Iminosugars As Glycosidase Inhibitors: Nojirimycin and Beyond

Arnold E. Stiitz

Iminosugars as Glycosidase Inhibitors Nojirimycin and Beyond

WILEY-VCH Inrinosiigars as Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

Further Titles of Interest

Yves Chapleur (Ed.) Carbohydrate Mimics - Concepts and Methods XXVIII, 604 pages with 375 figures and 52 tables 1998, cloth, ISBN 3-527-29526-7 H. van Bekkum, H. Roper, A. G. J. Vorhagen (Eds.) Carbohydrates as Organic Raw Materials III X, 315 pages with 156 figures and 68 tables 1996, cloth, ISBN 3-527-30079-1

Arnold E. Stiitz

Iminosugars as Glycosidase Inibitors Nojirimycin and Beyond

WILEY-VCH Weinheim · New York · Chichester Brisbane · Singapore · Toronto

Univ.-Prof. Dipl.-Ing. Dr. Arnold E. Stiitz Professor for Organic Chemistry Institute for Organic Chemistry Technical University Graz Stremayrgasse 16 A-8010 Graz Austria

This book was carefully produced. Nevertheless, authors, editor, and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

Cover Illustration: Convolvulus arvensis and selected structures of iminosugars that have been found in plants. Library of Congress Card No. applied for.

A catalogue record for this book is available from the British Library.

Deutsche Bibliothek Cataloguing-in-Publication Data: Stiitz, Anrold E.: Iminosugars as glycosidase inhibitors : Nojirimycin and beyond / Arnold E. Stiitz. - 1. Aufl. - Weinheim ; New York ; Chichester ; Brisbane ; Singapore : Wiley-VCH, 1999 ISBN 3-527-29544-5

© WILEY-VCH Verlag GmbH. D-69469 Weinheim (Federal Republic of Germany), 1999 Printed on acid-free and chlorine-free paper. All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form - by photoprinting, microfilm, or any other means - nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition: Hagedorn Kommunikation, D-68519 Viernheim. Printing: betz-druck GmbH, D-64291 Darmstadt. Bookbinding: J. Schaffer GmbH & Co. KG, D-67263 Griinstadt. Printed in the Federal Republic of Germany.

Preface

Glycoside hydrolases are a very important class of carbohydrate and glycoconjugate modifying enzymes. Their exploitation, for example, in bio- and food technology, wood chemistry, as well as in biochemical and medical research has become a major research area over the past three decades. Important contemporary applications also include their use for the enzymatic synthesis of a large variety of glycosides and complex oligosaccharides under environmentally benign conditions. Various types of inhibitors of these enzymes play a vital part in fundamental investigations into mechanistic aspects of enzymatic glycoside hydrolysis. Amongst the reversible inhibitors, sugar analogs with nitrogen instead of oxygen in the ring system have emerged as versatile tools for biochemists and cell biologists involved in basic and applied glycobiology and -technology. Some representatives of this class of compounds are already marketed as Pharmaceuticals against certain forms of diabetes, quite a few others exhibit promising anti-infective properties. The large variety of structures and their remarkable biological activities pose a pertinent challenge for scientists in many areas. When Dr. Kellersohn of VCH-Wiley suggested a book on this exiting topic, I aimed to tackle this challenge with some of the most renowned researchers in the field. In a team effort we have attempted to compile a well-balanced mixture of topics to provide a fairly comprehensive overview over the area we so much enjoy to work in. Many of the chapters provide a wide range of established information to help the reader getting aquainted with the topic as well as the relevant literature. Other chapters might be slightly controversial and are aimed to be a basis of discussion for the advanced. The close relationships between the natural products distribution and isolation, the synthetic chemists' work and biochemical research in the field have been well documented. I am grateful to all the authors for their faith in me and in the idea of this book. Dr. Anette Eckerle, Maike Petersen and all the other members of the publishers' team I would like to thank for their help and guidance throughout the technical parts of this project.

VI

Preface

At this point it is a great pleasure to highlight the most senior amongst the authors, Professor Hans Paulsen, whose early contributions to the chemistry and synthesis of the class of compounds under consideration have been vital for the rapid development of the field. Graz, September 1998

Arnold E. Stutz

Contents

1

2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.3 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5 2.4.6 2.4.7 2.4.8 2.5

The Early Days of Monosaccharides Containing Nitrogen in the Ring (Hans Paulsen) Taxonomic Distribution of Iminosugars in Plants and Their Biological Activities (Monique S. J. Simmonds, Geoffrey C. Kite, Elaine A. Porter) Introduction Structural Diversity of Iminosugars Piperidines Pyrrolidines Indolizidines Pyrrolizidines Ntfrtropanes Glycosides Chemotaxonomy Biological Interactions Inhibition of glycosidases Glycoprotein processing Anti-viral activity Lysosomal diseases Anti-cancer properties Effects on insects Nematicidal activity Plant growth regulatory activity Summary

1

8 8 8 8 11 12 13 14 15 15 17 19 19 20 21 22 22 25 26 26

VIII

Contents

3

Glycosidase Inhibition by Basic Sugar Analogs and the Transition State of Enzymatic Glycoside Hydrolysis (Gunter Legler) 3.1 Introduction 3.2 Transition State Structure and Inhibitory Potency 3.3 General Mechanisms of Enzymatic Glycoside Hydrolysis and Models of the Transition State 3.4 Basic Sugar Analogs as Glycosidase Inhibitors 3.4.1 Position of the basic (cationic) center 3.4.2 Basicity of the anomeric carbon atom 3.4.3 Geometry and charge distribution at the anomeric position 3.4.4 Hydroxylation, topography, ring size and flexibility as determinants of specificity 3.4.4.1 Five-membered azasugars 3.4.4.2 Seven-membered azasugars 3.4.4.3 Rigid azasugar derivatives 3.4.5 Interactions with the aglycon binding site 3.4.6 Hydrogen bond formation with the catalytic acid 3.5 Criteria for Transition State Resemblance 4 4.1 4.2 4.2.1 4.2.2 4.2.2.1 4.2.2.2 4.2.2.3 4.3 4.3.1 4.3.1.1 4.3.1.2 4.3.2 4.3.3 4.3.4 4.4

Synthetic Methods for the Preparation of Iminosugars (Barbara La Ferla, Francesco Nicotra) Introduction Iminosugars from 'True' Sugars Iminosugars from aminosugars Iminosugars via amination of 'true' sugars and subsequent cyclization Iminosugars via amination of the anomeric center Iminosugars via 'chain amination' Iminosugars from alditols Iminosugars from Noncarbohydrate Starting Materials Ex novo synthesis of iminosugars via acyclic stereoselection Formation of iminosugars by stereoselective chemical condensations Formation of iminosugars employing aldolases Iminosugars via cycloaddition reactions Iminosugars from cyclic dienes Iminosugars from other non-carbohydrate substrates Conclusions

31 31 32 34 38 38 42 44 49 50 53 54 56 59 61 68 68 69 71 72 72 76 79 81 81 81 84 85 86 88 90

Contents 5 5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.1.1 5.3.1.2 5.3.1.3 5.3.2 5.3.2.1 5.3.2.2 5.3.2.3 5.4 6 6.1 6.2 6.3 6.4 7 7.1 7.2 7.2.1 7.2.1.1 7.2.1.2 7.2.1.3 7.2.1.4 7.2.2 7.2.2.1 7.2.2.2 7.2.2.3 7.2.2.4 7.2.2.5 7.2.2.6

Iminosugars as Powerful Glycosidase Inhibitors Synthetic Approaches from AIdonolactones (Inge Lundt, Robert Madsen) Introduction Syntheses Iminosugars via iminoamides - strategy I Iminosugars via dibromoalditols - strategy II Iminosugars via lactams - strategy III Iminosugars via azidolactones - strategy IV Biochemical Evaluation l,4-Dideoxy-l,4-iminoalditols (pyrrolidines) Iminotetritols 2,5-Dideoxy-2,5-iminohexitols l,4-Dideoxy-l,4-iminohexitols l,5-Dideoxy-l,5-iminoalditols (piperidines) l,2,5-Trideoxy-l,5-iminopentitols l,5-Dideoxy-l,5-iminopentitols l,5-Dideoxy-l,5-iminoheptitols Conclusions Isoiminosugars: Glycosidase Inhibitors with Nitrogen at the Anomeric Position (Inge Lundt, Robert Madsen) The Anomeric Position Siastatin B and Analogs Isofagomine and Beyond Conclusions Synthesis and Biological Activity of Castanospermine and Close Analogs (Peter C. Tyler, Bryan G. Winchester) Introduction Synthesis of Castanospermine and Analogs Total syntheses Castanospermine and stereoisomers Pentahydroxyindolizidines Tetrahydroxyquinolizidines Miscellaneous Castanospermine analogs Syntheses from Castanospermine Selective protection of Castanospermine Reactions at C-6 Reactions at C-8 Reactions at C-7 Reactions at C-I Miscellaneous derivatives

IX

93 93 95 95 98 100 103 104 104 104 104 104 106 107 107 109 109 112 112 114 117 123 125 125 125 126 126 132 134 135 136 136 139 141 142 143 143

X

Contents

7.3 7.3.1 7.3.2 7.3.2.1 7.3.2.2 7.3.2.3 7.3.3 7.3.4 7.3.5 7.3.6

Biological Activities of Castanospermine and Derivatives Structure-activity relationships Enzymology of glucosidases Digestive glucosidases Lysosomal glucosidases Processing α-D-glucosidases Mechanism of action Anti-viral activity Anti-cancer activity Modulation of other cellular processes

8

Some Reflections on Structure-Activity Relationships in Glycosidase-Inhibiting Iminoalditols and Iminosugars (Arnold E. Stiitz) Introduction D-Galactosidase Inhibitors Piperidine derivatives 5 -Amino-5 -deoxy-o-galactose, 1,5 -dideoxy-1,5 -iminoD-galactitol (1-deoxygalactonojirimycin) and epimers Epimers of isofagomine Important pyrrolidine derivatives Tetrahydroxyazepanes Conclusions o-Mannosidase Inhibitors Piperidine derivatives Pyrrolidine derivatives 1,4-Dideoxy-1,4-imino-D-mannitol and derivatives Tetrahydroxyazepanes Bicyclic systems Swainsonine Miscellaneous Structure-activity relationships Inhibition of α-L-fucosidases by D-mannosidase inhibitors and related compounds o-Glucosidase Inhibitors Piperidine derivatives Nojirimycin and 1-deoxynojirimycin Glycosylated derivatives Structural alterations of ring substituents Castanospermine Summary Important pyrrolidine and pyrrolizidine derivatives 2,5-Dideoxy-2,5-imino-D-mannitol Australine Tetrahydroxyazepanes Isofagomine

8.1 8.2 8.2.1 8.2.1.1 8.2.1.2 8.2.2 8.2.3 8.2.4 8.3 8.3.1 8.3.2 8.3.2.1 8.3.3 8.3.4 8.3.4.1 8.3.4.2 8.3.5 8.3.6 8.4 8.4.1 8.4.1.1 8.4.1.2 8.4.1.3 8.4.2 8.4.2.1 8.4.3 8.4.3.1 8.4.3.2 8.4.4 8.4.5

144 145 146 146 147 148 149 149 150 151 157 157 159 159 159 160 160 161 161 161 161 162 162 164 164 164 165 165 167 168 168 168 169 170 172 173 173 173 175 176 176

Contents 8.4.6 8.4.7 8.4.8

Calystegines Common features and structure-activity relationships Conclusion and outlook

9

Potent Glycoside Inhibitors: Transition State Mimics or Simply Fortuitous Binders? (Stephen G. Withers, Mark Namchuk, Renee Most) Introduction Transition State Theory and Mimicry Glycosidase Mechanisms and Transition States Tight-Binding Glycosidase Inhibitors Probing Transition State Mimicry Nojirimycin tetrazoles Acarbose as a transition state analog? Castanospermine and deoxynojirimycin as transition state analogs? Conclusions

9.1 9.2 9.3 9.4 9.5 9.5.1 9.5.2 9.5.3 9.6

Iminoalditols as Affinity Ligands for the Purification of Glycosidases (Anna de Raadt, Christian Ekhart, Gunter Legler, Arnold E. Stutz) 10.1 Introduction 10.2 D-Glucosidases 10.2.1 Glucosidase I 10.2.2 Glucosylceramidase 10.2.3 Cytosolic β-glucosidase 10.2.4 Microsomal bile acid /?-glucosidase 10.3 o-Mannosidases 10.4 N-Acetyl-/J-D-hexosaminidases 10.5 α-L-Fucosidases 10.6 Miscellaneous 10.7 Conclusions

XI 177 177 180 188 188 189 191 194 197 197 199 202 203

10

11

Inhibitors of Glycoprotein Processing (Alan D. Elbein, Russell J. Molyneux) 11.1 Introduction 11.2 Chemistry of Alkaloid Glycosidase Inhibitors 11.2.1 Structural classes 11.2.2 Occurrence and isolation from natural sources 11.2.2.1 Occurrence 11.2.2.2 Isolation 11.3 Glycosidase Inhibition 11.3.1 Glycosidase inhibitory activity 11.3.2 Structure-activity relationships 11.3.3 Synthetic polyhydroxy alkaloids

207 207 208 208 209 209 210 210 211 212 213 214 216 216 216 216 218 218 219 220 220 222 224

XII

Contents

11.4 11.4.1 11.4.2 11.5 11.5.1 11.5.2 11.5.3 11.6 11.6.1 11.6.2 11.6.3

Biological Activity of Glycosidase Inhibitors Mammalian toxicity Therapeutic activity Processing of N-linked Oligosaccharides Introduction Biosynthesis of N-linked oligosaccharides Processing of N-linked oligosaccharides Inhibitors of N-linked Glycoprotein Processing Introduction Glucosidase inhibitors Mannosidase inhibitors

Tables of Glycosidase Inhibitors with Nitrogen in the Sugar Ring and Their Inhibitory Activities Index General Chemical compounds and substances Genera and species

224 224 226 227 227 229 230 236 236 236 240 253 391 391 393 396

List of contributors

A. D. Elbein University of Arkansas for Medical Sciences Department of Biochemistry/ Molecular Biology 4301 W. Markham Street Mail Slot 516 Little Rock, Arkansas 72205 USA G. Legler Institut fur Biochemie der Universitat KoIn Otto-Fischer-Str. 12-14 50674 KoIn Germany I. Lundt, Dr. R. Madsen Department of Organic Chemistry The Technical University of Denmark Building 201 DK-2800 Lyngby Denmark H. Paulsen Institut fur Organische Chemie Universitat Hamburg Martin-Luther-King-Platz 6 D-20146 Hamburg Deutschland

St. G. Withers, M. Namchuk, R. Mosi Department of Chemistry University of British Columbia Vancouver Brithish Columbia CANADA V6T IZl B. G. Winchester Institute of Child Health University of London 30 Guilford Street London WClN IEH United Kingdom R C. Tyler Carbohydrate Chemistry Group Industrial Research Limited PO Box 31310 Lower Hutt New Zealand M. S. J. Simmonds, G. C. Kite, E. A. Porter Royal Botanic Gardens, Kew Richmond Surrey United Kingdom

XIV

List of contributors

R Nicotra, B. La Ferla Universita degli Studi di Milano Dipartimento di Chimica Organica e Industriale Via Venezian 21 1-20133 Milano Italia R. J. Molyneux Western Regional Research Center ARS-USDA Albany, CA USA

A. de Raadt, C. W. Ekhart, M. H. Fechter, P. Hadwiger, E. Mlaker, A. E. Stiitz, A. Tauss, T. M. Wrodnigg Institut fur Organische Chemie Technische Universitat Graz Stremayrgasse 16 A-8010 Graz Austria

1

The Early Days of Monosaccharides Containing Nitrogen in the Ring HANS PAULSEN

The history of the first synthesis of monosaccharides with nitrogen in the ring as well as the discovery of their very strong inhibitory effect on glycoside-cleaving glycosidases much later is a fascinating chapter in carbohydrate chemistry [1,2]. Early attempts to replace the oxygen in pyranoses or furanoses by another hetero atom were made with sulfur. In 1961, the laboratories of J. C. Schwarz [3] and L. W. Owen [4] as well as of R. L. Whistler [5] succeeded independently and simultaneously in synthesizing 5-thio-o-xylopyranose, which contains sulfur instead of oxygen in the ring. The chemical properties of these sulfur-containing saccharides were very similar to those of the oxygen analogs. This is not surprising, since sulfur is divalent and, like oxygen, found in the second main group of the periodic table. The sulfur-containing sugars showed mutarotation and could be converted into the corresponding glycosides. R. L. Whistler then studied this substance class in detail [2]. Apart from sulfur-containing six-membered rings, sulfur-containing five-membered rings and also the corresponding sulfur-containing nucleosides could be synthesized [6]. The thio compounds did show one peculiarity; the sulfur in the ring could be oxidized to sulfoxides and sulfones. At that time we were interested in finding out whether the ring oxygen of pyranoses could also be replaced by nitrogen. The reactions in this case were not predictable, because nitrogen is trivalent and also has basic properties. Aldimines were also possible cyclization products. We first synthesized 5,6-diacetamidol,2-0-cyclohexylidene-5,6-dideoxyhexofuranoses of the n-gluco- and L-idoform, since it could be expected that the 5 -aminosugars would give an appropriate ring closure with nitrogen. Deblocking of the compounds under the required acidic conditions yielded, however, the 2-aminomethyl-5-hydroxypyridine in both cases [7]. This result showed that although a saccharide with nitrogen in the ring had apparently been formed as the intermediate product, it had been converted into the pyridine derivative under the acidic conditions with elimination of three water molecules. With this result the race was on to isolate the nitrogen-containing saccharides that had so far only been identified as intermediate products. We were more successful with 5-acetamido-l,2-O-cyclohexylidene-5-deoxy-Dxylofuranose. It could be deblocked under milder acidic conditions and we were able to isolate the crystalline 5-acetamido-5-deoxy-o-xylopypranose in 1962 [8]. Jniinosugfirs ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

2

1 The Early Days ofMonosaccharides Containing Nitrogen in the Ring

At the same time, the groups of J. K. N. Jones [9] and S. Hanessian [10] independently obtained the same crystalline substance by a slightly different pathway. In our enthusiasm we named this new substance class piperidinoses. Today, these compounds are often referred to by the trivial name azasugars. Melville L. WoIfrom, however, taught us better, i.e. that no new nomenclature was needed, as the rules already existing were fully adequate [2]. From the names 5-amino-5deoxyhexopyranose and 5-amino-5-deoxy-l,5-anhydrohexitol any chemist can deduce that these are compounds with nitrogen-containing rings. The N-acylated compounds with a ring nitrogen were interesting because the basicity of the nitrogen was weakened by the amide group. As a result, they exhibited chemical properties similar to those of the oxygen-containing compounds and, like them, could be used to synthesize glycosides [U]. Due to the partial carbonnitrogen double bond, however, the acetamido group was planar. This resulted in considerable steric hindrance for the /J-glycoside. Therefore the free sugar and the methylglycoside could be obtained only with axial groups at the anomeric center, i.e. in the α-glycosidic form [12]. Because of the partial carbon-nitrogen double bond, rotation of the amide group was hindered so that an Zs.Z-isomerism occurred, which could be detected very well in a saccharide derivative for the first time using temperature-dependent NMR spectroscopy [13]. This left the most important question unanswered, which was how monosaccharides would behave that had free, unsubstituted amino groups in the ring which apparently were extremely sensitive to acids. Corresponding N-alkylated derivatives were likewise converted to N-alkylpyridinium compounds with elimination of three water molecules on treatment with acid [14]. D. L. Ingles' observation that the aminosugar could be deblocked with sulfurous acid brought a breakthrough [15], With this method a sulfonic acid derivative of the aminosugar was isolated that could then be cleaved under mild alkaline conditions with barium hydroxide to saccharides with unsubstituted nitrogen in the ring. In 1965, my coworkers K. Todt and F. Leupold synthesized all the pentose isomers containing unsubstituted ring nitrogen and studied their properties [16, 17]. These compounds were stable only in neutral to weakly alkaline solution and eliminated water to give piperideines and final pyridines even in acid solution. Synthesis of glycosides was impossible for the same reason, since in methylglycosides methanol is cleaved directly. It was possible to easily convert all the compounds into the stable 1,5anhydropentitols with ring nitrogen by means of hydrogenation. Analogously, the 1,5-anhydrohexitols with ring nitrogen could be synthesized from 5-amino5-deoxyhexoses. Likewise, the glucose derivative of the antibiotic nojirimycin with ring nitrogen was isolated for the first time and subsequently synthesized by S. Inouye [18] via this sulfonic acid derivative. These results caused us to intensify our studies. K. Todt synthesized every 1,6anhydrohexose possible containing one or two nitrogen atoms in the ring [19] as well as a nitrogen-containing seven-membered ring sugar [2O]. In addition, we studied furanoses with ring nitrogen, some of which provided extremely complex dimerization products [21, 22]. Furthermore, saccharides containing one or two nitrogens in the ring could be synthesized with hydrazine [23]. In Canada, J. K. N. Jones, who personally in 1967 still impressively organized the 4th International

1 The Early Days of Monosaccharides Containing Nitrogen in the Ring

3

Conference on Carbohydrates in Kingston, Ontario, continued intensive collaboration with W. A. Szarek [24]. Unfortunately this collaboration was terminated by the very untimely death of J. K. N. Jones. Meanwhile S. Hanessian continued his intensive studies even after moving from Ann Arbor to Montreal [25]. At that time the main question remaining was how ketoses would behave. To avoid repetitious work we made an agreement with Steve Hanessian that he would work on the fructose [26] and we would take the sorbose. At the time we still had no idea that sorbose would later prove to be much more interesting. My co-worker, I. Sangster, converted the Reichstein intermediate of vitamin C synthesis, 2,3-0-isopropylidine-a-L-sorbofuranose, into a 6-amino-6-deoxy-Lsorbose [27]. He found that in concentrated hydrochloric acid solution this aminosugar formed an ammonium salt in which the sorbose is in the furanose form with oxygen in the ring. In neutral or weakly alkaline solution, however, a spontaneous extension of the sorbose ring took place to give the pyranose form with nitrogen in the ring. This form is in equilibrium with the corresponding water elimination product, the piperideine. With catalytic hydrogenation of this compound, he directly obtained the corresponding 2,6-anhydro-L-sorbitol with nitrogen in the ring. When this molecule is inverted it corresponds exactly to the 1,5-anhydro-Dglucitol, which is the deoxynojirimycin that would cause so much excitement later. This deoxynojirimycin synthesis was published in a review article in the Angewandte Chemie in 1966 [1] and in detail in Chemische Berichte in 1967 [27]. I. Sangster is a native of Scotland and a truly original chemist, who now lives in Jamaica. There he is not only affiliated with the sugar industry but also is the producer on the edge of the rain forest of an alcoholic beverage that captures all the flavors and pleasant scents of the forests of the Blue Mountains of Jamaica. On the label of the bottle of this wonderful beverage there is even a likeness, complete with tropical hat, of the man who first synthesized deoxynojirimycin. Since so much material had become known about the chemistry of saccharides with ring nitrogen, it seemed to M. L. Wolfrom that the time had come for a review article to appear in Advances in Carbohydrate Chemistry. We immediately agreed to write the article but asked if it would be possible to submit it in German. Wolfrom decided that this would not be a problem, since he had a number of German post-doctorate students who could easily translate the article into English. Later I personally had the opportunity to visit Wolfrom in Columbus, Ohio. Without any doubt Wolfrom was one of the dominant personalities in carbohydrate chemistry, who impressed me very much. He loved going on walks in the afternoon with his big dog. It was on one of these walks that I had the honour of accompanying him and I recognized that our article for Advances in Carbohydrate Chemistry had not been translated by a post-doc but by him personally. Wolfrom understood German very well. He found the article fascinating and was very impressed by the new possibilities of NMR spectroscopy. We had used this method in our studies on hindered rotation in saccharides with N-acetylated nitrogen in the ring. To be honest, I have to admit that the style of the English translation was even better than the style of the German original. With publication of the review in 1968 [2], the main features of the chemistry of saccharides with nitrogen in the ring had been fully elucidated so we turned our attention to new projects. The strong inhibitory effect of

4

1 The Early Days of Monosaccharides Containing Nitrogen in the Ring

deoxynojirimycin on glycosidases went unnoticed at that time. The many compounds with nitrogen-containing rings were not screened thoroughly, for at that time no one really expected these carbohydrate compounds to have a decisive biological effect. This did change dramatically many years later. In the years that followed, the scientists of the Research Center of Bayer AG in Wuppertal and especially the physician W. PuIs were on the look-out for new ways to treat diabetes mellitus. The carbohydrates in our diet consist mainly of starch and saccharose. In the small intestine, starch is degraded by an enzyme from the pancreas, α-amylase, to oligosaccharides, mainly maltose. The smaller carbohydrate building blocks, which also include saccharose, are degraded by intestinal enzymes, the α-glucosidases, to glucose and fructose. These monosaccharides are absorbed directly by the wall of the intestines and enter the blood stream. In diabetics, their absorption causes a sudden increase in the blood glucose level, so-called post-prandial hyperglycemia. Transient glucose peaks of this type can cause secondary diseases in the blood vessels, nerves, eyes and kidneys. W. PuIs and his co-workers hoped to slow down carbohydrate degradation by inhibiting the α-glycosidases and thus prevent peaks in the blood sugar levels. Therefore the problem was to find inhibitors for α-glucosidases that were suitable for a preventive therapy of this type. An appropriate test system was developed by W. PuIs and D. Schmidt and used to screen a series of substances available within Bayer AG or commercially for inhibitory properties towards α-glucosidase. Since no suitable substances were found, the search remained unsuccessful. In cooperation with W. Frommer and L. Miiller, fermentation broths of microorganisms were then studied in the hope of finding components that are active as inhibitors for α-glucosidase. This screening of culture filtrates revealed that micro-organisms of the Actinomycetales order mainly of the Actinoplanes genus produce α-amylase and α-glucosidase inhibitors. By improving the fermentation process it was then possible to obtain a preparation containing approximately 20 % of an α-amylase inhibitor. However, the structure of the active substance still remained largely unknown. It was B. Junge who then succeeded in isolating the main component of the active substance from the fermentation batches in pure form. The degradation experiments he conducted showed that a cyclitol unit, an aminosugar and glucose residues were the building blocks. The NMR spectra provided good information on how the building blocks were linked. At this time - it was 1973 - S. Schiitz, who was then the head of the pharmaceutical research laboratorium in Wuppertal, called, commenting cautiously: Carbohydrate chemistry has caught up with us. This was a comment I was also to hear later from other colleagues. He asked me to discuss the proposed structure in Wuppertal. B. Junge had truly done a very fine job. Excellent NMR spectra were available from which the structure of acarbose as we know it today could be unequivocally derived. It consists of an unsaturated cyclitol building block, valienamine with a 4-amino-4,6-dideoxy sugar and two glucose residues [28]. My co-worker F. Heiker then synthesized valienamine, a very difficult task at that time [29]. Because the separation of enantiomers was still not highly developed, we used the pure enantiomeric form of the natural product quebrachitol as

1 The Early Days of Monosaccharides Containing Nitrogen in the Ring

5

the starting product. We obtained this compound from waste water of natural latex production, which had been procured especially for us from a rubber plantation in Malaysia. Valienamine was also a component of the antibiotic validamycin A, which had been studied and synthesized by Seiichiro Ogawa [3O]. Valienamine and validamine belong to a class of compounds, the so-called carba-sugars, in which the oxygen of the pyranose ring is replaced by a carbon. The chemistry of the carba-sugars was studied in detail for many years by S. Ogawa [31]. In the process, nearly all possible isomers were prepared and S. Ogawa also succeeded in the total synthesis of acarbose [31]. Acarbose can bind to the active center of saccharase instead of the oligosaccharides originating from starch but cannot itself be cleaved by the enzyme. This saccharase inhibitor is hardly metabolized and is eliminated rapidly and completely by the kidneys. Meanwhile acarbose has been successfully introduced worldwide as Glucobay for the adjuvant therapy of diabetes mellitus [32]. In 1976, some time after my visit to Wuppertal in connection with acarbose, to my surprise I received another call from B. Junge. He told me that another substance had been isolated from the fermentation broths of a different Bacillus strain that was even more effective than acarbose in the inhibition of α-glucosidases. The active substance was a smaller molecule with a structure that could be elucidated relatively easily. It turned out to be the deoxynojirimycin that I. Sangster had already synthesized for us in 1966, a synthesis I even had trouble remembering, for it had actually been synthesized ten years before its important biological effect was recognized. The first 100 grams of deoxynojirimycin were then synthesized for testing from the Reichstein sorbose derivative according to our method. I. Sangster, by this time in Jamaica, would never have imagined when he synthesized it that his substance would later achieve such importance. This finding naturally stimulated further intensive synthetic studies in this area. Alternative deoxynojirimycin syntheses were developed and hundreds of N-substituted and C-branched derivatives of deoxynojirimycin were synthesized in the Wuppertal laboratories and screened for their biological activity. The best derivatives were nearly two orders of magnitude more effective than acarbose [33]. Because of its good in vivo efficacy, the N-hydroxyethyl derivative of deoxynojirimycin was finally chosen for further development and clinical testing. Today it is approved for treatment of non-insulin-dependent diabetes in Europe and the USA under the name Miglitol. Discovery of the inhibitory effect of deoxynojirimycin led simultaneously to other synthesis activities in a great number of laboratories world-wide. New, original syntheses of deoxynojirimycin were developed, and nearly every possible isomer as well as other derivatives of these were prepared [33]. These studies were also stimulated by the finding that even related hydroxylated bicyclic pyrrolidines like castanospermine and swainsonine have a strong inhibitory effect on glycosidases. P. R. Dorling [34] showed that especially swainsonine, which was first isolated by S. M. Colgate [35] in 1979 from the Australian plant Swainsonia canescens, exhibits a pronounced characteristic strong inhibitory activity towards amannosidases. Swainsonine was subsequently used by R. W. Jeanloz and C. D. Warren [36] in 1988 for an interesting experiment. At that time the extremely com-

6

1 The Early Days of Monosaccharides Containing Nitrogen in the Ring

plex oligosaccharide side chains of glycoproteins of the 'high mannose-type' could only be synthesized with extreme difficulty. Jeanloz and Warren treated sheep with swainsonine. As a result, the degradation of 'high mannose-type' oligosaccharide chains was inhibited and accordingly the level of 'high mannose' oligosaccharides in the urine of sheep increased. These oligosaccharides were then isolated by them from the urine and used for new chemical syntheses. Today, as a result of many synthetic efforts, numerous inhibitors are available that can be used selectively for the widest variety of glycosidases [37]. They are important aids in elucidating the biosynthesis and degradation especially of complex carbohydrate side chains like those present in glycoproteins and glycolipids. Meanwhile the chemical and biochemical findings are so extensive that they are hardly comprehensible. It is therefore all the more admirable that experts in the individual areas have combined to produce a comprehensive review of the findings. As a result, this book provides a representative cross-section of the whole field. One can only hope that the book will become a source of information of general interest to chemists and biochemists alike.

References [1] H. Paulsen, Angew. Chem., 1966, 78, 501-516; Angew. Chem., Int. Ed. Engl., 1966, 5, 495-510. [2] H. Paulsen, K. Todt, Adv. Carbohydr. Chem., 1968, 23, 115-232. [3] J. C. P. Schwarz, K. C. Yule, Proc. Chem. Soc,. 1961, 417. [4] T. J. Adley, L. N. Owen, Proc. Chem. Soc., 1961, 418. [5] D. L. Ingles, R. L. Whistler, J. Org. Chem., 1962, 27, 3896-3898. [6] B. Urbas, R. L. Whistler, J. Org. Chem., 1966, 31, 813-816. [7] H. Paulsen, Angew. Chem., 1962, 74, 585-586; Angew. Chem. Int. Ed. Engl, 1962, 1, 1454-1455; Liebigs Ann. Chem., 1993, 665, 166-187. [8] H. Paulsen, Angew. Chem., 1962, 74, 901-902; Angew. Chem. Int. Ed. Engl, 1962, 1, 597-598; Liebigs Ann. Chem., 1963, 670, 121-127. [9] J. K. N. Jones, W A. Szarek, Can. J. Chem., 1963, 41, 636-640. [10] S. Hanessian, T. H. Haskell, J. Org. Chem., 1963, 28, 2604-2610. [11] H. Paulsen, F. Leupold, Carbohydr. Res., 1966, 3, 47-57. [12] H. Paulsen, F. Leupold, Chem. Ber., 1969, 102, 2804-2821. [13] H. Paulsen, K. Todt, Chem. Ber., 1967, 100, 3385-3396; 3397-3407. [14] H. Paulsen, K. Todt, K. Heyns, Liebigs Ann. Chem., 1964, 679, 168-177. [15] D. L. Ingles, Chem. Ind. (London), 1964, 927-928. [16] H. Paulsen, F. Leupold, K. Todt, Liebigs Ann. Chem., 1966, 692, 200-214. [17] H. Paulsen, F. Leupold, Chem. Ber., 1969, 102, 2822-2834. [18] S. Inouye, T. Tsuruoka, T. Ito, T. Niida, Tetrahedron, 1968, 23, 2125-2144. [19] H. Paulsen, K. Todt, Chem. Ber., 1966, 99, 3450-3460. [20] H. Paulsen, K. Todt, Chem. Ber., 1967, 100, 271-279. [21] H. Paulsen, J. Briming, K. Heyns, Chem. Ber., 1969, 102, 459-468. [22] H. Paulsen, K. Propp, J. Briming, Chem. Ber., 1969, 102, 469-487. [23] H. Paulsen, G. Steinert, Chem. Ber., 1967, 100, 2467-2473; Chem. Ber., 1970, 103, 475-485. [24] W A. Szarek, J. K. N. Jones, Can. J. Chem., 1964, 42, 20-24. [25] S. Hanessian, Carbohydr. Res., 1965, 1, 178-180.

References

1

[26] S. Hanessian, Chem. Ind. (London), 1966, 2126-2127. [27] H. Paulsen, I. Sangster, K. Heyns, Chem. Ber., 1967, 100, 802-815. [28] E. Truscheit, W. Frommer, B. Junge, L. Muller, D. D. Schmidt, Angew. Chem., 1981, 93, 738-755; Angew. Chem., Int. Ed. Engl, 1981, 20, 744-761. [29] H. Paulsen, F. R. Heiker, Angew. Chem., 1980, 92, 930-931; Angew. Chem., Int. Ed. Engl., 1980, 79, 904-905; Liebigs Ann. Chem., 1981, 2180-2203. [30] T. Toyokuni, S. Ogawa, T. Suami, Bull Chem. Soc. Jpn., 1983, 56, 1161-1170. [31] T. Suami, S. Ogawa, Adv. Carbohydr. Chem. Biochem., 1990, 48, 21-90. [32] E. Truscheit, B. Hillebrand, B. Junge, L. Muller, W. PuIs, D. Schmidt, Prog. Clin. Biochem. Med., 1988, 7, 17-99. [33] B. Junge, M. Matzke, J. Stoltefuss, Handbook of Experimental Pharmacology, 1996, 779,411-482. [34] P. R. Dorling, C. R. Huxtable, S. M. Colgate, Biochem. J., 1980, 797, 649-651. [35] S. M. Colgate, P. R. Dorling, C. R. Huxtable, Austr. J. Chem., 1979, 32, 2257-2264. [36] S. Nakabayashi, C. D. Warren, R. W. Jeanloz, Carbohydr. Res., 1988, 774, 279-289. [37] G. Legler, Adv. Carbohydr. Chem. Biochem., 1990, 48, 319-384.

2

Taxonomic Distribution of Iminosugars in Plants and Their Biological Activities MONIQUE S. J. SlMMONDS, GEOFFREY C. KlTE and ELAINE A. PORTER

2.1 Introduction Recently there has been increasing interest in a heterogeneous group of hydrophilic plant alkaloids due to their potentially useful biological activities and their possible ecological and taxonomic significance [1-4]. The compounds concerned are simple hydroxylated derivatives of the monocyclic and bicyclic nitrogen-containing ring systems found in piperidine, pyrrolidine, indolizidine, pyrrolizidine and nortropane alkaloids [5,6]. Although these alkaloids have generally been grouped together, their relationships are largely conceptual, based on the fact that the hydroxyl groups are held in a fixed stereochemistry by the heterocyclic ring system and in a way which resembles the stereochemical positioning of the hydroxy groups in carbohydrates. Members of the group have been given various generic names in attempts to indicate this structural resemblance to sugars: iminosugars [7], polyhydroxy alkaloids [1], azasugars [8], aminosugars [9] or sugar-shaped alkaloids [10]. Deciding what constitutes an iminosugar and what does not is somewhat arbitrary, although attempts to define the group have been made [11]. Despite these problems, individual iminosugars are proving useful in helping to resolve the relationships of the plant species in which they occur (chemotaxonomy) and in having important biomedical and ecological properties (biological interactions).

2.2 2.2.1

Structural Diversity of Iminosugars Piperidines

Aminosugars such as D-glucosamine (2-amino-2-deoxy-D-glucopyranose) and D-galactosamine (2-amino-2-deoxy-D-galactopyranose), in which the hydroxyl Iminosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

2.2 Structural Diversity of Iminosugars group at C-2 of the monosaccharide is replaced by an amino group, are widespread in nature; in particular, the N-acetyl derivatives of 2-aminosugars are ubiquitous constituents of glycoproteins. By contrast, 5-substituted aminosugars with a full complement of hydroxyl groups are considerably rarer. Only three naturally occurring examples are known, these being the glucose, mannose and galactose derivatives which have been isolated from species of the Streptomyces [12-14]. As with sugar aldoses, these 5-substituted aminosugars exist primarily as cyclical structures to produce compounds (with both a and β anomers) that are analogous to the respective pyranoses but in which the ring oxygen of the monosaccharide is replaced by a nitrogen atom. Hence they may be termed iminosugars. The first natural occurring iminosugar to be discovered was the 5-substituted amino-glucose (5-amino-5-deoxy-D-glucopyranose). 'Glucosimine', as it might simply be called, was originally isolated as an antibiotic from Streptomyces roseochromogenes R-468 [15] but, following structural characterization [16], it was given the trivial name nojirimycin after its isolation from Streptomyces nojiriensis (Figure 2-1) [17, 18]. As well as having antimicrobial activity, nojirimycin was found to be a potent inhibitor of a-and /?-glucosidases [12], as might be expected from its structural mimicry of glucose. Indeed, it was the potent inhibition of glycosidases by culture broths of species of Streptomyces that led to the isolation of the two other iminosugars: the α-mannosidase inhibitor 5-amino-5-deoxy-D-mannopyranose (nojirimycin B) from Streptomyces lavandulae SF 425 [13] and the βgalactosidase inhibitor 5-amino-5-deoxy-D-galactopyranose (galactostatin) from Streptomyces lydicus PA-5726 [14]. Soon after the discovery of naturally occurring iminosugars in micro-organisms, alkaloid chemists began to isolate multiple hydroxylated piperidine alkaloids from plants. The first to be isolated was fagomine (2-hydroxy methyl-3,4-dihydroxypiperidine) from Fagopyrum esculentum (Polygonaceae) [19] followed by moranoline (2-hydroxymethyl-3,4,5-trihydroxypiperidine) from a species of Morus (Moraceae) [2O]. Yagi et al. [20] noted that moranoline was identical to the iminosugar nojirimycin except that it lacked the hydroxy group on the anomeric carbon. 1-Deoxynojirimycin had, in fact, already been synthesized [21] and was later also found to be produced by bacteria, Bacillus species [22] and S. lavandulae [23]. The name 1-deoxynojirimycin, usually abbreviated to DNJ, has therefore taken precedence such that when the 2-epimer was isolated from the legumes Lonchocarpus sericeus and L. costaricensis [24], it was later named as the nojirimycin derivative 1-deoxymannojirimycin or DMJ [2, 25]. This nomenclature was adopted to reflect the mannose stereochemistry of the hydroxy groups; other authors have preferred not to extend the inserted 'man' to 'manno' and refer to the compound as 1-deoxymannonojirimycin. More recently, higher homologs of nojirimycin have been discovered in which a hydroxymethyl group replaces the hydroxy group on the anomeric carbon. a-Homonojirimycin (α-HNJ) from Omphalea diandra (Euphorbiaceae) was the first to be isolated [26] and recently the 2-epimer, homomannojirimycin (HMJ, both a- and /?-anomers) was obtained from an aroid belonging to the genus Aglaonema together with a-3,4-diepi-homonojirimycin and a- and β- HNJ [27]. It is perhaps unfortunate that these compounds, which could be considered as either piperidine alkaloids or iminosugars,

10

2 Taxonomic Distribution of Iminosugars in Plants . HO

HO

OH

OH

α-D-glucose

a-D-mannose

OH a-mannojirimycin

α-nojirimycin

(α-norjirimycin B)

HO

HO 1 -deoxynojirimycin (DNJ)

1 -deoxy mannojirimycin (DMJ)

1 ,2-dideoxynojirimycin (fagomine)

HO "7-^^^ X-^NH / / uo-^/^^T~— — ^/ H0 I OH α-homonojirimycin Η0

HO °^7^^-^^>^^/NH /HO / HO^—i^^^~^—^J I OH a H

Figure 2-1. Structures of iminosugar piperidines (nojirimycins) and analogous monosaccharides. The views drawn are rotated 60° from that normally used for pyranoses so as to aid comparison

have adopted a nomenclature based on an antibiotic. However, if one is to pursue this line of nomenclature, a-3,4-dic/?i-homonojirimycin could be termed a-homogulonojirimycin and fagomine might also be considered to be 1,2-dideoxynojirimycin, although neither of these names have been applied. Excluding glycosides, the only other 'nojirimycin' known from natural sources is 3,4-diepi-fagomine (1,2-dideoxygulonojirimycin) [28]. Deoxyderivatives and homologs of galactostatin (galactonojirimycin) have yet to be discovered.

2.2 Structural Diversity of Iminosugars

2.2.2

11

Pyrrolidines

The suspicion that many other glycosidase-inhibiting alkaloidal analogs of sugars might occur naturally, in addition to hydroxylated piperidines, came following the reisolation of a pyrrolidine alkaloid, 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidine (DMDP), from species of the legume genus Lonchocarpus [29, 30] (Figure 2-2). DMDP was originally isolated from another legume, Derris elliptica, and its relative stereochemistry was assigned by NMR [31]. The absolute structure of DMDP was determined from the Lonchocarpus compound by demonstrating that it had an identical optical rotation to 2/?,5/?-dihydroxymethyl-3/?,4/?-dihydroxypyrrolidine produced by enantiospecific synthesis [32]; this is the enantiomer of the structure that had previously been chosen for DMDP [31]. DMDP was considered to be an analog of /?-D-fructofuranose, although it was demonstrated to be a glucosidase inhibitor [25]. The analogy of DMDP with /?-D-fructofuranose (more correctly 2-deoxy-/?-D-fructofuranose) has been cited on numerous occasions [2, 6, 10], but given the occurrence of α-homonojirimycin as the homolog of α-D-glucosimine (nojirimycin), DMDP might also be considered as the homolog of /?-D-arabinosimine (furanose form). The 1-deoxy derivative of o-arabinosimine has been isolated, almost simultaneously, both from the legume Angylocalyx boutiqueanus [33] and the fern Arachniodes standishii [34]. The correct stereochemistry of this compound was subsequently determined [35] and it has since been called D-ABl [2]. The 2-epimer of D-ABl (i.e. 1-deoxy o-ribosimine, D-RBl) is also known from nature, having been isolated from Morus alba [36], as is the 1,2-dideoxy derivative, which was obtained from Castanospermum australe [33] and subsequently became known by the code 'CYB-3' [2].

OH

α-D-arabinose

HN-_

α-D-ribose

·ΟΗ τ^ΟΗ

^ΟΗ

VXT"OH ΗΝ

·^-—^^-^\ΟΗ

D-AB 1

Figure 2-2. Structures of iminosugar pyrrolidines and possibly analagous monosaccharides. Drawn to show structural similarity of DMDP with australine (c/. Figure 2-4).

DMDP

CYB-S

12

2.2.3

2 Taxonomic Distribution of Iminosugars in Plants...

Indolizidines

The first of the polyhydroxylated bicyclic alkaloids discovered was swainsonine, a trihydroxyindolizidine alkaloid, isolated from the legume Swainsona canescens [37] (Figure 2-3). Subsequently seeds of another Australian legume, Castanospermum australe were found to contain another polyhydroxyindolizidine, castanospermine [38]. As with DMDP, in the original report on the structure of castanospermine one of the enantiomers was arbitrarily chosen, and this was subsequently revealed to be the wrong choice by synthesis [39]. The disposition of hydroxy groups in castanospermine is the same as in DNJ and biosynthesis of castanospermine from the corresponding carboxylic acid of DNJ (i.e. a trihydroxypipecolic acid) by condensation with one acetyl unit has been postulated [39]. However, an N-hydroxyethyl derivative of CYB-3 isolated from C. australe, might be an alternative precursor [4O]. Swainsonine has been shown to be synthesized from pipecolic acid in Rhizoctonia leguminicola [41] even though the C-I carbon in swainsonine has the opposite stereochemistry relative to the bridgehead carbon compared to the hydroxymethyl group in pipecolic acid from which it is derived; an epimerization at the bridgehead carbon occurs during biosynthesis. Two epimeric forms of castanospermine, the 6-epimer [42] and the 6,7-diepimer [40] occur with castanospermine in C. australe, together with the 7-deoxy derivative of 6-epi-castanospermine [43]. A deoxy derivative of swainsonine, 2-epi-lentiginosine, and its 2-epimer, lentiginosine, have also been isolated from Astragalus lentiginosus, a relative of Swainsona [44].

castanospermine

swainsonine

Figure 2-3. Structures of iminosugar indolizidines (castanospermines and

OH 6-ep/-castanospermine

6,7-diep/-castanospermine

2.2 Structural Diversity of Iminosugars

2.2.4

13

Pyrrolizidines

The first two polyhydroxypyrrolizidines to be discovered, australine and alexine, were isolated at about the same time from C. australe [45] and Alexa leiopetala, a member of a closely related genus [46] (Figure 2-4). Co-occurring with australine in C. australe are its 1- [47], 3- [48] and 7-epimers [49]. Since australine and alexine are epimeric to one another at the bridgehead position, it has been thought likely that a similar series of alexine epimers might also exist along with alexine in Alexa species [50], but these have yet to be reported. 1-epi-Australine has been detected in A. leiopetala [49] and the 3 -carboxylic acid derivative of australine, 7a-e/?i-alexaflorine (perhaps better named australiflorine), has also been isolated from A. grandiflora [51]. The most recent polyhydroxypyrrolizidine to be isolated, casuarine from, Casuarina equisetifolia. (Casuarinaceae), is also the most oxygenated, being the 6-hydroxy derivative of australine and consequently bearing five hydroxy groups [52]. Although alexine, casuarine and the australines can be considered as members of the large group of pyrrolizidine alkaloids, they differ from all other known members by having a carbon substituent at the C-3 position, rather than the more usual C-I carbon substituent on the necine base [53]. In addition, the structural similarity between australine and DMDP has been noted [45], australine being DMDP with one of the hydroxymethyl groups cyclized to the nitrogen via an acetyl group. It is not known whether this represents a possible biosynthetic route; Molyneux et al. [40] suggest that the N-hydro-

australine

alexine

1 -ep/-australine

7-ep/-australine

3-ep/-australine

casuarine

Figure 2-4. Structures of iminosugar pyrrolozidines (australines, alexine and

14

2 Taxonomic Distribution of Iminosugars in Plants...

xyethyl derivative of CYB-3, proposed as a precursor for castanospermine, could also be a precursor for australines via direct cyclization.

2.2.5

TVortropanes

Members of the most recently recognized group of iminosugars, the calystegines, are tri-, tetra- and pentahydroxy nortropane alkaloids isolated from roots and root exudates of Calystegia sepium (Convolvulaceae) during research to identify substances produced by plant roots that had selective effects on rhizosphere organisms [54] (Figure 2-5). The nomenclature used for calystegines is derived from their chromatographic behavior during the original isolation from C. sepium the tri- and tetrahydroxy calystegines were initially separated as two spots, A and B, after paper electrophoresis, and each of these was then resolved into their isomeric components by liquid chromatography, to give calystegines AI, A2, A3, A4, BI and BI [55]. Of these, only the structures of calystegines A3, BI and 62 have been elucidated [55]. Other calystegines have since been isolated, characterized and named in accordance with the existing system: calystegins AS and B3 from Physalis alkekengi (Solanaceae) [56] and calstegine B4 from a Scopolia japonica (Solanaceae) [57], and the pentahydroxy nortropanes, calystegine Ci from Morbus alba [36] and calystegine C 2 from Duboisa leichhardtii (Solanaceae) [58]. Calystegines have been differentiated from other tropane alkaloids by the lack of an N-methyl

calystegine A3

calystegine B2

calystegine C1

Figure 2-5.

calystegine A5

calystegine B3

calystegine C2

calystegine B4

calystegine B1

Structures of iminosugar rcortropanes (calystegines).

2.3 Chemotaxonomy

15

group and the presence of a hydroxy group at the bicyclic ring bridgehead. Like the hydroxy group on the anomeric carbon of nojirimycin, the bridgehead hydroxy group of a calystegine derives from cyclization of a tautomeric form, which for the calystegines is an appropriately hydroxylated 4-aminocycloheptanone [55, 59]. The corresponding alcohol of the monocyclic tautomer of calystegine AS (Le. l-amino-2,3,5-trihydroxycycloheptane) has been isolated from Physalis alkekengi and assumed to be a precursor or degradative product [6O]. Recently, a calystegine, named calystegine NI, has been isolated from Hyoscyamus niger (Solanaceae) in which the bridgehead hydroxy group is replaced by an amino group [61], although it is unclear whether this is an artefact created during isolation [59].

2.2.6

Glycosides

Glycosides are now known for four of the five classes of alkaloids, the exception being the indolizidines, but by far the most numerous are glycosides of the piperidines. There was a relatively early isolation of a /?-D-glucoside of fagomine from the legume Xanthocercis zambesiaca [62]. However, other glycosides remained elusive until recently when known iminosugar-producing plants were re-examined using improved isolation techniques. This resulted in seven glucosides and two galactosides of DNJ, and one glucoside of D-ABl being isolated from species of Morus [36, 63], a glucoside and galactoside of α-HNJ being characterized from a species of Aglaonema [27], and another glucoside of fagomine being obtained from X. zambesiaca [28]. There is only one example known from each of the polyhydroxylated pyrrolizidine and nortropane classes, these being glucosides of casuarine [64] and calystegine BI [65]. Finally, a glucoside of DMDP has been isolated, but not characterized, from Omphalea diandra [66].

2.3

Chemotaxonomy

Iminosugar-producing plant species have been found within several plant families: Moraceae, Euphorbiaceae, Leguminosae, Campanulaceae, Polygonaceae, Myrtaceae, Araceae, Hyacinthaceae, Casuarinaceae, Convolvulaceae, and Solanaceae [6]. This apparently wide and disjunct distribution may simply be a reflection of their heterogenous nature or that very few scientists have studied their distribution among related plant families. The chemotaxonomic surveys that have been performed among genera within a family indicate that the distribution of individual iminosugars is often supported by other taxonomic evidence, in particular DNA sequence analysis data. This suggests that they will provide good characters for modern methods of phylogenetic analysis, such as cladistic techniques. The most comprehensive family survey to be undertaken is that in Araceae. A screen of the family was undertaken following the discovery, by chance, of DMDP in species of Aglaonema [67]. This survey revealed the presence of

16

2 Taxonomic Distribution of Iminosugars in Plants...

DMDP, α-HNJ and epimers of HNJ in species of Aglaonema and Aglaodorum of the tribe Aglaonemateae, and Nephthytis, Anchomanes and Pseudohydrosme of the tribe Nephthytideae; Nephthytis species also contained DMJ [68]. A close relationship between Nephthytideae and Aglaonemateae had not been suggested by previous taxonomic treatments of the family based on morphologic characters, but a cladistic analysis of chloroplast DNA restriction site data did show unexpectedly strong support for these tribes being sister groups [69]. The taxonomy of these two tribes is now being re-examined by aroid systematists in view of the congruence of these two data sets. In Euphorbiaceae there is a similar occurrence of iminosugars in taxa that may be more closely related than was thought from morphological data. Here, DMDP was discovered in the neotropical liana Omphalea diandra as a result of a chemoecologic investigation into its relationship with the moth Urania fulgens (Uraniidae) which utilizes Omphalea as its larval foodplant [7O]. Subsequently the plant was shown to synthesize DMJ, DNJ and α-HNJ as well as DMDP and these compounds were also in species of Endospermum that were the larval foodplants of other members of Uraniidae [71]. Although a member of Euphorbiaceae, Endospermum is placed in the subfamily Crotonoideae while Omphalea is considered to be a member of Acalyphoideae [72]. The iminosugar data suggest that Omphalea is misplaced in Acalyphoideae, but whether molecular information, such as rbcL DNA sequence analysis, will support this hypothesis still needs to be established. A chemotaxonomic study of iminosugars in legumes showed an apparently unique occurrence of castanospermine in the monotypic Castanospermum australe from Australia and in geographically distant Alexa species from South America. Castanospermum was thought to be part of an 'Angylocalyx alliance' of the tribe Sophoreae (Leguminosae) and Alexa part of a 'Dussia alliance' of the same tribe [73], even though a highly anomalous pollen wall stucture occurs in both species [74]. Following the isolation of castanospermine and interest in its activity against HIV, this similarity in pollen morphology led to an examination of herbarium specimens of Alexa species and the discovery of their chemical similarity [75]. Subsequently the genera were shown to share 1-epi-australine and 7-epi-australine. It is probable that Castanospermum and Alexa are very closely related and the present disjunct distribution is the result of the break-up of Gondwana. The fact that, to date, no iminosugars have been reported from species within the Dussia alliance suggests that Alexa is misplaced in the Dussia alliance, whereas other members of the Anglyocalyx alliance contain iminosugars; for example fagomine is present in Castanospermum [45], Anglyocalyx [76] and Xanthoceris [28]. The distribution of DMDP in the large and probably unnatural legume genera Derris and Lonchocarpus, from which this iminosugar was first isolated, also provides data for future taxonomic treatments [3O]. The presence of DMDP correlates well with subgeneric sections proposed within the genera and indicates some misplaced species. It also suggests that two subgeneric groups, Derris section Paraderris and Lonchocarpus subgenus Phacelanthus, may be more closely related to each other than to other subgenera in Derris and Lonchocarpus as currently circumscribed. This relationship is supported by the distribution of rotenoids but

2.4 Biological Interactions

17

the significance of this phytochemical data awaits a modern systematic treatment of these taxonomically confused woody members of the subfamily Papillionoideae. Further evidence of the use of iminosugars in systematics comes from ongoing research in Myrtaceae and Hyacinthaceae. In Myrtaceae, the only report of imino sugars has been casuarine and its glucoside in Syzygium (=Eugenia) jambolana [64]. However, a more extensive survey is revealing that nearly all genera examined from one of the two subfamilies, Myrtoideae, contain iminosugars, whereas they are only present in two genera of the other subfamily, Leptospermoideae. The distribution of these, as yet uncharacterized, iminosugars does not correspond with the currently recognized subfamilial boundaries and could have taxonomic significance. Within Myrtoideae, the iminosugar data suggest an affinity between genera in a 'Eugenia alliance' with those in an 'Acmena alliance'; these alliances have not previously been considered to be closely related [77]. In Hyacinthaceae, iminosugars have been reported in Hyacinthus orientalis and members of several other genera [78]. As with Myrtaceae, the structures of the compounds concerned are still being elucidated. Hyacinthus orientalis appears to be unique in Hyacinthaceae in synthesizing a-HNJ [78]. In the past, many species have been placed in Hyacinthus but most of these have now been transferred to other genera. Only two species, other than H. orientalis, now remain, but the absence of α-HNJ in these perhaps indicates that H. orientalis may eventually prove to be monotypic [78]. Data from these systematic studies support the initial view that iminosugars occur in related taxa within otherwise unrelated groups. This gives the impression that there are 'hotspots' of iminosugar evolution. These hotspots can vary in taxonomic size from a subgenus (such as DMDP in subgenera of Lonchocarpus), through tribes (imino sugars in the aroid tribes Aglaonemateae and Nephthytideae) to most of a family (their occurrence in Hyacinthaceae). The existence of taxonomic hotspots may reflect the sporadic evolution of these compounds in plants or to problems in their isolation. It should be remembered that the group escaped detection in plants until the late 1970s and the occurrence of monoglycosides has only been realized in recent years. No diglycosides have yet been reported and one wonders whether this is simply because of the difficulty in detecting them by GC-MS, the standard method of analysis. As LC-MS becomes more available, we might find that oligoglycosides of iminosugars are more widespread than the aglycones.

2.4

Biological Interactions

To date, many reviews on the biological activity of iminosugars have concentrated on their biomedical properties which are usually ascribed to their ability to inhibit glycosidases [1,6]. This review covers the biomedical activities and less welldocumented ecological and pesticidal activities of these compounds.

[79, 80]

[80, 81]

[43] [45]

[55] [55] [55] [55]

[76]

[55] [55] [76] [76] [76] [55]

[42]

[42, 80]

[25]

[23, 81]

/9-gluc

[55] [55, 76] [55, 76] [76] [76] [55]

α-galac

[55]

[55] [55] [55]

[42] [43]

/7-galac

[37]

[18]

[13]

[84]

[43] [42]

[85]

[82]

α-mann /9-mann gluc-I

[42, 85]

[82]

glue -II

[5]

mann-II

Key: α-gluc = α-glucosidase; β-gluc = /?-glucosidase; α-galac = α-galactosidase; /?-galac = /?-galactosidase; α-mann = a-mannosidase; /?-mann = /?-mannosidase; gluc-I = glucosidase I; gluc-II = glucosidase II; mann-II = mannosidase-II. Trehalase and invertase have been included as examples of two a-glucosidases.

[56]

[6] [55]

[6]

[25]

[79, 80] [79, 80] [79, 80]

[26] [83] [25] [83] [79, 80]

[23]

[12,81] [13] [23, 79, 80] [23]

[12]

invertas

nojirimycin norjirimycin B DNJ DMJ HNJ AB-I DMDP CYB-3 castanospermine swainsonine o-ep/castanospermine 6J-diepiaustraline alexine casuarine calystegine A3 calystegine Bl calystegine B 2 calystegine B 3 calystegine B4 calystegine Cl

trehalase

a-gluc

imino sugar

Table 2-1. Range of glycosidase inhibitory activities associated with iminosugars. Numbers within [] indicate references containing information about the activity of each imino sugars.

