3iH38D 2NOi 6 (M + H+): 684.24724. Found: 684.24943. 23,4,6-Tetra-<3-acetyl-p-D-glucopyra^ (1 ->3)-2,4-di-0-acetyl-7V-benzyloxycarbonyl-1,1 -dideutero-1 -deoxyxylonojirimycin (2.65) Rf 0.22 (1:1 petroleum ether - ethyl acetate). 1 H-NMR (400 MHz, CDC13) 5 7.35-7.20 (m, 5 H, Ph), 5.13-4.98 (m, 4 H), 4.96-4.80 (m, 3 H), 4.80-4.72 (br s, 1 H), 4.52-4.40 (m, 2 H), 4.36 (dd, 1 H, J 4.3 Hz, J 12.3 Hz), 4.27 (dd, 1 H, J 4 , 5 a 5.4 Hz, J 5 a , 5 b 12.1 Hz, H-5a), 4.17-4.06 (br s, 1 H) 4.01 (dd, 1 H, J 2.3 Hz, J 12.4 Hz), 3.84-3.53 (m, 5 H), 2.77 (dd, 1 H, J4,5b 10.1 Hz, H-5b), 2.01-1.94 (m, 27 H, Ac). ES-LRMS: calc for C43H53D2N024Na (M + Na+): 995. Found: 995. LSIMS-LRMS: 995. 2,3,4,6-Tetra-0-acetyl-fi-D-glucopyranosyl-(l—>4)-2,3,6,-tri-(9-acetyl-P-D-glucopyranosyl-(1 -»4)-2,4-di-0-acetyl-Af-benzyloxycarbonyl-1 -1 -dideutero-1 -deoxyxylonojirimycin (2.70) Rf 0.18 (1:1 petroleum ether - ethyl acetate). 1 H-NMR (400 MHz, CDC13) 8 7.35-7.20 (m, 5 H, Ph), 5.23-5.15 (br s, 1 H), 5.125, 5.12 (2 t, 2 H, J 5.9 Hz, J 5.9 Hz), 4.87 (m, 2 H), 2.68 (br s, 1 H), 4.61 (d, 1 H, J 8.0 Hz), 4.57 (d, 1 H, J 7.9 Hz), 4.56 (dd, 1 H, J 2.0 Hz, J 12.5 Hz), 4.33 (dd, 1 H, J 4.5 Hz, J 12.5 Hz), 4.09 (dd, 1 H, J 5.1 Hz, J 11.9 Hz), 4.02 (dd, 1 H, J 2.1 Hz, J 12.4 Hz), 3. 80 9t, 1 H, J 5.6 Hz), 3.77-348 (m, 5 H), 2.1-1.95 (m, 27 H, Ac). LSIMS-LRMS: calc for C 4 3H 54D 2N0 24 (M + H+): 972.33143. Found: 972.33133. 2,3,4,6-Tetra-0-acetyl-b-D-glucopyranosyl-(l->4)-2,3,6,-tri-0-acetyl-b-D-glucopyranosyl-(l->4)-2,3,6,-tri-0-acetyl-b-D-glucopyranosyl-(l->4)-2,4-di-0-acetyl-N-benzyloxycarbonyl-1 -1 -dideutero-1 -deoxyxylonojirimycin (2.71) Rf 0.14 (1:1 petroleum ether - ethyl acetate). 1 H-NMR (400 MHz, CDC13) 8 7.45-95 7.20 (m, 5 H, Ph), 5.25-5.15 (br s, 1 H), 5.15-4.98 (m, 5 H), 4.92-4.73 (m, 4 H), 4.59 (d, 1 H, J 8.0 Hz), 4.50-4.34 (m, 4 H), 4.32 (dd, 1 H, J 4.4 Hz, J 12.4 Hz), 3.79 (t, 1 H, J 5.5 Hz), 3.77-3.50 (m, 6 H), 2.12-1.93 (m, 36 H, Ac). LSMS-LRMS: calc for CssHegDzNO^Na (M + Na+): 12823. Found: 1283. 3.2 Enzyme Kinetics 3.2.1 General Wild-type and mutant T. saccharolyticum P-xylosidases were isolated and purified by David Vocadlo, Department of Chemistry, UBC. Absorbance measurements were made on a Unicam 8700 UV-Visible spectrophotometer equipped with a circulating water bath. Phenyl P-D-xylopyranoside (PX) was provided by Dr. Lothar Ziser, UBC Department of Chemistry. Buffers and other substrates were purchased from Sigma Chemical Company and used without further purification. 3.2.2 Determination of Kinetic Parameters The rates of 3,4-dinitrophenyl-P-D-xyloside or phenyl-P-D-xyloside hydrolysis by wild-type P-xylosidase were measured at 37°C between pH 5.5 and 7.5 in 50 mM sodium phosphate or sodium citrate buffer in the presence of 0.01% BSA by spectrophotometrically observing the release of 3,4-dinitrophenolate at 400nm (e =11.05 cm'1 mM"1) or phenolate at 277 nm (e =1.063 cm"1 mM"1). Measurements were taken at six or more substrate concentrations ranging from 0.2 to 5 times the estimated K m value. Values of K m and kcat at each pH were determined by fitting the experimental data to the Michaelis-Menten equation using the program GraFit(75(9). 96 Kinetic parameters for T. saccharolyticum (3-xylosidase at pH 4.5, 5.0, 6.0, 6.5 and K values for 1-deoxyxylonojirimycin at these pH values were determined by Karen Rupitz. Kinetic parameters and Kj values for the Glul60Ala mutant of T. saccharolyticum P-xylosidasewere were also determined by Karen Rupitz using 3,4-DNPX as substrate. 