Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Human [alpha]-L-iduronidase : substrate synthesis and mechanistic analysis Nieman, Catharine E. 2000

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

Item Metadata

Download

Media
831-ubc_2000-0505.pdf [ 5.1MB ]
Metadata
JSON: 831-1.0061396.json
JSON-LD: 831-1.0061396-ld.json
RDF/XML (Pretty): 831-1.0061396-rdf.xml
RDF/JSON: 831-1.0061396-rdf.json
Turtle: 831-1.0061396-turtle.txt
N-Triples: 831-1.0061396-rdf-ntriples.txt
Original Record: 831-1.0061396-source.json
Full Text
831-1.0061396-fulltext.txt
Citation
831-1.0061396.ris

Full Text

H U M A N a-L-IDURONIDASE: SUBSTRATE SYNTHESIS AND MECHANISTIC ANALYSIS By Catharine E. Nieman B.Sc., McMaster University, 1997 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A September 2000 © Catharine E. Nieman, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of (^/[f/n/^4f(j The University of British Columoia Vancouver, Canada Date DE-6 (2/88) A B S T R A C T Glycosidases are enzymes responsible for hydrolysis of glycosidic bonds. On the basis of sequence sirnilarities, glycosidases have been assigned to a series of families, one of which, Family 39, contains both P-D-xylosidases and a-L-iduronidases. However, the level of sequence similarity between these two enzymes is relatively low, calling into question mechanistic conclusions made about the iduronidases on the basis of results from xylosidases. By being assigned to a family of retaining glycosidases, human a-L-iduronidase (IDUA) is also predicted to operate with net retention of anomeric configuration, and sequence alignments suggest that E299 is the catalytic nucleophile. When 4-methylumbelliferyl-a-L-idopyranosiduronic acid was incubated with IDUA in the presence of 3 M methanol, *H NMR analysis of the resulting products revealed the presence of methyl a-L-idopyranosiduronic acid, and a complete absence of the P-methanolysis product, thus proving IDUA to be a retaining glycosidase. In order to identify the catalytic nucleophile, 5-fluoro-a-L-idopyranosyluronic acid fluoride (5F-IdoAF) was used to label the enzyme via trapping of the intermediate. MS analysis of a peptic digest of an assay mixture containing IDUA and 5F-IdoAF permitted the identification of the labeled peptide 2 9 1 A D T P I Y N D E A D P L V G 3 0 5 , having m/z = 893 ± 1. Further analysis revealed that the label was attached to E299, proving it to be the catalytic nucleophile. This, together with the proof of retention of anomeric configuration, supports the assignment of IDUA to Family 39. ii To further probe the mechanism of this enzyme, we required a series of aryl a-L-iduronides of varying reactivities. Because the glycosylation chemistry of ido compounds is poorly understood, a methodology study was performed in which five glycosylation methods were tested for their ability to form the desired anomeric stereochemistry with a range of acceptors of varying pKa. The trichloroacetimidate method was found to produce the best results, and a series of aryl a-L-iduronides was synthesized. Kinetic parameters for their hydrolysis by LOUA were determined, and it was found that both kcai and kcai/Km were independent of leaving group pKa- This indicates that either glycosylation is rate-determining, with extremely efficient proton donation at the transition state, or that some other step, such as conformational change is rate-determining. iii T A B L E O F C O N T E N T S ABSTRACT ii T A B L E OF CONTENTS iv LIST OF FIGURES vii LIST OF TABLES ix LIST OF ABBREVIATIONS x ACKNOWLEDGEMENTS xi DEDICATION xii CHAPTER I INTRODUCTION 1 1.1 Human a-L-iduronidase 1 1.1.1 Glycosidases 2 1.1.2 The biological function of iduronidase 4 1.1.2.1 The natural iduronidase substrate 4 1.1.2.2 Biological role of iduronic acid 6 1.1.2.3 Mucopolysaccharidosis Type I 7 1.1.3 Glycosidase Family Assignment 8 1.1.4 Mechanistic analysis of retaining glycosidases 9 1.1.4.1 Possible mechanisms of retaining glycosidases 9 1.1.4.2 Oxacarbenium ion-like transition state 12 1.1.4.3 Mechanism-based inactivators 14 1.1.4.4 Linear free energy relationships 15 1.2 Substrate Synthesis 17 1.2.1 Synthetic strategies 18 iv 1.2.2 Glycosylation techniques 21 1.2.3 Selective Oxidation of C6 .....23 1.3 Objectives... 23 CHAPTER H RESULTS AND DISCUSSION: SYNTHESIS 25 2.1 Introduction 25 2.2 Large scale synthesis of L-idose 25 2.3 Glycosylation reactions 28 2.3.1 Cyanoethylidene glycosylation 28 2.3.2 Sulfoxide glycosylation 34 2.3.3 Silver triflate glycosylation 38 2.3.4 Trichloroacetimidate glycosylation 41 2.3.5 Fluorodinitrobenzene glycosylation 44 2.3.6 Summary of glycosylation reactions 45 2.4 Deprotection 46 2.5 Selective oxidation of C6 48 2.6 Conclusions and future work 53 CHAPTER m RESULTS AND DISCUSSION: ENZYMOLOGY 54 3.1 Introduction ...54 3.2 Stereochemistry of iduronidase-catalyzed reaction 54 3.2.1 Enzyme-catalyzed substrate hydrolysis 55 3.2.2 Enzyme-catalyzed substrate methanolysis 61 3.3 Identification of the catalytic nucleophile 64 3.3.1 Sequence alignments 65 3.3.2 Kinetics 67 v 3.3.3 Mass spectrometric analysis 74 3.4 Probing the rate-determining step 79 3.4.1 Brensted correlation 79 3.4.2 The effect methanol addition on rate 83 3.5 Conclusions and Future Work 84 CHAPTER IV MATERIALS AND METHODS 86 4.1 Synthesis 86 4.1.1 General materials and methods 86 4.1.2 General Procedures 88 4.1.2.1 General procedure - deprotection 88 4.1.2.2 General procedure - oxidation 88 4.1.3 Large scale synthesis of idose 89 4.1.4 Methodology study 95 4.1.4.1 Cyanoethylidene method 95 4.1.4.2 Sulfoxide method 98 4.1.4.3 Silver trifluoromethanesulfonate method 100 4.1.4.4 Trichloroacetimidate method 101 4.1.5 Large scale glycosylation reactions 102 4.1.6 Deprotections 108 4.1.7 Oxidations 110 4.2 Enzymology 113 4.2.1 General Procedures 113 4.2.2 Stereochemistry experiments 114 4.2.3 Sequence alignments 115 4.2.4 Kinetics with 5F-IdoAF 115 4.2.5 Labeling and mass spectrometric analysis 116 4.2.6 Linear free energy relationships 117 REFERENCES 119 vi LIST OF FIGURES Figure 1.1 Reaction catalyzed by a generic glycosidase 2 Figure 1.2 a and |3 descriptors for D and L sugars 4 Figure 1.3 Structure of the pre-heparin basic dimer unit 5 Figure 1.4 Proposed mechanism for a retaining a-L-iduronidase using an enzyme active site nucleophile 10 Figure 1.5 Proposed mechanism of jack bean N-acetyl-B-glucosaminidase 11 Figure 1.6 Possible mechanism of a retaining a-L-iduronidase using the C6 carboxylate as the nucleophile 12 Figure 1.7 conformations of glucuronic acid, xylose, and iduronic acid 14 Figure 1.8 Formation of iduronic acid from glucuronic acid by C5 epimerization 18 Figure 1.9 Formation of penta-O-acetyl-a-D-idopyranose from penta-O-acetyl-P-D-glucopyranose. 19 Figure 1.10 Baggett synthesis of penta-0-acetyl-a,|3-L-idopyranose 20 Figure 1.11 Konigs-Knorr glycosylation produces 1,2-trans glycosides as a result of anchimeric assistance 22 Figure 2.1 Alternate proposed synthesis of L-idose .26 Figure 2.2 Enforced formation of 1,2-trans glycosides 28 Figure 2.3 Mechanism of cyanoethylidene glycosylation 29 Figure 2.4 1,3-Diaxial interactions prevent the formation of the endo isomer of 3,4,6-tri-O-acetyl-1,2-0-( l-cyanoethylidene)-p-L-idopyranose 31 Figure 2.5 Formation of Phenyl 2,3,4,6-tetra-O-acetyl-l-sulfinyl-a-L-idopyranoside... 35 Figure 2.6 Mechanism of glycosylation by the sulfoxide method .37 Figure 2.7 Scavenging the phenylsulfenyl triflate intermediate of the sulfoxide glycosylation reaction 38 Figure 2.8 1,2-Trans-diaxial orientation of both P-D-mannosyl and a-L-idosyl bromides 39 Figure 2.9 Possible conformational changes of a-L-idosyl bromide preventing C2 acetoxy participation during Koenings-Knorr glycosylation reactions 40 Figure 2.10 Koenigs-Knorr-type chemistry using a soluble silver salt 41 vii Figure 2.11 Formation of Phenyl-2,3,4,6-tetra-0-acetyl-a-L-idopyranoside by the McWoroacetimidate method. 43 Figure 2.12 Formation of 2,4-dinitrophenyl glycosides using fluorodinitrobenzene 44 Figure 2.13 Crystal structure of 4'-chlorophenyl a-L-idopyranoside 47 Figure 2.14 Mechanism of TEMPO/HOBr oxidation 50 Figure 3.1 Monitoring the stereochemical outcome of glycosidase-catalyzed substrate hydrolysis by *H NMR 55 Figure 3.2 *H NMR analysis of the iduronidase-catalyzed substrate hydrolysis 57 Figure 3.3 Geometry of glycosidase active sites 60 Figure 3.4 Mutarotation and pyranose-furanose interconversion pass through a common intermediate 61 Figure 3.5 Representation of tic obtained from iduronidase-catalyzed MUI hydrolysis in the presence of varying concentrations of methanol 63 Figure 3.6 *H NMR analysis of iduronidase-catalyzed MUI hydrolysis in the presence of 3 M methanol 64 Figure 3.7 Primary sequence alignments of the family 39 glycosidases in the regions flanking the predicted nucleophiles and acid/base catalysts 66 Figure 3.8 Testing 5F-IdoAF as a time-dependent inactivator of iduronidase .68 Figure 3.9 Testing 5F-IdoAF as a competitive inhibitor of iduronidase 69 Figure 3.10 Testing 5F-IdoAF as a substrate 72 Figure 3.11 5-Fluoro glycosides release 2 fluoride ions per hydrolysis event 73 Figure 3.12 MS/MS analysis of daughter ions of the labeled peptide .76 Figure 3.13 Closer examination of MS/MS analysis of daughter ions of the labeled peptide 78 Figure 3.14 Br0nsted plots for a-L-iduronidase 81 Figure 3.15 Br0nsted plots for a-L-iduronidase alongside data for Agrobacterium {J-D-glucosidase 82 Figure 3.16 Dependence of rate on methanol concentration at saturating substrate concentration 84 viii LIST OF TABLES Table 2.1 Summary of results of large scale glycosylation reactions 45 Table 2.2 Summary of results of deprotection reactions 48 Table 2.3 Summary of results of oxidation reactions 52 Table 3.1 Daughter ions observed after fragmentation of labeled peptide, assuming cleavage at (C=0)-N bonds 77 Table 3.2 Kinetic parameters for a series of substrates of varying reactivities 80 ix L I S T O F A B B R E V I A T I O N S 5F-IdoAF 5-fluoro-a-L-idopyranosyluronic acid fluoride A angstrom Ac acetyl Asp, D aspartate CS chondroitin sulfate DABCO 1,4-diazabicyclo [2.2.2]octane D B U 1,8-diazabiclyclo [5.4.0]undec-7-ene D C E 1,2-dichloroethane DMF TV.iV-dimethyl formamide DNA deoxy ribonucleic acid DS dermatan sulfate DTBMP 2,6-di-fert-butyl-4-methyl pyridine E + electrophile E C M extracellular matrix endo-H endo-P-N-acetylglycosaminidase H FDNB fluoro-2,4-dinitrobenzene G A G glycosaminoglycan GalNAc N-acetyl-D-galactosamine GlcNAc N-acetyl-D-glucosamine GlcA D-glucuronic acid Glu,E glutamate H A hyaluronic acid HPLC high pressure liquid chromatography IdoA Iduronic acid IDUA human a-L-iduronidase MPS mucopolysaccharidosis MS mass spectrometry Ms methane sulfonyl MUI 4-Methylumbelliferyl-a-L-idopyranosiduronic acid NMR nuclear magnetic resonance spectroscopy Nu nucleophile PET positron emission tomography PNP 4-nitrophenol Py pyridine T C A trichloroacetimidate TEMPO 2,2,6,6-tetramethylpiperidinyl-1 -oxy tic thin layer chromatography Tr triphenylcarbenium xyl xylosidase x ACKNOWLEDGEMENTS I would like to express my appreciation to Dr. Stephen Withers for giving me the opportunity to work in his laboratory, for his guidance, insight and enthusiasm, and for providing me with a project that was both extremely challenging and interesting. I would also like to thank Dr. Robert Stick for his guidance and expertise in the synthetic aspects of this work in Steve's absence. I would like to thank all of the members of the Withers lab for their friendship, support and innumerable helpful discussions. Particularly, I would like to thank Alex Wong for providing the inactivator and for his insights into the chemistry of ido compounds, Shouming He for his work on the mass spectrometric characterization of the labeled peptide, David Vocadlo and Harry Brumer for their support and expertise, Hoa Ly for making it possible for me to have written my thesis from Michigan, and Karen Rupitz for her technical expertise. The work done by Dr. Lawrence Mcintosh on the proton NMR analysis of substrate hydrolysis is gratefully acknowledged, as is the crystallography work done by Brian Patrick. A special thanks is extended to the members of Dr. Edward Piers' lab for their synthetic expertise and insights, particularly to Shawn Walker, Sebastien Caille and Dr. James Nieman. I would like to thank Dr. John Hopwood and Dr. Lome Clarke for their collaboration on this project and for providing me with purified enzyme. xi For Jim and for my parents xii CHAPTER I INTRODUCTION 1.1 Human a-L-iduronidase It has long been known that carbohydrates play a pivotal role in the natural world. It has been estimated that in excess of 60 % of the carbon in the biosphere exists in the form of carbohydrate, much of which is found in cellulose, the primary structural material of plants. Carbohydrates have also long been studied in their role of bridging the gap between solar and chemical energy. In the process of photosynthesis, energy from the sun is used to combine carbon dioxide and water to form molecular oxygen and carbohydrate. These carbohydrates, and derivatives thereof, provide a key energy source for both plants and animals, through the processes of glycolysis and respiration. Only recently have the more subtle roles of carbohydrates begun to be realized and understood. Preceding one of the most famous scientific discoveries of the 20 t h century, the determination of the structure of DNA, it was revealed that the information storage molecule for all of life employs a carbohydrate-based polymer as its backbone. In cell membranes, anywhere from 5 to 40 % of the lipids are glycosylated, depending on the cell type(7), and while the purpose of glycolipids remains unclear, there are indications of involvement in cell-cell communication. Glycosylation of proteins can fundamentally influence many processes, including folding and assembly, targeting to the proper subcellular organelle and, when considering cell-surface glycoproteins, the carbohydrate moiety may also play a role in cell-cell communication^). Large polysaccharide 1 structures (most of which are covalently linked to proteins) comprise a large portion of the extracellular matrix in higher organisms, and are quite abundant in tissues such as skin, cartilage, cornea, and bone, providing strength, elasticity and lubrication. In addition, these complex structures also influence cell development, migration, shape, proliferation and metabolic function. In light of the vital and diverse roles of carbohydrates, it is not surprising that the enzymes involved in the formation and degradation of carbohydrate structures are becoming increasingly targeted in modern biochemical research. Of particular interest in this study is the human lysosomal glycosidase, a-L-iduronidase (3.2.1.76), an enzyme implicated in the fatal disease Hurler syndrome. With the rapidly rising interest in treating the disease by enzyme-replacement therapy, there is a concurrent increase in interest in the enzyme's catalytic mechanism. The investigation of this mechanism will be the focus of this work. 1.1.1 Glycosidases Glycosidases, or glycosyl hydrolases are enzymes that hydrolyze glycosidic linkages, cleaving the exocyclic C-0 bond that connects the anomeric carbon to the aglycone (R in Figure 1.1). ROH glycone aglycone Figure 1.1 Reaction catalyzed by a generic glycosidase 2 Glycosidases are classified according to three criteria: i) stereochemistry at the anomeric centre of the substrate: a-glycosidases hydrolyze glycosides with a-glycosidic linkages and, likewise, (3-glycosidases hydrolyze substrates having 3-linkages; ii) stereochemistry at the anomeric centre of the product, relative to the substrate: if the stereochemistry at the anomeric centre of the product is the same as in the substrate, the glycosidase is said to be retaining; if the stereochemistry is inverted, the glycosidase is said to be inverting. iii) glycone specificity: most glycosidases have a preferred substrate with which they interact most strongly, and the basis of that preference is derived primarily from interactions between the enzyme and the glycone portion of the substrate; the degree of specificity varies greatly from one glycosidase to the next. Iduronidase is an unusual example of an a-glycosidase, in that its substrate is more reminiscent of a (3-glycoside than an a-glycoside (Figure 1.2). This is because the glycone moiety of its substrate is an L-sugar, thus the a and B descriptors appear to be different from those applied to the more common D-sugars. Although the required anomeric configuration of the substrate is clear, the stereochemical outcome of the enzyme-catalyzed substrate hydrolysis has yet to be determined. Iduronidase has an extremely strong preference for a-L-iduronic acid (4) as the glycone moiety. Indeed, it has not been found to hydrolyze any substrates other than iduronides. Work done by Clements et al. revealed that the most important recognition features of the substrate are the C 6 carboxylate and the stereochemistry of C5(3). 3 a-D-glucuronic acid (1) p-D-glucuronic acid (2) p-L-iduronic acid (3) a-L-iduronic acid (4) Figure 1.2 a and P descriptors for D and L sugars. Note that the ido compounds are not shown in their lowest energy conformations. 1.1.2 The biological function of iduronidase Unlike many other glycosidases that are widely distributed in Nature, to date, the only known iduronidases are found in mammalian lysozymes. This would indicate that both the enzyme and its substrate play a highly evolved and very specific role in the organisms in which they are found. 1.1.2.1 The natural iduronidase substrate The extracellular matrix (ECM) is a complex tissue consisting of proteins, polysaccharides, glycoproteins and glycosaminoglycans (GAG's). Of particular interest to this study are the GAG's heparin, heparan sulfate and dermatan sulfate. Presently there is no agreed-upon criteria for distinguishing between heparin and heparan sulfate, therefore the term "heparin" will be used to represent both in this dissertation. 4 ,0 NH1 [GlcAp(1 -»4)GlcNAca(1 -»4)]n Figure 1.3 Structure of the pre-heparin basic dimer unit Heparin originates as a polymer of alternating D-glucuronic acid and D-N-acetyl glucosamine residues, linked as shown in Figure 1.3. A number of modifications to this polymer occur to generate the mature heparin. The GlcNAc residues can be N-deacetylated and N-sulfated. Subsequently, some, but not all, of the GlcA residues are epimerized at C5, to form L-iduronic acid (note that this changes the notation of the glycosidic linkage from P to a). Finally, the polymer is O-sulfated on C2 of IdoA residues and C6 of GlcN residues. Rare O-sulfation events can occur on C3 of GlcN residues and C2 or C3 of GlcA residues(^). Dermatan sulfate also originates as a polymer of disaccharides, but the constituents are glucuronic acid and N-acetyl galactosamine ([GlcAP(l-^3)GalNAcP(l-^4)]n). The modifications to this polymer to generate the mature form are similar to those of heparin, but N-deacetylation does not occur(5). The net effect of these modifications in both heparin and dermatan sulfate is the formation of highly complex polysaccharides. 5 1.1.2.2 Biological role of iduronic acid The conversion of D-glucuronic acid to L-iduronic acid within these polymers has interesting consequences. Research has found that GAG's containing IdoA have a much wider range of biological activity than those having only GlcA as their uronic acid component, such as chondroitin sulfate (CS) and hyaluronic acid (HA)(<5). The biological activities of these IdoA-containing GAG's include interactions with such proteins as thrombin, basic fibroblast growth factor, platelet factor 4, fibronectin and histones; also, interactions with whole cells affect such complex functions as cell growth and differentiation. These biological activities have been attributed to ionic interactions between the negatively charged GAG's and positively charged residues on the proteins. However, because CS and HA, which also contain negatively charged carboxylate and sulfate moieties but do not show any of these activities, the biological activity cannot be attributed to merely ionic interactions. In contrast to GlcA, which has a decided preference to sit in a 4 C i conformation(7), iduronic acid is a sugar with an extraordinarily large number of conformations having similar energy levels(c?-72), and has been found to exist in a lC4 or a 4 C i conformation, or even the 2So skew-boat within G A G polymers. The preferred conformation of the IdoA residue within the polysaccharide is influenced primarily by the degree of sulfation and the surrounding residues. It is proposed that the flexibility of IdoA within the polysaccharide chains accounts for the major difference in biological activities of different GAG's(73). 6 1.1.2.3 Mucopolysaccharidosis Type I Like most other tissues, the E C M is constantly undergoing breakdown and regeneration. The extracellular GAG's destined for catabolism are taken up by the cells via receptor-mediated endocytosis, and the resulting intracellular vesicles are fused with lysosomes, where degradation occurs (reviewed m(J4)). There are a large number of lysosomal glycosidases, each specific for a different glycosidic linkage. Because these enzymes operate sequentially, the lack of a specific enzyme results in the build-up of that enzyme's substrate, and an inability of the subsequent enzymes to perform their function. Known as lysosomal storage disorders, or mucopolysaccharidoses, these are degenerative genetic disorders that are detected in 1 in 10 000 live births. Mucopolysaccharidosis type I (MPS I) is a lack of a-L-iduronidase, the enzyme responsible for cleaving a-L-iduronic acid from the non-reducing ends of polysaccharide chains. The disease is observed at a frequency of 1 in 50 000 live births. In its most severe form, the disease is known as Hurler's syndrome. The symptoms of the disease include mental retardation, organomegaly, corneal opacity and skeletal deformities (dwarfism). Those afflicted usually do not survive beyond ten years of age. Patients suffering from Scheie syndrome, a less severe form of the disease, have normal intelligence and life expectancy, but still display the symptoms of skeletal deformities and corneal opacity. Currently, the only available treatment for the disease is allogenic bone marrow transplantation. This treatment is limited, however, by the availability of donors. Research into a more readily available treatment, in the form of enzyme-replacement 7 therapy is underway using a knock-out mouse model(75) and a naturally occuring canine model(7<5). 