2.4 Biological Interactions

2.4.1

19

Inhibition of glycosidases

Glycosidases are a class of enzyme which are extremely widespread in organisms [79, 8O]. These enzymes catalyze the hydrolysis of glycosidic bonds in carbohydrates and glycoconjugates. The net effect is the release of low-molecular weight monosaccharides and oligosaccharides [5]. Because glycosidases are important in many biochemical systems it is not surprising that compounds that inhibit them exhibit biological activity. Thus glycosidase inhibitors can, and do, cause detrimental effects but they can also be used as valuable tools when investigating the physiological role of glycosidase enzymes 1O]. The enzymatic activity of iminosugars has been attributed to their structural resemblance to simple sugars. The extent and specificity of the inhibition was thought to be dependent on the position, number and stereochemistry of the hydroxyl groups on the molecule. However, experimental data have shown that the chirality of the hydroxy groups on the iminosugars was not sufficient to predict their ability to inhibit enzymes [5]. Enzyme inhibition can be highly variable and influenced by the source of the enzyme as well as experimental conditions, such as pH [8O]. A summary of the types of glucosidase enzymes inhibited by some of the iminosugars is presented in Table 2-1.

2.4.2

Glycoprotein processing

A few iminosugars inhibit glycosidases involved in glycoprotein processing (Table 2-1) Glycoproteins belong to a class of glycoconjugates which also includes glycolipids and proteoglycans [87]. The most common class of glycoproteins are the Nlinked glycans which are formed via a common glycolipid precursor, Glcs-MangGlcNAc2-pyrophosphoryl-dolichol. This is transferred to asparagine residues of proteins. Subsequent trimming and processing of this precursor leads to a variety of N-linked glycans [88]. The processing is performed by glycosidase enzymes, which catalyze a series of specific reactions, ultimately yielding glycoproteins targeted to perform particular biological functions. The functioning of these glycoproteins will depend on the correct sequencing of their monosaccharides during processing. Enzymes involved in this processing include glucosidase I, which removes outer a-1,2-linked glucose units, and glucosidase II which removes a-1,3-linked glucose units. Further modification of proteins within a cell resides in the endoplasmic reticulum and Golgi apparatus and involves removal of the a-1,2-linked mannose units by specific mannosidases IA/B and removal of the a-1,3- and a-1,6linked mannose residues by mannosidase II [5]. These enzymes differ in their specificity; for example, glycosidases enzymes from the endoplasmic reticulum and Golgi apparatus are very distinct from those found in lysosomes. Malfunctioning of these enzymes can result in disease as well as the breakdown of the digestion of carbohydrates [5, 8, 9]. Iminosugars can selectively inhibit these enzymes and are currently being used to study the role of glycosidases in many cellular interactions including processing of virally encoded glycoproteins and metastasis.

20

2 Taxonomic Distribution of Iminosugars in Plants...

2.4.3

Anti-viral activity

A number of iminosugars have anti-viral activities, in particular, against Human Immunodeficiency Virus (HIV), the agent responsible for Aquired Immune Deficiency Syndrome (AIDS) [90-92]. HIV has two glycosylated envelope proteins, gp!20 and gp41. Glycoprotein gp!20 binds to the CD4 antigen of T4-lymphocytes and the transmembrane glycoprotein gp41 anchors the envelope to the viral membrane [93]. These glycoproteins and host CD4 surface receptors play an important part in viral absorption, penetration, syncytium formation and spread of the virus to adjacent cells. When the host cell and membranes of viral infected cell fuse, viral RNA is taken up into the host cell. Compounds that can disrupt the viral envelope and prevent viral-cell contact have a potential role against HIV infection. A lack of syncytia formation often indicates that the virus has been unable to bind to the CD4 receptor of the cell and entry to the cell has been inhibited [94]. Sunkara et al. [94] used Moloney murine leukemia (MoL) virus as a model for testing the activity of iminosugars against HIV. They showed that castanospermine and DNJ were active against the virus (LC5ol.2-2.5 μg/ml), whereas DMJ and swainsonine were not active at 100 μg/ml. Castanospermine is known to inhibit glucosidase I and DNJ inhibits glucosidase I and II, whereas DMJ and swainsonine inhibit mannosidases. This suggests that the removal of outer glucose by glucosidases I and II from asparagine-linked oligosaccharides could be critical for the viability of MoL virus. Human cytomegalovirus (CMV), a virus related to HIV, encodes a number of glycoproteins that are important for viral infectivity [95-97]. When the virus was placed in human embryo fibroblast cell cultures in the presence of castanospermine, deoxynojirimycin or DMDP it was observed that the growth of the infectious virus was blocked [98]. Analysis by SDS-PAGE showed that two particular glycoproteins, gp!30 and gp52, were important for the infectivity of CMV [98]. Gruters et al. [99] reported that when castanospermine was present in in vitro cultures of T-cells (C8166 cells) with HIV, the formation of syncytia did not occur at concentrations as low as 189 μg/ml. Further work showed that a derivative of castanospermine, 6-O-butanoyl-castanospermine (MDL-28,574) inhibited replication of HIV [92]. MDL-28,574 could also block the growth of Herpes simplex virus (HSV) in a mouse model, with an ICso of 100 μΜ [100], whereas the parent molecule castanospermine exhibited weak inhibition at 4 mM [101]. The target enzyme of the synthetic compound was identified as α-glucosidase and the effect was a modification of the viral glycoprotein coat [100, 102]. DNJ and castanospermine were shown to decrease the growth of Sindbis virus by inhibiting the removal of glucose from oligosaccharides which resulted in the disruption of protein folding [94, 103]. Although DNJ showed anti-viral activity, it is cytotoxic [93]. A number of alkyl derivatives of DNJ were tested against HIV and N-butyl-deoxynojirimycin was found to be the most active [91, 104]. NButyl-deoxynojirimycin was shown to be a potent α-glucosidase inhibitor and it is currently undergoing clinical trials as part of a combination therapy with zidovu-

2.4 Biological Interactions

21

dine, AZT (a recognized anti-HIVagent) [105]. Overall, these results indicate that the glycosidase inhibiting activity of these compounds could explain their antiviral activity. However, the results are not conclusive and it appears that, as yet, the mechanisms by which these compounds inhibit HIV is incomplete. For example, HNJ is a potent glucosidase inhibitor, showing more activity against glucosidase II than glucosidase I [82] and more activity against glucosidase I than castanospermine, yet, HNJ is ineffective against HIV even at high concentrations (500 μg/ml) [82]. Thus the inhibition of glucosidase I might not explain the anti-viral activity of compounds such as castanospermine and DNJ.

2.4.4

Lysosomal diseases

Livestock in Australia that consume species ofSwainsona suffer from a neurophysiological disorder called 'peastruck' [106]. In the western parts of America livestock can suffer 'locoism' by consuming locoweeds, which include species of Astragalus and Oxytropis [107]. These poisoned animals display neurological signs which resembled mannosidosis, the rare genetically determined lysosomal disease [108]. Dorling et al. [109] showed that the lymph nodes of sheep displaying locoism contained high levels of mannose-rich oligosaccharides. The imino sugar swainsonine was isolated from Swainsona canescens [37] and also Astragalus lentiginosus [110] and Oxytropis spp. [111]. Biochemical assays showed it inhibited a-mannosidase [110, 111] and was most likely responsible for locoweed poisoning, although the concentration of swainsonine in plants can be very low [111, 112]. Oxytropis ochrocephala and O. kansuensis contain 0.012 % and 0.021 % dry weight of swainsonine, respectively [111]. As swainsonine is both water and lipid soluble it is able to cross plasma membranes and accumulate within the lysosomes. Therefore if animals were feeding on plants containing swainsonine for a short period of time they could consume enough to cause neurophysiological disorders. More recently, two other iminosugars that inhibit glycosidases, lentiginosine and 2-e/?/-entiginosine, have been isolated from species of Astragalus and could contribute to the toxicity of the plants [44]. Swainsonine exhibits a degree of specificity in its activity since it inhibits lysosomal α-D-mannosidase and Golgi mannosidase II but has little or no effect on Golgi mannosidases IA or IB [113]. The affinity of swainsonine for a-o-mannosidase could be attributed to its close resemblance to the mannosyl cation, an intermediate in the enzymatic hydrolysis of mannosidases. However, we do not yet know what determines the specificity of swainsonine inhibition among different mannosidases. [114]. It is possible that the affinity of swainsonine is modulated by pH. The pH optimum for the different mannosidases vary; the lysosomal mannosidases have an acidic pH (pH 4), whereas the cytosolic mannosidases have a more neutral pH (pH 6.5).

22

2 Taxonomic Distribution of Iminosugars in Plants...

2.4.5

Anti-cancer properties

There is increasing evidence that oligosaccarides on the surface of tumor cells play an important role in malignant phenotype and tumor growth. Iminosugars have been tested for their ability to inhibit glycosylation or processing of asparginelinked oligosaccharides that could inhibit tumor growth and metastasis. To date, swainsonine has attracted the most attention, as it inhibits tumor growth and stimulates the immune response [105, 115]. It also enhances bone marrow cellularity, stimulates lymphocyte proliferation and caused a decrease in metastatic foci in the lungs [105,116]. These activities could occur if swainsonine interacted directly or indirectly with growth factors that have carbohydrate-binding properties. The early studies indicated that swainsonine activated natural anti-tumor immunity and enhanced the production of T-cells [117]. Swainsonine was found to enhance the activities of the mouse immune system in vitro 111]. Kino et al. [118] monitored the antibody response of sheep red blood cells (SRBC) in immunodeficient mice when treated with an immunosuppressive factor such as sarcoma 180 tumors. They showed the addition of swainsonine (3.7-100 mg/kg) restored the ability of the mice to produce antibody against SRBC. These results suggested that swainsonine could be used as a treatment for immunocompromised hosts [118]. Swainsonine is currently being evaluated in human patients as a chemotherapeutic against cancer. However, the results have so far been varied, with patients suffering adverse side effects [116]. These patients showed no signs of the immune-stimulatory activity observed in the model systems. The biomedical interest in these imino sugar have progressed from laboratory curiosities; however, variability in their activity have hampered their development in clinical applications. It is hoped that further research into their mode of action and the factors that influence their specificity will enable researches to optimize the potential biomedical uses of these compounds.

2.4.6

Effects on insects

Compounds in plants that inhibit carbohydrate metabolism in herbivores could be part of a very effective defense strategy. For example, Lonchocarpus seeds are resistant to attack by the seed predatory beetle Callosobruchus maculatus [119]. The imino sugar DMDP was isolated from these seeds and shown to be toxic to C. maculatus 119]. This toxicity was attributed to the glycosidase inhibitory activity of DMDP. However, some iminosugar-producing plants are utilized as a food source by specialist insects. This raises the question as to whether these specialist insects have adapted in some way to the imino sugars present in the plants. Larvae of uraniine moths feed only on certain species of Euphorbiaceae that contain imino sugars; for example, larvae of the colorful day-flying moth Urania fulgens are restricted to a few species of Omphalea such as O. diandra which contains DMDP, HNJ, DMJ, DNJ and glycosides of DMDP and HNJ [26, 71, 12O]. The lar-

2.4 Biological Interactions

23

vae selectively accumulate DMDP and HNJ and retain these to adulthood [71]. It has been hypothesized that the presence of imino sugars in U. fulgens protects the larvae and adult moths from predation [71]. Recent studies have shown HNJ to be a potent inhibitor of avian glucosides [120] but whether, this influences the behavior of birds that prey on U. fulgens is not known. The possible role of imino sugars in providing protection against bird predation was investigated by comparing day-flying with night-flying uraniid moths in Australia [71]. Larvae of the brown night-flying Lyssa macleayi fed on another euphorb, Endospermum medullosum, the foliage, which contained much lower levels of iminosugars than Omphalea queenslandiae, the larval food plant of the colorful day-flying Alcides metaurus. However, the hypothesis that L. macleayi could utilize E. medullosum as it did not require chemical protection against bird predation, possibly conferred by accumulating iminosugars, was not supported by the finding that adults of L. macleayi accumulated DMDP and HNJ to levels that were not greatly different to the levels of iminosugars found in A. metaurus [171]. Kite et al. [66] have investigated whether larvae of uraniid moths are unusual in their ability to accumulate iminosugars or whether this phenomenon occurs in other insects that have a less restricted diet but which also feed naturally on plants containing iminosugars. They found that iminosugars were absent or present only in trace amounts in seven generalist herbivores that had been observed to be feeding on O. diandra growing in Panama. These generalist herbivores were larvae of the lycaenid butterfly Panthiades ballus, larvae of the riodinid butterflies Theope virgilius and Nymula mycone, larvae from two limacodid moths (Sibine sp. and Phobetron sp.), adults of the bee Trigona fusipennis and adults of the leaf beetle Rhabdopteris fulvipes. Iminosugars were also not detectable or present at only trace levels in adults of the lycaenid and riodinids that were reared from larvae fed on a diet of O. diandra. The gut glycosidases of the generalist insects were less resistant to inhibition by imino sugars than the glycosidases of the specialist larvae of Urania fulgens [66]. DMDP showed the broadest spectrum of inhibition against the sucrase, maltase and trehalase activities in the generalist insects but the most potent inhibition was exhibited by a glucoside of HNJ against trehalase activity. Digestive trehalase activity in U. fulgens was also not resistant to the HNJ glucoside [66]. These studies indicate that the specialist U. fulgens larvae show some physiological adaptation to surviving on, a plant containing iminosugars. Whether these compounds also play a role in the host selection behavior of uraniid moths is not known. DMDP has been found to occur on the surface of O. diandra leaves and this would indicate that it could be detected by foraging insects and might be used by a specialist insect as a 'sign' stimulus to initiate feeding or oviposition. Iminosugars have also been shown to affect detrimentally the feeding behavior and development of many economically important pest insects. For example, Locusta migratoria, Schistocera gregaria, Spodoptera littoralis, S. exempta, Heliothis virescens [121] Acyrthosiphon psium [122], Callosobruchus maculatus, Tribolium confusum [119], Spodoptera littoralis, S.frugiperda, HeIicoverpa armigera, Heliothis virescens [123] and Myzus persicae [122, 124] can be killed or deterred from feeding by iminosugars.

24

2 Taxonomic Distribution of Iminosugars in Plants...

When the glycosidase inhibitory activity of a selection of iminosugars were tested against enzymes from a range of insect pests they differed in their specificity and level of activity [79, 80, 83]. Such differences could reflect the influence of pH on the iminosugars or it could indicate variations in enzyme binding sites among insects. The majority of these enzyme assays were undertaken on intact insects or removed guts but not on purified enzymes [8O]. Thus the crude nature of the substrate could contribute to the observed variation in results [80, 83]. Despite these limitations, DMDP was able to inhibit hydrolysis of more of the sugars than most of the other imino sugars [80, 83]. The level of activity varied among insect species but DMDP was usually more active against turanose and lactose hydrolysis than against trehalose or sucrose hydrolysis. The only sugar hydrolysis it did not inhibit was that of α-mannopyranoside. D-AB1 was also able to inhibit many glycosidases but was less active than DMDP. D-ABI was more active against insect enzymes than L-ABI the opposite occurred with mammalian enzymes. Castanospermine was also more active against mammalian enzymes than those from insects [83]. DNJ inhibited the hydrolysis of turanose and trehalose more than that of sucrose but this activity varied greatly among species of insects. HNJ had an activity profile similar to DNJ; it was more active against the hydrolysis of turanose than that of sucrose or trahalose. Alexine and australine (cp/alexine) did not inhibit many glycosidases, but showed some inhibition of lactose, trehalose and cellubiose hydrolysis. Overall, the levels at which the imino sugars inhibited α-glucosidases differed between insects and vertebrates: Insect enzymes were more sensitive to DMDP, D-ABI and less sensitive to castanospermine, DNJ, L-ABl and HNJ [8O, 83]. The development or growth inhibitory effects of the iminosugars could be attributed to their enzyme inhibitory activity [119, 121]. However, the mechanisms underlying the anti-feedant activity of the compounds are less clear. Electrophysiologic experiments have shown that neurones in sensilla on the mouthparts of caterpillars, such as Spodoptera, are responsive to iminosugars [123]. When these neurones are stimulated with solutions containing an iminosugar dissolved in an electrolyte, a positive dose-response in neurone activity occurs. The magnitude of this response, that is the number of impulses per second, usually correlates in a positive dose-dependent response with the antifeedant activity of the iminosugar. However, when the sensilla are stimulated with a solution that contains the iminosugars in combination with an electrolyte and sugar (sucrose, fructose or glucose), then a negative dose-response often occurs [123]. For example, if the solution contains fructose and then DMDP is added at increasing concentrations, the magnitude of the neural response decreases as the concentration of DMDP increases. This type of decrease in neural activity is termed 'peripheral interaction' [125, 126]. It is possible that the similarities in molecular structure of DMDP to fructose enables it to block temporarily the receptor sites responsive to fructose. In other experiments castanospermine was shown to interact with the responses of neurones to stimulation with glucose and sucrose, whereas swainsonine lowered the neural response to glucose [123]. The variation in the level and duration of the inhibition indicates some specificity in the interactions between the iminosugar and the sugar. These changes in neural activity were associated with changes in larval

2.4 Biological Interactions

25

feeding behavior. However, the effects on feeding behavior differed among the compounds. For example, when S. littoralis was exposed to glass-fiber discs treated with DMDP the duration of the first meal was similar to that of larvae exposed to control discs but the duration of subsequent meals decreased, although the interval between meals did not differ [123]. When larvae were exposed to discs treated with castanospermine the duration of the meals were shorter than occurred on the control discs. With swainsonine the duration of the first meal on the treated discs was significantly shorter than that on the control discs and the duration of time between meals was significantly longer. In electrophysiological experiments the overall neural response to swainsonine and castanospermine was greater than that to DMDP, although DMDP caused more of an interaction than either castanospermine or swainsonine [123]. This suggests that the mechanisms underlying the anti-feedant response of S. littoralis larvae to these iminosugars might differ. Whether the differences in the anti-feedant responses of other species of Lepidoptera to the iminosugars reflect variations in the neural responsiveness to the compounds is not known. S. frugiperda was deterred from feeding on discs treated with DMDP but not those treated with castanospermine. In contrast, Heliothis virescens and Helicoverpa armigera were both deterred by DMDP and castanospermine [123]. These larvae would not normally feed on plants containing these compounds. These results suggest that many insects that do not normally feed on plants that contain iminosugars are behaviorally responsive to the compounds, although their responses vary. In establishing the ecological importance of these compounds in plant-insect interactions it would be interesting to know if larvae of species of Urania that feed on plants such as Omphalea that contain some of these imino sugars compounds are behaviorally responsive to the compounds. If they are responsive, do the compounds stimulate receptors on deterrent or phagostimulant neurons?

2.4.7

Nematicidal activity

Plant parasitic nematodes cause widespread crop damage, in the region of 20 % of the world's total crop production, and are responsible for serious financial loss. DMDP was found to have a range of activities against several species of parasitic plant nematodes [127]. In vitro assays with potato cyst nematodes Globodera spp., showed DMDP inhibited the mobility of Globodera rostochiensis over a 72-hour period, but was less effective than the commercial nematicide, Oxamyl. When cysts of Globodera pallida were treated with DMDP, the proportion of live juveniles that emerged was reduced by 31 %. This reduction could be due to the enzyme inhibitory activity of DMDP: Osmotic changes involving trehalose occur when juveniles hatch from their eggs and DMDP is know to inhibit trehalase. Further experiments with potato cultivars that differ in their susceptibility to attack by nematodes showed that DMDP could increase the cultivars, resistance to cyst nematodes. Experiments with Xiphinema diversicaudatum showed that DMDP could decrease root galling but whether this is because the nematodes were deterred from feeding by DMDP is not known [128].

26

2 Taxonomic Distribution of Iminosugars in Plants...

Birch et al. [127] showed that DMDP has the potential to be used as a foliar spray, soil drench or seed coating. To date, the mechanisms underlying the nematicidal activity are speculative. DMDP may be acting directly on the nematodes by inhibiting enzymes or indirectly by triggering or enhancing the defense system in the roots of plants [127]. Further work is currently being undertaken in Costa Rica to establish if DMDP could be used as part of an integrated pest management strategy for the control of plant nematodes.

2.4.8

Plant growth regulatory activity

Castanospermine has been shown to be a potent plant growth regulator, inhibiting root elongation in dicotyledons by 50 % at a concentration of 300 ppb. In monocotyledons the effective dose (ED50) was 200 ppm. Thus castanospermine is 103 times less effective on monocot roots than on dicot roots. Swainsonine did not inhibit elongation in either root type at any concentration used [128]. It is not known if these compounds occur naturally in root exudates of iminosugar producing plants; if they do, these results indicate that some iminosugars might have allelopathic properties.

2.5

Summary

As a group of compounds the iminosugars have a diverse range of uses: as taxonomic characters, as pharmaceutical tools, as agrochemicals, and as stimuli in insect-plant interactions. With advances in isolation techniques and screening it is possible that more of these compounds will be isolated and developed as drugs. However, it is also important to study their functional role in plants and how they manipulate plant-herbivore interactions. Investigating such interactions could help us understand the modes of action of these interesting biologically active molecules.

Acknowledgements Thanks to R. Grayer for comments on the manuscript and to the many Herbarium staff at Kew who have provided taxonomic support. We especially thank previous Kew staff including L. Fellows, R. Nash, A. Watson and S. Evans for their help in isolating some of the iminosugars used in many of the studies reported in this review.

References

27

References [1] L. E. Fellows, Pestic. ScL, 1986, 17, 602-606. [2] L. E. Fellows, S. V. Evans, R. J. Nash, E. A. Bell, ACS Symp. Ser. 296, 1986, 72-78. [3] L. E. Fellows, G. C. Kite, R. J. Nash, M. S. J. Simmonds, A. M. Scofield in Plant Nitrogen Metabolism (Eds: J. E. Poulton, J. T. Romeo, E. E. Conn), 1989, Plenum Press, New York, p. 395-427. [4] M. S. J. Simmonds, L. E. Fellows, W. M. Blaney in New Crops for Food and Industry (Eds.: G. Wickens, N. Haq, P. Day), 1989, Groom Helm, England, p. 365-377. [5] A. D. Elbein, R. J. Molyneux in Alkaloids: Chem. Biol Perspect. (Ed.: S. W. Pelletier), Elsevier, Oxford, 1987, 5, 1-56. [61 R. J. Nash, A. A. Watson, N. Asano in Alkaloids: Chem. Biol. Perspect. II (Ed.: S.W. Pelletier), Elsevier, Oxford, 1996, p. 345-376. [7] K. S. Manning, D. G. Lynn, J. Shabanowitz, L. E. Fellows, M. Singh, B. D. Schrire, J. Chem. Soc. Chem. Comm., 1985, 127-129. [8] G. W. J. Fleet in Topics in Medicinal Chemistry (Ed.: P. R. Leeming), Cambridge. University Press, Cambridge, 1987, p. 149-161. [9] G. W. Fleet, L. E. Fellows, B. Winchester in Bioactive Compounds from Plants (Eds.: D. J. Chadwick, J. Marsh), John Wiley Sons, Chichester, 1990, p. 112-125. [10] L. E. Fellows, R. J. Nash, ScL Procress Oxford, 1990, 245-261. [11] R. J. Molyneux, Phytochem. Analy., 1993, 4, 193-204. [12] T. Niwa, S. Inouye, T. Tsuruoka, Y. Koaze, T. Niida, Agric. Biol. Chem., 1970, 34, 966-967. [13] T. Niwa, T. Tsuruoka, H. Goi, Y. Kodama, J. Itoh, S. Inouye, Y. Yamada, T. Niida, M. Nobe, Y. Ogawa, J. Antibiotics, 1984, 37, 1579-1587. [14] Y. Mikaye, M. Ebata, J. Antibiotics, 1987, 40, 122-123. [15] T. Nishikawa, N. Ishida, J. Antibiotics, 1965, 18, 132-133. [16] S. Inouye, T. Tsuruoka, T. Niida, J. Antibiotics, 1966, 19, 288-292. [17] N. Ishida, K. Kumagai, T. Niida, K. Hamamoto, T. Shomura, J. Antibiotics, 1967, 20, 62-65. [18] N. Ishida, K. Kumagai, T. Niida, T. Tsuruoka, H. Yumoto, /. Antibiotics, 1967, 20, 66-71. [19] M. Koyama, S. Sakamura, Agric. Biol. Chem, 1974, 38, 1111-1112. [20] M. Yagi, T. Kouno, Y. Aoyagi, H. Murai, Nippon Nogei Kagaku Kaishi, 1976, 50, 571-572. [21] H. Paulsen, I. Sangster, K. Heyns, Chem. Ber., 1967, 100, 802-815. [22] D. D. Schmidt, W. Frommer, L. Miiller, E. Truscheit, Naturwissenschaften, 1979, 66, 584-585. [23] S. Murao, S. Miyata, Agric. Biol. Chem, 1980, 44, 219-221. [24] L. E. Fellows, E. A. Bell, D. G. Lynn, F. Pilkiewicz, I. Miura, K. Nakanishi, /. Chem. Soc. Chem. Comm., 1979, 22, 977-978. [25] S. V. Evans, L. E. Fellows, T. K. M. Shing, G. W. J. Fleet, Phy to chemistry, 1985, 24, 1953-1955. [26] G. C. Kite, L. E. Fellows, G. W. J. Fleet, P. S. Liu, A. M. Scofield, N. G. Smith, Tetrahedron. Lett., 1988, 29, 6483-6486. [27] N. Asano, M. Nishida, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, J. Nat. Prod., 1997,60,98-101. [28] A. Kato, N. Asano, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, J. Nat. Prod., 1997, 60, 312-314. [29] A. D. Elbein, M. Mitchell, B. A. Sanford, L. E. Fellows, S. V. Evans, J. Biol. Chem. 1984, 259, 12409-12413. [30] S. V. Evans, L. E. Fellows, E. A. Bell, Biochem. System. Ecol, 1985, 13, 271-302. [31] A. Welter, J. Jadot, G. Dardenne, M. Marlier, J. Casimir, Phy to chemistry, 1976, 75, 747-749.

28 [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63]

2 Taxonomic Distribution of Iminosugars in Plants... G. W. J. Fleet, P. W. Smith, Tetrahedron, 1987, 43, 971-978. R. J. Nash, E. A. Bell, J. M. Williams, Phytochemistry, 1985, 24, 1620-1622. J. Furukawa, S. Okuda, K. Saito, S. I. Hatanaka, Phytochemistry, 1985, 24, 593-594. D. W. C. Jones, R. J. Nash, E. A. Bell, J. M. Williams, Tetrahedron. Lett., 1985, 26, 3125-3126. N. Asano, K. Oseki, E. Tomioka, H. Kizu, K. Matsui, Carb. Res., 1994, 259, 243-255. S. M. Colgate, P. R. Dorling, C. R. Huxtable, Aust. J Chem., 1979, 32, 2257-2264. L. D. Hohenschutz, E. A. Bell, P. J. Jewess, D. P. Leworthy, R. J. Pryce, E. Arnold, J. Clardy, Phytochemistry, 1981, 20, 811-814. R. C. Bernotas, B. Ganem, Tetrahedron Lett., 1984, 25, 165-168. R. J. Molyneux, Y. T. Pan, J. E. Tropea, M. Benson, G. P. Kaushal, A. D. Elbein, Biochem, 1991, 30, 9981-9987. C. Harris, B. C. Campbell, R. J. Molyneux, T. M. Harris, Tetrahedron Lett., 1988, 29, 4815-4817. R. J. Molyneux, J. N. Roitman, G. Dunnheim, T. Szumilo, A. D. Elbein, Arch. Biochem. Biophys., 1986, 257, 450-457. R. J. Molyneux, J. E. Tropea, A. D. Elbein, J. Nat. Prod., 1990, 53, 609-614. I. Pastuszak, R. J. Molyneux, L. F. James, A. D. Elbein, Biochemistry, 1990, 29, 1886-1891. J. Molyneux, M. Benson, R. Y. Wong, J. E. Tropea, A. D. Elbein, /. Nat. Prod., 1988, 51, 1198-1206. R. J. Nash, L. E. Fellows, J. V. Bring, G. W. J. Fleet, A. E. Derome, T. A. Hamor, A. M. Scofield, D. J. Watkin, Tetrahedron Lett., 1988, 29, 2487-2490. C. M. Harris, T. M. Harris, R. Molyneux, J. Tropea, A. Elbein, Tetrahedron Lett., 1989, 30, 5685-5688. R. J. Nash, L. E. Fellows, A. C. Plant, G. W. J. Fleet, A. E. Derome, P. D. Baird, M. P. Hegarty, A. M. Scofield, Tetrahedron, 1988, 44, 5959-5964. R. J. Nash, L. E. Fellows, J. V. Dring, G. W. J. Fleet, A. Girdhar, N. Ramsden, J. M. Peach, M. P. Heggarty, A. M. Scofield, Phytochemistry, 1990, 29, 111-114. R. J. Molyneux in Methods in Plant Biochemistry. Vol. 8, Alkaloids and Sulphur Compounds (Ed.: P. J. Waterman), Academic Press, London, 1993, p. 511-530. A. C. de S. Pereira, M. A. C. Kaplan, J. G. S. Maia, O. R. Gottlieb, R. J. Nash, G. W. J. Fleet, L. Pearce, D. J. Watkin, A. M. Scofield, Tetrahedron, 1991, 47, 5637-5639. R. J. Nash, P. I. Thomas, R. D. Waigh, G. W. J. Fleet, M. R. Wormald, P. D. de Q. Lilley, D. J. Watkin, Tetrahedron Lett., 1995, 35, 7849-7852. D. J. Robins in Methods in Plant Biochemistry. Vol. 8, Alkaloids and Sulphur Compounds (Ed.: P. J. Waterman), Academic Press, London, 1993, p. 175-195. D. Tepfer, A. Goldman, N. Pamboukdjian, M. Maille, A. Lepingle, D. Chevailer, J. Denarie, C. Rosenberg, /. Bact., 1988, 70, 1152-1161. A. Goldmann, M.-L, Milat, P-H. Ducrot, J.-Y. Lallemand, M. Maille, A. Lepingle, I. Charpin, D. Tepfer, Phytochemistry, 1990, 29, 2125-2128. N. Asano, A. Kato, K. Oseki, H. Kizu, K. Matsui, Eur. J. Biochem., 1995, 229, 369-376. N. Asano, A. Kato, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, Carb. Res., 1996, 293, 295-204. A. Kato, N. Asano, H. Kizu, K. Matsui, S. Suzuki, M. Arisawa, Phytochemistry, 1997, 45, 425-429. R. J. Molyneux, R. J. Nash, N. Asano in Alkaloids: Chem. Biol Perspect. Vol. 11 (Ed.: S. W. Pelletier), Pergamon, Oxford, 1996, Chapter 4. N. Asano, A. Kato, H. Kizu, K. Matsui, Phytochemistry, 1996, 42, 719-721. N. Asano, A. Kato, Y. Yokoyama, M. Miyauchi, M. Yamamoto, H. Kizu, K. Matsui, Carb. Res., 1996, 284, 169-178. S. V. Evans, A. R. Hayman, L. E. Fellows, T. K. M. Shing, A. E. Derome, G. W. J. Fleet, Tetrahedron Lett., 1985, 26, 1465-1468. N. Asano, K. Oseki, E. Tomioka, H. Kizu, K. Matsui, Carb. Res., 1994, 259, 243-255.

References

29

[64] M. R. Wormald, R. J. Nash, A. A. Watson, B. K. Bhadoria, R. Langford, M. Sims, G. W. J. Fleet, Carbohydr. Lett., 1996, 2, 169-174. [65] R. Griffiths, A. Watson, H. Kizu, N. Asano, H. Sharpe, M. Jones, M. Wormald, G. W. J. Fleet, R. J. Nash, Tetrahedron Lett., 1996, 37, 3207-3208. [66] G. C. Kite, A. M. Scofield, D. C. Lees, M. Hughes, N. G. Smith, /. Chem. Ecol., 1997, 23, 119-135. [67] J. V. Bring, G. C. Kite, R. J. Nash, T. Reynolds, Bot. J. Linn. Soc., 1995, 777, 1-12. [68] G. C. Kite, H. Sharp, P. Hill, P. Boyce, Biochem. System. Ecol. (In press). [69] J. C. French, M. C. Chung, Y. K. Hur in Monocotyledons: Systematics and Evolution (Eds.: P. J. Rudall, P. J. Cribb, D. F. Cutler, C. J. Humphries), Royal Botanic Gardens, Kew, 1995, p. 255-275. [70] J. M. Horn, D. C. Lees, N. G. Smith, R. J. Nash, E. A. Bell in Proc. 6th Inter. Symp. Plant-Insect Relationships (Eds: V. Labeyris, G. Fabres, D. Lachaise), Dr. W. Jink, Dordrecht, 1987, p. 394. [71] G. C. Kite, L. E. Fellows, D. C. Lees, D. Kitchen, G. B. Monteith, Biochem. Syst. Ecol, 1991, 19, 441-445. [72] G. L. Webster, Taxon, 1975, 24, 593-601. [73] R. M. Polhill in Advances in Legume Systematics Vol. 1 (Eds: R. M. Polhill, P. H. Raven), Royal Botanic Gardens, Kew, 1981, p. 213-230. [74] I. K. Ferguson, J. J. Skvarla in Advances in Legume Systematics Vol. 2 (Eds: R. M. Polhill, P. H. Raven), Royal Botanic Gardens, Kew, 1981, p. 859-896. [75] R. J. Nash, L. E. Fellows, J. V. Dring, C. H. Stirton, D. Carter, M. P. Hegarty, E. A. Bell, Phytochemistry, 1988, 27, 1403-1404. [76] R. J. Molyneux, Y. T. Pan, J. E. Tropea, A. D. Elbein, C. H. Lawyer, D. J. Hughes, G. W. J. Fleet, J. Nat. Prod., 1993, 56, 1356-1364. [77] B. G. Briggs, L. A. S. Johnson, Proc. Linn. Soc. New South Wales, 1979, 102, 157-257. [78] G. C. Kite, C. Selwood, P. Wilkin, M. J. S. Simmonds, Biochem. Syst. Ecol. (in press). [79] A. M. Scofield, P. Witham, R. J. Nash, G. C. Kite, L. E. Fellows, Comp. Biochem. Physiol, 1995, 112A, 187-196. [80] A. M. Scofield, P. Witham, R. J. Nash, G. C. Kite, L. E. Fellows, Comp. Biochem. Physiol., 1995, 112A, 197-205. [81] L. Muller in Biotechnology, (Eds.: H. J. Rehm, G. Reed), VCH, Verlagsgesellschaft, Weinheim, 1985, Chaper 18. [82] Y. Zeng, Glycobiology, 1997, 7 (2), 297-303. [83] A. M. Scofield, L. E. Fellows, R. J. Nash, G. W. J. Fleet, Life ScL, 1986, 39, 645-650. [84] B. Winchester, Biochem. Soc. Trans., 1992, 20, 699-710. [85] Y. T. Pan, H. Hidetaka, R. Saul, B. A. Sanford, R. J. Molyneux, A. D. Elbein, Biochemistry, 1983, 22, 3975-3984. [86] M. J. Schneider, F. S. Ungemach, H. P. Broquist, T. M. Harris, Tetrahedron, 1983, 39, 29-34. [87] J. Martin, in Harpers 19th Edition Review of Biochemistry, 1983, Chapter 33. [88] R. Kornfeld, S. Kornfeld, Ann. Rev. Biochem, 1985, 54, 631-664. [89] W. W. Chen, W. J. Lennarz, J. Biol Chem., 1978, 253, 5780-5784. [90] A. S. Tyms, E. M. Berrie, T. A. Ryder, R. J. Nash, P. M. Hegarty, D. L. Taylor, M. A. Mobberley, J. M. Davis, E. A. Bell, D. J. Jeffries, D. J. Taylor-Robinson, L. E. Fellows, Lancet II, 1987, 1025-1026. [91] G. W. J. Fleet, A. Karpas, R. A. Dewek, L. E. Fellows, A. S. Tyms, S. Petursson, S. K. Namgoong, N. G. Ramsden, P. W. Smith, J. C. Son, F. Wilson, D. R. Witty, G. S. Jacob, T. W. Rademacher, FEBS Lett., 1988, 237, 128-132. [92] D. L. Taylor, P. Sunkara, P. S. Liu, M. S. Kang, T. L. Bowlin, A. S. Tyms, AIDS, 1991, 5, 693-698. [93] L. Ratner, Aids. Res. Human Retroviruses, 1992, 8, 165-173. [94] P. S. Sunkara, T. L. Bowlin, P. Liu, A. Sjoerdsma, Biochem. Biophys. Res. Comm., 1987, 148, 206-210. [95] L. Pereira, M. Hoffman, N. Cremer, Infect. Immun., 1982, 36, 933-942.

30

2 Taxonomic Distribution of Iminosugars in Plants...

[96] W. J. Britt, Virology, 1984, 135, 369-378. [97] L. E. Rasmussen, R. M. Nelson, D. C. Kelsall, T. C. Merigan, Acad Sd., 1984, 81, 876-880. [98] C. D. L. Taylor, L. E. Fellows, G. H. Farrar, R. J. Nash, D. Taylor-Robinson, M. A. Mobberley, T. A. Ryder, D. J. Jeffries, A. S. Tyms, Antiviral Res., 1988, 10, 11-26. [99] R. A. Gruters, J. J. Neefjes, M. Tersmette, R. E. Y. de Goede, A. TuIp, H. G. Huisman, F. Miedema, H. L. Ploegh, Nature, 1987, 330, 74-77. [100] S. P. Ahmed, R. J. Nash, C. G. Bridges, D. L. Taylor, M. S. Kang, E. A. Porter, A. S. Tyms, Biochem. Biophys Res. Comm., 1995, 208, 267-273. [101] G. S. Jacob, M. L. Bryant, in Anti-AIDS drug development: challenges, strategies and prospects (Eds.: P. Mohan, M. Baba) Harwood Academic Publishers, 1995, 65-91. [102] C. G. Bridges, S. P. Ahmed, M. S. Kang, R. J. Nash, E. A. Porter, A. S. Tyms, Glycobiology, 1995, 5, 249-251. [103] D. L. Taylor, L. E. Fellows, G. H. Farrar, R. J. Nash, D. Taylor-Robinson, M. A. Mobberley, T. A. Ryder, D. J. Jeffeies, A. S. Tyms, Antiviral Res., 1988, 10, 11-26. [104] A. Karpas, G. W. J. Fleet, R. A. Dwek, S. Petursson, S. K. Namgoong, N. G. Ramsden, G. S. Jacob, T. W. Rademacher, Proc. Natl. Acad. Sd., 1988, 85, 9229-9233. [105] G. S. Jacob, Current Biology Structural Biol, 1995, 5, 605-611. [106] S. L. Everest, Poisonous Plants of Australia, Angus and Robertson Sydney, 1981, p. 481. [107] C. R. Huxtable, P. R. Dorling, Am. J. Pathol, 1982, 107, 124-127. [108] H. Broquist, Ann. Rev. Nutr., 1985, 5, 391-409. [109] P. R. Dorling, C. R. Huxtable, S. M. Colegate, Biochem. J., 1980, 797, 649-652. [110] R. J. Molyneux, L. F. James, Science, 1982, 276, 190-191. [Ill] G. R. Cao, S. J. Li, D. X. Duan, R. J. Molyneux, L. F. James, K. Wang, L. long, in Poisonous Plants. Proc. 3rd Int. Symp., Iowa State Press, 1992, p. 117-121. [112] D. P. R. Tulsiani, H. P. Broquist, L. F. James, O. Touster, Arch. Biochem. Biophys., 1984, 232, 76-85. [113] D. R. Tulsiani, O. Touster, J. Biol. Chem., 1983, 258, 7578-7584. [114] R. J. Molyeux, L. F. James, K. E. Panter, M. H. Ralphs, Phytochem. Analy., 1991, 2, 125-129. [115] S. Fujieda, I. Noda, H. Saito, T. Hoshino, M. Yagita, Arch. Otolaryngel. Head Neck Surg., 1994, 120, 389-394. [116] P. E. Goss, J. Baptiste, B. Fernandes, M. Baker, J. W. Dennis, Cancer Res., 1994, 54, 1450-1453. [117] M. Hino, O. Nakayama, Y. Tsurumi, K. Adachi, T. Shibata, H. Terano, M. Kohsaka, H. Aoki, H. Imanaka, J. Antibiot., 1985, 38, 926-935 [118] T. Kino, N. Inamura, K. Nakahara, S. Kiyoto, T. Goto, H. Terano, M. Kohsaka, H. Aoki, H. Imanaka, /. Antibiot., 1985, 38, 936-940. [119] S. V. Evans, A. M. R. Gatehouse, L. E. Fellows, Entomol. Exp. App., 1985, 37, 257-261. [120] G. C. Kite, J. M. Horn, J. T. Romeo, L, E. Fellows, D. C. Lees, A. M. Scofield, N. G. Smith, Phytochemistry, 1990, 29, 103-105. [121] M. S. J. Simmonds, W. M. Blaney, L. E. Fellows, J. Chem Ecol, 1990,16, 3167-3177. [122] W. M. Blaney, M. S. J. Simmonds, S. V. Evans, L. E. Fellows, Entomol. Exp. Appl, 1984, 36, 209-216. [123] D. L. Dreyer, K. C. Jones, R. J. Molyneux, /. Chem. Ecol., 1985, 11, 1045-1050. [124] A. A. Watson, M. S. J. Simmonds, E. Porter, W. Robertson, A. N. E. Birch, I. Geoghegan, W. M. Blaney, Phytochem. Soc. Europ. The Netherlands April, 1992; poster. [125] M. S. J. Simmonds, W. M. Blaney, Symp. Biol. Hung., 1990, 39, 17-27. [126] B. K. Mitchell, J. F. Sutcliffe, Physiol. Entomol., 1984, 9, 57-64. [127] A. N. E. Birch, W. M. Robertson, I. E. Geoghegan, W. J. McGavin, T. J. W. Alphey, M. S. Phillips, L. E. Fellows, A. A. Watson, M. S. J. Simmonds, L. E. Porter, Nematologica, 1993, 39, 521-535. [128] K. L. Stevens, R. J. Molyneux, J. Chem. Ecol., 1988, 14, 1467-1473.

3

Glycosidase Inhibition by Basic Sugar Analogs and the Transition State of Enzymatic Glycoside Hydrolysis GUNTER LEGLER

3.1

Introduction

The transition state of a chemical reaction can be regarded as a thermodynamic state populated by an activated molecular species in which the identity of the reactants merges into that of the products [1]. It differs from ordinary molecules in that it is not stable with respect to motions along the reaction co-ordinate but decomposes at the rate of weak molecular vibrations (« lO^s"1) into products or back to the reactants. This frequency and the population density of the transition state determine the experimental rate constant k of the reaction. Assuming the transition state to be in thermal equilibrium with the ground state k can, in principle, be calculated from the difference of the standard free energies AG* between the transition and ground states by Eq. (1) and (2) 7

T

(1) k = — exp (-AG*/RT)

kB h

Boltzmann's constant Planck's constant

(2) AG* = AH* - AS*T

AG* free energy of activation AH* heat (enthalpy) of activation AS* entropy for activation

Because of the involved calculations required to obtain AG* even for simple gas phase reactions, equ. (1) and (2) are normally used the other way around, i.e. to calculate AG*, AH*, and AS* from observed reaction rates. Application of Eq. (1) for the calculation of k for reactions in solution meets with even larger difficulties because free energies depend strongly on solvent-solute interactions which may even affect the reaction pathway. For enzyme-catalyzed reactions a multitude of interactions of the substrate with the active site have to be taken into account. Their energetics could, in principle, be derived from X-ray diffraction data of enzyme crystals complexed with a suitable inhibitor. Translating these data into AG* or AH*, however, requires a model of the geometry and charge distribution of the transition state which is still based on chemical reasoning on reaction pathIminosngars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

32

3 Glycosidase Inhibition by Basic Sugar Analogs...

ways of corresponding non-enzymatic reactions. In the case of glycoside hydrolysis, this model is assumed to resemble a glycosyl cation, as was first proposed for lysozyme in 1967 by Phillips and co-workers [2]. Experimental support for the required sp2-like planar geometry at the anomeric carbon atom came from the strong inhibition of glycosidases by the corresponding glycono-l,5-lactones and from secondary α-deuterium isotope effects on the reaction rate which indicated a change of the hybridization of C-I from sp3 to sp2 in the rate-limiting step (see ref. [3-5] for reviews). Calculations by Warshel [6] on the lysozyme model involved the electrostatic force field of all permanent and induced dipoles and of the two carboxyl groups (Asp-52 and Glu-35) surrounding the subsites D and E of the active site cleft. Stabilization based on ΔΗ* of the cationic transition state for bond cleavage between sites D and E was found to amount to 12 kcal/mol relative to ΔΗ* calculated for the same reaction of a disaccharide model with two attached carboxyl groups. This would correspond to a rate enhancement factor of 7· 108. It was concluded that lowering of ΔΗ* by the suitably oriented dipoles surrounding the active site would provide a much better stabilization of a highly polar transitions state than the fluctuating dipoles of bulk water. When values for ΔΟ* are compared [6b] the rate enhancement factor is reduced to ^ 106 on account of the large (more positive) Δ8* of the reference model. Actual enhancement factors cannot be given because there are no experimental data on the hydrolysis of the (hypothetical) reference compound or the cleavage of an N-acetylglucosaminide bond under neutral conditions.

3.2

Transition State Structure and Inhibitory Potency

The large number of calculations required to obtain transition state energies, even with an empiric model to start with, have made more popular a different approach to transition state structures which is based on the inhibitory strength of substraterelated inhibitors. As was first pointed out by Pauling [7] the active site of enzymes might have envolved towards optimal complementarity to the transition state rather than to the substrate in its ground state. Interactions with such an active site would stabilize the transition state, thereby lowering ΔΟ*, and enhancing the reaction rate. Based on the Eyring postulate of an equilibrium between the ground and transition states (equilibrium constant A^*) and the direct relation of transition state population with reaction rates, Wolfenden [8] has derived an equation which gives the virtual dissociation constant ^TS of the transition state complex relative to the enzyme-substrate complex in the ground states KS (Eq. (3), Scheme 3-1): Thus, ^TTS should be smaller than Ks by the same factor knon/kcat as the rate of the non-catalyzed reaction (knon) is enhanced by the enzyme. This means that a hypothetical compound having exactly the same structure as the transition state but lacking its reactivity should inhibit the enzyme with a K{ that is smaller than ^s (or Km) by the same factor. Of course, it is impossible to synthesize such a com-

3.2 Transition State Structure and Inhibitory Potency

]

33

(3)

K.

^s

k

*cat

ES*

*~

P + E

Scheme 3-1.

pound but one can judge the resemblance of a proposed transition state with the 'real' counterpart from the extent by which K1 of an analog approaches K^sThere are two not uncommon facts which may cause an underestimation of the rate enhancement factor kcat/knon: i) The non-catalyzed reaction is too slow to be detectable, even at temperatures approaching the limit of stability of the substrate. Thus only an upper limit for &n0n can be given. ii) In cases where the enzymatic reaction proceeds by a multi-step pathway (see below) the rate constant for the bond-breaking step may be larger by an unknown factor than the observed value for kcat which is determined by the slowest step of the pathway. As found by the author [9], kn0n could only be determined for glycosides having an aglycon with pKa <7.5, for example, 4-nitrophenol or 4-methylumbelliferone. Values for &non extrapolated from 100 0C to 25 0C compared with kcat for two /?-glucosidases gave rate enhancement factors of *** 1014. The majority of glycosidases have values for kcat with 'good' substrates in the same range as the enzymes studied in ref. [9] (10 to 1000 s"1). Rate enhancement factors up to 1014 are thus quite common and are probably even larger with aliphatic glycosides where knon could not be determined. It follows from this magnitude of enzymatic rate enhancement that non-covalent inhibitors which come near Wolfenden's criterion for transition state resemblance would appear to cause an irreversible inhibition for the following reasons: The rate of complex formation between a low-molecular weight compound and a protein cannot be faster than the diffusion-controlled limit, Le. ^1O7 M"1 s-1 [1O]. As values for K5 fall in the range of 10~2 to 10~5 M, a perfect transition state mimic should have K{ «-* 10~16 M. With K{ = k0ff/kon the dissociation rate £0ff of the enzyme-inhibitor complex would be « 10~9 s"1, i.e. the complex would have a half-life in the absence of excess inhibitor of ^6OO days.

34

3 Glycosidase Inhibition by Basic Sugar Analogs...

3.3

General Mechanisms of Enzymatic Glycoside Hydrolysis and Models of the Transition State

The hydrolysis and alcoholysis (transglycosylation) of glycosides can be regarded as a nucleophilic displacement reaction at the anomeric carbon atom of the glycon moiety. It has been shown by polarimetry, 1H-NMR spectroscopy, and product analysis of transglycosylation experiments that glycosidases can be classified according to the anomeric configuration of the product sugar relative to the substrate in two groups (see [3] for a review): i) 'Retaining' glycosidases transfer the sugar to water or an alcohol to give a product with the same anomeric configuration as the substrate. As proposed by Koshland [11] the action of these glycosidases involves a double displacement with two consecutive inversions. In the first step the aglycon is released and a nucleophile of the enzyme forms a glycosyl-enzyme intermediate with inverted configuration. The sugar is then cleaved from the intermediate with another inversion of configuration at C-I by an activated water or alcohol molecule; the stereospecificity being provided for by the geometry of the activation process and the accessibility of the anomeric center. There are thus at least two transition states on the reaction pathway (Scheme 3-2). Presumably, the general acid AH donating a proton to the glycosidic oxygen in the first step acts as a general base in the second. Depending on aglycon structure and leaving group propensity the first or the scond step may be rate limiting. In a few cases, for example almond /?-glucosidase acting on aryl βglucosides, a conformational change of the enzyme may be rate limiting [12].

Scheme 3-2. Models of the first (a) and second transition states (b) of a retaining /?-glucosidase. Dashed lines symbolize bonds being made or broken. Experimental support for the sp2-like hybridization of the glucosyl unit in (b) comes from studies of α-deuterium secondary kinetic isotope effects with substrates where deglycosylation is rate limiting [13]. For α-glucosidases the spatial orientation of AH and carboxylate are inverted.

3.3 General Mechanisms of Enzymatic Glycoside Hydrolysis...

35

ii) 'Inverting' glycosidases catalyze the hydrolysis of their substrates with the stereochemical course implied by their name. The reaction can be regarded as direct displacement of the aglycon by an activated water molecule [U]. In this respect they resemble the hydrolysis ot the glycosyl-enzyme intermediate of retaining glycosidases with the additional feature of enzyme-assisted aglycon departure. The geometry of this feature in conjunction with that of water activation provides for the stereospecificity. Models of the transition states are depicted in Scheme 3-3. In contrast with the transglycosylation reactions of most 'retaining' glycosidases, no glycosyl transfer from a glycoside to an alcohol has been observed with an 'inverting' enzyme (for exeptions with glycosyl fluorides of 'wrong' anomeric configuration, see [14]). As transglycosylation products are also substrates of the glycosidases which formed them, the 'inverting' enzyme would have to act on a substrate having the 'wrong' anomeric configuration and thereby violate the principle of anomeric specificity. The design and synthesis of inhibitors based on models of the transition state appear profitable for the following reasons: they will advance our understanding of mechanisms of catalysis on an empiric basis because the effects of systematic variations of structural details on inhibitory potency will permit refinements of the model and add experimental support as the values for KJK1 approach Wolfenden's criterion. It is also a rational approach towards effective inhibitors for biochemical and biomedical applications (cf. Chapters 9 and 11). The specificity of an inhibitor will increase with its inhibitory potency because differences in the

Scheme 3-3. Models of the transition state of an inverting-a-glucosidase. a) The reaction proceeds by an SN2-like mechanism as suggested by Koshland [U]. b) Proton-assisted aglycon departure preceeds bond formation with water. The glycosyl cation-like intermediate is based on secondary kinetic isotope effects with glucoamylases [15] which are of similar magnitude as with retaining glycosidases. The catalytic base A~ and the acid AH have been identified as carboxylate and carboxylic groups with several inverting glycosidase by site-directed mutagenesis and X-ray crystallography (see [16] for a review).

36

3 Glycosidase Inhibition by Basic Sugar Analogs...

response of particular enzymes become more pronounced at the lower concentrations permitted by the more potent inhibitors. Only a minor proportion of known inhibitors can, by virtue of a suitably placed substituent on a glycon analog, be regarded as true mimics of the first transition state of 'retaining' glycosidases (Scheme 3-2a). As will be shown later, the interaction of this substituent with the aglycon binding site can make a large contribution to the binding energy. The majority of inhibitors, however, consist only of an analog of the glycon moiety of the substrate. They are thus, strictly speaking, (partial) mimics of the transition state of the second bond-breaking step (Scheme 3-2b) which lack the equivalent of the incoming water molecule. Evaluation of inhibitors by comparing KJK1 with kcai/knon is fairly straightforward with analogs of the first transition state, but meets with a number of difficulties with analogs of the second. Values of kcat can, at least in principle, be obtained with substrates having an aglycon with good leaving group propensity. The cleavage of the intermediate as in Scheme 3-2b differs from the normal pathway of carboxylic ester hydrolysis in that the O-acyl rather than the 0-alkyl bond is broken. Thus, 1-O-acyl glycosides could be suitable models to obtain £non provided they are cleaved at the same bond as the intermediate. Appropriate data are lacking, however. A rough estimate of knon can be based on the neutral hydrolysis of glycosides with an aglycon of an acidity comparable with carboxyl groups. The uncatalyzed hydrolysis of /?-D-galactosides with 2,4-dinitrophenol (pKa. 4.08) and 3,4dinitrophenol (p^a 6.58) proceeds at 25 0 C with k = 4.8 - lO-V^lY] and 1.0 · 10"8S"1 [18], respectively. As values for &cat of the deglycosylation step are in the range of 5 s"1 (calculated for the mammalian cytosolic /?-glucosidase//?-galactosidase from [19]) to 1300 s"1 (for /?-galactosidase from E. coli [17]) kcai/kn0n is in the range of 109 to 1011, which is more than three orders of magnitude smaller than the rate enhancement factor for the first step. This is mainly due to the much larger knon of the second step; in addition, the first step is rate limiting for the vast majority of natural substrates. There is, thus, no evolutionary pressure to increase the rate of the second. Another problem is presented by the definition and determination of K8 because the glycosyl enzyme is both substrate and catalyst in the second step. Lienhard [20] has discussed this problem with enzymes forming a covalent intermediate for chymotrypsin where the active site serine first forms an acyl derivative with peptides and amino acid esters which is then hydrolyzed in the second step. His example, however, refers only to the first step where Ks can be determined as in Scheme 3-1. The value of knon was obtained from a model reaction (N-acetylserine being acylated by an acid amide whith relase of NHs) which can hardly be realized in practice because of the unfavorable equilibrium. Whereas values of knon for the second step can be obtained from the hydrolysis of O-acylserine derivatives we cannot use K1 of the second product, the free sugar in the case of glycosidases, for Ks. The complex formed in the determination of Ki (product) will, of course, contain an extremely small (though unknown) proportion of the covalent intermediate. However, for comparison with kcat/knon we should use \IK\ (product) to replace Ks. Whereas for the first step a small Ks favors the reaction, the reverse is true for the second step. In the opinion of the author, the case is unsettled as present.

3.3 General Mechanisms of Enzymatic Glycoside Hydrolysis...

37

The synthesis of a complete transition state analog for 'inverting' glycosidases presents a challenge that has as yet to be met. The difficulties are not due to the requirement for a penta-co-ordinated, bipyramidal equivalent of the anomeric carbon atom as in the original Koshland model (Scheme 3-3a) but are caused by the specificity of the majority of the enzymes belonging to this class. Glucoamylases, /?-amylases, and several cellulases have oligo- and polysaccharides as natural substrates and most of them have active sites consisting of barrel-shaped protein folds with several contiguous subsites adapted to the monosaccharide units of their substrates [16]. A high occupancy of these sites is required not only for efficient catalysis but also for tight binding of inhibitors. The transition state of 'inverting' glycosidases seems to resemble the Phillips mechanism for lysozyme [2] in that a glycosyl oxocarbenium ion is formed on the acid-catalyzed departure of the aglycon (Scheme 3-3b). Experimental support for a change from a sp3- to an sp2-like hybridization of the anomeric carbon comes from measurements of α-tritium kinetic isotope effects [15] and from the inhibition of glucoamylase by o-glucono-l,5-lactone (^ 1.08 mM [21]) which is 120-fold better than by a//?-D-glucopyranose and 500-fold better than by methyl a-D-glucopyranoside. This may not appear much but is similar to most other α-specific glycosidases [4]. Evidence fot the stabilization of a positive charge at the catalytic site comes from the powerful inhibition by basic analogs of D-glucose and of malto-oligosaccharides. Glucoamylase from Aspergillus niger is inhibited by 1-deoxynojirimycin (1) with Kv 2.1μΜ and by acarbose (2) with K1 <0.006 nM [22]. As maltoheptaose

38

3 Glycosidase Inhibition by Basic Sugar Analogs...

binds with Ks mM [23], the evaluation of 2 according to Wolfenden gives KJK1 > 107 which is one of the largest values for KJKi known in this series of compounds. On the other hand, as fccat values for glucoamylases acting on malto-oligosaccarides are «4 O2 s"1 and the uncatalyzed hydrolysis of oligosaccharides cannot be measured (&non < 10"12 s"1 [9]) we get kcat/knon > 1014, which is more than seven orders of magnitude larger than KJK1 for one of the best inhibitors of glucoamylase known to date. The much larger A^-values (relative to K1 of 2) of D-g/wo?dihydrocarbose (K1 14 nM) and of a-methylacarvioside (3, 1.6 μΜ) demonstrate the importance of a more planar ring geometry and of interactions with the aglycon binding sites, respectively [22]. /?-Amylase, which cleaves ^-maltose from the non-reducing end of a1,4-glucan chains is not inhibited by 1 and 4 [24] because it requires interactions with two glucose units from the cleavage site.