3.2.3 Inhibition of T. saccharolyticum P-Xylosidase K; values for each inhibitor were determined by measuring rates at a single substrate concentration at or near its K m in the presence of different concentrations of the inhibitor. Measurements were taken at five or more inhibitor concentrations which were varied so as to bracket the observed Kj. The K; values were determined by plotting the experimental data in the form of a Dixon plot (rate1 versus inhibitor concentration); the slope and intercept of which were calculated by linear regression analysis. Assuming competitive inhibition, Kj was obtained from the intercept of the line with 1/Vm a x. The error associated with the linear fit was taken to represent experimental error. The K value for inhibition of T. saccharolyticum P-xylosidase by 1-deoxyxylonojirimycin was determined at pH 5.5 using phenyl p-D-xyloside as substrate. Measurements of V m a x and K m ( a p p ) were taken at a series of inhibitor concentrations so as to bracket the approximate Kj value. The value of Kj was determined by a direct fit of the experimental data to an expression for competitive inhibition using GraFit(730). 97 References 1. Tolan, J. S., Olson, D., and Dines, R. E. (1996) in Enzymes for Pulp and Paper Processing (Jeffries, T. W., and Viikari, L. , Eds.) pp 25-35, American Chemical Society, Washington, DC. 2. Goss, P. E. , Baker, M . A., Carver, J. P., and Dennis, J. W. (1995) Clinical Cancer Research 1, 935-944. \ 3. Fleet, G. W. J., Karpas, A., Dwek, R. A., Fellows, L. E. , Tyms, A. S., Petursson, S., Namgoong, S.K., Ramsden, N. G., Smith, P. W., Son, J. C , Wilson, F., Witty, D. R., Jacob, G. S., and Rademacher, T. W. (1988) FEBS Lett. 237, 128-132. 4. Tyms, A. S., and Taylor, D. 1. (1988) Journal of Antimicrobial Chemotherapy 22, 271-273. 5. Ratner, L . (1992) AIDS Research and Human Retroviruses 8, 165-173. 6. von Itzstein, M . , Wu, W.-Y., and Jin, B. (1994) Carbohydr. Res. 259, 301-305. 7. von Itzstein, M . , Wu, W. Y., and Kok, G. B. (1993) Nature 363, 418-423. 8. Wade, R. C. (1997) Structure 5, 1139-1145. 9. Silagy, C. A., Campion, K., Keene, O., Cooper, C , Bolton, P., Watts, R., Lincoln, P., Liaw, T., Melbourne, U. o., Narayan, K., Hospital, B. H , and Delooze, F. (1998) Lancet 352, 1877-81. 10. Sulzenbacher, G., Driguez, H , Henrissat, B., Schulein, M . , and Davies, G. J. (1996) Biochemistry 35, 15280-15287. 11. Henrissat, B., and Bairoch, A. (1993) Biochem. J. 293, 281-288. 12. Henrissat, B. (1991) Biochem. J. 280, 309-316. 13. Sinnott, M . L . (1990) Chem. Rev. 90, 1171-1202. 14. Koshland, D. E. (1953) Biol. Rev. 28, 416-436. 15. Legler, G. (1990) Adv. Curb. Chem. Biochem. 48, 319-384. 16. McCarter, J. D., Adam, M . J., and Withers, S. G. (1992) Biochem. J. 286, 721-727. 17. Namchuk, M . N., and Withers, S. G. (1995) Biochemistry 34, 16194-16202. 18. Anderson, W. F., Grutter, M . G., Remington, S. J., Weaver, L. H., and Matthews, B. W. (1981)/. Mol. Biol. 147,523. 19. Blake, C. C. F., Mair, G. A., North, A. C. T., Philips, D. C , and Sharma, V. R. (1967) Proc. R. Soc. London Ser. B 167, 365. 20. Jacobsen, R. H., Zhang, X. J., Dubose, R. F., and Matthews, B. W. (1994) Nature 369, 761-764. 21. Derewenda, U., Swenson, L. , Green, R., Wei, Y., Morosoli, R., Shareck, F., Kluepfel, D., and Derewenda, Z. S. (1994) J. Biol. Chem. 269, 20811-2-814. 98 22. White, A., Withers, S. G., Gilkes, N. R., and Rose, D. R. (1994) Biochemistry 33, 12546-12552. 23. Davies, G., and Henrissat, B. (1995) Structure 3, 853-859. 