1.1.3 Glycosidase Family Assignment In spite of the push towards using iduronidase as a therapeutic agent, very little is known about the activity of iduronidase on a mechanistic level. One of the first clues about the mechanism of a glycosidase is derived from its assignment to a glycosidase family. On the basis of amino acid sequence homology, Henrissat has divided all known glycosidases into a set of families(77, 18), and enzymes grouped together usually share a common mechanism. In 1990, the iduronidase gene was mapped to chromosome 4pl6.3 (19, 20), and in 1991, the cDNA was isolated and cloned(27). With the protein sequence in hand, iduronidase was assigned to Glycosyl Hydrolase Family 39. It is a small family, currently composed only of ot-L-iduronidases and a small number of (3-D-xylosidases, with only 8 members in total. Interestingly, a high degree of sequence similarity is found among the iduronidases and also among the xylosidases, but only a low level of similarity between the two groups of enzymes, making this family assignment somewhat suspect. Family 39 glycosidases are predicted to be retaining enzymes. The first direct evidence of this was found in 1996 when Armand et al. examined the stereochemical outcome of the reaction catalyzed by a family 39 xylosidase isolated from Thermoanaerobacterium saccharolyticum(22), and found that the reaction proceeds with net retention of anomeric configuration. The assignment of iduronidase to the same family as another retaining enzyme implies, but does not prove, that iduronidase is also 8 retaining. The low sequence similarity in the region encompassing the predicted nucleophile brings this assumption into question. 1.1.4 Mechanistic analysis of retaining glycosidases Once the stereochemical outcome of the glycosidase-catalyzed reaction is known, it is possible to further analyze the catalytic mechanism. This often involves the use of inhibitors and/or inactivators, and a variety of substrates. 1.1.4.1 Possible mechanisms of retaining glycosidases In 1953, Koshland(23) proposed that retaining glycosidases use a two-step mechanism in which the stereochemistry at the anomeric centre is inverted twice, resulting in net retention of stereochemistry. In the first step, a nucleophile in the enzyme active site (an aspartate or glutamate residue) attacks the anomeric centre, displacing the aglycone and forming a covalent glycosyl-enzyme intermediate. In the second step, water attacks the anomeric centre, displacing the enzyme nucleophile, regenerating the enzyme and producing the free sugar (Figure 1.4). In addition to the nucleophile, a second Asp or Glu residue is conserved in the active sites of nearly all known retaining glycosidases. This residue acts as a general acid catalyst in the first step of hydrolysis, protonating the leaving aglycone, and a general base catalyst in the second step, deprotonating the water, thus enhancing its nucleophilicity. 9 Figure 1.4 Proposed mechanism for a retaining a-L-iduronidase using an enzyme active site nucleophile 10 Although most retaining glycosidases use an enzyme active-site nucleophile, there are a few exceptions. There is substantial evidence suggesting that jack bean N-acetyl-B-glucosaminidase employs an anchimeric assistance mechanism in which the C2 acetamido functionality acts as the nucleophile, forming an oxazoline intermediate^-^), which then undergoes hydrolysis (Figure 1.5). Figure 1.5 Proposed mechanism of jack bean N-acetyl-B-D-glucosaminidase(24) The iduronidase substrate also contains a potential nucleophile: the C6 carboxylate of iduronic acid. The presence of this functionality and the poor sequence alignment around the proposed nucleophiles of iduronidases and xylosidases makes anchimeric assistance an alternate possibility for the mechanism of this enzyme (Figure 1.6). 11 Figure 1.6 Possible mechanism of a retaining a-L-iduronidase using the C6 carboxylate as the nucleophile 1.1.4.2 Oxacarbenium ion-like transition state As shown in Figures 1.4 and 1.6, regardless of the overall mechanism, both propose passage through oxacarbenium ion-like transition states. Positive charge that develops on 12 the anomeric centre as the aglycone begins to depart is stabilized, in part, by the endocyclic oxygen of the sugar, imparting partial double-bond character to the Cl -05 bond. This requires a conformational change in the sugar in order to allow the anomeric centre to take on sp2 character. The sp2 character of the transition state has been demonstrated by the use of kinetic isotope effects in a number of glycosidases(25, 26). The conformation adopted at the transition state is such that 05, C5, C l and C2 are all in the same plane. This can be achieved in one of two ways. In most glycosidases, it is believed that the sugar adopts a flattened half chair conformation. There are, however, two examples of a sugars taking on the unusual conformation of a 2,5-boat (2'5B) when bound to the enzyme: Bacillus circulans xylanase and Bacillus agaradhaerens xylanase, both from glycosidase family 11(27, 28). The reason this conformation is so seldom observed, or even considered as a reasonable transition state geometry is that most common sugars (D-glucose, D-galactose, D-mannose) possess an equatorial C5 hydroxymethyl group (or in the case of uronic acids, a carboxylate group). In a 2 , 5 B conformation, this substituent would be forced into a pseudo-axial orientation causing destabilizing flagpole interactions with the C2 hydrogen. Xylose (5) does not have a C5 substituent, and thus there is no such highly unfavourable interaction. While no crystal structure data are available for any of the family 39 enzymes, it is quite possible that the xylosidases in this family also induce a 2 ' 5 B conformation of the substrate in the transition state. Interestingly, while L-idose (6) and L-iduronic acid (4) possess a C5 substituent, in the 2 ' 5 B conformation, the substituent will be pseudo-equatorial, thus will not contribute to destabilizing flagpole interactions (Figure 1.7). It is possible that these unique 13 properties and the conformational flexibilities of ido and xylo compounds are major contributing factors to the homology of these two types of seemingly unrelated enzymes. H O H H O H H O H p-D-glucuronic acid (2) p-D-xylose (5) a-L-iduronic acid (4) Figure 1.7 2 , 5 B conformations of D-glucuronic acid, D-xylose, and L-iduronic acid 1.1.4.3 Mechanism-based inactivators An important advance in the study of retaining glycosidases has been the development of mechanism-based inactivators (2-deoxy-2-fluoro or 5-fluoro glycosides), which take advantage of the oxacarbenium ion-like transition state. The substitution of fluorine for hydrogen or hydroxyl adjacent to either the anomeric centre or the endocyclic oxygen increases transition state energies for both the glycosylation and deglycosylation steps of the hydrolysis reaction by inductively destabilizing the forming oxacarbenium ion. By placing a very good leaving group (fluoride or 2,4-dinitrophenol) at the anomeric centre of either type of inactivator, the glycosylation step is accelerated, but the rate of the deglycosylation step is unaffected. The net result is accumulation of the covalent glycosyl-enzyme intermediate. 5-Fluoro glycosides are generally effective inhibitors of both a - D - and P-D-glycosidases. In contrast, the 2 deoxy-2-fluoro glycosides tend to inactivate only P-D glycosidases. It is thought that this is because the partial positive 14 charge developed by B-glycosidases is localized primarily on the anomeric carbon, whereas for a-glycosidases, the partial positive charge is more evenly distributed over both the anomeric centre and the endocyclic oxygen, thus the 2-F substitution maximally affects B-glycosidases. These fluoro-sugars have been used to label and identify the catalytic nucleophiles of several glycosidases(29-3<5). I*1 several of these cases, the glycosyl-enzyme intermediate was quite long-lived, having a half-life on the order of hours to days. Using these inactivators, the possibility exists to track enzymes within living organisms. By synthesizing 18F-labeled inactivators, and introducing these compounds into a test animal, the biodistribution can be observed through the use of positron emission tomography, also known as "PET" imaging. This technology has tremendous potential in the study of enzyme replacement therapy as a treatment for MPS-1. Presently, the only means available for tracking the distribution of enzyme in vivo is to sacrifice the test subjects and analyze various tissues for iduronidase activity. PET imaging, coupled with the use of 18F-labeled inactivators, has the potential of eliminating the need to euthanize laboratory animals for this purpose. 1.1.4.4 Linear free energy relationships Linear free energy relationships employ structure-activity correlations to probe reaction mechanisms. When systematic perturbations in the structure of the reactants cause changes in the equilibrium or rate of the reaction in question, evidence concerning the rate-detenrdning step(s), existence of intermediates and other mechanistic features can often be obtained. A linear free energy relationship exists if systematic changes in reactant structure (such as altering the electronegativity of substrate substituents) change 15 the rate (k) or equilibrium (K) of one reaction in a way that is closely correlated in a second reaction, modified only by a factor that can be attributed to the difference in the two substrates. That is, log(kX)2/k0,2) = log(kx,i/ko,i) x constant (Equation 1.1) and/or log(Kx,2/Ko,2) = logOKjci/Kci) x constant (Equation 1.2) where kx and K x refer to the rate and equilibrium, respectively, of the reaction for the substituted compound, and ko and Ko refer to the unsubstituted compound. Because log(k) is the free energy of activation, and log(K) is the standard free energy of the reaction, the term "linear free energy relationship" is appropriate. The strength of the linear free energy relationship as a tool for elucidating reaction mechanism lies in the correlation that is sometimes observed between rate (k) and equilibrium (K). When represented graphically, this correlation fits the equation: log(k/ko) = m log(K/Ko) (Equation 1.3) where the slope, m, expresses the degree of charge development in the transition state, with a positive slope representing an electron deficient transition state, and a negative slope representing an electron sufficient transition state. The details regarding the history and derivation of linear free energy relationships can be found in most physical organic chemistry texts; many useful references and reviews of the subject are referenced within §2.2 of Lowry and Richardson's text(37). The Brensted catalysis law describes a specific type of linear free energy relationship involving proton transfer in which the rate of proton donation is directly related to the acidity of the proton donor, or 16 k = C K a a (Equation 1.4) or log(k) = a logfK*) + log(C) (Equation 1.5) where k is the rate of dissociation, is the acid dissociation constant, and C is a constant of proportionality. The proportionality constant a indicates the degree to which the rate of dissociation is dependent upon K»; the closer a becomes to unity, the more complete proton transfer is at the transition state. Glycoside hydrolysis is generally thought to involve general acid catalysis with concurrent breakage of the glycosidic linkage. In this scenario, the general acid catalyst remains constant (a carboxylic acid residue in the enzyme active site), but the conjugate base (the aglycone moiety of the substrate) can vary. In this scenario, one would investigate a relationship between the rate of hydrolysis and the basicity of the leaving group, rather than the acidity of the general acid. Because acidity and basicity are intimately related, one can continue to use K a in equations 1.3 and 1.4. The major difference between the glycosidase scenario and the original application of the Brensted equation is the meaning of a. In both cases, it represents the dependence of rate on K a , however, in the case of glycoside hydrolysis, the more complete the proton transfer is at the transition state, the closer the value of a becomes to zero (i.e. the less influence leaving group basicity has on the rate). 1.2 Substrate Synthesis Presently, there is only one commercially available substrate for a-L-iduronidase: methylumbelliferyl-a-L-iduronide. However, in order to study the enzyme on a 17 mechanistic level, a variety of substrates and inactivators/inhibitors are required. This clearly makes it necessary to investigate the synthesis of a variety of substrates for this enzyme. 1.2.1 Synthetic strategies There are two potential approaches to the synthesis of the desired substrates. Perhaps the most direct route to the desired a-L-iduronides is one that mimics Nature's methodology. Iduronic acid can be obtained from glucuronic acid by inversion of C5, which, in Nature, is done within G A G polymers by D-glucuronyl C5-epimerase(4). In the lab, this has been accomplished through photobromination of C5 of protected B-D-glucuronides, followed by reductive inversion, using tributyltin hydride(35) (Figure 1.8). Although the yield for the reduction step is quite low (-30%), the remaining 70% is primarily the starting glucuronide, and can be cycled through the reaction process again. The major drawback to this strategy is the final deprotection of the acid functionality. Cleaving esters requires relatively harsh conditions, under which the glycosidic bonds of many of the labile iduronides we are interested in would be cleaved. 7 8 9 Figure 1.8 Formation of L-iduronic acid from D-glucuronic acid by C5 epimerization 18 O A c O A c 11 Figure 1.9 Formation of penta-O-acetyl-ct-D-idopyranose (11) from penta-0-acetyl-B-D-glucopyranose (10) Alternatively, it is possible to start with idose, form the glycoside, deprotect, and, as the final step, selectively oxidize C6 to a carboxylic acid. Idose, however, is neither a commercially available nor a naturally occurring sugar, and must, therefore, be synthesized. a-D-Idose pentaacetate (11) is readily obtained in very large quantities. Starting with B-D-glucose pentaacetate (10) in methylene chloride, the addition of a catalytic amount of antimony pentachloride causes a series of acetate migrations around 19 the sugar ring. Every stereocentre on the ring is inverted with the exception of C5, producing the D-ido product, which is less soluble than the other intermediates and crystallizes out of solution, thus driving the equilibrium of acetoxonium salts to the ido form. (Figure 1.9). Hydrolysis of the acetoxonium salt and subsequent acetylation produces a-D-idose penta-O-acetate. This is, unfortunately, the enantiomer of the required starting material. The same reaction can be performed on ct-L-glucose to form L-idose, but the cost of sufficient quantities of L-glucose is, at present, prohibitively high. 1) NaOMe, MeOH 2) HCI, rfcO.lOO °C 3) Ac 2 O,Py,0 °C AcOH 18 19 Figure 1.10 Baggett synthesis of penta-0-acetyl-a,B-L-idopyranose 20 The most direct (and financially feasible) route to large quantities of L-idose is a procedure that starts with D-glucurono-6,3-lactone (12), and produces the desired compound in 9 steps, and 20% overall yield(39) (Figure 1.10). The hydroxyl groups on C l and C2 are protected in the form of a ketal, leaving only the C5 hydroxyl group free. This group is then converted to a tosylate, which, after reductive lactone opening and protection of the resulting alcohols, is displaced with acetate in an SN2 manner, thus epimerizing C5 and converting the compound from the D-gluco to the L-ido form. Rearrangement to the pyranose form yields the desired compound. 1.2.2 Glycosylation techniques One of the greatest challenges of making a series of iduronides, once the starting sugar is in hand, is the formation of the glycosidic linkage with the correct stereochemistry. One of the most widely used methods for making 1,2-trans glycosides is known as the Koenigs-Knorr method(</0). A glycosyl halide, having a participating group protecting the C2 hydroxyl, is reacted with a silver salt and the desired acceptor. As a halide-silver bond begins to form, and the halide is pulled from the anomeric centre, its departure is assisted by the endocyclic oxygen, forming an oxacarbenium ion intermediate. The C2 acetoxy group, being at a much higher local concentration than the acceptor, intercepts the intermediate, forming the more stable acetoxonium ion intermediate. This directs the acceptor to attack at the anomeric centre from the opposite side, enforcing the formation of the 1,2-trans glycoside (Figure 1.11). 21 Figure 1.11 Koenigs-Knorr glycosylation produces 1,2-trans glycosides as a result of anchimeric assistance Work previously done in the Withers lab by Evelyn Rodriguez, in agreement with literature data, shows that, under Koenigs-Knorr conditions, L-idose does not form the desired 1,2-trans glycosides, but, rather, the undesired 1,2-cw glycosides(47) or the orthoester(3c9). a-L-Idosides have been most consistently synthesized by a method known as a melt. In this method, the acceptor is melted in the presence of either a Lewis or Brensted acid, and the peracetylated sugar is added. The reaction is run under vacuum to remove the acetic acid liberated during the reaction. Using this technique, Baggett reports yields in the neighbourhood of 40% for the a-anomer, and 5% for the B-anomer(3°). While this is an improvement over the complete absence of the desired anomer that results from Koenigs-Knorr chemistry, this technique is limited to acceptors that are thermally stable and have sufficiently low melting points that the procedure doesn't cause the thermal 22 decomposition of the donor sugar. These criteria exclude many of the low pKa acceptors of interest. Also, the resulting anomeric mixture creates the need for tedious purification of the desired idoside. 1.2.3 Selective Oxidation of C6 The requirements for the final step of the synthesis are quite stringent. The oxidation must be selective for C6, leaving the remaining three free hydroxyl groups untouched, and the conditions of the reaction must be mild enough to leave the glycosidic bond intact. While this may seem a daunting task, there is a very useful procedure that has been proven successful for several different sugars(42-44). The reaction, which is usually complete in only 30 minutes, uses 2,2,6,6-tetramethylpiperidinyl-l-oxy (TEMPO), sodium hypochlorite and sodium bromide, and pH 10-11 carbonate buffer at 0 °C. de Nooy and Besemer obtained quite impressive results with both glucose and galactose, forming the corresponding uronic acids in excess of 90%, irrespective of anomeric configuration and substituent(42). 1.3 Objectives The focus of this project will be on the characterization of human a-L-iduronidase. We intend to determine whether the substrate hydrolysis it catalyzes proceeds with net retention or inversion of anomeric configuration. If we find that the enzyme is, indeed, a retaining glycosidase, we hope to be able to identify the catalytic nucleophile, whether it is an enzyme active site residue or the C6 carboxylate of the substrate itself. We would 23 like to further our understanding of this enzyme by examining the rate-deterniining step of the enzyme-catalyzed substrate hydrolysis. In order to achieve this final goal, we will require a series of substrates of varying reactivities. Because the current proven methods for synthesizing iduronides are not applicable to many of the more reactive compounds that we would like to synthesize, a detailed study of the glycosylation chemistry of L -idose will be necessary. 24 CHAPTER II R E S U L T S A N D D I S C U S S I O N : S Y N T H E S I S 2.1 Introduction With only one substrate commercially available for iduronidase, the extent to which the kinetic mechanism can be investigated is quite limited. Substrates for iduronidase present a significant synthetic challenge as L-idose, the required starting material, is not commercially available and glycosylation reactions with this sugar are poorly understood. The work detailed within this chapter addresses these issues. 2.2 Large scale synthesis of L-idose While there are many published syntheses of idose(5P, 45-47), the synthesis developed by Baggett et al. (Figure 1.10) was chosen due to its relative apparent ease of application in large scale synthesis. Because this is a well-documented reaction sequence, I will limit the discussion to the few steps I found to be problematic. The only step in the sequence that causes serious difficulties is the actual C5 epimerization step, in which a tosylate is replaced with an acetate via an SN2 mechanism, thus inverting the stereochemistry from the D-gluco to the h-ido form. This reaction requires heating to 140 °C in acetic anhydride for 48 hours. Not surprisingly, these harsh conditions cause significant thermal decomposition. In addition, the starting material and product have nearly identical chromatographic behaviour, making both monitoring the reaction and purifying the product extremely difficult. 25 Figure 2.1 Alternate proposed synthesis of L-idose It was thought that the main cause of the problems encountered in this step were primarily due to a combination of the poor nucleophilicity of acetate and the mediocre leaving group ability of tosylate. Replacement of the tosylate with a better leaving group, such as triflate or mesylate presents the possibility of using milder conditions for the reaction. However, this also creates the possibility of 5,6-epoxide formation during reductive lactone opening. The alternative is to perform the displacement reaction earlier in the synthesis, while the lactone is still intact. Then, reduction of the lactone with lithium borohydride will ensure that the C3 acetate is also reduced, forming 1,2-0-isopropylidene-B-L-idofuranose (20), which can be transformed in two steps to L-idose pentaacetate (19), as in the Baggett synthesis. Not only would this proposal eliminate the problematic step, it would also shorten the synthesis by three steps (Figure 2.1). 