3.4 Basic Sugar Analogs as Glycosidase Inhibitors The first publications on the powerful inhibition of glycosidases by basic sugar analogs date back to the isolation and synthesis of the antibiotic nojirimycin (5amino-5-deoxy-D-glucopyranose, 4) 30 years ago [25, 26]. Five years later, another class of glycosidase inhibitor was introduced by Lai and Axelrod [27] who reported that glycosylamines (5) have ^-values in the sub-millimolar range with a- and yff-glycosidases corresponding in glycon specificity with their structure. Whereas the inhibitory potency of 4 was tentatively ascribed to its (reversible) dehydration to the imine and resemblance of the iminonium ion with the putative glycosyl cation intermediate of glycoside hydrolysis [28], Lai and Axelrod proposed that glycosyl amines would interact with the carboxyl groups of the catalytic site and thus be bound more tightly than the parent hexoses. Subsequent studies have shown that the following features of basic sugar analogs are of importance for their inhibitory potency: 1) Position of the basic (cationic) center 2) Basicity 3) Geometry and charge distribution at the anomeric position 4) Hydroxylation pattern, ring size and flexibility as determinants of specificitiy 5) Interactions with the aglycon binding site 6) Hydrogen bond formation with the catalytic acid

3.4.1

Position of the basic (cationic) center

The basic sugar analogs of Type 1, 4 and 5 inhibit a- and ^-glycosidases from a few hundred up to 106-fold better than the corresponding hexoses (Table 3-1). A generally accepted explanation is the formation of an intimate ion pair consisting of the protonated inhibitor and a carboxylate group of the catalytic site. As ionic inter-

39

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

actions are weak in aqueous solutions the large enhancement of the inhibition by the replacement of the ring oxygen with -NH- requires that the access of water is greatly restricted to this part of the active site. This was confirmed by inhibition studies in buffers of various concentrations (5 to 300 mM) where K1 was constant or varied in the same way as Km of the neutral substrates [4]. Exceptions to the above generalizations are a weak inhibition of some α-glycosidases by glycosylamines (probably due to the preponderance of the ^-configurated anomer) and lack of enhanced inhibition of the N-acetylglucosaminidase from Asp. niger where the basic sugar analogs inhibit no better than N-acetylglucosamine [29]. As this enzyme also differs from other N-acetylglucosaminidases in not being inhibited Table 3-1. Glycosidase inhibition by N-substituted glycosylamines and type 1 and type 23 azasugars (^-values in μΜ) Inhibitor

Inhibitor

^i

Ki

α-Glucosidase I, bovine ER [106]

1

3

a-Glucosidase II, bovine ER [106]

1

120

α-Galactosidase, coffee beans [39]

GaI-NH2 gal-l

7.5 0.0016

α-Mannosidase, bovine ER [107]

man-l

7

man-l-CHi man-l-Cp

75 140

α-L-Fucosidase, bovine liver [41]

nor-fuc-l

2.2

nor>c-l-(CH3)2 nor-/wc-l-Cp

305 18

Invertase, yeast [83]

23

3.5

23-CH3 23-(CH3)2

20 130

/?-Glucosidase, almonds [45]

GIc-NH2

210

GIc-NH-Bn

0.32

/?-Glucosidase, human, lysos. [105]

GIc-NH2 1

350 65

GIc-NH-Ci2H25 1-Ci2H25

0.00015 0.0046

/?-Gluco-//?-Galactosidase bovine, cytosolic [75]

GIc-NH2

240

1

210

gal-l

570

GIc-NH-C4H9 GIc-NH-Ci2H25 1-C4H9 1-Ci2H25' gal-l-C7H15

0.7 0.0015 850 3.8 0.6

GaI-NH-C7-Hi5 gal-l-C7-Hi5

0.0057 0.26

NAc-I-Cp NAc-l-(CH3)2

0.4 4.0

/?-Galactosidase, E. coli [39, 49]

GaI-NH2 gal-l

/?-AT-Acetylhexosaminidase [38] bovine

NAc-I

7 12.5

0.76

1-CH3 1-(CHs)2 1-Cp 1-CH3 1-Cp gal-l-C7Hl5

0.3 0.4 8 130 500 3.2

Abbrevations: ER, endoplasmatic reticulum; Cp, 5-carboxypentyl; GaI-NH2, o-galactosylamine; GIc-NH2, D-glucosylamine; nor-/wc-l, nor-fuco 1 (l,5-dideoxy-l,5-imino-D-arabinitol; NAc-I, 7V-acetylglucosamine analog of 1.

40

3 Glycosidase Inhibition by Basic Sugar Analogs... OH H

,OH

OHH

Scheme 3-5.

by N-acetylconduramine B trans-epoxide [30] it could well be that it acts by a different ('inverting'?) mechanism. The data summarized in Table 3-1 reveal that sugar analogs with an endocyclic basic (cationic) center (type 1 and 4) are up to 104-fold better inhibitors than glycosylamines (type 5). It is tempting to speculate that this reflects an adaptation of the active sites to a glycosyl cation-like transition state as depicted in Scheme 3-2. The endocyclic location of the positive charge in the protonated type 1 and 4 inhibitors bears a greater resemblance with the charge distribution in the glycosyl cation. If this is correct we have to assume that the flexibility of the carboxylate group participation in ion pair formation is insufficient for an adjustment to the exocyclic cationic center of the glycosylamines. Evidence for the restricted mobility of the carboxylate group comes from studies with 2-amino-2-deoxyhexoses in glycosyl methylamines 6 which inhibit no or at most 40-fold better than the corresponding hexoses. In some cases, a weak inhibition by the type 6 compounds might be due to their high basicity (pKa > 9) as discussed in Section 3.4.2. An interesting aspect of how the position of the basic center affects the inhibitory potency of sugar analogs was adressed by Jespersen et αϊ. [31] who synthesized isofagomine (7) where the anomeric carbon atom rather than the ring oxygen is replaced by the -NH-group. Whereas 7 inhibited α-glucosidase from yeast about 3-fold weaker than 1 it was an almost 500-fold better inhibitor for /?-glucosidase from almonds (K^ 0.1 μΜ vs. 47 μΜ with 1 at pH 6.8). The large inhibition enhancement caused by the positional shift of the -NH-group is surprising because 7 lacks the OH-group on C-2. A comparison of 1 with fagomine (8) shows that the deoxygenation of C-2 causes a 280-fold increase of K1 [32]. A nojirimycin-like 1azasugar might thus have a K[ in the subnanomolar range, provided that the contributions to the binding energy of individual substitutents are additive as proposed by Kajimoto et al. [32]. The structural motif of 1-azasugars was subsequently extended to o-galactose [33], D-glucuronic acid [34], L-fucose [35, 36], and N-acetylgalactosamine [37]. The data summarized in Table 3-2 show that the shift of the basic center from the position of the ring oxygen to that of the anomeric carbon can have much larger effects with many enzymes than the shift to the exocyclic position with glycosylamines (Table 3-1). With both a- and/?-specific enzymes there are examples with only small effects of the positional shift. But, whereas the inhibition by the type 7 compounds can be up to 1000-fold larger than that of type 1 with/?-specific enzymes, the inhibition of a several α-glycosidases is up to 10 000-fold weaker by the type 7 inhibitors. With bovine

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

41

/?-N-acetylglucosaminidase the different positions of the basic center are probably without effect because this enzyme (Hexosaminidase A) has a 6-fold higher Km for D-g/wcoconfigured substrates than for D-ga/acto-configured ones [38]. The similar inhibition of glucoamylase and of yeast α-glucosidase by 1 and 7 lend additional support for a catalytic environment and transition state of inverting glycosidases which resemble the model in Scheme 3-3b. The results reported for the inhibition of almond /?-glucosidase by D-ga/acto-isofagomine requires special comments because this enzyme has a 5- to 10-fold higher affinity for 4-nitrophenyl- and -thiophenyl /?-o-glucosides than for D-galactosides [44a-d]. Whereas the ^i-values with the Ό-gluco- and v-galacto-l reflect this relationship, those for the type 7 compounds resemble the inhibition by D-glucosyl and o-galactosylamine (^ 2.6 and 1.7 mM, respectively [44b]). As &cat-values for o-glucosides and D-galactosides are very similar one may speculate that both the type 5 and type 7 compounds which have their basic center near or at the position of the anomeric carbon show a greater resemblance with the location of the positive charge in the transition state than the type 1 inhibitors. Dong et al. [3Ic] have ascribed the stronger inhibition of almond /?-glucosidase by isofagomine to a closer position of the carboxylate to the basic (cationic) center than in the complex with 1-deoxynojirimycin. If we take account of the different degrees of ionization of 1 and 7 (see Section 3.4.2) the positional difference of 1.46 A of the NH-groups has caused the K1 of 1 to be 30 000-fold larger than the K1 of 7 corresponding to ΔΔΟ° = 6.2 kcal/mol. This large effect is unlikely to be caused exclusively by charge-charge interactions. Probably, a large contribution to the binding energy comes from interactions which do not only depend on distance but also on the mutual orientation of the partners, such as hydrogen bonds. Table 3-2. Inhibition constants K1 (μΜ) for 1-aza- (type 7) and 5-azasugars (type 1). Structures correspond to the glycon specificities of the enzymes except where indicated otherwise. Enzyme (source) /?-Glucosidase (almonds) pH 6.8 α-Glucosidase (yeast) pH 6.8 Glucoamylase (Asp. awamori) /?-Galactosidase (Asp. oryzae) α-Glucosidase (coffee beans) /?-Glucosidase (almonds)b /?-Glucuronidase (bovine liver) α-L-Fucosidase (bovine) pH 6.8 α-L-Fucosidase (human) pH 7.5 a-TV-Acetylgalactosaminidase (chicken) /i-Af-Acetylglucosaminidase (bovine) a b c d

Type 7

0.11 86 3.7 0.004 -70a 0.06C 0.079 8.4 6.7 0.46 -0.13d

Typel

[3Ic] [3Ic] [3Ic] [33] [33] [33] [34] [35] [36] [37] [37]

47 25 10 0.0016 540 80 0.0013 0.004 0.6

estimated from IC50 = 200 μΜ with S/^m = 2 inhibition by Ό-galacto-configured type 7 and 1 inhibitors estimated from IC50 = 0.19 μΜ with S/Km = 2 inhibition by W-acetylgalactosamine analog of 7; K1 estimated from ICso = 0.42 μΜ

[3Ic] [3Ic] [3Ic] [39] [39] [40] [41] [42]

[38]

42

3 Glycosidase Inhibition by Basic Sugar Analogs...

This would also take account of the relatively large K[ of o-glucosylamine (2600 μΜ) in which the basic center has the same distance from the anomeric position as in 1. Similar considerations would also apply to other systems listed in Tab. 3-2 where the Type 1 inhibitors have up to 40 000-fold smaller ^-values than type 7.

3.4.2

Basicity of the anomeric carbon atom

Azasugars and glycosylamines are of moderate basicity which ranges from pKa 5.1 to 5.6 for nojirimycins (type 4) and glycosylamines (type 5) and from pKa 6.3 to 7.5 for 1-deoxynojirimycins (type 1). The L-/wco-analog of 1 has pKa 8.4 [42]; unsubstituted glycosylmethylamines (type 6) have pKa 9.1 [43]. No p^a-data seem to have been published for type 7 azasugars. Except for the N-acetylgalactosamine analog of 7 their substitution patterns of the imino group resemble bis(2hydroxyethyl)amine, i.e. they are likely to have pKa « 8.9. Two possible mechanisms can be considered for ion pair formation from the protonated inhibitor and an active site carboxyl group: (i) the basic inhibitor equilibrates with the aqueous solvent according to its pKa and the cation associates with the ionized carboxyl group; (ii) the enzyme binds the neutral form which then becomes protonated by an active site carboxyl group. This may seem trivial and the two mechanisms indistinguishable, but experimental data have shown that one prevails almost to the exclusion of the other [4]. Protonation of the inhibitors by water or a carboxyl group requires that they are of sufficient basicity with a lower limit at pKa «* 4. In fact, the low inhibitory potency of N-arylglucosylamines (pKa 1.5 [45]) and N-acyl-1 [46] was taken as evidence that ion pair formation makes an important contribution to the binding energy. A too-large basicity (pKa > 9) will be detrimental for the inhibition of the class (ii) enzymes because only trace amounts of the free base of the inhibitor are present under the usual assay conditions. N-Oxides of tertiary amine derivatives of azasugars represent a special group of basic inhibitors which mostly show a lower inhibitory potency than their parent compounds. They are zwitterionic, weakly basic compounds (pKa < 4 [47]). The first example was castanospermine N-oxide (9, [48]) which inhibited almond βglucosidase with K{ 760 μΜ (2500 μΜ in ref. [32]) vs. 1.5 μΜ with castanospermine itself. With the N-oxide of N-methyl-1 (10) the inhibition of the same enzyme was only 2-fold lower but that of yeast α-glucosidase was more than 30-fold lower than with 1 [32]. The N-oxide of a pseudo-disaccharide based on isofagomine (11 [3Ic]) was a 6-fold better inhibitor for the almond enzyme than the parent compound wheras no difference was seen with yeast α-glucosidase. Glucoamylase, too, was almost equally well inhibited by 11 and the non-oxidized inhibitor. The results published in [32] and [3Ic] show that the oxidation of the amino group causes only a moderate (< 10-fold) decrease or increase in inhibition of

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

43

VX. .g

HO OMe

10

11

the almond /J-glucosidase, irrespective of the position and orientation of the oxygen (axial 'up' in 10, axial 'down' in 11). As castanospermine (pKa 6.1) is of similar basicity as the other inhibitors, the «103 -fold lower affinity of its N-oxide must be caused by detrimental interactions of the equatorial oxygen. As steric crowding can be ruled out (N-methyl-1 inhibits only 3-fold less than 1 [32]), it appears likely that the electrostatic repulsion of the negative oxygen by a closely positioned carboxylate group is responsible for the weak inhibition. Assignment of glycosidases to class (i) or (ii) is straightforward in cases where structurally equivalent neutral and permanently cationic sugar-based compounds like C-glycosyl benzene derivatives and glycosyl pyridinium salts are available. Whereas class (i) enzymes, e.g. /?-glucosidase AS from Asp. wentii [9] will be strongly inhibited by the cationic glycosides, members of class (ii), like /?-glucosidase from almonds [44d, 45] or /?-galactosidase from E. coli [49] are of similarly poor susceptibility to both neutral and cationic types of sugar derivatives. In some cases, the quaternary ammonium salts obtained by N-dimethylation of azasugars have the same or a moderately reduced inhibitory potency compared with that of the parent compounds. A large impairment of the inhibition by this modification my be due to steric effects or to the presence of a class (ii) enzyme. The pH-dependence of the inhibition by basic and cationic glycon derivatives is governed by the correct ionization state of both the inhibitor and the carboxyl group(s) responsible for strong binding [4]. On starting at low pH, l/K{ increases with increasing pH with both class (i) and (ii) enzymes. With the former and with permanently cationic inhibitors, plots of log (I/Ki) = pKi vs. pH are often linear with a slope + 1.0, reflecting the ionisation of the carboxyl group required for ion pair formation. The slope decreases to ± O on passing the pKa of this carboxyl group. The same plot is obtained with basic inhibitors if K1 is based on the concentration of the protonated form [9]. Interestingly, initial slopes of + 1.0 are quite common, indicating that other carboxyl groups are too far away to have a measurable effect on ion pair formation. Interpretation of the pH-dependence with class (ii) inhibitors is more difficult because inhibitor and carboxyl groups often have similar values for pKa and their ionizations overlap. The increase of I/Ki on the low pH side can sometimes be tracted to the deprotonation of the inhibitor. With ^-(D-galactosylmethyl)amine, for example, Bemiller et al. [43] obtained pH-independent /^-values when the free base was used for the calculations which were almost identical with those of

44

3 Glycosidase Inhibition by Basic Sugar Analogs...

D-galactosylamine (pKa 6.1) and galacto-l (pKa 7.1). A decrease in \ΙΚ·γ at higher pH may be related to the deprotonation of the carboxyl group required for the protonation of the inhibitor [5O]. In general, the interpretation of the pH-dependence requires detailed studies with inhibitors of widely differing basicity and comparison with the effects of pH on the kinetic constants kcat and Km [44d]. The inability of class (ii) glycosidases to bind cationic sugar derivatives better than their neutral analogs is marked contrast to their high affinity for basic glycon or substrate-related compounds. The molecular basis of this property is still an open question. Models proposed for /?-glucosidase [51] with a cationic acid or for /?-galactosidase from E.coli [49] with a cationic amino acid close to the active site carboxylate have as yet to be verified or disproved.

3.4.3

Geometry and charge distribution at the anomeric position

The strong inhibition of 'retaining' ^-glycosidases (not α-glycosidases) by hexono-l,5-lactones was at first ascribed [52] to the resemblance of their half-chair conformation and planar geometry at C-I with the putative glycosyl oxocarbenium ion intermediate of glycoside hydrolysis. Geometric factors alone, however, hardly contribute to the binding energy as shown by the low affinity of /?-galactosidase from E.coli for 2,6-anhydro-l-deoxy-D-ga/acto-hept-l-enitol [53] (12, Km 60 mM) which is less than for ligands with tetrahedral 4Ci-conformation (K1 (α/β-Όgalactose) 34 mM; Km (methyl/?-o-galactoside 8 mM). Instead, ion-dipole interactions with the carbonyl group and possibly a hydrogen bond with the partially negatively charged carbonyl oxygen supply the main part of the binding energy. The inhibitory potency of hexono-l,5-lactones (^-values extending into the submicromolar range), and their lability with respect to hydrolysis to weakly inhibitory acids (half-life « 30 min at 25 0C and pH 6-7) and isomerization to the 1,4lactones have stimulated research efforts aimed at the synthesis of stable and perhaps more potent analogs. This was first recognized with the corresponding lactams (13 [26]) with the Ό-gluco-analog and subsequently with other sugar analogs for which nojirimycin isomers had been synthesized. The inhibition by lactams was found to be similar to or often weaker (up to 20-fold) than by the corresponding lactones [4]. /?-glucosidase from almonds seems to be an exception for which Dale et al. [44d] and Hoos et al. [55] report a 5-fold and 3.2-fold better inhibition by o-gluconolactam, respectively. Studies with D-galactono-l,5-lactam [56] where the corresponding lactone turns rapidly into the 1,4-isomer have confirmed the strong inhibition of the ^-specific and weak inhibition of α-specific enzymes with this type of inhibitor. Hydroximolactones and -lactames, also called lactone oximes (14) and lactam oximes (15), represent another group of hexopyranose analogs with sp2-hybridization of C-I and an exocyclic double bond. The D-g/wco-analog of the former was introduced by Beer and Vasella [54] and that of the latter by Tong et al. [58]. The N-acetyl-glucosamine analog of 14 [57] and the D-raanno-analog of 15 [59] have also been prepared and evaluated as glycosidase inhibitors.

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

13

45

14 OH

HO\

ISa

/OH N H

15b

Scheme 3-7.

The inhibition by 14 was found to be 10- to 25-fold lower than by the lactone with /?-glucosidases from almonds and Agrobacterium faecalis (K1 4.3 mM and 0.03 mM, respectively) and was also weak with α-glucosidase from yeast (J^i 6.8 mM) [55], With the N-acetylglucosaminidases from bovine kidney, jack beans, and from the fungus Mucor rouxii the inhibition by the corresponding hydroximolactone was only 2-fold lower than by the lactone itself [57]. In comparison with the lactones there are thus no additional polar interactions with complementary groups of the active site; the detrimental effect of the N-hydroxyl group seen with /?-glucosidases could be due to steric effects. The inhibition of glucosidases by 15, on the other hand, was better than by Dgluconolactone and -lactam. This was most pronounced with yeast a-glucosidase with K{ 2.9 μΜ, i.e. 2300-fold better than by 14. The D-manno-analog of 15 was even a much better inhibitor for α-mannosidase from jack bean (Ki 0.15 μΜ) than the o-raanno-analogs of 1 and 4, for which Ki 60 μΜ and 1.12 μΜ have been reported [6O]. As hydroximolactams are weakly basic (pKa 4.8 [55] or 5.6 [68]), their high affinity with α-specific enzymes was ascribed to the formation of a strong hydrogen bond with the catalytic acid [55] which is of insufficient acidity for complete proton transfer and ion pair formation. Note that the type 15 inhibitors can exist in two tautomeric forms (ISa and 15b). Whereas Papandreou et αϊ. [59] favored the tautomer with the endocyclic double bond it was shown by Hoos et αϊ. [61] by theoretical calculations and 15N-NMR-spectroscopy that hydroximolactams exist almost exclusively as shown for 15a. These authors have also shown that the imino group is Z-configured and that protonation occurs on the exocyclic nitrogen. In contrast with lactones and lactams their hydroximo derivatives permit the attachment of substituents on the N-hydroxyl group for interactions with the aglycon binding site (c.f. 3.4.5). The inhibition by N-phenylcarbamoyl derivatives of 14 [55] and its N-acetylglucosamine analog [57] points to considerable differences in the size and the polar and hydrophobic properties of this site. The enhancement of the inhibition by the substituent was 100-fold with /?-glucosidase from almonds

46

3 Glycosidase Inhibition by Basic Sugar Analogs...

whereas it was > 1000-fold by the smaller N-benzyl substituent of o-glucosylamine [45]. With /?-glucosidase from Agrobacterium faecalis a 200-fold and with α-glucosidase from yeast a 2000-fold better inhibition was seen with the N-phenylcarbamoyl derivative. Nojiritetrazoles (16) are non-basic azasugar derivatives with an sp2 geometry of C-I and an exocyclic double bond, but with a more rigid 4Hs half-chair conformation than 13, 14 and 15. Analogs of o-glucose [62], D-mannose [63], and N-acetylglucosamine [64] have been prepared. Their inhibitory strength is similar to that of the hydroximolactams except with yeast α-glucosidase which was only weakly inhibited (^ 1.3-5.6 mM). The moderate to good inhibition indicates that enzyme-inhibitor (or -substrate) interactions cannot be very close with ring oxygen or -NH- of the lactam derivatives as hardly any steric effects are caused by the tetrazole ring. Even though their /^-values do not extend beyond the μΜ range, the tetrazoles can be considered as transition state mimics in that pj^i for the Dgluco- and o-raannoconfigured isomers with six a- and /?-glucosidases and -mannosidases correlated well with their specific acitivities expressed as log (Vmax/^m) (more appropriate would have been log (A:cat/^m) to correct for differences in molecular mass). An important result with respect to the direction of the catalytic glycoside protonation came from studies with nojiritriazole 17 and -imidazole 18 [65]. In the paradigmatic model for the first transition state of retaining glycosidase (Scheme 3-2a) the acid AH donates its proton to the glycoside oxygen from above the pyranose ring. As already discussed, sp2-hybridized sugar analogs derive their inhibitory potency from polar interactions rather than from geometric resemblance with the glycosyl oxocarbenium ion. With the non-basic ligands the strongest interactions are hydrogen bonds between suitably aligned donor and acceptor pairs. The strong inhibition by 16 and an even larger by 18 [66] were in marked contrast with the lack of inhibition (A^ > 10 mM except with /?-glucosidase from Caldocellum saccharolyticum where 17 was 400-fold less potent than 16). As the only common feature of non-basic good inhibitors with sp2-geometry of C-I is a non-bonding electron pair in an in-plane orbital of the exocyclic heteroatom, the authors conclude that a strong hydrogen bond between the heteroatom and the catalytic acid AH makes a substantial contribution to the binding energy (Scheme 3-9a). This is in keeping with the good inhibition by the hydroximolactams 15 where the exocyclic nitrogen with an 'in-plane' orientation of the nonbonding electron pair has been shown to be the more basic [61]. A similar hydrogen bond with the glycosidic oxygen of substrates could lead to the catalytic pro-

,OH

OH / -^N

/^M

^N,

/

HO-V^^-^N HO

16 Scheme 3-8.

,1Λ^-«-Λ^-^Νν Η0\^ ^ HO

17

I '

OH . .^^--ν-Λ^'-Ν HO-V^^-^N HO

18

47

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

O

:

Q

Y _J

Scheme 3-9. a) hydrogen bond formation in-plane with the pyranose ring with 16; b) proton transfer to the glycosidic oxygen perpendicular to the ring (Koshland model); c) proton transfer to the glycosidic oxygen in-plane with the pyranose ring (Adapted from [65], with permission)

ton transfer during hydrolysis (Scheme 3-9c). This model would also give an explanation for the poor susceptibility of α-glycosidases to the inhibition by glyconolactone, lactams and their hydroximo derivatives because the catalytic acid would be too far away from the lone electron pairs if it is close to the axial glycosidic oxygen. Glyconamidines (19) have both a planar geometry and a strong basic center at C-I and although proposed as good glycosidase inhibitors by Reese et al. [28] in 1971, were only synthesized in 1990 by Ganem and co-workers ([59] and references therein). The Ό-gluco-, O-galacto-, and D-raannoanalogs of 19 and of the amidrazone 20 were synthesized and tested with the corresponding enzymes from almonds, bovine kidney, jack beans, and coffee beans. Inhibition constants were in the μΜ range but comparison with nonbasic and moderately basic sugar analogs [67] showed that the combination of a flat C-I geometry and a positive charge under the assay conditions (pKa (19) 10.6; pKa (20) 8.7) gave, at most, a 10 -fold better inhibition than with hexonolactones and type 4 compounds. As discussed in Section 3.4.6, this is probably due to the inability of strongly basic and thus fully protonated inhibitors to act as hydrogen bond acceptors with the catalytic acid. Glyconamidines can be modified for possible interactions with the aglycon site by substituents on the exocyclic nitrogen. This has been realized for O-manno-19 with the N^benzyl and N^o-deoxy-o-mannopyranoside derivative [68, 69] and for O-gluco-19 with the N1 -butyl and N^dodecyl derivatives [67]. A strong enhancement of the inhibition by the substituents was only seen with mammalian cytosolic and lysomal /?-glucosidases which have an extended hydrophobic agly-

OH

OH

19 Scheme 3-10.

OH

20

OH

H

21

OH OH

H

22

48

3 Glycosidase Inhibition by Basic Sugar Analogs...

con site and which were inhibited by the dodecyl derivative of 19 with K1 values in the subnanomolar range [67]. In the protonated form of 19 the C-I double bond is delocalized between N-I and N-2. Molecular modeling and NMR studies on the D-manno-derivatives have shown [69] that the most stable conformation (E or Z) of the substituent depends on its chemical structure. The ^-configured N-benzyl derivative showed only a moderate similarity with the (calculated) conformation of the transition state of phenyl α-D-mannoside hydrolysis. Another way to integrate the geometry and electric charge of the glycosyl oxocarbenium ion into a stable compound was realized with the cyclic guanidines 21 [70] and 22 [71]. Both were synthesized with substituents on the exocyclic nitrogen: R = 4'-nitro- and 4'-aminophenyl and 4'- and 6'-linked o-glucose with 21 and R = benzyl and O-benzyl with 22. All compounds were strongly basic (p^a>10) except for the O-benzyl derivative of 22 which had pKa 6.8. In spite of their large p^a-values and the lack of hydroxyl groups on the C-2 and C-3 positions, the type 21 guanidines with phenyl substituents and with the 'correct' hydroxyl configuration were good inhibitors for /?-glucosidase from almonds (Ki 30 μΜ) and /?-galactosidase from E. coli (K-v 80 μΜ). The strong inhibition may seem surprising as two hydroxyl groups are missing and both enzymes are strongly inhibited by basic sugar-based inhibitors and only weakly by cationic ones [44d, 45, 49]. In addition 4D-g/wc<9'-21 exists predominantly in the E4-conformation (2Ib) rather than in the 4E conformation (2Ia) of the glucosyl cation. On the other hand, the type 21 guanidines linked with D-glucose were only weak inhibitors

21a

22a

H+

OΛ

ΗΟ

V-"

H0 21b Scheme 3-11.

22b

>

'NH

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

49

for /?-glucosidase (JBTi 10 mM) and /?-galactosidase (^i 5.5 mM). The effect of aglycon-related substituents on the inhibitory potency will be discussed in Section 3.4.5. The type 22a guanidino sugars which have the gem NH-C-OH structure were stable only at high pH as uncharged species. Protonation caused a reversible isomerization to the non-inhibitory guanidinyl furanose 22b. Half-time for the loss of inhibitory potency was 8.7 min at pH 7.5. Initial rates of coffee bean «-galactosidase measured with 22 (R = 0-benzyl) preincubated at pH 10.5 added to substrate strongly buffered at pH 7.5 and 6.0 gave K1 6.4 μΜ, respectively. From the lower inhibition at pH 6.0 the authors concluded that the inhibitor is bound as free base which is protonated by a carboxyl group of the active site. Calculated with the concentration of the free base (from pKa 6.8). The ^-values are 5.3 μΜ (at pH 7.5) and 2.1 μΜ (at pH 6.0), respectively. The lower inhibition at pH 7.5 reflects the deprotonation of the carboxyl group responsible for protonation of the inhibitor. The benzyl derivative of 22 preincubated at pH 13 gave K14.9 μΜ at pH 7.5. If it were also bound as free base, the K{ of the latter would be 0.012 μΜ. This over400-fold better inhibition relative to the 0-benzyl derivative might point to a close fit within the aglycon binding site. /?-Galactosidases from Asp. oryzae and Asp. niger were also inhibited by the type 22 compounds, with Ki ranging from 13 μΜ to 62 μΜ.

3.4.4

Hydroxylation, topography, ring size and flexibility as determinants of specificity

The specificity of glycosidases comprises both binding and catalytic action on their substrates. Accordingly, the specificity constant kcat/Km which includes substrate affinity as well as catalytic efficiency is used to quantify enzyme specificity. Both are determined by numerous weak interactions with the hydroxyl groups and other substituents, e.g. acetamido or methyl groups of the glycon moiety and, generally to a lesser extent, with structural features of the aglycon. Molecular details of enzyme-inhibitor and enzyme-substrate interactions have been obtained for a number of glycosidases by X-ray crystallography and site-directed mutagenesis (see [16] for a review). However, most of these enzymes act on oligo- and polysaccharides. Thus, for the majority of glycosidases such information has to be obtained from kinetic studies with systematically modified substrates and inhibitors. In order to judge the specificities of inhibitors, those of the test enzymes must be known, as overlapping specificities are not uncommon. On the one hand, there are enzymes like /?-galactosidase from E. coli [72] or /?-glucosidase AS from Asp. 4 6 wentii [73] with a 10 -fold and 10 -fold discrimination between /?-galactosides and -glucosides, respectively. On the other hand, an almost complete lack of discrimination between Ό-gluco- and D-ga/acto-configured substrates is shown by mammalian lysosomal N-acetylhexosaminidases [38, 74] and by mammalian 'broad specificity' cytosolic /?-glucosidases (human [19], cow [75], rabbit [76]) where the lack of specificity extends to the substituents on C-5. Based on kcat/

50

3 Glycosidase Inhibition by Basic Sugar Analogs...

Km, /?-D-fucosides and α-L-arabinosides are even better substrates than D-glucoand D-galactosides [19]. A commerical preparation ot the latter enzyme of bovine origin has erroneously been termed 'β-galactosidase' [62, 77] which caused an unwarranted assignment of a broad specificity to serval inhibitors with the Dg/wco-configuration [32, 58]. /?-Glucosidases from sweet almonds [44], which are favorite test enzymes for testing glycosidase inhibitors, and from Agrobacterium faecalis [78] show an intermediate degree of O-gluco-/O-galacto-discnminauon. Comparison of data for the almond enzyme from different laboratories, however, is rendered difficult, because various iso-forms exist which are generally not specified by the suppliers. For example, published values for relative specificity constants kcat/Km (glucoside)/£cat/^m(galactoside) range from 3.8 [44b] to 23 [44a] and inhibition constants for 1-deoxynojirimycin from 0.32 mM [44d] to 1.7 mM [79] and for 23 (Tab. 3-3) from 1.7 μΜ [83] to 160 μΜ [79]. 3.4.4.1

Five-membered azasugars

The first reported five-membered azasugar, 2,5-dideoxy-2,5-imino-D-mannitol, 23, was isolated from a leguminous plant in 1976 [80] but the wide inhibitory potential of this class of compounds was recognized only in 1985 [81]. At first sight, the type 23 and related compounds bear less structural resemblance with the pyranosyl cation-like transition state than the corresponding inhibitors with a hexapyranose-like structure. However, as shown in Table 3-3, they inhibit many glycosidases more than 100-fold better than the corresponding 1-deoxynojirimycins 1. This is most pronounced with α-glucosidase from yeast and 26, /?-glucosidases from almonds and Agrobacterium faecalis (K{ (1) 12 μΜ and K[ (23) 0.2 μΜ [87]), and α-mannosidase from jack beans and 24. As proposed by Sinott [3], this can be rationalized by the better resemblance of hydroxyl group orientation in the half-chair furanose structures with the transition state than in the chair conformation of the type 1 inhibitors. Note that bond angles in the pentagon are 108°. The pyrrolidine-type inhibitors could thus adopt a planar conformation but this would put hydrogens and substituents in an eclipsed position. To relieve the resulting strain (5.7 kcal/mol in cyclopentane) these compounds adopt conformations resembling an envelope or twisted half-chair. Support for Sinnott's explanation of larger inhibitory strength of five-membered azasugars comes from results with the cyclopentylamine derivatives 27 (mannostatin A [88] and 28 [89]. They inhibit jack bean α-mannosidase even better than 24 and 25 with Ki 0.07 μΜ and « 0.03 μΜ (estimated from ICs0), respectively.

27 Scheme 3-12.

28

29

30

1.9 [86]

C

>1,000 [81]

0.5 [85]

- 80 [81]

-220 [81]

-250 [81]

24

I

-7 [81]

-70 [81]

-0.1 [81]

130 [84]

> 1,000 [81]

25

V\IX jft

λ HO J — N

H

H )— N

- 0.1 [81]

-50 [81]

- 100 [81]

26

HO

(Q^ c\IX

~^\

I

a

0.14 [29]

d

68 [60]

b

13 [84]

a

0.0016 [39]

250 [83],

0

HO NH

26 [82]

6.5 [82]

67 [82]

17 [79] 13 [82]

4.8 [79]

29 [82]

37

^^ OH

^\

V_ HO Y"

0.44 [79]

13 [4a] 25 [83]

1,700 [79] 320 [44d]

1

HoW OH

N

H

:<9^ >

H °\

— ) approximate K1 estimated from ICso assuming competitive inhibition and [S] = Km', a ) with Ό-galacto analog of 1; b) with Ό-manno analog of 1; c) with 1-acetamido-l-deoxy analog of 23 (l-acetamido-l,2,5-trideoxy-2,5-imino-Dmannitol); d)with 2-acetamido-2-deoxy analog of 1

/^-W-Acetylglucosaminidase (jack bean)

α-Mannosidase (jack bean)

/?-Galactosidase (Asp. niger)

α-Galactosidase (coffee beans) > 1,000 [81]

44 [83]

/?-Gluco-//?-Galactosidase cytosolic (bovine kidney)

-1.6 [81] 0.73 [83]

-4 [81] 1.7 [83] 160 [79]

23

\

H

HO / — N \/ OHV V lX \IX

HO-)

*— OH

/?-Glucosidase (almonds)

α-Glucosidase (B. stearothermoph.)

α-Glucosidase (yeast)

Enzyme

HO

/ OH\ Ι^ίΐλ \ c\IX

~Λ/— N

Η

HO

Table 3-3. Inhibition constants K1 (μΜ) of five-, six-, and seven-membered azasugars. Structures were drawn in an orientation giving maximal overlap of -NH- and, where applicable, of -CH2-OH corresponding to C- 6 of hexoses.

a

I; ^H

I

f1

OQ

^.

OQ

Co

4--,

52

3 Glycosidase Inhibition by Basic Sugar Analogs...

The pyrrolidine-type azasugar 29 bearing some structural resemblance with Lrhamnose and L-fucose has been synthesized by Provencher et al [9O]. In aqueous solution it is in a pH-dependent equilibrium with the nojirimycin-like hydration product 30 and a dimer (not shown). α-L-Rhamnosidase from Penicillium decumbens was inhibited by 29/30 with Ki 0.14 μΜ with a slow approach (Vn « 2 min) to the inhibition. The authors have ascribed this to a reaction of 29 with a nucleophile of the active site. However, as the residual enzyme activity was ^ 40 % after « 15 min in the presence of 0.16 mM inhibitor, a covalent inactivation appears unlikely. A slow approach to the inhibition is a widespread phenomenon with glycosidase inhibitors of various types, including many which are not susceptible to a covalent reaction under physiologic conditions (see [4] for discussion). The inhibition of α-L-rhamnosidase by 31 with £4 5.5 μΜ is in accord with most other glycosidases in that nojirimycins are better inhibitors than their 1-deoxy derivatives (c.f. Section 3.4.6). Furthermore, both 29/30 and 31 inhibited a-L-rhamosidase better than the L-rhamno analog of 1 (^i 62 μΜ [90] and K1 36 μΜ [91]) even though they lack the C-2 hydroxyl group. A larger inhibitory potency of pyrrolidine-type compounds with appropriate substitution pattern is, however, not a general feature. For a substantial number of enzymes, six-membered azasugars are more powerful inhibitors than their five-membered ring counterparts. Examples are: α-galactosidase (coffee beans) is inhibited 70-fold better by Ό-galacto-l than by 25; /?-glucosidase (Asp. wentii) 30-fold better by 1 than by 23 [83]; N-acetylglucosaminidase 13-fold better by the 2-acetamido analog of 1 than by the 1-acetamido analog of 23 (Table 3-3). The different response to the altered structures may reflect differences in active site geometry and transition state structure which also show up in the inhibition of α-glucosidases and α-mannosidases by rigid analogs of hydroxylated pyrrolidines and piperidines (c.f. Section 3.4.4.3). Except for N-acetylhexosaminidases, which probably act with acetamido group participation [30, 92] there is no experimental evidence for other catalytic mechanisms than discussed in the introduction.

)H

31

OH

32

OH H

HO°

35 Scheme 3-13.

36

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

53

A better inhibition by the six-membered, substrate-related azasugars than by five-membered ones is also seen with α-L-fucosidase. The enzyme from bovine kidney was inhibited (at pH 5.5) by 31 with Ki 1.6 μΜ and by 32 with K1 1.4 μΜ [90] whereas the L-/wco-analog of 1 was a 950- and 170-fold better inhibitor, respectively [41, 42]. Three azasugars derived from 32 by configurational inversion of the OH on C-3 and, in addition, of the methyl or hydroxymethyl group inhibited with ^-values ranging from 4 μΜ to 22 μΜ [90], These small effects of structural modifications are in striking contrast with the effects of alterations of similar extent on the inhibition by six-membered azasugars. L-rhamno-deoxynojirimycin (33) can be regarded as an analog of the L-/wc6>-isomer (34) where the configurations of C-2 and C-4 are inverted. Compound 33 (^i 2.2 μΜ at pH 6.7 [91]) was 1600-fold less potent than 34 (K{ 0.0014 μΜ [41]). Furthermore, the D-manno-analog of 1, (35) which differs from 34 by the absence of the CHsand an additional -CH2OH group when bound in the same orientation, has K1 9.5 μΜ [41] which is similar to Λ^-values of 35 and 36 in relation to that of 34 point to a tight fit of the CHs-group within the glycon binding site of a-L-fucosidase which probably cannot be realized with the five-membered 32 and its derivatives. The loose fit of the latter compounds seems to permit sufficient interactions with substituents in a 'wrong' orientation so as to cause a leveling of the inhibitory strength to about 10-fold larger than L-fucose and about 300-fold larger than methyl α-L-fucoside [4I]. 3.4.4.2

Seven-membered azasugars

Polyhydroxylated azepanes are more flexible than the corresponding pyrrolidines and piperidines and can adopt several puckered low-energy conformations (see ref. [82] for results of molecular modeling studies). They are thus able to adapt to the space filling and polar requirements of glycosidases having different glycon specificities as shown in Table 3-3 for compound 37 [79, 82]. In some cases, for example with /?-glucosidase from almonds and /?-galactosidase from Asp. niger, 37 inhibits even better than the glycon-specific type 1 piperidines, a finding which may have a similar explanation as proposed for strongly inhibiting pyrrolidines. The isomeric tetrahydroxy azepane 38, having two pairs of vicinal hydroxyl groups, was weakly active or non-inhibitory with the enzymes listed in Table 3-3 except for bovine N-acetylhexosaminidase (Ki 4.6 μΜ [82]; Ki (2-acetamidodeoxy-1) 0.14 μΜ [29, 38]). Conflicting results were reported for a-L-fucosidase from bovine kidney for which Merrer et al. [79] gave #1 (37) « 1000 μΜ and K{ (38) 28 μΜ whereas Qian et al. [82] gave data for %-inhibition from which K1 (37) « 125 μΜ and Kj (38) « 1000 μΜ can be estimated. Nevertheless, both sets of data demonstrate that a methyl group and appropriately oriented hydroxyl groups are indispensible for tight inhibition of α-L-fucosidase. Whereas the spatial arrangement of hydroxyl groups of glycon-based azasugars resembles the transition state only to a moderate extent it should be possible for open-chain analogs to adopt a conformation with a close correspondence to the glycon binding site. That this is indeed possible has been shown with α-glucosidase from yeast and

54

3 Glycosidase Inhibition by Basic Sugar Analogs...

HOx

OH

HOx

<

"HO^

HO

39

40

41

Scheme 3-14.

its inhibition by open-chain, partly truncated analogs of nojirimycin and its 1deoxyderivative [93]. Compounds 39, 40, and 41 were similar or up to 4-fold better inhibitors (Ki 4 μΜ) than 1 and 4. In order to appreciate these results the large entropy loss from freezing out four or five rotational degrees of freedom should be taken into account when these flexible inhibitors bind to the enzyme, which puts them at a disadvantage over their cyclic analogs. According to Page and Jencks [94] the entropy loss per degree of freedom amounts to about 5 cal/Kmol, corresponding to an increase of the binding energy of 6-7 kcal/mol. Thus, if these open-chain amino alcohols could be fixed in the conformation they adopt in the active site they would inhibit with ^-values in the subnanomolar range. It remains to be tested if other glycosidases respond in the same way to open-chain azasugars of appropriate structure. 3.4.4.3

Rigid azasugar derivatives

Bicyclic azasugar derivatives like swainsonine (42), australine (43), castanospermine (44), and calystegine BI (45) differ from monocyclic analogs mainly by the restriction of conformational flexibility of the polyhydroxylated ring by a four(42), three- (43, 44), or two-carbon bridge (45). Inhibition studies (reviewed in [4a]) with 42 and several α-mannosidases and with 44 and various a- and /?-glucosidases revealed that these bicyclic azasugars were, in the majority of cases, better inhibitors than their monocyclic counterparts, with ^-values down to 0.5 nM. H

HO /\

HO^

HO

H

-

°

/

HO -\\

^\ ' \

HO

42 Scheme 3-15.

43

44

45

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

55

With a few enzymes, however, for example α-mannosidase from almonds, Golgi mannosidase I involved in glycoprotein biosynthesis, as well as a-glucosidase from yeast, no inhibition could be observed with up to 1 mM of 42 and 44, respectively. As bicyclic azasugars are of similar basicity as the monocyclic inhibitors (for example pKa (42) 7.4 [95]; pKa (44) 6.1 [48]) their interactions with the active site will resemble those discussed in Section 3.4.2, including interactions of the hydroxyl groups with complementary functionalities responsible for the glycon specificity of the enzymes. From the good inhibition of α-mannosidases by 42 and of most a- and /?-glucosidases by 44 we can conclude that there is sufficient room on one side of the glycon binding site to accommodate the €4- and Cs-bridge of the inhibitors. This is probably not the case with glycosidases which are strongly inhibited by monocyclic azasugars but not measurably by their bicyclic counterparts. Another explanation for the inability of some glycosidases to bind rigid azasugars with correct hydroxylation could be an active site which has envolved towards a transition state with spatial requirements which can be fullfilled by the more flexible monocyclic azasugars but not by the rigid bicyclic ones. In many cases, the bridged azasugars inhibit much better than the monocyclic analogs, a feature which is probably related to a better resemblacne of their hydroxyl group topography with the transition state and to entropic factors resulting from the rigid structure of the former. This latter aspect is most obvious for the equivalent of the -CHiOH group on C-5 of the substrate sugar which is presumably fixed in a position resembling that in the transition state. Enlargement of the Csbridge of 44 by a -CH2-group caused an increase of K1 for /?-glucosidase from almonds from 1.5 μΜ to 260 μΜ [96]. Even larger effects resulted from four other configurational inversions of hydroxyl groups, including those which gave homo-castanospermine analogs of o-mannose and D-galactose [97]: inhibiton of glucosidases was completely abolished and no inhibition could be observed with up to 2 mM inhibitor with α-mannosidase and /J-galactosidase. Exploratory studies by Cenci di Bello et al. [95] and by Elbein et al. [98] on the effects of configurational alterations of 42 on the inhibition of various (crude) amannosidases from human liver have shown that the epimerization of C-8, C-Sa and C-2 caused a more than 1000-fold reduction of the inhibitory potency. Similar studies with 44 and glucoamylase from Asp. niger, an inverting enzyme, by MoIyneux et al. [99] showed that this enzyme responded much less (3- to 10-fold) to epimerizations at C-6 and C-I. No inhibition of α-mannosidase (jack bean) was observed with 6-epi-44, even though it has the D-mannoconfiguration. This is in accordance with a preference of α-mannosidase for azasugars with vicinal cishydroxyl groups in a flatterend ring system as in 24, 25 and 42 [97]. Australine (43) can be considered a rigid analog of 2,5-dideoxy-2,5-imino-Dmannitol (23) with the C-5 hydroxy!methyl group fixed by an ethylene bridge. It is a good inhibitor (K1 1.5 μΜ vs. K1 (1) 2.1 μΜ [22]) [100] but in contrast to 23 it does not inhibit a- or /?-glucosidase from yeast and almonds, respectively. Studies on structure-function relations [97] showed that 7-epi-44 was a 12-fold better inhibitor for glucoamylase than 43 whereas enlargement of the monohydroxylated ring by a -CH2-group was slightly detrimental (^i 4.5 μΜ).

56

3 Glycosidase Inhibition by Basic Sugar Analogs...

Calystegines were first described by Molyneux et al [101] as glycosidase inhibitors and studied in detail by Asano et al. [102]. They can be regarded as bridged analogs of 1-deoxynojirimycin (1) or of isofagomine (7) as indicated in the formula scheme for the most potent /?-glucosidase inhibitor of this class, calystegine 62 (45). Its hydroxylation pattern resembles that of D-glucose and it inhibits /?-glucosidases from almonds and Caldocellum saccharolyticum with K{ 1.2 μΜ and 0.55 μΜ, respectively, which is ^100-fold better than 1. α-glucosidase inhibition, on the other hand, was weak for the enzyme from rice (Ki ^ 30 μΜ; Ki (1) 0.01 μΜ) or could not be detected with the yeast enzyme (Ki >1000 μΜ). Surprisingly, calystegine 62 was also a good inhibitor for α-galactosidase from coffee beans (Ki 0.86 μΜ) and Asp. niger (Ki 2.3 μΜ) and for yff-galactosidase from rat intestine (Ki « 3 μΜ). The inhibition for /?-galactosidase from bovine liver can be ascribed to the lack of discrimination between o-galactosides and o-glucosides by this enzyme. Other calystegines isolated from Convolvulaceae and Solanaceae have the same positioning of hydroxyl groups as 62 but differ from it by epimerization at C-2 (B3), deoxygenation of C-2 (A5) or C-4 (A3) or by hydroxylation of C-6 (Bi and Ci). These structural alterations have only moderate effects on the inhibitory potency except those at C-2 where epimerization greatly reduces it and deoxygenation abolishes it completely. As epimerization of C-2 would change the D-g/wc0-resemblance of B2 to Dgalacto if it is considered an analog of 1, its effect on the inhibition of galactosidases by B3 (K\ > 1000 μΜ) would rule out this analogy. The lack of kinetic data with 2- and 4-deoxy-/?-glucosides and -galactosides with the enzymes employed for comparison with the calystegines does not permit a decision whether they bind in an orientation resembling that of isofagomine or differently. The model proposed by Asano et al. [102b] for calystegines BI and Ci bound to /?-glucosidase features a hydrogen bond of the catalytic acid with the hydroxyl group on C-6 but does not account for electrostatic or hydrogen bond interactions with the amino group and it also fails to explain the inhibition shown by B2 which is only about 2-fold weaker than with its C-6-hydroxylated analogs. The hitherto reported data show that enzymes having the same glycon specificity have a much wider spread in their susceptibility to inhibition by rigid bicyclic azasugar derivatives than by monocyclic ones. This may partly reflect steric effects of the glycon site with the bridging structure but also small differences in transition state topology which are, in cases with strong inhibition, well matched by the rigid inhibitors.

3.4.5

Interactions with the aglycon binding site

As shown in Schemes 3 -2a and 3 - 9, the aglycon R is part of the transition state and its interactions with the corresponding part of the active site make a substantial contribution to catalysis. This is most pronounced with substrates having an aglycon with poor leaving group propensity, e.g. oligo- and polysaccharides and glycosides with aliphatic alcohols. It is with these glycosides where a high aglycon specificity is observed, presumably, because the optimal orientation of the catalytic

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

57

groups with respect to the cleavage site requires a close interplay of numerous weak interactions between enzyme and substrate. Glycosides with good leaving groups are not that demanding in this respect. In addition to the enzymes involved in the degradation of starch and cellulose, the following examples illustrate a large contribution of aglycon interactions to the catalytic activity: /?-glucosidase A^ from Asp. wentii cleaves cellobiose ^ 10-fold better (based on &Cat/^m) than 4nitrophenyl /?-D-glucoside [9]; mammalian lysosomal /?-glucosidase has a « 60-fold higher activity with its natural substrate glucocerebroside than with 4-methylumbelliferyl /?-D-glucoside [103]; α-mannosidases IA and IB involved in the biosynthesis of N-linked glycoproteins are specific for a-l,2-oligomannosides of the glycoprotein precursors but they are inactive against 4-nitrophenyl a-D-mannoside. On the other hand, α-mannosidase II of the same pathway cleaves a-1,3and a-l,6-oligomannosides but is also active against the synthetic substrate [104]. In contrast with the large number of inhibitors designed for interactions with the glycon binding site and the catalytic groups much fewer studies have been published which address interactions with the aglycon site as well. As the aglycon moieties of most natural substrates require rather lengthy syntheses or are unknown, simple alkyl derivatives of D-glucosylamine and o-galactosylamine were the first inhibitors prepared for studies of aglycon site interactions. The results revealed unexpectedly large enhancements of the inhibitory potency, e.g. «4000-fold with N-benzyl D-glucosylamine and /?-glucosidase from almonds [45]; 12000fold with N-heptyl o-galactosylamine and /?-galactosidase from E. coli [49]; and «406 -fold with N-dodecyl D-glucosylamine and human lysosomal /?-glucosidase [105]. Whereas the latter result confirms an extended hydrophobic cleft to accommodate the sphingosine and fatty acid chains of glucocerebroside the data for βgalactosidase from E. coli were unexpected because this enzyme has lactose as its natural substrate (see [4b] for a detailed discussion). N-Substituents on 1-deoxynojirimycin-type of inhibitors cause with few exceptions, much smaller enhancements of inhibitory potency than with glycosylamines or are even detrimental (Tab. 3-1). This is understandable because the position of the substituent differs from that of the aglycon. The inhibition will be impared if there are close interactions of the endocyclic -NH- group with the glycon site. With enzymes having an extended hydrophobic aglycon site, e.g. mammalian cytosolic /?-glucosidase//?-galactosidase [75], a long aliphatic chain as in N-dodecyl-1 may overcome steric repulsions near the glycon site by folding back to distant parts of the aglycon site (Table 3-1). The inhibition by azasugars of the isofagomine type (7) could be expected not to be impaired by N-substituents because their attachment site is the same as that of the aglycon. Studies with the pseudo-disaccharide analog 46 [3Ic] showed the inhibition of yeast α-glucosidase not to be affected but almond /?-glucosidase and yeast isomaltase were 20- and 13-fold less strongly inhibited than by 7 itself. Note, that with both enzymes the aglycon substituent is a good mimic of the natural substrates, which are the 1,6-/?- and 1,6-a-linked glucose units of gentiobiose and isomaltose, respectively. Glucoamylase, on the other hand, an inverting exo-a-glucosidase specific for 1,4-a- and 1,6-α-linked glucose units, was inhibited 60-fold better by 46 than by 7.

58

3 Glycosidase Inhibition by Basic Sugar Analogs...

The reasons for the failure of the N-substituent of 46 to contribute to the inhibition of 'retaining' a- and /?-glucosidases is still subject to speculation. The preferred conformation of 46 is probably the one shown in the formula because treatment with H2U2 gave exclusively the N-oxide 11 with the axial oxygen and methylation with CHsI gave mainly (75 %) the quaternary ammonium salt having an axial methyl group. The '/?-like' conformation of 46 could provide an explanation for the results with α-glucosidase and isomaltase if their active sites are complementary to an early (i.e. substrate-like) transition state but it fails to account for the results with /?-glucosidase from almonds. Another puzzling result with the latter enzyme was obtained with the N-oxide 11 which inhibited 8-iold-better than 46 even though its axial oxygen would be repelled by the catalytic carboxylate on the α-face as depicted in Scheme 3-2a and which is supposed to be the target for the inactivation of this enzyme by conduritol B-epoxide and its 6-bromo derivative [4a]. However, the correlation of inhibition data is difficult to assess because the first bond-breaking step with the aryl glucosides employed for testing is not rate limiting [12]. Substituents for aglycon site interactions have also been linked to inhibitors having a planar geometry at C-I, for example the non- or weakly basic gluconhydroximolactone 14 and -lactam 15 by conversion to their N-phenylcarbamate derivatives 47 [54, 55] and 48 [55] and by synthesis of the stongly basic N^benzyl-omannonamidine 49 (not shown) and of N^alkyl-o-gluconamidines 50 [77]. The N-phenylcarbamoyl group contributed strongly («100 -fold) to the inhibition of α-glucosidase (yeast) and /?-glucosidase (almonds) and from Agrobacteriumfaecalis (20-fold) but only weakly when linked to the lactam [55]. A possible explanation for this difference could be that the non-basic 14 is a much weaker inhibitor (K{ in the mM range) than the moderately basic 15 (pKa 4.8; K1 in the μΜ range). Reduction of the basicity of 15 by conversion to the phenylcarbamate could largely offset the enhancing effect of aglycon site interactions seen with 47. Marked differences were seen in the response of N-acetylglucosaminidases to the N-phenylcarbamate substitent of the N-acetylglucosamine analog of 14 [57]. Whereas the inhibition of the enzyme from bovine kidney and jack beans were 4- and 6-fold enhanced by possible aglycon site interactions, that of the enzyme from the fungus Mucor rouxii was 500-fold larger than by the parent compound (^i (N-phenylcarbamoyl derivative) 0.04 μΜ vs. K1 (oximolactone) 21.5 μΜ). The authors ascribed this to a more hydrophobic aglycon site of the fungal enzyme.

OHV

Scheme 3-16.

O^ ΓΛ

II

41

X =O

48

X = NH

I

OH

50

R = C4H7 R = C12H25

3.4 Basic Sugar Analogs as Glycosidase Inhibitors

59

Interactions with the aglycon site by N^substituents of amidines turned out to be negligable with jack bean α-mannosidase and 49 which had the same K1 as the D-manntf-analog of 20 (K-, 0.20 μΜ [68b] vs. K10.17 μΜ [59]). The authors explain the lack of aglycon site interactions to an orientation of the benzyl substituent which differs significantly from that of the phenyl group in the (calculated) transition state of phenyl α-D-mannoside hydrolysis. It could be argued that the enzyme has an aglycon site complementary to hydrophilic oligomannoside substrates, but Ό-manno-amidine having 1-O-methyl 6-deoxy-a-D-mannose linked to N1 had K1 2.6 μΜ. In contrast with 49 the latter derivative was Z-configured rather than E at the (partial) exocyclic C-N1 double bond [69b]. An interesting point with the ^-substituted mannoamidines was their inhibition of glycosidases with different glycon specificity, e.g. /?-glucosidase from almonds which had K1 (49) 5.0 μΜ whereas D-g/wco-amidine 19 had K1 10 μΜ [59]. On the other hand, no inhibition (^i > 1000 μΜ) was observed with mammalian cytosolic /?-glucosidase//?-galactosidase (referred to as /?-galactosidase from bovine kidney in [68b]). A poor glycon specificity has also been described [59] for the O-glucoand D-manno-configured amidines and amidrazones acting on /J-mannosidase (jack bean). This aspect will be discussed with respect to the catalytic specificity in Section 3.5. A marked contribution to the inhibition by the butyl substituent in 50 was seen only with bovine cytospolic /?-glucosidase//?-galactosidase which was inhibited 130-fold better than by amidrazone 20 used as the reference. No enhancement of the inhibition was seen with /?-glucosidases from Asp. wentii and sweet almonds [67]. The dodecyl substituent in 50, on the other hand, gave a 105-fold enhancement with the cytosolic and lysosomal /?-glucosidases which both have an extended hydrophobic aglycon binding site. Contrary to expectations about contributions of the sp2-geometry on positive charge at C-I, the inhibition was not much better than by the corresponding N-alky 1 D-glucosylamines. The failure of the substitutents attached to isofagomine and mannoamidine to contribute to the inhibition is presumably due to the requirement of a close fit within the aglycon site which cannot be matched by the conformations acessible to their somewhat rigid structures. The alkyl chains linked to o-glycosyl- and Dgalactosylamine appear to be of sufficient flexibility to make even non-specific interactions with /?-glucosidase and -galactosidase favorable. The large contribution of the benzyl substituent to the inhibition of almond /?-glucosidase (Tab. 3-1) can be assigned to an adaptation of the aglycon site to the natural substrate amygdalin, /?-gentiobiosyl Γ-cyanobenzylalcohol.

3.4.6

Hydrogen bond formation with the catalytic acid

It is now generally agreed upon that a half-chair conformation is of prime importance for strong inhibition due to the sp2-hybridization of the anomeric carbon and the capacity for ionic interactions with the carboxyl groups of the active site, even though these two features and the hydroxyl group topography have been emphasized differently [32, 59]. Contributions to the binding energy by these fac-

60

3 Glycosidase Inhibition by Basic Sugar Analogs...

tors alone, however, do not adequately account for the relatively weak inhibition by strongly basic, Le. fully protonated and by permanently cationic sugar derivatives (see Section 3.4.2). Even when strong basicity is combined with sp2-hybridization, the inhibition is not much better, or as for α-galactosidase and a-L-fucosidase, even weaker than by moderately basic reference compounds (Table 3-4). This also holds for the N1-alky 1 D-gluconamidines 50 which had ^-values similar with those of the corresponding N-alkyl o-glucosylamines [67]. As shown in Schemes 3-2a and 3-3b, the transition state includes the transfer of a proton to the glycosidic oxygen by the catalytic acid -AH. With the enzyme-inhibitor complex this could result in protonation of the inhibitor and formation of an ion pair or, depending of the basicity of the substituent on C-I in relation to the acidity of -AH, formation of a hydrogen bond. As the catalytic center is strongly shielded from the aqueous environment [4a] both possibilities can make a substantial contribution to the binding energy. In order to act as hydrogen bond acceptor the non-bonding electron pair of the inhibitor must be within reach of -AH which includes distance (< 3.0 A) as well as a nearly co-linear orientation between the AH-donor bond and the doubly occupied acceptor orbitals. Inspection of Table 3-5 shows a similar or better inhibition by the weakly basic type 4 azasugars with sp3-configurated C-I than by the sp2configurated amidrazones used as reference which are almost fully protonated Table 3-4. Glycosidase inhibition by sugar analogs of different C-I geometry and basicity (^-values in μΜ, p^a-values are for the g/wc6>-configured compounds, data without reference are from [67]. Enzymes

Hexono-1,5lactone P^a <0

/?-Glucosidases sweet almonds, pH 5.0 Asp. wentii, pH 5.0 bovine, lysosom. pH 5.0 bovine, cytosol. pH 7.0

a

5.0

6.5 1.8

a-Galactosidase coffee bean, pH 6.0 a-Mannosidase jack bean a-L-Fucosidase bovine, pH 6.0 a

b c

Typel aza-sugar ptfa 6.3

48 0.30b 150 210 [75] 0.0016b [39]

120 [107]

68 [60] 0.0027 [41]

Type 4 aza-sugar p^a 5.3 b

1.3 0.3C [66] 0.07b [9] 0.8b 42

Hexonoamidrazone P^a 8.4

b

4.7 8.4 [59] 0.031b 3.3 19b

0.0007b [39]

8.3 [59]

1.6b [60]

0.17 [59] 0.82 [108]

Dale et al. [44d] have probably studied a different isoenzyme because they report ^-values for «//^-glucose, 1, and D-glucono-l,5-lactone which were from 4- to 40-times larger than found in [67]. Slow approach to the inhibition equilibrium Ki of g/wco-imidazol 18, estimated from IC5o [66], pK& (18) approx. 5 to 6.