24. Sulzenbacher, G., Schulein, M . , and Davies, G. J. (1997) Biochemistry 36, 5902-5911. 25. McCarter, J. D., and Withers, S. G. (1994) Curr. Opin. Struct. Biol. 4, 885-892. 26. Davies, G. J., Mackenzie, L. , Varrot, A., Dauter, M . , Brzozowski, A. M . , Schulein, M. , and Withers, S. G. (1998) Biochemistry 37, 11707-11713. 27. Macleod, A. M . , Lindhorst, T., Withers, S. G., and Warren, R. A. J. (1994) Biochemistry 35,6371-7376. 28. Kempton, J. B., and Withers, S. G. (1992) Biochemistry 31, 9961-9969. 29. Tao, B. Y., Reilly, P. J., and Robyt, J. F. (1989) Biochim. Biophys. Acta 995, 214-220. 30. Herrchen, M . , and Legler, G. (1984) Eur. J. Biochem. 138, 527-53. 31. Gebler, J. C , Aebersold, R., and Withers, S. G. (1992) J. Biol. Chem 267, 11126-11130. 32. Withers, S. G., Street, I. P., Bird, P., and Dolphin, D. H. (1987) J. Am. Chem. Soc. 109, 7530-7531. 33. Withers, S. G., Rupitz, K., and Street, I. P. (1988) J. Biol. Chem. 263, 7929-7932. 34. Withers, S. G., and Aebersold, R. (1995) Protein Sci. 4, 361-372. 35. Withers, S. G., Warren, R. A. J., Street, I. P., Rupitz, K., Kempton, J. B., and Aebersold, R. (1990) J. Am. Chem. Soc. 112, 5887-5889. 36. Withers, S. G., and Street, I. P. (1988) J. Am. Chem. Soc. 110, 8551-8553. 37. Bols, M . (1998) Acc. Chem. Res. 31, 1-8. 38. Street, I. P., Kempton, J. B., and Withers, S. G. (1992) Biochemistry 31, 9970-9978. 39. Braun, C , Brayer, G. D., and Withers, S. G. (1995) J. Biol. Chem. 270, 26778-26781. 40. McCarter, J. D., and Withers, S. G. (1996) J. Am. Chem. Soc. 118, 241-242. 41. Withers, S. G., Namchuk, M . , and Mosi, R. (1999) in Imino Sugars as Glycosidase Inhibitors (Stiitz, A. E. , Ed.) pp 188-206, Wiley-VCH, New York. 42. Wolfenden, R. (1969) Nature 223, 704-705. 43. Pauling, L . (1946) Chem. Eng. News 24, 1375-1377. 44. Wolfenden, R. (1976) Annu. Rev. Biophys. Bioeng. 5, 271-306. 45. Wolfenden, R. (1972) Acc. Chem. Res. 5, 10-18. 46. Lienhard, G. E. (1972) Annu. Rep. Med. Chem. 7, 249-258. 47. Ganem, B., and Papandreou, G. (1991) J. Am. Chem. Soc. 113, 8984-8985. 48. Ermert, P., Vasella, A., Weber, M . , Rupitz, K , and Withers, S. G. (1993) Carbohydr. Res. 250, 113-128. 99 49. Ezaki, S. (1940) J. Biochem. (Tokyo) 32, 104-111. 50. Horikoshi, K. (1942) J. Biochem. (Tokyo) 35, 39-42. 51. Hackert, M . L. , and Jacobson, M . I. (1971) Acta. Crystallogr., Ser. B 27, 203-209. 52. Walaszek, Z., Horton, D., and Ekiel, I. (1982) Carbohydr. Res. 106, 193-201. 53. Nishikawa, T., and Ishida, N. (1965) / . Antibiot. Ser. A. 18, 132-133. 54. Stilts, A. E . (1999) in Imino Sugars as Glycosidase Inhibitors (Stuts, A. E . , Ed.) pp 157-185, Wiley - V C H , New York. 55. Vasella, A., Ermert, P., Hoos, R., Naughton, A. B., Rupitz, K., Thiel, W., Weber, M . , Weber, W., and Withers, S. G. (1994) in Complex Carbohydrates In Drug Research (Bock, K., and Clausen, H. , Eds.) pp 134-150, Copenhagen. 56. Mader, M . M . , and Bartlett, P. A. (1997) Chem. Rev. 97, 1281-1300. 57. Namchuk, M . , Braun, C , McCarter, J. D., and Withers, S. G. (1996) in Biomedical Frontiers of Fluorine Chemistry pp 279-293. 58. Namchuk, M . (1994) Ph.D. Thesis, University of British Columbia. 59. Dale, M . P., Ensley, H. E. , Kern, K., Sastry, K. A. R., and Dyers, L . D. (1985) Biochemistry 24, 3530-3530. 60. Morrison, J. F., and Walsh, C. T. (1988) Adv. Enzymol. Relat. Areas. Mol. Biol. 61, 201-301. 61. Legler, G. (1999) in Imino Sugars as Glycosidase Inhibitors (Stiitz, A. E. , Ed.) pp 31-67, Wiley - V C H , Weinhein. 