26 The mesylate was readily formed in 74 % yield. However, the reactants at the proposed inversion step deteriorated extremely rapidly (approximately 2 minutes) into a complex mixture of products, and this route was abandoned. The final step of the Baggett synthesis is the acetolysis and acetylation of 1,6-anhydro L-idose (18) (Figure 1.10). The expected final product is L-idose penta-0-acetate (19). However, when this reaction was performed on a large scale, a mixture of products was obtained: L-idose penta-O-acetate, as we had expected, as well as the 2,3,4,6-tetra-O-acetate (23). Analysis of the reaction mixture by tic showed only a single product, indicating that the second product formed during the workup procedure. Normally, once the reaction is complete, water is added, and the mixture is stirred on ice to hydrolyze the acetic anhydride. The solution is then extracted repeatedly with dichloromethane, after which the pooled organic layers are neutralized by washing with saturated sodium bicarbonate. Because of the large volumes involved, after the first sodium bicarbonate wash, I found it necessary to divide the dichloromethane layer into two portions. While the first portion was being neutralized, the second portion remained in dichloromethane containing acetic acid, sulfuric acid and water. These conditions are sufficient for anomeric deprotection, and this was likely the cause of the product mixture. Had it not been necessary to divide the dichloromethane layer during workup, this problem would likely have been avoided. However, it was readily rectified by reacetylation of the deprotected portion of the product using acetic anhydride/pyridine. 27 2.3 Glycosylation reactions Key to the project was the development of glycosylation reactions that would not only give the desired anomeric configuration, but also be applicable to acceptors having a wide range of pKa values. With very little information in the literature regarding glycosylation reactions with idose, a thorough investigation of possible glycosylation methods was necessary. 2.3.1 Cyanoethylidene glycosylation In the late '70's and early '80's, Kochetkov developed a novel strategy for synthesizing disaccharides having exclusively a 1,2-trans configuration (reviewed in (48)). In order to obtain absolute stereospecificity in the reaction, he developed a synthesis in which the donor exists as a stable bicyclic structure, with the C2 protecting group (normally an acetate) linked to the anomeric centre(4P). By using R = C N (Figure 2.2), the donor is sufficiently stable that the collapse of the 5-membered ring requires electrophilic attack at the cyanide group, concurrent with nucleophilic attack at the anomeric centre, in effect, a "push-pull" mechanism (Figure 2.2), thus enforcing the desired 1,2-trans configuration of the resulting glycoside. Figure 2.2 Enforced formation of 1,2-trans glycosides 28 Triphenylcarbenium salts were used as the electrophiles, most commonly as their perchlorate or tetrafluoroborate salts. In non-carbohydrate applications, hydroxyl groups have been used successfully as the nucleophiles(50). However, the strong acids that are formed as byproducts (HCIO4, HBF4) make this problematic for systems containing acid-sensitive functionalities. Kochetkov solved this problem by protecting the nucleophile as a trityl ether. The resulting byproduct, rather than being a strong acid, was simply the regenerated triphenylcarbenium salt. This had the added advantage of allowing the use of only catalytic amounts of the salt. The other byproduct, triphenylcarbenium cyanide (TrCN), formed by the reaction of the initial byproduct (TrNC) with another molecule of the trityl salt, was found to crystallize out of solution, thus driving the reaction forward (Figure 2.3)(48). Figure 2.3 Mechanism of cyanoethylidene glycosylation 29 While Kochetkov had great success with this chemistry using D-glucose, D-galactose, D-mannose(49), and their corresponding uronic acids(57), this chemistry had not been attempted with idose. The first concern was whether the initial cyanoethylidene could be formed. Per-O-acetylated L-idose was transformed to the bromide, as described in §4.1.4.1, with a nearly quantitative yield. The crystalline product was dissolved in dry acetonitrile, to which recrystallized, dry potassium cyanide was added. After 24 hours at room temperature, the desired bicyclic product (24) was isolated. While Kochetkov reports yields of 91%, 95% and 78% for the gluco, galacto and manno compounds, respectively, we were unable to reproduce these results with the gluco compound, obtaining only -60%, and the ido compound was obtained in only 59% yield. As a distinctive *H NMR resonance characteristic of a cyanoethylidene product, Kochetkov cites a slight upfield shift (0.1 - 0.3 ppm) of a single methyl group, relative to the methyl groups of the acetates. Examination of the NMR data for the L-idocyanoethylidene (24) also showed an upfield shift for one methyl group. However, this shift was also observed in the protected L-idosyl bromide, as well as in reported a-L-idosides, so even as a preliminary indication of formation of the desired product, this was not applicable to the ido system. Conclusive evidence of the formation of the cyanoethylidene was obtained by 1 3 C NMR. One of the four carbonyl carbons in the 165-170 ppm range present in the starting material was absent in the product, and was replaced by a peak at 99 ppm, corresponding to the new quaternary carbon and a new peak at 116 ppm corresponding to the cyanide group of the product. The peak at 99 ppm also clearly shows that the product is the desired bicyclic compound, rather than the idosyl cyanide. 30 Also in contrast to Kochetkov's work, in which the cyanoethylidene products were obtained as almost equal mixtures of the endo and exo diastereomers, only a single diastereomer of the L-ido compound was obtained. Various N M R experiments were performed in an attempt to determine which diastereomer was formed, but conclusive data were not obtained. The idocyanoethylidene was not obtained as a crystalline compound, but as a syrup, and attempts to crystallize it failed; it was thus impossible to determine the stereochemistry through x-ray crystallography. CH 3 CH 3 C H 3 ido gluco galacto manno Figure 2.4 1,3-Diaxial interactions prevent the formation of the endo isomer of 3,4,6-tri-O-acetyl-1,2-0-(l -cyanoethyhdene)-B-L-idopyranose While conclusive evidence of the stereochemistry of the new compound was lacking, it is possible to hypothesize as to which diastereomer was more likely to be formed. It was clear, based on lH NMR data, that the L-idosyl bromide sits in a * C 4 conformation, and likewise, also based on ! H NMR data, we know that the cyanoethylidene product also sits in a lC4 conformation, based on the small coupling constants observed for H1-H2, H2-H3 and H3-H4 (2 to 3 Hz). It is reasonable to assume that the acetoxonium ion intermediate also exists in a lC4 conformation. If this is the case, the endo face of the carbocation is partially shielded by the C3 acetate, in a manner akin to a 1,3-diaxial interaction, while the exo face is unobstructed. Thus, we can 31 hypothesize that the exo product is formed preferentially. Since glucose, galactose and mannose all sit in a 4 C i conformation, there are no pseudo-l,3-diaxial interactions to consider, which is consistent with the diastereomeric mixtures obtained (Figure 2.4). The other challenge to this synthetic approach was the formation of the trityl ethers of the acceptors of interest. Kochetkov's work was limited to carbohydrate acceptors, while our interest lies in acceptors having significantly lower pK» values. Forming the trityl ether of a phenol would appear to be quite trivial: triphenylmethyl chloride reacted with the desired phenol, in the presence of a base, such as triethylamine. The desired product is readily formed, but, depending on the pKa of the phenol, isolation can be challenging. For high pKa phenols, rapid workup followed by rapid chromatography using a very non-polar solvent system, such as 20:1 petroleum ether: diethyl ether, was a satisfactory means of purification. Purity was judged solely by tic, as N M R techniques revealed a mass of peaks all falling at nearly the same chemical shift. The first test of the cyanoethylidene glycosylation reaction was done using phenyl trityl ether as the acceptor. Because of the extreme water-sensitivity of the trityl catalyst, scrupulously dry conditions are required for the reaction. In order to ensure dryness of the reagents, both the cyanoethylidene and trityl ether were freeze dried repeatedly from benzene. The cyanoethylidene was dissolved in dry 1,2-dichloroethane (DCE) and freshly activated 4A molecular sieves were added. Triphenylcarbenium tetrafluoroborate and the trityl ether were each dissolved in dry D C E , and added to the cyanoethylidene. Tic revealed that after 5 hours at 55 °C, the reaction was complete. Purification and characterization revealed that the reaction had, indeed, produced exclusively the desired 1,2-trans idoside (25) in a respectable 46% yield. To my knowledge, this was the first 32 time this glycosylation chemistry had been applied to reactions other than the synthesis of oligosaccharides. Having shown that the reaction produced exclusively the a-L-idoside, the task was then to move to acceptors of lower pKa. Para-nitrophenol (PNP) was the next acceptor tested. It became apparent that isolation of the PNP-trityl ether would be extremely difficult, if not impossible, and an alternate route had to be established. Generation of the trityl ether in such a way as to eliminate the need for isolation and purification appeared to be the only viable solution. It was thought that this could be achieved by combining the potassium salt of PNP with an excess of trityl tetrafluoroborate. It was hoped that the byproduct of the reaction, potassium tetrafluoroborate, would not interfere with the glycosylation reaction. The potassium salt of PNP was generated by reacting the phenol with potassium hydride in tetrahydrofuran. The addition of ether to the solution caused the deep yellow-orange salt to precipitate out. The precipitate was collected by vacuum filtration and dried in vacuo over P2O5. To form the trityl ether, 1.1 eq phenolate and 1.5 eq trityl salts (relative to the amount of donor to be used in the glycosylation reaction) were weighed into separate flasks, under an atmosphere of argon. Each was dissolved in a minimal amount of dry DMF. The phenolate salt solution was a deep yellow-orange colour, while the solution of the trityl salt was a deep yellow. When the trityl salt solution was added to the phenolate solution, the orange colour disappeared instantaneously, indicating that the phenolate had been consumed to form the ether. Once the DMF had been removed in vacuo, the combination of the trityl ether, trityl tetrafluoroborate and potassium tetrafluoroborate was dissolved in dry D C E and added to the glycosyl donor. While the yield of this 33 reaction was not determined, as it was done on an extremely small scale, tic and *H NMR analysis of the crude product clearly indicated the formation of the desired 1,2-trans-idoside (26). The procedure described above was repeated using 3,4-dinitrophenol. Even after 48 hours at 70 °C, only starting materials were observed. The complete lack of reaction can be attributed to the already poor nucleophilicity of 3,4-dinitrophenol, which is further diminished in the form of the trityl ether. Clearly this reaction sequence is very pKa-sensitive, having a cutoff around pKg 6. 2.3.2 Sulfoxide glycosylation In the total synthesis of vancomycin, Kahne also experienced difficulties with unreactive phenols in Koenigs-Knorr-type glycosylation methods(52). While the phenol he was attempting to use as an acceptor (2,6-dimethoxyphenol) had a relatively high pKa, its substitution at the 2 and 6 positions offered sufficient steric hindrance at the reaction centre to inhibit glycosylation. Sulfoxide glycosylation chemistry had previously been shown to be applicable in cases in which the acceptor is quite unreactive. However, the presence of a small participating group, such as an acetate, on C2 of the donor, caused the formation of the orthoester, rather than the desired 1,2-trans glycoside(55). While Lewis acids are known to suppress orthoester formation, Kahne's attempts to perform the sulfoxide glycosylation with boron trifluoride added to the reaction mixture failed to produce the desired product. This was thought to be because of the poor nucleophilicity of the acceptor. Kahne found that the addition of 2,6 di-ter/-butyl-4-methylpyridine (DTBMP), a proton-specific base, in addition to the Lewis acid produced the desired 1,2-trans glycoside in respectable yields(52). 34 Since we also were struggling to form 1,2-trans glycosides of unreactive phenols, we sought to apply this chemistry to the ido system. Starting with the penta-O-acetate (19), thiophenyl idoside (27) was formed in 54 % yield by the addition of thiophenol and stannic chloride. The product was obtained as an anomeric mixture, with the a-anomer being predominant. For the sake of simplifying the characterization of later products, only the a-anomer was carried on to further reactions. Oxidation of the sulfide using /weta-chloroperoxybenzoic acid yielded a mixture of products, with the major product being one of two possible diastereomers of the desired sulfoxide product (28), which was obtained in 62 % yield (Figure 2.5). The absolute stereochemistry of the sulfur centre was not determined. PhSH SPh Ac ; C , OAc _ ^ O A c Q n r i * AcCX OAc J AcO ° A S P h OAc OAc 19 OAc OAc 27 MCPBA CH2CI2 AcO. Figure 2.5 Formation of Phenyl 2,3,4,6-tetra-O-acetyl-l-sulfinyl-a-L-idopyranoside (28) 35 The glycosylation reaction was carried out, as described by Thompson et al.(52), using phenol as the acceptor. The sulfoxide and DTBMP were repeatedly freeze-dried from benzene to ensure dryness. A mixture of the sulfoxide, DTBMP and freshly activated 4 A MS in dry dichloromethane was cooled to -78 °C. Following the addition of triflic anhydride, the reaction was warmed to -60 °C, to allow the formation of the triflate derivative of the glycosyl sulfoxide, which then converts to the glycosyl triflate (29) (Figure 2.6a) (54). Once again at -78 °C, a mixture of the acceptor (phenol) and the Lewis acid (boron trifluoride diethyletherate) in dichloromethane was added, and the reaction mixture was brought to -10 °C (Figure 2.6b). The mixture of products obtained was quite interesting. By tic, the reaction appeared to have gone quite cleanly, and a compound running as a single spot on tic was purified chromatographically, giving an apparent yield of 42%. However, analysis of the *H NMR spectrum of the sample revealed a 70:30 mixture of compounds. The major component was the desired phenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (25). Surprisingly, the minor component was phenyl 2,3,4,6-tetra-O-acetyl-l-thio-a-L-idopyranoside (27). As it was clear from NMR and micro analysis that the sulfoxide was not contaminated with unreacted sulfide, this unexpected byproduct must have formed during the glycosylation reaction. 36 Figure 2.6 Mechanism of glycosylation by the sulfoxide method, a) Formation of the glycosyl triflate intermediate; b) formation of the desired glycoside While Kahne does not ever report obtaining thioglycosides as byproducts from the sulfoxide glycosylation method, the formation of these compounds is not unprecedented. In a recent publication, Crich notes that, when using methanol as the acceptor in several sulfoxide glycosylation reactions, in addition to obtaining both the desired 1,2-trans glycoside and the orthoester, he also obtained the thioglycoside in up to 19% yield(54). 37 This unexpected product is a result of coupling of the donor sulfoxide with thiophenol, rather than with the desired acceptor. The formation of thiophenol is thought to occur through a series of complex disproportionation reactions involving the highly reactive benzenesulfenyl triflate byproduct(54). In the same manner as described above for phenol, 3,4-dinitrophenol was used as the acceptor in this reaction. The only product obtained was the thiophenyl idoside (27). In 1999, Kahne developed a means to suppress the formation of sulfenate-based byproducts, by introducing a sulfenate scavenger into the reaction mixture. He found that 4-allyl-l,2-dimethoxybenzene was quite proficient at intercepting the sulfenate byproduct (Figure 2.7), thereby, suppressing the formation of undesired byproducts(55). While this has the potential of making the sulfoxide glycosylations cleaner, it further complicates an already complex reaction. In the event that other routes to the desired products failed, we planned to revisit this chemistry. MeCv ^ ^ PhSOTf J J I  MeO M e o Y ^ s p h aqueous MeO^ ^ O H workup MeO v SPh Figure 2.7 Scavenging the phenylsulfenyl triflate intermediate of the sulfoxide glycosylation reaction 2.3.3 Silver triflate glycosylation In glycosylation reactions involving glycosyl halides as the donors, the use of insoluble silver salts, such as silver carbonate and silver silicate are known to favour direct inversion, and disfavour anchimeric assistance(56). However, in the synthesis of 1,2-cis mannosides through a-D-mannosyl bromide, the use of insoluble silver salts does 38 not ensure the formation of the desired product; in order to obtain reasonable yields, the use of non-participating C2 protecting groups is also necessary. Surprisingly, in light of the above expectations, previous work done in the Withers lab showed that treatment of 2,3,4,6-tetra-O-acetyl-a-L-idopyranosyl bromide (30) with silver carbonate and para-nitrophenol resulted in the formation of the B-L-idoside, which is consistent with direct Walden inversion rather than the expected participation of the C2 acetoxy protecting group. While 1,2-cw-mannosides are considered among the most difficult glycosides to form, the formation of 1,2-cw-idosides was remarkably readily achieved. This is particularly surprising when one considers the fact that, as in the ct-D-mannosyl bromide case, the C2 acetoxy group and the bromide in a-L-idosyl bromide exist in a fraws-diaxial orientation, ideal for anchimeric assistance (Figure 2.8). Br O A c O A c 30 manno ido Figure 2.8 1,2-Trans-diaxial orientation of both P-D-mannosyl and a-L-idosyl bromides In order to examine possible explanations for this apparent inconsistency, one must consider the mechanism of the reaction. Silver, acting as an electrophile, pulls the halide from the anomeric centre, concurrent with nucleophilic attack at the anomeric centre by either the C2 acetoxy group, as is observed in the manno case, or, in the ido case, the acceptor. It is also important to remember that the silver salt in this reaction is insoluble, and the halide is interacting with one particular silver ion on a large silver carbonate surface. In the manno case, this poses no problem; however, in the ido case, the C3 39 acetoxy group could create significant steric hindrance. Having tremendous conformational flexibility, it is possible that a steric conflict between the silver salt and the C3 acetoxy group, combined with the decreased anomeric effect (resulting from partial C-Br bond breakage) would be sufficient to cause the sugar to ring-flip to a lower energy conformation in which the C2 acetoxy group is no longer optimally aligned for participation (Figure 2.9). This could favour external nucleophilic attack, rather than anchimeric assistance. Figure 2.9 Possible conformational changes of a-L-idosyl bromide preventing C2 acetoxy participation during Koenings-Knorr glycosylation reactions In an effort to encourage participation of the C2 acetoxy group, a soluble silver salt (silver triflate) was used. Encouraging results were achieved when using phenol as the acceptor. In 37 % yield, orthoester 31 was obtained (Figure 2.9), showing clearly that the C2 acetoxy group is participating. However, we were quite surprised to see that, in a 40 parallel reaction using phenol and silver carbonate, the same orthoester was obtained, in contrast to the results previously obtained with /7-nitrophenol. From these data alone, it is difficult to explain these results. However, further work with a series of various acceptors may reveal that the identity of the product formed in this reaction depends upon the pKa of the acceptor. By changing the solvent from acetonitrile to dichloromethane, using silver triflate as the silver source, we were able to form the desired 1,2-trans idoside rather than the orthoester (Figure 2 .10) . While the effect of nitrile solvents on glycosylation reactions is well documented(57), it is unclear by what mechanism acetonitrile enhances, while dichloromethane suppresses, orthoester formation. O A c O A c 25 Figure 2.10 Koenigs-Knorr-type chemistry using a soluble silver salt As with the preceding reactions, this methodology proved to be unsuccessful in forming idosides of acceptors having low pKa values. 2.3.4 Triehloroacetimidate glycosylation Another commonly used glycosylation method for unreactive acceptors is the triehloroacetimidate method. This glycosylation chemistry is widely applicable as it can 4 1 be tailored to suit many different cases. Unlike glycosyl bromides, glycosyl trichloroacetimidates can be formed under conditions which can favour the preferential formation of either the a- or P-anomer. This means that in cases in which the C2 protecting group is nonparticipating, both the a- and P-glycosides are accessible through this donor. Because we have shown through the other glycosylation methods attempted that C2 acetoxy participation almost always occurs, we set out to form the a-L-ido-trichloroacetimidate (TCA), which should be more readily displaced by anchimeric assistance than the P-TCA. At -45 °C, in dichloromethane, 2,3,4,6-tetra-(9-acetyl-a,P-L-idopyranose (23) was reacted with trichloroacetonitrile in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to produce a 7:4 ratio of a:P-anomers in 94% yield. The two anomers were crudely separated chromatographically. First to elute was the a-anomer (32), and it was isolated approximately 90% pure (with the remaining 10 % being the p-anomer (33)); the remaining product was pooled as a 2:1 ratio of P:a anomers. Both mixtures yielded exclusively the desired a-L-idoside (25) when reacted with phenol in dichloromethane at 0 °C, in the presence of boron trifluoride diethyl etherate, and both in approximately 70% yield (Figure 2.11). Because the stereochemistry at the anomeric centre of the donor trichloroacetimidate had no effect on the anomeric stereochemistry of the product, in later reactions, the donor was used as a mixture of diastereomers. 42 OPh A c O . 9Ac J OAc OAc 25 Figure 2.11 Formation of Phenyl-2,3,4,6-tetra-0-acetyl-a-L-idopyranoside (25) by the trichloroacetimidate method As was done in the previous glycosylation studies, this reaction was then tested with 3,4-dinitrophenol as the acceptor. It was with great jubilation that the desired product (34) was obtained anomerically pure in 53% yield. The trichloroacetimidate glycosylation chemistry was also applied to glycosylations using 3,5-dinitrophenol and 4-nitrophenol as the acceptors. 