3.5 Criteria for Transition State Resemblance

61

under the assay conditions. A possible explanation would be the formation of a strong hydrogen bond of -AH with the exocyclic substituent on C-I of the type 4 inhibitors. This would also hold for the weaker inhibition of /?-glucosidases by Dglucono-l,5-lactone which is expected to be a less efficient acceptor. Support for this hypothesis comes from the data for the type 1 azasugars which inhibit ^-specific enzymes up to 300-fold weaker than their type 4 counterparts because they lack the substituent on C-I (see [4a] and Table 3-1). Important information on the orientation of -AH with respect to the inhibitor and the contribution of the hydrogen bond to the binding energy was provided by Vasella's group [65] from studies on the inhibition of four /^-specific enzymes by nojiritetrazol 16, nojiritriazol 17, and their D-manno-isomers (see Section 3.4.3). In contrast with the freely rotating C-I substituent of the type 4 azasugars and of glycosylamines, the potential hydrogen bond acceptors are fixed 'in-plane' with the fused sugar ring. The results of this study were in contradiction with the paradigmatic lysozyme model (Scheme 3-2a) where proton transfer to the glycosidic oxygen is perpendicular to the pyranose ring. The ^-values for 16 with /?-glucosidases from almonds and C. thermocellum as well as cytosolic /?-glucosidase//?galactosidase from bovine kidney ranged from 1.5 to 40 μΜ. For /?-mannosidase from snail, K1 was 160 μΜ. The triazol 17 was non-inhibitory up to 8000 μΜ, equivalent to a K1 larger than 15 000 μΜ. As shown in Scheme 3-9, the catalytic acid -AH must be located 'in-plane' with the pyranose ring for efficient hydrogen bond formation. Its contribution to the binding energy may amount to > 5.4 kcal/ mol as estimated from K1 values of 16 and 17: This is in the upper range for hydrogen bonds of this type. Details of charge distribution at the active site, solvent access, and other features of enzymes susceptible to strong inhibition by cationic sugar derivatives (for example /?-glucsidase AS from Asp. wentii [9] and mammalian lysosomes [107], α-glucosidase I from the endoplasmatic reticulum [106], and bovine N-acetyl-/?glucosaminidase [38]) which could distinguish these enzymes from the majority of other glycosidases are still unknown.

3.5

Criteria for Transition State Resemblance

Inhibitory strength. The strength of an inhibitor expressed by its binding constant UK1 in relation to substrate affinity is a prime criterion to judge transition state resemblance. However, a meaningful evaluation and interpretation often meets with considerable difficulties. A comparison of KJKi with kCaJkunCat according to Wolfenden [8] is not possible for many enzymes because kuncat is too slow to be measured or kcat is determined by the deglycosylation step or a conformational change of the enzyme. Only a lower limit can be given in these cases for ^cat/^uncat·

As ^-values < 10~9 M are quite rare and values for K8 range from 10~2 to 10~5 M, KJKi ^ 107 for the most favorable cases which is, by orders of magnitude,

62

3 Glycosidase Inhibition by Basic Sugar Analogs...

only half-way to the rate acceleration factor kcat/kuncai. The reasons for this shortcoming are manifold. The models of the transition state in Schemes 3-2a and 3-3b can be considered only rough approximations because details like bond lengths, bond angles, and orientation of the hydroxyl groups and the leaving aglycon are not precisely specified. As transition state stabilization requires many weak interactions with complementary groups of the active site, a perfect match will be difficult to achieve with a stable compound. Additional problems arise from the great variability with respect to active site structure as demonstrated by the wide range of ^-values for a specific type of inhibitor acting on different enzymes with the same glycon specificity (see [4] and Tables in this chapter). The fulfillment of the following criteria, or rather its lack, may point to specific reasons why it is still a long way to the 'perfect' transition state mimic. Specificity. The catalytic specificity of glycosidases expressed by kcat/Km (cf. Section 3.4.4) depends, in addition to the stabilization of the transition state by the catalytic machinery, on specific interactions with the sugar hydroxyl groups and, to a lesser extent, on the structural and electronic features of the aglycon. Transition state resemblance of a given type of inhibitor should thus show up in the extent to which variations of glycon structure are reflected in its inhibitory strength with enzymes acting on substrates having the same aglycon linked to different sugars. Thus, 1/K1 should depend in the same way on glycon structure as kcai/ Km. A systematic study of this aspect has been carried out by Ermert et αϊ. [63] with Ό-gluco- and D-manno-configurated nojiritetrazoles 16 which fulfilled this specificity criterion quite well (see Section 3.4.3). Kajimoto et αϊ. [32] have distinguished ground state from transition state binding by the lesser importance of hydroxyl group topology for the high affinity of inhibitors with half-chair conformation and a positive charge at C-I than for fullchair azasugars and substrates. They were led to this conclusion by the 'broad specificity' of D-glucoaminidine 19 and -amidrazone 20 which inhibited /?-glucosidase from almonds and bovine cytosolic /?-glucosidase//?-galactosidase as well as jack bean α-mannosidase with K1 « 10 μΜ [58]. The D-manno-analog of 20, on the other hand was more specific in that it had K1200 μΜ, and 0.17 μΜ with the /?-glucosidases and α-mannosidase, respectively [59]. The discrimination factor, however, displayed by the almond enzyme for the OH-group on C-2 is larger than 3000 based on kcat/Km for 4-nitrophenyl /?-D-gluco- and -mannosides [44d]. There is, thus, no reason to correlate broad specificity with transition state resemblance as proposed in [32]. 4 Inhibition studies with azasugars in Ci conformation where hydroxyl groups have been replaced by hydrogen or fluorine, have revealed marked differences in the effect of the replacement on inhibitory potency compared with its effect on substrate hydrolysis. The N-methyl-6-deoxy derivative of 1, for example, is a 3-fold weaker inhibitor for almond /?-glucosidase than the N-methyl derivative, whereas deoxygenation at C-6 caused 3-fold better hydrolysis of phenyl /?-D-glucoside [109] and a 50-fold better hydrolysis of the corresponding 4-nitrophenyl /?-Dgalactosides by this enzyme [44d]. Deoxygenation of C-2 gives fagomine and results in a 280-fold reduction of l/K{ [32]. This is about one-fourth of the effect of deoxygenation of C-2 on kcat/Km [UO]. Note that the lower inhibition by fago-

3.5 Criteria for Transition State Resemblance

63

mine is partly due to its greater basicity which results in a larger proportion of the noninhibitory cation. The effects of the replacement of -OH by -F at C-2, C-3 and C- 6 of 1 on its inhibition of six a- and /?-glucosidases were studied by Andersen et al. [87]. They were smallest for the C- 6 and largest for the C-2 replacement but were still much smaller than those which the same structural alteration had on substrate hydrolysis, where such comparisons could be made. As with 2-deoxygenation the replacement of OH by -F had no or only a small effect on substrate affinity, fccat was mainly 8 affected and was reduced up to 10 -fold [78]. Dependence of UKi onpH. The presumed catalytic mechanism involves proton transfer from the catalytic acid and stabilization of the (partial) positive charge on the anomeric carbon by a carboxylate. The pH-dependence of kcat/Km is thus governed by the p^Ta-values of these two groups in the free enzyme and the enzyme transition state complex, respectively. The pH-dependence of 1 IK\ of a transition state mimic should, therefore, resemble that of kcai/Km. Such studies have been made and the effect of pH on the concentration of the inhibiting species (protonated inhibitor or free base) has been factored out; no, or only a poor, correlation between l/K{ and kcat/Km could be seen (see [4b] for a more detailed discussion). An enzyme inhibitor system where the criterion of pH-dependence of UK1 seems to be fulfilled was described by Axamawaty et al. [50] who studied the inhibition of a fungal α-L-arabinofuranosidase by l,4-dideoxy-l,4-imino-L-arabinitol. The inhibition could be described by the interaction of the catalytic acid (pKa 5.9) with the unprotonated form of the inhibitor (pKa 7.58). A consequence of this relation between the acid AH and the basic form of the inhibitor is an unfavorable ionization state even at the maximum of the inhibition (Ki 1.3 μΜ at pH 6.85). Optimal ionization of both groups would have resulted in K1 0.02 μΜ. Slow inhibition. Some glycosidase inhibitors with K1 in the lower μΜ range and most of those with K{< 10 nM show a slow approach to the inhibition equilibrium [4a]. The rate constants vary from 10~4 M-1S"1 to < 10~3 M1 s"1, i.e. well below the diffusion-controlled limit. As measurements are usually made with inhibitor concentrations I < 10 K[, rate determinations do not require rapid reaction techniques. A possible model for this widespread phenomenon (see [111] for a detailed discussion) comprises a conformation of the enzyme which is complementary with the ground state of the substrate and which can rapidly adopt a conformation complementary to the transition state after formation of the initial enzyme-substrate complex. Transition state-like inhibitors bind slowly because only a minute fraction of the enzyme is present in the latter conformation, thus greatly reducing successful encounters. As there are no clear-cut correlations between ^i and the rate of inhibitor binding and slow binding is also observed with azasugar derivatives like swainsonine and castanosperine [4a] which have only a moderate similarity with the current models of the transition state, it is not clear to what extent the slow binding of strong inhibitors can be used as criterion for transition state resemblance.

64

3 Glycosidase Inhibition by Basic Sugar Analogs...

References [1] H. Eyring, J. Chem. Phys. 1935, 3, 107-110; M. G. Evans, M. Polany, Transact. Faraday Soc. 1935, 875-880. [2] a) D. C. Phillips, ScL American 1966, 275(5), 78-90; b) C. C. F. Blake, G. A. Mair, A. T. C. North, D. C. Phillips, V. R. Sharma, Proc. Royal Acad. Sd. (London), Ser. B. 1967, 167, 378-387. [3] M. L. Sinott, Chem. Rev. 1990, 90, 1171-1202. [4] a) G. Legler, Adv. Carbohydr. Chem. Biochem. 1990, 48, 319-384; b) G. Legler in Carbohydrate Mimics (Ed.: Y. Chapleur), VCH, Weinheim, 1997, Chapter 24. [5] V. L. Schramm, B. A. Horenstein, I. C. Klein, /. Biol Chem. 1994, 269, 18269-19262. [6] A. Warshel, Proc. Natl Acad. ScL USA 1978, 75, 5250-5256, Biochemistry, 1981, 20, 3167-4177; Proc. Natl. Acad. ScL USA 1984, 87, 444-448. [7] L. Pauling, Chem. Eng. News 1946, 24, 1375-1378; Nature (London) 1948, 767, 707-710. [8] A. Wolfenden, Ace. Chem. Res. 1972, 5, 10-18. [9] G. Legler, Forschungsber. Nordrhein-Westfalen, Nr. 2846, Westdeutscher Verlag, OpIaden, FRG, 1979; G. Legler, M. L. Sinott, S. G. Withers, /. Chem. Soc., Perkin Trans. II 1980, 1376-1382. [10] M. Eigen, G. Hammes, Adv. Enzymol. 1963, 25, 1-38. [11] D. E. Koshland, Biol Rev. 1953, 28, 416-426. [12] R. L. Nath, H. N. Rydon, Biochem. J. 1954, 1-10. [13] M. L. Sinnott, I. J. L. Souchard, Biochem. J. 1973, 133, 89-98. [14] E. J. Hehre, C. F. Brewer, D. S. Genghof, J. Biol. Chem. 1979, 254, 5942-6954. [15] M. Matsui, J. S. Blanchard, C. F. Brewer, E. J. Hehre, /. Biol. Chem. 1989, 264, 87148716. [16] B. Svensson, M. Sogaard, J. Biotechnol. 1993, 29, 1-37. [17] D. Cocker, M. L. Sinnott, /. Chem. Soc., Perkin Trans. /71975, 1391-1395. [18] M. L. Sinnott, S. G. Withers, O. Viratelle, Biochem. J. 1978, 775, 539-546. [19] a) L. B. Daniels, R J. Coyle,Y.-B. Chiao, R. B. Glew, J. Biol. Chem. 1981, 256, 1300413013; b) V. Gopalan, D. S. van der Jagt, D. Libell, R. H. Glew, J. Biol. Chem. 1992, 267, 9629-9638. [20] G. E. Lienhard, Science 1973, 180, 149-154. [21] H. Hiromi, A. Tanaka, M. Ohnishi, Biochemistry 1982, 27, 102-107. [22] B. Svensson, M. R. Sierks, Carbohydr. Res. 1992, 227, 29-44. [23] M. R. Sierks, C. Ford, P. J. Reilly, B. Svensson, Protein Eng. 1989, 2, 621-626. [24] E. Truscheit, W. Frommer, B. Junge, L. Muller, D. Schmidt, Angew. Chem. 1981, 93, 738-755; Angew. Chem., Int. Ed. Engl. 1981, 20, 744-761. [25] H. Ishida, K. Kumagai, T. Tsuruoka, H. Yumoto, J. Antibiotics, Ser. A 1967, 20, 60-70. [26] S. Inouye, T. Tsuruoka, T. Ito, T. Niida, Tetrahedron 1968, 23, 2125-2144. [27] H.-Y. L. Lai, B. Axelrod, Biochem. Biophys. Res. Commun. 1973, 54, 463-468. [28] E. T. Reese, F. W. Parrish, M. Ettlinger, Carbohydr. Res. 1971, 18, 381-388. [29] E. Kappes, G. Legler, /. Carbohydr. Chem. 1989, 8, 371-388. [30] G. Legler, R. Bollhagen, Carbohydr. Res. 1992, 233, 113-123. [31] a) T. M. Jespersen, W. Dong, T. Skrydstrup, I. Lundt, M. BoIs, Angew. Chem., Int. Ed. Engl. 1994, 37, 1778-1779; b) T. M. Jespersen, M. BoIs, M. R. Sierks, T. Skrydstrup, Tetrahedron 1994, 50, 13449-13460; c) W. Dong, T. Jespersen, M. BoIs, T. Skrydstrup, M. R. Sierks, Biochemistry 1996, 35, 2788-2795. [32] T. Kajimoto, K. K.-C. Liu, R. L. Pederson, Z. Zhong, Y. Ichikawa, J. A. Porco, Jr., C.-H. Wong, /. Am. Chem. Soc. 1991, 113, 6191-6202. [33] Y. Ichikawa, Y. Igarashi, Tetrahedron Lett. 1995, 36, 4585-4586. [34] Y. Igarashi, M. Ichikawa, Y. Ishikawa, Tetrahedron Lett. 1996, 37, 2707-2708. [35] Y. Igarashi, M. Ichikawa, Y. Ishikawa, Bioorg. Med. Chem. Lett. 1996, 6, 553-558. [36] A. Hansen, T. M. Tagmose, M. BoIs, Tetrahedron 1997, 53, 697-706.

References

65

[371 Y. Nishimura, T, Satoh, T. Kudo, S. Kondo, T. Takeushi, Bioorg. Med. Chem. 1996, 4, 91-96. [38] G. Legler, E. Liillau, E. Kappes, F. Kastenholz, Biochim. Biophys. Acta 1991, 1080, 89-95. [39] G. Legler, S. Pohl, Carbohydr. Res. 1986, 755, 119-129. [40] I. Cenci di Bello, P. Dorling, L. Fellows, B. Winchester, FEBS-Lett. 1984, 776, 61-64. [41] G. Legler, A. E. Stiitz, H. Immich, Carbohydr. Res. 1995, 272, 17-30. [42] B. Winchester, C. Barker, S. Baines, G. Jacob, S. K. Nagoong, G. W. J. Fleet, Biochem. J. 1990, 256, 277-285. [43] J. N. BeMiller, R. J. Gilson, R. W. Myers, M. M. Satoro, B. Yadav, Carbohydr. Res. 1993, 250, 93-100. [44] a) A. K. Grover, R. J. Cushley, Biochim. Biophys. Acta 1977, 482, 109-124; b) D. E. Walker, B. Axelrod, Arch. Biochem. Biophys. 1978, 187, 102-107; c) L. Kiss, L. K. Berki, P. Nanasi, Biochem. Biophys. Res. Commun. 1981, 98, 792-799; d) M. P. Dale, H. E. Enyley, K. Kern, K. A. R. Sastry, L. D. Byers, Biochemistry 1985, 24, 3530-3539. [45] G. Legler, Biochim. Biophys. Acta 1978, 524, 94-101. [46] G. Legler, H. Liedtke, Biol. Chem. Hoppe-Seyler 1985, 366, 1113-1122. [47] T. D. Stuart, S. Maeser, /. Am. Chem. Soc. 1924, 46, 2583-2590. [48] R. Saul, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys. 1984, 230, 668-673. [49] G. Legler, M. Herrchen, Carbohydr. Res. 1983, 776, 95-103. [50] M. T. H. Axamawaty, G. W. J. Fleet, K. A. Hannah, S. G. Namgoong, M. L. Sinnott, Biochem. J. 1990, 266, 245-249. [51] G. Legler, Biochim. Biophys. Acta 1977, 529, 94-101. [52] D. H. Leaback, Biochem. Biophys. Res. Commun. 1968, 32, 1025-1023. [53] M. Brockmann, J. Lehmann, Carbohydr. Res. 1977, 53, 21-31. [54] D. Beer, A. Vasella, HeIv. Chim. Acta 1986, 69, 267-270. [55] R. Hoos, A. Vasella, K. Rupitz, S. G. Withers, Carbohydr. Res. 1977, 298, 291-298. [56] G. Legler, F. Kastenholz, unpublished data. [57] M. Horsch, L. Hoesch, A. Vasella, D. M. Rast, Eur. J. Biochem. 1991, 797, 815-818. [58] M. K. Tong, G. Papandreou, B. Ganem, /. Am. Chem. Soc. 1990, 772, 6137-6139. [59] G. Papandreou, M. K. Tong, B. Ganem, /. Am. Chem. Soc. 1993, 775, 11682-11690. [60] G. Legler, E. Julich, Carbohydr. Res. 1984, 728, 61-72. [61] R. Hoos, A. B. Naughton, W. Thiel, A. Vasella, W. Weber, K. Rupitz, S. G.Withers, HeIv. Chim. Acta 1993, 76, 2666-2686. [62] P. Ermert, A. Vasella, HeIv. Chim. Acta 1991, 74, 2043-2053. [63] P. Ermert, A. Vasella, M. Weber, K. Rupitz, S. G. Withers, Carbohydr. Res. 1993, 250, 113-128. [64] T. D. Heightman, P. Ermert, D. Klein, A. Vasella, HeIv. Chim. Acta 1996, 79, 514-519. [65] T. D. Heightman, M. Locatelli, A. Vasella, HeIv. Chim. Acta 1996, 79, 2190-2200. [66] K. Tatsuta, S. Miura, S. Ohtam, H. Gunji, /. Antibiot. 1995, 48, 286-289. [67] G. Legler, M.-T. Finken, Carbohydr. Res. 1996, 292, 103-115. [68] a) Y. Bleriot, A. Genre-Grandpierre, C. Tellier, Tetrahedron Lett. 1994, 35, 1867-1870; b) Y. Bleriot, T. Dintinger, A. Genre-Grandpierre, M. Padrines, C. Tellier, Bioorg. Med. Chem. Lett. 1995, 5, 2655-2660. [69] a) Y. Bleriot, T. Dintinger, N. Guillo, C. Tellier, Tetrahedron Lett. 1995, 36, 5175-5178; b) Y. Bleriot, A. Genre-Grandpierre, A. Imberty, C. Tellier, J. Carbohydr. Chem. 1996, 75, 985-1000. [70] a) J. Lehmann, B. Rob, Liebigs Ann. Chem. 1994, 805-809; b) J. Lehmann, B. Rob, H.-A. Wagenknecht, Carbohydr. Res. 1995, 278, 167-180. [71] J.-H. Jeong, S. W. Murray, S. Takayama, C.-H. Wong, J. Am. Chem. Soc. 1996, 118, 4227-4234. [72] K. Wallenfels, R. Weil in The Enzymes, 3rd ed., 1972, (Ed.: P. D. Boyer), Vol. VII, 618-665. [73] K.-R. Roeser, G. Legler, Biochim. Biophys. Acta 1981, 657, 321-333. [74] H.-J. Kytzia, K. Sandhoff, J. Biol. Chem. 1985, 260, 7568-7575.

66

3 Glycosidase Inhibition by Basic Sugar Analogs...

[75] G. Legler, E. Bieberich, Arch. Biochem. Biophys. 1988, 260, 427-438. [76] M. Paez de Ia Cadena, J. Rodriguez-Berrocal, J. A. Cabezas, N. Perez-Gonzales, Biochimie 1986, 68, 251-260. [77] G. Legler, M.-T. Finken, S. Felsch, Carbohydr. Res. 1996, 292, 91-101. [78] N. M. Namchuk, S. G. Withers, Biochemistry 1995, 34, 16194-16202. [79] Y. Ie Merrer, L. Poitout, J.-C. Depezay, I. Dosbaa, S. Geoffrey, M. J. Fogliatti, Bioorg. Med. Chem. 1997, 5, 519-533. [80] A. Welter, G. Dardenne, M. Marlier, J. Casimir, Phytochemistry 1976, 25, 757-748. [81] G. W. J. Fleet, S. J. Nicholas, P. W. Smith, S. V. Evans, L. E. Fellows, R. J. Nash, Tetrahedron Lett. 1985, 26, 3127-3130. [82] X. Qian, F. Moris-Varas, M. C. Fitzgerald, C.-H. Wong, Bioorg. Med. Chem. 1996, 4, 2055-2069. [83] G. Legler, A. Korth, A. Berger, C. Ekhart, G. Gradnig, A. E. Stiitz, Carbohydr. Res. 1993, 250, 67-73. [84] L. H. Fotch, C.-H. Wong, Tetrahedron Lett. 1994, 35, 3481-3483. [85] G. W J. Fleet, P. W. Smith, S. V. Evans, L. E. Fellows, /. Chem. Soc., Chem. Commun. 1984, 1240-1241. [86] Y. Takaoka, I. Kajimoto, C.-H. Wong, J. Org. Chem. 1993, 58, 4809-4812. [87] S. M. Andersen, M. Ebner, C. W. Ekhart, G. Gradnig, G. Legler, I. Lundt, A. E. Stiitz, S. G. Withers, T. Wrodnigg, Carbohydr. Res. 1997, 301, 155-166. [88] a) J. E. Tropea, G. P. Kaushal, I. Pastuszak, M. Mitchell, T. Aoyagi, R. J. Molyneux, A. D. Elbein, Biochemistry 1990, 29, 10062-10069; b) Y.-T. Pan, G. P. Kaushal, /. Biol. Chem. 1992, 267, 8313-8318. [89] R. A. Farr, M. P. Peet, M. S. Kang, Tetrahedron Lett. 1990, 31, 7109-7112. [90] L. Provencher, D. H. Steensma, C.-H. Wong, Bioorg. Med. Chem. 1994, 2, 1179-1194. [91] P. Zhou, H. M. Salleh, P. C. M. Chan, G. Lajoie, J. E Honek, P. T. C. Nambiar, O. P. Ward, Carbohydr. Res. 1993, 239, 155-166. [92] S. Knapp, D. Vocadlo, Z. Gao, B. Kirk, J. Lou, S. G. Withers, /. Am. Chem. Soc. 1996, 118, 6804-6805. [93] P. A. Fowler, A. H. Raines, R. J. K. Taylor, E. J. T. Chrystal, M. B. Gravestock, /. Chem. Soc., Perkin Trans. 11994, 2229-2235. [94] M. I. Page, W. P. Jencks, Proc. Natl. Acad. Sd. USA 1971, 68, 1678-1683. [95] I. Cenci di Bello, G. W. J. Fleet, S. K. Namgoong, K.-I. Tadano, B. Winchester, Biochem. J. 1989, 259, 855-861. [96] G. Gradnig, A. Berger, V. Grassberger, A. E. Stutz, G. Legler, Tetrahedron Lett. 1991, 32, 4889-4892. [97] W. H. Pearson, E. J. Hembre, J. Org. Chem. 1996, 61, 5537-5548. [98] A. D. Elbein, T. Szumilo, B. A. Sanford, K. B. Sharpless, C. Adams, Biochemistry 1987,2(5,2602-2510. [99] R. Molyneux, Y. T. Pan, J. E. Tropea, M. Benson, G. P. Kaushal, A. D. Elbein, Biochemistry 1991, 30, 9981-9987. [100] R. Nash, L. E. Fellows, J. V. Dring, G. W. J. Fleet, A. Girdhar, N. G. Ramsden, J. M. Peach, M. P. Hagerty, A. M. Scofield, Phytochemistry 1990, 29, 111-116. [101] R. J. Molyneux, Y. T. Pan, A. Goldman, D. A. Tepfer, A. D. Elbein, Arch. Biochem. Biophys. 1991, 304, 81-88. [102] a) N. Asano, E. Tomioka, H. Kizu, K. Matsui, Carbohydr. Res. 1994, 253, 235-245; b) N. Asano, A. Kato, K. Oseki, H. Kizu, K. Matsui, Eur. J. Biochem. 1995, 229, 369-376. [103] G. A. Grabowski, K. Osiecki-Newman, T. Dinur, D. Fabbro, G. Legler, S. Gatt, R. J. Desnick, J. Biol. Chem., 1986, 261, 8263-8269. [104] D. R. P. Tulsiani, S. C. Hubbard, P. W Robbins, O. Touster, J. Biol. Chem. 1982, 257, 3660-3668. [105] P. Greenberg, A. H. Merrill, D. C. Liotta, G. A. Grabowski, Biochim. Biophys. Acta, 1990, 1039, 12-20. [106] H. Hettkamp, G. Legler, E. Bause, Eur. J. Biochem. 1984, 142, 85-90.

References [107] [108] [109] [110] [Ill]

67

S.-C. Li, Y.-T. Li, J. BioL Chem. 1970, 245, 5153-5160. D. J. Schedler, B. R. Bowens, B. Ganem, Tetrahedron Lett. 1994, 35, 3845-3848. B. Helferich, T. Kleinschmidt, Hoppe-Seyler's Z. physiol. Chem. 1968, 349, 25-27. G. Legler, Acta Microbiol Acad. ScL Hung. 1975, 22, 403-409. J. F. Morrison, C. T. Walsh, Adv. Enzymol. 1988, 67, 201-301.

4

Synthetic Methods for the Preparation of Iminosugars BARBARA LA FERLA and FRANCESCO NICOTRA

This chapter reviews many of the multifarious methods available for the synthesis of iminosugars, and attempts to classify these methods in different categories, depending on the starting materials employed and on the strategies adopted. The methods are divided into those employing carbohydrate as starting materials, and those starting from other substrates. For each method different procedures have been reported, and for each procedure one or two examples have been selected. The aim of this chapter is to give an insight into the various possible strategies that can be used to synthesize iminosugars, and not to describe exhaustively the field, extensively reporting similar approaches.

4.1 Introduction 'Sugar mimics' of natural origin, such as castanospermine, swainsonine or deoxynojirimicin (Figure 4-1), in which a nitrogen substitutes the ring oxygen, have shown interesting biologic and pharmacologic properties. They act as inhibitors of carbohydrate-processing enzymes, and this behavior results in a variety of potentially therapeutic activities such as anti-viral [1-4], immunomodulator [3], anti-tumoral [5], anti-diabetic [3] or anti-hyperglycemic [3]. These observations have stimulated the interest of many synthetic chemists towards this class of compounds, generally defined imino-sugars or aza-sugars. Multifarious iminosugars of natural and synthetic origin have been produced in the last decade, and some of them have shown promising pharmaceutic properties,

Figure 4.1. Some iminosugars of natural origin. Iminosugars ns GlycosUlase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

4.2 Iminosugars from 'True'Sugars

69

for example as anti-AIDS agents. A great variety of synthetic approaches have been used to assemble these molecules, ranging from chemical to enzymatic methods, and employing from sugars to benzene as starting materials. The aim of this chapter is to classify the different synthetic procedures in homogenous groups, in an effort to rationalize the numerous examples and the great variety of methods reported in the literature. The description of each reported method will not be exhaustive, rather the chapter will give a general overview of the different approaches. Due to the impossibility of reporting all the literature in the field, the choice of the examples for each synthetic method is subjective and arbitrary, and does not imply any priority.

4.2

Iminosugars from 'True' Sugars

The close structural relation between iminosugars and 'true' sugars has suggested to use largely available carbohydrates as starting material for the synthesis of their imino analogs. This is in fact the way through which nature effects the biosynthesis of natural occurring iminosugars. For example mannonojrimicin biosynthesis occurs starting from D-fructose in one of the possible ways reported in Scheme 4-1 [6]. In this scheme, the primary hydroxyl group of fructose is oxidized to aldehyde, the amino function is introduced by reductive amination of one of the two carbonyl groups, and the cyclization occurs by nucleophilic attack of the amine to the carbonyl group, with formation of a hemiaminal. The synthetic procedures exploited to convert 'true' sugars into iminosugars require in general the same steps shown in the biosynthetic approach: the introduction of an amino function and the subsequent cyclization. Often iminosugars lack the carbonyl function of the parent sugar, the hemiaminal function being reduced to afford a more stable polyhydroxylated pyrrolidine or piperidine (see the examples in Figure 4-1). As a consequence there is no difference, in terms of functional group, between the two carbon atoms linked to the (:κ οκ =ο 2

=ο

HO — —OH

— OH -OH

CH2OH

<:ΗΟ

CH2OH

CH2OH

H2N-

*.

HO-

H2N-

HO-

—OH

Scheme 4-1. Possible biosynthetic pathways to mannonojirimicin.

-OH CH2OH

CHO

70

4 Synthetic Methods for the Preparation of Iminosugars

OR I

H

ORj OH

\

OR

OR ^

deoxynojirimicin(R = H)

Scheme 4-2. General Scheme ίθΓ the

COn-

version of a sugar into an iminosugar

nitrogen of most iminosugars. Therefore in the synthesis of these molecules, the amino group can be introduced on either of the two carbon atoms, provided that the cyclization is then effected on the other. Starting from o-glucose, for example, an iminosugar can be obtained by introduction of the amino group at C-5, and subsequent cyclization by nucleophilic attack on C-I, or vice versa by introduction of the amino group at C-I, and subsequent cyclization to C-5 (Scheme 4-2). The cyclization of the animated open-chain intermediates shown in Scheme 4-2 requires the attack of the amino group on an electrophilic carbon, such as a carbonyl group (A, E, F in Scheme 4-3), a mesylate, tosylate or triflate (C, H; LG = leaving group) or a double bond (I) activated by an electrophile such as iodine or mercuric salts. One of the two steps required for the synthesis of iminosugars starting from true sugars, the introduction of the amino function, can be omitted if the target molecule is obtainable by elaboration of an available aminosugar.

Scheme 4-3. Different procedures to obtain an iminosugar (deoxynojirimicin) by cyclization of aminated open-chain intermediates.

4.2 Iminosugars from 'True'Sugars

4.2.1

71

Iminosugars from aminosugars

Iminosugars have been prepared taking advantage of the amino group of aminosugars. This approach is convenient when the structure of the target molecule is related to that of a commercially available or easily obtainable aminosugar. The amino substituent of the starting sugar is exploited in the cyclization, which in general requires a proper functionalization of the carbon atom in γ- or ^-position, to afford the cyclic amine. An interesting example of this approach is reported in Scheme 4-4, which describes the synthesis of 6, an imino analog of L-fucose, starting from D-galactosamine 1 [7]. In this example the amino group of o-galactosamine 1 effected the cyclization on the non-reducing end of the starting sugar, which was oxidized to carboxylic acid. The carbonyl group of D-galactosamine was, in contrast, reduced to a methyl group, which became the C-6 of 6. This compound is an imino analog of L-fucose, that showed inhibition towards human L-fucosidase. Another interesting example of synthesis of iminosugars starting from available aminosugars is shown in Scheme 4-5 [8]. In this case, the aminosugar involved in the aminocyclization is N-carbobenzyloxyneuraminic acid 8, which was prepared enzymatically by the NANA-aldolasecatalyzed reaction of N-carbobenzyloxy-o-mannose 7 and pyruvate. Catalytic hydrogenation of 8 directly afforded the pyrrolidine derivative 9 through deprotection of the amino group, which is in the correct position to effect a nucleophilic attack on the anomeric carbonyl group. The so-obtained five-membered cyclic hemiaminal (in brackets in Scheme 4-5), disfavored in the equilibrium, underwent dehydration to afford an intermediate imine, which was further hydrogenated. The conversion of the primary hydroxyl group of 9 into a tosylate, which required previous protection of the amino group and subsequent deprotection, finally allowed a further cyclization with formation of the second ring. Reduction of the carboxylic acid of the bicyclic structure 12 afforded 3-(hydroxymethyl)-6-epicastanospermine 13.

Scheme 4-4. (a) CbzCl; (b) BnOH, CH3COCl; (c) Pt, O2, H2O-EtOH; (d) 1:4 0 TFAiH2O, 9O C; (e) EtSH, HCl; (f) Raney-Ni EtOH; (g) TMSCl, (TMS)2NH; (h) Lawesson's Reagent, then CH3OH-HCl; (i) NH2NH2, CH3OH, O 0 C.

72

4 Synthetic Methods for the Preparation oflminosugars

-X\Ί NW

H(T

—

-X\^

IKT

13 12

Scheme 4-5.

4.2.2

Iminosugars via amination of 'true' sugars and subsequent cyclization

Following the general Scheme 4-2, the synthesis of iminosugars via animation of true sugars and subsequent cyclization can be classified as follows: (i) procedures that involve amination of the anomeric carbon, an easy process which takes advantage of the reactivity of the aldehydic function of the sugar; and (ii) procedures that require chain amination, so exploiting the reactivity of the anomeric carbon in the subsequent cyclization. 4.2.2.1

Iminosugars via amination of the anomeric center

The anomeric center of a sugar can be easily aminated by treatment with ammonia or a primary amine, to afford a glycosylamine (15 in Scheme 4-6), which is in equilibrium with the open-chain imine 16. Reduction of the imine 16, or reaction with an organometallic reagent, affords an aminoalditol 17, the aminocyclization of which generates an iminosugar 19 or 21. Scheme 4-6 describes different cyclization procedures. One possibility consists of the conversion of the free hydroxyl group, originally involved in the hemiacetalic linkage, into a leaving group (a in Scheme 4-6). The amino function effects a nucleophilic substitution on the carbon bearing the leaving group, with inversion of its configuration, affording an iminosugar. Another possibility (b in Scheme 4-6) consists of the formation of an epoxide, obtained with inversion of configuration at the carbon atom originally involved into the hemiacetalic linkage. The amino group effects the cyclization

4.2 Iminosugars from 'True'Sugars

73

Scheme 4-6. Different cyclization procedures of amino sugars obtained from true sugars by attack on the epoxide, regenerating the original stereochemistry. Another possible cyclization approach (c in Scheme 4-6) lies in the formation of a double bond by reductive elimination of an α-halohydrin. Activation of the double bond with electrophiles such as mercuric salts causes an aminocyclization. In this case the stereochemical outcome of the reaction depends on, which face of the double bond is preferentially attacked by the electrophile. An example of approach (a) (Scheme 4-6) is reported in Scheme 4-7. 2,3,5-TriO-benzyl-D-arabinose 22 reacted with benzylamine to afford the corresponding glycosylamine 23. Reaction of 23 with a Grignard reagent stereoselectively gave the acyclic aminoalcohol 24, the cyclization of which by treatment with triflic anhydride afforded the iminosugar 25 [9]. The possibility of effecting the cyclization by oxidation of the hydroxyl group originally involved in the hemiacetalic linkage, and subsequent reduction of the obtained cyclic hemiaminal, has also been investigated [1O]. However, in the example reported in Scheme 4-7, the use of pyridinium chlorochromate resulted in an oxidative degradation, affording the lactam 26 which was then reduced with diborane to the iminosugar 27 [U]. An example of approach (b) is reported in Scheme 4-8 [12]. Condensation of 2,3,4 tri-O-benzyl-o-glucopyranose 28 with benzylamine afforded the corresponding glucosylamine 29 which was then reduced and trifluoroacetylated giving the

Scheme 4-7.

74

4 Synthetic Methods for the Preparation of Iminosugars ΗΟ

-Λ Ό

BnNH2

«Ο-, BnO^V-A_—O

BnOA.^---^---^^ OBn

LLiAlH 4 2. (CF3CO)2O

«Ο ηΟ ~

Β

BnOj ΟΗ

28

^β^ΝΗΒη

29 1. IBuMe2SiCI, Imidazole 2. MsCl 3. Bu4NF-THF; MeONa, MeOH

32

amide 30. Selective protection of the primary hydroxyl group of 30, mesylation of the secondary hydroxyl group and subsequent deprotection of the primary, afforded the epoxide 31 with inversion of the configuration at C-5. The amino group was then regenerated by reduction of the trifluoroacetamido group of 31, to afford the aminoepoxide 32 which cyclized spontaneously to a mixture of piperidine 33 and azepane 34 in a ratio of 45:55. Approach (c) is exemplified in Scheme 4-9. Methyl 2,3,4-tri-0-benzyl-6bromo-6-deoxy-D-glucopyranose 35 was refluxed with zinc dust in n-propanolwater in the presence of benzylamine and NaBHsCN. The reaction afforded directly the unsaturated amine 36 by reductive elimination of the haloether and reductive animation of the obtained aldehydic intermediate. Treatment of 36 with mercuric trifluoroacetate finally gave a mixture of the two possible epimeric iminosugars 37 and 38 in ratio of 6:4 [13]. To introduce the amino group required for the formation of an imino sugar, the anomeric center has been aminated also by reaction with hydroxylamine. The stable open-chain oxime so obtained can be easily manipulated at the free hydroxyl group, for example to convert it into a leaving group, so allowing the formation of the cyclic structure. In the case reported in Scheme 4-10, reaction of 2,3,5-tri-Obenzyl-L-arabinose 39 with hydroxylamine afforded the oxime 40. The free hydroxyl group of 40 was then converted into a mesylate by treatment with methanesul-

Zn, PrOH, H2O BnNH2

35

38

Scheme 4-9.

37:38

=6:4

x=Br H H 0

^

4.2 Iminosugars from 'True'Sugars

NH2OHHCl J MeONa MeOH >- BnO— HO

HOB»

OBn 39

-OBn

40

75

CN

Py

Β Βηο-Γ° "

MsO—

I

1OBn

41

NH 42

Scheme 4-10.

phonyl chloride, a reaction that simultaneously converts the oxime into the corresponding nitrile 41. Finally, selective reduction of the nitrile of 41 afforded the iminosugar 42 [14] (Scheme 4-10). Glyconolactones have been also used in the synthesis of iminosugars. The lactone in fact can be easily converted into the corresponding amide by treatment with an amine, and the amide can be cyclized to afford a lactam, or eventually reduced to the corresponding amine. In the example shown in Scheme 4-11, 2,3,5-tri-O-benzyl-D-arabinolactone 43 was converted into the amide 44 by treatment with benzylamine. The amide was then easily transformed into the iminosugar 45 by conversion of the free hydroxyl group into a mesylate and by subsequent reduction with borane [15]. Alternatively, the open-chain hydroxyamide (47 in Scheme 4-12) obtained by reaction of the gluconolactone 46 with a primary amine, was oxidized to the corresponding ketone 48. The amidic hydrogen of 48 is nucleophilic enough to effect the addition to the carbonyl group, affording 49. Catalytic hydrogenation of 49 afforded 50, the reduction of which with LiAlH4 finally gave the iminosugar 51 [16].

76

4 Synthetic Methods for the Preparation of Iminosugars

4.2.2.2

Iminosugars via 'chain animation'

The synthesis of iminosugars from 'true' sugars can be effected by amination of the sugar substrate 52 in position γ or δ with respect to the carbonyl group. The amino group, once introduced, reacts with the carbonyl function affording a cyclic hemiaminal 54. Reduction of the hemiaminal directly affords the iminosugar 56, in a process which involves an intermediate cyclic imine 55, generated from the hemiaminal 54 by elimination of a molecule of water (Scheme 4-13). The amino group has been introduced in the sugar skeleton by conversion of a hydroxyl into a leaving group, substitution with an azide, and finally reduction. An elegant example of this approach is reported in Scheme 4-14 [17]. In this case the readily available 6,6'-dichloro-6,6'-dideoxysucrose 57 was converted into the diazide 58, the hydrolysis of which afforded 6-azido-6-deoxy-D-glucose 59 and 6-azido-6-deoxy-o-fructose 60. Catalytic hydrogenation of 60 afforded the iminosugar 61, whereas 59 was converted into 60 by treatment with glucose isomerase. It is noteworthy that this procedure does not require protecting groups. Alternatively, the amino group has been introduced by oxidation of a hydroxyl group and subsequent reductive amination, as reported in Scheme 4-15 [18]. In both cases, the reductive step, which is done by catalytic hydrogenation, generates the amino group and effects the reduction of the cyclic aminal intermediate. CHO RO-J

CHO

R(

RO-

-N

RO-

OR |-OH CH2OR 52

„

ικ>.../ωΛ-*

k-OR H2N-J CH2OR

R0

OH

54

53

ROM

"\

/

RO

55

OR

κ υ

"'\/ RO

56

OR

Scheme 4-13.

Scheme 4-14. CHO HOHO-OH

CH2OH 62

BnNH2 NaBH3CN AcOH, MeOH

ΙΤΛ H0

..... HO

63

Scheme 4-15.

77

4.2 Iminosugars from 'True' Sugars Scheme 4-16.

(a) phthNH, DEAD, PPh3; (b) HgCl2, CaCO3, MeCN/H2O; ((Cj\ (Z)( \

(MOMO)CH=CHCH2dBlpc2, BF3Et2O then H2O2 NaHCO3; (d) MsCl, Et3N; (e) MeNH2; (f) CbzCl, NaHCO3; (g) BH3 THF; (h) H2O2 NaHCO3; (i) H2, Pd/ /^ /·\ ΓΤ01 /1 \ · C; (j) HClaq; (K) ion exchange.

L-Arabinose

ς

,

HO

SEt -^^\^^/ Y^ γ^

OBn 64

M

„ ,

^-»-

-

i

SEt

HO, , Λ Λ > Η . Γ T OH ^ Λ / \ —/ 69

I

66

65

,

f

Bn0

OBn

ΓΥ ^^™ -V^A I » OH cbz OMOM

k N

QH

OBn

OBn

OMOM

Cbz 67

68

Some examples of synthesis of iminosugars via 'chain animation' do not involve the carbonyl group of the starting sugar in the cyclization step. In order to obtain more complex, or bicyclic iminosugars, the carbonyl function of the starting sugar in fact has been used for the elongation of the sugar skeleton. In Scheme 4-16 a route for the synthesis of several isomers of castanospermine is described. In this case L-arabinose was converted into the dithiane 64. The free primary hydroxyl group of 64 was transformed into a phthalimido group, and the aldehyde was regenerated by deprotection, affording 65. The aldehydo group was then exploited to elongate the molecule by reaction with a chiral allylborane. The homoallylic alcohol 66 was so formed stereoselectively. Mesylation of the hydroxyl function of 66 and deprotection of the amino group afforded the cyclic iminosugar 67. Finally, hydroboration-oxidation of the double bond of 59 and a further mesylation gave rise to the second ring of the target molecule 68 [19]. In Scheme 4 -17, the anomeric carbon of the starting sugar, originally masked as double bond by Wittig reaction and then regenerated, was once more exploited in the elongation step. In this case the formation of the bicyclic structure occurred through the azido-epoxy-tosylate 76, which upon reduction of the azide directly afforded 77 [2O]. In other cases the cyclization step has been effected, neglecting the carbonyl group of the starting sugar, and exploiting the conversion of a proper hydroxyl group of the sugar into a leaving group. This is the case in the example reported in Scheme 4-18 [21], in which the free hydroxyl group of the starting glyconolactone 79 was first converted into the azide 80. The lactone was then selectively reduced affording the azidodiol 81, which in turn was selectively protected at the primary hydroxyl group, mesylated at the secondary hydroxyl group and then submitted to selective removal of one isopropylidene to afford 82. The diol-mesylate 82, treated with barium methoxide, gave the epoxide 83 with inversion of configuration of the carbon atom bearing the leaving group. Finally, reduction of the azido group afforded the iminosugar 84 through epoxide opening and restoration of the original configuration at the carbon atom involved in the cyclization.

78

4 Synthetic Methods for the Preparation of Iminosugars OH BnO/ BnO BnO

BnO 70

71

™

75

76 g

I Bn0

OH

f

/

Scheme 4-17. (a) Ph3P+CH3 Br, BuLi,; (b) Tf2O then Bu4NN3; (c) O3, Sudan III 0,1%, -78 0C then Me2S; (d) Ph3P+(CH2)3OH Br, KN(SiMe3)2, Me3SiCl, -78 0C then HC1;

^ 78

Bn

° \^N^y

(O PTsCl; (g) H2, Pd/C, Et 2 O/EtOH(2:l);(h)H 2 ,

BnO^

Pd/C,

77

EtOH.

1. Ph2ButSiCl, Imidazole 2. MsCl 3.AcOH

[Si]O

84

Scheme 4-18.

82

79

4.2 Iminosugars from 'True'Sugars 4.2.2.3

Iminosugars from alditols

Readily available alditols have been extensively used to prepare iminosugars. One procedure to convert alditols into iminosugars is based on the tosylation of the primary hydroxyl groups of the starting molecule, and subsequent reaction of the ditosylate with an amine, as reported in Scheme 4-19. The procedure has been applied to xylitol, L-arabinitol and o-arabinitol, to prepare different iminosugars. The alditols were first protected at the primary hydroxyl groups as ditrityl derivatives. The secondary hydroxyl groups were then protected as benzyl ethers and detritylation afforded the diols 85, which were converted into the ditosylates 86. Finally, the ditosylates were treated with a primary amine, so affording the iminosugars 87 [22]. In a similar approach, 2,3,5-tri-0-benzyl-D-arabinitol 88, easily obtained by reduction of 2,3,5-tri-O-benzyl-D-arabinose, was converted into a cyclic sulfate 89. Reaction of the sulfate 89 with a primary amine afforded an aminosulfate 90, which underwent aminocyclization when treated with BuLi. Alternatively, the cyclic sulfate 89 was treated with lithium azide, and the obtained azidosulfate 91 was converted into the azidomesylate 92. Finally, the reduction of 92 afforded the iminosugar 93 (Scheme 4-20) [23]. In the example reported in Scheme 4-19 the two primary hydroxyl groups of the alditol were involved in the aminocyclization, whereas in the example reported in

Scheme 4-19.

, ν

Γ

J

\ /Μι

BnNH2

,

OH

93

Scheme 4-20.

γ

[

^NH2Bn

80

4 Synthetic Methods for the Preparation of Iminosugars

D-Mannitol

Pd/CA \

/

OH

50

Psi

MeOH HCI

:i Ph^

97

^O""

96

Scheme 4-21.

Scheme 4-20 a primary and a secondary hydroxyl group are involved. Examples are also proposed in which the aminocyclization occurs involving two secondary hydroxyl groups of the starting alditol. This is the case in the example reported in Scheme 4-21, in which the readily available l:3,4:6-di-0-benzylidene-D-mannitol 94 was converted into the iminosugar 97 by oxidation of the two free secondary hydroxyl groups, and reductive amination of the obtained diketone, which is present in the hydrated cyclic diketaIicform95[24]. Dihalogenated alditols such as 99 have been obtained, in the deprotected form, by reduction of the easily available corresponding glyconolactones [25]. This approach opens the way to the synthesis of different iminosugars simply by treatment of these dihalogenated alditols with ammonia. It is noteworthy that the procedure does not require any protective step (Scheme 4-22). I—OH

Br aqNH 3

—OH —OH

Br

98

Br

HO— —OH I—OH

— Br

99

Scheme 4-22.

4.3 Iminosugars from Noncarbohydrate Starting Materials

81

4.3 Iminosugars from Noncarbohydrate Starting Materials Although the similarity between iminosugars and true sugars induces use of the latter as starting materials to prepare the former, this is not always the most efficient procedure. Many efforts have been devoted towards the synthesis of iminosugars from noncarbohydrate starting materials. In these strategies the polyhydroxylated chiral skeleton of the target molecule has been stereoselectively built up by chemical or enzymatic procedures from shorter, easily available starting materials. The chemical methods are mainly based on the acyclic stereoselection in condensation reactions which involve chiral aldehydes. Alternatively, cycloaddition reactions have been employed, and the obtained adducts have been properly manipulated to afford the iminosugars. Furthermore, iminosugars have also been obtained by stereoselective functionalization of carbocyclic and heterocyclic intermediates. The enzymatic procedures are mainly based on the construction of the sugar-like skeleton, employing aldolases.

4.3.1

Ex novo synthesis of iminosugars via acyclic stereoselection

The chiral polysubstituted skeleton of sugars as well as iminosugars has been widely adopted as a stimulating target in acyclic stereoselection studies. Interesting results have been obtained in the stereoselective elongation of polyhydroxylated chiral aldehydes with different chemical reagents, or exploiting aldolases and dihydroxyacetone phosphate in chemoenzymatic processes. 4.3.1.1 Formation of iminosugars by stereoselective chemical condensations The skeleton of iminosugars can be stereoselectively built up by chemical procedures which elongate enantiomerically pure starting materials such as amino acids; this is the case of the synthesis reported in Scheme 4-23. Protected L-serinal 101 reacted with the stabilized ylid 102 to effect a 3-carbon atoms elongation with formation of the a,/?-unsaturated ketone 103. Stereoselective hydroxylation of the double bond of 103, reduction of the ketone and conversion of the thiazole ring into an aldehyde, afforded the protected aminosugar 104. Deprotection of 104 finally afforded the iminosugar 105 [26]. In a different approach L-serine was protected and reduced to the serinal 106, which was elongated by stereoselective reaction with propargylic ethyl ester anion, to afford 107 (Scheme 4-24) [27]. The triple bond of 107 was then reduced with Lindlar's catalyst, affording the Z-alkene 108. Treatment of 108 with trifluoroacetic acid, which removed the acetalic and Boc protecting groups, afforded, in basic conditions, the unsaturated lactam 109. Protection of the amino and hydroxyl

82

4 Synthetic Methods for the Preparation of Iminosugars

L-Serine

\^

X

CHO 102

101

103

1. OsO4, NMMO 2. DMP, TsOH 3. NaBH4 4. tBuMe2SiCl, Imidazole 5. MeI; NaBH4; HgCl2

Scheme 4-23. QBz

^CO2Et

L-serine

** /

\

—-—*·

/

\

V

^J

^

107

9»

TJ HN

/

?

k/k/o" λ^

in

OH

OH

?BI

^- \l /^-τ ι />^^

r

\^

'^^/^

\

el no

°\\ /

QBz OH

'

k/k HN

\

de

\

\

CO2Et

/

^

J-

Ii 109

Scheme 4-24. (a) HOCCO2Et, BuLi, HMPT; (b) BzCl; (c) Lindlar's cat.; (d) Et2O/H2O/TFA (1:1:3); (e) EtOAc, sat. NaHCO3; (f) PhH, DMP, PPTS; (g) Ace+ tone/H2O (1:1), OsO4, NMMO then BH3SMe2 then dowex H .

groups of 109, stereoselective syn-dihydroxylation of the double bond, and final deprotection, afforded the iminosugar 111. Starting from L-alanine the iminosugar 117, related to L-fucose, has been prepared exploiting the stereoselective Sharpless osmylation. In this synthesis, reported in Scheme 4-25 [28], the benzophenone imine of L-alanine methyl ester 112 was reduced with diisobutylaluminium hydride, and the obtained aldehyde was reacted with an E'-lithio-allyl ether to afford the /?-hydroxy-y,<5-unsaturated imine 113 with excellent stereoselection (20:1). The free hydroxyl group of 113 was protected as pivaloyl ester, and the double bond was osmylated applying Sharpless conditions to afford the iminoalditol 115. Deprotection and selective oxidation of the primary hydroxyl group of 115 with TEMPO, and final deprotection, afforded the iminosugar 117.

83

4.3 Iminosugars from Noncarbohydrate Starting Materials Ij

OH OM

i

'

N=CPh

>

LDIBAL

Λ

~LK~^r^OR '

2

OPiv x/\.

OSiMe2t-Bu

Y '^ N=CPh2

112

J

PivCl

^^

OSiMe2I-Bu

/X

"

"Ύ " J=CPh2

113

114 Sharpless Osmylation

1.NaBH3CN 2.Bu 4 NOH

PlV

¥ J^

+ Ph2=N

*

OH

/\^

LBu 4 NF ^O 2. NaOCl, TEMPO

OH

C Ξ

I i Ph2=N

116

Ξ OH

115

Scheme 4-25.

In a different approach, the D-serine-derived reagent 119, was exploited in the synthesis of the iminosugar 122 (Scheme 4-26) [29]. 119 was condensed with (7?)-isopropylideneglyceroyl chloride 118, and the obtained ketone 120 was stereoselectively reduced with L-Selectride® to the corresponding alcohol, which spontaneously gave, after appropriate protection-deprotection steps, the lactone 121. This molecule has an amino and a hydroxyl group in the proper positions to afford the iminosugar 122, according to a sequence reported by Fleet [30], in which the amino group originated from D-serine, effects a nucleophilic substitution on the alcoholic carbon atom originally β to the carboxyl group of glyceroyl chloride. An interesting alternative is reported in Scheme 4-27. In this case the starting material was the benzimine of isopropylidene-L-glyceraldehyde 123, and the amino group required for the formation of the iminosugar was obtained by stereoselective attack of reagent 124, a useful synthon for a 4-carbon atoms elongation [31]. The obtained a,/?-unsaturated lactone 125 was then stereoselectively dihydroxylated to afford the protected aminolactone 126. Deprotection of the amino group of 126 gave rise to a lactam, the reduction of which finally afforded the imimosugar 127. All these examples describe the synthesis of iminosugars by stereoselective elongation of small, easily available chiral synthons, exploiting different chemical /NHBoc IZn^ ^Y

7—0 <

I

/O O

L2Bn

y^-0 119

θ(

λ

Pd

^

. I CO2Bn

O

120

H

°'

NH

Scheme 4-26.

-

IBuMe2SiO

122

CO2H

1. L-selectride then I2, MeOH 2. TBSCl

HO \

°

c> ^xx

o

121

84

4 Synthetic Methods for the Preparation of Iminosugars °

OTMS 124

123

BF 3 Et 2 O 1. CbzCl 2. KMnO4 then DMP, TsOH 1. H2, Pd(OH)2

CbzNH

ο

Scheme 4-27.

procedures. These synthetic approaches also require, in general, the stereoselective dihydroxylation of a double bond, in order to create the polyhydroxylated skeleton of the iminosugar with the required stereochemistry. On the basis that, in nature, the sugar skeleton is stereoselectively constructed from smaller molecules in aldolase-catalyzed condensation reactions, the idea of preparing iminosugars using aldolases was also explored. 4.3.1.2

Formation of iminosugars employing aldolases

The polyhydroxylated skeleton of iminosugars has been stereoselectively built up exploiting the ability of aldolases to condense DHAP with different aldehydes. It has been shown that α-azido- or a-hydroxy-/J-azidoaldehydes are good substrates for these enzymes (Scheme 4-28), affording γ- and <5-azidoketones respectively. These azidoketones were easily transformed into five- or six-membered ring iminosugars by catalytic hydrogenation. As seen before, the catalytic hydrogenation converts the azide into an amine and reduces the aminal to the cyclic imine. Using fructose 1,6-diphosphate aldolase (FDP aldolase), a largely available and inexpensive enzyme, the two stereocenters are formed with Si-Si stereoselectivity, as shown in Scheme 4-28 [32-35]. Other aldolases, namely L-fuculose !-phosphate aldolase and L-rhamnulose 1-phosphate aldolase, that generate other possi-

N3 I

1. DHAP, FDP aldolase 2. acid phosphatase

·3

9Η

I

_

^

OH

128

O

129

N3x

I •

131 1J1

Scheme 4-28.

^

QH

OH

«N J

,,

I ~

OH

O

132

4.3 Iminosugars from Noncarbohydrate Starting Materials OH

85

OH ~

134

r.u 9H

!37

1) DHAP, Fuc-lP aldolase 2) acid phosphatase

OH

OH

N;

I38

Scheme 4-29.

ble stereoisomers, have been used in the synthesis of different stereoisomers of the iminosugar reported in Scheme 4-28 [32]. Interestingly, in these enzymatic syntheses, it is not necessary to use an enantiomerically pure starting material. In the presence of a racemic aldehyde the enzyme in fact is able to effect a kinetic resolution. Scheme 4-29 shows an example of this resolution; in this case a racemic mixture of 3-azido-2-hydroxypropanal is treated with dihydroxyacetone phosphate in the presence of rhamnulose 1-phosphate aldolase or fuculose 1-phosphate aldolase. As can be observed from the scheme, the /^-aldehyde is a better substrate for rhamnulose 1-phosphate aldolase whereas the 5-aldehyde reacts faster in the condensation catalyzed by fuculose 1-phosphate aldolase [35]. By employing aldolases, a great variety of iminosugars have been synthesized and tested as inhibitors of glycosidases.

4.3.2 Iminosugars via cycloaddition reactions It is now well established that rare and modified sugars can be prepared by DielsAlder reactions which exploit furan as diene. Following the same approach, different iminosugars have been prepared. In the example reported in Scheme 4-30 [36], the Diels-Alder adduct 142 of furan and the 1-cyanovinytester of (IS)-camphanic acid was converted into the ketone 143. The ketone was then transformed in three steps into the bromolactone 144. Opening of the bromolactone and conversion of the bromine into an amino group, afforded the aminosugar 145 which was then easily converted into the iminosugar 146 by deprotection. Following this procedure different iminosugars of biological relevance, such as castanospermine and derivatives have been synthesized [37].

86

4 Synthetic Methods for the Preparation of Iminosugars

146

145

'

Scheme 4-30. (a) ZnI2; (b) N-^BDMS-N-Me-trifluoroacetamide; (c) Br2; (d) CF3COOH, Na2HPO4; (e) CH2=CHCH2OH, CH3SO3H; (f) EtOH/H2O (9:1), [Rh(PPh3)SCl], DABCO; (g) CsN3 then PhCH2Br; (h) LiAlH4; (i) HCl.