62. Schweden, J., Borgmann, C , Legler, G., and Bause, E. (1986) Arch. Biochem. Biophys. 248, 335-340. 63. Segal, P., Feig, P. U., Schernthaner, G., Ratzmann, K. P., Rybka, J., Petzinna, D., and Berlin, C. (1997) Diabetes Care 20, 687-91. 64. Wolffenbuttel, B. H., and Graal, M . B. (1996) Postgrad Med J 72, 657-62. 65. Karpas, A., Fleet, G. W. J., Dwek, R. A., Petursson, S., Namgoong, S. K., Ramsden, N. G , Jacob, G. S., and Rademacher, T. W. (1988) Proc. Natl. Acad. Sci. USA 85, 9229-9233. 66. Tan, A., van den Broek, L. , van Boeckel, S., Ploegh, H., and Bolscher, J. (1991) J. Biol. Chem. 266, 14504-14510. 67. Hughes, A. B., and Rudge, A. J. (1994) Natural Product Reports, 135-162. 68. Johnson, C. R., and Johns, B. A. (1997) J. Org. Chem. 62, 6046-6050. 69. Ichikawa, Y., Igarashi, Y., Ichikawa, M . , and Suhara, Y. (1998) J. Am. Chem. Soc. 120, 3007-3018. 70. Ichikawa, Y., and Igarashi, Y. (1995) / . Am. Chem. Soc. 36,4585-4586. 71. Jespersen, T. M . , dong, W., Sierks, M . R., Skrydstrup, T., Lundt, I., and Bols, M . (1994) Angew. Chem. Int. Ed. Engl. 33, 1778-1779. 100 72. Bols, M . , Hazell, R. G., and Thomsen, I. B. (1997) Chem. Eur. J. 3, 940-947. 73. Dong, W., Jespersen, T., Bols, M . , Skrydstrup, T., and Sierks, M . R. (1996) Biochemistry 35, 2788-2795. 74. Kajimoto, T., Liu, K. K. C , Pederson, R. L. , Zhong, Z., Ichikawa, Y., Jr, J. A. P., and Wong, C. H. (1991) J. Am. Chem. Soc. 113, 6191-6202. 75. Hadfield, A. T., Harvey, D. J., Archer, D. B., MacKenzie, D. A., Jeenes, D. J., Radford, S. E. , Lowe, G., Dobson, C. M . , and Johnson, L. N. (1994) J. Mol. Biol. 243, 856-872. 76. Beer, D., and Vasella, A. (1986) Helv. Chim. Acta. 69, 267-270. 77. Hoos, R., Vasella, A., Rupitz, K., and Withers, S. G. (1997) Carbohydr. Res. 298, 291-298. 78. Hoos, R., Naughton, A. B., Thiel, W., Vasella, A., Weber, W., Rupitz, K., and Withers, S. G. (1993) Helv. Chim. Acta. 76, 2666-2686. 79. Papageorgiou, A. C., Oikonomakos, N. G., and Leonidas, D. D. (1989) Arch. Biochem. Biophys. 272, 376-385. 80. Therisod, H., Letourneux, Y., and Therisod, M . (1998) Bioorg. Med. Chem. Lett. 8, 371-372. 81. Vonhoff, S., Heightman, T. D., and Vasella, A. (1998) Helv. Chim. Acta. 81, 1710-1725. 82. Papandreou, G., Tong, M . K., and Ganem, G. (1993) J. Am. Chem. Soc. 115, 11682-11690. 83. Legler, G., and Finken, M . T. (1996) Carbohydr. Res. 292, 103-115. 84. Pan, Y. T., Kaushal, G. P., Papandreou, G., Ganem, B., and Elbein, A. D. (1992) J. Biol. Chem. 267, 8313-8318. 85. Ermert, P., and Vasella, A. (1991) Helv. Chim. Acta 74, 2043-2053. 86. Panday, N., Granier, T., and Vasella, A. (1998) Helv. Chim. Acta. 81, 475-490. 87. Granier, T., Panday, N., and Vasella, A. (1997) Helv. Chim. Acta. 80, 979-987. 88. Heightman, T. D., Ermert, P., Klein, D., and Vasella, A. (1995) Helv. Chim. Acta 78, 514-532. 89. Heightman, T. D., Locatelli, M . , and Vasella, A. (1996) Helv. Chim. Acta 79, 2190-2200. 90. Bleriot, Y., Genre-Grandpirre, A., Imberty, A., and Tellier, C. (1996) J. Carbohydr. Chem. 15, 985-1000. 91. Withers, S. G., Rupitz, K., Trimbur, D., and Warren, R. A. J. (1992) Biochemistry 31, 9979-9985. 92. Thiem, J. (1995) FEMS Microbiol. Rev. 16, 193-211. 93. Gijsen, H., Qiao, L. , Fitz, W., and Wong, C. H. (1996) Chem. Rev. 96, 443-473. 94. Cote, G. L. , and Tao, B. Y. (1990) Glycoconj. J. 7, 145-162. 101 95. Prade, H., Mackenzie, L. F., and Withers, S. g. (1998) Carbohydr. Res. 305, 371-381. 96. Kawaguchi, T., Sugimoto, K., Hyashi, H., and Arai, M . (1996) Biosci. Biotech. Biochem. 60, 344-346. 