43 2.3.5 Fluorodinitrobenzene glycosylation When working with electron-deficient phenols, such as 2,4-dinitrophenol, it may not be wise to attempt to use them as nucleophiles. An alternative approach may be to form the glycosidic linkage via nucleophilic aromatic substitution, using the sugar as the nucleophile. Following the approach developed by Koeners et al. for other sugars(5#), 2,3,4,6-tetra-(9-acetyl-a,B-L-idopyranose was reacted with fluorodinitrobenzene (FDNB) in the presence of l,4-diazabicyclo[2.2.2]octane (DABCO) (Figure 2.12). The reaction yielded an anomeric mixture of products, with the major product being the desired a-anomer (35), which was obtained in a disappointing 17% yield. Moderately increased yields could be made possible by increasing the chromatographic resolution of the two anomers. Br 36 35 Figure 2.12 Formation of 2,4-dinitrophenyl glycosides using fluorodinitrobenzene 44 2.3.6 Summary of glycosylation reactions Using a variety of the glycosylation methods described above, the desired series of idosides was formed in yields ranging from 17% to 72% (Table 2.1). It appears that the triehloroacetimidate method was by far the most effective, producing good yields and anomerically pure products. The silver triflate method also produced good yields, but yielded approximately 10% of the undesired P-L-idoside, thereby making tedious purification of the a-anomer necessary. Acceptor pKa Glycosylation Method Yield Compound # 4-methoxyphenol 11.06 silver triflate 45% 37 phenol 9.99 silver triflate 37% 25 4-chlorophenol 9.38 silver triflate 26% 38 3-nitrophenol 8.39 triehloroacetimidate 37% 39 4-nitrophenol 7.18 triehloroacetimidate 54% 26 3,5-dinitrophenol 6.45 triehloroacetimidate 72% 40 3,4-dinitrophenol 5.36 triehloroacetimidate 53% 34 2,4-dinitrophenol 3.96 fluorodinitrobenzene 17% 35 Table 2.1 Summary of results of large scale glycosylation reactions 45 2.4 Deprotection Deprotection of acetate-protected idosides was achieved readily by following a modification of the Zemplen method(5P). Sodium methoxide was generated in situ using sodium in methanol. While only catalytic amounts of methoxide are required in this procedure, rapid deprotection was achieved, with no methanolysis of the idosides, by using 4 equivalents of methoxide. The reactions were complete within 15 minutes, and the solutions were rapidly neutralized by the addition of Amberlite® TR120 resin. Purification was achieved by chromatography and/or crystallization, depending on the idoside. It was found that idosides of phenols having para substituents were highly crystalline, while phenyl idoside and idosides having ortho or meta substituents on the aglycone tended to remain as syrups, and consistently formed oils when crystallization was attempted. As with the fully protected ce-L-idosides, *H NMR data indicated that the deprotected a-L-idosides also have a strong preference for the conformation, based on the very small coupling constant observed for J\t2 (~3 Hz). This was confirmed by solving the crystal structure for the /?-chlorophenyl idoside (42) (Figure 2.13). In agreement with the proton NMR, data, the crystal structure clearly shows a nearly perfect *C4 chair. The anomeric effect, in combination with the large tendency of the C5 hydroxymethyl group to adopt an equatorial position, is sufficient to force the C2, C3 and C4 hydroxyl groups, as well as the large /rarra-chlorophenyl substituent at the anomeric centre into axial orientations. 46 Figure 2.13 Crystal structure of 4'-chlorophenyl a-L-idopyranoside Methoxide deprotection of the idosides was successful for all except the most labile idosides (3,4- and 2,4-dinitrophenyl), which were subjected to acidic deprotection. Acetyl chloride was dissolved in dry methanol, to give a 4% solution of HC1, and the protected idoside was added. The reaction mixture was maintained at 4 °C until tic showed the reaction to be complete (12 to 36 hours). In both cases, tic indicated that the reactions had proceeded relatively cleanly, with the majority of the desired product remaining intact. However, as previously noted, idosides of phenols having substituents in the ortho and/or meta positions tend not to crystallize, and these two compounds were no exception. Unfortunately, they also proved to be too labile to purify chromatographically, and so, even though we were finally able to develop a means of forming idosides of low pK» phenols, we were unable to carry them through the 47 remainder of the synthesis to form substrates for the kinetic investigation of iduronidase The results are summarized in Table 2.2. Compound # (starting material) Aglycone Deprotection method Yield Compound # (product) 37 4-methoxyphenol methoxide 76% 41 25 Phenol methoxide 69% 42 38 4-chlorophenol methoxide 86% 43 39 3-nitrophenol methoxide 86% 44 26 4-nitrophenol methoxide 78% 45 40 3,5-dinitrophenol methoxide 74% 46 34 3.4-dinitrophenol HCl/MeOH - -35 2,4-dinitrophenol HCl/MeOH - -Table 2.2 Summary of results of deprotection reactions 2.5 Selective oxidation of C6 The final step in the synthesis of the substrates for a-L-iduronidase is the selective oxidation of C6 from the primary alcohol to the acid, while leaving the remaining 3 secondary alcohols untouched. In addition, the reaction conditions must be sufficiently mild to leave the glycosidic linkage intact. These conditions are met in the TEMPO oxidation(^i). While there is debate in the literature over the extent of selectivity for primary over secondary alcohols, work done by de Nooy et al. addresses these issues and 48 examines the kinetics and mechanism of this reaction using methyl a-D-glucopyranoside as the substrate(42). They found the optimal conditions to involve the use of 6.5 milli-equivalents of TEMPO, 0.2 equivalents of NaBr, and 3 equivalents of NaOCl. The temperature was maintained at ~ 1.5 °C, and the pH at 10. Under these conditions, they obtained methyl a-D-glucuronate in quantitative yield. Equally impressive results were obtained for several other substrates (reviewed in(44)). They found that increased amounts of NaBr(45), increased temperature(</3) or decreased pH(42) caused a decrease in selectivity for primary alcohols. The oxidizing agent in this reaction is the alkyl nitrosonium ion, formed from the reaction of the TEMPO radical with hypobromite, as outlined below(43): HOCl + Br -» HOBr + Cf HOBr + 2 TEMPO 2TEMPO + + HBr T E M P O + + OFT + RCH 2 OH -> TEMPOH + RCHO + H 2 O T E M P O + + OFf + TEMP OH -» 2 TEMPO + H 2 0 The rate-determining step was found to be the oxidation of the substrate. The byproduct, TEMPOH, undergoes a disproportionation reaction with a second nitrosonium ion (TEMPO 4) to regenerate the TEMPO catalyst. The mechanism of nitrosonium-mediated oxidation is detailed in Figure 2.14. In the second step of the oxidation of the substrate (from the aldehyde to the acid), the aldehyde is in the hydrated form. The hydration step was also found to be rapid in comparison with the oxidation step(42). 49 Figure 2.14 Mechanism of TEMPO/HOBr oxidation Based on the extremely positive results obtained by de Nooy et al, we sought to apply this reaction to the ido system. We first tested the reaction with /?ara-nitrophenyl P-D-glucoside. The conditions used were as described by de Nooy, with the exception that, in the absence of a pH-stat to maintain the pH between 10 and 11, the reaction was run in a 1 M buffer composed of a mixture of sodium carbonate and sodium bicarbonate, pH 10.5. As described in the literature, the ! H N M R of the crude product showed a remarkably clean transformation to the glucuronide, with no other detectable products, after a reaction time of 60 minutes. 50 However, when this reaction was tested with idosides, the results were dramatically different. The oxidation was first attempted with para-methoxyphenyl a-L-idoside (41). In order to monitor the reaction by tic, it was necessary to manually maintain the pH at -10.5 by the periodic addition of sodium hydroxide, rather than through the use of a buffer. The reaction was sluggish, and after several hours at 0 °C, tic analyses showed that the majority of the starting material was unreacted. In an attempt to increase the reaction rate, the temperature was increased to 22 °C. Not only did this not affect the rate of disappearance of starting material, as judged by tic, the solution gradually took on a brownish shade. Proton NMR analysis revealed a complex mixture. Similar results were obtained using /weta-nitrophenyl a-L-idoside (44). After much investigation, through altering the reaction conditions, including temperature, ratios of reagents and reaction time, conditions were established under which a small amount of desired product was obtained. The reactions were run at 0 °C in 1 M sodium carbonate/sodium bicarbonate buffer, pH 10.5, using 0.2 equivalents of TEMPO, 4 equivalents of NaOCl and 0.1 equivalents of sodium bromide, relative to 1 equivalent of the appropriate idoside. After 75 minutes, the excess oxidizing agent was quenched by the addition of ethanol, the pH was reduced to 7-8, and the TEMPO was removed by extraction with ether. Purification of the desired compounds was achieved by size exclusion chromatography using Biogel P2 resin, eluting with 25 mM ammonium bicarbonate. This protocol was successfully applied to the /?-chloro-, m-nitro-, /?-nitro-, and 3,5-dinitrophenyl idosides. Oxidation of the /?ara-methoxyphenyl derivative required a decreased amount of TEMPO (0.04 equivalents) and an increased reaction time (36 51 hours). Both *H and 1 3 C NMR confirmed the formation of the desired compound. However, in contrast to the other derivatives, purification from the buffer salts was not achieved by size exclusion chromatography. Thus, it was not possible to use this compound in the enzymology studies. Neither oxidation protocol, when applied to the phenyl derivative, was successful. In both cases, the only identifiable product obtained was unreacted starting material, and that, only in approximately 10 % yield. These observations are also in sharp contrast to de Nooy's observation that, while anomeric configuration caused a small change in reaction rate, the anomeric substituent did not have any significant effect(42, 43). Compound # (starting material) Aglycone Yield Compound # (product) 41 4-methoxyphenol _* 47 42 phenol -43 4-chlorophenol 23% 48 44 3-nitrophenol 7.5% 49 45 4-nitrophenol 6.0% 50 46 3,5-dinitrophenol 12% 51 *the desired product was observed by H NMR, but we were unable to purify it from the buffer salts Table 2.3 Summary of results of oxidation reactions The yields for this final step of the synthesis of iduronidase substrates ranged from a disappointing 23% down to 6%, with two substrates not forming any of the desired 52 product (Table 2.3). It is unclear why the ido system presents such difficulties in this reaction. 2.6 Conclusions and future work Many interesting glycosylation methodologies were studied in an effort to obtain highly reactive ct-L-idosides. The triehloroacetimidate method was found to meet the requirements outlined in §2.3, making it possible to access many of the previously unobtainable idosides in the protected form. However, deprotection and purification of the more labile compounds continues to be problematic. In addition, the TEMPO oxidation methodology that has had tremendous success in other carbohydrate systems performed abysmally with idosides. While time did not permit this work within this dissertation, it would be interesting to investigate the reasons why idosides pose such difficulty in this oxidation. Clearly, if work is to continue in this area, a superior oxidation method must be developed. 53 CHAPTER TH R E S U L T S A N D D I S C U S S I O N : E N Z Y M O L O G Y 3.1 Introduction As interest in human a-L-iduronidase as a therapeutic agent for the treatment of MPS-1 is growing, questions regarding its mechanism and activity are becoming increasingly relevant to the treatment of this disease. The work described in this chapter is an investigation into the catalytic mechanism of a-L-iduronidase. The stereochemical outcome of the enzyme-catalyzed reaction, the identity of the catalytic nucleophile and kinetic details regarding the rate-determining step are all addressed. 3.2 Stereochemistry of iduronidase-catalyzed reaction The stereochemical outcome of the reaction catalyzed by a glycosidase is one of the key features used in its classification. Iduronidase has been placed in a family of retaining glycosidases on the basis of the retention of configuration observed for a Family 39 P-D-xylosidase, thus implying that the reaction it catalyzes proceeds with net retention of anomeric configuration. However, there are no empirical data to support this premise, and there is reason to question this assignment given the significant difference in primary structure of P-D-xylosidases and a-L-iduronidases. 54 3.2.1 Enzyme-catalyzed substrate hydrolysis The stereochemical outcome of glycosidase-catalyzed reactions is generally best determined by lH NMR analysis of the enzyme-catalyzed reaction. In this experiment, a spectrum of the intact substrate is taken, after which, enzyme is added to the sample and a series of spectra are obtained until the substrate has been completely consumed and the mixture of products has reached equilibrium. Monitoring of the chemical shift of the anomeric proton and its coupling constant to H2 in real time allows for the determination of the stereochemistry of the initial product formed in the reaction. Figure 3.1 Monitoring the stereochemical outcome of glycosidase-catalyzed substrate hydrolysis by lH NMR Free sugars undergo mutarotation in aqueous media, but this process is normally quite slow relative to the enzyme-catalyzed glycoside hydrolysis if sufficient enzyme is used (Figure 3.1). This results in accumulation of the initial product for a sufficient period of time to allow for lH NMR characterization and assignment of stereochemistry. At neutral pH, mutarotation has been observed to be quite slow, which is not unexpected, as it is both an acid- and base-catalyzed process. Because it was known from previous work that the pH optimum for human iduronidase is 4.5(5), there was concern that at such 55 a low pH mutarotation would be too fast for this experiment to be feasible. However, in 1997, Howard et al. successfully determined the stereochemistry of two different mannosidases using this technique, also working at pH 4.5. He found that mutarotation was not observed until approximately 20 minutes after enzyme addition, and complete equlibration required approximately 90 minutes(o'O). In addition, Wong et al. determined the stereochemical outcome of the human lysosomal B-glucuronidase-catalyzed reaction by this method(55). This is of particular interest because not only was the assay run at pH 4.8, but the substrate, glucuronic acid, is structurally very similar to the iduronidase substrate, and thus, these compounds might be expected to have comparable mutarotation rates. With substantial literature precedent and armed with literature data on the *H NMR spectrum of L-iduronic acid in aqueous solution(67), we attempted to determine the stereochemistry of the iduronidase-catalyzed reaction by this real-time NMR technique. Ideally, the experiment is set up such that sufficient enzyme is added to completely hydrolyze all of the substrate within the first minute or two of the reaction, in order to obtain a strong signal for the anomeric proton of the initial product before mutarotation becomes a factor. This proved to be problematic for our system since iduronidase is quite a slow enzyme (kcat for methylumbelliferyl a-L-iduronide is 2.7 s"1; see §3.4), and very little of the enzyme was available to us. In order to obtain reasonable signal to noise, we required a minimum concentration of 1 mM substrate, and 300 ug enzyme was used to achieve complete turnover in approximately 20 minutes (in contrast to the mannosidase experiment, which used 8 or 27 mM substrate and 5 or 0.16 mg enzyme, respectively). 56 substrate 1.4 min 2.5 min 5 min 10 min 20 min "1 I r 5.5 ~ i 1 1 1 1 1 1—i 1 1 r 5.0 4.5 ppm Figure 3.2 *H NMR analysis of the iduronidase-catalyzed substrate hydrolysis. H i . s : anomeric proton of the substrate (MUI); H 5 . s : H5 of MUI; H , v anomeric proton of the a-pyranose form of iduronic acid; H i ^ p : anomeric proton of the B-pyranose form; H ^ f : anomeric proton of the a-furanose form; H , . p f : anomeric proton of the B-furanose form 57 Because of the concern regarding mutarotation rates, following the addition of enzyme to the substrate mixture in the NMR tube, everything was done as rapidly as possible, in an effort to obtain the first spectrum before mutarotation was observable. The enzyme was added, rapidly mixed, the sample was placed back in the spectrometer and the magnetic field shimming was adjusted, all in under a minute. The first spectrum was obtained using only eight scans, which, while resulting in poor signal-to-noise, minimized the time between enzyme addition and obtaining the first spectrum. However, in spite of the fact that only 1.4 minutes had elapsed from the time the enzyme was added to the time the first spectrum was obtained, the anomeric protons of both the a - and P-anomers were visible in similar intensity. By 5 minutes, in addition to the a - and P-pyranose products, the a - and p-furanose products also appeared. The substrate was completely consumed after approximately 20 minutes, at which time, there also existed an equilibrium mixture of the four aforementioned products (Figure 3.2). There are two possible interpretations of these results. Either the mutarotation rate for iduronic acid is significantly faster than that of mannose and glucuronic acid at the same pH (so much so that it exceeds what can be observed on the experimental timescale) or iduronidase catalyzes the formation of both the a- and P-anomers, meaning that it is both an inverting and retaining enzyme. A glycosidase catalyzing hydrolysis with both retention and inversion of anomeric stereochemistry in substrate hydrolysis is unprecedented, which is not to say that it is impossible. However, examination of the conditions required for such an event to occur strongly suggest that this is not likely. Several crystal structures of both retaining and inverting glycosidases have been solved, and the geometry of the active sites of the two groups of enzymes differs in one dramatic 58 feature: the distance between the two conserved carboxylate residues. In both classes of enzyme, one of the carboxylic acids must be close enough to the glycosidic bond to be capable of protonating the leaving aglycone, to assist in its departure. However, the position of the second carboxylate is dependent on the type of reaction catalyzed. In retaining glycosidases, the carboxylate must be close enough to the anomeric centre to act as the catalytic nucleophile and form a covalent bond to C l of the glycone. In contrast, in inverting glycosidases, the second carboxylate is not a nucleophile, but, rather, a general base. Its role is not to form a covalent link to the glycone, but, rather, to deprotonate a water molecule situated between the carboxylate and the anomeric centre. This means that there must be a larger space between the anomeric centre and the carboxylate in inverting glycosidases than in retaining glycosidases (Figure 3.3). Generally, the distance between the conserved carboxylates in retaining glycosidases is approximately 5.5 A, while in inverting enzymes, it is 6.5 to 9.5 A. Based on these mutually exclusive geometric requirements, it is unlikely that one active site would have the capacity to catalyze both types of glycoside hydrolysis. This does not rule out the possibility that iduronidase has two active sites, each catalyzing one type of reaction. If an enzyme has two active sites, these sites can behave either independently of one another, or co-operatively. Co-operative action of active sites is reflected in the kinetics, and this behaviour was not observed for iduronidase. The scenario of independently acting active sites is somewhat more difficult to rule out, but, given that this scenario is seldom observed in glycosidases, the chances of this being the explanation are remote, particularly when one considers that the two active sites would catalyze different reactions. 59 " o o T 6.5 - 9.5 A Retaining glycosidase Inverting glycosidase Figure 3.3 Geometry of glycosidase active sites While it was rather unexpected, in light of literature precedent to the contrary, that the mutarotation of iduronic acid, even at pH 4.5, would be so fast, it seems to be the more plausible explanation. Evidence lending support to this is the appearance of the oc-and P-furanose forms after only 5 minutes. The formation of these 5-membered rings follows a mechanism that is, initially, identical to the initial steps of mutarotation (Figure 3.4). The first step of both mutarotation and pyranose-furanose interconversion is ring-opening via protonation of the endocyclic oxygen and the subsequent cleavage of the C l -05 bond (Figure 3.4a). Mutarotation then involves rotation about the C1-C2 bond, followed by nucleophilic attack of 05 at the CI protonated aldehyde (Figure 3.4b). Pyranose-furanose interconversion also proceeds from the open-chain form shown in Figure 3.4a. Rotation about the C3-C4 bond aligns 04 for attack on the CI protonated aldehyde to achieve ring closure (Figure 3.4c). 60 That the furanose forms of iduronic acid appear so quickly following substrate hydrolysis lends further support for the rapid mutarotation explanation of the unexpected hydrolysis data. It is possible that the higher rates of mutarotation of iduronic acid could be a consequence of the destabilized ground state of iduronic acid, relative to the sugars in the aforementioned examples, due to its (formally) axial C5 carboxylic acid substituent. H O Figure 3.4 Mutarotation and pyranose-furanose interconversion pass through a common intermediate, illustrated with the acid-catalyzed mechanism, a) formation of the common intermedate; b) ring closure resulting in mutarotation; c) ring closure resulting in pyranose-furanose interconversion. 