4.3.3

Iminosugars from cyclic dienes

The possibility of converting double bonds into diols stereoselectively, has stimulated the interest on the synthesis of sugar analogs, among them iminosugars, from cyclic, unsaturated starting materials. For example, cyclopentadiene has been chemoenzymatically converted into iminosugars [38, 39], as shown in Scheme 4-31 for the synthesis of 1-deoxynojirimycin 152 [38]. Enzymatic asymmetrization and selective protective group manipulation allowed the transformation of cyclopentadiene in the enone 148, which was then α-iodinated, reduced, and the allylic alcohol protected, affording the vinyl iodide 149, that was then transformed into the alcohol 150. Ozonolysis and reductive amination followed by deprotection afforded the final product 152.

chemoenzymatic synthesis

147

!ι o^/X

».»»

„

148

Η

Ϊ

OH

152

η

11 149

r

BS

OTBS

γ — y\ τ J

,

σΤ

OH

15

151

°

0

Scheme 4-31. (a) NaBH4, CeCl3 7H2O, -78 C; (b) TBSCl, Imidazole; (c) CO 0 0 (1 atm.), Bu3SnH, Pd(Ph3)4 then NaBH4, CeCl3 7H2O, -78 C; (d) O3, -78 C then DMS; (e) BnNH3Cl, NaBH3CN; (f) IN HCl; (g) H2, 30 psi, Pd/C.

4.3 Iminosugars from Noncarbohydrate Starting Materials

OH

OH

OTBS

159

87

OTMS 158

157

Scheme 4-32. (a) DMP, H+ then MCPBA/phosphate buffer; (b) LiCl, ethyl acetoacetate, NaN3; (c) HMDS/TMSC1; (d) O3, -78 0C; (e) NaBH4, O 0 C, H2, Pd/C; (f) TBSCl, DiPEA.

In a similar approach halocyclohexadiene-cis-diols such as 154, synthetised from chlorobenzene via enzymatic hydroxylation, have been used for the synthesis of iminosugars. An example is reported in Scheme 4-32 [4O]. The cis-l,2-dihydroxy-3-chlorocyclohexa-3,5-diene 154 was converted into the epoxide 155 which was stereoselectively transformed into the azidoalcohol 156. Ozonolysis of 156, followed by oxidative cleavage, gave the hydroperoxide 157, which was reduced with sodium boron hydride and then catalytically hydrogenated to afford the lactam 158. Finally, 158 was easily converted into the iminosugar 159 by reduction and deprotection. Heterocyclic dienes already containing the nitrogen atom in the ring have also been used as precursors of imino sugars, as in the example shown in Scheme 4-33. Starting from hydropyridine derivative 160, the piperidine triol 162 has been obtained directly by treatment of 160 with osmium tetroxide and catalytic hydrogenation [41]. 9H

QH

H0 OsO4, NMMO

v /^\^N

^S^

OH

« ™/^ H ,Pd/C 2

H(V

^^

CO2Me

Scheme 4-33.

160

i6i

162

°H

88

4 Synthetic Methods for the Preparation of Iminosugars

4.3.4

Iminosugars from other non-carbohydrate substrates

Iminosugars can be considered, in many cases, as polyhydroxylated pyrrolidines or piperidines, with high chirality content. This has suggested to synthetic chemists that all types of chiral synthons may be used, especially if easily available, for the synthesis of these compounds. In the previous paragraphs we have shown some examples in which amino acids have been used for this purpose; indeed, these compounds have been used extensively as starting materials as they are easily available in an enantiomerically pure form. In this paragraph some examples of the synthesis of iminosugars are reported, which use chiral synthons and cannot be classified among the methods already described. The following examples simply show the potential in using basic chiral starting materials. Scheme 4-34 shows how epicastanospermine 169 and castanospermine 170 have been produced starting from dimethyl L-tartrate 163 [42]. In this multistep synthetic scheme the two carboxylic groups of dimethyl L-tartrate 163 were manipulated selectively to afford, in six steps, the a,/?-unsaturated alcohol 164. The double bond of this molecule was then functionalized by Sharpless stereoselective epoxidation, and the epoxide treated with Et2AlN(CH2Ph)2 to afford an aminoalcohol that was converted into 165. The primary hydroxyl group of 165 was oxidized, and the obtained aldehyde submitted to condensation with ethyl acetate, which stereoselectively afforded 167 and a small amount of its diastereomer 168. Both 167 and 168 are useful precursors of bicyclic iminosugars,

163

MOMQ

OH

'°\^/^\ /^\

,COOEt

MOMQ

OH

o H >,

1

"^^^

/ /

^-^/-

\0

NBn2

NBn2

/\

OTBDMS

OTBDMS

89

167

:

11

HQ

HQ

OH

\^

H

169

^^

Α

*° °

\

/\ \ / O^N

NBn2

OTBDMS

8 steps

H

1

^LiN(SiMe3)Z Ε

χ

I66

6 steps

HO '

MOMQ COOEt

1

HO ''

\^

no

OH f

7

^

Scheme 4-34.

4.3 Iminosugars from Noncarbohydrate Starting Materials

89

CCl3

OH

O^ ^N

V^ V L — ,' νTT 171

ecu

172

1

1. HCl 6N, MeOH, r.t. 0 2. NaHCO3, MeOH, SO C; 0 BoC2O, O C

QH /\

173

I74

^NHBoc

1?5

Scheme 4-35.

in having the amino group in the correct position to effect a six-membered cyclization on the primary alcoholic carbon atom, and a five-membered cyclization on the carbonyl group. Exploiting this requirement, the iminosugars 1-epicastanospermine 169 and castanospermine 170 were obtained respectively. Another example of synthesis of iminosugars exploits the €2 symmetrical and meso-iodoamino alcohols 172 and 177, obtained from 3,4-dihydroxyhexa-l,5dienes 171 and 178 by double iodoamination of the corresponding trifluoroacetoimidates. Cyclization of these iodoamino alcohol afforded different iminosugars 173-179, as reported in Schemes 4-35 and 4-36 [43]. Finally, cyclitols of natural origin, such as L-quebrachitol 180, have been also used as starting materials for the synthesis of iminosugars. In this case the chiral polyhydroxylated cyclohexane structure was properly functionalized with the introduction of an amino group and a carbonyl group, and the latter was used to cut the carbocyclic structure by Baeyer-Villiger oxidation, according to Scheme 4-37 [44].

__ ,

1. HCl 6N, MeOH, r.t.

177

Scheme 4-36.

178

/

F

179

90

4 Synthetic Methods for the Preparation of Iminosugars OMe I

OMe

HO

^^

^OH

,,OH

-^ χ O1

H

° 180

. . ,,OH / \ ,."

____^ -*.

^ ^\

^\^° ^ 181

182 I l.H2,Raney.Ni 1 2. CF3COOEt, Et3N 3. TEMPO, NaBrO2

OMe

A^

OMe O V^O

MCPBA

o?"""X^\NHCOCF -6

QI ι

\^/

^NHCOCF3

3

—V—ο

184

183

1. NaBH4 2. (BoC)2O

/ \ lOMe

Ol 1) H2, Pd/CHnUl nH-.Pd/C i 2) SO2 gas

I

Γ

OH ^OH >

/^\ / \ ^^

' ! l t ' ^^^^ H2,Raney-Ni

O3S^^^N+/^^CH2OH H2

187

4.4

1 ^N^^CH CH2( 2OH H

188

Conclusions

This review has been written with the aim of showing how many different approaches have been developed to synthesize iminosugars, from all possible starting materials, the effort being to rationalize, if possible, and classify these different procedures. An in-dept description of specific procedures and particular targets is left to other authors of this book.

References

91

Abbreviations CbzCl DABCO DEAD DIBAL DiPEA DMP DMS DMSO HMDS HMPT MCPBA MOM NMMO PCC PPTS Sudan III 'BDMS TEMPO TFA TMEDA TMSCl

tert-butoxycarbonyl anhydride carbobenzoxy chloride l,4-diazabicyclo[2.2.2]octane diethylazodicarboxylate diisobuthylaluminiumhydride diisopropylethylamine 2,2 -dimethoxypropane dimethyl sulfide dimethyl sulfoxide hexamethyldisilazane hexamethylphosphorous triamide 3-chloroperbenzoic acid methoxymethyl 4 -methy lmorpholine- 4 -oxide pyridinium chlorochromate pyridinium toluene- 4 -sulfonate l-[[4-(phenylazo)phenyl]azo]-2-naphtalenol tert-butyldimethylsilyl (2,2,6,6-tetramethylpiperidine-N-oxyl) trifluoroacetic acid tetramethylethylendiamine trimethylsilylchloride

References [1] R. A Gruters, J. J. Neefjies, M. Tersmette, R. E. Y. de Goede, ATuIp, H. G. Huisman, R Miedema, H. L. Ploegh, Nature, 1987, 330, 74-77. [2] G. W. J. Fleet, A. Karpas, R. A. Dwek, L. E. Fellows, A. S. Tyms, S. Petursson, S. K. Namgoong, N. G. Ramsden, P. W. Smith, J. C. Son, F. Wilson, D. R. Witty, G. S. Jacob, T. W. Rademacher, FEBS Letters, 1988, 237, 128-132, and ref. therein. [3] L. A. G. M. van der Boek, D. J. Vermaas. B. M. Heskamp, C. A. A. van Boeckel, M. C. A. A. Tan, J. G. M. Bolsher, H. L. Ploegh, F. J. van Kemenade, R. E. Y. de Goede, F. Miedema, Red. Trav. Chim. Pays-Bas, 1993, 772, 82-94, and ref. cited therein. [4] C. G. Bridges, S. P. Ahmed, M. S. Kang, R. J. Nash, E. A. Porter, A. S. Tyms, Glycobiology, 1995, 5, 249-253. [5] R. PiIi, J. Chang, R. A. Partis, R. A. Mueller, F. J. Chrest, A. Passaniti, Cancer Res., 1995, 55, 2920-2926, and ref. therein. [6] D. J. Hardick, D. W. Hutchinson, S. J. Trew, E. M. H. Wellington, Tetrahedron, 1992, 48, 6285-6296. [71 D. J. A. Schedler, B. R. Bowen, B. Ganem, Tetrahedron Lett., 1994, 35, 3845-3848. [8] R Zhou, H. M. Salleh, J. F. Honek, /. Org. Chem., 1993, 58, 264-266. [9] L. Cipolla, L. Lay, F. Nicotra, C. Pangrazio, L. Panza, Tetrahedron, 1995, 57, 4679-4690.

92

4 Synthetic Methods for the Preparation of Iminosugars

[10] L. Cipolla, F. Nicotra, C. Pangrazio, Gazzetta Chim. Ital, 1996, 726, 663-666. [11] L. Lay, F. Nicotra, A. Paganini, C. Pangrazio, L. Panza, Tetrahedron, 1993, 34, 45554558. [12] R. C. Bertonas, B. Ganem, Tetrahedron Lett., 1984, 25, 165-168. [13] R. C. Bertonas, B. Ganem, Tetrahedron Lett., 1985, 26, 1123-1126. [14] J. G. Buchanan, K. W. Lumbard, R. J. Sturgeon, D. K. Thompson, R. H. Wightman, /. Chem. Soc. Perkin Trans I, 1990, 699-706. [15] Q. Meng, M. Hesse, HeIv. Chim. Acta, 1991, 74, 445-450. [16] H. S. Overkleeft, J. van Wiltenburg, U. K. Pandit, Tetrahedron Lett., 1993, 34, 25272528. [17] A. de Raadt, A. E. Stutz, Tetrahedron Lett., 1992, 33, 189-192. [18] E. W. Baxter, A. B. Reitz, Bioorg. Med. Chem. Lett., 1992, 2, 1419-1422. [19] K. Burgess, D. A. Chaplin, I. Henderson, Y. T. Pan, A. D. Elbein, /. Org. Chem., 1992, 57, 1103-1109. [20] W. H. Pearson, J. V. Hines, Tetrahedron Lett., 1991, 32, 5513-5516. [21] P. M. Myerscough, A. J. Fairbanks, A. H. Jones, I. Bruce, S. S. Choi, G. W. J. Fleet, S. S. Al-Daher, I. C. di Bello, B. Winchester, Tetrahedron, 1992, 48, 10177-10190. [22] A. E. Mc Caig, B. Chomier, R. H. Wightman, J. Carbohydr. Chem., 1994, 13, 397-407. [23] P. A. M. Van der Klein, W. Filemon, H. J. G. Broxterman, A. G. Van der Marel, J. H. Van Boom, Synt. Comm., 1992, 22, 1763-1771. [24] W. Zou, W. A. Szarek, Carbohydr. Res., 1993, 242, 311-314. [25] I. Lundt, R. Madsen, Synthesis, 1995, 787-793. [26] A. Dondoni, P. Merino, D. Perrone, Tetrahedron, 1993, 49, 2939-2956. [27] H. J. Altenbach, K. Himmeldirk, Tetrahedron Asymm., 1995, 6, 1077-1080. [28] D. Sanies, R. PoIt, Synlett, 1995, 552-554. [29] R. F. W. Jackson, A. B. Rettie, Tetrahedron Lett., 1993, 34, 2985. [30] B. P. Bashyal. H-F. Chow, G. W. J. Fleet, Tetrahedron, 1987, 43, 423-430. [31] G. Rassu, L. Pinna, P. Spanu, M. Culeddu, G. Casiraghi, G. Gasparri Fava, M. Belicchi Ferrari, G. Pelosi, Tetrahedron, 1992, 48, 727. [32] C. H. Wong, R. L. Halcomb, Y. Ichikawa, T. Kajimoto, Angew. Chem. Int. Ed. Engl., 1995, 34, 412-432 [33] A. Straub, F. Effenberger, P. Fischer, /. Org. Chem., 1990, 55, 3926-3932. [34] R. R. Hung, J. A. Straub, G. Whitesides, /. Org. Chem., 1991, 56, 3849-3855. [35] K. K. C. Liu, T. Kajimoto, C. Lihren, Z. Zhong, Y. Ichikawa, C. H. Wong, J. Org. Chem. 1991, 56, 6280-6289. [36] Y. Auberson, P. Vogel, Angew. Chem. Int. Ed. Engl., 1989, 28, 1498-1499. [37] J. L. Reymond, A. A. Pinkerton, P. Vogel, J. Org. Chem., 1991, 56, 2128-2135. [38] C. R. Johnson, A. Golebiowski, E. Schoffers, H. Sundram, M. P. Braun, Synlett, 1995, 313-314. [39] C. R. Johnson, B. M. Nerurkar, A. Golebiowski, H. Sundram, J. L. Esker, J. Chem. Soc. Chem. Commun., 1995, 1139-1140. [40] T. Hudlicky, J. Rouden, H. Luna, /. Org. Chem., 1993, 58, 985-987. [41] T. Tschamber, F. Backenstrass, M. Neuburger, M. Zehnder, J. Streith, Tetrahedron, 1994, 50, 1135-1152. [42] H. Ina, C. Kibayashi, J. Org. Chem., 1993, 58, 52-61. [43] S. H. Kang, D. H. Ryu, Tetrahedron Lett., 1993, 38, 607-610. [44] N. Chida, T. Tanikawa, T. Tobe, S. Ogawa, /. Chem. Soc. Chem. Commun., 1994, 12471248.

5

Iminosugars as Powerful Glycosidase Inhibitors - Synthetic Approaches from Aldonolactones INGE LUNDT and ROBERT MADSEN

5.1 Introduction Since iminosugars were discovered to be glycosidase inhibitors their synthesis has attracted major attention [1]. Carbohydrates have been the most common starting materials due to their closely related structural motif. However, multiple protection-deprotection sequences have often been necessary to prepare specific compounds, and this has stimulated many efforts to develop novel syntheses of iminosugars. In recent years aldonolactones have emerged as valuable components in carbohydrate synthesis [2, 3]. Aldonolactones/aldonic acids constitute a more diverse chiral pool of compounds than reducing sugars. From each reducing sugar, three to four aldonolactones/aldonic acids can be prepared and crystallized in one step by either anomeric oxidation, Kiliani chain elongation or Humphlett oxidative degradation. Although protection-deprotection sequences on aldonolactones can give rise to many iminosugars, we were attracted to a different strategy. We have developed one-step procedures for regioselective functionalization of aldonolactones to give crystalline compounds from which iminosugars can be prepared in only 2-4 steps [2,4-11]. This constitutes the most general nonenzymatic approach to iminosugars from carbohydrate precursors without the use of protecting groups. Regioselective functionalization of aldonolactones is possible because of the enhanced reactivity of the a and the co positions. Treatment of aldonolactones/aldonic acids with hydrogen bromide in acetic acid affords bromodeoxyaldonolactones [2]. In this strongly acidic medium acetoxonium ions can be formed. Subsequent opening of these with bromide ions gives acetylated bromohydrins. The ability to form these acetoxonium ions determines where bromine is introduced. Bromine is always introduced in the primary position (w position) although prolonged treatment might be necessary on tetrono- and pentonolactones. When the a- and /?hydroxy groups are cis oriented, bromine is always introduced in the a-position with inversion of configuration. Sometimes a-bromides can also be formed from lactones having the a- and /?-hydroxy groups trans oriented, i.e. on glucono-, xylono-, and threonolactone [2] (Scheme 5-1). Iminosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

94

5 Iminosugars as Powerful Glycosidase Inhibitors... ' Pyridine

<x,(o-Dibrc>mo-a,a>-dideoxy aldonolactones

η = 0,1,2 Aldonolactones

^

^

Br OH L(i]

-

0

ω-Bromodeoxy-aldonolactones JsC|

Pyridine

LBr 2,4-Dibromo-2,4-dideoxytetronic acid methyl esters

Tetronolactones

α,ω-Ditosyl-aldonolactones

Scheme 5-1.

Regioselective bromination of the primary position in aldonolactones can also be achieved with carbon tetrabromide and triphenylphosphine in pyridine [12] or thionyl bromide in dimethyl formamide [13]. Attempts to tosylate regioselectively the primary hydroxy group were less successful due to competing reaction of the ahydroxy group. Instead, α,ω-ditosylated aldonolactones can be prepared by treatment with 2.0-2.3 eq. of tosyl chloride in pyridine [14]. The reaction is particularly efficient for lactones having the a- and /?-hydroxy group cis oriented and also cis to the side chain, giving high yields of crystalline α,ω-ditosylated aldonolactones. This ditosylation method complements the dibromination with HBrAcOH. Both methods introduce «,^-leaving groups, but with different stereochemistry at C-2. Tosylation retains the stereochemistry at C-2 while HBr-AcOH inverts it. The preparation of iminosugars from these functionalized lactones requires at least two transformations, ring closure to form the pyrrolidine/piperidine ring and reduction of the lactone group. To carry out these two transformations we have developed four different strategies [4-11] (Scheme 5-2). Strategy I relies on direct ring-closure of dibromodideoxy-/ditosyl hexono- and heptonolactones with ammonia to form iminoamides. These are subsequently reduced to pyrrolidines and piperidines of type A. Strategy II is closely related by interconverting the two steps. Dibromodideoxy hexono- and heptonolactones are reduced to the corresponding dibromoalditols which on treatment with ammonia give the same pyrrolidines/piperidines of type A. When functionalized pentonolactones and tetronic acid esters are treated with ammonia, lactams are obtained as shown in strategy III. Subsequent reduction of the lactam then gives pyrrolidines and piperidines of type B. Strategy IV differs from the previous routes by introducing the nitrogen by virtue of an azide instead of ring closure with ammonia. Dibromodideoxyhexonolactones give rise to 2-azido lactones which by azide reduction and ring closure yields iminoacids. Reduction of the acid then affords pyrrolidines of type C. Strategies I-IV all have a number of points in common. All ring closure reactions to the pyrrolidine/piperidine ring proceed through epoxides which are opened

5.2 Syntheses Br OH

^

ο

Br OH

=0

^

I—I—S^

QH

τ""0

V-Y HO

"

^\

NH2

95

N

/~R

W

HO

OH

lminoacid: R = COOH „ C : R = CH2OH

Strategy I

H

Strategy III

H2NOC

X=OH, Br, OTs n=0

(^

I— Lactam: R1 R = O I * B: R, R = H, H x = Br

\-4.

°(T^"^ HO

I Strategy Il η = 1,2

I Strategy III n = -1

COOMe

U-Br

-

UvOH L-Br

OH

A

Dibromoalditol

Scheme 5-2.

by ammonia (intermolecular) or by an amino group (intramolecular). All reactions are easily performed on large, gram scale, using cheap reagents and give rise to crystalline products. In the following text the four strategies will be discussed in detail.

5.2

Syntheses

5.2.1

Iminosugars via iminoamides - strategy I

The cheapest aldonolactone is D-gluconolactone from which dibromomannonolactone 1 is available upon reaction with hydrogen bromide in acetic acid. Direct treatment of 1 with saturated aqueous ammonia (Scheme 5-3) gave iminoamide 2a [4]. The mechanism was elucidated by monitoring the reaction by 13C-NMR spectroscopy. The spectra revealed immediate formation of epoxyamide Ia which is converted rapidly into diepoxyamide Ib. The next product observed was iminoamide 2a, presumably formed through opening of the 5,6-epoxide in Ib by ammonia to afford aminoepoxide Ic. This is, however, not observed in the 13 C-NMR spectra and most likely cyclizes immediately to form the pyrrolidine ring by exo-opening of the 2,3-epoxide. Iminoamide 2a can be converted into iminoallitol 3 by a one-pot procedure involving hydrolysis to the acid 2b, esterifica-

96

5 Iminosugars as Powerful Glycosidase Inhibitors...

NaBH4 OH

HO/ '

syrup 52% cryst. 30%

Scheme 5-3.

tion and reduction with sodium borohydride to yield pyrrolidine 3. All the reactions are easily performed on large, gram-scale and pyrrolidines 2a and 3 can be isolated by direct crystallization of the crude products [4]. Iminoallitol 3 has previously been prepared from L-gulonolactone in 5-6 steps using protection/deprotection techniques [15, 16]. This concept can be extended to several differently functionalized aldonolactones. Ditosylmannonolactone 4 also gave iminoamide 2a on treatment with ammonia through the same cascade of epoxide intermediates (Scheme 5-3). Dibromogluconolactone 5 available from D-mannonolactone gave iminoamide 6a on treatment with ammonia [4] (Scheme 5-4). Inverting the stereochemistry at C-2

syrup 69%

CONH2

CONH2

CONH2

O^ — OH — OH -Br

5a

Scheme 5-4.

5.2 Syntheses

97

does not greatly affect the ring closure. By 13C-NMR, epoxides 5a and 5b are observed as well as their conversion into iminoamide 6a through aminoepoxide 5c; the latter could, however, not be observed. One-pot hydrolysis, esterification and reduction gave iminotalitol 7 [4]. This pyrrolidine has been previously prepared from o-mannose in 5-6 steps [16-18]. The ring closure could also be performed on dibromoheptonolactones as shown with 8 [6] (Scheme 5-4). Here ammonia at 6O 0 C gave better results than aqueous ammonia and furnished iminoamide 9. This reaction was also shown to proceed through a cascade of epoxides. Iminoamide 9 was then converted into the corresponding iminoheptitol 10 by the same one-pot procedure as described above [6]. The ring closure reaction cannot be extended to all dibromohexono- and -heptonolactones as shown with dlbromoaltronolactone 11 [7]. Thus treatment of 11 with ammonia, aqueous or liquid, at different temperatures always furnished a complex mixture of compounds dominated by tetrahydrofuran derivatives. One of the problems seems to be the highly reactive bromine in the 2-position. However, because the ammonia ring closure reaction has to proceed through epoxides, these epoxides can be preformed prior to the ammonia reaction and thus prevent some of the side reactions. Treatment of 11 with the weak base potassium fluoride gave a 5:2 mixture of epoxylactones 12a and 12b (Scheme 5-5). Reaction with aqueous ammonia then gave only one product, iminoamide 13a which hydrolyzed slowly to iminoacid 13b. Monitoring the ring closure by 13C-NMR revealed formation of the same epoxy intermediates as described above. Lactonization of iminoacid 13b gave 14 which was reduced to iminogulitol 15 [I]. This pyrrolidine has previously been prepared from D-glucose in 7 steps [19]. Although preforming the epoxides from the dibromoaldonolactones can be a very efficient way of avoiding side reactions in the cyclization with ammonia, it does add steps. This has stimulated several efforts to develop other techniques to convert dibromolactones into imino compounds and reduce them to iminosugars. One solution emerged as a reversal of the order of the two steps.

H2N-J

Scheme 5-5.

98

5 Iminosugars as Powerful Glycosidase Inhibitors...

5.2.2

Iminosugars via dibromoalditols - strategy II

Dlbromoaldonolactones are easily reduced to dibromoalditols with sodium borohydride. This diminishes the reactivity of the C-2 bromine and dibromoalditols are less prone to give tetrahydrofuran derivatives on treatment with base. Indeed, dibromohexitols and -heptitols are excellent substrates for pyrrolidine and piperidine formation with aqueous ammonia. However, in contrast to the dibromolactones, ring closure of dibromoalditols occurs through different cascades of epoxides, depending on the ease by which a 2,3-epoxide is formed. Reaction of dibromomannitol 16 with aqueous ammonia gave iminoallitol 3 as the only product [4] (Scheme 5-6). Monitoring the cyclization by 13C-NMR spectroscopy revealed the immediate formation of 2,3-epoxide 16a which rapidly was converted into diepoxide 16b. Iminoallitol 3 was observed as the next product, presumably through opening of the 5,6-epoxide in 16b by ammonia to give 16c followed by exo-opening of the 2,3-epoxide. Aminoepoxide 16c was not observed in the 13C-NMR spectra [4]. When the C-2 isomeric dibromoglucitol 17 was reacted with aqueous ammonia, iminotalitol 7 was obtained as the only product [4] (Scheme 5-6). However, 13C-NMR revealed a different pattern in formation

aqNH3 syrup 73% cryst.51%

\OH IX

HO/ '

\OH

HO/

HO

— OH — OH

syrup 76% cryst. 48%

' -OH

NH3

— OH '

— OH — OH -NH2

Scheme 5-6.

5.2 Syntheses

99

— OH

L

OH

^-/

L— OH

,-=0 Γ

— Br

NaBH4 cryst.58%

™Π

aq NH3 1 cryst.51%

I—OH -OH

19 I

NH3

l—OH

r—OH

— Br

-Br

HO — HO—I

I— OH,

Scheme 5-7.

x°

""" 193

L-NH2

of the intermediate epoxides. In this case, a 5,6-epoxide 17a was the first product formed, followed by formation of the diepoxide 17b. Thus a 2,3-cis epoxide is less favored than a 2,3 -trans epoxide. Iminotalitol 7 was then formed and again the presumed intermediate aminoepoxide 17c could not be detected in the 13C-NMR spectra [4]. When the C-4 isomeric dibromogalactitol 19 was treated with aqueous ammonia, iminoiditol 20 was obtained as the sole product [5] (Scheme 5-7). The 13 C-NMR spectra revealed immediate formation of 5,6-epoxide 19a. However, because galactitols are very stable in a planar zig-zag conformation where the C-2 bromine and C-3 hydroxy group are syn, formation of a 2,3-epoxide is very difficult. This results in the formation of aminobromide 19b which is slowly converted into iminoiditol 20, through the intermediate aminoepoxide 19c, which was not observed in the 13C-NMR spectra [5]. These ring closure reactions can also be extended to dibromoheptitols to form piperidines. Thus, dibromo-D-g/yccro-L-g/wco-heptitol 22 gave piperidine 23 as the only product on treatment with ammonia [6] (Scheme 5-8). Monitoring the reaction by 13C-NMR revealed the formation of a similar cascade of epoxides as observed for the hexitols above. However, contrary to the hexitols, in this case the amino epoxide 22c could be observed in the spectra. Formation of the six-membered piperidine ring is kinetically slower than formation of the five-membered pyrrolidine ring. Analogously, dibromo-D-glycero-O-galacto-heptitol 25 was converted into piperidine 26 with aqueous ammonia [6] (Scheme 5-8). As observed previously with galactitol 19, the formation of the 2,3-epoxide is also here somewhat sluggish which permits the formation of both diepoxide 25b and aminobromide 25c. Aminoepoxide 25d is then observed in the spectra followed by exoopening of the 2,3-epoxide to give piperidine 26. In general, cyclization of dibromohexitols and -heptitols with aqueous ammonia occurs very cleanly through a cascade of epoxides. High yields of the crude products are obtained and after crystallization the iminosugars can be isolated in about 50 % yield. This constitutes the shortest way to synthesize polyhydroxypyrrolidines and -piperidines with a 1,2-dihydroxyethyl side chain.

100

5 Iminosugars as Powerful Glycosidase Inhibitors...

aqNH3

NaBH4 Br/

~

cryst. 79%

syrup 97% cryst. 47%

HO —

^

HO — -OH

L-Br

22

HO —

HO —

HO —

HO-

-°H

0

·cryst. 86%

HO —

syrup 87% cryst.

HO — — OH

— OH

— OH

— OH -Br 24

\ -Br HO—

NH3 O' -^

5-8.

5.2.3

Iminosugars via lactams - strategy III

Contrary to dibromohexono- and -heptonolactones the corresponding dibromotetronic acid esters and dibromo/ditosylpentonolactones give aminolactams on treatment with ammonia. These four- and five-carbon aldonic acid esters and lactones are too small to form pyrrolidine and piperidine rings on direct treatment with ammonia, but rather five- and six-membered lactams become thermodynamically favored instead. These reactions also proceed through epoxides, the opening of which determines the stereochemical outcome. With methyl esters of dibromotetronic acids it was necessary to preform the 2,3epoxides in order to obtain acceptable yields in the reaction with ammonia. Thus, dibromo-threonic and -erythronic acid methyl ester 27 and 31 gave the cis-epoxide 28 and the trans-epoxide 32, respectively, when treated with potassium carbonate in acetone [9] (Scheme 5-9). Treatment of cis-epoxide 28 with dry ammonia at 60 0C furnished the 2-aminolactam 29. Additional studies revealed that after formation of amide 28a, substitution of the primary bromide takes place to give

5.2 Syntheses

101

CO2Me ι ομΟΗ acetone Γ~ syrup 98%

CO2Me L Br

CO2Me K

2C°3

L0H^nT cryst 81%

LBr

31

'"

-

d

%

LBr

R " qNH3

^T65%

BHs-Me 2 S

^

VH^HO/^

7^7^

\^

>

32

^

33

JNH3

po "—Br

Scheme 5-9

320

>

\ H0

/

*

34

Vx^ ^\

CONH2

d

N.

<^ ^>=0

CONH2 NHs

-*

CONH2 NHs

4 ^O I— NH2

32b

HO-] H2N^ L-MH2

320

28b. Ring closure then gives epoxylactam 28c which is opened exclusively at C-2 to give aminolactam 29. Reduction of the lactam with borane dimethylsulfide complex afforded the aminopyrrolidine 30. On the other hand, trans-epoxide 32 furnished 3-aminolactam 33 on treatment with liquid ammonia at 9O 0 C. Although the mechanism has not been fully elucidated, the reaction most probably proceeds through amide 32a and aminoamide 32b. Because of the trans-epoxide, cyclization to a lactam cannot occur before the epoxide has been opened with ammonia. This opening occurs predominant at C-3, possibly through the formation and opening of a 3,4-aziridine. Reduction of the lactam then gave the diastereomeric aminopyrrolidine 34 [9]. Dibromo/ditosylpentonolactones also yield aminolactams on reaction with ammonia through a very similar mechanism. Dibromoxylonolactone 35 afforded very cleanly aminolactam 36 with liquid ammonia [10, 20] (Scheme 5-10). Moni13 toring the reaction by C-NMR showed immediate formation of diepoxyamide 35a. Opening of the 4,5-epoxide gave aminoepoxide 35b, which cyclizes to epoxylactam 35c. Opening of the 2,3-epoxide occurs again exclusively at C-2 to afford aminolactam 36. Reduction with borane dimethylsulfide subsequently gave aminopiperidine 37 [1O]. Ditosylribonolactone 38 afforded aminolactam 39 which was somewhat difficult to crystallize [1O]. Again, 13C-NMR showed immediate formation of diepoxyamide 38a which on further reaction with ammonia gave aminoepoxide 38b. Because the epoxide is trans, a lactam cannot be formed before the epoxide has been opened at C-2. Diaminoamide 38c subsequently cyclizes to aminolactam 39. Reduction then yielded aminopiperidine 40 [1O]. Monobromopentonolactones also give lactams on treatment with ammonia. Reaction of 5 -bromoarabinonolactone 41 with aqueous ammonia afforded lactam

102

5 Iminosugars as Powerful Glycosidase Inhibitors. BH3 Me2S cryst. 71%

NY" HO N

/ <

NH

'

cryst. 58%^ {\°" HO NH2 37

1 NH3

—

I NH3

CONH2

CONH2 NH3

NH

X

o

^ /

^y^Q

O^ ^O X

—

— OH -NH 2 35b

3 Sa

HO^^^^^T —

35c

HqNH3

/

cryst. 19%

v!

ΗΟ

BH3Me2S

==

fv

y

cryst. 61°/Γ

(

Ί

NH,

HO

JNH3

CONH2

CONH2 NH

y

ζ

CONH2 NH3

ζ

-NH2 — OH

^o V^

)8a

— OH ^O -NH2 38b

— OH -NH, 38c

Scheme 5-10.

—I

42a isolated as its crystalline isopropylidine derivative 42b [8] (Scheme 5-11). Monitoring the reaction by 13C-NMR revealed immediate formation of bromoamide 41a which was rapidly converted into epoxyamide 41b. Opening of the epoxide gave aminoamide 41c which cyclized to lactam 42a. Reduction of the isopropylidine lactam 42b with sodium borohydride/trifluoroacetic acid then gave iminoarabinitol 43 [8]. This piperidine has previously been prepared in nine steps from methyl a-D-mannopyranoside [21] and four steps from D-arabinose [22].

Μμ

j

aqNH3

NH N== O

/

cryst. 50%

==

Pv

\y

NaBH4 CF

3CO°H

cryst. 66%

a : R =H b : R1R = CMe2

CONH2

CONH2

CONH2

NH3 IH

UOH

LBr

K0

^

""

HO — —OH

-°H

LNH

.

2

Scheme 5-11.

103

5.2 Syntheses

5.2.4

Iminosugars via azidolactones - strategy IV

Although ammonia is very effective as nitrogen source for preparation of iminosugars, it does have a number of limitations. Only certain iminosugars are available using ammonia because all reactions have to proceed through epoxides. Traditionally the most popular nitrogen source for iminosugar synthesis has been the azido group. This can easily be reduced to an amine by many different methods and although it does add steps, the reactions are high-yielding. Azides can also be introduced on dibromoaldonolactones, and by taking advantage of the already developed aminoepoxide ring closure to pyrrolidines/piperidines, new iminosugars can be prepared. Due to the enhanced reactivity of the 2-position in aldonolactones, azides can be regioselectively introduced here by one of two methods. Direct substitution of the bromine at C-2 in dibromoaldonolactones gives 2-azidolactones [23]. In some cases this azide substitution gives rise to both C-2 epimeric azidolactones [24]. To circumvent this problem a 2,3-epoxide can be prepared and then opened with trimethylsilyl azide to give the corresponding 2,3-tran5-2-azidolactone [25, 26]. These two strategies are shown with dibromomannolactone 1 giving rise to azidolactones 44a and 48a (Scheme 5-12). The compounds were not isolated, but subsequently hydrogenated to give crystalline aminolactones 44b and 48b [2, 25], Dissolving each of these aminolactones in excess aqueous potassium hydroxide gave the aminoepoxides 44c and 48c, respectively. However, as observed previously [4, 5, 7], such 5-amino-l,2-epoxide systems are not stable in aqueous base, since they immediately cyclize in a 5-exo mode to give pyrrolidines. Thus, 45 and 49 were formed as the only products [2, 25, 26]. No piperidines, coming from 6-endo opening of the epoxides by the 2-amino group, were observed. Esterification and reduction of the carboxylic acid of 45 and 49 then gave bishydroxymethylpyrrolidines 46 and 50 (Scheme 5-12) [2, 25]. Other methods for preparation of 2,5-dideoxy-2,5-imino-D-glucitol (46) [27] and the corresponding L-iditol (50) [28] have been described. COOKEtOOC

H2NNaN3 -^5J55N*

/IHO^

-^0

ί

L-nw -OH

EtOH

HO

KF

acetone

H2-Pd/C I EtOH, HCI |

46

a: X - N 3 ^b : χ= NH2

COOK— -NH2

^-

H+

HO — — OH

^"

?/

HO

H2-Pd/C ιa: X- N 3 EtOH1HCI I_ ..b.x = NH2

Scheme 5-12.

^^ I

H

HO-K

ΐΝαο,, 4

H0

^

HO/

104

5 Iminosugars as Powerful Glycosidase Inhibitors...

5.3

Biochemical Evaluation

5.3.1 5.3.1.1

l,4-Dideoxy-l,4-iminoalditols (pyrrolidines) Iminotetritols

1,4-Iminotetritols have generally not attracted much attention as glycosidase inhibitors due to the size of the molecules, and their having only two chiral centers. By using strategy III, amino iminotetritols 30 and 34 (Scheme 5-9) together with their enantiomers are easily available [9]. Interestingly, compound 30 inhibited a-omannosidase from jack bean, showing a K1 of 40 μΜ. None of the other enzymes was inhibited by 30 or by any of the three other stereoisomers in the μΜ range. The aminohydroxypyrrolidine 30 inhibits α-mannosidase [9] in the range of the mannosidase inhibitor 1-deoxymannojirimycin (^i 68 μΜ) [29]. 5.3.1.2

2,5-Dideoxy-2,5-iminohexitols

A convenient synthesis using strategy IV (Scheme 5-2) led to the two 2,5-iminohexitols 46 and 50 having Ό-gluco- and L-iifo-configurations, respectively (Scheme 5-12). We have tested the D-g/wco-isomer 46 towards the enzymes in human liver (Table 5-1) and found a strong inhibition towards /?-glucosidase (89%), a-L-fucosidase (86%), and β-xylosidase (86%) [25]. Compound 46 has been synthesized previously [27] and shown to be a competitive inhibitor of a-glucosidase (brewer's yeast) (Ki 2.8 μΜ), /?-glucosidase (almond) (Ki 19 μΜ), and agalactosidase (green coffee bean) (K1 50 μΜ) [27 a,b]. Likewise, the /do-configurated pyrrolidine 50 has previously been synthesized [27d, 28]. Further inhibition data are given in the Tables of chapter 11. 5.3.1.3

l,4-Dideoxy-l,4-iminohexitols

1,4 -Dideoxy-1,4 -iminohexitols prepared by strategies I and II (Schemes 5-2-5-6) are shown collectively in Figure 5-1 together with isomeric compounds obtained by the same strategy. A few of the compounds have previously been tested as inhibitors towards glycosidases. The inhibitory activity of the nine 1,4-iminohexitols shown in Fig. 5-1 have been tested towards enzymes in human liver (Table 5-1) [7], but very few exhibited specific enzyme inhibition. The L-iJo-configured 1,4-iminohexitol 55 was a potent α-D-galactosidase inhibitor while the two with D- (53) or L- (54) galacto configuration did not inhibit any galactosidase. Compound 20, the D-ido-isomer, was a moderate inhibitor of a-L-fucosidase. The most interesting features were the inhibition of various α-mannosidases by 3 (L-allo) and 7 (o-talo). Lysosomal α-mannosidase was inhibited by both compounds (for both Ki - 120 μΜ) [15] and they were both able to alter the metabolism of N-linked glycans in cell cultures in a way distinct from that of swainsonine, another inhibitor of α-mannosidase. From a study of the oligosaccharides that

29

NI

33

NI

[7]

86

NI

[25]

/?-xylosidase

a-L-arabinosidase

References:

NI: no inhibition

27

27

[7]

67

NI

86

20

a-L-fucosidase

/?-hexosaminidase

NI

NI

54

/?-galactosidase

NI

a-galactosidase

NI

89

NI

30

a-glucosidase

/?-glucosidase

NI

NI

10

37

/?-mannosidase

97

NI

55

17

a-mannosidase (pH 6.5)

86

NI

13

NI

a-mannosidase (pH4)

3

51

50

46

Enzyme

[7]

[7]

NI

NI

NI 18

24

20

27

45

NI

NI

NI NI

26

NI

NI 29

NI

15

92 NI

NI

52

56

7

[7]

NI

NI

31

30

NI

[7]

NI

NI

23

16

NI

27

17

NI

48

11

NI

NI

NI

53

14

10

13

NI

15

[7] [7]

23

NI [7]

62

NI

NI

NI

NI

69

35

NI 27

95

NI

NI

NI

NI

NI

55

NI

13

12

10

NI

27

13

NI

NI

23

36

NI

NI

NI

NI NI

20

54

[8]

83

NI

NI

NI

NI

61

NI

NI

[8]

NI

20

97

NI

NI

38

NI

NI

NI

NI

22 20

43

58

[8]

NI

NI

NI

NI

NI

97

18

NI

37

NI

59

Inhibitor (Figure 5-1 and Figure 5-2)

[8]

23

NI

20

NI

NI

76

NI

NI

NI

NI

NI

60

[10]

[10]

NI

NI

21 29

20

NI

NI

NI

12

NI

NI

NI

NI

37

NI

NI

48

NI

11

NI

NI

NI

NI

40

[8]

NI

NI

33

90

NI

NI

10

NI

NI

NI

NI

10

[8]

NI

22

35

84

43

NI

32

NI

NI

NI

NI

23

[8]

NI

NI

26

91

65

NI

NI

NI

NI

NI

NI

26

Table 5-1. Glycosidase inhibitor activities of 1,4 -dideoxy-1,4 -iminohexitols and 1,5-dideoxy-1,5-iminopentitols and -heptitols (Figure 5-1 and Figure 5-2) towards enzymes in human liver (percentage inhibition [I] 1.0 mM, [S] 0.5 mM).

106

5 Iminosugars as Powerful Glycosidase Inhibitors.,

& L-aulo 15 [7]

54 [5]

20 [5]

55 [5]

Figure 5-1.

accumulate in cells in the presence of 3 and 7 it appears that they inhibit lysosomal α-mannosidase rather than the processing α-mannosidases I and II. Both compounds may have application as selective inhibitors of intracellular a-mannosidases, while the lack of a similar effect of 15 is surprising. Compound 15 (L-gulo) differs from the strong α-mannosidase inhibitor, l,4-dideoxy-l,4-iminoD-mannitol only by the configuration at C-5 [7].

5.3.2

l,5-Dideoxy-l,5-iminoalditols (piperidines)

56 [11]

60 [8]

i\_ /

0

> - NHΓ OH 40[1O]

2

37[1O]

61[1O]

62[1O]

ι— OH — OH

J-H
HO ""

HO "γββ^^^

HO \^^^^/

OH

Figure 5-2.

5.3 Biochemical Evaluation 5.3.2.1

107

l,2,5-Trideoxy-l,5-iminopentitols

The two 2-deoxy-l,5-iminopentitols 56 and 57 (Figure 5-2) have been synthesized using strategy III [11] from the corresponding 5-bromo-2,5-dideoxypentono-l,4lactones [30] and they have been tested towards commercially available glycosidases. It was found that l,2,5-tndeoxy-O-erythro-l,5-immopenutol (56) inhibited /?-glucosidase (almonds) with a K1 of 6 μΜ and the corresponding Ό-threo isomer the same enzyme with a ^ of 64 μΜ [U]. Neither 56 nor 57 inhibited a-glucosidase, but the t/zreo-isomer 57 inhibited /?-mannosidase with K1 20 μΜ. None of the other enzymes tested was inhibited to any extent (Table 5-2). 5.3.2.2

1,5-Dideoxy-1,5-iminopentitols

All four possible stereoisomeric 1,5-iminopentitols 58, 43, 59 and 60 (Figure 5-2) were synthesized conveniently using strategy III (Scheme 5-11) [8]. The O-arabino- (43) [30], the xylo (59) [21,31] and the L-arabino (60) [21] iminopentitols were known compounds, while details of the nfco-isomer (58) were published [33] simultaneously with our paper [8]. The four compounds (Figure 5-2) were tested towards human liver enzymes (Table 5-1). Interestingly, 59, the meso iminopentitol with xy/o-configuration was a potent /?-glucosidase inhibitor (97 %), having the structural motif of glucose. Altering the configuration at C-3, the ribo-analog 58, also a meso compound, was a somewhat weaker inhibitor of the same enzyme (61 %) (Table 5-1). For comparison we tested the four compounds towards commercially available glycosidases (Table 5-2). The jty/o-isomer 59 inhibited βglucosidase from almonds with a K1 of 34 μΜ while the n&o-isomer 58 showed a somewhat stronger inhibition (K1 12 μΜ) towards the same enzyme, similar to the value reported elsewhere (ICso 8.8 μΜ, [32]). Comparison between data for inhibition of /?-glucosidase from human liver (Table 5-1) and from almonds (Table 5 -2) reveals the xylo isomer 59 to be a more efficient inhibitor in the former system while the ribo isomer 58 inhibits the almond /?-glucosidase more strongly. The values in Table 5-2 furthermore show that the 2-deoxy-O-threo- analog 57 was an inhibitor having a K1 in the same range as 59, indicating that the 2-OH group might be less important for the activity with almond /?-glucosidase. l,5-Dideoxy-l,5-imino-L-arabinitol (60) showed moderate inhibition (76%, Table 5-1) towards α-galactosidase in human liver. Compound 60, having the same configuration at C-2, C-3 and C-4 as D-galactose, was a stronger inhibitor of α-galactosidase from green coffee beans with K\ 15 μΜ (Table 5-2). The deoxygalactostatin inhibits the latter enzyme with K{ 1.6 μΜ [29]. In contrast, the iminoL-arabinitol (60) inhibited the /?-galactosidase only moderately (^i 56 μΜ) whereas the ribo analog 58 was a stronger inhibitor of this enzyme (Ki 2 μΜ) (Table 5-2). Also the O-arabino (43) and the xylo (59) isomers inhibited /?-galactosidase moderately (Table 5-2). The /?-mannosidase was inhibited moderately by three of the four isomeric 1,5-iminopentitols: Ό-arabino (o-lyxo) (43), the xylo (59) and the L-arabino (60) with K1 53, 43, and 33 μΜ, respectively. Furthermore Table 5-1 shows that 43 was the only one of the four isomeric 1,5-iminopentitols which inhibited α-L-fucosidase from human liver (97 %) confirming the O-arabino-con-

NI 6 NI NI NI NI NI [11]

α-glucosidase (baker's yeast)

/?-glucosidase (almonds)

α-mannosidase (jack bean)

/?-mannosidase (snail)

α-galactosidase (green coffee bean)

/?-galactosidase (E. coli)

α-L-fucosidase (bovine kidney)

References:

NI: no inhibition: K1 > 100 μΜ

56

[U]

NI

NI

NI

20

NI

64

NI

57

[10]

NI

2

NI

NI

NI

12

NI

58

[10]

30

67

NI

53

NI

NI

NI

43

[10]

NI

71

NI

43

NI

[10]

NI

56

15

33

NI

NI

NI

NI 34

60

59

[10]

NI [10]

NI

NI NI

NI

68

NI

NI

NI

37

NI

56

NI

3

NI

40

Inhibitor (Figure 5-2)

Glycosidase inhibitor activities (Xi/μΜ) of l,5-dideoxy-l,5 -iminoalditols.

Enzyme

Table 5-2.

[10]

NI

62

72

19

NI

NI

NI

61

[10]

NI

NI

57

12

NI

NI

NI

62

[25]

3

29

NI

NI

NI

NI

NI

10

[25]

16

23

NI

NI

NI

NI

NI

23

[25]

9

30

NI

NI

NI

NI

NI

26

δ >ί

IS;

IS

3

O

3 ^ Co

O

IS"

O,

O OO

5.4 Conclusions

109

figuration as the minimum structural motif necessary for inhibition of this enzyme ([8] and references cited therein). Likewise, l,5-dideoxy-l,5-imino-D-arabinitol (43) inhibited α-L-fucosidase from bovine kidney (Ki 30 μΜ) (Table 5-2). The 2-amino-l,5-imino-l,2,5-trideoxy-D-ribitol (40) and -o-xylitol (37), together with the O-arabino- (61) and L-.ry/0-isomer (62) (Fig. 5-2) have been synthesized using strategy III (Scheme 5-10) [1O]. The compounds 40 and 37 were tested towards human liver enzymes, but did not exhibit any significant glycosidase inhibitory activities (Table 5-1). In contrast, the ri&o-isomer 40 was found to be a potent inhibitor of /?-glucosidase from almonds (Ki = 3 μΜ) [1O]. Both 40 and the .ryfo-analogue 37 showed an inhibition of /?-mannosidase with Ki 56 and 68 μΜ, respectively (Table 5-2). 2-Amino-l,5-imino-D-arabinitol (61) inhibited /?-mannosidase (Ki 19 μΜ) and α-galactosidase (K1 72 μΜ). In addition, the L-xylo isomer 62 also inhibited these two enzymes with Ki 12 μΜ and 57 μΜ, respectively. Differences between the inhibition of human liver enzymes and other enzymes tested were again found when comparing the similar glycosidases. However, the inhibitory activities measured towards human liver enzymes are more important with regard to identifying specific inhibitors for enzymes involved in human diseases. 5.3.2.3

l,5-Dideoxy-l,5-iminoheptitols

Using strategies I or II, the three 1,5-iminoheptitols (Figure 5-2) 10 (Scheme 5-4), together with 23 and 26 (Scheme 5-8) have been synthesized and tested as glycosidase inhibitors. As anticipated all were powerful inhibitors (> 85 % inhibition) of the human liver a-L-fucosidase [8] (Table 5-1). Our anticipation was based upon the absolute configuration of the hydroxyl-bearing carbon atoms of the compounds, all being O-arabino [8]. They all inhibited this enzyme to the same extent as the 1,5-imino-D -arabinitol 43 discussed above, indicating that substitution of 43 with a 1,2-dihydroxyethyl group either at C-I (10 and 26) or at C-5 (23) does not influence the inhibitory potency (Table 5-1). We also found that a-L-fucosidase from bovine kidney was inhibited strongly (Table 5-2) (K1 3, 16 and 9 μΜ, respectively). The three iminoheptitols were moderate inhibitors towards /?-galactosidase (Ki 29, 23, and 30 μΜ, respectively).

5.4

Conclusions

The synthesis of iminosugars from regioselectively activated aldonolactones [2] has proved to be very efficient, and has given rise to 28 different types of iminosugars without the use of any protecting group strategy. The iminosugars prepared range from amino/hydroxy-substituted tetritols to heptitols including both pyrrolidine as well as piperidine types of compounds. Many of the compounds were tested towards enzymes in human liver as well as towards commercial enzymes and results are shown collectively in Tables 5-1 and 5-2. Several novel inhibitors

110

5 Iminosugars as Powerful Glycosidase Inhibitors .

β-glucosidase:

56

H

6

58

OH

12

40

NH2

3

β-mannosidase:

57

H

61

20

OH

62 OH

19

H

12

a-L-fucosidase:

K, (μΜ) 58

2

R1

R2

H

43 23

^

OH OH

26

H

10

H

H

HO" °H

H

H

R3

K1 (μΜ)

H

30

H

16

H

tgd

9 3

-.ρ,.

-

^,

^,

Figure 5-3. Comparison of Structures of Strong Glycosidase Inhibitors (Table 5-2) [10, 11, 25].

have been discovered, particularly of /?-D-glucosidase, /?-D-mannosidase and a-Lfucosidase. Figure 5-3 shows a comparison of the structures of the most remarkable inhibitors synthesized.

References [1] See for example, A. de Raadt, C. W. Ekhart, M. Ebner, A. E. Stiitz in, Top. Curr. Chem., (Ed. H. Driguez, J. Thiem), Springer-Verlag, Heidelberg, 1997, 787, 157-186; C. H. Wong, R. L. Halcomb, Y. Ichikawa, T. Kajimoto, Angew. Chem., 1995, 702, 453-474; Angew. Chem. Int. Ed. EngL, 1995, 34, 412-432; L. A. G. M. van den Broek, D. J. Vermaas, B. M. Heskamp, C. A. A. van Boeckel, M. C. A. A. Tan, J. G. M. Bolscher, H. L. Ploegh, F. J. van Kemenade, R. E. Y. de Goede, E Miedema, Reel. Trav. Chim. Pays Bas, 1993, 772, 82; K. Burgess, I. Henderson, Tetrahedron, 1992, 48, 4045-4066. [2] I. Lundt in, Top. Curr. Chem., (Ed.: H. Drigues, J. Thiem), Springer-Verlag, Heidelberg, 1997, 187, 117-156. [3] R. M. de Lederkremer, O. Varela, Adv. Carbohydr. Chem. Biochem., 1994, 50, 125-209; G. W. J. Fleet in, Antibiotics and Antiviral Compounds: Chemical Synthesis and Modification (Eds.: K. Krohn, H. A. Kirst, H. Maas), VCH, Weinheim, 1993, p. 333-342. [4] I. Lundt, R. Madsen, Synthesis, 1993, 714-720. [5] I. Lundt, R. Madsen, Synthesis, 1993, 720-724. [6] I. Lundt, R. Madsen, Synthesis, 1995, 787-794. [7] I. Lundt, R. Madsen, S. Al Daher, B. Winchester, Tetrahedron, 1994, 50, 7513-7520. [8] M. Godskesen, I. Lundt, R. Madsen, B. Winchester, Bioorg. Med. Chem., 1996, 4, 1857-1865. [9] G. Limberg, I. Lundt, J. Zavilla, Synthesis, 1998, in press. [10] M. Godskesen, I. Lundt, Submitted.

References

111

[11] L. Hyldtoft, M. Godskesen, I. Lundt, Submitted. [12] I. Lundt, H. Frank, Tetrahedron, 1994, 50, 13285-13298; H. KoId, I. Lundt, C. Pedersen, Acta Chem. Scand., 1994, 48, 675-678. [13] V. Bouchez, I. Stasik, D. Beaupere, R. Uzan, Carbohydr. Res., 1997, 300, 139-142. [14] I. Lundt, R. Madsen, Synthesis, 1992, 1129-1132. [15] S. Al Daher, G. W. J. Fleet, S. K. Namgoong, B. Winchester, Biochem. J., 1989, 258, 613-615. [16] J. G. Buchanan, K.W. Lumbard, R. J. Sturgeon, D. K. Thompson, R. H. Wightman, /. Chem. Soc., Perkin Trans. 1, 1990, 699-706. [17] H. Setoi, H. Kayakiri, H. Takeno, M. Hashimoto, Chem. Pharm. Bull, 1987, 35, 3995-3999. [18] G. W. J. Fleet, J. C. Son, D. S. C. Green, I. Cenci di Bello, B. Winchester, Tetrahedron, 1988, 44, 2649-2655. [19] G. N. Austin, P. D. Baird, G. W. J. Fleet, J. M. Peach, P. W. Smith, D. J. Watkin, Tetrahedron, 1987, 43, 3095-3108. [20] M. BoIs, I. Lundt, Acta Chem. Scand., 1991, 45, 280-284. [21] R. C. Bernotas, G. Papandreou, J. Urbach, B. Ganem, Tetrahedron Lett., 1990, 31, 3393-3396. [22] G. Legler, A. E. Stutz, H. Immich, Carbohydr. Res., 1995, 272, 17-30. [23] K. Bock, I. Lundt, C. Pedersen, Acta Chem. Scand., 1987, B41, 435-441. [24] M. BoIs, I. Lundt, Acta Chem. Scand., 1988, B42, 67-74. [25] I. Lundt, unpublished results. [26] G. Mikkelsen, T. V. Christensen, M. BoIs, I. Lundt, Tetrahedron Lett, 1995, 36, 6541-6544. [27] a) T. Kajimoto, L. Chen, K. K.-C. Liu, C.-H. Wong, J. Am. Chem. Soc., 1991, 113, 6678-6680; b) K. K.-C. Liu, T. Kajimoto, L. Chen, Z. Zhong, Y. Ichikawa, C.-H. Wong, J. Org Chem., 1991, 56, 6280-6289; c) G. Legler, A. Korth, A. Berger, C. Ekhart, G. Gradnig, A. E. Stutz, Carbohydr. Res., 1993, 250, 67-77; d) A. B. Reitz, E. W. Baxter, Tetrahedron Lett., 1990, 31, 6777-6780. [28] A. Dureault, M. Portal, J. C. Depezay, Synlett, 1991, 225-226; T. K. M. Shing, Tetrahedron, 1988, 44, 7261-7264. [29] G. Legler, Adv. Carbohydr. Chem. Biochem., 1990, 48, 319-384. [30] K. Bock, I. Lundt, C. Pedersen, Carbohydr. Res., 1981, 90, 17-26. [31] H. Paulsen, F. Leupold, K. Todt, Liebigs Ann. Chem., 1966, 692, 200-214. [32] Y. Igarashi, M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem. Lett., 1996, 6, 553-558.

6

Isoiminosugars: Glycosidase Inhibitors with Nitrogen at the Anomeric Position INGE LUNDT and ROBERT MADSEN

6.1

The Anomeric Position

The success of iminosugars as glycosidase inhibitors has been largely attributed to their resemblance to the transition state for glycoside cleavage [1]. However, the exact nature of this transition state continues to be the subject of much discussion [2]. The mechanism for enzymatic hydrolysis of glycosides has been thoroughly studied [3]. Although some inverting glycosidases are known, most glycosidases are retaining and yield the product, glycose, in the same anomeric configuration as the starting glycoside. This implies a double displacement mechanism at the anomeric center and the formation of an intermediate covalent glycosyl-enzyme species. In all cases examined so far this species is a glycosyl ester, formed by acid-catalyzed substitution at the anomeric center with an enzyme carboxylate from an aspartate or glutamate residue. This acyl species is then hydrolyzed by a base-catalyzed substitution, giving rise to the glycose product and returning the enzyme to its original protonation state (Figure 6-1). Both the formation and the hydrolysis of the glycosyl species proceed through distorted oxo-carbonium ion-like transition states [3, 4]. Recent mechanistic studies on several ^-glycosidases indicated substantial bond cleavage and positive charge development at the anomeric center in these two transition states [5]. These observations have to be compared with studies on the nonenzymatic, acid-catalyzed hydrolysis of glycosides. It is generally accepted that protonation occurs reversibly at the exocyclic oxygen atom, followed by Cl-oxygen bond breaking [6]. A detailed study of kinetic isotopic effects has shown that the resulting positive charge at Cl on the glycon is not completely delocalized by one of the ring oxygen lone pair of electrons [7], the reduced orbital overlap produced being largely attributed to the constraints imposed by the pyran ring. As a result, there oppear to be reasons for believing that iminosugars with nitrogen at the anomeric position will also function as glycosidase inhibitors. Such iminosugars, here termed isoiminosugars, are defined by moving the ring nitrogen in iminosugars to the anomeric position (Figure 6-2). Iminosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

6.1 The Anomeric Position

113

Figure 6-1. Catalytic mechanism for retaining glucosidases.