97. Mackenzie, L . F., Wang, Q., Warren, R. A. J., and Withers, S. G. (1998) J. Am. Chem. Soc. 120, 5583-5584. 98. Wang, Q., Graham, R. W., Trimbur, D., Warren, R. A. J., and Withers, S. G. (1994) J. Am. Chem. Soc. 116, 11594-11595. 99. Joseleau, J. P., Comtat, J., and Ruel, K. (1992) in Xylans and Xylanases (Visser, J., Beldman, G., Someren, M . A. K. v., and Voragen, A. G. J., Eds.) pp 1-15, Elsevier, Amsterdam. 100. Kirk, T. K., and Jeffries, T. W. (1996) in Enzymes for Pulp and Paper Processing (Jeffries, T. W., and Viikari, L . , Eds.) pp 2-14, American Chemical Society. 101. Armand, S., Vieille, C , Gey, C , Heyraud, A., Zeikus, J. G., and Henrissat, B. (1996) Eur. J. Biochem 236, 706-713. 102. Lee, Y. E. , and Zeikus, J. G. (1993) J. Gen. Microbiol. 139, 1235-1243. 103. Vocadlo, D. J., Mackenzie, L . F., He, S., Zeikus, G. J., and Withers, S. G. (1998) Biochem. J. 335, 449-455. 104. Vocadlo, D. (1999) personal communication. 105. Ziser, L. , and Withers, S. G. (1994) Carbohydr. Res. 265, 9-17. 106. Ziser, L. , Setyawati, L, and Withers, S. G. (1995) Carbohydr. Res. 274, 137-153. 107. Godskesen, M . , Lundt, I., Madsen, R., and Winchester, B. (1996) Bioorg. Med. Chem. 4, 1857-1865. 108. Kisfaludy, L . , and Laszlo Otvos, J. (1987) Synth. Commun., 510. 109. Kisfaludy, L. , Mohacsi, T., Low, M . , and Drexler, F. (1979) / . Org. Chem. 44, 654-656. 110. Williams, S.( 1999) unpublished results. 111. Mitchell, E. P., Withers, S. G , ERmert, P., Vasella, A. T., Garman, E. F., Oikonomakos, N. G., and Johnson, L. N. (1996) Biochemistry 35, 7341-7355. 112. Granier, T., Gaiser, G., Hintermann, L. , and Vasella, A. (1997) Helv. Chim . Acta. 80, 1443-1456. 113. Wessel, H.-P., Iversen, T., and Bundle, D. R. (1985) / . Chem. Soc. Perkin Trans. 1, 2247-2250. 114. Iversen, T., and Bundle, D. R. (1981) J. Chem. Soc. Chem. Comm., 1240-1241. 115. Jensen, H. S., Limberg, G., and Pedersen, C. (1997) Carbohydr. Res. 302, 109-112. 116. Scriven, E . F. V. , and Turnbull, K. (1988) Chem. Rev. 88, 351-367. 117. Gajda, T., and Zwierzak, A. (1981) Synth. Commun., 1005-1008. 102 118. Koziara, A., Zawadzki, S., and Zwierzak, A. (1979) Synthesis, 527-529. 119. Vonhoff, S., and Vasella, A. (1999) personal communication. 120. Malet, C , and Planas, A. (1998) FEBS Letters 440, 208-212. 121. Puis, J., Borchmann, A., Gottschalk, D., and Wiegel, J. (1988) Meth. Enzymol. 160, 528-536. 122. Bray, M . R., and Clarke, A. J. (1990) Biochem. J. 270, 91-96. 123. Hirsch, J., and Kovac, P. (1982) Chem. Zvesti 36, 125-131. 124. Kovac, P. (1980) Chem. Zvesti 34, 234-240. 125. Mechaly, A., Belakhov, V., Shoham, Y., and Baasov, T. (1997) Carbohydr. Res. 304, 111-115. 126. Dixon, M . , Webb, E. C , Thorn, C. J. R., and Tipton, K. F. (1979) Enzymes, 3 ed., Academic Press, New York. 127. Cornish-Bowden, A. (1976) Principles of Enzyme Kinetics, Butterworths, Toronto. 128. Knowles, J. R. (1976) CRC. Critical Reviews Biochemistry, 165-172. 129. Jacqueline Wicki. (1999) unpublished results. 130. Leatherbarrow, R. J. (1994) Erithacus Software Ltd, London, U.K. 131. Eyring, H. (1935) Chem. Rev. 17, 65-77. 132. Morgan, D. P., Scholtz, J. M . , Ballinger, M . , Zipkin, I., and Bartlett, P. A. (1991) J. Am. Chem. Soc. 113, 297-307. 133. Bartlett, P. A., and Marlowe, C. K. (1983) Biochemistry 22, 4618-4624. 134. Hanson, J. E. , Kaplan, A. P., and Bartlett, P. A. (1989) Biochemistry 28, 6294-6305. 135. Bartlett, P. A., and Giangiordano, M . A. (1996) J. Org. Chem. 61, 3433-3438. 