3.2.2 Enzyme-catalyzed substrate methanolysis One of the few strategies that can be used to overcome the problem of rapid mutarotation is to change the pH at which the reaction is run. However, the extent to 61 which the pH can be altered is limited by the pH range in which the enzyme retains activity. Increasing the pH from 4.5 to 5.0 caused a moderate decrease in enzyme activity, but had no observable effect on the rate of mutarotation. The other strategy involves forcing the enzyme to form a compound that is incapable of mutarotation. To do this, the enzyme must employ a nucleophile other than water, such as methanol. While the resulting methyl glycoside is also a potential substrate for the enzyme, the turnover of this product would be expected to be very slow relative to the initial substrate (in this case, the methylumbelliferyl glycoside). The success of this approach depends on the extent to which the enzyme tolerates the presence of organic solvents in the buffer, as the presence of organic solvents, even in small amounts, causes some enzymes to denature. Being one of the lysosomal enzymes, which are generally quite robust, we were hopeful that iduronidase would retain its activity even in the presence of methanol. Methanolysis reactions were monitored by tic, and Figure 3.5 shows a representation of the results obtained using varying concentrations of methanol. All reactions were run for 35 minutes at 37 °C, using 5 mM substrate (methylumbelliferyl-ct-L-iduronide). The first lane, (0 M methanol) shows that under these conditions, the reaction goes nearly to completion, leaving only a small amount of intact substrate (Rf = 0.39). The other two spots are accounted for as the reaction products. The high-running fluorescent spot (Rf = 0.79) is 4-methylumbelliferone and the lower running spot is the free iduronic acid (Rf = 0.21). The mixture of ct/B-pyranose and a/B-furanose compounds show identical chromatographic behaviour, likely due to rapid equilibration between the different forms. The addition of methanol to the reaction mixture caused an 62 interesting change in the tic. A fourth spot appeared, which was only very slightly less polar than the free iduronic acid (Rf = 0.27). As with the free iduronic acid, this spot was neither U V active, nor fluorescent. These characteristics are consistent with what would be expected for the methyl glycoside of iduronic acid. It was also noted that increasing the concentration of methanol caused the intensity of this spot to increase, with a concurrent decrease in intensity of the free iduronic acid. Figure 3.5 shows the results up to 3 M methanol, at which point the methanolysis product appeared to be approximately 50% of the product formed. The methanol concentration was further increased to 6 M, with only a moderate increase in the proportion of methanolysis product obtained. Fluroescent (methylumbelliferone) U V active, char (substrate) Char only (iduronic acid) [MeOH] (M) Figure 3.5 Representation of tic obtained from iduronidase-catalyzed MUI hydrolysis in the presence of varying concentrations of methanol (run in 3:1:1 butanol: acetic acid: water) The above experiment (using 3 M methanol) was repeated on a larger scale, and *H NMR analysis of the product mixture is shown in Figure 3.6. Shown in blue is the mixture of the cc/B-L-idopyranuronate and a/B-L-idofuranuronate. The anomeric protons are labeled, and clearly correspond with the anomeric protons observed in the real-time hydrolysis experiment (Figure 3.2). However, consistent with the results observed by tic, a fifth product was observed, in approximately 50% yield (shown in red). 63 o o o CS> d > O 0 1.5 3 p p m Figure 3.6 X H N M R analysis of iduronidase-catalyzed M U I hydrolysis in the presence of 3 M methanol The chemical shifts and coupling constants of this new product are in agreement with those published for the sodium salt of methyl a-L-idopyranosiduronic acid(67). In the spectrum, there was no evidence of any of the B product. These data indicate that iduronidase is a retaining enzyme. 3.3 Identification of the catalytic nucleophile Having determined that substrate hydrolysis occurs with net retention of anomeric configuration, it then became necessary to distinguish between the two possible mechanisms by which this might occur: an enzyme active site nucleophile (Figure 1.4) or anchimeric assistance, using the C6 carboxylate as the nucleophile (Figure 1.6). 64 3.3.1 Sequence alignments Family 39 comprises all known a-L-iduronidases and a selection of B-D-xylosidases. As discussed in §1.1.3, the xylosidases have a high degree of sequence similarity, as do the iduronidases; however, the homology between the two groups of enzymes is considerably lower. One of the key applications of glycosidase sequence alignments is in the prediction of the catalytic residues, based on amino acid conservation. As shown in Figure 3.7a, alignment of the sequences around the proposed acid/base residue shows a very high degree of homology, and there is a very high probability that the prediction of El78 as the catalytic acid/base residue of human iduronidase is correct. The case is quite different, however, for the sequence alignment surrounding the proposed catalyic nucleophiles. Looking at the amino acid sequences in this section, it appears as though Family 39 could be broken down into two subfamilies: the xylosidases (shown in black in Figure 3.7) and the iduronidases (blue in Figure 3.7). The xylosidases in this region are nearly identical, but the similarity to the iduronidases (which are also highly homologous to one another) is minimal. The only exception to this is the xylosidase from Caldicellulosiruptor saccharolyticus (xyl-B), which has only moderate homology to the remainder of the xylosidases, but also has very low homology to the iduronidases. In fact, only one of the four sequence alignment matrices within the ClustalW sequence alignment program will align the xylosidase nucleophiles with the iduronidase nucleophiles proposed by Henrissat(62) (Figure 3.7b), and this alignment requires the lowest stringency settings within the program. The other alignment matrices 65 > H fa fa w w « « w w P o < < Q Q fa o o ex a j en EH 8-1 s s > > fa * O co > fa fa fa fa fa * fa P p P R !> SI EG H j K E? E? g Q n Q • ft Pi ft ft ft ft ft Pi * 3 S * 3 3 3 3 3 3 * H H H H EH fa fa fa fa fa fa w w * fa fa fa fa fa & fa & * ft ft ft J fa 2 3 3 3 3 3 3 * P EH fa !4 <! pi co CO CO > W EV W HV r> * [Tj fa Q « fa CO EH < H H w H J CD o o O o o o o * >n >< >H >H >H Pi ft1 a! ft' CO CO P W CO O O o H H H H H H n H * fa fa fa fa >H >< BB Pi K CO « CO K s 2 CT\ CTi in CTi CM o <-H ro CO ro i n «3 . H i-l i H i H rH i H i H U) ID M C-CN CN CN xyl) P 1 i-l >f pq i i H >1 um ( 01 c. CO 0 •H 0 O 4J •H . •H >l 4J 4-J M .—. i~H 0 \ >H 0 — - O <0 £ u u JT, <a CO <a 0 jr, 0 • 0 M 0 UJ a 0 0 CO w (0 <c CO D. CO (Ct p 0 9 9 in S U T J •H •H 0 0 . >H >H <D -—- id td <D Q) a. X ! 2 4J -U 3 4J 3 to T i 0 0 0 U •H •H ID <B ••H >H ••H U •—• •—• J3 J5 w UJ CO <ti 0 0 0 CU 0 • M CO CO U >H 4J M M 3 CD CD CO 3 M UJ UJ M i-H § 3 q Cl M CO M O ^ ta CD d) ^1 CO 0 0 0 M O 3 CO -H. ••H. W >H • H TI • H 0 0) cu M 0 co S X ! J5 10 IB H! 3 0 EH EH O cq O O xyl) P 1 i-H >1 m i >1 ) um CO CO u 3 3 O O •H , •H ^ 4J rH 4J M ,—. M >-H O i-H \ ^ O O « % In in (0 CO 10 O -CJ 3 jr. O O •-H 0 HI ft O ••H 0 CO CO (0 X ! <a CO Q< CO (0 O ,d 9 9 s lH .—. •H ••H. 0 in 0 •H . — . aj iH 4J <D ^—• (0 3 <D CU Q< X ! Q . •6 -U 4-1 3 • u 3 CO •ri O O in 0 M •M • H •—-10 10 •-H in • H In •—• ^1 )^ co i0 CO <B 0 0 0 0) 0 ••H CO CO H in i-H 4J i-H i-H 3 C| fl) <U 3 CO 3 • H i-H CD 10 (0 M i-H 3 ••H C| q i-H co •-H NH O Q< 10 (0 Q) 3 CU co <f 0 0 O O 3 CO • H M • H CO in T3 ••H T3 "H 0 cu i-H O M C| CO S, XI (0 (0 <0 10 3 p EH EH o oq o o &5 place aspartate-301 or alanine-314 in this position; some even suggest that there is no catalytic nucleophile in the iduronidases, aligning the xylosidase nucleophiles with a gap in the iduronidase sequences. This lack of homology in the nucleophile region causes some doubt as to whether the prediction of E299 as the human iduronidase nucleophile is correct. Given the disparity between the alignment results, the proposal of iduronidase using an anchimeric assistance mechanism rather than one involving an active site nucleophile becomes all the more reasonable. 3.3.2 Kinetics To probe the mechanistic controversy, the potential inactivator 5-fluoro-a-L-idopyranosyluronic acid fluoride (5F-IdoAF) (Figure 3.8a), which was generously provided by A.W. Wong, was investigated. As described in §1.1.4.3, glycosides fluorinated at the 2 or 5 position have been used to label the catalytic nucleophiles of several glycosidases. 5F-IdoAF was predicted, with substantial literature precedent, to behave likewise. To test this, the enzyme was incubated with 5F-IdoAF, and aliquots were taken at specific times following the addition of 5F-IdoAF to the enzyme, added to a substrate mix, and assayed for residual iduronidase activity. In this type of experiment, inactivators cause the residual enzyme activity to decrease according to single exponential kinetic behaviour, with higher inactivator concentration leading to more rapid loss of activity. The pattern shown for 5F-IdoAF, when tested as an inactivator for iduronidase, was not consistent with time-dependent inactivation. Rather than decreasing activity with time, the enzyme activity remained constant following the addition of inactivator (Figure 67 3.8). However, the observed enzyme activity decreased with increased concentration of 5F-IdoAF, with no time-dependent loss of activity thereby indicating competitive inhibition by the material carried over into the assay mix, rather than time-dependent inactivation. Since the concentration of 5F-IdoAF in the assay mixes would be relatively low, it must bind quite tightly to the enzyme, a) b) | H O A 0 O H 0.8 •E 0.6 E 3 0.4 (U 0.2 0.2 mM 5 F - l d o A F -• 1 mM 5 F - l d o A F J i l i I i_ 2 4 6 time (minutes) Figure 3.8 Testing 5F-IdoAF as a time-dependent inactivator of iduronidase. a) structure of 5F-IdoAF; b) kinetic data obtained Therefore, the compound was tested as a reversible inhibitor. To do this, the kinetic parameters Km and Vmax were first determined for the substrate of choice (MUI). They were found to be 15 uM and 12 umol/(min!mg) respectively (Figure 3.9a). The substrate concentration was held constant at 50 uM, and the concentration of 5F-IdoAF was varied. The resulting data were analyzed by plotting 1/v vs. [I]. Kj was determined as the intersection of the lines 1 • = [/] Km [S]+Km vapP " JKiVmsK[S] Vmax[S] (Equation 3.1) and 68 v = (Equation 3.2) V max Solving this system of equations, where y = l/v0, gives the following relationship: [I] = -K; (Equation 3.3) and the intersection of the two lines gives the Kj value (Figure 3.9b). The Kj was found to be 1.2 uM. K m = 1 5 M M V m a x = 12 umol/(min-mg) i . i . i • i . 0 20 40 60 80 100 120 140 160 [Substrate] (uM) ^» 0.2 -2 0 2 4 6 8 10 12 14 16 [Inhibitor] (uM) Figure 3.9 Testing 5F-IdoAF as a competitive inhibitor of iduronidase. a) Kn/Vmax determination for MUI in the absence of inhibitor; the line represents the fit to the Michaelis Menten expression using the parameters shown; b) determination of K; Such a low Ki value would be quite surprising if it reflected purely binding; a more plausible explanation is that the enzyme is hydrolyzing the compound slowly, and there is significant accumulation of some intermediate along the reaction pathway. Therefore, the compound was also tested as a substrate (Figure 3.10) by directly monitoring fluoride release using a fluoride electrode. In order to determine the K„, and Vmax for any substrate, one must measure the initial velocity of the reaction at various concentrations, 69 generally ranging from approximately 0.1 to 10 times the Km value, allowing hydrolysis of only approximately 10 % of the substrate. Given that the Kj was found to be 1.2 uM, which is a reasonable approximation of the Km should the compound actually act as a substrate (see discussion below), and that the detection limit of the fluoride electrode is also approximately 1 uM, it was impossible to determine Km and Vmax by conventional methods. However a value for Vmax can be obtained by measuring rates at a very high concentration of 5F-IdoAF. We used 1 mM 5F-IdoAF (which is lOOOx Km), at which concentration, v 0 (the observed rate) is equal to 0.999 x Km (Equation 3.4). v" = ^ ""Inl (Equation 3.4) Km+[S] FmaxjlOOO/G,) V°~ Km+lOOOKm Vo = 0.999F max The reason Km can be approximated from the Kj determination is as follows. Km is defined as a combination of microscopic rate constants, reflecting the breakdown of the ES complex: k i k 2 E + S - " ES *~ E + P K m = ( k - i + k 2 ) / k i (Equation 3.5) When a competitive inhibitor is present, it binds only to the free enzyme, and Ki expresses the dissociation of the EI complex, as a true equilibrium constant: 70 E+S + I ES E + P EI (Equation 3.6) Kj = [E][rj / [EI] However, if the compound that was tested as an inhibitor is actually a substrate, as may be the case for 5F-IdoAF, the situation changes as illustrated below: E+S + I ES E + P EI E + I K m * = (k.,*+ k2*) / kj * (Equation 3.7) and the measured K; is actually reflective of the combination of rate constants used to determine Km, rather than being a true equilibrium constant. Thus, the Km value for 5F-IdoAF (should it behave as a substrate, not just a reversible inhibitor) was found to be 1.2 uM. 71 The resulting graph, obtained by measuring the release of F , had two interesting features. First, although the assay mixture was preincubated at 37 °C prior to the addition of enzyme, there was a significant lag period before a linear rate was observed. At the present time, we have no explanation for this phenomenon. The slope representing Vmax was taken from the latter, linear portion of the graph (Figure 3.10a), and kcat was determined to be 0.83 s'1 (taken as an average of duplicate runs). If the rate had been taken from the earliest part of the curve, a value of kcat = 0.38 s"1 would have been obtained. This 2-fold difference makes no significant difference to our interpretation. 0 10 20 30 40 24 26 28 30 32 34 36 38 40 Time (min) Time (min) Figure 3.10 Testing 5F-IdoAF as a substrate, a) iduronidase was added following a preincubation period, and the reaction was monitored until all substrate had been consumed; b) magnification of t = 24 to 40 minutes (from 3.10a) It is important to note when calculating the turnover rate for this compound that, unlike the 2-deoxy-2-fluoro inactivators, which release a single fluoride ion per hydrolysis event, the 5-fluoro compounds, following enzyme-catalyzed hydrolysis, undergo a rapid base-catalyzed elimination resulting in the release of 2 fluoride ions per hydrolysis event (Figure 3.11). Values obtained for kcat are corrected for this. 72 The second interesting feature of the graph shown in Figure 3.10a is magnified in Figure 3.10b. Instead of seeing a gradual decrease in rate as the substrate is consumed, there is a rather abrupt transition from the observed steady-state rate to a rate of zero. This is indicative of a very low Km, in agreement with previous results, implying accumulation of an intermediate; presumably, this intermediate is either the glycosylated enzyme, or the enzyme bound very tightly, albeit non-covalently to a reactive intermediate, such as the lactone illustrated in figure 1.6. Figure 3.11 5-Fluoro glycosides release 2 fluoride ions per hydrolysis event Although this proposed inactivator did not behave exactly as we had hoped, the indication of the accumulation of a reaction intermediate holds great potential for distinguishing between the two possible mechanisms. This was exploited as follows. 73 3.3.3 Mass spectrometric analysis Mass spectrometric analysis is, perhaps, the most straightforward method by which the mechanism of iduronidase can be elucidated. If substrate hydrolysis involves a covalent glycosyl-enzyme intermediate, and if this is the compound that is accumulating during the hydrolysis of 5F-IdoAF, it may be observable by MS. Because the iduronidase we are working with was expressed from mammalian cells (Chinese Hamster Ovary cells)(<55), mass spectrometric analysis was complicated by the complex glycosylation pattern. Iduronidase has six N-glycosylation sites, some of which are susceptible to endo-B-N-acetylglycosaminidase H (endo H) cleavage. However, the remainder of the carbohydrate moieties contain fucose at the proximal GlcNAc, making endo H ineffective(64), thereby making the completely deglycosylated enzyme unattainable. The inability to fully deglycosylate, or reduce the glycosylation pattern to one that is uniform across the enzyme population, made the mass spectrum of the intact protein uninterpretable. Therefore, it was necessary to assume that a covalent intermediate was, indeed, being formed during the reaction with 5F-IdoAF. Without direct validation of this assumption, we proceeded to proteolytic digestion of the "labeled" enzyme, and the resulting peptides were separated by HPLC. The isolation of a peptide covalently linked to 5-fluoro iduronic acid would conclusively prove the mechanism shown in Figure 1.4, in which the enzyme employs an active site nucleophile, while the absence of such a labeled peptide neither disproves this mechanism, nor proves the mechanism of anchimeric assistance (Figure 1.6). In order to identify a labeled peptide by MS, it is necessary to have a control, in which the enzyme is digested under the same conditions, but in the absence of 5F-IdoAF 74 (making the assumption that the same proteolysis pattern would occur in both the presence and absence of 5F-IdoAF). Indeed, comparison of the two digests revealed the presence of a labeled peptide. At 33.5 minutes on the total ion current chromatogram of the HPLC purification of the labeled peptide digest, a peptide having m/z = 893±1 was found, this fragment being absent in the unlabeled digest. Treatment of this peptide with ammonium hydroxide caused m/z to decrease by 97, corresponding to a loss of a label with a molecular mass of 194 g/mol, observed in the M + 2H* form. The labeled peptide was then subjected to sequencing by MS/MS (Figure 3.12). Table 3.1 details the peaks observed in the mass spectrum and their sequence assignments. The observed B ions containing the label were ADTPIYNDEA (m/z = 1286), ADTPIYNDEAD (m/z = 1401), ADTPIYNDEADP (m/z = 749), ADTPIYNDEADPL (m/z = 1610), and ADTPIYNDEADPLV (m/z = 855). These data indicate that the label is located within the ADTPIYNDEA section of the peptide. Only one Y" ion containing the label was observed. NDEADPLVG. Together with the B ion data, this indicates that the label was localized to the tetrapeptide NDEA. Alanine-300 can be ruled out, as it does not have a functional group capable of acting in a nucleophilic manner, and asparagine 297 can most likely be ruled out on the basis of precedent: all known retaining glycosidases (using an enzyme-active site nucleophile) employ either aspartate or glutamate as the nucleophile. We can therefore conclude that the label is localized to either aspartate-298 or glutamate-299. 75 m/z Fragment m/z Fragment 175s V G 855d ADTPIYNDEADPLV* 288s ADT 892d ADTPIYNDEADPLVG* 386s ADTP 1123s N D E A D P L V G * 498s ADTPI 1286s ADTPIYNDEA* 661s ADTPIY 1401s ADTPIYNDE A D * 749d ADTPI YNDEADP* 1610s ADTPIYNDEADPL* Table 3.1 Daughter ions observed after fragmentation of labeled peptide, assuming cleavage at (C=0)-N bonds; s: singly charged, d: doubly charged, *: labeled. In order to show conclusively whether D298 or E299 carries the label, one of two peptides must be observed. If the label is on D298, the labeled fragment ADTPrYND* would have to be observed (with * representing the label). This fragment, having m/z = 1084, was not observed in the mass spectrum (Figure 3.13a), nor was the unlabeled fragment E A D P L V G , with m/z = 700 (Figure 13b). In order to definitively show that the label was on E299, the labeled fragment * E A D P L V G would have to be observed. This labeled fragment has m/z = 895. This creates a most unfortunate situation. The intact, labeled peptide appears at m/z = 893, and has quite a strong signal, thus the peak we are searching for (m/z = 895) could quite easily be lost in the shoulder of the more intense peak of the intact peptide. 77 a) 50000 » 40000 o ~ 30000 to c 2 £ 20000 10000 1009.5 » 1094.5 1 1 2 2 m/z, amu 1100 b) 6 0 0 0 0 -50000 -CO Q . 4 0 0 0 0 -O 30000 -CO c © 2 0 0 0 0 -10000- 702.0 686.5 I kk A w\IU/V /U 700 m/z, amu Figure 3.13 Closer examination of MS/MS analysis of daughter ions of the labeled peptide, a) expansion of the region m/z = 1000 to 1130; b) expansion of the region m/z = 660 to 720; expansion of the region m/z = 879 to 905 with increased source voltage 78 In order to explore the possibility that this is what was happening, the voltage difference between the two quadrupoles (Q2 and QO) was adjusted, in an effort to fragment a greater portion of the parent peptide, thereby decreasing the intensity of the m/z = 893 peak. The results, shown in Figure 3.13c, show a small peak at m/z = 895, indicating that the label is, indeed, localized to E299. This is in agreement with the sequence alignments that predicted E299 to be the catalytic nucleophile. These data conclusively show that iduronidase does not act by anchimeric assistance, but, as do most retaining glycosidases, employs an enzyme active site catalytic nucleophile, which is glutamate 299. 3.4 Probing the rate-determining step A great deal of information regarding the mechanism of glycosidases can be obtained through the identification and characterization of the rate-determining step of substrate hydrolysis. One of the most useful tools for these investigations involves the analysis of linear free energy relationships. 3.4.1 Bronsted correlation With a series of substrates of varying reactivity in hand, we set out to measure the kinetic parameters for their hydrolysis in hopes of generating insight into the rate-limiting step. Each substrate was subjected to enzyme-catalyzed hydrolysis under identical conditions (temperature, pFL buffer concentration, etc.). The parameters Km and kcat were measured for each substrate (Table 3.2), and we then sought to determine if there was a 79 correlation between the catalytic constants and the pKa values of the leaving groups (a Brensted correlation). The presence of such a correlation would indicate that glycosylation is the rate-determining step and provide insight into the degree of charge development at the transition state for this step. Aglycone p K a kcat Km (uM) lOg(kcat) l o g ( k c a t / K m ) 4-chlorophenol 9.38 4.4 44 0.64 -1.0 3-nitrophenol 8.39 2.2 19 0.34 -0.94 methylumbelliferone 7.50 2.7 6.1 0.43 -0.35 4-nitrophenol 7.18 4.2 17 0.62 -0.61 3,5-dinitrophenol 6.45 1.5 13 0.18 -0.94 Table 3.2 Kinetic parameters for a series of substrates of varying reactivities The catalytic constant, kcat, is a first-order rate constant; it is a combination of all first order rate constants following formation of the ES complex, whether these steps be chemical or otherwise. When one of these rate constants is significantly lower than the rest, this constant will dominate, and be the observed catalytic constant. The parameter kcat is also referred to as the "turnover number" as it represents the number of substrate molecules converted to product per enzyme active site per unit time. The kcat values observed for iduronidase were quite low, and varied very little from one substrate to the next. The most reactive of the substrates, surprisingly, was found to be the substrate with the highest pK a leaving group, the /?