X = C-subst. Y = OH or H Z = NHAc or H

Figure 6-2.

Isoiminosugars are able to develop a positive charge at the anomeric position and should therefore also be good transition-state mimics and able to form an ion pair with the enzyme carboxylate. Only six-membered isoiminosugars will be discussed here. Five-membered isoiminosugars have been reported, but did not show any significant inhibition of glycosidases [8]. The first reported isoiminosugar was siastatin B 1, isolated from a micro-organism in 1974 and shown to be a strong inhibitor of N-acetylneuraminidase [9]. Many siastatin B derivatives have since been prepared and tested as glycosidase inhibitors. However, it was not until 1994 that the full potential of placing the nitrogen at the anomeric center was realized following the synthesis of isofagomine 2, which to date is the most potent inhibitor of /?-glucosidase from almonds. [10, 11] (Figure 6-3). Since then a number of different isoiminosugars have been prepared, several of which have shown remarkable inhibition of ^-glycosidases. This is in contrast to normal six-membered iminosugars which typically inhibit α-glycosidases very well, but not the corresponding /?-glycosidases.

114

HO

6 Isoiminosugars: Glycosidase Inhibitors "with Nitrogen.

J

~Ί

K^

/

Η

Χ

<(°»

™

NHAc

Siastatin B

lsofagomine

ι

2

6.2

Figure 6-3.

Siastatin B and Analogs

At the time of its isolation from Streptomyces in 1974 [9], only the relative configuration of siastatin B was determined. The absolute configuration was established by Nishimura et al. in 1988 through total synthesis from D-ribonolactone 3 [12]. The synthesis involved a total of 24 steps and is the only synthesis of siastatin B (Scheme 6-1) known to date. The key steps involved formation of the piperidine ring, creation of the aminal structure, and introduction of the carboxyl group. The piperidine ring was formed by ring expansion of azidolactone 4 to give lactam 5. The aminal structure was created by Mitsunobu reaction of hemiaminal 6 with

BSO

Ν

/0Η

PMhNH

TBSO

3 steps

\

Λ

Scheme 6-1.

—»· -^

Synthesis of siastatin B.

—*-

j /

Nx

\OH

s

\\

6.2 Siastatin B and Analogs

115

phthalimide. The final introduction of the carboxyl group was carried out through a Nef reaction of nitro compound 10. The generated aldehyde was not purified, but subsequently oxidized further to the carboxylic acid, isolated as the MEM ester 12. Reduction with sodium borohydride gave alcohol 13 which was converted into siastatin B in the final three steps. The same route was used to prepare the enantiomer of siastatin B [12]. The total synthesis of siastatin B provides a number of intermediates that may be converted into analogs of siastatin B. By using these intermediates, Nishimura etal. have prepared more than 40 analogs of siastatin B, and tested these as glycosidase inhibitors [13-2O]. Compounds 14 to 24 (Figure 6-4) constitute some of the more potent analogs that have shown strong inhibition of several glycosidases (Table 6-1). For inhibition of N-acetylneuraminidase, the presence of both the acetamide and carboxyl group in siastatin B 1 is important, as alkylation of the amino group enhances the activity. If the acetamide in siastatin B is replaced by a trifluoroacetamide several very potent inhibitors of /?-glucuronidase are obtained. In addition, compounds 18 and 19 showed inhibitory activity for tumor metastasis in mice [17, 21]. These trifluoroacetamide analogs also inhibit a- and /?-glucosidases. Reducing the carboxyl group in siastatin B to a hydroxymethyl group creates a very potent inhibitor of N-acetyl-/?-glucosaminidase and N-acetyl-agalactosaminidase. However, inverting the stereochemistry of the two hydroxy groups or the acetamide generally gives less inhibition of the glycosidases. Y. Nishimura et al have also prepared analogs 25 and 26 (Figure 6-5) which showed no activity against the glycosidases in Table 6-1 [22, 23]. However, 25

COOH HO

OH

XT"

NH

/"

J

ν

Y OH

N

(

jJjHAc

NHAc

14

N—Bn

XT/

OH

NHAc

15

16

O2N

OH

N

/

NH

HOHO ' ' OH

NH

<( OH

/

NHAc

22

^

NH

NHAc

23

Figure 6-4. Siastatin analogs as glycosidase inhibitors (Table 6-1).

>100

[12-14]

/?-Amylase (sweet potato)

References:

NI: no inhibition

15.5

/?-Glucuronidase (bovine liver)

13

α-Glucosidase (baker's yeast) >100

>100

NI

Af-Acetyl-a-galactosaminidase >100 (Chicken liver)

^-Glucosidase (almonds)

NI

>100

7V-Acetyl-/?-glucosaminidase (Bovine epididymis)

[13]

>100

12

5.3

[13]

64

>100

12

[15]

18

NI

NI

4

36

1.3

Λ^-Acetyl-^-glucosaminidase (Equine kidney)

1.8

3.5

16

6.3

14

15

Af-Acetylneuraminidase (Streptococcus)

20

14

50

1

W-Acetylneuraminidase (Clostridium perfrigens)

Enzyme

[16]

NI

28.5

NI

NI

NI

NI

17

[17, 18]

0.008

19

40

NI

18

[19]

>100

0.02

>100

7.7

NI

NI

19

Inhibitor (Figure 6-4)

>100 10

>100 16.8

[19]

>100

60

[19]

1.9

NI

NI

2.2

NI

21

NI

20

Table 6-1. Glucosidase inhibitor activities (ICso^g/mL) of siastatin B (1) and analogs (14-24).

>100

>100

[14]

8.6

>100

[14]

>100

>100

0.27 >100

>100

23

0.42

22

[16]

>100

NI

>100

2.5

NI

NI

24

^

OQ

δ·

OQ

ON

117

6.3 Isofagomine and Beyond

I NH

^^

Figure 6-5.

COOH

HO

J

KoH HO

NHAc

, AcNH NH

NL-J/

Figure 6-6. showed marked inhibitory activity against influenza virus N-acetylneuraminidases and significant inhibition of influenza virus infection in vitro [22]. Compound 26 had significant inhibitory activity against experimental pulmonary metastasis in mice at concentrations where cell growth was not affected [23]. Recently, three novel diastereomers of siastatin B were isolated from the culture filtrate of Streptomyces mobilis SANK 60192 [24]. Only the relative configurations of 27, 28, and 29 have been determined (Figure 6-6). However, because SANK 60192 also produces siastatin B, compounds 27-29 are assumed to be epimers of siastatin B [24]. While 27 and 28 showed no significant inhibitory activity, 29 showed marked inhibition against bovine liver /?-glucuronidase and tumor cell heparanase with ICso values of 1.6 and 12 μΜ, respectively [25].

6.3 Isofagomine and Beyond The discovery of isofagomine in 1994 as one of the most potent /?-glucosidase inhibitors resulted in an increased interest in glycosidase inhibitors with nitrogen at the anomeric position. The synthesis of isofagomine 2 was accomplished in 10 steps from levoglucosan 30 using a ring-closing reductive amination of dialdehyde 34 with ammonia as the key step [10] (Scheme 6-2). Isofagomine shows significant inhibition towards several glycosidases, in contrast to fagomine which shows only weak inhibition of some a-glucosidases [26]. When compared with deoxynojirimycin, isofagomine is a slightly weaker inhibitor of α-glucosidase, but a much stronger inhibitor of /?-glucosidase (Table 6-2). In fact, isofagomine is currently the most potent inhibitor of /?-glucosidase from almonds reported to date [U]. Placing nitrogen at the anomeric position seems greatly to enhance inhibition against certain glycosidases. To examine this idea further Ichikawa et al. have prepared a number of isoiminosugars and tested them as glycosidase inhibitors. The

118

0

6 Isoiminosugars: Glycosidase Inhibitors with Nitrogen.

Η0

η

35

Η0

η

η

dcneme o-z.

2

Isofagomine

Fagomine

Synthesis of isofagOITime.

Table 6-2. Glycosidase inhibitor activities (Κ^/μΜ [U]) of isofagomine, deoxynojirimycin, and deoxymannojirimycin. Enzyme

Inhibitor „0

ΗΟ-η

/ /OH

X

N

Η0\β^β^

Isofagomine α-Glucosidase (yeast)

86

HO^

J

NH

\

H

NJ°

f/

HO ^L··.---.-· CDH

Deoxynojirimycin

/

NI

0.11

47

300

7.2

11

490

Glucoamylase (Asp. awamori) α-Mannosidase (jack bean)

3.7

270

\

Deoxymannojirimycin

/?-Glucosidase (almonds) Isomaltase (yeast)

770

NH

H

k? ?y HONLJ/

25

9.8

J

66 280

NI: no inhibition

first reported inhibitor was 5 -hydroxyisof agomine 39 [27], which was prepared from diisopropylidenemannofuranose 36 in a total of seven steps using a reductive amination of benzyl azidofuranoside 38 as the key step [27] (Scheme 6-3). Using a similar overall strategy, Ichikawa et al. have also prepared isoiminosugars 40-43 [28-31] (Figure 6-7 and 6-8).

119

6.3 Isofagomine and Beyond

5 steps

Scheme 6-3.

HOHO

Figure 6-7. Analogs of isofagomine as glycosidase inhibitor (Table 6-3).

/

\

N

HU

/I

OH

/

\

\

/n

Λ \ JxL—/" H0 , HO

When tested as glycosidase inhibitor, 5 -hydroxyisofagomine 39 proved to be a strong inhibitor of /?-glucosidase (^i 4.3 μΜ) and a weaker inhibitor of a-glucosidase (Table 6-3). This follows the same pattern as observed with isofagomine, although the inhibition is weaker. The galacto analog of isofagomine 40 was an extremely potent inhibitor of /?-galactosidase (Ki 4.1 μΜ) and a weak inhibitor of α-galactosidase. This is in contrast to deoxygalactostatin (Figure 6-8) which is a very powerful inhibitor of coffee bean α-galactosidase (K1 1.6 μΜ), but a weaker inhibitor of several /?-galactosidases [32]. Galacto analog 40 was also a very potent inhibitor of /?-glucosidase in the same range as isofagomine. The 5-hydroxy analog 41 was a much less potent inhibitor than 40. As also observed with 39 the 5hydroxy group seems to decrease binding to the enzyme active site. Fucose analog 42 which has also been prepared by BoIs et al. [33], was an inhibitor of a-L-fucosidase (Ki 8.4 μΜ). In contrast to the other isoiminosugars, 42 is a weaker inhibitor than the parent deoxyfuconojirimycin, which displayed K1 0.029 μΜ against bovine kidney a-L-fucosidase [34]. Glucuronic acid-derived isoiminosugar 43 was a potent inhibitor of /?-glucuronidase (bovine liver, K1 0.079 μΜ), but has not been tested against other glycosidases. The inhibition is 1000-fold stronger than that reported for naturally occurring uronic acid analog 44 (Figure 6-8) against human liver /?-glucuronidase (Ki 80 μΜ) [35]. COOH

((OH

iN_/

H

Deoxygalactostatin

Figure 6-8.

Deoxyfuconojirimycin

NH

120

6 Isoiminosugars: Glycosidase Inhibitors with Nitrogen...

Table 6-3. Glycosidase inhibitor activities (ICso/μΜ) of isofagomine analogs. Inhibitor (Scheme 6-3 and Figure 6-7)

Enzyme

α-Glucosidase (baker's yeast) /?-Glucosidase (almonds)

39

40

41

42

380

NI

NI

NI

420

500

610

NI

0.19

6

α-Galactosidase (coffee beans)

NI

/?-Galactosidase (Asp. oryzae)

NI

α-Mannosidase (jack beans)

NI

200 0.012

17.5

NI

NI NI

α-L-Fucosidase (bovine kidney)

26

References:

[27]

[28]

[29]

[30]

NI: no inhibition

Π ^°\ 1

O=0

Br—-

NHs

/

\

- Η0^_>0

BH 3 -SMe 2

Χ~

\

Ήθ<ί-->

η

Scheme 6-4.

One of the most efficient general procedures for the synthesis of iminosugars has been the ring closure of functionalized aldonolactones. By branching at C-2 of unprotected aldonolactones or of the ring-closed 1,5-lactams, this procedure can also be extended to preparation of isoiminosugars. Thus, ring closure of the 5-bromo-2,5-dideoxy-D-pentonolactones 45 and 48 [36] using ammonia gave the 2-deoxylactams 46 and 49, respectively [37] (Scheme 6-4). Alkylation at C-2 of these unprotected lactams did not proceed satisfactorily, but the lactams were reduced to the iminosugars 47 and 50, respectively, and, for comparison, tested as glycosidase inhibitors (Table 6-4). In contrast, alkylation of the unprotected 5-bromolactones 45 and 48 proceeded smoothly to give the C-2-methylated lactones 51 and 54, respectively, with high stereoselectivity [37] (Scheme 6-5), the 3-OH and the incoming electrophile being trans oriented. Ring expansion via reduction of the corresponding C-5-azidolactone gave the crystalline C-2-methylated lactams 52 and 55, respectively. Reduction of the lactam function afforded methylated piperidines, the 6-deoxyisofagomine (53), with D-jcyfo-configuration, and the isomeric compound having L-rito-configuration (56) [37].

6.3 Isofagomine and Beyond

X

^4

Si. _ I

I (D

N

ω

-O

8 S S^ 28 ^

1 I 1 " §

C-

121

122

6 Isoiminosugars: Glycosidase Inhibitors with Nitrogen...

Br—ι

then MeI

ο

. T^ ^>= \ QH Γ~ 53

6-Deoxy-isofagomine

Scheme 6-5. Synthesis of 6-deoxyisofagomine and analog.

The isoiminosugars 53 and 56 and the iminosugars 47 and 50 were tested towards the commercial glycosidases (Table 6-4). It was found [37] that 6-deoxyisofagomine (53) inhibited /?-glucosidase with K1 3 μΜ and 56 the same enzyme with Ki 2 μΜ, which might be somewhat surprising for 56, having L-ribo-configuration. Interestingly, neither compound inhibited α-glucosidase. Thus, both isoiminosugars showed the same inhibition profile, although not the same potency, as isofagomine. The C-5 isomer of 56, the D-/y;c0-configured isoiminosugar, has been synthesized by Wong and co-workers using aldolase coupling reactions [38], but to our knowledge has not been tested as a glycosidase inhibitor. Among the other enzymes tested, the isofagomine analog 53 inhibited a-L-fucosidase to the same degree as the analog with L-/yxo-configuration (42) (compare Tables 6-3 and 6-4), while the L-ri&o-isomer 56 inhibited α-mannosidase with K1 of 86 μΜ. Interestingly, the two iminosugars with O-threo- (47)- and D-eryi/iro-configuration (50) both inhibited /?-glucosidase with K1 64 μΜ and 6 μΜ, respectively, 50 thus being an inhibitor in the same range as 6-deoxyisofagomine. The syntheses of isofagomine (Scheme 6-2) and of other isoiminosugars (Scheme 6-1) suffer from the many steps involved, including necessary protecting/deprotecting steps. Our approach [37] to isoiminosugars involved simple alkylation of easily available C-5 activated unprotected 2-deoxyaldono-l,4-lactones [39]. This method can be extended by use of other electrophilic reagents.

References

6.4

123

Conclusions

The most striking feature observed when comparing iminosugars and isoiminosugars as glycosidase inhibitors is the effectiveness by which isoiminosugars (piperidine analogs) inhibit their corresponding /?-glycosidases in favor of the similar aglycosidases. Iminosugars, both the piperidine as well as the pyrrolidine analogs, inhibit the α-glycosidases to a greater extent than the corresponding /?-glycosidases. This striking observation might reflect a significant difference in the charge distribution in the transition states for the enzyme-catalyzed hydrolysis of β- and a-glycosides. Although these observations relating to isominosugars might help to shed some light upon the detailed mechanism and the transition state of the enzyme-catalyzed cleavage of glycosides, more studies are still needed to obtain a deeper understanding of this important biological process [40, 41].

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

[11] [12] [13] [14] [15] [16] [17] [18]

G. Legler, Adv. Carbohydr. Chem. Biochem., 1990, 48, 319-384. M. L. Sinnott, Adv. Phy. Org. Chem., 1988, 24, 113-204. M. L. Sinnott, Chem. Rev., 1990, 90, 1171-1202. J. B. Kempton, S. G. Withers, Biochemistry, 1992, 31, 9961-9969. M. N. Namchuk, S. G. Withers, Biochemistry, 1995, 34, 16194-16202; D. TuIl, S. G. Withers, Biochemistry, 1994, 33, 6363-6370. M. L. Sinnott, Bioorg. Chem., 1993, 21, 34-40. A. J. Bennet, M. L. Sinnott, J. Am. Chem. Soc., 1986, 108, 7287-7294. M. BoIs, M. P. Persson, W. M. Butt, M. J0rgensen, P. Christensen, L. T. Hansen, Tetrahedron Lett., 1996, 37, 2097-2100. H. Umezawa, T. Aoyagi, T. Komiyama, H. Morishima, M. Hamada, T. Takeuchi, J. Antibiot., 1974, 12, 963-969. T. M. Jespersen, W. Dong, M. R. Sierks, T. Skrydstrup, I. Lundt, M. BoIs, Angew. Chem., 1994, 106, 1858-1860; Angew Chem. Int. Ed. Engl, 1994, 33, 1778-1779; T. M. Jespersen, M. BoIs, M. R. Sierks, T. Skrydstrup, Tetrahedron, 1994, 50, 1344913460. W. Dong, T. Jespersen, M. BoIs, T. Skrydstrup, M. R. Sierks, Biochemistry, 1996, 35, 2788-2795. Y. Nishimura, W-M. Wang, S. Kondo, T. Aoyagi, H. Umezawa, J. Am. Chem. Soc., 1988, 110, 7249-7250; Y. Nishimura, W-M. Wang, T. Kudo, S. Kondo, Bull. Chem. Soc. Jpn., 1992, 65, 978-986. T. Kudo, Y. Nishimura, S. Kondo. T. Takeuchi, J. Antibiot., 1993, 46, 300-309. Y. Nishimura, T. Satoh, T. Kudo, S. Kondo, T. Takeuchi, Bioorg. Med. Chem., 1996, 4, 91-96. T. Kudo, Y Nishimura, S. Kondo, T. Takeuchi, J. Antibiot., 1992, 45, 1662-1668. T. Kudo, Y Nishimura, S. Kondo, T. Takeuchi, J. Antibiot., 1992, 45, 954-962. Y Nishimura, T. Kudo, S. Kondo, T. Takeuchi, T. Tsuruoka, H. Fukuyasu, S. Shibahara, J. Antibiot., 1994, 47, 101-107. T. Satoh, Y. Nishimura, S. Kondo, T. Takeuchi, Carbohydr. Res., 1996, 286, 173-178.

124

6 Isoiminosugars: Glycosidase Inhibitors with Nitrogen...

[19] Y. Nishimura, T. Kudo, S. Kondo, T. Takeuchi, /. Antibiot., 1992, 45, 963-970. [20] T. Satoh, Y. Nishimura, S. Kondo, T. Takeuchi, M. Azetaka, H. Fukuyasu, Y. lizuka, S. Ohuchi, S. Shibahara, J. Antibiot., 1996, 49, 321-325. [21] Y. Nishimura, T. Satoh, S. Kondo, T. Takeuchi, M. Azetaka, H. Fukuyasu, Y. lizuka, S. Shibahara, J. Antibiot., 1994, 47, 840-842. [22] Y Nishimura, Y. Umzawa, S. Kondo, T. Takeuchi, K. Mori, I. Kijima-Suda, K. Tomita, K. Sugawara, K. Nakamura, J. Antibiot., 1993, 6, 1883-1889. [23] Y. Nishimura, T. Satoh, H. Adachi, S. Kondo, T. Takeuchi, M. Azetaka, H. Fukuyasu, Y. lizuka, /. Am. Chem. Soc., 1996, 118, 3051-3052. [24] T. Takatsu, M. Takahashi, Y Kawase, R. Enokita, T. Okazaki, H. Matsukawa, K. Ogawa, Y. Sakaida, T. Kagasaki, T. Kinoshita, M. Nakajima, K. Tanzawa, J. Antibiot., 1996, 49, 54-60. [25] Y. Kawase, M. Takahashi, T. Takatsu, M. Arai, M. Nakajima, K. Tanzawa, /. Antibiot., 1996, 49, 61-64. [26] A. M. Scofield, L. E. Fellows, R. J. Nash, G. W. J. Fleet, Life ScL, 1986, 39, 645; S. V. Evans, A. R. Hayman, L. E. Fellows, T. K. M. Shing, A. E. Derome, G. W. J. Fleet, Tetrahedron Lett, 1985, 26, 1465-1468. [27] M. Ichikawa, Y. Igarashi, Y Ichikawa, Tetrahedron Lett., 1995, 36, 1767-1770. [28] Y Ichikawa, Y. Igarashi, Tetrahedron Lett., 1995, 36, 4585-486. [29] M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem., 1995, 3, 161-165. [30] Y Igarashi, M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem. Lett., 1996, 6, 553-558. [31] Y Igarashi, M. Ichikawa, Y Ichikawa, Tetrahedron Lett., 1996, 37, 2707-2708. [32] R. C. Bernotas, M. A. Pezzone, B. Ganem, Carbohydr. Res., 1987, 767, 305-311; G. Legler, S. Pohl, Carbohydr. Res., 1986, 755, 119-129. [33] A. Hansen, T. M. Tagmose, M. BoIs, Tetrahedron, 1997, 53, 697-706; A. Hansen, T. M. Tagmose, M. BoIs, Chem. Commun., 1996, 2649-2650. [34] D. R Dumas, T. Kajimoto, K. K.-C. Liu, C.-H. Wong, D. B. Berkowitz, S. J. Danishefsky, Bioorg. Med. Chem. Lett., 1992, 2, 33-36. [35] I. Cenci di Bello, P. Dorling, L. Fellows, B. Winchester, FEBS Lett., 1984, 776, 61-64. [36] K. Bock, I. Lundt, C. Pedersen, Carbohydr. Res., 1981, 90, 17-26. [37] L. Hyldtoft, M. Godskesen, I. Lundt, unpublished results. [38] L. Chen, D. P. Dumas, C.-H. Wong, J. Am. Chem. Soc., 1992, 774, 741-748. [39] I. Lundt, Topics in Current Chemistry, 1997, 787, 117-156, Springer Heidelberg. [40] S. G. Withers, R. Albersold, Protein Science, 1995, 4, 361-372. [41] S. G. Withers, Pure & Appl. Chem., 1995, 67, 1673-1682.

7

Synthesis and Biological Activity of Castanospermine and Close Analogs PETER C. TYLER and BRYAN G. WINCHESTER

7.1 Introduction Castanospermine [(15,65,77?,8/?,8a/?)-tetrahydroxyoctahydroindolizidine] 1, was originally isolated from the seeds of the Moreton Bay chestnut Castanospermum australe [1] and subsequently from another legume Alexa leiopetala [2]. It is one of a number of plant derived polyhydroxyalkaloids, many of which have been shown to be potent glycosidase inhibitors [3-5]. Castanospermine is a potent inhibitor of several glucosidases [6-14] including some involved in lysosomal glycoprotein processing [5-19]. A diverse plethora of biological properties have been ascribed to Castanospermine including anti-viral [20-32], anti-cancer [33-37] and anti-malarial [38] activity as well as anti-inflammatory [39, 40], immunosuppressant [41-44] and anti-diabetic [45, 46] properties. The source of this alluring array of biological activity has been presumed - but not always proven - to be castanospermine's glycosidase inhibitory properties. This review will concentrate on Castanospermine and 'close' analogs. We will not cover all polyhydroxyindolizidines, and swainsonine and its analogs are specifically excluded. Trihydroxyindolizidines that can be readily considered as castanospermine analogs are covered as are some pentahydroxyindolizidines (2-hydroxycastanospermine analogs) and tetrahydroxyquinolizidines. Methods used for the synthesis of Castanospermine and analogs will be discussed followed by an assessment of the relative glycosidase inhibitory properties from the data available and the structural features required for potent glycosidase inhibition.

7.2

Synthesis of Castanospermine and Analogs

Early synthetic approaches to Castanospermine and analogs have been the subject of an excellent review [47] and only subsequent material will be considered here. Jniinosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

126

7 Synthesis and Biological Activity of Castanospermine...

7.2.1

Total syntheses

7.2.1.1

Castanospermine and stereoisomers

Syntheses of Castanospermine and its stereoisomers (tetrahydroxyindolizidines) have commonly utilized carbohydrate starting materials with multi-step sequences being the rule. Some recent efforts, however, have used Sharpless epoxidations to generate at least some of the chiral centers but the overall syntheses were still very long. In one case [48] the lactol 2 was converted by an iterative Wittig protocol and subsequent tosylation into the diene 3 (Scheme 7-1). Desilylation and asymmetric Sharpless epoxidation afforded epoxide 4 as a single isomer which upon azide displacement of the tosylate gave 5. Asymmetric dihydroxylation of the remaining double bond with OsCU in the presence of chiral ligand f(DHQ)2-PHAL], afforded the pair of diols 6 and 7 in a ratio of 10:1 while use of the alternative ligand [(DHQD)2-PHAL] generated a ratio of 1:20. Reduction of the azide moiety in 6 led to the indolizidine lactam 8, (isolated as the tetraacetate 9), but in 13 % yield from 5. Reduction of the lactam and deacetylation then afforded Castanospermine 1. The same procedure applied to 7 gave 6,7-diepicastanospermine 10. Another synthesis [49] of Castanospermine that features a Sharpless asymmetric epoxidation started with allylic alcohol 11 (derived from dimethyl L-tartrate) which led

1

HO '

OTIPS

1 X = H, Y = OH

10 χ = OH, Y = H

Scheme 7-1.

7.2 Synthesis of Castanospermine and Analogs

V^^OH

L

I OTBDMS

127

—> Sharpless Epoxidation

OTBDMS 14 X = OH, Y = H 15X = H1Y = OH MOMO

17R=H 18R = OAc1OCOPr

QMOM

OTBDMS

Scheme 7-2.

19X = OH, Y = H

20 X = H, Y = OH

to epoxide 12 (Scheme 7-2). Regioselective opening of the epoxide and subsequent standard manipulation led to aldehyde 13 which was exposed to lithiated ethyl acetate to give predominantly 14 along with small amounts of epimer 15. Desilylation of 14, tosylation of the primary hydroxyl group and hydrogenolysis of the N(Bn)2 moiety generated the indolizidine lactam 16 which on acid hydrolysis afforded 1-epicastanospermine 17. Acylation of 16 followed by acid hydrolysis gave access to the 1-0-acyl-l-epicastanospermines 18. Alternatively, ester reduction and then selective silylation of the isomeric mixture 14, 15 gave 19 and 20. Mitsunobu inversion of the major isomer 19 gave 20 which was converted (desilylation, tosylation of the primary hydroxy groups, hydrogenolysis and acid hydrolysis) into castanospermine. Other syntheses of castanospermine have employed carbohydrate starting materials. A Reformatsky reaction was applied to L-idurono-3,6-lactone derivative 21 followed by conventional processing to give the octose epimers 22 and 23 in a 4:1 ratio (Scheme 7-3) [5O]. These were converted into 1-epicastanospermine 17 and castanospermine 1 respectively via replacement of the silyl ether with a tosylate leaving group, reduction of the azide and pyrrolidine ring formation followed by hydrolysis and reductive amination to generate the indolizidine ring system.

128

7 Synthesis and Biological Activity of Castanospermine... CH2N3

H— — H — OTBDMS TBDMSO"

22 X = OH, Y = H 23 X = H, Y = OH

Scheme 7-3.

Another long synthesis [51] of castanospermine starts with the gluconolactam derivative 24, which is readily available from 2,3,4,6-tetra-O-benzyl-D-glucose by way of reductive amination of a 5-keto-gluconamide [52, 53]. N-Allylation of 24 followed by selective acetolysis and saponification gave 25 which on oxidation and Wittig reaction afforded the diene 26 (Scheme 7-4). This was subjected to an olefin metathesis reaction with the ruthenium carbene complex 27 generating

QBn

OBn

QBn

QH

PPR3 PPR3 27 R = Cyclohexyl Reagents: i, CH2=CHCH2Br, aq KOH1 CH2CI2, Bu4NI; ii, FeCI3, Ac2O; iii, NH3, MeOH; Iv, Dess Martin; v, Ph3P=CHCO2Me; vi, 27, toluene, reflux, 48h; vii, OsO4, NMO; viii, SOCI2, Et3N; Ix, RuCI3, NaIO4; x, NaBH4, then 20%H2SO4, Et2O; xi, BH3-SMe2; xii, H2, Pd/C.

Scheme

7-4.

7.2 Synthesis of Castanospermine and Analogs

129

the unsaturated indolizidine 28. Dihydroxylation of the double bond gave a mixture of diols which were isolated as their cyclic sulfates. The major cyclic sulfate 29, was selectively reduced and then further standard manipulations afforded castanospermine 1. The unnatural enantiomer of castanospermine has been synthesized [54] from the o-xylose derivative 30. Addition of viny!magnesium bromide to 30 and protection of the diol product as the bis-MOM ether gave 31 as the major product, which on ozonolysis afforded the hexose 32. Allylation of this aldehyde could be controlled to produce either of the stereoisomers 33 or 34 as the major product. With Hiyama-Nozaki [55, 56] conditions (CrCls/LiAH-U, allyl bromide) the product ratio was 85:15 in favor of 33, whereas with chelation-controlled addition (allyltributyltin, MgB 12) the product was 95:5 in favor of 34. Benzylation of 33, followed by ozonolysis, reduction, and Mitsunobu amination of the resulting alcohol afforded the phthalimido compound 35. Following removal of the MOM protecting groups, the primary hydroxy group was selectively silylated and mesylation of the other hydroxy group followed by deprotection of the amine generated the pyrrolidine 36. Subsequent desilylation and chlorination of the alcohol induced spontaneous cyclization to produce the indolizidine 37, which on debenzylation afforded (-)-castanospermine 38. The same process applied to 34 afforded (-)l-epicastanospermine 39 (Scheme 7-5). QBn QBn MOMC

MOMi ... * '"

"TT - 'OBn OMOM iv

32

°Γν

33 X = H1Y = OH 34 X = OH, Y = H OBn OBn

Bn

O

39

Scheme 7-5.

Reagents: i, CH2=CHMgBr, THF; ii, MOMCI, IPr2NEt; iii, O3, CH2CI2; Iv, CH2=CHCH2Br, CrCI3, LiAIH4; v, CH2=CHCH2SnBu3, MgBr2; vi, NaH, BnBr; vii, LiAIH4; viii, phthalimide, Ph3P, DEAD; ix, HCI, MeOH; x, tBUPh2SiCI, imidazole; xi, MsCI, py; xii, N2H4-H2O, EtOH; x»», Bu4NF; xiv, Ph3P, CCI4, CH3CN, Et3N; xv, H2, Pd/C.

130

7 Synthesis and Biological Activity of Castanospermine.. 3n QBn QMOM ^CHO

OBn

MOMO

40

QBn ~

vii, viii

j

π

^^

*"

SnBu3

QH -

H

Reagents: i, MgBr2, CH2CI2; ii, MsCI, Et3N; iii, MeNH2, EtOH; \v, aq HCI, THF; v, CbzCI, NaHCO3; vi, BH3-THF then H2O2, NaHCO3; vii, TsCI, DMAP, Et3N; viii, H2, Pd/C.

Scheme 7-6.

The above approach used two successive organometallic additions to control two of the chiral centers, starting with a pentose aldehyde. Others have generated these two chiral centers in a single step from pentose derivatives [47, 57]. In a further elaboration of this approach the chelation controlled addition of an allylstannane to the arabinose derivative 40 afforded a single product 41 (Scheme 7-6) [58]. Conventional processing led to the piperidine 42 which after hydroboration gave 43. Selective tosylation of the primary hydroxy group followed by hydrogenolysis then produced 1,6-diepicastanospermine 44. A similar strategy has been used in the allylboration of D-arabinose derivative 45 with (Z)-S-(methoxy)methoxyallyldiisopinocampheylborane 46 to give a single product 47 (Scheme 7-7) [59]. Hydroboration of this followed by mesylation of the resultant diol and then a double displacement with benzylamine afforded the pyrrolidine 48. Selective hydrolysis of the terminal isopropylidene moiety, tosylation of the primary hydroxy group and hydrogenolysis generated the indolizidine ring system which after deprotection gave 6,8a-diepicastanospermine 49. Potentially all stereoisomers of castanospermine are available in this way depending on the pentose starting material. An elegant triple reductive amination approach to castanospermine has been described [6O]. Allylation of the readily available D-glucose derivative 50 generated a 9:1 ratio of stereoisomers, and benzylation of the major isomer afforded 51 (Scheme 7-8). A clever manipulation of this material using iodonium dicollidine perchlorate then produced the tetrahydrofuran 52, which without purification was treated with zinc to regenerate the olefin 53 with the C-5 -hydroxy group selectively exposed for oxidation. Swern oxidation of 53 followed by ozonolysis of the olefin and hydrolysis of the dimethylacetal moiety produced the dialdosulose deri-

7.2 Synthesis of Castanospermine and Analogs

131

PMOM

QH ~. H

OH

v-viii

49 Reagents: i, -78OC then NaOAc, H2O2; ii, (cyclohexylfeBH, then H2O2, NaOAc; Hi, MsCI, py; iv, BnNH2, 7QOC; v, aq TFA; vi, TsCI,

Scheme 7-7.

PV·' νίί· Η2>Pdblack;νί|ί'Η+ resin.a
QBn

QBn

BnOs

BnO""

BnO1"'

ix /55 R = Bn

C

^1

Scheme 7-8.

R=H

Reagents: i, CH2=CHCH2Br, Sn1 aq CH3CN1,»)); ii, NaH, BnBr; iii, IDCP, CH2CI2, MeOH; iv, Zn, EtOH; v, Swern; vi, O3, then Ph3P; vii, aq HCI, MeOH; viii, 1.3eq NH4HCO2, NaBH3CN, MeOH;

'*, Pd/c, HCOOH, MeOH.

132

7 Synthesis and Biological Activity of Castanospermine...

vative 54 as a mixture of lactol isomers. Triple reductive amination of this material successfully afforded tetra-O-benzylcastanospermine 55 (53 % yield) which provided castanospermine after transfer hydrogenolysis. 7.2.1.2 Pentahydroxyindolizidines The search for castanospermine analogs with an improved biological profile has led to the synthesis of a number of 2-hydroxycastanospermine isomers (pentahydroxy indolizidines). The uronolactone 56 (prepared via an aldol reaction between 2,3-O-isopropylidene-D-glyceraldehyde and a lactone derived from the Diels-Alder adduct of furan with 1-cyanovinyl (IR )-camphanate) [61], has been employed in the synthesis of some pentahydroxyindolizidines. Alcoholysis of lactone 56 with benzyl alcohol produced either the benzyl ester 57 or its C-5 epimer 58 as the major product depending on the carefully selected reaction conditions (Scheme 7-9a) [62, 63]. Silylation of the diols 57 and 58 and hydrogenolysis of the benzyl esters gave the carboxylic acids which were treated with diphenylphosphoryl azide [64] followed by benzyl alcohol affording the benzylcarbamates 59 and 60 respectively. Desilylation and hydrogenolysis of 59 generated the piperidine ring system 61, which after acid hydrolysis was treated under Mitsunobu conditions to afford the pentahydroxyindolizidine 62. Similar treatment of 60 gave the isomeric pentahydroxyindolizidine 63. This methodology was also used to convert the uronolactone 64 into the pentahydroxyindolizidines 65 and 66 [63] (Scheme 7-9b). Other pentahydroxyindolizidines have been prepared from the octonolactone derivative 67 [65]. Reduction of the lactone 67 (Scheme 7-10) [66] followed by mesylation of the resultant diol gave the 1,4-dimesylate 68 which suffered a

''"Y OH 57 X = CO 2 Bn 1 Y=H 58 X = H, Y = CO2Bn

OH OH " H " vi.vii 59X = NHCO2Bn, Y = H 60 X = H, Y = NHCO2Bn

62

63

Reagents: i, CsF, BnOH, CCI4, 0<>C; ii, CsF, BnOH, DMSO; iii, TBDMSOTf, 2,6-lutidine; iv, H2, Pd/C; v, (PhO)2PON3, then BnOH; vi, Bu4NF; vii, H2, Pd/C, EtOAc; viii, aq TFA; ix, Ph3P, DEAD

· py-

Scheme 7-9a.

7.2 Synthesis of Castanospermine and Analogs

133

65 Sa(R)

Scheme 7-9b.

"

Λ>

66

8a(S)

double displacement with benzylamine affording the pyrrolidine 69. Selective hydrolysis of the terminal isopropylidene unit, mesylation of the resulting primary hydroxy group, followed by hydrogenolysis and deprotection then gave rise to 6-£p/-2-(S)-hydroxycastanospermine 70. Attempts to synthesize 2-(S)-hydroxycastanospermine 71 from a suitably protected derivative of 70 were unsuccessful [66] and thus the octonolactone 67 was converted into triflate 72 which gave rise to azide 73 by way of an epoxide intermediate. After reduction of the lactone in 73 and mesylation of the resultant diol, the azide moiety was reduced to an

TBDMSO"' 71

Scheme 7-10.

Reagents: i, LiBH4; ii, MsCI, py, DMAP; Hi, BnNH2; iv, TsOH, MeOH; v, H2, Pd, EtOH, NaOAc; vi, aq TFA; vii, aq HOAc, then TBDMSCI, imidazole, DMF, then Tf2O, py; viii, Bu4NF; ix, NaN3,

DMF; x, TBDMSOTf, Py.

134

7 Synthesis and Biological Activity of Castanospermine...

amine and a spontaneous double cyclization occurred, displacing both mesylates, generating the indolizidine 74. Deprotection then afforded 2-(S)-hydroxycastanospermine 71. Octonolactone 75 was also converted into 6-epi-2-(/?)-hydroxycastanospermine 76 in an analogous manner [66].

76

7.2.1.3

Tetrahydroxyquinolizidines

Some examples of ring-expanded homologues of castanospermine, tetrahydroxyquinolizidines, have been prepared. In the first synthesis of such compounds [67], the Grignard reagent derived from 2-(2-bromoethyl)-l,3-dioxolane was —OH —OTBDMS

Scheme 7-11.

8 4 R = Bn1 9a(S) 8 5 R = Bn, 9a(R) 86 R = H,9a(S) 87 R = H, 9a(R) Reagents: i, Ph3P=CH(CH2J3CI; ii, Ph3P, DEAD, HN3; iii, MCPBA; iv, H2, Pd/C, Et2O, EtOH, then K2CO3, EtOH, reflux; v, H2, Pd/C,

MeOH, HCI.

Scheme 7-12.

7.2 Synthesis of Castanospermine and Analogs

135

added to lactol Π to give the epimeric mixture 78 (Scheme 7-11). The desired epimer (which was the minor isomer) was converted into azide 79 which after acid hydrolysis and hydrogenolysis afforded the quinolizidine 80. In a different approach, [68] Wittig olefination of the o-arabinose derivative 81 gave the 9-carbon alditol 82 (Scheme 7-12) which gave access to the epimeric mixture of epoxy-azides 83. Reduction of the azide allowed a double cyclization to give the quinolizidines 84 and 85 which were deprotected by hydrogenolysis affording 86 and 87. 7.2.1.4

Miscellaneous castanospermine analogs

In an interesting approach to castanospermine analogs, the neuraminic acid derivative 88 (Scheme 7-13) was subjected to hydrogenolysis affording the pyrrolidine 89 directly, isolated as the carbamate 90 [69]. After methyl ester formation the major isomer 91 could be separated. Tosylation of the primary hydroxy group in 91 followed by hydrogenolysis achieved cyclization to the indolizidine system and reduction of the ester afforded 6-epi-3-(hydroxymethyl)castanospermine 92. The hepturonate derivative 93 (obtained from D-xylose) has been converted into nitrone 94 (Scheme 7-14) [70], via formation of an oxime, and in the presence of methyl acrylate bicyclic 95 was obtained. Reduction of the N-O bond afforded lac tarn 96 and then the indolizidine 97. The synthesis of some deoxycastanospermine analogs have been described, including 1-deoxycastanospermine 98 and l-deoxy-8a-£/?i-castanospermine 99 [71] as well as 6,8-dideoxycastanospermine 100, [72] l-deoxy-7-epi-castanospermine 101 and l,7-dideoxy-7-fluorocastanospermine 102 [73]. The thiazolidine 103 is available directly from D-arabinose (Scheme 7-15) [74] and activation of the primary hydroxy group afforded the castanospermine analogs 104 as a mixture of 8a-epimers. The same procedure has since been used to make other stereoisomers of 104 [75].

OH

f% /°\ CbzHN—(

V-/ H6

~

COOH V

_J^_ OH

HC

88

iv-vi

Scheme 7-13.

91 92 Reagents: i, H2, Pd/C, H2O; ii, CbzCI, NaHCO3; iii, HCI, MeOH; iv » TsC|. PY; v« "2, Pd/C, MeOH, NaOAc; v, NaBH4, H2O.

136

7 Synthesis and Biological Activity of Castanospermine.

O2Me

Scheme 7-14.

98 X = OH, Y = H 101 X = H1Y = OH 102 X = F, Y = H QH

Reagents: i, HSCH2CH2NH2, py; ii, Ph3P, CCI4, Et3N.

7.2.2

Scheme

7-15.

Syntheses from castanospermine

While considerable effort has gone into the total synthesis of castanospermine analogs, the ready availability of castanospermine by extraction and purification [45, 76] from the seeds of Castanospermum australe make the natural product itself a useful starting material and a number of analogs have been synthesized in this way. 7.2.2.1

Selective protection of castanospermine

Selective protection of the hydroxy groups of castanospermine has been studied in order to allow the efficient synthesis of analogs. The methods now published allow the specific isolation of any one of the four hydroxy groups in castanospermine with minimal effort. The direct acylation of castanospermine has been studied. In pyridine, with limited quantities of benzoyl chloride, castanospermine furnished the 6,7-dibenzoate 105 in moderate yield, [30, 35, 77] and similar treatment with benzyl chloro-

7.2 Synthesis of Castanospermine and Analogs

137

formate afforded a low yield of the 6,7-bis(benzylcarbonate) 106 [78]. The corresponding 6-O and 7-0-monoesters are also available in low yield from these reactions. The 6,7- dibenzoate 105 has been used to prepare some 6- and 7-monoesters [79] via the 1,8-acetals 107 and 108 which were prepared by acetal formation followed by saponification. Selective acylation followed by acid hydrolysis then furnished the 6-substituted compounds 109. The 6-O-benzy!carbonate 110 prepared in this way could be esterified and then deprotected by hydrogenolysis and acid hydrolysis to give the 7-0-monoesters 111. Some l-<9-acylcastanospermine derivatives have been prepared using the protease enzyme subtilisin as catalyst. A subsequent lipase-catalyzed acylation of these compounds gave 1,7-diesters which were selectively hydrolyzed with subtilisin to give 7-0-monoesters [8O]. In an organotin-mediated acylation of castanospermine, the adduct from dibutyltin oxide and castanospermine in methanol was treated with acyl halides and triethylamine affording moderate yields of 6-0-monoesters [81]. In contrast, the adduct from bis(tributyltin) oxide and castanospermine in cold toluene was selectively acylated with acyl halides to give high yields (85-90 %) of the 6-esters 109 directly [82]. The corresponding 6-O-benzy!carbonate 112 was formed in the same way and acetylated in situ (Scheme 7-16) to give 113 which on hydrogenolysis afforded 1,7,8-tri-O-acetylcastanospermine 114. This triacetate was also available, along with lesser amounts of 1,6,8-tri-(9-acetylcastanospermine 115, by a selective deacetylation of tetra-O-acetylcastanospermine. When excess acyl halide was used with the bis(tributyltin) oxide adduct above, a mixture of the corresponding 6monoester and 1,6-diester was obtained and more complete acylation was not possible. However, using dibutyltin oxide under the same conditions, with 2 equivalents of benzoyl chloride or pivaloyl chloride the 1,6-di-O-acylcastanospermines 116 were produced in high yield [83]. With three equivalents of benzoyl chloride a mixture of the 1,6,7- and 1,6,8-tri-O-benzoylcastanospermines 117 and 118

105 106 109 111

R1 s R2 = Bz R1 = R2 = CO2Bn R1 = Bz1 COPr, R2 = H R1 = H 1 R 2 = Bz1COPr

107

R1 = R2 = H, R31R4 = Me 108 R1 = R2 = H, R31R4 = (CH2J5 no R1 = CO2Bn, R2 = H, R31R4 = (CH2J5

PAc AcC

81 . HO1"' 112 R=H 113 R = Ac

Scheme 7-16.

Reagents: i, (Bu3Sn)2O, toluene, reflux, then BnOCOCI, -20<>C to RT, then Ac2O; ii, H2, PdIC, EtOH, EtOAc.

138

7 Synthesis and Biological Activity of Castanospermine...

BzO1'' 117

Reagents: i, Bu2SnO, toluene, reflux, then BzCI, -7QOC to RT.

Scheme

7-17.

were obtained in a 3:1 ratio [84]. Under these conditions it appears that the reactivity to acylation is O-6 > O-l > O-l > O-8. However, when trichloroethylchloroformate was the acylating agent the single product was l,6,8-tri-O-(2,2,2-trichloroethoxycarbonyl)-castanospermine 119 [83]. Other results have suggested that the product ratio may depend on isomerisation of the initially formed products [85]. When benzylation of castanospermine was attempted using dibutyltin oxide in refluxing toluene, the first reaction on addition of benzyl bromide was quaternisation of the nitrogen. Then a slow reaction took place over three days in refluxing toluene to give (after dequaternisation) 1,6,7-tri-O-benzylcastanospermine 120 in 50% yield, (contaminated with a little of the corresponding 1,6,8-tribenzyl ether 121), as well as 15 % of the separable rearranged material 122 (Scheme 7-18) [84]. The rearrangement is presumed to take place via the transient epoxide intermediate 123.

iOBn

RO" 123

Reagents: i, Bu2SnO, toluene, reflux, then Bu4NBr, BnBr, reflux 3 days; ii, LiSPr, DMSO.

Scheme 7-18.

7.2 Synthesis of Castanospermine and Analogs H

139

OTBDMS

HO1" BzO1'' 125 R = H1R11R2 = Me 126 R = H1R11R2 = (CH2J5 127 R = B^R11R2 = Me 1 2 8 R = Bz, R11R2 = (CH2)5 AcC

?A°H ~ -

PAC

MsO1'''

Figure 7-1.

129 R = BZ, coc(CH3)3

130

Surprisingly, when silylation was attempted using bis(tributyltin) oxide in refluxing toluene, only the l-O- terf-butyldimethylsilylcastanospermine 124 was obtained [83] (Figure 7-1). Acylation using this organotin reagent has afforded only the corresponding 6-esters and the cause of this difference in selectivity is not readily apparent. The formation of castanospermine O-acetals (isopropylidene, benzylidene, cyclohexylidene) has been unsuccessful under all standard conditions when applied directly to castanospermine. However the 6-benzoate 109 (R=Bz) was readily transformed into the 1,8-O-acetals 125 and 126 [82] and, as mentioned above, the 6,7-dibenzoate also afforded the 1,8-acetals 127 and 128 [79]. The 1,6-di-O-acyl derivatives 116 could be selectively deacylated (KCN, MeOH or NHs, MeOH) to give the 1-esters 129. Initially isomerisation occurred to give mixtures of 1,6- 1,7- and 1,8-diesters, followed by gradual deacylation [83, 86]. A facile acyl migration of 6-O-acetylcastanospermine has also been observed [87]. 7.2.2.2 Reactions at C-6 Displacement reactions applied to 1,7,8-tri-O-acetylcastanospermine 114 and its derived mesylate 130 have afforded a number of selectively modified derivatives. Treatment of the alcohol 114 with diethylaminosulfur trifluoride (DAST) generated the product of displacement with retention of configuration 131 as well as the rearranged material 132 (Scheme 7-19) [82]. Clearly the nitrogen participates to form the aziridinium ion intermediate 133. The same aziridinium ion was an intermediate when displacement reactions were applied to mesylate 130. Depending on the nucleophile, varying amounts of the products of displacement with retention or rearrangement were obtained. By this means a number of analogs 134 and 135 have been prepared [82]. Interestingly, when the 6-chloro compound 134 (X=Cl) was allowed to stand in aqueous phosphate buffer the 6-O-phosphate 134 (X=OPO3H2) was formed [88], with the aziridinium ion 133 again being a presumed intermediate.

140

7 Synthesis and Biological Activity of Castanospermine. -OAC

Reagents: i, DAST, CH2CI2.

133

Scheme 7-19.

C H2X

134

"

135 X = NH2, NHAc, Cl, CN, NHR R = Me, AIIyI, "Butyl, 2-methylpropyl, 1-methylpropyl, 1-methylbutyl, 2-methoxyethyl

MOM

X'''' 136 X = OMs, R11R2 = Me 137 X = OMs, R^R2 = (CH2)5 139 X = I1R^R2 = (CH2)S

l

MOMC

138 X = N(CH2CH2OH)2, NEt2, NHBn, N(Et)Bu

°

CH2=CH''' 140

O -

LJ

^ " R 142 X = H, Y = OH

143 X,Y='O

144

R =

"'

Me> CH=CH2

Figure 7-2.

7.2 Synthesis of Castanospermine and Analogs

141

The 1,8-acetals 125 and 126 were converted into mesylates 136 and 137. Displacement reactions applied to these compounds afforded products of displacement with retention of configuration, but without any rearrangement [82]. Sufficient nitrogen participation to explain the retention of configuration must be invoked, but not enough to allow the formation of rearrangement products. The analogs 134 were also available in this way - usually in better overall yield - and some other tertiary amines 138 were prepared. The iodide 139 was prepared, by displacement of the mesylate 137, and when treated with vinylmagnesium bromide afforded the branched-chain derivative 140 which was deprotected to triol 141 (Figure 7-2). Whereas all attempted oxidations applied to 1,7,8-tri-O-acetylcastanospermine 114 only furnished starting material or degradation products, Swern oxidation of the l,8-(9-cyclohexylidene derivative 142 successfully generated ketone 143. This gave access to the 6-^pi-castanospermine derivatives 144 [82]. Attempts to prepare a C-6 deoxy analogue were unsuccessful. Radical reduction of thiocarbonyl derivatives of alcohol 114 only returned the alcohol and reduction of iodide 139 or l,7,8-tri-O-acetyl-6-bromo-6-deoxycastanospermine by hydrogenolysis or radical or hydride reduction were unsuccessful. 7.2.2.3

Reactions at C-8

Displacement reactions have been applied to mesylates 145 and 146, and again the nitrogen participates so that products of displacement with retention of configuration and rearrangement are formed, the aziridinium ion 147 being an intermediate [84]. Thus tribenzoate 145 readily gave azides 148 and 149, which gave access to 150-153. The mesylate 154 was available from 122 and in displacement reactions afforded the same products as 146 via the same aziridinium ion intermediate 147. The nitriles 155 and 156 were synthesized in this way and the deoxyfluoro compounds 157-160 were obtained from DAST treatment of alcohols 120 or 122. Reduction of the nitriles gave the amines 161 and 162. The tribenzyl ether

145 R = Bz, X = OMs 146R = Bn,

Figure 7-3.

<ΛΟ 148 150 151 155 157 159 161 163 164 166 168

R0,. T I

>t^ ^J Vr\ ^+

B =PM5 M RO''''^^ ^7 R = Bz, X = N3 147 R = H1X = NH2 R = H, X = NHAc R = Bn, X = CN R = Bn, X = F R = H, X = F R = Bn, X = CH2NH2 R = H1X = OCOC(CHa)3 R = Bn1X = OMe R = H1X = OMe R = Bz, X = OCS-imidazole

169 R = X = H

)

149 R = Bz, X = N3 152R = H,X = NH 2 153 R = H, X = NHAc 154 R = Bn 1 X = OMs R = Bn, X = CN 156 158R = Bn,X = F 160 R = H, X = F 162 R = Bn, X = CH2NH2 165R = Bn, X = OMe 167 R = H, X = OMe

142

7 Synthesis and Biological Activity of Castanospermine...

120 has been acylated and then hydrogenolyzed to give 8-0-pivaloylcastanospermine 163, and other 8-0-substituted derivatives are available in this way. The 8-0methyl ether was prepared, but to avoid problems of quaternisation of the nitrogen the quaternary ammonium salts formed in the preparation of 120 and 122 were Omethylated and then dequaternised in the usual way to give 164 and 165 and after hydrogenolysis, 166 and 167. The thiocarbonylimidazole derivative 168 could be reduced in this case, with tributyltin hydride, to give after saponification, 8-deoxycastanospermine 169 [84] (Figure 7-3). 7.2.2.4

Reactions at C-7

Chemistry at this centre generally proceeded without obvious participation of the nitrogen [83]. The triflate ester 170, derived from alcohol, 115 was readily displaced by acetate or fluoride anions to give, after saponification, 7-e/?i-castanospermine 171 and 7-deoxy-7-£pi-7-fluorocastanospermine 172. Base treatment of the triflate 170 afforded only the 6,7-anhydro derivative 173. Attempts to effect a double inversion at C-7 by treatment of triflate 170 with sodium iodide (to give iodide 174) followed by various nucleophiles did not afford discree t products. However the epi-mesylate 175 could be displaced by azide or chloride anions to give, after further manipulation, the 7-deoxy-7-substituted castanospermines 176-178. Oxidation of alcohol 115 readily afforded the ketone 179 - in contrast to the unsuccessful attempts to oxidize at C-6 or C-8. Treatment of this ketone with methylmagnesium bromide generated 7-C-methylcastanospermine 180 and the 7-epimer 181 in a 4:1 ratio (Figure 7-4).

RO" 170 R = Ac 1 X = OTf 177

R-UY

178 R = K; X = ci "~

]75 R = ACjX 1 -H, Y ==OMs

181 R = H^X = Me1V = OH OAc O; Ω AcO1''

^

b

179

Figure 7-4.

7.2 Synthesis of Castanospermine and Analogs 7.2.2.5

143

Reactions at C-I

There has been limited selective chemistry performed at this centre and, as expected, the nitrogen moiety was not observed to participate in displacement reactions. While an enzyme catalyzed selective 1-0-acylation of castanospermine has been reported [80], preparatively useful quantities of the 1-esters 129 were best obtained by selective saponification of the corresponding 1,6-diesters 116 [83, 86]. Protection of the triol and deacylation afforded 6,7,8-tri-O-methoxymethylcastanospermine 182. This was also readily available from the 1-0-silyl ether 124. Swern oxidation of the alcohol 182 furnished the ketone 183 and selective reduction generated the protected 1-^pi-compound 184 which, after acid hydrolysis, gave 1-cpi-castanospermine 17 [83]. Treatment of the alcohol 182 or the \-epialcohol 184 separately with DAST and then aqueous acid afforded 1-deoxy-l-epi1-fluorocastanospermine 185 and 1-deoxy-l-fluorocastanospermine 186 respectively. 1-O-Methylcastanospermine 187 was also prepared by methylation of alcohol 182 - without quaternisation of the nitrogen - and acid hydrolysis. Addition of methylmagnesium bromide to the ketone 183 stereoselectively afforded, after deprotection, 1-C-methylcastanospermine 188 [83] (Figure 7-5).

OMOM MOMO. ^AJ

MOMO"" \^

Jt-Y

^7

182 X = OH,Y = H 184 X = H, Y = OH MOMC

Figure 7-5.

7.2.2.6

HO^ ^v?

J..-.Y

HO"

OMOM ο -. H // - Ι

J85 186 187 188

Χ-Η,Υ-F X = F, Y = H X = OMe, Y = H X = OH, Y = Me

MOMO1'' 183

Miscellaneous derivatives

The synthesis of some 0-glucosyl-castanospermine derivatives has been reported [89, 9O]. Condensation of imidate 189 with the 6-OH compound 190 (prepared from the 1,8-acetal 110) produced the 6-O-a-glucoside 191 after deprotection. Similarly the 7-0-a-glucoside 192 was synthesized from the tribenzoate 118. Treatment of the same tribenzoate 118 with the O-acetyl imidate 193 generated the 7-0-/?-glucoside 194 after deprotection. Glucosylation of the 6,7-dibenzoate 105 with imidate 189 gave a mixture of products affording the 8-0-a- and -/?-glucosides 195 and 196 as well as the 1-0-a-glucoside 197 after deprotection, but in low yields (Figure 7-6).

144

7 Synthesis and Biological Activity of Castanospermine.

190

OR3

1

R1O 191 192 194 195 196 197

R1 = α-D-Glcp, R2-R4 = H R2 = α-D-Glcp, R1,R3,R4 = H R2 = β-D-Glcp, Ri-R3-R4 = H R3 = α-D-Glcp, R11R-^R4 = H R3 = β-D-Glcp, Ri1R21R4 = H R4 = a-D-Glcp, R1-R3 = H

* IgUFC 7-0.

Treatment of castanospermine with periodic acid followed by sodium borohydride has afforded the substituted pyrrolidine triol 198, isolated as the triacetate 199 [91].

7.3 Biological Activities of Castanospermine and Derivatives The biological properties of castanospermine can be understood in terms of inhibition of the interaction between the glucosyl residue of a glycoconjugate and an enzyme or other protein. Two main factors contribute to this inhibition: (i) the strong structural resemblance between the four substituents at C-6, C-7 and C-8 and C-Sa on the piperidine ring of castanospermine and the four substituents on C-2, C-3, C-4 and C-5 of the pyranose ring of D-glucose [92] and (ii) the metabolic inertness of castanospermine in animal cells. Glucosyl residues are found in animals in dietary polysaccharides of plant and animal origin, as transient intermediates in the biosynthesis of asparagine-linked glycans of glycoproteins and as constituents of glycosphingolipids, 0-linked glycans of glycoproteins, glycogen and collagen. It was established soon after its isolation [1] that castanospermine is a potent and specific inhibitor of a- and /?-D-glucosidases from a wide range of organisms and sub-cellular locations e.g. lysosomal a- and /?-D-glucosidases [6], glycoprotein processing a-o-glucosidases [17], digestive a- and /?-D-glucosidases [93], the broad specificity cytosolic /?-D-glucosidase [12] and plant thioglucosidase (myrosinase) [8].

7.3 Biological Activities of Castanospermine and Derivatives

145

The biological consequences and therapeutic potential of inhibition of these enzymes have been studied at the level of an individual enzyme, viruses, cells in culture and in whole organisms [4, 15, 94, 95]. Various naturally occurring or chemically synthesized derivatives of castanospermine have also been investigated to help to define the critical structural features responsible for the specificity and potency of the inhibition [11, 16, 83, 96-98]. The effect of castanospermine or a derivative on an enzyme activity in living cells will depend not only on the potency of the inhibition towards the enzyme but also on the accessibility of the enzyme to the inhibitor. Many chemical modifications have been made to castanospermine to enhance its uptake into cells and intracellular transport under physiological conditions [11, 16].