136. Mosi, R. (1998) Ph.D. Thesis, University of British Columbia. Appendix A: Transition State Analogy 104 A Probing Transition state Analogy Many compounds have been put forth as transition state analogues for the enzyme which they inhibit. The fact is that even when an inhibitor has been carefully designed to display important aspects of the proposed transition state, strong inhibition alone is not confirmation that the compound is binding through mimicry of the transition state. Most "transition state analogues" have been labelled as such simply by rationalisation of their tight binding(56). Confirmation of transition state analogy requires a rigorous approach, which will be discussed here. For the sake of clarity, a brief introduction to transition state theory as it relates to enzyme catalysis will also be presented. A.l Transition state Theory Our current understanding of transition state theory is based on Eyring's proposals(iii) and elaborated upon by many workers including Wolfenden, Lienhard and Bartlett(44, 46, 56). In a chemical reaction, a high energy species referred to as an "activated complex" (or transition state) is formed. This species is said to exist in a pseudo equilibrium with the ground state form of the reactants and products, and the rate of reaction is dictated by the rate at which this highest energy transition state decomposes to products as described in the Eyring equation (Equation A.l) . Equation A. 1 k = KvK* The rate of reaction (k) is a function of the pseudo equilibrium constant between the transition state and ground state (K*), the transmission coefficient (K) which is near unity in most cases, and the frequency of normal vibration of the scissile bond (v). K* in turn is 105 related to the energy difference between the ground state substrate and transition state (AG*) by Equation A.2. Equation A.2: K t = e ( - A G * / R T ) Transition state theory has been applied to enzyme catalysed reactions, where catalysis is achieved by preferential binding of the transition state over the ground state conformation of the substrate(s). Although enzyme catalysis is a multistep process the relationship between the catalysed and uncatalysed reactions can be expressed in its simplest form through the thermodynamic cycle shown in Figure A. 1(44, 45). The dissociation constants K S , KTS, and pseudo equilibrium constants and ATcat* are related according to Equation A.3. E+'S E-S E + SH m K. r E-S^ TS E + P E + P Figure A. 1: The thermodynamic cycle Equation A.3: K TS Kun Acat Recognising that KCJ and KUn are proportional to kcat and kun respectively by the Eyring equation (Equation A.l) and rearranging gives an estimate of the dissociation constant of the transition state (KTS) as a function of substrate binding (Ks), the catalysed (kcat) and uncatalysed (kun) reaction rates as expressed in Equation A.4. If we assume that the 106 Michaelis constant, K m , is a reasonable approximation of Ks, then binding of the transition state can be calculated from Equation A.5. A perfect transition state analogue would bind a factor of k c a t /k u n better than the substrate. • Equation A.4: K T S = K s u n ^cat Equation A.5: K T S = Ki = K ^cat A.2 Modified Substrates and Inhibitors to Probe Transition state Mimicry Wolfenden recognised an important consequence of the relationships described above: the inhibition constant (Kj) of any inhibitor can be related to the transition state dissociation constant (KTs) by Equation A.6, where d is a simple numerical constant. This can be re-expressed