-chlorophenol, with kcat = 4.4 s"1, while the least reactive, having a kcat of 1.5 s"1, was the 3,5-dinitrophenol compound, which had the lowest pK a leaving group. The three remaining substrates had kcat values scattered in 80 between, with no apparent trend (Figure 3.14a), these being a total spread of kcat values of less than three-fold. Figure 3.14 Bransted plots for both the a) first- and b) second-order rate constants The second-order rate constant, kcat/Km, is also referred to as the specificity constant, and is the rate constant relating the reaction rate to the concentration of free enzyme and free substrate, not to the ES complex. More specifically, reflects the first irreversible step in substrate hydrolysis. The plot of log(kcat/Km) vs. pKa had a slightly different shape, but showed the same overall lack of a trend (Figure 3.14b). Three of the substrates (/?-chlorophenyl, /w-nitrophenyl and 3,5-dinitrophenyl) had nearly identical kcat/Km values, while the /?-nitrophenyl and methylumbelliferyl substrates had siginificantly higher second order rate constants. However, with no overall trend, these data seem to indicate no dependence on leaving group ability, or a slope of zero on the Bransted plots. When graphing data having a slope of zero, it is difficult to know what a reasonable scale is. As plotted in Figure 3.14, the data points appear to be completely random, and even assigning a slope of zero is questionable. To put this data in a reasonable context, it has been graphed alongside the work done by Kempton (Figure 3.15). Her work 81 investigated the rate-limiting step for aryl glycoside hydrolysis catalyzed by Agrobacterium B-glucosidase(26), the prototypical glycosidase in the Withers lab. (The data on these graphs span a much wider range of pKa values, as the B-glucosides are much easier to obtain synthetically.) Within this context, the amount of scatter in both the iduronidase plots appears much more reasonable, and thus, so does assigning a slope of zero to both plots. This co-plot also illustrates how very low both the first- and second-order rate constants for the iduronidase-catalyzed hydrolysis actually are. a) b) o 1 1 1 r O 2 h o 1 h oo _i L 4 h 1 £ 2 1 6 8 PKa 10 o o i 1 i 1 -o I • I I • • • • 1 1 4 6 8 10 PKa Figure 3.15 Bransted plots for a-L-iduronidase (!) and Agrobacterium B-D-glucosidase (#). a) first-order rate constants; b) second-order rate constants That the first-order rate constant (1^) doesn't depend on the pK a of the leaving group can mean two different things. First, it could indicate that glycosylation is not rate-determining, and some other step in the hydrolysis reaction is the slow step. The second possibility is that glycosylation is indeed rate-determining, but that charge development in the transition state is minimal due to extremely efficient proton donation by the catalytic acid/base residue. 82 The observation that the second-order rate constant (kcat/Km) does not depend on the pKa of the leaving group (Figure 3.15b) is somewhat more enkghtening, as this parameter reflects the first irreversible step in substrate hydrolysis. In glycoside hydrolysis, the first irreversible step is generally the cleavage of the glycosidic bond. These data suggest a number of possibilites. As stated above, it is possible that glycosylation is, indeed, rate-limiting, but that proton donation is extremely efficient. It is possible, albeit not probable, that some step prior to bond cleavage is rate-limiting, such as a conformational change of the substrate. It is also improbable, but not impossible, that glycosylation and the preceding steps do not represent the first irreversible step in substrate hydrolysis, but that deglycosylation is the first irreversible step. 3.4.2 The effect methanol addition on rate If the glycosylation step is not rate-determining in kcat, the most likely candidate for the rate-determining step is deglycosylation, the only other chemical step in the mechanism. In cases where deglycosylation is rate-determining, rate acceleration can often be achieved by the addition of a superior nucleophile, such as methanol or dithiothreitol, to the reaction mixture(2<5), since this accelerates the rate-limiting step, lending to faster overall turnover. We have shown earlier that iduronidase is extremely tolerant of methanol, and that at relatively low concentrations of methanol, very high ratios of methanolysis to hydrolysis products can be formed (§ 3.2.2). However, when methanol was added to the assay mixture, no rate enhancement was observed, indicating that deglycosylation was not rate-determining (Figure 3.16). Indeed, a slight rate reduction was seen, as shown below. 83 Figure 3.16 Dependence of rate on methanol concentration at saturating substrate concentration The methanolysis data indicate convincingly that deglycosylation is not rate-limiting. This, together with the Bronsted plot of kcat/Km vs. pKa suggest that the most likely scenario is one in which glycosylation is rate-hmiting, but extremely efficient proton transfer minimizes the build-up of negative charge at the transition state, thus nunimizing the effect of the leaving group pKa. Proving that glycosylation is indeed the rate-determining step is best done by measuring kinetic isotope effects, however, the substrates required to do this work present yet another significant synthetic challenge which was not undertaken within this work. 3.5 Conclusions and Future Work Human-a-L-iduronidase was proven to be a retaining enzyme, and, in spite of the poor sequence alignments with many of the other family 39 glycosidases, the prediction of Glu299 as the catalytic nucleophile was shown to be correct. Preliminary mechanistic 84 investigations indicated that deglycosylation was not rate-determining, and that glycosylation, with extremely efficient proton transfer at the transition state may be rate determining. It would be interesting to expand the Brensted plots to include more substrates, with a wider range of leaving-group pKa values. However, this requires further synthetic investigation to develop these compounds. Also of great interest in this study would be the measurements of kinetic isotope effects to determine conclusively whether glycosylation is, indeed, rate-limiting; this work also requires further synthetic investigation. Having shown that iduronidase employs an enzyme-active site nucleophile, as do the other family 39 members, the question of the relationship between the family 39 xylosidases and iduronidases is again raised. As discussed in § 1.1.4.2 there exists the possibility that the conformational flexibility of both iduronic acid and xylose, and their ability to take on a 2 ' 5 B conformation is the foundation for their assignment to the same glycosidase family. The answer to this question lies in x-ray crystallographic studies. Currently, there are no crystal structures available for any family 39 member, however, efforts are ongoing in this area, particularly with human a-L-iduronidase because of its clinical significance^). 85 C H A P T E R IV MATERIALS AND METHODS 4.1 Synthesis 4.1.1 General materials and methods All *H nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 400 MHz WH-400 instrument and chemical shifts are given in parts per million. Samples were referenced to (CHC13: 7.24, CD 3 OD: 3.30, CeDe: 7.15). 1 3 C NMR are proton-decoupled and were run on either a Varian XL-300 or a Bruker AM-400 spectrometer, at 75 or 100 MHz, respectively. Melting points were determined on a Laboratory Devices Mel-temp II melting point apparatus, and are uncorrected. Micro-analyses were performed by Mr. Peter Borda, Microanalytical laboratory, University of British Columbia (UBC). Mass spectral analysis of synthesized compounds was performed at the U B C Mass Spectrometry Facility. Low and high resolution desorption chemical ionization 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 (LSTMS-LRMS and LSIMS-HRMS) were performed on a Kratos-Concept U H mass spectrometer equipped with a cesium-ion gun using glycerol and methanol as the matrix. Low and 86 high resolution fast atom bombardment mass spectrometry (FAB-LRMS and FAB-FfRMS) were performed using a Kratos Concept HHQ. Reagents were either of reagent, certified or spectral grade, and no further purification was performed unless otherwise stated. Phenol and /7-chlorophenol were distilled, potassium cyanide was recrystallized from ethanol and water, and was dried under vacuum at 78 °C. Solvents were dried as follows: methanol was distilled from magnesium and iodine, dichloroethane was distilled from phosphorus pentoxide; dichloromethane, pyridine and acetonitrile were distilled from calcium hydride; sequential drying over 3 A molecular sieves was performed for A^Af-dimethylformamide. Dried solvents were used for all reactions, unless otherwise stated. Thin layer chromatography (tic) was used to follow all reactions, unless otherwise stated. Tic separations were performed using either E M Science or Merck Kieselgel 60 F-254 aluminum-backed analytical plates. Compounds were visualized with ultra violet light (where possible) and/or stained with either 10% sulfuric acid in methanol, or 10% ammonium molybdate, 2 M sulfuric acid in water. For fully acetylated sugars and protected glycosides, the tic's were run in 1:1 ethyl acetate: petroleum ether. Deprotected sugars were run in 17:2:1 ethyl acetate: methanol: water. Oxidation reactions were not monitored by tic. Sili-Cycle silica gel, 230-400 mesh, was used for all column chromatography. 87 4.1.2 General Procedures 4.1.2.1 General procedure - deprotection Deprotection of idosides was carried out according to Zemplen(5P) with minor modifications. The appropriate acetylated idoside was suspended in methanol (50 mL/g), and sodium (4 eq.) was added. While the sodium was reacting to form hydrogen gas and sodium methoxide, an atmosphere of argon was generated in the reaction flask, which was cooled on ice. Once the sodium had completely reacted, the reaction mixture was stirred at room temperature until reaction was complete, as indicated by tic. The reaction was usually complete within 15 minutes. The methoxide was neutralized using Amberlite® IR-120 resin and immediately filtered and concentrated in vacuo. The resulting syrup was purified by silica gel flash column chromatography (5% methanol in ethyl acetate) to yield the desired idoside. 4.1.2.2 General procedure - oxidation To a solution of 1 M sodium carbonate/sodium bicarbonate (2 mL, pH 10.5) was added sodium bromide (0.1 equivalents), sodium hypochlorite (4 equivalents, using a 6% solution), and 2,2,6,6-tetramethylpiperidinyl-l-oxy (TEMPO) (0.2 equivalents). Complete dissolution of the TEMPO required 30 to 60 minutes. This mixture was then cooled to 0 °C and added to an ice-cold solution of the appropriate idoside (1 equivalent) in deionized water. The reaction mixture was stirred on ice for 75 minutes, after which time, ethanol (50 uL) was added. After stirring for an additional 15 minutes, the pH was adjusted to between 7 and 8 by dropwise addition of a 3 M solution of HC1. The solution was extracted with diethyl ether (2 x 1 mL) to remove the TEMPO. The resulting 88 aqueous layer was loaded directly onto a BioGel P2 column (1.8 x 50 cm), and was eluted with 25 mM ammonium bicarbonate. Fractions containing the desired compound were pooled and lyophilized repeatedly from deionized water. 4.1.3 Large scale synthesis of idose The synthesis of per-O-acetylated L-idose was accomplished according to the work ofBaggettera/. (39). J, 2-O-Isopropylidene- a-D-glucurono-6,3-lactone (13) D-Glucurono-6,3 -lactone (12) (126 g, 715 mmol) was dissolved in acetone (3 L) containing sulfuric acid (40 mL). After stirring for 8 hours, sodium bicarbonate (120 g) was added to the solution and the mixture was stirred overnight. Solids were removed by gravity filtration, and the acetone by evaporation in vacuo. The desired product was obtained by crystallization from isopropanol (103 g, 476 mmol, 67 %). *H NMR (CDCb) 5 5.96 (d, 1 H, Jl>2 = 3.7 Hz, H(l)), 4.90-4.92 (m, 1 H, H(4)), 4.78-4.81, (m, 2 H, H(2), H(3)), 4.49 (d, 1 H, JSA= 4.4 Hz, H(5)), 2.69 (bs, 2 H, OH), 1.49, 1.32 (2 x s, 6 H, 2 x CH 3). J, 2-0-Isopropylidene-5-0-toluene-p-sulfonyl-a-D-glucurono-6,3-lactone (14) Protected lactone 13 (98.9 g, 457 mmol) was dissolved in pyridine (1 L) and the solution was cooled on ice. Para-toluene sulfonyl chloride (96 g, 503 mmol) was added portionwise over 4 hours while keeping the reaction mixture cold. After 18 hours on ice, the pyridine was removed in vacuo and the residue dissolved in dichloromethane (2.5 L). This organic solution was washed with 1 M HC1 (3 x 300 mL). The organic layer was 89 dried over M g S 0 4 and treated with activated charcoal. The solids were removed by gravity filtration and the solvent by evaporation in vacuo. The product was crystallized from acetone (132 g, 356 mmol, 78 %). ! H NMR (CDC13) 5 7.87 (d, 1 H, J = 8.4 Hz H(2'), H(6')), 7.34 (d, 2 H, J= 8.4 Hz H(3'), H(5')), 5.94 (d, 1 H, Jh2 = 3.7 Hz, H(l)), 5.20 (d, 1 H, J 3,4 = 4.0 Hz, H(3)), 4.94 (dd, 1 H, J4,i = 4.0, J4,5 = 2.9 Hz, H(4)), 4.80 (d, 1 H, JsA = 2.9 Hz H(5)), 4.75 (d, 1 H, J2,i = 3.7 Hz, H(2)), 2.42 (s, 3 H, C H 3 of toluene group) 1.46, 1.30 (2 x s, 6 H, 2 x CH 3). I, 2-0-Isopropylidene-5-0-toluene-p-sulfonyl-a-D-glucofuranose (15) Tosylate 14 (132 g, 356 mmol) was dissolved in 2.5 L dioxane containing glacial acetic acid (300 mL) and the solution was cooled on ice. Sodium borohydride (130 g, 3.43 moles) was added portionwise over 5 hours, maintaining the temperature of the reaction mixture at 0 °C. The reaction mixture was allowed to warm to room temperature as it was stirred overnight, then was poured into ice water (2 L), which was stirred until no more gas evolved. The dioxane/water mixture was repeatedly extracted with dichloromethane (total volume used: 9.5 L). The organic phase was concentrated in vacuo, and the product was crystallized from chloroform/petroleum ether. After one crop, further crystallization from the mother liquor could not be achieved. The remaining product (-80 g) was subjected to silica gel flash column chromatography (1:1 ethyl acetate: petroleum ether). Pure fractions were concentrated in vacuo to yield a colourless syrup. Crystallization was then accomplished from the aforementioned solvent system (95.7 g, 256 mmol, 72 %).lHNMR (CDC13) 6 7.81 (d, 2 H, J= 8.2 Hz, H(2'), H(6')), 7.36 (d, 2 H, J= 8.2 Hz, H(3'), H(5')), 5.88 (d, 1 H, J u = 3.6 Hz, H(l)), 4.79-4.84 (m, 1 H, H(5)), 4.55 (d, 1 H, J2,i = 3.6 Hz, H(2)), 4.33 (dd, 1 H, J 3 m = 4.2, J3,4 = 2.4 Hz, H(3)), 90 4.25 (dd, 1 H, J4,3 = 2.4, J4y5 = 9.3 Hz, H(4)), 3.67-3.81 (m, 2 H, H(6a), H(6b)), 3.19 (d, 2 H, J = 4.2 Hz, 2 x OH), 2.44 (s, 3 H, C H 3 of toluene group), 1.45, 1.29 (2 x s, 6 H, 2 x CH 3). 3,6-Di-O-acetyl-l,2-0-isopropylidene-5-0-toluene-^-sulfonyl-a-D-glucofuranose (16) Diol 15 (95.7 g, 256 mmol) was dissolved in acetic anhydride (750 mL) and the solution was cooled on ice while pyridine (250 mL) was added dropwise over 3 hours. After stirring the reaction mixture for an additional 2 hours, the solvents were evaporated in vacuo. The residue was dissolved in dichloromethane (1.5 L) and washed with water (2 x 250 mL) and brine (250 mL). The aqueous fractions were pooled and extracted with 500 mL dichloromethane. The organic layers were pooled, dried over MgS04, filtered and concentrated in vacuo. The product was obtained as a white crystalline solid, and no further purification was carried out (113 g, 247 mmol, 97 %). *H NMR (CDC13) 8 7.74 (d, 2 H, J= 8.2, H(2'), H(6')), 7.31 (d, 2 H, J= 8.2 Hz, H(3'), H(5')), 5.83 (d, 1 H, Jh2 = 3.7 Hz, H(l)), 5.18 (d, 1H, J3A = 3.1 Hz, H(3)), 5.09 (ddd, 1 H, J5A = 8.4, JiM = 5.8, J5jSb = 2.3 Hz, H(5)), 4.48 (m, 2 H, H(2), H(6b)), 4.36 (dd, 1 H, J 4,3 = 3.1, J4,5 = 8.4 Hz, H(4)), 4.15 (dd, 1 H, y 6 a,5 = 5.8, J 6 a , 6 b = 12.8 Hz, H(6a)), 2.42 (s, 3 H, C H 3 of toluene), 2.13, I. 96 (2 x s, 6 H, 2 x OAc), 1.45, 1.27 (2 x s, 6 H, 2 x CH 3). 3,4,6-Tri-0-acetyl-l,2-0-isopropylidene-B-L-idofuranose (17) Because of the difficulties with this reaction, described in §2.2, the starting material was divided into two portions, and reacted separately. Tosylate 16 (40 g, 87.2 mmol;) was dissolved in acetic anhydride (400 mL) and potassium acetate was added (38 g, 387 mmol), and the reaction mixture was stirred at 140 °C for 48 hours, after which the 91 resulting tar-black mixture was cooled to 0 °C. Water (200 mL) was added, the mixture was stirred for 2 hours and extracted with chloroform (3 x 500 mL). The combined organic layers were washed with several portions of saturated sodium bicarbonate. The neutralized solution was dried over MgS0 4 , filtered and concentrated. The residue was passed through a plug of silica gel using ethyl acetate as the eluent to remove the majority of the black contaminants. A second reaction was set up using 73 g (159 mmol) of tosylate 16, and the other reagents were scaled up accordingly. This reaction was run for 24 hours, in hopes of decreasing the amount of thermal decomposition of the starting materials and products. Since the starting material and products have almost identical chromatographic characteristics, it was not possible to tell, until after work-up, that although the decomposition had been decreased, the desired reaction had only proceeded to 75 % completion. The products obtained from both of the above reactions were isolated by crystallization from ethanol, silica gel flash chromatography of the mother liquor (3:2 diethyl ether: petroleum ether), followed by repeated recrystallization from ethanol (44.4 g, 128 mmol, 52 %; 62% when recovered starting material is accounted for). lH NMR (CDC13) 6 5.90 (d, 1 H, J u = 3.5 Hz, H(l)), 5.34 (ddd, 1 H, J5A = 8.0 Hz, J5M = 3.6, J5,6h = 6.1, H(5)), 5.19 (d, 1 H, J3A = 3.3 Hz, H(3)), 4.49 (d, 1 H, J2,i = 3.5 Hz, H(2)), 4.39 (dd, 1 H, J4,3 = 3.3, J4,5 = 8.0 Hz, H(4)), 4.31 (dd, 1 H, J 6 a,5 = 3.6, ,4,6b = 12.2 Hz, H(6a)), 3.94 (dd, 1 H, J6b,5 = 6.1, y6b,6a = 12.2 Hz, H(6b)), 2.10, 2.09, 2.02 (3 x s, 9 H, 3 x OAc), 1.50, 1.29 (2 x s, 6 H, 2 x CH 3). 92 2,3,4-Tri-O-acetyl-l, 6-anhydro-fi-L-idopyranose (18) Sodium (883 mg, 0.038 mmol) was added portionwise to cooled methanol (500 mL) under an atmosphere of argon. Once the sodium had completely reacted, the solution was warmed to room temperature and idofuranose derivative 17 (44.41 g, 0.128 mmol) was added. After stirring for 90 minutes, the solution was neutralized with dilute HC1 and the solvent was evaporated in vacuo. The residue was dissolved in water (600 mL) containing HC1 (40 mL of 12 M solution). This reaction mixture was stirred at 100 °C for 2 hours after which time the water was evaporated in vacuo. The residue was dissolved in pyridine (400 mL) and cooled on ice. Acetic anhydride (200 mL) was added dropwise, and the reaction mixture was allowed to warm to room temperature. After stirring for 3 hours, the solvent was evaporated in vacuo, and the product was isolated by silica gel flash chromatography (3:2 diethyl ether: petroleum ether), followed by crystallization from ethanol. Pure fractions were pooled and concentrated in vacuo (21.72 g, 75.4 mmol, 58 % over 3 steps). J H NMR (CDC13) 6 5.42 (d, 1 H, J 1 > 2 = 1.8 Hz, H(l)), 5.30 (t, 1 H, y 3 > 2 = AA = 8.6 Hz, H(3)), 5.05 (ddd, 1 H, = 8.6, J4>5 = 4.5, JAJA> = 1.1 Hz, H(4)), 4.82 (dd, 1 H, J2,i = 1.8, J2,3 = 8.6 Hz, H(2)), 4.58 (t, 1 H, J 5 , 4 = J5Ja> = 4.5 Hz, H(5)), 4.14 (d, 1 H, y 6 a , 6 b = 8.0 Hz, H(6a)), 3.76 (ddd, 1 H, J6BM = 8.0, J6b,s = 5.1, J6BA = 1.1 Hz, H(6b)), 2.04, 2.02, 1.99 (3 x s, 9 H, 3 x OAc). J, 2,3,4,6-Penta-0-acetyl-a,p-L-idopyranose (19) Anhydro sugar 18 (21.72 g, 75.4 mmol) was dissolved in acetic anhydride (400 mL) and a mixture of acetic acid (200 mL) and sulfuric acid (6 mL) was added dropwise. The reaction mixture was stirred for 1 hour, after which time it was poured into ice water (200 mL) and stirred for 30 minutes, then the resulting solution was extracted with 93 dichloromethane (4 x 400 mL). The combined organic layers were washed with several portions of saturated sodium bicarbonate until neutral. The organic layer was dried over MgS04, filtered and concentrated in vacuo. A mixture of the title compound and the hemiacetal 23 was obtained. Purification was achieved by column chromatography (1:1 ethyl acetate: petroleum ether). Pure fractions were pooled and concentrated in vacuo. (12.50 g, 32.0 mmol, 42 % (19); 6.26 g, 18.0 mmol, 24 % (23)). *H NMR for 19 (CDC13) (2:1 ratio of a.B anomers) 8 6.05 (d, 1 H, Jia = 2.3 Hz, H(lp)), 6.03 (s, 1 H, H( l«) ) , 5.23 (t, 1 H, J 3 ,4 = Ai = 4.8 Hz, H(3P)), 5.05 (dt, 1 H, J = 3.2, 3.2, 1 Hz), H(2a) or H (3a)), 4.99 (dd, 1 H, J 2 , i = 2.3, J 2 > 3 = 4.9 Hz, H(2P)), 4.91 (t, 1 H , J = 3.0, H(3) or H(2)), 4.89 (dd, 1 H, J4,3 = 4.8, J4,5 = 3.2 Hz H(4P)), 4.86 (m, 1 H, H(4a)), 4.44 (ddd, 1 H, J= 6.3, 6.3, 2.4 Hz, H(5a)), 4.35 (ddd, 1 H, J5,4= 3.2, Js^Jsja = 6.9 Hz, H(5P)), 4.46 - 4.42 (m, H(6a«), H(6ap), H(6ba), H(6ap)), 2.13 - 2.00 (m, 10 x OAc); J H NMR for 23 (CDC13) (1.3:1 ratio of a:0 anomers) 8 5.20 (t, 7 = 3.9 Hz), 5.15 (dd, Jh2 = 2.4, Jl>OH = 5.3 Hz, H(l«)) , 5.1 (dd, Jia = 1.7, Jxm = 8.1 Hz, H(l p)), 5.07 (t, J= 4.0 Hz), 4.92 (t, J = 2.75 Hz), 4.82 - 4.80 (m), 4.71 - 4.64 (m), 4.57 - 4.53 (m), 4.26 - 4.13 (m, H(6a„), H(6a,3), H(6ba), H^ap)), 3.80 (d, JOH,i = 8.1 Hz, OHp), 3.48 (d, y 0 H , i = 5.3 Hz, OH«), 2.17 - 1.99 (8 x OAc). 94 4.1.4 Methodology study 4.1.4.1 Cyanoethylidene method 2,3,4,6-Tetra-O-acetyl-a-L-idopyranosyl bromide (30) A mixture of the a and B anomers of peracetylated L-idose (19) (1.78 g, 4.57 mmol) was added to a solution comprising dichloromethane (25 mL) and acetic anhydride (0.5 mL). Once the starting material had dissolved, 45% hydrogen bromide in glacial acetic acid (5 mL) was added and the flask was stoppered. After 2 hours, the reaction was judged by tic to be complete. The following steps were done in rapid succession in order to prevent decomposition of the product. The solution was diluted to 100 mL with dichloromethane, and washed with ice-cold, saturated sodium bicarbonate (4 x 30 mL). The organic layer was dried over MgS0 4, filtered, and concentrated in vacuo to yield the product as a white, crystalline solid (1.764 g, 4.29 mmol, 94%). *H NMR (CDCI3) 8 6.31 (s, 1 H, H(l)), 5.09-5.12 (m, 1 H, H(4)), 4.95-4.97 (m, 1 H, H(2)), 4.90-4.92 (m, 1 H, H(3)), 4.61 (m, 1 H, Hz, H(5)), 4.23 (dd, 1 H, J6a,5 = 5.5, J6a>6b = 11.5 Hz H(6a)), 4.17 (dd, 1 H, J6bf5 = 7.3, y6b,6a =11.5, H(6b)), 2.14, 2.10, 2.08, 2.06 (4 x s, 12 H, 4 x OAc). 