7.3.1

Structure-activity relationships

The four chiral centers at C-I, C-6, C-7 and C-Sa have been systematically modified by inversion of the configuration, removal, replacement or esterification of the hydroxyl groups [11, 16, 82-86, 96, 97]. All of these modifications decreased the inhibition of lysosomal α-D-glucosidase and/or the processing α-D-glucosidase in vitro showing the importance of the correct number and stereochemistry of the hydroxyl groups for the specific inhibition of glycosidases [11, 16, 83]. Some of these modifications created analogs of other monosaccharides e.g. 6-epi-castanospermine, which is related to o-mannose in the same way as castanospermine is to o-glucose. 6-E/7/-castanospermine is a good inhibitor of human cytosolic and neutral α-D-mannosidases but does not inhibit the lysosomal a-D-mannosidase [U]. This differential inhibition can be explained by the preferential recognition of different conformations of o-mannose by the different isoenzymes of a-D-mannosidase [94], All the derivatives of 6-^pi-castanospermine also have the minimal structural features for the inhibition of α-L-fucosidase but those with a substituent configuration analogous to the /?-anomeric position in L-fucose did not inhibit the enzyme. l-Deoxy-6,8a-diepicastanospermine, which has four chiral centers identical with α-L-fucose, was a very potent inhibitor of α-L-fucosidase (Ki 1.3 mM) [U]. Although most substitutions abolished the inhibition of α-D-glucosidase, substitution of the 6-OH with Cl, F or CN only decreases the inhibition very slightly (unpublished results). Substitution of the 6-OH of castanospermine with a free or acylated amino group produced excellent inhibitors of N-acetyl-/?-D-glucosaminidase activity (/?-hexosaminidase) (unpublished results). Contraction of the sixmembered ring to form australine decreases the inhibition of α-D-glucosidase I and II and particularly of /?-D-glucosidase [99]. Most of the C-6 or C-7 ester derivatives of castanospermine are as good or better inhibitors of the processing α-D-glucosidase I in cells in culture than castanospermine, although they are inactive towards the enzyme in vitro. This is because they are taken up more efficiently into the cells on account of their increase in hydrophobicity and are then hydrolyzed in situ to the active parent compound. This is the basis of the use of these derivatives as prodrugs. Derivatives with nonhydrolysable groups such as amides or esters are not active in vitro or in vivo.

146

7 Synthesis and Biological Activity of Castanospermine...

The N-oxide and N-methyl derivatives of castanospermine retain some activity towards a-D-glucosidase, 50 and 80%, respectively at 1 mM concentration. An active N-oxide of swainsonine has been isolated from extracts of locoweed [10O].

7.3.2

Enzymology of glucosidases

Castanospermine is highly specific for inhibition of α-and /?-o-glucosidases. This specificity makes it a useful tool for comparing the kinetics and mechanism of action of different α-and /?-D-glucosidases in vitro [12, 101]. It has become the yardstick by which novel natural and synthetic a- and /?-o-glucosidase inhibitors are judged e.g. nectrisine [102] or calystegins [103]. The high affinity of a-D-glucosidases for castanospermine has been exploited in an ingenious manner to isolate liver Golgi endomannosidase by affinity chromatography [104]. Inclusion of castanospermine in the buffer prevented binding of α-o-glucosidases to the α-linked glucosyl residue of the ligand, Glcal^SMan. The specificity and potency of castanospermine have also been used to improve the efficiency of an assay for seminal a-D-glucosidase [105]. 7.3.2.1

Digestive glucosidases

Consumption of the chestnut-like seeds of the Moreton Bay chestnut Castanospermum australe has been reported to lead to gastrointestinal problems and occasional deaths among animals and humans in its native Australia [106]. The cause of death was probably a combination of the effects of castanospermine, especially if consumption was prolonged. However, the acute gastrointestinal problems can be attributed to the very potent inhibition of mammalian intestinal disaccharidases by castanospermine with sub micromolar values of K1 for the inhibition of sucrase and maltase. This would lead to decreased absorption of monosaccharides and osmotic diarrhoea [46, 107-11O]. The controlled inhibition of the digestion and absorption of dietary carbohydrates in the small intestine has therapeutic potential in the treatment of diabetes and obesity. Castanospermine can delay the hyperglycemic response to oral sucrose in normal and diabetic (streptozoticin-induced) rats [45, 111]. Glucosyl derivatives of castanospermine have been synthesized to obtain selectivity of inhibition of oligosaccharidases, but it has been difficult to predict the specificity of these pseudo-disaccharides [112]. Two other a-o-glucosidase inhibitors, the pseudotetrasaccharide acarbose (BAYg5421) and Miglitol ((Nhydroxyethyl)-l-deoxynojirimycin, BAY ml099) are in clinical use or undergoing clinical trials for treatment of non-insulin-dependent and insulin-dependent diabetes mellitus [113-115]. Acarbose, unlike Miglitol, is not absorbed appreciably into the bloodstream and therefore its action is largely confined to the intestine. Its main side effect is moderate diarrhoea and associated flatulence. Castanospermine would be expected to cause systemic complications as well as diarrhoea because of its absorption into the bloodstream, especially if high concentrations are used.

7.3 Biological Activities of Castanospermine and Derivatives

147

There are marked differences between species in the specificity of inhibition of digestive disaccharidases by castanospermine [3,4, 116-121]. For example, castanospermine is a good inhibitor (K{ 0.8 mM) of the soluble midgut trehalase from the larvae of the gypsy moth [118] but it does not inhibit trehalase of the aphid (Acyrthosiphon pisum, Harris) [119]. This differential inhibition is potentially the basis of selective insecticides. Interestingly, castanospermine is a very active feeding deterrent to the aphid and to nymphs of the migratory locust, Locusta migratoria L but not to nymphs of the desert locust Schistocerca gregaria Forsk [12O]. This feeding deterrence may not result from inhibition of glycosidases but from recognition of the castanospermine by taste receptors on sensillae on the insect mouthparts, which normally recognize sugars and evoke a response to eat. The molecular basis of this signaling is under investigation as it offers the possibility of crop protection. This property of castanospermine and certain other naturally occurring aminosugars such as DMDP (2/?,5/?-dihydroxymethyl-3/?,4/?dihydroxypyrrolidine) might explain their presence in plants. It also illustrates another very important potential application of castanospermine and derivatives as inhibitors or stimulators of processes mediated by carbohydrate-recognizing proteins such as lectins or transporters as opposed to inhibition of enzymes. Inhibition of α-D-glucosidase activity by castanospermine has also been demonstrated to be useful in regulating the biological effects of the stable ascorbate prodrug, 2-0-a-D-glucopyranosyl-L-ascorbic acid (AA-2G) [121-124], Addition of castanospermine to cultures of human fibroblasts and murine splenocytes or human peripheral blood lymphocytes inhibits the AA-2G-induced synthesis of collagen and production of antibodies, respectively. This suggests that the AA-2G is hydrolyzed by a cellular α-D-glucosidase to release ascorbate, possibly after transport to the lysosomes. Perfusion studies using the guinea pig small intestine showed that AA-2G disappeared from the perfusate but that intact AA-2G was not detected in the portal vein. These observations indicated that AA-2G was hydrolyzed by brush border α-D-glucosidase activity and the released ascorbate taken up. The loss of AA-2G from the perfusate and absorption of ascorbate was completely prevented by the addition of castanospermine, presumably through inhibition of the brush-border a-D-glucosidases [123]. These results suggest that the concentration of active L-ascorbate in humans and human cells, which like guinea pigs lack the pathway for synthesizing L-ascorbate from D-glucuronate, could be modulated by castanospermine. 7.3.2.2

Lysosomal glucosidases

The possibility of castanospermine affecting lysosomal α-D-glucosidase was shown in the above experiments. Intraperitoneal injection of castanospermine at greater than 1 mg/g body weight for 3 days into rats leads to intralysosomal accumulation of glycogen, as in Pompe disease, the lysosomal storage disease resulting from a genetic deficiency of lysosomal α-D-glucosidase [13]. Prolonged administration of castanospermine at 143.6 mg/kg/day for 28 days via sub-cutaneous osmotic minipumps did not cause clinical signs in rats but microscopic examination showed degenerative vacuolation of hepatocytes and skeletal myocytes as in

148

7 Synthesis and Biological Activity of Castanospermine...

Pompe disease [2O]. There was also mild vacuolation of the renal tubular and thyroid follicular epithelia. In these experiments with relatively high doses of castanospermine, gastrointestinal problems were avoided by removing sucrose from the diet. Although disruption of lysosomal turnover of glycogen is theoretically a hazard of any therapeutic use of castanospermine, it is probable that the doses used for therapeutic purposes will be too low to induce lysosomal storage [126]. Castanospermine also inhibits lysosomal /?-o-glucosidase (glucocerebrosidase), the enzyme deficient in the most common lysosomal storage disease, Gaucher disease [12]. However, the K1 for human fibroblast lysosomal /?-o-glucosidase is 1 mM, much higher than that for the corresponding α-D-glucosidase of 0.1 mM. Although fibroblasts in culture appear to be permeant to castanospermine, there was no evidence of accumulation of glucocerebroside [12]. Fibroblasts from patients with Gaucher disease type 2, the most severe form of the disease, also do not accumulate glucocerebroside. This is either due to the low turnover of glucosylceramide in fibroblasts or to diversion of undegradable glucosylceramide into biosynthetic pathways. There is a 1.5 to 2 fold increase in some other glycosphingolipids in Gaucher fibroblasts [127]. Two novel glycosphingolipids appeared in normal fibroblasts cultured in the presence of castanospermine, supporting this explanation. However, it is also possible that they arose from inhibition of the broad specificity /?-D-glucosidase, which is competitively inhibited by castanospermine but with a much higher K{ of 40 mM [12]. 7.3.2.3

Processing a-o-glucosidases

Glycoproteins play an important role in many cellular processes. These functions depend on the wide variety of asparagine N-linked glycans found on mature glycoproteins. This variety of structure results from the processing in the endoplasmic reticulum and Golgi apparatus of a common oligosaccharide precursor, which is transferred en bloc from a lipid carrier to newly synthesized proteins (for a review of glycoprotein biosynthesis see [128]). The first steps in the processing of N-linked glycans are the removal in the endoplasmic reticulum of the outermost a-l-^2 linked glucose by α-D-glucosidase I followed by the removal of the two a-1—>3 linked glucose residues by α-D-glucosidase IL Both of these processing α-D-glucosidases are inhibited by castanospermine and this has been shown to prevent processing of N-linked glycans and to lead to the production of glycans retaining the outer α-linked glucose residues [17]. This property of castanospermine has attracted much attention because manipulation of the processing of glycoproteins could be exploited for the prevention of disease or enhancement of a beneficial process. In consequence the effects of castanospermine and derivatives on many cellular processes, such as intracellular transport and targeting of proteins, cell surface receptors, cell-cell recognition, cell adhesion, viral replication (including HIV), the immune response, fertility and metastasis have been studied. It is impossible in a review of this length to describe comprehensively all the research carried out to investigate these effects, much of which has been covered in previous reviews [15, 94, 95, 107]. Therefore, emphasis will be placed on novel applications and recent attempts to understand the mechanism of action of castanospermine.

7.3 Biological Activities of Castanospermine and Derivatives

7.3.3

149

Mechanism of action

The precise effect of inhibition of α-D-glucosidases I and II by castanospermine on the structure of N-linked glycans, as opposed to cellular processes dependent on glycosylation, has only been thoroughly investigated in a few cases. The existence of an endomannosidase, which can remove the three α-D-linked glucose residues and a mannose residue by an alternative route, may be able to compensate for the inhibition of the α-D-glucosidases in some cells [129, 13O]. Cells also have a quality control process in the endoplasmic reticulum to ensure that only correctly glycosylated and folded proteins are transported to the Golgi for further processing and distribution to their site of action [131-134]. The molecular chaperones calnexin [135] and calreticulin [136] bind to monoglucosylated glycoproteins in the endoplasmic reticulum to ensure correct folding by preventing aggregation, oxidation, substitution and degradation of partially folded glycoproteins. Therefore, inhibition of α-D-glucosidase I by castanospermine would be expected to inhibit the binding of calnexin and calreticulin to glycoproteins with a consequent decrease in the efficiency of maturation of the glycoproteins and loss of function. This has been shown to be the case for a range of glycoproteins and cellular functions and may be the primary mechanism of action of castanospermine [137-142].

7.3.4

Anti-viral activity

Castanospermine was first reported to inhibit replication of the human immunodeficiency virus-1 (HIV-I) in CD4+ T cell cultures and to prevent syncitium formation at concentrations non-cytopathic to lymphocytes in 1987 [31, 32, 143]. The inhibition was reported to be synergistic with zidovudine (AZT, 3'-azido-3'-deoxythymidine) [144, 145] and other drugs [146]. A comparison of the anti-HIV activity of a series of castanospermine analogues revealed that 6-O-butanoylcastanospermine [BuCast, 109 (R=COPr)] was 20-30 times more active than castanospermine despite having an ICso 10 times greater than castanospermine for inhibition of a-Dglucosidase I [24, 146]. The explanation of this paradox is that BuCast is taken up into cells and absorbed more rapidly into the blood stream than castanospermine [147]. It is subsequently rapidly hydrolyzed enzymically to castanospermine i.e. it is a prodrug. Administration of the prodrug may avoid some of the gut toxicity of the parent compound. BuCast is currently in clinical trial as a potential antiHIV agent. The expression of the integrin LFA-I (CD 18/CD Ua), which may decrease cell adhesion of uninfected mononuclear leukocytes [148] is also decreased by BuCast and this may play a role in the prevention of transfer of HIV-1 from cell to cell [23]. Subsequently, a series of analogs of castanospermine was synthesized and tested for their inhibition of α-D-glucosidase I in vitro and in cells [16]. Although the analogs were weaker inhibitors against the purified enzyme in vitro, those with lipophilic side chains were more effective in the cells. Substitution of oxygen for methylene groups in the alkyl chains of N-decyldeoxynojirimycin was found to decrease its detergent properties without affecting

150

7 Synthesis and Biological Activity of Castanospermine...

the intracellular blocking of glycoprotein processing [149]. Such modification may also be applicable to castanospermine. The precise mechanism by which ao-glucosidase inhibitors decrease HIV-infectivity is not known but experiments with the closely related N-butyldeoxynojirimycin show that it impairs viral entry at a post-CD4 binding step [15O]. The prodrug, BuCast, may have several other important applications. Oral treatment of mice infected with herpes simplex virus-1 with 6-0-butanoyl castanospermine decreased infection in the brain and delayed development of lesions [21]. It also blocked growth of herpes simplex virus-2 [2O]. The basis of the activity of castanospermine against herpes simplex virus is probably blocking of the association of herpes simplex viral glycoproteins with calnexin [151].

7.3.5

Anti-cancer activity

Both catabolic and processing glycosidases are involved in the transformation of normal cells to cancer cells and in tumor cell invasion and migration. Many tumor cells display aberrant glycosylation due to an altered expression of glycosyltransferases [152] and it has been known for a long time that the levels of glycosidases are elevated in the sera of many patients with different tumors [153]. Secreted glycosidases may be involved in the degradation of the extracellular matrix in tumor cell invasion [154]. The lysosomal system is extremely active in cancer cells presumably reflecting the enhanced turnover of glycoproteins and other molecules. Castanospermine and other glycosidase inhibitors are being investigated as potential anti-cancer agents [155, 156]. Castanospermine has been shown to affect many of the properties of tumor cells in vitro and in vivo [34, 36, 37, 157]. Increased formation of new capillaries, angiogenesis, is a feature of many pathological processes including tumor growth. Castanospermine was found to inhibit tumor growth and angiogenesis in nude mice infected with EHSBAM tumor cells which form highly vascularized tumors [33]. Altered glycoproteins were found on endothelial cells, which had a decreased ability to migrate and invade the basement membrane. These experiments suggest that specific cell surface oligosaccharides are involved in angiogenesis and that inhibition of their formation by castanospermine may be a way of preventing tumor growth. The adhesion of human myeloma cells to endothelial cells [158] is enhanced by castanospermine whereas the interaction of integrins on carcinoma cells with fibronectin was decreased [159]. Castanospermine also impairs the transformation of chicken embryo fibroblasts by the virus, env-sea, by blocking transport to the cell surface of proteins encoded by the env-sea oncogene [16O]. Thus the consequences of inhibition of processing α-D-glucosidase by castanospermine in cancer pathology are diverse.

7.3 Biological Activities of Castanospermine and Derivatives

7.3.6

151

Modulation of other cellular processes

The decrease in the appearance on the cell surface of specific carbohydrate structures or glycoproteins by inhibition of processing α-D-glucosidase by castanospermine is being investigated as a means of modulating other important cellular processes. The anti-inflammatory properties of castanospermine in an adjuvantinduced rat model of arthritis have been attributed to prevention of the expression of leukocyte cell surface-bound enzymes or of adhesion molecules involved in the capture and retention of the leukocytes in the inflamed tissue [161]. Both of these changes would affect the passage of the leukocytes through the sub-endothelial basement membrane. Subsequent work showed that although castanospermine selectively inhibited the phorbol myristate-induced heparanase and sulfatase activity of endothelial cells, it did not inhibit the constitutive expression of enzymes for degradation of the extracellular matrix by non-stimulated cells [39]. The alteration in the expression of cell surface glycoproteins induced by castanospermine may be useful in modulating the immune response. The reduction in expression of adhesion molecules can prolong heart allograft survival in rats [41] and pancreatic duodenal allograft survival is prolonged in an experimental form of treatment for diabetes in rats [162]. Castanospermine can act synergistically with other drugs in prolonging allografts [42, 43]. Other cellular processes affected by the castanospermine-induced alteration in the expression of cell surface glycoproteins include interleukin-4 induced macrophage fusion and the formation of giant cells involving the mannose receptor [163], fusion of myoblasts to form multinucleated myotubes [164], oligodendrocyte differentiation in cell culture [165, 166], neurite outgrowth in vivo during development and regeneration [167] and gap junction formation [168]. Although glycosidases are very abundant in the male reproductive tract, and have a characteristic distribution along the epididymis, their function is not fully understood. Continuous administration of castanospermine to rats suppressed the epididymal α-D-glucosidase within 2 days but only decreased fertility transiently [169]. It was concluded that the epididymal α-D-glucosidase does not play a crucial role in the development of sperm fertility but may be involved in preparation of spermatozoa for storage. Castanospermine and its analogs have become standard tools for investigating properties of individual glycoproteins and cellular processes dependent on glycosylation. The response of an animal cell to castanospermine will depend on its function, physiological state, tissue environment and the concentration of castanospermine to which it is exposed. As most cells contain several susceptible glucosidases with different sub-cellular locations, pH-optima and inhibition constants, the effect may be complicated. However the extracellular and resultant intracellular concentration of castanospermine required to cause inhibition of the processing glucosidases appears to be much lower than that required to inhibit the lysosomal enzymes. Therefore, therapeutic applications exploiting inhibition of the processing or digestive glucosidases may not be complicated by induced storage of glycogen.

152

7 Synthesis and Biological Activity of Castanospermine...

References [1] L. D. Hohenschutz, E. A. Bell, P. J. Jewess, D. R Leworthy, R. J. Pryce, E. Arnold, J. Clardy, Phytochem., 1981, 20, 811-814. [2] R. J. Nash, L. E. Fellows, J. V. Bring, C. H. Stirton, D. Carter, M. P. Hegarty, E. A. Bell, Phytochem., 1988, 27, 1403-1404. [3] L. E. Fellows, G. W. J. Fleet, in Natural Products Isolation, (Eds.: G. H. Wagman, R. Cooper), Elsevier, 1989, p. 539-559. [4] L. E. Fellows, Pestic. ScL, 1986, 17, 602-606. [5] A. D. Elbein, R. J. Molyneux, in Alkaloids: Chemical and Biological Perspectives, (Ed.: S. W. Pelletier), John Wiley & Sons, New York, 1987, p. 1-54. [6] R. Saul, J. P. Chambers, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys, 1983, 227, 593-597. [7] P. L. Durham, J. E. Poulton, Plant PhysioL, 1989, 90, 48-52. [8] A. M. Scofield, J. T. Rossiter, P. Witham, G. C. Kite, R. J. Nash, L. E. Fellows, Phytochem., 1990, 29, 107-109. [9] R. Saul, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys., 1984, 230, 668-675. [10] A. M. Scofield, P. Witham, R. J. Nash, G. C. Kite, L. E. Fellows, Comp. Biochem. PhysioL, A: PhysioL, 1995, II2A, 187-196. [11] B. G. Winchester, I. Cenci di Bello, A. C. Richardson, R. J. Nash, L. E. Fellows, N. G. Ramsden, G. W. J. Fleet, Biochem. J., 1990, 269, 227-231. [12] I. Cenci di Bello, D. Mann, R. J. Nash, B. G. Winchester, in Lipid Storage Disorders Biological and Medical Aspects, (Eds.: R. Salvayre, L. Douste-Blazy, S. Gatt), Plenum Press, New York, 1988, p. 635-641. [13] R. Saul, J. J. Ghidoni, R. J. Molyneux, A. D. Elbein, Proc. Natl. Acad. ScL USA, 1985, 82, 93-97. [14] A. D. Elbein, CRC Crit. Rev. Biochem., 1984, 16, 21-49. [15] A. D. Elbein, FASEB. J., 1991, 5, 3055-3063. [16] M. S. Kang, P. S. Liu, R. C. Bernotas, B. S. Harry, P. S. Sunkara, Glycobiology, 1995, 5, 147-152. [17] Y. T. Pan, H. Hori, R. Saul, B. A. Sanford, R. J. Molyneux, A. D. Elbein, Biochemistry, 1983, 22, 3975-3984. [18] G. Palamarczyk, A. D. Elbein, Biochem. J., 1985, 227, 795-804. [19] H. Hori, Y. T. Pan, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys., 1984, 228, 525-533. [20] S. P. Ahmed, R. J. Nash, C. G. Bridges, D. L. Taylor, M. S. Kang, E. A. Porter, A. S. Tyms, Biochem. Biophys. Res. Commun., 1995, 208, 267-273. [21] C. G. Bridges, S. P. Ahmed, M. S. Kang, R. J. Nash, E. A. Porter, A. S. Tyms, Glycobiology, 1995, 5, 249-253. [22] D. L. Taylor, M. S. Kang, T. M. Brennan, C. G. Bridges, P S. Sunkara, A. S. Tyms, Antimicrob. Agents Chemother., 1994, 38, 1780-1787. [23] C. G. Bridges, T. M. Brennan, D. L. Taylor, M. McPherson, A. S. Tyms, Antiviral Res., 1994, 25, 169-175. [24] D. L. Taylor, P. S. Sunkara, P. S. Liu, M. S. Kang, T. L. Bowlin, A. S. Tyms, Aids, 1991, 5, 693-698. [25] R. M. Ruprecht, S. Mullaney, J. Anderson, R. Bronson, /. Acq. Imm. Def. Synd., 1989, 2, 149-157. [26] P. S. Sunkara, D. L. Taylor, M. S. Kang, T. L. Bowlin, P. S. Liu, A. S. Tyms, A. Sjoerdsma, The Lancet, 1989, 1206. [27] G. W. J. Fleet, A. Karpas, R. A. Dwek, L. E. Fellows, A. S. Tyms, S. Petursson, S. K. Namgoon, N. G. Ramsden, P. W. Smith, FEBS Lett., 1988, 237, 128-132. [28] A. Karpas, G. W. J. Fleet, R. A. Dwek, S. Petursson, S. K. Namgoong, N. G. Ramsden, G. S. Jacob, T. W. Rademacher, Proc. Natl. Acad. ScL USA, 1988, 85, 9229-9233.

References

153

[29] D. L. Taylor, L. E. Fellows, G. H. Farrar, R. J. Nash, D. Taylor-Robinson, M. A. Mobberley, T. A. Ryder, D. J. Jeffries, A. S. Tyms, Antiviral Res., 1988, 10, 11-26. [30] P. S. Liu, T. Bowlin, EPA 309952, 1989. [31] R. A. Gruters, J. J. Neefjes, M. Tersmette, R. E. Y. de Goede, A. TuIp, H. G. Huisman, F. Miedema, H. L. Ploegh, Nature, 1987, 330, 74-77. [32] B. D. Walker, M. Kowalski, W. C. Goh, K. Kozarsky, M. Krieger, C. Rosen, L. Rohrschneider, W. A. Haseltine, J. Sodroski, Proc. Natl Acad. ScL USA, 1987, 84, 8120-8124. [33] R. PiIi, J. Chang, R. A. Partis, R. A. Mueller, F. J. Chrest, A. Passaniti, Cancer Res., 1995, 55, 2920-2926. [34] M. A. Spearman, B. C. Ballon, J. M. Gerrard, A. H. Greenberg, J. A. Wright, Cancer Lett., 1991,60, 185-191. [35] S. P Sunkara, EPA 0373663, 1989. [36] G. K. Ostrander, N. K. Scribner, L. R. Rohrschneider, Cancer Res., 1988, 48, 1091-1094. [37] M. J. Humphries, K. Matsumoto, S. L. White, K, Olden, Cancer Res., 1986, 46, 5215-5222. [38] A. J. Bitonti, A. Sjoerdsma, P. P. McCann, EPA 423728, 1991. [39] M. R. Bartlett, W. B. Cowden, C. R. Parish, J. Leukocyte Biol, 1995, 57, 207-213. [40] D. O. Willenborg, C. R. Parish, W. B. Cowden, /. Neurological ScL, 1989, 90, 77-85. [41] A. D. Hibberd, P M. Grochowicz, Y. C. Smart, K. M. Bowen, D. A. Clark, W. B. Cowden, D. O. Willenborg, Transplant. Proc., 1997, 29, 1257-1258. [42] P M. Grochowicz, A. D. Hibberd, K. M. Bowen, D. A. Clark, G. Pang, W. B. Cowden, T. C. Chou, L. K. Grochowicz, Y. C. Smart, Transplant. Proc., 1997, 29, 1259-1260. [43] P. M. Grochowicz, A. D. Hibberd, K. M. Bowen, D. A. Clark, G. Pang, L. K. Grochowicz, D. O. Willenborg, W. B. Cowden, Transplant. Proc., 1995, 27, 355-356. [44] A. D. Hibberd, P. M. Grochowicz, Y C. Smart, K. M. Bowen, D. A. Clark, B. Purdon, D. O. Willenborg, W B. Cowden, Transplant. Proc., 1995, 27, 448-449. [45] B. L. Rhinehart, K. M. Robinson, A. J. Payne, M. E. Wheatley, J. L. Fisher, P. S. Liu, W. Cheng, Life ScL, 1987, 41, 2325-2331. [46] G. Trugnan, M. Rousset, A. Zweibaum, FEBS Lett., 1986, 795, 28-32. [47] K. Burgess, I. Henderson, Tetrahedron, 1992, 48, 4045-4066. [48] N. S. Kim, J. R. Choi, J. K. Cha, J. Org. Chem., 1993, 58, 7096-7099. [49] H. Ina, C. Kibayashi, J. Org. Chem., 1993, 58, 52-61. [50] V. Grassberger, A. Berger, K. Dax, M. Fecher, G. Gradnig, A. E. Stuetz, Liebigs Ann. Chem., 1993, 379-390. [51] H. S. Overkleeft, U. K. Pandit, Tetrahedron Lett., 1996, 37, 547-550. [52] H. S. Overkleeft, J. van Wiltenburg, U. K. Pandit, Tetrahedron Lett., 1993, 34, 2527-2528. [53] H. S. Overkleeft, J. van Wiltenburg, U. K. Pandit, Tetrahedron, 1994, 50, 4215-4224. [54] J. Mulzer, H. Dehmlow, J. Buschmann, P. Luger, /. Org. Chem., 1992, 57, 3194-3202. [55] Y. Okude, S. Hirano, T. Hiyama, H. Nozaki, J. Am. Chem. Soc., 1977, 99, 3179-3181. [56] T. Hiyama, Y. Okude, K. Kimura, H. Nozaki, Bull Chem. Soc. Jpn., 1982, 55, 561-568. [57] K. Burgess, D. A. Chaplin, I. Henderson, Y T. Pan, A. D. Elbein, J. Org. Chem., 1992, 57, 1103-1109. [58] K. Burgess, D. A. Chaplin, Tetrahedron Lett., 1992, 33, 6077-6080. [59] P. K. Jadhav, F. J. Woerner, Tetrahedron Lett., 1994, 35, 8973-8976. [60] H. Zhao, D. R. Mootoo, J. Org. Chem., 1996, 61, 6762-6763. [61] S. Jeganathan, P. Vogel, J. Org. Chem., 1991, 56, 1133-1142. [62] Y. Chen, P. Vogel, Tetrahedron Lett., 1992, 33, 4917-4920. [63] Y. Chen, P. Vogel, J. Org. Chem., 1994, 59, 2487-2496. [64] K. Ninomiya, T. Shioiri, S. Yamada, Tetrahedron, 1974, 30, 2151-2157. [65] A. A. Bell, R. J. Nash, G. W. J. Fleet, Tetrahedron: Asymm., 1996, 7, 595-606. [66] A. A. Bell, L. Pickering, A. A. Watson, R. J. Nash, R. C. Griffiths, M. G. Jones, G. W. J. Fleet, Tetrahedron Lett., 1996, 37, 8561-8564.

154

7 Synthesis and Biological Activity of Castanospermine...

[61] G. Gradnig, A. Berger, V. Grassberger, A. E. Stuetz, G. Legler, Tetrahedron Lett., 1991, 32, 4889-4892. [68] W. H. Pearson, E. J. Hembre, Tetrahedron Lett., 1993, 34, 8221-8224. [69] P. Zhou, H. M. Salleh, J. F. Honek, /. Org. Chem., 1993, 58, 264-266. [70] P. Herczegh, I. Kovacs, L. Szilagyi, T. Varga, Z. Dinya, E Sztaricskai, Tetrahedron Lett., 1993, 34, 1211-1214. [71] S. F. Martin, H. J. Chen, C. P. Yang, J. Org. Chem., 1993, 58, 2867-2873. [72] M. Maggini, M. Prato, M. Ranelli, G. Scorrano, Tetrahedron Lett., 1992, 33, 6537-6540. [73] C. Lee, K. Y. Sim, J. Zhu, Tetrahedron, 1992, 48, 8541-8544. [74] A. H. Siriwardena, A. Chiaroni, C. Riche, D. S. Grierson, J. Org. Chem., 1992, 57, 5661-5666. [75] D. Marek, A. Wadouachi, R. Uzan, D. Beaupere, G. Nowogrocki, G. Laplace, Tetrahedron Lett., 1996, 37, 49-52. [76] P. S. Liu, B. L. Rhinehart, EPA 202661, 1986. [77] P. S. Liu, J. K. Daniel, B. L. Rhinehart, EPA 297534, 1989. [78] N. G. Landmesser, H. Tsui, C. R. King, L. A. Paquette, Synth. Commun., 1996, 26, 2213-2221. [79] P. S. Liu, W. J. Hoekstra, C. R. King, Tetrahedron Lett., 1990, 31, 2829-2832. [80] A. L. Margolin, D. L. Delinck, M. R. Whalon, /. Am. Chem. Soc., 1990, 772, 2849-2854. [81] W. K. Anderson, R. A. Coburn, A. Gopalsamy, T. J. Howe, Tetrahedron Lett., 1990, 31, 169-170. [82] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, Tetrahedron, 1994, 50, 2131-2160. [83] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, O. Hartley, B. G. Winchester, Tetrahedron, 1997, 53, 245-268. [84] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, Tetrahedron, 1995, 57, 12611-12630. [85] J. M. Mason, P. C. Tyler, Unpublished results. [86] R. H. Furneaux, J. M. Mason, P. C. Tyler, Tetrahedron Lett., 1994, 35, 3143-3146. [87] A. Gopalsamy, W. K. Anderson, Tetrahedron Lett., 1991, 32, 4669-4670. [88] R. H. Furneaux, J. M. Mason, P. C. Tyler, Tetrahedron Lett., 1995, 36, 3055-3058. [89] B. L. Rhinehart, K. M. Robinson, C. R. King, P. S. Liu, Biochem. Pharmacol, 1990, 39, 1537-1543. [90] P. S. Liu, C. H. King, Synth. Commun., 1992, 22, 2111-2116. [91] W. K. Anderson, P Tan, Synth. Commun., 1994, 24, 3081-3089. [92] A. Hempel, N. Camerman, D. Mastropaolo, A. Camerman, J. Med. Chem., 1993, 36, 4082-4086. [93] A. M. Scofield, L. E. Fellows, R. J. Nash, G. W. J. Fleet, Life Sd., 1986, 39, 645-650. [94] B. Winchester, G. W. J. Fleet, Glycobiology, 1992, 2 199-210. [95] G. S. Jacob, Curr. Opin. Struct. Biol, 1995, 5, 605-611. [96] D. Hendry, L. Hough, A. C. Richardson, Tetrahedron Lett, 1987, 28, 4597-4600. [97] D. Hendry, L. Hough, A. C. Richardson, Tetrahedron Lett., 1987, 28, 4601-4604. [98] R. J. Molyneux, J. E. Tropea, A. D. Elbein, /. Nat. Prod., 1990, 53, 609-6 14. [99] J. E. Tropea, R. J. Molyneux, G. P. Kashal, Y. T. Pan, M. Mitchell, A. D. Elbein, Biochemistry, 1989, 28, 2027-2034. [100] R. J. Molyneux, L. F. James, Science, 1982, 190-191. [101] S. van Weely, M. Brandsma, A. Strijland, J. M. Tager, J. M. Aerts, Biochim. Biophys. Acta, 1993, 1181, 55-62. [102] E. Tsujii, M. Muroi, N. Shiragami, A. Takatsuki, Biochem. Biophys. Res. Commun., 1996, 220, 459-466. [103] R. J. Molyneux, Y. T. Pan, A. Goldmann, D. A. Tepfer, A. D. Elbein, Arch. Biochem. Biophys., 1993, 304, 81-88. [104] S. Hiraizumi, U. Spohr, R. G. Spiro, J. Biol. Chem., 1994, 269, 4697-4700.

References

155

[105] T. G. Cooper, C. H. Yeung, D. Nashan, F. Jockenhovel, E. Nieschlag, J. AndroL, 1993, 13, 297-305. [106] S. L. Everist, Poisonous Plants of Australia, Angus Robertson, Sydney, 1974, 283-285. [107] G. S. Jacob, M. L. Bryant, Perspectives in Drug Discovery and Design, 1993, 7, 211-224. [108] A. M. Scofield, L. E. Fellows, R. J. Nash, G. W. J. Fleet, Life ScL, 1986, 39, 645-650. [109] Y. T. Pan, J. Ghidoni, A. D. Elbein, Arch. Biochem. Biophys., 1993, 303, 134-144. [110] S. V. Kyosseva, Z. N. Kyossev, A. D. Elbein, Arch. Biochem. Biophys, 1995, 316, 821-826. [Ill] K. M. Robinson, E. W. Heineke, M. E. Begovic, /. Nutr., 1990, 120, 105-111. [112] B. L. Rhinehart, K. M. Robinson, C. H. King, P. S. Liu, Biochem. Pharmacol, 1990, 39, 1537-1543. [113] J. A. Balfour, D. McTavish, Drugs, 1993, 46, 1025-1054. [114] R. F. Coniff, J. A. Shapiro,T. B. Seaton, Arch. Intern. Med., 1994, 154, 2442-2448. [115] J. P. J. E. SeIs, P. J. Klingma, B. H. R. Wolffenbuttel, P. P. C. A. Menheere, J. H. Branholte, A. C. N. Kruseman, Netherland J. Med., 1994, 44, 198-201. [116] B. C. Campbell, R. J. Molyneux, K. C. Jones, J. Chem. Ecol, 1987, 13, 1759-1770. [117] L. E. Fellows, G. Kite, R. Nash, M. Simmonds, A. Scofield, in Plant Nitrogen Metabolism., (Eds.: J. E. Poulton, J. T. Romeo, E. E. Conn), Plenum Press, 1989, 395-427. [118] A. P. Valaitis, D. F. Bowers, Insect. Biochem. MoI. Biol, 1993 23, 599-606. [119] D. L. Dreyer, K. C. Jones, R. J. Molyneux, J. Chem. Ecol., 1985, 11, 1045-1051. [120] R. J. Nash, S. V. Evans, L. E. Fellows, E. A. Bell, in Plant Toxicology, (Eds.: A. A. Seawright, M. P. Hegarty, L. F. James, R. F. Keeler), Queensland Poisonous Plants Committee, Yeerongpilly, Australia, 1985, p. 309-314. [121] I. Yamamoto, N. Muto, K. Murakami, J. Akiyama, J. Nutr., 1992, 722, 871-877. [122] M. Tanaka, N. Muto, E. Gohda, I. Yamamoto, Jpn. J. Pharmacol., 1994, 66, 451-456. [123] H. Wakamiya, E. Suzuki, I. Yamamoto, M. Akiba, N. Arakawa, J. Nutr. ScL VitaminolTokyo, 1995, 41, 265-272. [124] M. Yanagida, E. Gohda, I. Yamamoto, Life. ScL, 1996, 59, 2075-2081. [125] B. L. Stegelmeier, R. J. Molyneux, A. D. Elbein, L. F. James, Vet. PathoL, 1995, 32, 289-298. [126] B. L. Rhinehart, M. E. Begovic, K. M. Robinson, Biochem. Pharmacol., 1991, 41, 223-228. [127] M. Saito, A. Rosenberg, J. Biol. Chem., 1985, 260, 2295-2300. [128] R. Kornfeld, S. Kornfeld, Annu. Rev. Biochem., 1985, 54, 631-664. [129] S. E. H. Moore, R. G. Spiro, J. Biol. Chem., 1990, 265, 13104-13112. [130] C. Rabouille, R. G. Spiro, /. Biol. Chem., 1992, 267, 11573-11578. [131] M. Sousa, M. A. Ferrero-Garcia, A. J. Perodi, Biochemistry, 1992, 31, 97-105. [132] N. Ahluwalia, J. J. Bergeron, I. Wada, E. Degen, D. B. Williams, J. Biol. Chem., 1992, 267, 10914-10918. [133] C. Hammond, I. Braakman, A. Helenius, Proc. Natl. Acad. ScL USA, 1994,91,913-917. [134] J. D. Oliver, F. J. van-der-Wal, N. J. Bulleid, S. High, Science, 1997, 275, 86-88. [135] F. E. Ware, A. Vassilakos, P. A. Peterson, M. R. Jackson, M. A. Lehrman, D. B. Williams, J. Biol. Chem., 1995, 270, 4697-4704. [136] I. Wada, S. Imai, M. Kai, F. Sakane, H. Kanoh, J. Biol. Chem., 1995, 270, 2029820304. [137] K. P. Kearse, D. B. Williams, A. Singer, EMBO J., 1994, 13, 3678-3686. [138] C. Hammond, A. Helenius, Science, 1994, 266, 456-458. [139] Q. Zhang, M. Tector, R. D. Slater, J. Biol. Chem., 1995, 270, 3944-3948. [140] D. N. Hebert, B. Foellmer, A. Helenius, EMBO J., 1996, 75, 1961-1968. [141] A, Vassilakos, M. F. Cohen-Doyle, P. A. Peterson, M. R. Jackson, D. B. Williams, EMBO J., 1996, 75, 1495-1506. [142] B. Sadasivan, P. J. Lehner, B. Ortmann, T. Spies, P. Cresswell, Immunity, 1996, 5, 103-114.

156

7 Synthesis and Biological Activity of Castanospermine...

[143] A. S. Tyms, E. M. Berrie, T. A. Ryder, R. J. Nash, M. P. Hegarty, D. L. Taylor, M. A. Mobberley, J. M. Davis, E. A. Bell, D. J. Jeffries, D. Taylor-Robinson, L. E. Fellows, The Lancet, 1987, 1025-1026. [144] V. A. Johnson, B. D. Walker, M. A. Barlow, T. J. Paradis, T. C. Chou, M. S. Hirsch, Antimicrob. Agents Chemother., 1989, 33, 53-57. [145] A. M. Sorensen, C. Nielsen, L. R. Mathiesen, J. O. Nielsen, J. E. Hansen, Scand. J. Infect. Dis., 1993, 25, 365-371. [146] D. C. Montefiori, W. E. Robinson Jr., W. M. Mitchell, AIDS Res. Hum. Retroviruses, 1989, 5, 193-203. [147] M. S. Kang, Glycobiology, 1996, 6, 209-216. [148] C. G. Bridges, D. L. Taylor, M. Kang, T. M. Brennan, A. S. Tyms, Glycobiology, 1995, 5, 243-247. [149] A. Tan, I. Van den Broek, J. Bolscher, D. J. Vermaas, L. Pastoors, C. van Boeckel, H. Ploegh, Glycobiology, 1994, 4, 141-149. [150] P. B. Fischer, M. Collins, G. B. Karlson, W. James, T. D. Butters, S. I. Davis, S. Gordon, R. A. Dwek, R M. Platt, J. ViroL, 1995, 5791-5797. [151] Y. Yamashita, M. Yamada, T. Daikoku, H. Yamada, A. Tadauchi, T. Tsurumi, Y. Nishiyama, Virology, 1996, 225, 216-222. [152] S. Hakomori, Cancer Res., 1985, 45, 2405-2414. [153] J. W. Woollen, P. Turner, CUn. Chim. Acta, 1965, 12, 671-683. [154] R. J. Bernacki, M. J. Niedbala, W. Korytnyk, Cancer and Metastasis Reviews, 1985, 4, 81-102. [155] K. Olden, P. Breton, K. Grzegorzewski, Y. Yasuda, B. L. Cause, O. A. Oredipe, S. A. Newton, S. L. White, Pharmac. Ther., 1991, 50, 285-290. [156] P. E. Goss, J. Baptiste, B. Fernandes, J. W. Dennis, Cancer Res., 1994, 54, 1450-1457. [157] A. Hadwiger, H. Niemann, A. Kabisch, H. Bauer, T. Tamura, EMBO J., 1986, 5, 689-694. [158] T. Nishiura, T. Karasuno, H. Yoshida, H. Nakao, M. Ogawa, Y. Horikawa, M. Yoshimura, Y. Okajima, Y. Kanakura, Y. Matsuzawa, Blood, 1996, 88, 3546-3554. [159] S. Ringeard, J. Harb, F. Gautier, J. Menanteau, K. Meflah, J. Cell. Biochem., 1996, 62, 40-49. [160] A. J. Crowe, M. J. Hayman, Cell Growth Differ., 1993, 4, 403-410. [161] D. O. Willenborg, C. R. Parish, W. B. Cowden, Immunol. Cell. Biol, 1992, 70, 369-377. [162] P. M. Grochowicz, K. M. Bowen, A. D. Hibberd, W. B. Cowden, D. O. Willenborg, Transplant Proc., 1992, 24, 2295-2296. [163] A. K. McNaIIy, K. M. DeFife, J. M. Anderson, Am. J. Pathol, 1996, 149, 975-985. [164] J. C. Jamieson, S. Wayne, R. S. BeIo, J. A. Wright, M. A. Spearman, Biochem. Cell. Biol, 1992, 70, 408-412. [165] S. Ishii, J. J. Volpe, Dev. Neurosci., 1992, 14, 221-229. [166] N. R. Bhat, P. Zhang, J. Neurosci. Res., 1994, 39, 1-10. [167] C. Abotomey, W. B. Cowden, C. E. Hill, Brain. Res. Dev., 1994, 79, 101-114. [168] Y. Wang, P. P. Mehta, Eur. J. Cell. Biol., 1995, 67, 285-296. [169] C. H. Yeung, T. G. Cooper, /. Reprod. Fertil, 1994, 102, 401-410.

8

Some Reflections on Structure-Activity Relationships in Glycosidase-Inhibiting Iminoalditols and Iminosugars ARNOLD E. STUTZ

8.1 Introduction Since Paulsen's fundamental synthetic and analytical work [1] on sugars with nitrogen instead of oxygen in the ring and the first discovery of such a natural product [2], numerous iminosugar-related natural products with glycosidase-inhibitory properties have been discovered or synthesized. Triggered by the range of exciting biochemical properties of iminosugars, many groups have invested tremendous effort in this field and contributed to the swiftly emerging research area related to the chemical and biochemical synthesis of these substances and non-natural derivatives thereof. Consequently, several hundreds of compounds have been discovered or synthesized over the past 30 years. Pioneering syntheses have been conducted by Fleet and his group and Legler and co-workers, as well as Wong and his team, just to mention a very few of the international leaders in this context. These researchers and other synthetically oriented groups, frequently in collaboration with eminent biochemists such as Elbein, Winchester, or Withers, have contributed to a vast array of biochemical data on the class of compounds under consideration. The quantity of crystallographic data relating to glycosidases and their active sites is only slowly increasing due to the inherent problems experienced with the crystallization of these frequently membrane-bound and sensitive enzymes. Consequently, information for rationalizing inhibitor properties and subsequent drug design is still scarce. Thus, chemists have attempted to define the structure-activity relationships of glycosidase inhibitors by exploiting the rapidly increasing amount of information on enzyme inhibitory activities. The aim of this research is to gain a better understanding of enzymatic glycoside hydrolysis and the design of novel derivatives with possibly advantageous biological properties in terms of activities as well as selectivities. Due to the large variety of glycosidases, as well as assay conditions (or the lack of specification of the latter), reported in the literature, these tasks have been hampered by difficulties in comparing and meaningfully interpreting inhibitory data from different sources. Even comparing activities of the same compound or closely related derivatives can prove problematic. Powerful inhibitors of a-L-fucosidases, D-galactosidases or Iminosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

158

8 Some Reflections on Structure-Activity Relationships...

Figure 8-1.

D-hexosaminidases exhibit inhibitory constants or related values in the nanomolar range, while o-mannosidase inhibitors are found to be efficient up to the 10 micromolar level. Parameters such as temperature, pH value, buffer system employed, or even the presence of ions such as chloride, can significantly influence the outcome of kinetic measurements. Consequently, proven inhibitors, such as 1-deoxynojirimycin (1) in the case of D-glucosidase inhibition, should always be employed as standards to validate screening procedures. Additional problems in comparing data arise from the various ways in which researchers choose to publish their results, such as inhibitory constants (Ki)9 inhibitory concentration (for example, ICso), or percent inhibition at a defined inhibitor concentration such as 1 mM. Similar glycosidase inhibitory activities of inhibitors can arise from either the similar arrangement of functional groups around the same carbon framework or, more importantly, by the similar three-dimensional alignment of functional groups attached to structurally different carbon skeletons. The latter relationship of active compounds, although sometimes obvious to the beholder, can frequently be not easily appreciated at first glance. Furthermore, some of the molecules under consideration appear to have (at least) two means of binding to susceptible active sites. A noteable example is found with D-mannosidase inhibitors which are also usually good inhibitors of α-L-fucosidases. In selected cases, the application of simple symmetry operations suggests that such alternative binding modes should not be entirely disregarded for cases not experimentally established such as with D-glucosidases and some of their inhibitors. Such alternative binding modes might also be found to be of significance for the work of chemists synthesizing non-natural analogs of inhibitors which are regio- and stereoselectively modified at 'strategic' positions to improve or enhance their biochemical properties. In the following, from the viewpoint of a preparatively oriented organic chemist, selected and hopefully meaningful inhibitory data of a range of typical inhibitors and important derivatives frequently dealt with in the literature will be briefly discussed and suggestions made for the interpretation of these data. For inhibitors of o-galactoses, only limited and fairly obvious conclusions on structure-activity relationships will be mentioned as a basis for discussion. D-Mannosidases and their inhibitors have attracted considerably more interest in the past and allow a somewhat better understanding of the stereochemical requirements for good inhibition, as demonstrated by available data and previous discussions of this topic based on inhibitory information as well as computer-aided molecular modeling. Due to the comparably larger amount of information available for D-glucosidase inhibitors, as well as the possibility of less obvious three-dimensional interactions in the respective active site, these compounds will be dealt with in more detail than previous groups of inhibitors deemed important as examples in this context.

159

8.2 D-Galactosidase Inhibitors

It should be noted that, inherently, such considerations always contain elements of uncertainty and are flawed by over-simplification. In the case of D-glucosidase inhibitors, such as castanospermine, some conclusions drawn are merely speculative at this point in time, but hopefully will be supported experimentally in the near future.

8.2 D-Galactosidase Inhibitors 8.2.1

Piperidine derivatives

8.2.1.1

5-Amino-5-deoxy-D-galactose, l,5-dideoxy-l,5-imino-D-galactitol (1-deoxygalactonojirimycin) and epimers

In 1980, Paulsen and co-workers reported the synthesis of l,5-dideoxy-l,5-iminoo-galactitol (3) the epimer at C-4 of 1-deoxynojirimycin and, in preliminary screening, reported its inhibitory activity against pig gut /?-galactosidases [3]. A new synthetic approach and an in-depth investigation into the glycosidase inhibitory properties of 5-amino-5-deoxy-D-galactose (2) and 1-deoxy derivative (3) was carried out by Legler and Pohl [4]. These workers found high activities for both compounds against a variety of a- as well as /?-galactosidases, with K1 values well in the nanomolar range. Ganem and co-workers [5] synthesized 3 using a different approach which employed their intramolecular aminomercuration protocol and found excellent inhibitory activity of α-galactosidase from green coffee beans as well as human placental ceramide trihexosidase (!€50 4 nM at pH 5). In 1987, 'galactonojirimycin' (5-amino-5-deoxy-D-galactose, 2) was discovered in fermentation broths of Streptomyces lydicus PA-5725 and named galactostatin by its Japanese discoverers [6], who also noted the comparably low toxicity of this compound. At first sight, and somewhat surprising for the investigators [7], 1,5-dideoxy1,5-imino-D-allitol (4), an epimer of compound 3 at C-3 and C-4, was found to H

?

,OH

^OH

π H

HOA^-^A, Η0\

°

R

R=OH

2

R=H

3

|

OH

OH

OH 4

° OH Figure 8-2.

6

R =H

7

R = OH 8

160

8 Some Reflections on Structure—Activity Relationships...

exhibit comparably low but distinct inhibitory activity with bovine liver cytosolic /?-galactosidase (ICso 300 μΜ) as well as rat intestine lactase (27 μΜ). Even more active was the corresponding 2-deoxy analog, l,2,5-trideoxy-l,5-imino-D-n7?<9hexitol 5 (Ki 1.9 μΜ with the lactase and 1.5 μΜ with the cytosolic enzyme) [7]. Another diastereomer of compound 3, l,5-dideoxy-l,5-imino-L-mannitol (6) was also discovered to be a good inhibitor of α-galacosidase from Aspergillus niger (Ki 8 μΜ) [8]. This activity was convincingly rationalized by the close similarity of this molecule with inhibitor 3 after lateral inversion of the conventionally drawn structure (Figure 8-2). 8.2.1.2

Epimers of isofagomine

Structurally related to D-galactose, epimer 7 [9] of isofagomine [10] was synthesized shortly after the patent compound and found to be a powerful inhibitor of /?-galactosidase from A. oryzae (ICso 12 nM at pH 6.8) and from almonds (0.19 μΜ). This observation demonstrated the higher susceptibility of /?-glycosidases for inhibitors with the basic nitrogen in the C-I position of their natural substrates. Somethat less potent, was the corresponding derivative with a tertiary hydroxyl group at the branching point of the carbon chain 8 [U].

8.2.2

Important pyrrolidine derivatives

2,5-Dideoxy-2,5-imino-D-galactitol (9) was synthesized by Wong and co-workers [12] by employing previously established chemical and chemoenzymatic approaches. This compound exhibited powerful activity against a-galactosidase from green coffee beans (Ki 0.05 μΜ at pH 5.5), albeit more than one order of magnitude less than the corresponding piperidine-derived l,5-dideoxy-l,5-imino-Dgalactitol (2). 2,5-Dideoxy-2,5-imino-D-mannitol (10) was shown to be about as powerful as 1,2,5 -trideoxy-1,5 -imino-D-ribo-hexitol (5, ICso around 3 μΜ with rat intestine lactase and bovine liver cytosolic /?-glucosidase) [7]. l,4-Dideoxy-l,4-imino-D-

10

12

11

13

Figure 8-3.

8.3 D-Mannosidase Inhibitors

161

lyxitol (11) was prepared by Fleet and his group [13] and found to be a good inhibitor of the α-galactosidase from coffee beans (^i 0.1 μΜ). l,4-Dideoxy-l,4-imino-oribitol (12), yet one more example for an 'upside-down' binding mode, showed good activity with rat intestine lactase (ICso 17 μΜ) (s. picture 8-1, page 163) [7]. l,4-Dideoxy-l,4-imino-L-iditol (13), synthesized by Lundt and her group [14], gave 95 % inhibition at 1 mM with human liver lysosomal/?-galactosidase.

8.2.3

Tetrahydroxyazepanes

Recently, the symmetric tetrahydroxyazepane derivatives l,6-dideoxy-l,6-iminoo-galactitol (14) and -L-iditol (15) were reported [15] to be inhibitors of a-galactosidase from green coffee beans (Ki 220 and 67 μΜ, respectively) and Aspergillus niger /?-galactosidase (50 and 6.5 μΜ, respectively).

Figure 8-4.

8.2.4

Conclusions

The examples under consideration suggest that inhibitors of o-galactosidases require at least three hydroxyl groups around the respective ring system in order to be active. Due to the fact that significant numbers of compounds bearing nonnatural substituents have not been synthesized to date, conclusions can only be drawn as to the vital importance of two cis-oriented hydroxyl groups mimicking the diol system of 3-OH and 4-OH of D-galactopyranose. In some cases, the equivalent of 2-OH is absent in the inhibitor, while in other compounds no apparent substitute for 6-OH can be detected. This suggests that either of these hydroxyl groups can be modified or removed, albeit on account of the pronounced inhibitory activity of the parent compound. Furthermore, in general it can be concluded that piperidine derivatives are more active than stereochemically comparable systems featuring a five- or seven-membered ring framework.

8.3 D-Mannosidase Inhibitors 8.3.1

Piperidine derivatives

The closest structural relatives of D-mannopyranose in the realm of compounds under consideration are 5-amino-5-deoxy-D-mannose (mannojirimycin, 16),

162

8 Some Reflections on Structure-Activity Relationships... 0

OH

Ηί

R = OH 16 R = H 17

18

OH

f

19

Figure 8-5.

R=F 20 R = OEt 21

nu

22

which exhibits a ^i value of 1.2 μΜ with jack bean α-mannosidase at pH 5.5, and 1deoxymannojirimycin (17). This epimer of compound 1 is a natural product first identified in the legumes Lonchocarpus sericeus and L. costariciensis [16]. Due to its interesting glycosidase inhibitory properties (K{ ^8O μΜ with jack bean amannosidase at pH 5.5), several syntheses of this compound have been devised [17]. The inherent close structural analogy with L-fucose also makes it a comparably poor α-L-fucosidase inhibitor (K1 9.5 μΜ with α-L-fucosidase from bovine epididymis at pH 6; for comparison: 1-deoxy-L-fuconojirimycin, 18, exhibits K1 0.003 μΜ)) [18]. Non-natural analogs bearing the characteristic 2,3 -cis orientations of substituents include compounds such as 6-deoxy [19] (no inhibitory data given), 4deoxy (19) [20] (K{ 620 μΜ with jack bean α-mannosidase at pH 5.5), 2-deoxyfluoro (20) [19], as well as 2-O-ethyl derivatives (21) [19]. None of these substances has been found to exhibit enzyme inhibitory activity close to that the parent compound. l,5-Dideoxy-l,5-imino-D-arabinitol or (likewise) -lyxitol (22), lacking the hydroxymethyl group at C-5, is also devoid of mannosidase inhibitory activity [21].

8.3.2

Pyrrolidine derivatives

8.3.2.1

l,4-Dideoxy-l,4-imino-D-mannitol and derivatives

The pyrrolidine-type iminosugar l,4-dideoxy-l,4-imino-D-mannitol (23) was found to exert powerful inhibitory against α-mannosidase from jack bean (ICso 0.5 μΜ at pH 4.5, K1 0.76 μΜ) as well as other enzymes [22]. Subsequently, this derivative was found to be more potent against this particular enzyme than swainsonine.

8.3 D-Mannosidase inhibitors

163

Similarity high activities were found for the 6-deoxy derivative (24) with jack bean a-mannosidase (K, 0.5 uM at pH 5.0) [23] as well as both the 6-deoxy and the 6-deoxyfluoro (25) analogs with human a-mannosidases [24]. A «fj/--derivative which lacks C-6, 1.4-dideoxy-I,4-imino-D-lyxitol (11), can also be regarded as a monocyclic highly truncated swainsonine analog. Under the same conditions, this derivative had an IC.so of only 14 uM with jack bean a-mannosidase, but was

164

S Some Reflections on Structure-Activity Relationships. •

OH

R = OH 23 R = H 24

R =F

26

27

25

-p.

,

0 Figure 8-6.

found to be a potent inhibitor of α-galactosidase from green coffee beans [13]. Its epimer at C-3, l,4-dideoxy-l,4-imino-D-arabinitol (26), exhibited a Ki value of 100 μΜ. These data demonstrate the importance of two vicinal cis-oriented hydroxyl groups for effective inhibition [13].

8.3.3

Tetrahydroxyazepanes

Recently, it was demonstrated that selected tetrahydroxyazepanes such as 1,6dideoxy-l,6-imino-D-glucitol (27) and the corresponding L-ido-configured analog 15 inhibit α-mannosidase from jack bean with A^-190 μΜ and 26 μΜ, respectively, at pH 6.8 [15]. For the Ό-gluco epimer a similar value was obtained by other workers (Ki 310 μΜ at pH 5.5) [25]. At a pH of 4.5, compound 27 was found to be inactive [26].

8.3.4

Bicyclic systems

8.3.4.1

Swainsonine

Swainsonine (28) was first isolated [27] from the Australian legume Swainsona canescens and discovered as the causative agent of locoism in cattle having been fed with Astragalus species [28]. This trihydroxylated indolizidine alkaloid was found to be a highly potent inhibitor of D-mannosidases from a variety of sources (Ki values in the micromolar and submicromolar range) [29, 3O]. This activity was contributed to by the arrangement of functional groups putatively mimicking the stereochemistry as well as charge distribution of a 'flap-up' D-mannopyranosyl oxocarbonium ion [31]. Some epimers of swainsonine have also been found to exhibit high mannosidase inhibitory against α-mannosidase from human liver [32-34]. Due to swainsonine's

HO, 28

R = H, R1 = OH

R = OH, R1 = H

29

so

Figure 8-7.

165

8.3 D-Mannosidase Inhibitors

interference with mannosidases of glycoprotein trimming, the compound exhibits interesting biological activities. 8.3.4.2

Miscellaneous

Interestingly, jack bean α-mannosidase was inhibited by the tetrahydroxyquinolizidine 29 with a K1 of 360 μΜ (pH 5.5). As the corresponding epimer at C-9 (30) exhibited K{ 520 μΜ under the same conditions [20], the better interaction of the former with the enzyme under consideration could be attributed to the cis-relationship of the hydroxyl groups at C-I and C-9.