in logarithmic form as shown in Equation A.7. Since log(d ku n) is a constant, log Kj and log (K m /k c a t ) are directly related(44). Equation A.6: K; = d • K T S = d • K m — ^cat Equation A.7: Log Ki = log 1c V c a t J + log(d-k u n) As a result, structural changes which alter binding of the transition state and thereby affect the reaction rate should have the same effect on the binding of a transition state analogue. This provides a means to test transition state mimicry. If the same alterations are made in a series of substrates and a series of inhibitors, and the resulting kinetic parameters 107 are plotted as log Kj versus log (Km/k c a t), then a linear relationship with a slope of unity will be seen, if the inhibitors do indeed mimic the transition state. This assumes that the structural alterations do not have an effect on the uncatalysed reaction, and that the rate determining step in the catalysed process is unchanged. This approach has been applied to a number of enzyme systems, including thermolysin(i 52, 133), carboxypeptidase A(134), and pepsin(755). The correlation of kinetic parameters for a series of inhibitors and substrates of pepsin is shown in Figure A . l . As expected for a series of transition state analogues, there is a linear relationship between log Kj and log(Km/k c a t) in Figure A.la. The lack of a meaningful correlation between log Kj and log K m in Figure A. lb indicates that these inhibitors are not binding through substrate analogy (755). -7 -6 -5 -4 -3 -2 -5 -4 -3 l o g ^ / k a , (M-8) l O g l ^ O M ) Figure A.2: a) log Kj versus log (Km/k c a t) for a series of substrates and phosphonic acid inhibitors of pepsin, b) log K, versus log K m for the same series(reprinted from(755)). Transition state mimicry by 1-deoxynojirimycin and castanospermine on Agrobacterium sp. P-glucosidase was also tested in this way(77, 28). Kinetic data from a 108 number of substrate analogues and correspondingly modified inhibitors was collected. As shown in Figure A.3 the plots of ln(l/Kj) versus ln(kcat/Km) for castanospermine and 1-deoxynojirimycin derivatives showed no correlation in either case, indicating that these two inhibitor types are not transition state analogues. Instead, their binding is the result of interactions between a positively charged nitrogen centre and a negatively charged residue in the active site(47). A) B) 0 2 4 6 8 10 0 2 4 6 8 10 In (k c a/KJ ln (k c a/KJ Figure A.3 Linear free energy relationships for binding of a series of correspondingly substituted DNP glycosides and inhibitors with Abg P-glucosidase. A) ln (1/KO for substituted 1-Deoxynojirimycins versus (ln kcat/Km) for DNP glycosides, B) ln (1/Kj) for substituted Castanospermines versus (ln kcat/Km) for corresponding DNP glycosides, (reproduced from data in(5S)). A.3 Probing Transition state Mimicry Using Enzyme Mutants An alternative approach which takes advantage of the same relationships involves the preparation of a series of enzyme mutants rather than a collection of modified substrates and inhibitors. By mutating amino acid residues which are important to binding and catalysis, or have an effect on the conformation of the active site, the environment will be changed. Changes in the chemical environment should have the same influence on transition state 109 analogue binding and reaction rate. This approach is advantageous as it allows a greater range of inhibitors to be tested without the laborious preparation of analogous substrates and inhibitors. It