3,4,6-Tri-O-acetyl-l, 2-0-(l-cyanoethylidene)-B-L-idopyranose (24) Bromide 23 (53 mg, 0.128 mmol) was dissolved in acetonitrile (2 mL). This solution was transferred to a flask containing potassium cyanide (44 mg, 1.09 mmol) and freshly activated 4 A molecular sieves. The reaction mixture was stirred, protected from light, for 24 hours. The solids were removed by gravity filtration, and the solvent by 95 rotary evaporation. The residue was dissolved in CHCI3 (15 mL), and washed with water ( 2 x 5 mL), then the organic layer was dried over MgSCv, filtered and concentrated in vacuo. The desired product was purified by silica gel flash column chromatography in 1:2 ethyl acetate: petroleum ether. A single diastereomeric product was isolated as a colourless syrup (27 mg, 0.076 mmol, 59%). *H NMR (CeDe) 5 5.43 (t, 1 H, J= 2.2 Hz), 4.86 (d, 1 H, J 1.2 = 2.6 Hz, H(l)), 4.80-4.82 (m, 1 H), 4.11 (dd, 1 H, y 6 a,5 = 6.1, y 6 a,6b = 11.5 Hz, H(6a)), 4.17 (dd, 1 H, J6b,5 = 6.8, J 6 b, 6a = 11.5 Hz, H(6b)), 3.77 (ddd, 1 H, J5A = 1.6, J 5 >6a = Jsja, = 6.1 Hz, H(5)), 3.70-3.73 (m, 1 H), 1.62, 1.57, 1.53, 1.51 (4 x s, 12 H, 4 x OAc), 1 3 C NMR (75 MHz, CDC13) 8 169.17, 168.68, 168.17, 116.24, 99.51, 96.39, 73.86, 68.74, 67.87, 67.73, 52.92, 24.48, 20.63, 20.60; Exact mass calc'd. for C15H23N2O9: 357.1403; found: 357.1403. Triphenylmethylphenyl ether (52) Phenol (1.00 g, 1.06 mmol) and triphenylmethyl chloride (2.95 g, 1.06 mmol) were dissolved in dichloromethane (50 mL) and triethylamine (4.43 mL, 3.18 mmol) was added. After 2 hours, the reaction mixture was diluted to 100 mL and washed with 1 M HC1 (2x30 mL), followed by brine (30 mL). The organic phase was dried over MgS04, filtered, and concentrated in vacuo. Purification was achieved by silica gel flash chromatography in 18:1 petroleum ether: diethyl ether. The product was obtained as a white amorphous solid, and was carried on to the following reaction with no further characterization. 96 Phenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (25) by the cyanoethylidene method Cyanoethylidene 24 (110 mg, 0.307 mmol) and trityl ether 51 (114 mg, 0.339 mmol) were dissolved in benzene (2 mL) in separate flasks, and freeze-dried three times, in order to ensure dryness. Freshly activated 4 A molecular sieves were added to the flask containing 24, followed by dichloroethane (2 mL). After stirring at room temperature for 0.5 h, 25 and triphenylcarbenium tetrafluoroborate (55 mg, 0.166 mmol), each dissolved in dichloroethane (1 mL), were transferred to the sugar-containing flask. After stirring at 55 °C for 5 hours, the reaction mixture was cooled to room temperature, filtered and diluted to 25 mL with dichloromethane. The solution was washed with NaOH ( 2 x 5 mL, 1 M solution), followed by brine (5 mL), then the organic layer was dried over MgS04, filtered and concentrated in vacuo. The residue was subjected to silica gel flash column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and concentrated in vacuo to yield the desired product as a colourless syrup (60 mg, 0.14 mmol, 46%). *H N M R (CDC13) 8 7.29 - 7.24 (m, 2 H, H(2'), H(6')), 7.05 - 7.00 (m, 3 H, H(3'), H(4'), H(5')), 5.52 (s, 1 H, H(l)), 5.08 - 5.07 (m, 2 H), 4.99-5.01 (m, 1 H), 4.86-4.92 (m, 2 H), 4.22 (AB multiplet, 2 H, J6&,5 = 7.0, = 5.2, 76a,6b = 11.4 Hz, H(6a, 6b)), 2.17, 2.10, 2.07, 2.01 (4 x s, 12 H, 4 x OAc), 1 3 C N M R (100 MHz, CDC13) 8 170.33, 169.65, 169.12, 169.03, 155.93, 129.44, 122.69, 116.73, 95.90, 67.21, 66.99, 66.59, 65.25, 52.00, 20.74, 20.72, 20.60, 20.46; exact mass calc'd. for C20H28O10: 442.1713; found: 442.1710. 97 4.1.4.2 Sulfoxide method Phenyl 2,3,4,6-tetra-O-acetyl-l-thio-a-L-idopyranoside (27) A mixture of the a and P anomers of peracetylated L-idose (19) (100 mg, 0.256 mmol) was dissolved in dichloromethane (15 mL) and the solution was cooled to 0 °C. Thiophenol (29 uL, 0.282 mmol) and stannic chloride (21 uL, 0.18 mmol) were added. After stirring on ice for 2 hours, the reaction mixture was diluted to 30 mL with dichloromethane and washed with 1 M HC1 (2 x 10 mL). The organic layer was dried over MgSO-*, fitlered and concentrated in vacuo. The resulting crude ct/p mixture was subjected to column chromatography, using 3:1 diethyl ether: petroleum ether as the eluent. Pure fractions were pooled and concentrated in vacuo. The desired a anomer 0 was isolated as a colourless syrup (61 mg, 0.138 mmol, 54 %). J H N M R (CDC13) 8 7.48-7.52 (m, 2 H , H(2'), H(6')), 7.25-7.31 (m, 3 H , H(3'), H(4'), H(5')), 5.45 (s, 1 H , H(l)), 5.04 (ddd, 1 H , J = 3.3, 3.3, 1.0 Hz), 4.99-5.01 (m, 1 H), 4.86-4.92 (m, 2 H), 4.22 (AB multiplet, 2 H , J 6 a , 5 = 7.0, J 6 b > 5 = 5.2, «/6a,6b = 11.4 Hz, H(6a, 6b)), 2.17, 2.10, 2.07, 2.01 (4 x s, 12 H , 4 x OAc), 1 3 C N M R (100 MHz, CDC13) 8 170.43, 169.63, 169.14, 168.62, 134.62, 131.74, 128.99, 127.78, 85.43, 68.39, 66.40, 66.370, 65.39, 62.30, 20.79, 20.78, 20.66, 20.62. Phenyl 2,3,4,6-tetra-O-acetyl-l-sulfinyl-a-L-idopyranoside (28) A solution of 27 (61 mg, 0.138 mmol) in dichloromethane (2 mL) was cooled to -78 °C and /weta-chloroperbenzoic acid (24 mg, 80-85 % pure, 0.139 mmol) was added. The reaction mixture was stirred for 2 hours as it warmed to room temperature, after 98 which time, methyl sulfide was added (10 uL, 0.136 mmol). The solution was poured into saturated sodium bicarbonate (5 mL) and the aqueous layer was extracted with dichloromethane (3x5 mL). The organic layers were pooled, dried over MgS04, filtered and concentrated in vacuo. The residue was subjected to sillica flash column chromatography (3:2 ethyl acetate: petroleum ether) (39 mg, 0.085 mmol, 62%). Pure fractions were pooled and concentrated in vacuo, yielding the desired compound as a white solid. lU NMR (CDC13) 6 7.69-7.72 (m, 2 H, H(2'), H(6')), 7.49-7.52 (m, 3 H, H(3'), H(4'), H(5')), 5.52-5.55 (m, 1 H, H(2)), 5.22-5.25 (m, 1 H, H(3)), 4.97-5.00 (m, 1 H, H(4)), 4.65-4.71 (m, 2 H, H(5), H(l), 4.24 (dd, 1 H, J6a,5 = 7.7,. = H.8 Hz, H(6a)), 4.12 (dd, 1 H, J6h,5 = 4.5, J6b,6a = 11.8 Hz, H(6b)), 2.18, 2.11, 2.08, 1.93 (4 x s, 12 H, 4 x OAc), 1 3 C NMR (75 MHz, CDC13) 8 170.30, 169.46, 168.66, 142.65, 131.55, 129.19, 124.39, 93.48, 69.66, 65.84, 65.79, 62.87, 62.28, 20.81, 20.72, 20.63, 20.55; C2oH24OioS calc: C, 52.43; H, 5.30; S, 7.02; found: C, 52.43; H, 5.33; S, 7.01 %. Phenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (25) by the sulfoxide method Sulfoxide (28) (61 mg, 0.134 mmol) and 2,6-di-ferf-butyl-4-methyl pyridine (27.5 mg, 0.134 mmol) were combined and twice dissolved in and freeze-dried from benzene (2 mL). Freshly activated 4 A molecular sieves and 4 mL dichloromethane were added to the residue, and the mixture was stirred for 45 minutes at room temperature, and cooled to -78 °C. Trifluoromethane sulfonic anhydride (22.5 uL, 0.134 mmol) was added, the mixture was stirred at -60 °C for 20 minutes, and the solution was once again cooled to -78 °C. Boron trifluoride diethyl etherate (84 uL, 0.669 mmol) and phenol (13 mg, 0.138 mmol) were combined and dissolved in dichloromethane (1 mL). The resulting solution was added to the sulfoxide solution. The reaction mixture was warmed to -47 °C and 99 stirred for 10 minutes, after which time, the reaction mixture was poured into saturated sodium bicarbonate (20 mL). The aqueous layer was extracted with dichloromethane (3 x 15 mL). The combined organic layers were dried over MgSC>4, filtered and concentrated in vacuo. The resulting syrup was chromatographed (3:2 diethyl ether: petroleum ether) and an inseparable mixture of products (phenyl and thiophenyl) was obtained as a colourless syrup. 4.1.4.3 Silver trifluoromethanesulfonate method Phenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside by the silver trifluoromethanesulfonate method Idosyl bromide (30) (1.385 g, 3.377 mmol) and phenol (634 mg, 6.74 mmol) were dissolved in dichloromethane (30 mL). Freshly activated 4 A molecular sieves were added, and the mixture was stirred at room temperature for 15 minutes. The mixture was cooled to -10° C and silver trifluoromethanesulfonate (952 mg, 3.71 mmol) was added. The reaction mixture was stirred, protected from light, and the temperature was allowed to slowly increase to 15 °C over 2 hours. The solids in the reaction mixture were removed by gravity filtration, and the resulting solution was diluted to 200 mL with dichloromethane, which was extracted with 1 M NaOH (2 x 30 mL), followed by brine (30 mL). The organic layer was dried over MgSCv, filtered and concentrated in vacuo. The residue was purified by silica gel flash column chromatography (solvent B) to yield product (326 mg, 37 %; 279 mg a+B mixture). When the reaction was initially done on a small scale (0.109 mmol), the yield was substantially better (50%, a-anomer only). 100 4.1.4.4 Triehloroacetimidate method 2,3,4,6-Tetra-0-acetyl-a,B-L-idopyranose (23) Idosyl bromide (30) (2.096 g, 5.10 mmol) was dissolved in acetone (20 mL), and 180 uL (10 mmol) water was added. Silver carbonate (1.66 g, 6.00 mmol) was added, and the reaction mixture was stirred, protected from light, for 2 hours. The reaction mixture was filtered, using Celite® filtering agent, and concentrated. The residue was purified by silica gel flash column chromatography to yield the colourless, syrupy product as an anomeric mixture. 2,3,4,6-Tetra-O-acetyl-a, and 8-L-idopyranosyl triehloroacetimidate (33) Hemiacetal 23 (1.588g, 4.56 mmol) was dissolved in dichloromethane (20 mL) and cooled to -45 °C. Trichloroacetonitrile (2.02 mL, 20.2 mmol) and 1,8-diazabicyclo[5,4,0]undec-7-ene (DBU) (28 uL, 0.202 mmol) were added and the reaction mixture was stirred for 2 hours, allowing the temperature to increase to -20 °C. The solution was diluted to 100 mL with dichloromethane, and washed with 1 M HC1 (20 mL), followed by saturated sodium bicarbonate (20 mL). The organic layer was dried over MgS04, filtered and concentrated in vacuo. Purification by silica gel flash column chromatography yielded the desired product as a pale yellow syrup, in a 7:4 ratio of cc:P anomers. (2.014 g, 4.09 mmol, 90%); lH NMR for a anomer (CDC13) 6 8.71 (s, IH, NH), 6.24(s, 1 H, H(l)), 5.05 - 5.02, 4.99 - 4.96, 4.90 - 4.88 (3 x m, 3 H, H(2), H(3), H(4)), 4.58 (ddd, 1 H, J5A = 1.8, J5M = 5.4, J5,6b = 6.6 Hz, H(5)), 4.21 (dd, J6*,s = 5.4, y 6 a , 6 b = 11.6 Hz, H(6a)), 4.13 (dd, J^s = 6.6, = 11.6 Hz, H(6b)); 2.10 (m, 9 H, 3 x 101 OAc), 1.99, (s, 3 H, OAc); exact mass calc'd for C 2oH 2oN 3 5Cl2 3 7Cl 2Oio: 529.9782; found: 529.9789. Phenyl-2,3,4,6-tetra-0-acetyl-a-L-idopyranoside (25) by the trichloroacetimidate method Trichloroacetimidate 33 (45 mg, 0.091 mmol) and phenol (17 mg, 0.183 mmol) were dissolved in dichloromethane (2 mL), and freshly activated 4 A molecular sieves were added. After stirring for 20 minutes, the mixture was cooled to 0 °C and boron trifluoride diethyl etherate (58 uL, 0.457 mmol) was added. The reaction mixture was stirred on ice for 2 hours, after which time triethylamine (19 uL, 0.0136 mmol) was added. The solvent was evaporated in vacuo, and the residue was subjected to silica gel flash column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and concentrated in vacuo to yield the desired product as a colourless syrup (27 mg, 0.064 mmol, 70%). 4.1.5 Large scale glycosylation reactions 4'-Methoxyphenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside Idosyl bromide (30) (1.764 g, 4.29 mmol) was dissolved in dichloromethane (20 mL) and freshly activated 4 A molecular sieves were added. After stirring at room temperature for 15 minutes, the mixture was cooled to - 1 0 °C. Silver trifluoromethane sulfonate (1.21 g, 4.72 mmol) was added, and the reaction mixture was stirred at - 1 0 °C for 1 hour, protected from light. The solids were removed by gravity filtration, and the resulting solution was diluted to 200 mL with dichloromethane, washed with 1 M NaOH (3 x 50 mL) and brine (50 mL). The organic layer was dried over MgS04, filtered and 102 concentrated. Purification of the resulting solid was accomplished by recrystallization from ethyl acetate/hexanes to yield a white crystalline solid (868 mg, 1.91 mmol, 45 %). J H NMR (CDCb) 5 6.96, 6.80 (AA'XX 1 system, 4 H, H(2'), H(3'), H(5'), H(6')), 5.39 (bs, 1 H, H(l)), 5.05-5.07 (m, 1 H, H(2) or H(3)), 5.03-5.05 (m, 1 H, H(3) orH(2)), 4.92-4.94 (m, 1 H, H(4)), 4.56 (dt, 1 H, JIA = 1.8, J5M = A&> = 6.3 Hz, H(5)), 4.16-4.18 (m, 2 H, H(6a), H(6b)), 3.75 (s, 3 H, OCH 3 ) , 2.13, 2.12, 2.10, 1.92 (4 x s, 12 H, 4 x OAc ) , 1 3 C NMR (75 MHz, CDC13) 5 170.46, 169.77, 169.25, 167.15, 155.23, 149.99, 118.10, 114.50, 96.74, 67.10, 66.94, 66.55, 65.00, 62.13, 55.64, 20.85, 20.706, 20.63; C 2 iH 2 60n calc: C, 55.50; H, 5.77; found: C, 52.29; H, 5.71 %. 4'-Chlorophenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (38) Idosyl bromide (30) (1.546 g, 3.76 mmol) was dissolved in dichloromethane (20 mL) and freshly activated 4 A molecular sieves were added. /7-Chloro phenol (967 mg, 7.52 mmol) was added, and the mixture was stirred for 15 minutes before cooling to -10 °C. The reaction mixture was protected from light and silver trifluoromethane sulfonate (1.063 g, 4.136 mmol) was added. After stirring for 1 hour at -10 °C, the solids were removed by gravity filtration. The resulting solution was diluted to 100 mL with dichloromethane, washed with 1 M NaOH (3 x 30 mL) and brine (30 mL). The organic phase was dried over MgSCv, filtered and concentrated in vacuo. The product was purified by silica gel flash column chromatography (3:2 diethyl ether: petroleum ether), followed by crystallization from ethyl acetate/hexanes to yield the desired compound as an off-white crystalline solid (454 mg, 26%). *H N M R (CDC13) 8 7.23, 6.97 (AA'XX system, 4 H, H(2'), H(3'), H(5'), H(6')), 5.46 (bs, 1 H, H(l)), 5.06-5.08 (m, 1 H, H(3)), 5.02-5.04 (m, 1 H, H(2)), 4.91-4.94 (m, 1 H, H(4)), 4.56 (dt, 1 H, J5,4 = 2.0, = = 103 6.6 Hz, H(5)), 4.16 (d, 2 H, H(6a), H(6b)), 2.14, 2.12, 2.11, 1.90 (4 x s, 12 H, 4 x OAc), 1 3 C NMR (75 MHz, CDC13) 8 170.41, 169.70, 169.22, 169.04, 154.39, 129.39, 127.77, 118.07, 95.95, 66.90, 66.65, 66.31, 65.21, 62.04, 20.85, 20.81, 20.69, 20.58; C20H23CIO10 calc: C, 52.35; H, 5.05; found: C, 52.57; H, 5.15 %. 3 '-Nitrophenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside Idosyl bromide (30) (1.647 g, 4.01 mmol) was dissolved in dichloromethane (20 mL) and freshly activated 4 A molecular sieves were added. After stirring for 15 minutes, the mixture was cooled to -10 °C and 3-nitrophenol was added. The reaction mixture was stirred for 1 hour at -10 °C. The solids were removed by gravity filtration, and the resulting solution was diluted to 200 mL with dichloromethane, washed with 1 M sodium hydroxide (3 x 50 mL) and brine (50 mL). The organic layer was dried over MgS04, filtered, and concentrated in vacuo. The residue was subjected to silica gel flash column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and concentrated to yield the desired compound as an off-white/yellow syrup. (868 mg, 1.91 mmol, 48 %). J H NMR (CDC13) 8 7.88-7.94 (m, 2 H, H(2'), H(4'), 7.45 (dd, 1 H, JVJB = = 8.2 Hz, H(5')), 7.34 (ddd, IH, JE,Y = 8.2, Je,r = 2.3, JEA = 1 Hz, H(5'), 5.57 (s, 1 H, H(l)), 5.07-5.10 (m, 1 H, H(3)), 5.03-5.05 (m, 1 H, H(2)), 4.91-4.93 (m, 1 H, H(4)), 4.56 (ddd, 1 H, J5A = 2.1, J5M = 6.3, J5,6H = 5.1 Hz, H(5)), 4.11-4.20 (m, 2 H, H(6a), H(6b)), 2.15, 2.12, 2.11, 1.85 (4 x s, 12 H, 4 x OAc), 1 3 C NMR (75 MHz, CDCI3) 8 170.47, 169.73, 169.32, 169.04, 156.44, 169.17, 130.20, 123.16, 117.76, 111.91,96.20, 66.75, 77.56, 66.16, 65.57, 62.13, 20.93, 20.85, 20.74, 20.54; C2oH 2 3N0 1 2 calc: C, 51.18; H, 4.94; N, 2.98; found: C, 51.47; H, 5.07; N, 2.78%. 104 4'-Nitrophenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (26) 4-Nitrophenol (382 mg, 2.75 mmol) was dissolved in hot dichloroethane (30 mL), and was sonicated until the phenol had completely dissolved. The solution was transferred to a flask containing trichloroacetimidate 33 (903 mg, 1.83 mmol) and freshly activated 4 A molecular sieves. After stirring at room temperature for 20 minutes, the mixture was cooled to 0 °C and boron trifluoride diethyl etherate (1.16 mL, 9.15 mmol) was added. The reaction mixture was stirred on ice for 90 minutes, quenched with triethylamine (580 uL, 4.12 mmol), filtered and diluted to 200 mL with dichloromethane. The solution was extracted with 1 M sodium hydroxide (50 mL portions, until aqueous layer is no longer yellow) and brine (30 mL). The organic layer was dried over MgSCu, filtered and the solvent was removed in vacuo. The residue was purified by silica gel flash column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and concentrated in vacuo to give the anomerically pure product (466 mg, 0.993 mmol, 54 %). Proton NMR was in agreement with published data (39). 3 \ 5 '-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (39) 3,5-Dinitrophenol (621 mg, 3.37 mmol) was suspended in hot dichloroethane (40 mL) and sonicated until completely dissolved. The solution was transferred to a flask containing the trichloroacetimidate 33 (1.107 g, 2.25 mmol) and freshly activated 4 A molecular sieves. After stirring at room temperature for 20 minutes the reaction mixture was cooled to 0 °C, and boron trifluoride diethyl etherate was added (1.42 mL, 11.25 mmol). After stirring on ice for 2 hours, triethylamine (710 uL, 5.05 mmol) was added the solids were removed by gravity filtration and the resulting solution was diluted to 200 mL with dichloromethane, and washed with 1 M sodium hydroxide (50 mL portions, 105 until aqueous layer is no longer yellow) and brine (50 mL). The organic layer was dried over MgS04, filtered and concentrated in vacuo. The residue was purified by silica gel flash column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and the solvent was evaporated in vacuo, yielding the desired product as a pale yellow syrup (828 mg, 1.61 mmol, 72 %). *H NMR (CDC13) 6 8.73 (t, 1 H, J = 2.0 Hz H(4')), 8.24 (d, 2 H, J= 2.0 Hz H(2'), H(6')), 5.66 (s, 1 H, H(l)), 5.12-5.14 (m, 1 H, H(3)), 5.05-5.07 (m, 1 H, H(2)), 4.93-4.95 (m, 1 H, H(4)), 4.47 (m, 1 H, H(5)), 4.21 (dd, 1 H, J 6 a , 5 = 7.6, J6a,6b = 11.7 Hz, H(6a)), 4.15 (dd, 1 H, J6b,s = 5.0, J 6 b j 6 a = 11.8 Hz, H(6b)), 2.18, 2.15, 2.14, 1.87 (4 x s, 12 H, 4 x OAc); 1 3 C NMR (100 MHz, CDC13) 8 170.26, 169.53, 169.21, 168.81, 157.04, 149.13, 117.37, 112.71, 96.74, 66.31, 66.20, 65.93, 65.72, 61.98, 20.82, 20.68, 20.59, 20.37; exact mass calc'd. for C20H26O14N3: 532.1415; found: 532.1412. 3',4'-Dinitrophenyl2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (35) 3,4-Dinitrophenol (695 mg, 3.77 mmol) was suspended in hot dichloroethane (40 mL) and was sonicated until completely dissolved. The solution was transferred to a flask containing triehloroacetimidate 33 (1.240 g, 2.51 mmol) and freshly activated 4 A molecular sieves. After stirring the mixture at room temperature for 20 minutes it was cooled to 0 °C, and the boron trifluoride diethyl etherate was added (1.59 mL, 12.55 mmol). After stirring on ice for 2 hours, triethylamine (788 uL, 5.66 mmol) was added, the solids were removed by gravity filtration and the resulting solution was diluted to 200 mL with dichloromethane, washed with 1 M sodium hydroxide (50 mL portions, until aqueous layer is no longer yellow) and brine (50 mL). The organic layer was dried over MgS04, filtered and concentrated in vacuo. The residue was purified by silica gel flash 106 column chromatography (3:2 diethyl ether: petroleum ether). Pure fractions were pooled and concentrated in vacuo, yielding the desired product as an off-white foam (1.03 g, 2.00 mmol, 53 %). ! H NMR (CDC13) 6 8.01 (d, 1 H, Jyfi = 8.9 Hz, H(5')), 7.47 (d, 1 H, J?,& = 2.5 Hz, H(2')), 7.31 (dd, 1 H, Je,y = 8.9, Je,T = 2.5 Hz, H(6')), 5.62 (s, 1 H, H(l)), 5.09-5.12 (m, 1 H, H(3)), 5.02-5.04 (m, 1 H, H(2)), 4.92-4.94 (m, 1 H, H(4)), 4.39-4.44 (m, 1 H, H(5)), 4.12-4.20 (m, 2 H, H(6a), H(6b)), 2.15, 2.13, 2.13, 1.91 (4 x s, 12 H, 4 x OAc); 1 3 C NMR (100 MHz, CDC13) 8 170.34, 169.53, 169.20, 168.77, 159.48, 145.17, 136.21, 127.20, 119.70, 112.72, 96.42, 66.38, 66.15, 66.12, 65.79, 61.99, 20.79, 20.68, 20.60, 20.45; exact mass calc'd for C20H26N3O14: 532.1415; found: 532.1413. 2',4'-Dinitrophenyl 2,3,4,6-tetra-O-acetyl-a-L-idopyranoside (35) Hemiacetal 23 (105 mg, 0.301 mmol) was dissolved in AfAf-dimethyl formamide (2 mL) and fluoro-2,4-dinitrobenzene (59 mg, 0.32 mmol) and 1,4-diazabicyclo[2.2.2]octane (DABCO) (106 mg, 0.94 mmol) were added. After stirring for 2 hours, the reaction mixture was diluted to 40 mL with dichloromethane and was washed with brine (2x15 mL), 1 M HC1 (2 x 10 mL), saturated sodium bicarbonate (20 mL) and water (20 mL). The organic layer was dried over MgS04, filtered and concentrated in vacuo. The resulting syrup was subjected to column chromatography (solvent B). The first compound to elute was the desired a-anomer (28 mg, 0.054 mmol, 17%). lH NMR (CDCI3) 8 8.71 (d, 1 H, Jy,y = 2.8 Hz, H(3')), 8.40 (dd, 1 H, Jy,6< = 9.3, J5%y = 2.8 Hz, H(5')), 7.50 (d, 1 H, J&ty = 9.3 Hz, H(6')), 5.78 (s, 1 H, H(l)), 5.07-5.10 (m, 1 H, H(3)), 4.98-5.02 (m, 1 H, H(2)), 4.90-4.93 (m, 1 H, H(4)), 4.46-4.53 (m, 1 H, H(5)), 4.18 (dd, 1 H, J 6 a , 5 = 7, J6a,6b = 11.8 Hz, H(6a)), 4.13 (dd, 1 H, y 6 b j 5 = 5.6, J6bM = 11.6Hz,H(6b)), 2.21,2.16, 2.15, 1.91 (4xs, 12 H, 4 x OAc). 107 4.1.6 Deprotections Deprotections were done according to the general procedure outlined in §4.1.2.1 4'-Methoxyphenyl a-L-idopyranoside (41) Protected idoside 37 (800 mg, 1.76 mmol) was reacted with sodium (162 mg, 7.04 mmol) in methanol, according to the general procedure. Silica gel flash column chromatography produced a white solid (385 mg, 1.34 mmol, 76%). *H NMR (CD3OD) 8 7.09, 6.83 (AA'XX1 system, 4 H, H(2'), H(3'), H(5'), H(6')), 5.27 (d, 1 H, Jia = 3.1 Hz, H(l)), 4.22 (dt, 1 H, J= 2.6, 6.0 Hz), 3.84 (t, 1 H, J = 5.2 Hz), 3.70-3.78 (m, 7 H), 1 3 C NMR (75 MHz, D 20) 8 154.94, 150.63, 119.35, 115.17, 100.38, 72.14, 71.74, 71.06, 70.29, 59.84, 55.93; Ci 3 Hi 8 0 7 calc: C, 54.54; H, 6.34; found: C, 52.58; H, 6.44 %; exact mass calc'd. for Ci 3 Hi 9 0 7 : 287.1130; found: 287.1129. Phenyl a-L-idopyranoside (42) Protected idoside 25 (440 mg, 1.037 mmol) was suspended in methanol and sodium (95 mg, 4.15 mmol) was added, according to the general procedure. The desired product was obtained as a colourless syrup (184 mg, 0.718 mmol, 69 %). *H NMR (CD3OD) 8 7.27 (m, 2 H), 7.15 (m, 2 H), 6.98 (t, 1 H, J= 7.4 Hz), 5.41 (d, 1 H, Jia = 3.4 Hz, H(l)), 4.20 (dt, 1 H, J = 2.7, 6.0 Hz), 3.86 (t, 1 H, J = 5.0 Hz), 3.72-3.80 (m, 4 H), 1 3C NMR (100 MHz, CD3OD) 8 158.69, 130.37, 123.23, 118.19, 101.07, 72.43, 72.01, 71.59, 71.29, 61.99; exact mass calc'd. for Ci 2 Hi 7 0 6 : 257.1025; found: 257.1026. 