also avoids a change in the uncatalysed mechanism that may occur with poorly chosen modified substrates and inhibitors. This technique was used to demonstrate that acarbose (Figure A.4) is a transition state analogue of cyclodextrin glycosyltransferase (CGTase) from Bacillus circulans(136). Mosi and coworkers screened this inhibitor against a collection of CGTase active site mutants using a-D-glucosyl fluoride and cc-D-maltotrisyl fluoride as substrates(7.36). Figure A.4 Acarbose. As shown in Figure A.5, there was good correlation between log(Km/kcat) and log Ki for acarbose using a-glucosyl fluoride as substrate. A slope of 2.2 and correlation of r = 0.98 was observed. That the slope is greater than unity does indicate less than perfect analogy, but these results nevertheless indicate a substantial amount of transition state mimicry by acarbose. A correlation between log Ki and log Km for the substrate also indicates some substrate analogy (Figure A.5). Either by the preparation of suitably modified substrates and inhibitors or using enzyme mutants it is possible to confirm transition state mimicry by a putative analogue. 110 Which approach is preferable depends on the availability of enzyme mutants with measurable activity, the ease of preparation of modified inhibitors and substrates as well as other aspects specific to the case at hand. A B log K, log K, Figure A.5 Linear free energy relationship between kinetic parameters for the inhibitor acarbose and a-glucosyl fluoride as substrate with a series of mutants of CGTase. A) log (Km/kc a t) for a-glucosyl fluoride vs log Kj for acarbose.; B) log K m for aGF vs log K; for acarbose. (Reprinted from(735)). Appendix B: Graphical Representation of Kinetic Data 112 B .3 | S 2 4 6 8 10 [Phenyl Xyloside] m M 70 60 50 40 30 20 10 0 1 ^* ' 1 1 1 1 1 1 1 1 1 1 1 s 1 1 1 1 1 I 1 -20 -10 0 10 20 30 40 50 60 [l-Deoxyxylonojirimycin] | i M Figure B. 1 B Pi D 40 30 20 10 1 1 1 1 . 1 1 / 1 . : i : i i i : -100 0 100 200 300 400 [l-Deoxyxylonojirimycin] uM -20 0 20 40 60 80 100 [l-Deoxyxylonojirimycin] u M A) Determination of K m and V r a a x for T. saccharolyticum p-xylosidase using PX as substrate at pH 5.5, 50 mM Citrate, 0.01% BSA, T = 37°C. B-D) Range finder Kj plots for T. saccharolyticum P-xylosidase inhibition by 1-deoxyxylonojirimycin at various pH values. In all experiments: 0.01% BSA, T = 37°C.B) pH 5.5, 50 mM Citrate, [Enz] = 0.046 mg/ml, [PX] = 2.1 mM C) pH 7.0, 50 mM NaPi [Enz] = 0.046 mg/ml, [PX] = 4.4 mM. D) pH 7.5, 50 mM NaPi, [Enz] = 0.046 mg/ml, [PX] = 4.4 mM. 113 B .s oi 1> 500 1000 1500 2000 [Z-1-deoxyxylonojirimycin] u M -200 0 200 400 600 [Xylonojirimycin Tetrazole] u M Oi D 22 oi -2 0 2 4 6 [Xylonolactam] m M 10 -i i i i 1 i i i i i 1 i T l M l M , , i i i i 1 i i i i 1 i i i • 1 i 1 1 1 1 -10 -5 0 5 10 15 20 [N-Benzyl-ldeoxyxylonojirimycin] mM c E [A^-Acetyl-l-deoxyxylonojirimycin] m M .S E [A'-formyl-l-deoxyxylonojirimycin] m M Figure B. 2 K; plots for T. saccharolyticum |3-xylosidase with various inhibitors. In all experiments pH = 5.5, 50 mM Na Citrate, 0.01% BSA, T = 37°C. A) 7V-Benzyloxycarbonyl-1-deoxyxylonojirimycin: [Enz] = 0.033 mg/ml, [PX] = 2.1 mM. B) D-Xylonolactam: [Enz] = 0.0138 mg/ml, [PX] = 2.1 mM. C) D-xylonojirimycin tetrazole: [Enz] = .031 mg/ml, [PX] = 2.1 mM. D)/V-Benzyl-1-deoxyxylonojirimycin: [Enz] = .033 mg/ml, [PX] = 2.6 mM. E) TV-Acetyl-1-deoxyxylonojirimycin: [Enz] = .033 mg/ml, [PX] = 4.6 mM. F) TV-Formyl-l-deoxyxylonojirimycin: [Enz] = .031 mg/ml, [PX] = 2.6 mM.