108 4'-Chlorophenyl a-L-idopyranoside (43) Protected idoside 38 (425 mg, 0.926 mmol) was dissolved in methanol and treated with sodium (162 mg, 7.04 mmol) as detailed in the general procedure. Silica gel flash column chromatography, followed by crystallization from water gave the product as an off-white crystalline solid (232 mg, 0.798 mmol, 86 %). ! H NMR (CD3OD) 5 7.25, 7.14 (AA'XX1 system, 4 H, H(2'), H(3'), H(5'), H(6')), 5.37 (d, 1 H, J u = 3.0 Hz, H(l)), 4.15 (dt, 1 H, J= 2.9, 6.0 Hz), 3.84 (t, 1 H, J= 5.0 Hz), 3.70-3.76 (m, 4 H), 1 3 C NMR (75 MHz, CD 3OD) 8 157.40, 130.22, 128.06, 119.63, 101.10, 72.39, 71.93, 71.89, 71.21, 61.87; exact mass calc'd. for Ci 2 Hi 6 0 6 3 5 Cl: 291.0635; found: 290.0633. 3'-Nitrophenyl a-L-idopyranoside (44) Protected idoside 39 (708 mg, 1.51 mmol) was deprotected with sodium (69 mg, 3.01 mmol) in methanol, as described in the general procedure. Chromatographic purification yielded the desired product 0 (390 mg, 1.29 mmol, 86 %) as a pale yellow syrup. *H NMR (CD3OD) 8 8.03 (t, 1 H, JTfi = J?,* = 2.2 Hz, H(2')), 7.88 (ddd, 1 H, J = 1.4, 2.2, 7.9 Hz, H(4')), 7.56 (ddd, 1 H, J= 1.4, 2.2, 7.9 Hz, H(6')), 7.52 (t, 1 H, J= 7.9 Hz, H(5')), 5.52 (d, 1 H, J 1 > 2 = 3.3 Hz, H(l)), 4.16 (dt, 1 H, J= 3.0, 6.0 Hz), 3.87 (t, 1 H, J = 5.2 Hz), 3.72-3.78 (m, 4 H), 1 3 C NMR (100 MHz, CD3OD) 8 159.12, 150.52, 131.28, 124.47, 117.78, 112.88, 101.15, 72.39, 72.23, 71.86, 71.22, 61.82; exact mass calc'd for Ci 2 Hi 6 NOi 4 : 302.0875; found: 302.0871. 109 4'-Nitrophenyl a-L-idopyranoside (45) Protected idoside 26 (455 mg, 0.969 mmol) was dissolved in methanol, and sodium (89 mg, 3.88 mmol) was added and the reaction was carried out according to the general procedure. The desired product was isolated as a pale yellow syrup (227 mg, 0.754 mmol, 78 %). ! H NMR (CD3OD) 6 8.20 (d, 2 H, J = 9.5 Hz), 7.31 (d, 2 H, J= 9.5 Hz), 5.56 (d, 1 H, Jia = 3.3 Hz, H(l)), 4.1 (dt, 1 H, 7= 2.9, 6.0 Hz), 3.86 (t, 1 H, J= 5.2 Hz), 3.77-3.71 (m,4H). 3',5'-Nitrophenyl a-L-idopyranoside (46) Protected idoside 40 (769 mg, 1.49 mmol) was suspended in methanol (20 mL) and sodium was added (137 mg, 5.98 mmol). Reaction and workup according to the general procedure yielded the desired product as a pale yellow syrup (385 mg, 1.11 mmol, 74 %). *H NMR (CD3OD) 8 8.61 (t, 1 H, J4>,& = J4;r = 1.9 Hz, H(4')), 8.39 (dd, 2 H, 7= 1.9 Hz, H(2'), H(6')), 5.63 (d, 1 H, Jy,2 = 3.3 Hz, H(l)), 4.13 (ddd, 1 H, J= 2.7, 5.1, 4.9 Hz), 3.87 (t, 1 H, J= 5.2 Hz), 3.72-3.81 (m, 4 H); 1 3 C NMR (100 MHz, CD3OD) 8 159.50, 150.56, 118.59, 112.68, 101.40, 72.68, 72.29, 71.65, 71.20, 61.82. 4.1.7 Oxidations Oxidations were done according to the general procedure outlined in §4.1.2.2. 4'-Methoxyphenyl a-L-idopyranosiduronate (ammonium salt) (47) 4-Methoxyphenyl-a-L-idopyranoside 41 (30 mg, 0.105 mmol) was oxidized according to the general procedure, with the following exceptions. The amount of TEMPO was decreased to 0.04 equivalents (0.6 mg, 0.004 mmol) and the reaction was 110 allowed to stir for 36 hours at 0 °C. The desired product was obtained as a mixture with the buffer salts, as the column failed to provide adequate separation for this compound. Characterization was performed on the material as isolated, without further purification, and a yield was not determined. *H NMR (D20) 8 7.25 (d, 2U,J = 9.2 Hz), 7.04 (d, 2 H, J= 9.2 Hz), 5.46 (d, 1 H, Jh2 = 4.5 Hz, H(l)), 4.61 (d, 1 H, J = 4.0 Hz), 3.96-4.00 (m, 1 H), 3.85-3.89 (m, 4 H), 3.77-3.81 (m, 1 H); 1 3 C NMR (100 MHz, D 20) 8 176.31, 155.06, 150.85, 119.24, 115.52, 100.39, 72.04, 71.51, 70.88, 70.78, 56.36. 4'-Chlorophenyl a-L-idopyranosiduronate (ammonium salt) (48) Idoside 42 (60 mg, 0.206 mmol) was oxidized as described in the general procedure, except that a larger volume of deionized water (2.5 mL) was required to dissolve the idoside. Because the final volume of the reaction mixture was greater than the maximum recommended loading volume for the column, the solution was divided in half, and the column was run twice. The desired product was isolated as a white amorphous solid (15 mg, 0.047 mmol, 23%). *H NMR (D20) 8 7.36 (d, 2 H, J = 8.9 Hz), 7.22 (d, 2 H, J= 8.9 Hz), 5.47 (d, 1 H, J1>2 = 5.4 Hz, H(l)), 4.51 (d, 1 H, J5,4 = 4.4 Hz, H(5)), 3.89 (dd, 1 H, J4,5 = 4.4, J4,3 = 7.1 Hz, H(4)), 3.78 (t, 1 H, J3A = J3,2 = 7.1 Hz, H(3)), 3.69 (dd, 1 H, J2,3 = 7.1, J2A = 5.4 Hz, H(2)); 13C NMR (100 MHz, D20) 8 175.98, 155.40, 129.85, 127.64, 118.68, 99.18, 72.32, 71.94, 71.40, 70.85; Exact mass calc'd for Ci 2Hi 2 3 5C10 7: 303.0271; found: 303.0277. 3'-Nitrophenyl a-L-idopyranosiduronate (ammonium salt) (49) Idoside 44 (40 mg, 0.133 mmol) was oxidized as described in general procedure. The desired product was obtained as an off-white amorphous solid (3.3 mg, 0.010 mmol, 111 7.5 %). *H NMR (D20) 5 8.22 (t, 1 H, JVfi = JTA = 2.1 Hz, H(2')), 7.92-7.97 (m, 1 H, H(4')), 7.59-7.65 (m, 1 H, H(6')) 7.56 (t, 1 H, Jy,& = JyA = 8.2 Hz, H(5')), 5.54 (d, 1 H, Ji, 2 = 5.6 Hz, H(l)), 4.51 (d, 1 H, J5A = 4.7 Hz, H(5)), 3.89 (dd, 1 H, J4,5 = 4.8, J4,3 = 7.6 Hz, H(4)), 3.77 (t, 1 H, J3A = J3,2 = 7.6 Hz, H(3)), 3.71 (dd, 1 H, J2,3 = 7.6, J 2 > i = 5.6 Hz, H(2)); 1 3 C NMR (100 MHz, D 20) 8 176.01, 157.06, 149.08, 130.80, 123.92, 118.10, 111.74, 98.74, 72.44, 72.00, 71.45, 70.81; exact mass calc'd for Ci 2 Hi 2 N0 9 : 314.0512; found: 314.0505. 4'-Nitrophenyl a-L-idopyranosiduronate (ammonium salt) (50) Idoside 26 (58 mg, 0.192 mmol) was oxidized according to the general procedure with the exception that, due to the relative insolubility of the idoside, a larger volume of water (2.5 mL) was required to dissolve it. Because the final volume of the reaction mixture was greater than the maximum recommended loading volume for the column, the solution was divided in half, and the column was run twice. The desired iduronide was obtained as a pale yellow amorphous solid, which still contained substantial buffer salt contamination. This product was repurified using the same technique (3.8 mg, 0.012 mmol, 6%). ! H NMR (D20) 8 8.23 (d, 2 H, J = 9.2 Hz), 7.37 (d, 2 H, J= 9.2 Hz), 5.60 (d, 1 H, JU2 = 5.2 Hz, H(l)), 4.47 (d, 1 H, J5A = 4.5 Hz, H(5)), 3.88 (dd, 1 H, J4,5 = 4.5, J4,3 = 6.5 Hz, H(4)), 3.77 (m, 1 H, H(3)), 3.72 (dd, 1 H, 72>3 = 7.2, J2,i = 5.2 Hz, H(2)); 1 3 C NMR (100 MHz, D 20) 8 175.94, 162.10, 142.70, 126.40, 116.94, 98.503, 72.21, 72.08,71.16,70.73; 112 3',5'-Nitrophenyl a-L-idopyranosiduronate (ammonium salt) (51) Idoside 46 (60 mg, 0.173 mmol) was oxidized according to the general procedure. The desired iduronide was obtained as an off-white amorphous solid (8 mg, 0.021 mmol, 12%). ] H NMR (D20) 8 8.72 (s, 1 H, H(4')), 8.54 (s, 2 H, H(2'), H(6')), 5.57 (d, 1 H, Jia = 4.8 Hz, H(l)), 4.49 (d, 1 H, J5A = 4.6 Hz, H(5)), 3.86-3.90 (m, 1 H), 3.70-3.77 (m, 2 H); 13C NMR (100 MHz, D20) 5 175.945, 157.69, 149.20, 117.92, 113.14, 98.88, 72.46, 72.08, 71.30, 70.76; Exact mass calc'd for CnHnNjOn: 359.0362; found: 359.0362. 4.2 Enzymology 4.2.1 General Procedures Human a-L-iduronidase was generously provided by Dr. John J. Hopwood of the Centre for Medical Genetics at the Women's and Children's Hospital. Adelaide, Australia. The enzyme was obtained as described in (63), and stored in pH 7.0 phosphate buffer containing 0.5 M NaCl. All absorbance measurements were made using a Pye-Unicam PU-8800 spectrophotometer, maintained at 37 °C through the use of a circulating water bath. Quartz cells with a 1 cm path length were used for all measurements. Fluorescence measurements were made at room temperature on either a Perkin Elmer Luminescence Spectrometer LB50B or a Perkin Elmer Luminescence Spectrometer LS 5B, attached to a PE7500 professional computer, using plastic cells that allow low wavelength transmission. The cells were oriented such that the excitation pathlength was 1 cm, and 113 the emission pathlength was 0.5 cm. Standard curves were constructed for each instrument, and each batch of stop buffer used. 4-Methylumbelliferyl-a-L-idopyranosiduronic acid (cyclohexylammonium salt) (MUI) was obtained from Sigma. 5-Fluoro-a-L-idopyranosiduronic acid fluoride was generously provided by A.W. Wong. All other substrates were obtained synthetically, as detailed outlined in § 4.1. Km and V m a x parameters were determined by non-linear regression using the GraFit 4.0 software. 4.2.2 Stereochemistry experiments For the J H NMR-monitored substrate hydrolysis experiment, the assay buffer consisted of 20 mM 3,3-dimethyl glutarate buffer, made up in D2O. The desired pD of 4.5 was obtained by titration such that the reading on the pH meter was 4.1 (based on the equation pD = pH + 0.4, where pH is the reading on a pH meter). 1 mg MUI was dissolved in a portion of the deuterated buffer, and the solution was repeatedly lyophilized from 99.96% D 20. The final lyophilization was from 99.996% D 2 0, into which the residue was also redissolved. The enzyme was transferred to the deuterated buffer using a 50 kDa molecular mass cutoff centrifugal concentrator (Amicon Microcon-50). The hydrolysis experiments were performed at 25 °C using a Varian UNITY Spectrometer, operating at 500 MHz. A spectrum of the substrate was first obtained, and following enzyme addition, spectra were obtained at approximately 1 minute intervals. Substrate methanolysis was first tested on a small scale. The products were separated by tic on aluminum-backed silica plates, using 3:1:1 butanol: acetic acid: water. 114 The plates were dried and visualized first with UV light, and then by staining with ammonium molybdate. For the NMR-scale reaction, the reaction was run in 20 mM 3,3-dimethyl glutarate, pH 4.5 containing 3.0 M methanol, 1 mM MUI and 0.016 ug iduronidase in a total volume of 600 uL. The reaction was incubated at 37 °G for 60 minutes and the enzyme was removed by centrifugation through a 50 kDa molecular mass cutoff centrifugal concentrator (Amicon Microcon-50). The filtrate was lyophilized and redissolved in 99.96% D 2 0 several times. The final lyophilization was done from 99.996% D 2 0, into which the residue was redissolved. The *H NMR spectrum of the resulting solution was obtained on a Bruker WH-400 spectrometer at 400 MHz at room temperature. 4.2.3 Sequence alignments Sequence alignments for the family 39 members were performed using the Clustalw alignment program (found at http://www2.ebi.ac.uk/clustalw/), sponsored by the European Bioinformatics Institute. Each of the four available matrices (id, pam, blosum and md) was tested at both the highest and lowest stringencies. The matrix id, at the lowest stringency settings was used to generate Figure 3.7. The md matrix was not successful at aligning the sequences, and only reported error messages. 4.2.4 Kinetics with 5F-IdoAF When tested as an inactivator, 5F-IdoAF (0.2 mM and 1 mM) was incubated in Buffer A (100 mM 3,3-dimethyl glutarate, pH 4.5 containing 0.1 % BSA) with 0.075 ug iduronidase (100 u.L total volume) at 37 °C. Residual enzyme activity was tested periodically by diluting 10 uL aliquots of the inactivation mix into 90 uL substrate mix 115 (containing and 50 mM MUI in Buffer A). Five minutes following the addition of the aliquot of the inhibitor mix to the substrate mix, an 80 uL aliquot of the assay was diluted into 1.4 mL of stop buffer (100 mM glycine carbonate, pH 10.7), and the fluorescence was measured. To test 5F-IdoAF as a substrate, fluoride release was monitored using an Orion 96-09 combination fluoride ion-selective electrode fitted to a signal amplifier box, which was connected to a personal computer. Data was collected using Logger Pro™ (Vernier Software Inc). All measurements were made at 37 °C. Buffer A containing 1 mM 5F-IdoAF was incubated at 37 °C for 10 minutes, monitoring fluoride release to determine the rate of spontaneous hydrolysis. Iduronidase, which had been preincubated at 37 °C, was added. Hydrolysis was monitored until all substrate was consumed. 4.2.5 Labeling and mass spectrometric analysis Iduronidase (30 ug) was transferred to sodium acetate buffer (100 mM, pH 4.5) using a 50 kDa molecular mass cutoff centrifugal concentrator (Amicon Microcon-50). 5F-IdoAF was added to a concentration of 12.5 mM and the mixture was incubated at room temperature for 30 s. The pH was brought to approximately 2 by the addition of 4 uL phosphate buffer (pH 1.68, 2 M) and pepsin (1.5 ug) was added immediately. After 1 hour at 37 °C, the digestion mixture was frozen. An unlabeled control was also performed using 10 ug of iduronidase and 0.5 ug pepsin in the absence of 5F-IdoAF. The peptide fragments were purified by reverse-phase high pressure liquid chromatography on an Ultrafast Microprotein Analyzer which was directly interfaced with the mass spectrometer (PE-Sciex API 300 triple quadrupole, equipped with an Ionspray ion source). The pepsin digests were loaded directly onto a C-l8 column 116 (Reliasil, 1 x 150 mm), which had been equilibrated with solvent A (0.05 % trifluoroacetic acid (TFA) and 2% acetonitrile in water). The peptides were eluted with a 60 minute gradient to 100 % solvent B (0.045 % TFA and 80 % acetonitrile in water), followed by 100 % solvent B for 2 min. The solvents were pumped at a constant flow rate of 50 uL/min. In the single quadrupole mode (LCMS) the quadrupole was scanned over a m/z range of 400 to 1800 Da, with a step size of 1.5 Da, and a dwell time of 0.5 ms. The orifice energy was 45 V, and the ion source voltage was 5.5 kV. In the tandem MS daughter-ion scan mode, the spectrum was obtained by selectively introducing the parent ion (m/z = 893) from the first quadrupole (Ql) into the collision cell (Q2), and observing the product ions in the third quadrupole. The scan range for the Q3 was 50 to 1806, the step size was 0.5 Da, and the dwell time was 1 ms. The ion source voltage was 5 kV and the orifice energy was 45 V; Q0 = -10 V, IQ2 = -48 V. To increase the extent of fragmentation of the parent ion, the voltage difference between Q2 and Q0 was altered. 4.2.6 Linear free energy relationships All assays were performed at 37 °C in Buffer C (20 mM 3,3-dimethyl glutarate, pH 4.50, containing 0.01 % BSA). Due to the insensitivity of the assays, it was not possible to limit the hydrolysis of substrate to the ideal 10 %, however, for each substrate, it was shown that the rate of substrate hydrolysis was constant over the time period of the assay. Once linearity had been proven, the rates were obtained by a 2-point stopped assay. The spectrophotometer was "zeroed" on a quartz cuvette containing 130 uL Buffer D (1 M glycine-carbonate, pH 10.7). An aliquot of the appropriate volume (50 to 240 uL, depending on the extinction coefficient of the substrate) was taken from an assay mixture 117 2 minutes following the addition of enzyme to the assay mixture. After reading the absorbance, the process was repeated to obtain a reading 10 minutes following enzyme addition. The difference between the two absorbances gave the rate in absorbance units per 8 min. The concentration of substrate stocks was determined by total hydrolysis. An aliquot of each substrate was incubated with 0.4 ug iduronidase for 6 hours. An additional 4 ug enzyme was added and incubation was continued for another hour, after which time, the assay mixture was diluted into buffer D and the absorbance was read. Each assay was performed in triplicate, and the results were averaged. To determine the wavelength at which the absorbance readings were to be taken, and the extinction coefficients of the phenols released, solutions of the phenols were made to known concentrations. To ensure accurate weighing, the phenols were first dried under vacuum over P2Os for 24 hours, protected from light. Each stock phenol was diluted into a combination of Buffers C and D that was representative of the buffer mixture present in the cuvette for the absorbance readings in the assays, and wavelength scans were performed. The wavelength at which maximal absorbance occurred for each phenol, outside of the absorbance range of proteins (> 290 nm), was used for all further readings. In the same buffer composition, at the predetermined wavelength, the absorbance was read for three different concentrations of each phenol. On the graph of Absorbance Units vs. [phenol] the slope provides the extinction coefficients (M"1). 118 REFERENCES 1. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K , and Watson, J. D. (1989) Molecular Biology of the Cell, 2 ed., Garland Publishing, Inc., New York. 2. Gahmberg, C. G., and Tolvanen, M. (1996) Trends in Biochem. Sci. 21, 308-311 3. Clements, P. R., Muller, V., and Hopwood, J. J. (1985) FEBSLett. 152, 29-34. 4. Lindahl, U., Kusche, M., Lidholt, K., and Oscarsson, L.-G. (1989) Ann. N.Y. Acad Sci. 556, 36-50. 5. Kjellen, L., and Lindahl, U. (1991) Annu. Rev. Biochem 60, 443-475. 6. Ofosu, F. A., Danishefsky, I., and Hirsh, J. (1989) Heparin and Related Polysaccharides: structure and activities, New York Academy of Sciences, New York. 7. Whitfield, D. M., and Sarkar, B. (1991) J. Inorg. Biochem. 41, 157-170. 8. Ernst, S., Venkataraman, G., Sasisekharan, V., Langer, R., Cooney, C. L., and Sasisekharan, R. (1998)7. Am. Chem. Soc. 120, 2099-2107. 9. Inoue, Y., Inouye, Y., and Nagasawa, K. (1990) Biochem. J. 265, 533-538. 10. Rao, V. S. R., Balaji, P. V., and Qasba, P. K. (1995) Glycobiology 5, 273-279. 11. Sanderson, P. N , Hucherby, T. N , and Nieduszynski, I. A. (1987) Biochem. J. 243, 175-181. 12. Ferro, D. R., Provasoli, A., Ragazzi, M., Torri, G., Casu, B., Gatti, G., Jacquinet, J.-C , Sinay, P., Petitou, M., and Choay, J. (1986) J. Am. Chem. Soc. 108, 6773-6778. 13. Casu, B., Petitou, M., Provasoli, M., and Sinay, P. (1988) Trends in Biochem. Sci. 13, 221-225. 14. Kresse, H , and Glossl, J. (1987) Adv. Enzymol. Relat. Areas Mol. Biol. 60, 217-311. 15. Russell, C , Gendson, G., Jevon, G., Matlock, T., Yu, J., Aklujkar, M., Ng, K.-Y., and Clarke, L. A. (1998) Clin. Genet. 53, 349-361. 16. Shull, R. M., Kakkis, E. D., McEntee, M. F., Kania, S. A., Jonas, A. J., and Neufeld, E. F. (1994) Proc. Natl. Acad Sci. 91, 12937-12941. 17. Henrissat, B. (1991) Biochem. J. 280 (Pt 2), 309-316. 18. Henrissat, B., and Bairoch, A. (1993) Biochem. J. 293 (Pt 3), 781-788. 19. Scott, H. S., Ashton, L. J., Evre, H. J., Baker, E., Brooks, D. A., Callen, D. F., Sutherland, G. R., Morris, C. P., and Hopwood, J. J. (1990) Am. J. Hum. Genet. 47, 802-807. 20. MacDonald, M. E., Scott, H. S., Whaley, W. L., Pohl, T., Wasmuth, J. J., Lehrach, H., Morris, C. P., Frischauf, A.-M., Hopwood, J. J., and Gusella, J. F. (1991) Somatic Cell and Molecular Genetics 17, 421 -425. 21. Scott, H. S., Anson, D. S., Orsborn, A. M., Nelson, P. V., Clements, P. R., Morris, C. P., and Hopwood, J. J. (1991) Proc. Natl. Acad Sci. 88, 9695-9699. 22. Armand, S., Vieille, C , Gey, C , Heyraud, A., Zeidus, J. G., and Henrissat, B. (1996) Eur. J. Biochem. 236, 701-713. 23. Koshland, D. E. (1953) Biological Reviews 28, 416-436. 119 24. Knapp, S., Vocadlo, D., Gao, Z., Kirk, B., Lou, J., and Withers, S. G. (1996) 7. Am. Chem. Soc. 118, 6804-6805. 25. Sinnott, M. L., and Souchard, I. J. L. (1973) Biochem. J. 133, 89-98. 26. Kempton, J. B., and Withers, S. G. (1992) Biochemistry 31, 9961-9969. 27. Sidhu, G., Withers, S. G., Nguyen, N. T., Mcintosh, L. P., Ziser, L., and Brayer, G. D. (1999) Biochemistry 38, 5346-5354. 28. Sabini, E., Sulzenbacher, G., Dauter, M., Dauter, Z., Jorgensen, P. L., Schulein, M., Dupont, C., Davies, G. J., and Wilson, K. S. (1999) Chemistry and Biology 6, 483-492. 29. Howard, S., He, S., and Withers, S. G. (1998) 7. Biol. Chem. 273, 2067-2072. 30. McCarter, J. D., Burgoyne, D. L., Miao, S., Zhang, S., Callahan, J. W., and Withers, S. G. (1997) J . Biol. Chem. 272, 396-400^ 31. Miao, S., Ziser, L., Aebersold, R., and Withers, S. G. (1994) Biochemistry 33, 7027-7032. 32. Miao, S., McCarter, J. D., Grace, M. E., Grabowski, G. A., Aebersold, R., and Withers, S. G. (1994)7. Biol. Chem. 269, 10975-10978. 33. Tull, D., Withers, S. G., Gilkes, N. R., Kilburn, D. G., R.A.J., W., and Aebersold, R. (1991)7. Biol. Chem. 266, 15621-15625. 34. Vocadlo, D. J., MacKenzie, L. F., He, S., Zeikus, G. J., and Withers, S. G. (1998) Biochem. J. 335, 449-455. 35. Wong, A. W., He, S., Brubb, J. H., Sly, W. S., and Withers, S. G. (1998) 7. Biol. Chem. 273, 34057-34062. 36. McCarter, J. D., and Withers, S. G. (1996) 7. Biol. Chem. 271, 6889-6894. 37. Lowry, T. H., and Richardson, K. S. (1987) Mechanism and Theory in Organic Chemistry, HarperCollinsPwZ>//sAer.s, New York. 38. Chiba, T., and Sinay, P. (1986) Carbohydr. Res. 151, 379-389. 39. Baggett, N , Samra, A. K., and Smithson, A. (1983) Carbohydr. Res. 124, 63-75. 40. Igarashi, K. (1977) Adv. Carbohydr. Chem. Biochem. 34, 243-283. 41. Perchemlides, P., Osawa, T., Davidson, E. A., and Jeanloz, R. W. (1967) Carbohydr. Res. 3, 463-477. 42. de Nooy, A. E. J., Besemer, A. D., and van Bekkum, H. (1995) Tetrahedron 51, 8023-8032. 43. de Nooy, A. E. J., Besemer, A. D., and van Bekkum, H. (1995) Carbohydr. Res. 269, 89-98. 44. de Nooy, A. E. J., Besemer, A. D., and van Bekkum, H. (1995) Synthesis Stuttgart 70,1153-1174. 45. Adinolfi, M., Barone, G., De Lorenzo, F., and Iadonisi, A. (1999) Synlett 3, 336-338. 46. Dromowicz, M., and Koll, P. (1998) Carbohydr. Res. 308, 169-171. 47. Ferrier, R. J., and Tyler, P. C. (1980)7. Chem. Soc, Perkin Trans. 1, 1528-1534. 48. Kochetkov, N. K. (1987) Tetrahedron 43, 2389-2436. 49. Betaneli, V. I., Ovchinnikov, M. V., Backinowsky, L. V., and Kochetkov, N. K. (1979) Carbohydr. Res. 68, C11-C13. 50. Meerwein, H., Hederich, V., H., M., and Wunderlich, K. (1960) Lieb. Ann. 635, 1-21. 120 51. Betaneli, V. L, Litvak, M. M , Backinowsky, L. V., and Kochetkov, N. K. (1981) Carbohydr. Res. 94, C1-C4. 52. Thompson, C , Ge, M , and Kahne, D. (1999) J. Am. Chem. Soc. 121, 1237-1244. 53. Seeberger, P. H., Eckhardt, M , Gutteridge, C. E., and Danishefsky, S. J. (1997) J. Am. Chem. Soc. 119, 10064-10072. 54. Crich, D., and Sanxing, S. (1997) J. Am. Chem. Soc. 119, 11217-11223. 55. Gildersleeve, J., A., S., Sakurai, K., S., R., and Kahne, D. (1999) J. Am. Chem. Soc. 121, 6176-6182. 56. van Boeckel, C. A. A., and Beetz, T. (1985) Reel. Trav. Chim. Pays-Bos 104, 171-173. 57. Schmidt, R. R. (1994) Adv. Carbohydr. Chem. Biochem. 50, 21-123. 58. Koeners, H. J., de Kok, A. J., Romers, C , and van Boom, J. H. (1980) Reel. Trav. Chim. Pays-Bos 99, 355-362. 59. Zemplen, G., and Pascu, E. (1929)Ber. Dtsch. Chem. Gesell. 62, 1613. 60. Howard, S., Braun, C , McCarter, J., Moremen, K. W., Liao, Y.-F., and Withers, S.G. (1997) Biochem. Biophys. Res. Commun. 238, 896-898. 61. Whitfield, D. M., Birnbaum, G. I., Pang, H , Baptista, J., and Sarkar, B. (1991) J. Carbohydr. Chem. 10, 329-348. 62. Durand, P., Lehn, P., Callebaut, I., Fabrega, S., Henrissat, B., and Mornon, J.-P. (1997) Glycobiology 7, 277-284. 63. Unger, E. G., Durrant, J., and Hopwood, J. J. (1994) Biochem. J. 304, 43-49. 64. Zhao, K.-W., Faull, K. F., Kakkis, E. D., and Neufeld, E. F. (1997) J. Biol. Chem. 272, 22758-22765. 65. Ruth, L., Eisenberg, D., and Neufeld, E. F. (2000) Acta Crystallogr., Sect D: Biol. Crystallogr. 56, 524-528. 121 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061396/manifest

Comment

Related Items