8.3.5

Structure-activity relationships

Structure-activity relationships in D-mannosidase inhibitors have been discussed by Winkler [35] as well as Winkler and Holan [3Ia] on the basis of computeraided molecular modeling. This topic has also been addressed by van den Broek and co-workers [36], Asano and his group [7], and Winchester and colleagues [24] based on their inhibition of mammalian enzymes. All of these groups agree on the vital importance of inhibitors to mimick closely the steric properties of the putative 'flap-up' mannopyranosyl oxocarbonium ion in order to exhibit good inhibition [31, 35]. It is interesting to note that most of the excellent D-mannosidase inhibitors under consideration require only three hydroxyl groups for their respective activity in addition to a somewhat extended carbon framework. For example, the hydroxyl groups of the powerful inhibitor swainsonine (28) superimpose nicely on 2-OH, 3-OH and 6-OH of the mannopyranosyl oxocarbonium ion [3Ia]. The same is true for l,4,6-trideoxy-l,4-imino-D-mannitol (24) and its 6deoxyfluoro analog 25, the hydroxyl group at C-5 of the inhibitors matching position 6-OH. (Pictures 8-2.) In these three cases, 4-OH of the oxocarbonium ion appears not to take part in the binding/recognition process in the active site of the enzyme as none of these inhibitors bears an equivalent substituent. It is also interesting to compare the mannosidase inhibitory power of a compound which is one carbon shorter than the previously mentioned derivatives, l,4-dideoxy-l,4-imino-D-lyxitol (11), with compound 28 as well as structures 24 and 25. Despite the fact that these compounds are isosteric and the numbers and positions of hydroxyl groups are perfectly superimposable, iminopentitol 11, is about four orders of magnitude less active than swainsonine and 30 times less powerful with α-mannosidase from jack bean than the 1,4-deoxyiminomannitols under consideration, employing the same assay conditions.

HO. HC

Figure 8-8.

31

166

8 Some Reflections on Structure-Activity Relationships

8.3 D-Mannosidase Inhibitors

Figure 8-9.

167

32

Although 1-deoxymannojirimycin (17) in the conformationally preferred 4Ci chair does not fit the 'flap-up' oxocarbonium ion template very well, the boat conformation with the substituents at C-3 as well as C- 4 in axial positions superimposes perfectly. Not quite clear is the reason for the lack of activity [20] of 1,4,5trideoxy-l,5-imino-D-/;y;c0-hexitol (31), the 4-deoxy derivative of 1-deoxynojirimycin, which nicely fits the proposed steric requirements. The only significant difference is the somewhat out-of-plane position of the ring nitrogen and a slightly different orientation of the free lone electron pair. In all the other inhibitors this lone electron pair is positioned axially and, in 1-deoxymannojirimycin and related compounds, points across the ring towards 3-OH. Inspection of Dreiding models as well as computer-aided superpositions strongly support this view. Although not obvious at first sight, a relationship between the conformationally highly flexible L-ido-configured tetrahydroxyazepane 15 with the mannopyranosyl oxocarbonium ion can be established by matching 2-OH, 3-OH, as well as 4-OH of the seven-membered ring with 2-OH, 3-OH, and 6-OH of the template. In this orientation, 5-OH of the inhibitor is fairly remote from the areas under consideration and probably not contributing to the biological activity. This hypothesis certainly needs to be confirmed by probing the inhibitory power of the corresponding 5 -deoxy-o-jryfo-configured iminoheptitol 32. One shortcoming of such an approach would be the lack of €2 -symmetry in the deoxy derivative 32 which would probably reduce any activity by at least a factor of two, as was observed in a different context with glucosidase inhibitor 2,5-dideoxy-2,5-imino-Dmannitol (10) and derivatives lacking the symmetry of the parent compound [37].

8.3.6

Inhibition of α-L-fucosidases by D-mannosidase inhibitors and related compounds

α-L-Fucosidases are among the most susceptible enzymes for inhibitors of the class of compounds under consideration. The L-/wco-configured analog of 1-deoxynojirimycin, l,5-dideoxy-l,5-imino-L-fucitol (18) is one of the most potent inhibitors of these enzymes reported, exhibiting Ki values of 3-5 nM. Such outstanding inhibition constants reflect some of the strongest interactions of carbohydrate analogs with proteins known to date. 1-Deoxy-L-fuconojirimycin (18) was first synthesized by Fleet and co-workers in 1985 [38]. Any of the modifications of the methyl side chain at C-5 (extension or removal [39]) or epimerization at one of the chiral centers reduces the inhibitory power by at least three orders of magnitude. Nevertheless, as exemplified with 1-deoxymannojirimycin (17), many D-mannosidase inhibitors also inhibit L-fucosidases with ^-values well in the micromolar range [7].

168

8 Some Reflections on Structure-Activity Relationships . . .

17

18

Figure 8-10.

Generally, pyrrolidine derivatives are less potent inhibitors than six-membered ring systems [4O]. For inhibition in the nanomolar range, nonpolar interactions of the methyl side chain in compound 18 with the enzyme's active site have been reported to be essential [18]. In addition, the presence and correct positioning of all three secondary hydroxyl groups is required. This can be seen from the comparably less efficient inhibitory power of deoxy derivatives, diastereomers as well as ring-contracted analogs and others [4O]. Interestingly, some tetrahydroxylated azepane derivatives, particularly with the Ό-gluco as well as Ό-manno configuration, have recently been reported to exhibit good α-L-fucosidase inhibitory powers with J^i values the μΜ range [15, 26]. As with l,5-dideoxy-l,5-imino-D-arabinitol 22 [18, 21], these activities can be attributed to the correct orientations of three out of four functional groups each mimicking the required Ό-arabino configuration of hydroxyl groups around a fucopyranose ring.

8.4 D-Glucosidase Inhibitors 8.4.1

Piperidine derivatives

8.4.1.1

Nojirimycin and 1-deoxynojirimycin

5-Amino-5-deoxy-D-glucose (Nojirimycin, 33) was first found as a natural product in the fermentation broths of a Streptomyces species and reported to be an antibiotic substance [2, 41]. It was realised that soon that compound 33 is a potent reversible inhibitor of /?-glucosidases and also strongly inhibits a-glucosidases [42, 43]. By systematic screening for microbial inhibitors of mammalian a-glucosidases, it was shown that this compound was also produced by a variety of Bacillus strains [44]. Despite the fact that nojirimycin is a potent glucosidase inhibitor, its applications are limited and hampered by an inherent lack of chemical and biochemical stability. The 1-deoxy derivative, 1-deoxynojirimycin (l,5-dideoxy-l,5imino-D-glucitol, 1), is the paradigmatic representative of the class of D-glucosi-

R = OH 33

R =H

ι

R = (CHz)2OH

R = H-BU

^

35

Figure 8-11.

8.4 D-Glucosidase Inhibitors

169

dase inhibitors under consideration. Before it was discovered as a natural product, it had already been synthesized by Paulsen and co-workers [45] in an entirely different context from a suitable L-sorbose derivative. Later, a Japanese group in the course of the structure elucidation of nojirimycin also prepared this compound [2, 41]. Soon thereafter, it was also found to be a constituent of the root bark of the mulberry tree (Morus bombycis) [46] and consequently named moranoline by Japanese researchers [47]. Compound 1 has been found to be a highly efficient reversible inhibitor of a wide range of a- and /J-glucosidases as well as trehalases [48]. According to general belief [30], 1-deoxynojirimycin is a transition-state analog of the putative cyclic oxocarbonium ion generated in the active site by protonation of the glycosidic bond of the substrate. On the basis of kinetic measurements, the opinion that it is a product analog, at least for /?-glucosidase from almonds, has also been put forward [49]. First introduced as a promising lead compound for novel approaches in the treatment of certain forms of diabetes and related metabolic disorders, it was also discovered to be active against retro viruses [5O]. This activity was rationalized by its interference with glucosidases I and II, two enzymes located in the endoplasmatic reticulum, which catalyze the first two steps of glycoprotein trimming of the highly glycosylated viral envelope protein gp 120, thus inhibiting viral attachment to the CO4 receptors of T-lymphocytes. Several syntheses of 1-deoxy nojirimycin have been reported over the 30 years since its discovery. Of these, one of the most efficient and elegant remains the four-step chemoenzymatic approach devised by the Bayer group in 1981, who employed a Gluconobacter suboxidans-catalyzed oxidation of N-benzyloxycarbonyl-1-amino-l-deoxy-D-glucitol at C-5 as the key step [51]. Numerous N-alkylated derivatives of 1-deoxynojirimycin have been synthesized in the search for improved activities and/or selectivities. Indeed, two of these compounds have gained considerable exposure. N-hydroxyethyl-1-deoxynojirimycin (Miglitol 34) is now a commercial product with anti-diabetic properties while N-butyl-1-deoxy nojirimycin (35) could be shown to have superior anti-retroviral activity [52, 53] when compared with the unsubstituted parent compound. Furthermore, this compound exhibited inhibitory power against glucosyl-transferases involved in glycosphingolipid biosynthesis [54]. Recently, the extremely high activity of N-modified derivatives with strongly lipophilic silicon-containing substituents has been reported [55]. Some other N-alkylated analogs were employed in affinity chromatographic separation and purification of glucosidases. 8.4.1.2

Glycosylated derivatives

A range of 0-glycosylated 1-deoxynojirimycins has been found as natural products or were synthesized by chemical or enzymatic methods, including the 2-O-a-, 2-Ο-β-, and 3-O-a-o-glucopyranosyl [56], the 3-Ο-β- [57] and 4-O-a[56] and/?-o-glucopyranosyl [56, 58], as well as the 3-O- and the 4-O-/?-D-galactopyranosyl [57, 59, 60] derivatives and others [60, 61]. Many of these have been found to inhibit selected D-glucosidases, but are generally not very active against standard microbial, and plant-derived glucosidases, employed in inhibition assays.

170

S Some Reflections on Structure—Activity Relationships...

8.4.1.3

Structural alterations of ring substituents

Modifications at C-I A structurally close relative of 1-deoxynojirimycin, 2,6-dideoxy-2,6-imino-D-g/};cero-L-gwfo-heptitol (a-homonojirimycin, 36) was synthesized by Liu [62] and soon thereafter identified as a constituent of the neotropical liana, Omphalea diandra L. [63]. this compound was shown as a powerful inhibitor of mouse gut α-glucosidases and human intestinal a- glucosidases. Synthetic bicyclic analogs [64] were found to be potent inhibitors of α-glucosidase I from pig kidneys and comparable with castanospermine in their potency. The corresponding epimer (37), coined '/M-homonojirimycin', is a synthetic product with low activity against yeast α-glucosidase as well as /?-glucosidase from almonds [65]. OH

OH OH ΗΟ^ί^

ΗΟ

36

37

OH

R =H

R =F

38

40

39

R =H 41 R =F 42 R = OMe 43

-τ.·

ο ·· *

Figure 8-12.

44

OH

HC

45

46

Figure 8-13.

Modifications at C-2 The epimer of compound 1, 1-deoxymannojirimycin (17), is a good inhibitor of Dmannosidases. The 2-deoxy derivative, fagomine (38), has been discovered as a natural product [66]. The 2-deoxyfluoro analog of 1-deoxynojirimycin (39) was prepared in the hope that this 'polar' deoxygenation might lead to a more selective

171

8.4 D-Glucosidase Inhibitors

inhibitor [37, 67]. The fluorine atom is assumed to be a hydrogen bond acceptor [68] in the active site and the inductive effect of the strongly electronegative element was expected to influence the protonation of the ring nitrogen with a view to activity at lower pH values. None of these and other modifications at C-2 such as the C-2-methyl derivative 40 [69] has been reported to exhibit any appreciable activity against the standard D-glucosidases usually probed. Modifications at C-3, C-4, and C-S Modifications at C-3 have been rarely reported. The 3-deoxy derivatives 41 was synthesized but no activity was reported for this compound. 3-Deoxyfluoro (42) [25, 67, 70] as well as the 3-O-methyl (43) analogs [25, 71] have been found to be devoid of glucosidase inhibitory activity. This is also found to be the case for the 3-epi-compound-(4). The D-ga/acto-configured epimer (3) as well as 1,4dideoxy-4-fluoronojirimycin (44) [72] were also found to be poor glucosidase inhibitors. l,5-Dideoxy-l,5-imino-L-iditol (45), the epimer at C-5, was found to be a weak, noncompetitive inhibitor of α-glucosidase from yeast [73]. 1,5Dideoxy-l,5-imino-D-xylitol (46) [21], lacking the hydroxymethyl side chain, did not show any appreciable inhibitory activity. Modifications at C-6 1,6-Dideoxynojirimycin (47) was prepared by a chemoenzymatic approach and reported to be a poor inhibitor of yeast α-glucosidase and almond ^-glucosidase [19]. In contrast to this finding, the 6-deoxyfluoro derivative 48 as well as (6S)6-C-ethyl-deoxynojirimycin (49) have been found to exhibit appreciable activities against α-glucosidase from yeast (Ki 29 μΜ and 60 μΜ, respectively, at pH 6) [74]. Furthermore, these compounds were about half as active as the parent substance with almond ^-glucosidase (Ki 600 μΜ each at pH 5). Neither the N-methyl derivative of compound 49 nor the (6/?)-epimer exhibited biological activity [2O]. Some other 6-substituted and chain-extended derivatives have been synthesized either from simple sugars [75] or by degradation of castanospermine [76]; for example, the 6-azidodeoxy (50) and aminodeoxy analogs [77], but no data on inhibitory activities have been reported. OH

Figure 8-14.

R =H R =F

47 48

R = N3 so

49

51

172

S Some Reflections on Structure-Activity Relationships...

Open-chain analogs A range of 'seco derivatives' of 1 -deoxynojirimycin, such as compound 51, have been prepared and tested. Surprisingly, selected examples exhibited noteable glucosidase inhibitory activities [78]. Summary It appears that alterations of functional groups along the carbon backbone, other than modifications at C-6 are deleterious to the glucosidase inhibitory activity of 1-deoxynojirimycin, and modifications at C-6 usually lead to a general reduction in inhibitory power.

8.4.2

Castanospermine

Castanospermine (52), a natural product of Castanospermum australe [79], the Australian or Morreton Bay chestnut tree, is structurally the most closely related compound to 1-deoxynojirimycin (1). The most significant difference is the locked dihedral angle of the hydroxyl group in the five-membered ring of the former as compared with the freely rotating hydroxymethyl moiety at C-5 of the latter. The exciting chemistry and biochemistry of Castanospermine is treated in other chapters in this book. In contrast to expectations and taking into account the similar structural features of 1-deoxymannojirimycin and 6-cpi-castanospermine (53) [80], the latter is not an α-D-mannosidase inhibitor but is active against the £jto-a-D-glucosidase amyloglucosidase. Other Castanospermine derivatives have either been found in nature or were synthesized chemically. Structure-activity relationships have been discussed by Winchester, Fleet and their colleagues [81] as well as by other researchers. Tyler and co-workers have synthesized a wide range of Castanospermine derivatives, modified at positions 1, 6, 7, and 8, which correspond to C-6, C-2, C-3, and C-4, respectively, in 1-deoxynojirimycin [82].

52

54

53

OH

57

Figure 8-15.

8.4 D-Glucosidase Inhibitors

173

Most derivatives modified at positions 1, 7, or 8 turned out to be comparably weak inhibitors, or were almost devoid of any activity. Some modifications at C6 - corresponding to C-2 in 1-deoxynojirimycin - such as the 6-deoxyfluoro derivative 54 (A^ 50 μΜ with human liver enzymes [83]) exhibited interesting activity. Rearranged compound (55) nicely superimposes onto 7-deoxycastanospermine, the hydroxyl group of the six-membered ring taking the position of 6-OH of the parent molecule and the two secondary alcohol groups of the five-membered ring subunit mimicking 8-OH and 1-OH. A related bicyclic compound, coined ,glc-swainsonine' (56), also bearing only 3 hydroxyl groups but featuring an alltrans relationship of these substituents around the ring system has been reported by Elbein and co-workers to be a good amyloglucosidase inhibitor (Ki 50 μΜ) [84]. A compound analogous to castanospermine, quinolizidine 30, was synthesized and shown to be a good inhibitor of a- and /?-glucosidases, albeit not as potent as the parent compound [85]. The epimer 29 [86], corresponding to l-e/?i-castanospermine (57), was established to be devoid of any glucosidase inhibitory power but was found to be an inhibitor of jackbean α-mannosidase (Ki 360 μΜ at pH 5.5). 8.4.2.1

Summary

Interestingly, any modification at C-I including deoxygenation, epimerization deoxyfluorination has been shown to be deleterious to the inhibitory power of the molecule. Alterations at C-7 and C-8, these positions being comparable with C-3 and C-4 in 1-deoxynojirimycin, strongly reduce or destroy the biological activity of the castanospermine derivative. Variations of the substituent as well as the configuration at C-6, which is equivalent to C-2 in 1-deoxynojirimycin, seem to be permissible, though a distinct loss of inhibitory properties is observed. From biological activities reported for selected C-I-, C-7-, and C-8-modified analogs as well as rearranged products, it may be concluded that human lysosomal aglucosidase does not require all four hydroxyl groups of the parent compound for recognition/binding of the inhibitor molecule. Furthermore, results suggest that the hydroxyl group at C-7 is not essential for the inhibitory activity of the active compounds with this particular enzyme. This is also supported by the high activity of 7-£/?i-castanospermine against the same enzyme.

8.4.3

Important pyrrolidine and pyrrolizidine derivatives

8.4.3.1

2,5-Dideoxy-2,5-imino-D-mannitol

2,5-Dideoxy-2,5-imino-D-mannitol (10) was discovered as a natural product in 1976 [87] and shown to be a powerful inhibitor of a large range of a- as well as /?-glucosidases, even surpassing the activity of 1-deoxynojirimycin which is frequently used as a standard [88]. At first sight, this molecule does not closely resemble o-glucopyranosides or the structural types of inhibitors discussed so far. Inspection of Dreiding models and computer-assisted molecular modeling show that compound 10 and 1-deoxynojirimycin are practically isosteric, the hydroxyl

175

8.4 D-Glucosidase Inhibitors

group at C-I of the former taking the place of 2-OH in the latter [37]. Several diastereomers have recently been synthesized such as the epimer at C-2, 2,5-dideoxy2,5-imino-D-glucitol (58) [89, 9O]. However, this compound did not exhibit the pronounced activity of compound 10. Other chemical alterations include, for example, the 1-0-methyl- (59), [37], the 1-deoxy- (60) [91], the 1-aminodeoxy(61) [92] as well as the 1-deoxyfluoro (62) [37] derivatives. None of these derivatives exhibited the high activity of the parent compound. Interestingly, against some enzymes the derivatives at C-I showed about onethird to on half of the activity of compound 10 while other o-glucosidases were found to be inhibited to a much lesser extent. This fact was attributed to the two possible binding modes of the !-modified derivatives [37]. In the case where the non-natural substituent takes the position of 6-OH, considerable inhibitory power should remain, as can be expected by comparison with 6-modified deoxynojirimycin analogs [74]. Conversely, when the molecule is bound to the enzyme with the deoxygenated carbon in the position of 2-OH, inhibitory activity might be abolished to the same extent as in 2-modified 1-deoxynojirimycins, as mentioned previously. The inhibitory activity measured would therefore correspond to an equilibrium mixture of these orientations in the active site. Interestingly, l,4-dideoxy-l,4-imino-D-arabinitol (26) exhibits a K1 value of 0.18 μΜ with aglucosidase from yeast and was also found to be a powerful inhibitor of mammalian a-glucosidases [7]. 8.4.3.2

Australine

Australine (63), like castanospermine a product of the Australian tree Castanospermum australe, is related to 2,5-dideoxy-2,5-imino-D-mannitol (10) the same way as is castanospermine (52) to 1-deoxynojirimycin. In this pyrrolizidine alkaloid, one of the hydroxymethyl groups is locked, which abolishes the C2-axis of symmetry and probably reduces the ability of this compound to adapt to the steric requirements of certain enzymes. Consequently, australine is a less powerful and universal inhibitor of glucosidases than the pyrrolidine analog 10. As with C-Imodified derivatives of 10, two modes of binding to a given active site of a glucosidase are possible (Picture 5).

„

HO-I

r

OH

Figure 8-17.

^π

-

1

Γ\/

ViH ^)-OH

176

8 Some Reflections on Structure-Activity Relationships...

8.4.4

Tetrahydroxyazepanes

Recently, much to the surprise of many workers in the field, it was demonstrated that the symmetric seven-membered ring l,6-dideoxy-l,6-imino-L-iditol (15) is a powerful inhibitor of several glycosidases [15, 26]. For a- and /?-glucosidases, Ki values in the lower micromolar range were observed. Computer-aided visualization [15] revealed that the flexible seven-membered ring system can be superimposed onto several other potent glycosidase inhibitors such as 1-deoxynojirimycin (1), 2,5-dideoxy-2,5-imino-D-mannitol (10), or isofagomine (64), the steric requirements of the backbone as well as the positions of functional groups nicely matching these 'templates'. It should be noted that, in symmetric 1,6-dideoxy1,6-imino-L-iditol (15), the hydroxyl group at C-5 is likely to be axially oriented and parallel to the lone electron pair on the ring nitrogen and consequently comparable with 1-OH in castanospermine (52). In this context it should also be noted that the Ό-gluco epimer (27) is almost devoid of glucosidase inhibitory power and rather a moderate inhibitor of D-mannosidases. Similar to 2,5-dideoxy-2,5-imino-D-mannitol (10), the symmetric nature of compound 15 appears to allow two (albeit identical) binding modes. Furthermore, binding properties as well as affinity to more than one family of glycosidases might be due to the conformational flexibility of the molecule which would allow it to adapt to the electronic and steric requirements of different enzyme active sites.

8.4.5

Isofagomine

Isofagomine (64) was first synthesized by BoIs and co-workers [1O]. This compound was found to be a potent inhibitor of a- (from yeast, K1 around 80 μΜ at pH 6.8) and, in particular, of /?-glucosidases (from almonds, K1 0.1 μΜ, ρΗ 6.8) [93] with Ki values in the range of those reported for 1-deoxynojirimycin (K{ 25 and 38 μΜ, respectively, at pH 6) [94]. This was attributed to the basic nitrogen replacing the anomeric carbon in this molecule which was believed to be located more closely to one or both of the carboxylates/carboxylic acids in the active site of the /^-glucosidase. Isofagomine closely resembles the siastatins [95]. Structurally closely related compounds with pronounced glucosidase inhibitory activity such as compound 65 have been synthesized by Ichikawa and co-workers [96].

R =H

64

R = OH 65

Figure 8-18.

8.4 D-Glucosidase Inhibitors

111

8.4.6 Calystegines Recently, in the course of a program to investigate the interactions of plants with soil bacteria, three new alkaloids were isolated from the roots of Calystegia sepium. The structures of these compounds, calystegines BI (66), 62 (67), as well as AS (68) were determined and found to be members of the nortropane family [97]. Their common structural feature is a hemiaminal moiety formed by intramolecular amination of a highly functionalized /-amino cycloheptanone system. Synthetic studies were conducted based on various approaches [98, 99], for example with the aid of the Ferrier reaction [100], by a chemoenzymatic approach utilizing cycloheptatriene as well as via a hetero Diels-Alder cycloaddition [101]. Further members of the calystegine family were discovered in the leaves and roots of various Morus species together with a range of iminosugars [102] as well as in several other plant families such as Solanaceae (Atropa, Datura, Duboisia, Hyoscyamus, Scopolia). Surprisingly, calystegines are also found in food such as potatoes and egg-plant [103]. Of the currently known calystegines, the most powerful in terms of glucosidase inhibitory activity are calystegines BI and 62 as well as Ci (69). Whereas calystegines 62 and Ci are closely related featuring an all-trans relationship of hydroxyl groups around the six-membered ring, thus resembling-deoxynojirimycin or l,6-dideoxy-l,6-imino-L- iditol (15), calystegine BI contains structural elements of the powerful /^-glucosidase inhibitor isofagomine (64). HQ HO;

66

67

68

69

HQ

Figure 8-19.

8.4.7

Common features and structure-activity relationships

Due to the structural diversity of the glucosidase inhibitors under consideration, common features with respect to the three-dimensional arrangement of functional groups around the ring systems appear difficult to identify. However, relationships of 1-deoxynojirimycin, castanospermine, as well as homonojirimycins with o-glucose, glucosides or the D-glucose derived putative cyclic oxocarbonium ion can be easily deduced. Less apparent are the structural similarities of 2,5-dideoxy-2,5-

178

8 Some Reflections on Structure-Activity Relationships...

imino-D-mannitol (10) and its bicyclic analog, australine, with the substrates or intermediates/transition states. As can be visualized with the aid of molecular modeling, these five-membered ring inhibitors superimpose nicely with 1- deoxynojirimycin or castanospermine (Picture 8-3). Consequently, the inhibitory activities observed for the pyrrolidine and, albeit to a lesser extent, for the pyrrolizidine are not too remote from the formerly addressed piperidine and indolizidine relatives. The striking activities of the L-zW0-configured azepane as well as calystegines 62, BS and Ci against the enzymes under consideration seem more difficult to rationalize in the same type of model. Selected conformations of 1,6-dideoxy1,6-imino-L-iditol exhibit similarity with the spatial distribution of functional groups in 1-deoxynojirimycin and castanospermine. (Picture 8-4, s. page 182). Superposition of an energetically preferred [15] conformer revealed that the alltrans arrangement of the four hydroxyl groups nicely mimicks 2-OH to 4-OH of 1-deoxynojirimycin or a related D-glucopyranosyl cation, the axially oriented hydroxyl group taking the position of 1-OH in castanospermine. This appears to be an important structural feature for good activity as neither 1-deoxy- (70) nor 1epi-castanospermines (57) as well as l,6-dideoxy-l,6-imino-D-glucitol 27 [15, 25, 26], the epimer of 15 at C-5, exhibit any noteworthy D-glucosidase inhibitory power. Not surprisingly, by structural comparison, the latter was found to inhibit jack bean α-mannosidase (Ki 310 μΜ at pH 5.5). In the case of 1-deoxynojirimycin, some enzymes could be expected to prefer a similar dihedral angle between the freely rotating hydroxymethyl group and the C-5-N-bond as in castanospermine. Disregarding the conformational mobility of the monocyclic molecules under consideration, a general feature of most good inhibitors appears to be the alltrans relationship of four hydroxyl groups with respect to the ring system which seems to be essential for efficient inhibitory power. One of the few exceptions in this context would be isofagomine (64). Another crucial structural characteristic of all inhibitors seems to be the axial orientation of the lone electron pair of the ring nitrogen. In addition, exceptional inhibitors feature a C2-axis of symmetry as seen in l,6-dideoxy-l,6-imino-L-iditol (15) and 2,5-dideoxy-2,5-imino-D-mannitol (10). In all cases, derivatives with disturbed symmetry exhibit less inhibitory power than the symmetric parent compounds (Figure 8-16 compare Picture 8-3). This could suggest that the !-modified derivatives are either bound with the non-natural substituent at the O-6 position giving inhibition or turned through 180° with the substituent in the O-6 position thus lacking inhibitory activity, as is known from the 2-deoxy derivative of 1-deoxynojirimycin. Such an element of symmetry is, at first sight, lacking in the structures of 1-deoxynojirimycin as well as castanospermine and the calystegines. However, after carrying out a set of simple operations, a type of 'hidden' symmetry seems evident in the latter types of inhibitors, which probably allows two different binding orientations to the active site of a susceptible enzyme. In the most simple case, a 180° rotation of 1-deoxynojirimycin around an axis through the C-2-C-3- as well as the C-5-N-bonds allows superposition of hydroxyl groups in these particular orientations of the molecule, OH-2 taking the place of OH-6 and OH-3 interchanging its place with OH-4 (Picture 8-5).

8.4 D-Glucosidase Inhibitors

179

Figure 8-20.

The same is true for castanospermine, provided that a pyramidal inversion at the ring nitrogen can take place. After lateral inversion, OH-I occupies the position of OH-6, and OH-7 the position of OH-8, and vice versa (Figure 8-20, Picture 8-6). Several of the expected consequences of such a hypothetical change in orientation at an active site of a glucosidase are supported by experimental data. For example, all !-modified derivatives of castanospermine known to date are almost devoid of inhibitory power against D-glucosidases, and so are the known 2-modified 1deoxynojirimycin derivatives. In contrast, 6-modified castanospermines still bear interesting D-glucosidase inhibitory activities, albeit less pronounced than the parent compound. This is also generally true for 1-deoxynojirimycin analogs with modifications at position C-6. Could this mean that castanospermine derivatives, at least with some of the enzymes under consideration, might be bound in the alternative conformation/orientation mentioned above and depicted in Picture 8-6? (s. page 183). Here it can be seen that OH-I is in the place of OH-2 of 1-deoxynojirimycin and OH-6 is in the position of the primary OH-6 of 1-DNM. This hypothesis seems to be further supported by the following examples: 6-£/?i-castanospermine (53) was reported not to be a D-mannosidase inhibitor, as could be expected from the structural analogy with o-mannose or 1-deoxymannojirimycin, but was found to inhibit amyloglucosidase, an exo-a-glucosidase, as mentioned earlier. On the other hand, a /i0mo-castanospermine (l/?,2/?,3S,9S,9a/Metrahydroxyquinolizidine, (30) is, like castanospermine, an active ^-glucosidase inhibitor. However, the corresponding 9-e/?i-derivative 29, related to l-^pi-castanospermine with cis-oriented hydroxyl groups at C-I and C9, is almost inactive against glucosidases but was found to be an inhibitor of jack bean α-mannosidase (Ki 360 μΜ). Furthermore, the 2-deoxyfluoro derivative of 1-deoxynojirimycin is almost inactive whereas the corresponding castanospermine analog, 6-deoxyfluorocastanospermine, exhibited considerable activity against human lysosomal a-glucosidase. Conversely, the 6-deoxyfluoro derivative (48) of 1-deoxynojirimycin is a good inhibitor of α-glucosidase from yeast (K1 19 μΜ at pH 6) and an inhibitor of /?-glucosidases from A. wentii (380 μΜ at pH 5) as well as almonds (Ki 600 μΜ). In contrast, its bicyclic analog, 1-deoxyfluorocastanospermine (71), was reported to be virtually inactive against glucosidases.

Figure 8-21.

180

8 Some Reflections on Structure-Activity Relationships. OH

64

OH

26

Figure 8-22.

One inhibitor that does not immediately fit this albeit simplified picture is isofagomine, which has the ring nitrogen in the anomeric position, lacks OH-2 and therefore bears only three hydroxyl groups available for interactions in the active sites of /?-glucosidases. Structurally related is l,4-dideoxy-l,4-imino-D-arabinitol (26), a strong inhibitor of selected α-glucosidases with respect to the number and relative orientations of hydroxyl groups.

8.4.8

Conclusion and outlook

Reviewed data distinctly support the view that inhibitors possessing suitable threedimensional fit and geometry can be bound to the active sites of different glycosidases in possibly more than one orientation, and thereby are able to adjust to the stereochemical as well as electronic requirements of the respective glycosidase. In the case of molecules exhibiting a C2-axis of symmetry, such as compounds 10 and 15, this enables the inhibitors to have enhanced activity as either of two (identical) orientations towards the active site leads to productive interaction and subsequent inhibition. Accordingly, disturbance of such symmetry, seen with glucosidase inhibitors australine (s. picture 8-7, page 183) as compared with 2,5dideoxy-2,5-imino-D-mannitol (10) as well as with C-1-modified derivatives of the latter, results in inhibitory activities that are reduced by at least a factor of approximately two in cases of enzymes that do not require 6-OH for efficient inhibition. In cases of enzymes showing strong interactions with the 6-OH of glucosides, such derivatives lose most of their inhibitory power. Moreover, castanospermine and analogs thereof exhibit inhibitory activities which are difficult to rationalize on the basis of comparison with monocyclic, closely related derivatives such as 1-deoxynojirimycin. An examination of derivatives at various positions, most tellingly at carbons C-I and C-6, appears to suggest that by pyramidal inversion at the ring nitrogen, the resulting diastereomer might also fit into the active site, albeit with hydroxyl groups interchanged in their positions. In the example under consideration, as well as with other related inhibitors, a synergism of factors could trigger preference for one or the other binding mode such as steric requirements of the active site, basicity of the ring nitrogen, the sum or selected polar and non-polar interactions within the active site and, most importantly, the availability of intermolecular hydrogen bonding. Unambiguous experimental evidence for such alternative binding has not been put forward, to date, but the increasing availability of crystallographic data of glucosidase/inhibitor complexes such as those published by Honzatko and co-workers [104] will be a means of obtaining such crucial structural data. This may become significant in the understanding of how to exploit enzymatic glycoside hydrolysis to the full, as well as in the design

Acknowledgments

181

and syntheses of novel inhibitors with superior characteristics for medicinal and other specific purposes. A phenomenon worth mentioning in this context is the slow onset of inhibition with many of the inhibitors under consideration (see Chapter 3). Equilibria are frequently reached on a time scale of several minutes. This observation was rationalized by a necessary conformational change of the active site from its ground state into a less-favored conformation, and may also result from a stepwise mutual adjustment of conformations of both the inhibitor and the enzyme in order for optimal interaction, to occur. Taking into account problems in comparing important inhibitors with the same enzyme, and the evolutionary adjustment of enzymes from various sources to the specific needs of their respective tasks and environments, an attempt to conceive a valid general picture of structure-activity relationships would be met by severe difficulties. Nevertheless, in the search for common features and systematic approaches, considerations such as those outlined above, despite being oversimplified and, thus, biased, may be helpful contributions to be either proved or disproved with the aid of in-depth investigations. Basic and applied research on iminoalditols and related compounds with respect to glycosidase inhibition has emerged over the past decades as a result of synergistic efforts of scientists over a large range of expertise. Moreover, following the pioneering BAYER compounds mentioned in the introduction, this will be the basis of future developments and hopefully lead to carbohydrate-derived Pharmaceuticals. In the absence of structural X-ray crystallographic data on many important glycosidases, investigators must continue to rely on conclusions drawn from comparative inhibitory data of suitable inhibitors, as well as their inspiration and creativity in establishing and validating structure-activity relationships.

Acknowledgments Professors Giinter Legler and Steve Withers are thanked for their vital input into long term collaborations with the author. Essential financial support by the Austrian Fonds zur Forderung der Wissenschaftlichen Forschung as well as the Jubilaumsfonds der Osterreichischen Nationalbank is greatly appreciated.

182

8 Some Reflections on Structure—Activity Relationships

E E

Acknowledgments

183

184

8 Some Reflections on Structure-Activity Relationships...

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

[16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34]

H. Paulsen, K. Todt, Adv. Carbohydr. Chem. Biochem., 1968, 23, 116-232. S. Inouye, T. Tsuruoka, T. Ito, T. Niida, Tetrahedron, 1968, 24, 2125-2144. H. Paulsen, Y. Hayauchi, V. Sinnwell, Chem. Ber., 1980, 113, 2601-2608. G. Legler, S. Pohl, Carbohydr. Res., 1986, 755, 119-129. R. C. Bernotas, M. A. Pezzone, B. Ganem, Carbohydr. Res., 1987, 167, 305-311. Y. Miyake, M. Ebata, J. Antibiot., 1987, 40, 122-123; Y. Miyake, M. Ebata, Agric. Biol. Chem., 1988, 52, 153-158; Y. Miyake, M. Ebata, Agric. BM. Chem., 1988, 52, 1649-1654 N. Asano, K. Oseki, H. Kizu, K. Matsui, J. Med. Chem., 1994, 37, 3701-3706. P. Zhou, H. M. Salleh, P. C. M. Chan, G. Lajoie, J. F. Honek, P. T. C. Nambiar, O. P. Ward, Carbohydr. Res., 1993, 239, 155-166. Y. Ichikawa, Y. Igarashi, Tetrahedron Lett., 1995, 36, 4585-4586. T. M. Jespersen, W. Dong, M. R. Sierks, T. Skrydstrup, I. Lundt, M. BoIs, Angew. Chem., 1994, 706, 1858-1860. M. Ichikawa, Y. Ichikawa, Biorg. Med. Chem., 1995, 3, 161-165. Y-F. Wang, Y Takaoka, C.-H. Wong, Angew. Chem., 1994, 106, 1343-1345. G. W J. Fleet, S. J. Nicholas, P. W. Smith, S. V. Evans, L. E. Fellows, R. J. Nash, Tetrahedron Lett., 1985, 26, 3127-3130. I. Lundt, R. Madsen, S. Al Daher, B. Wichester, Tetrahedron, 1994, 50, 7513-7520. F. Moris-Varas, X.-H. Qian, C.-H. Wong, J. Am. Chem. Soc., 1996, 118, 7647-7652; X. Qian, F. Moris-Varas, M. C. Fitzgerald, C.-H. Wong, Bioorg. Med. Chem., 1996, 4, 2055-2069; X. Qian, F. Moris-Varas, C.-H. Wong, Bioorg. Med. Chem. Lett., 1996, 6, 1117-1122. L. E. Fellows, E. A. Bell, D. G. Lynn, F. Pilkiewicz, I. Miura, K. Nakanishi, J. Chem. Soc., Chem. Commun., 1979, 977-978. A. de Raadt, A. E. Stiitz, Tetrahedron Lett., 1992, 33, 189-192 and ref. cited therein. G. Legler, A. E. Stiitz, H. Immich, Carbohydr. Res., 1995, 272, 17-30. T. Kajimoto, K. K.-C. Liu, R. L. Pederson, Z. Zhong, Y. Ichikawa, J. A. Porco, Jr., C.-H. Wong, /. Am. Chem. Soc., 1991, 113, 6187-6196. G. Legler, unpublished results. Private communication to the author. R. C. Bernotas, G. Papandreou, J. Urbach, B. Ganem, Tetrahedron Lett., 1990, 31, 3393-3396. B. P. Bashyal, G. W. J. Fleet, M. J. Gough, P. W. Smith, Tetrahedron, 1987, 43, 30833093 and ref. cited therein. M. J. Eis, C. J. Rule, B. A. Wurzburg, B. Ganem, Tetrahedron Lett., 1985,26,5397-5398. B. Winchester, S. Al Daher, N. C. Carpenter, I. Cenci di Bello, S. S. Choi, A. J. Fairbanks, G. W. J. Fleet, Biochem. J., 1993, 290, 743-749. G. Legler, A. E. Stutz, unpublished data. Y. Le Merrer, L. Poitout, J.-C. Depezay, I. Dosbaa, S. Geoffroy, M.-J. Foglietti, Bioorg. Med. Chem., 1997, 5, 519-533. S. M. Colgate, P. R. Dorling, C. R. Huxtable, Aust. J. Chem. 1979, 32, 2257-2264. R. J. Molyneux, F. James, Science, 1982, 276, 190-191. R R. Dorling, C. R. Huxtable, S. M. Colgate, Biochem. J., 1980, 797, 649-651 G. Legler, Adv. Carbohydr. Chem., 1990, 48, 319-384 and ref. cited therein. a) D. A. Winkler, G. Holan, J. Med. Chem., 1989, 32, 2084-2089; b) R. A. Fair, A. K. Holland, E. W. Huber, N. P. Peet, R M. Weintraub, Tetrahedron, 1994, 50, 1033-1044. K. Tadano, Y. Hotta, M. Morita, T. Suami, B. Winchester, Bull. Chem. Soc. Jpn., 1987, 60,3667-3671. K. Tadano, Y. Iimura, Y. Hotta, C. Fukabora, T. Suami, Bull. Chem. Soc. Jpn., 1986, 59, 3885-3892. K. Tadano, M. Morita, Y. Hotta, S. Ogawa, B. Winchester, I. Cenci di Bello, /. Org. Chem., 1988, 53, 5209-5215.

References

185

[35] D. A. Winkler, J. Med. Chem., 1996, 39, 4332-4334. [36] L. A. G. M. van den Broek, D. J. Vermaas, B. M. Heskamp, C. A. A. van Boeckel, M. C. A. A. Tan, J. G. M. Bolscher, H. L. Ploegh, F. J. van Kemenade, R. E. Y. de Goede, F. Miedema, Red. Trav. Chim. Pays-Bas, 1993, 772, 82-94. [37] S. M. Andersen, M. Ebner, C., W. Ekhart, G. Gradnig, G. Legler, I. Lundt, A. E. Stiitz, S. G. Withers, T. Wrodnigg, Carbohydr. Res, 1997, 301, 155-166 and ref. cited therein. [38] G. W. J. Fleet, A. N. Shaw, S. V. Evans, L. E. Fellows, J. Chem. Soc., Chem. Commun., 1985, 841-842. [39] H. Paulsen, M. Matzke, B. Orthen, R. Nuck, W. Reutter, Liebigs Ann. Chem., 1990, 953-963. [40] A. de Raadt, C. W. Ekhart, M. Ebner, A. E. Stiitz, Topics Curr. Chem, 1997, 787, 157186. [41] S. Inouye, T. Tsuruoka, T. Niida, /. Antibiot., 1966, 79, 288-291. [42] T. Niwa, S. Inouye, T. Tsuruoka, Y. Koaze, T. Niida, AgHc. Biol. Chem., 1970, 34, 966967. [43] E. T. Reese, F. W. Parrish, M. Ettlinger, Carbohydr. Res., 1971, 18, 381-388. [44] D. D. Schmidt, W. Frommer, L. Miiller, E. Truscheit, Naturwissenschaften, 1979, 66, 584-585. [45] H. Paulsen, I. Sangster, K. Heyns, Chem. Ber., 1967, 100, 802-811. [46] K. Daigo, Y. Inamori, T. Takemoto, Chem. Pharm. Bull., 1986, 34, 2243-2246. [47] Y. Yoshikuni, AgHc. Biol,. Chem., 1988, 52, 121-128 and ref. cited therein. [48] S. Murao, S. Miyata, AgHc. Biol. Chem., 1980, 44, 219-221. [49] M. P. Dale, H. E. Ensley, K. Kern, K. A. R. Sastry, L. D. Byers, Biochemistry, 1985, 24, 3530-3539. [50] P. S. Sunkara, T. L. Bowlin, P. S. Liu, A. Sjoerdsma, Biochem. Biophys. Res. Commun., 1987, 148(1), 206-210. [51] G. Kinast, M. Schedel, Angew. Chem., 1981, 93, 799-801. [52] A. Karpas, G. W. J. Fleet, R. A. Dwek, S. Petursson, S. K. Namgoong, N. G. Ramsden, G. S. Jacob, T. W. Rademacher, Proc. Natl. Acad. Sd. USA, 1988, 85, 9229-9233. [53] P. B. Fischer, G. B. Karlsson, R. A. Dwek, F. M. Platt, J. Virol., 1966, 70, 7153-7160. [54] F. M. Platt, G. R. Neises, R. A. Dwek, T. D. Butters, /. Biol. Chem., 1994, 269, 83628365. [55] B. Lesur, J.-B. Ducep, M.-N. Lalloz, A. Ehrhard, C. Danzin, Bioorg. Med. Chem. Lett., 1997, 7, 355-360. [56] N. Asano, K. Oseki, E. Kaneko, K. Matsui, Carbohydr. Res, 1993, 258, 255-266. [57] M. Kiso, H. Katagiri, H. Furui, A. Hasegawa, J. Carbohydr. Chem., 1992, 11, 627-644. [58] Y. Ezure, AgHc. Biol. Chem., 1985, 49, 2159-2165. [59] C.-H. Wong, T. Krach, C. Gautheron-Le Narvor, Y. Ichikawa, G. C. Look, F. Gaeta, D. Thompson, K. C. Nicolaou, Tetrahedron Lett., 1991, 32, 4867-4870; C. GautheronLe Narvor, C.-H. Wong, /. Chem. Soc., Chem. Commun., 1991, 1130-1131; C.-H. Wong, Y. Ichikawa, T, Krach, C. Gautheron-Le Narvor, D. P. Dumas, G. C. Look, /. Am. Chem. Soc., 1991, 113, 8137-8145. [60] M. Kojima, T. Seto, Y. Kyotani, H. Ogawa, S. Kitazawa, K. Mori, S. Maruo, T. Ohgi, Y. Ezure, Biosci. Biotech. Biochem., 1966, 60, 694-696. [61] M. Kiso, K. Ando, H. Furui, A. Hasegawa, J. Carbohydr. Chem., 1993, 12, 985-990; M. Kiso, H. Furui, K. Ando, H. Ishida, A. Hasegawa, /. Carbohydr. Chem., 1994, 13, 163-174; M. Kiso, K. Ando, H. Inagaki, H. Ishida, A.Hasegawa, Carbohydr. Res., 1995, 272, 159-178; T. Kawaguchi, K. Sugimoto, H. Hayashi, M. Arai, Biosci. Biotech. Biochem., 1996, 60, 344-346. [62] P. S. Liu, J. Org. Chem., 1987, 52, 4717-4721. [63] G. C. Kite, L. E. Fellows, G. W. J. Fleet, P. S. Liu, A. M. Scofield, N. G. Smith, Tetrahedron Lett., 1988, 29, 6483-6486. [64] P. S. Liu, R. S. Rogers, M. S. Kang, P. S. Sunkara, Tetrahedron Lett., 1991, 32, 58535856. [65] K. E. Holt, F. J. Leeper, S. Handa,. J. Chem. Soc. Perkin Trans 1, 1994, 231-234.

186

8 Some Reflections on Structure-Activity Relationships...

[66] S. V. Evans, A. R. Hayman, L. E. Fellows, T. K. M. Shing, A. E. Derome, G. W. J. Fleet, Tetrahedron Lett., 1985, 26, 1465-1468. [67] D. P. Getman, G. A. de Crescenco, Eur. Pat. Appl. EP 410,953 (1990). [68] T. Tsuchiya, Adv. Carbohydr. Chem. Biochem., 1990, 48, 91-277. [69] I. K. Khanna, R. M. Weier, J. Julien, R. A. Mueller, D. C. Lankin, L. Swenton, Tetrahedron Lett., 1996, 37, 1355-1358. [70] C.-K. Lee, H. Jiang, L. L. Koh, Y. Wu, Carbohydr. Res., 1993, 239, 309-315. [71] L. A. G. M. van den Broek, D. J. Vermaas, B. M. Heskamp, C. A. A. van Boeckel, M. C. A. A. Tan, J. G. M. Bolscher, H. L. Ploegh, F. J. van Kemenade, R. E. Y. de Goede, F. Miedema, Reel. Trav. Chim. Pays-Bas., 1993, 772, 82-94. [72] G. A. de Crescenco, D. R Getman, Eur. Pat. Appl. EP 481,950 (1992). [73] R A. Fowler, A. H. Haines, R. J. K. Taylor, E. J. T. Chrystal, M. B. Gravestock, Carbohydr. Res.,1993, 246, 377-381. [74] A. Berger, K. Dax, G. Gradnig, V. Grassberger, A. E. Stiitz, M. Ungerank, G. Legler, E. Bause, Bioorg. Med. Chem. Lett., 1992, 2, 27-32. [75] A. Kilonda, F. Compernolle, S. Toppet, G. J. Hoornaert, Tetrahedron Lett., 1994, 35, 9047-9050. [76] J. B. Bremner, B. W. Skelton, R. J. Smith, G. J. Tarrant, A. H. White, Tetrahedron Lett., 1996, 37, 8573-8576. [77] A. Kilonda, F. Compernolle, S. Toppet, G. J. Hoornaert, J. Chem. Soc., Chem. Commun., 1994, 2147-2148. [78] R A. Fowler, A. H. Haines, R. J. K. Taylor, E. J. T. Chrystal, M. B. Gravestock, /. Chem. Soc., Perkin Trans. 1, 1994, 2229-2235. [79] L. D. Hohenschutz, E. A. Bell, P. J. Jewiss, D. P. Leworthy, R. J. Pryce, E. Arnold, J. Clardy, Phy to chemistry, 1981, 20, 811-814. [80] R. J. Molyneux, J. N. Roitman, G. Dunnheim, T. Szumilo, A. D. Elbein, Arch. Biochem. Biophys., 1986, 257, 450-457 [81] B. G. Winchester, I. Cenci di Bello, A. C. Richardson, R. J. Nash, L. E. Fellows, N. G. Ramsden, G. Fleet, Biochem. J., 1990, 269, 227-231. [82] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, Tetrahedron, 1994, 50, 21312160; R. H. Furneaux, J. M. Mason, R C. Tyler, Tetrahedron Lett., 1994, 55, 31433146; R. H. Furneaux, G. J. Gainsford, J. F. Mason, R C. Tyler, O. Hartley, B. G. . Winchester, Tetrahedron, 1995, 57, 12611-12630; R. H. Furneaux, G. J. Gainsford, J. M. Mason, R C. Tyler, O. Hartley, B. G. Winchester, Tetrahedron, 1997, 53, 245-268 [83] P. Tyler, private communication to the author. [84] A. D. Elbein, T. Tsumilo, B. A. Sanford, K. B. Sharpless, C. Adams, Biochemistry, 1987, 26, 2502-2510. [85] G. Gradnig, A. Berger, V. Grassberger, A. E. Stiitz, G. Legler, Tetrahedron Lett., 1991, 32, 4889-4892. [86] Gradnig, G. Legler, A. E. Stiitz, unpublished results. [87] A. Welter, G. Dardenne, M. Marlier, J. Casimir, Phytochemistry, 1976, 2, 747-749. [88] P. J. Card, W. D. Hitz, J. Org. Chem., 1985, 50, 891-893. [89] K. K.-C. Liu, T. Kajimoto, L. Chen, Z. Zhong, Y. Ichikawa, C.-H. Wong, J. Org. Chem., 1991, 56, 6280-6289. [90] G. Legler, A. Korth, A. Berger, C. Ekhart, G. Gradnig, A. E. Stiitz, Carbohydr. Res., 1993, 250, 67-77. [91] Y.-F. Wang, D. P. Dumas, C.-H. Wong, Tetrahedron Lett., 1993, 34. 403-406 [92] T. M. Wrodnigg, A. E. Stiitz, S. G. Withers, Tetrahedron Lett., 1997, 38, 5463-5466; T. M. Wrodnigg, A. E. Stiitz, S. G. Withers, unpublished results. [93] W Dong, T. Jespersen, M. BoIs, T. Skrystrup, M. R. Sierks, Biochemistry, 1996, 35, 2788-2795. [94] Comparison of values is complicated by the fact that authors measure under different experimental conditions such as pH, temperature as well as concentrations of solutions employed.

References

187

[95] H. Umezawa, T. Aoyagi, T. Komiyama, H. Morishima; M. Hamada, T. Takeuchi, J. Antibiotics, 1974, 27, 963-969; Y. Nishimura, T. Satoh, T. Kudo, S. Kondo, T. Takeuchi, Bioorg. Med. Chem., 1996, 4, 91-96 and references cited there. [96] M. Ichikawa, Y. Igarashi, Y Ishikawa, Tetrahedron Lett., 1995, 36, 1767-1770. [97] R-H. Ducrot, J. Y Lallemand, Tetrahedron Lett., 1990, 31, 3879-3882. [98] R-H. Ducrot, J. Beauhaire, J. Y. Lallemand, Tetrahedron Lett., 1990, 31, 3883-3886; O. Duclos, A. Dureault, J. C. Depezay, Tetrahedron Lett., 1992, 33, 1059-1062; O. Duclos, M. Mondange, A. Dureault, J. C. Depezay, Tetrahedron Lett., 1992, 33, 8061-8064; R-D. Boyer, R-H. Ducrot, V. Henry on, J. Soulie, J.-Y Lallemand, Synlett, 1992, 357-359; R-D. Boyer, J.-Y. Lallemand, Synlett, 1992, 969-971. [99] C. R. Johnson, S. J. Bis, J. Org. Chem., 1995, 60, 615-623. [100] R-D. Boyer, J.-Y. Lallemand, Tetrahedron, 1994, 50, 10443-10458. [101] J. Soulie, T. Paitg, J.-R Betzer, J.-Y Lallemand, Tetrahedron, 1996, 52, 15137-15146. [102] N. Asano, E. Tomioka, H. Kizu, K. Matsui, Carbohydr. Res., 1994, 253, 235-245; N. Asano, K. Oseki, E. Tomioka, H. Kizu, K. Matsui, Carbohydr. Res., 1994, 259, 243255. [103] N. Asano, A. Kato, M. Miyauchi, H. Kizu, T. Tomimori, K. Matsui, R. J. Nash, R. J. Molyneux, Eur. J. Biochem., 1997, 248, 296-303 and lit. cited therein. [104] R. M. S. Harris, A. Aleshin, A. Golubev, L. M. Firsov, R. B. Honzatko, Biochemistry, 1993, 31, 1618-1626.

9

Potent Glycosidase Inhibitors: Transition State Mimics or Simply Fortuitous Binders? STEPHEN G. WITHERS, MARK NAMCHUK and RENEE Mosi

9.1

Introduction

Arguably the best approach to design of, as opposed to screening for, tight inhibitors is through consideration of the structure of the transition state ordinarily stabilized by the enzyme. The underlying concept here is that originating with Pauling [1, 2], and substantially elaborated by Jencks [3] which argues that enzymatic catalysis has its root in the specific stabilization afforded to the transition state by interactions derived at the active site of the enzyme. These are typically non-covalent interactions which may be present at the ground state but are optimized only at the reaction transition state. In this way the reaction transition state is selectively stabilized, thus the activation energy lowered. The concept of transition state analog design is based upon the notion of constructing a stable molecule which mimics the structure of the transition state, both structurally and electronically. Such a molecule should bind tightly to the enzyme since it is able to recruit the binding interactions truly intended for stabilization of the reaction transition state. The major attraction of this approach is that the interaction energies available are enormous, as therefore are the potential binding affinities. Indeed it can be shown that a perfect transition state analog should bind more tightly to the enzyme than the ground state species (enzyme/substrate affinity) by a factor equal to the ratios of the rate constants for the enzyme-catalyzed and uncatalyzed reactions [4-8]. These factors can be huge, ranging up to 1017 fold in some cases [9]. Of specific importance here is that very recent studies by Wolfenden show that these factors for glycosidases are in the order of 1016 fold (R. Wolfenden, personal communication). Transition state analog inhibitors for glycosidases therefore have enormous potential, with K\ values of 10~20 M being theoretically possible. The argument is therefore that good transition state analogs should function as tight binding inhibitors. However the converse is not necessarily true since not all tight binding inhibitors are necessarily transition state analogs. Nonetheless, there is a tendency to conclude that any tight binding inhibitor must be a transition state analog. Such conclusions can be extremely misleading if efforts are then focussed on increasing inhibitor affinity through 'improving' the structure to Jniinosugars ns Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

9.2 Transition State Theory and Mimicry

189

make it more closely resemble that of the specific substrate in question. To address this question some diagnostic tests need to be applied to determine whether or not an inhibitor is truly a transition state analog. These tests, and their use to probe transition state mimicry in glycosidases, are the focus of this chapter.

9.2 Transition State Theory and Mimicry In order to understand the basis of these tests it is important to first review some of the concepts of transition state theory and its application to the design of transition state analogs. This topic has been the subject of an excellent, recent review [1O]. Transition state theory, originating with Eyring, tells us that if we consider a simple reaction involving a single barrier, then the rate of the reaction depends upon the 'equilibrium constant' (K*) between the ground state and the 'activated complex' (essentially the transition state), as well as the rate of decomposition of this activated complex to products. The expression below relates the rate constant fc, to this equilibrium constant K* and factors describing the rate of decomposition, namely v, the frequency of the normal vibrational mode of the scissile bond and K, a transmission coefficient whose value is ordinarily unity.

(9-1)

k = KV&

The value of K* is simply related to the activation energy through the Boltzmann relationship. Catalysis by enzymes can therefore be thought of, in the simplest terms, according to the thermodynamic box shown below in which E and S represent free enzyme and free substrate, ES represents the enzyme/substrate (Michaelis) complex, ES* represents the enzyme/substrate complex at the transition state and S* represents the free substrate in this same transition state structure. This thermodynamic cycle provides the relationship shown in Equation 9-2 between the ground state dissociation constant KS and the theoretical dissociation constant for the transition state complex A>rs through the relative values of the pseudo equilibrium constants for the formation of the activated complexes Kun* and KC^. <9 2)

-

n ^

Figure 9-1. Thermodynamic cycle relating enzyme-catalyzed and spontaneous reactions.

ES

φ E + S

** E + P

190

9 Potent Glycosidase Inhibitors: Transition State Mimics . . .

As noted above, these pseudo equilibrium constants are directly proportional to the rate constants for each process. Therefore, if we can assume that the transmission coefficients and normal mode frequencies are comparable for the two processes (spontaneous and enzyme-catalyzed hydrolysis), then we can write Equation 9-3 which relates the binding affinity of the transition state to that of the ground state through the relative rate constants.

This expression is used to estimate the affinity of a perfect transition state analog by assuming that, to a first approximation, K8 can be replaced by Km, the Michaelis constant. K^s will be the Ki value of interest. Under these conditions the Ki value for the inhibitor can be calculated from the rate constant for spontaneous hydrolysis, £un, and the reciprocal of the second order rate constant for reaction of free enzyme and free substrate, kcat/Km, as shown in Equation 9-4.

Ki = Km

(9-4)

This analysis therefore shows us that transition state analog inhibitors should bind very much more tightly than the ground state, by a factor of kcai/kun. We need now to see how this analysis can be further used to provide a rigorous criterion of whether or not an inhibitor is truly mimicing the transition state. The relevant approach was first pointed out by Wolfenden [U]. He noted that changes in the structure of the substrate/inhibitor or enzyme should affect both the binding of the inhibitor and the binding of the reaction transition state (thus the rate of the enzymatic reaction) in equivalent ways. Thus changes in the free energy of binding of the inhibitor should be reflected in changes in the activation free energy of the catalyzed reaction. This is best expressed in the form of a linear free energy relationship as shown below.

log K1 α log

(9-5)

The slope of this line should be unity for a good transition state analog. Such a slope indicates that at the site of substitution the changes in binding energy effected by the substitution are identical to the changes in transition state stabilization. Any smaller or greater slope indicates a less perfect correlation. There are two ways in which these changes can be effected in a controlled manner. One approach involves first synthesizing a series of modified substrates along with a series of identically modified inhibitors. The kinetic parameters (&cat and Km) for hydrolysis of each of these substrates and for the inhibition (K1 values) by each of the modified inhibitors are then determined with the wild type enzyme. The alternative approach involves changing the enzyme itself by mutation. The residues mutated should be those thought to play a role in substrate recognition

9.3 Glycosidase Mechanisms and Transition States

191

and catalysis, though these are not necessarily those directly interacting. They could well be residues more deeply buried in the protein, but still playing a role in the correct formation of the active site. Indeed it is not essential to know the three-dimensional structure of the protein in order to do this, as the mutations could be any which leave the structure otherwise unchanged, but which have an effect upon the activity. With a series of mutants in hand, K1 values for the inhibitor of interest are then determined for each protein. These K{ values are then compared with a series of kinetic parameters (kcat/Km values) determined for each mutant using a defined substrate that is similar in dimensions to the inhibitor in question. There are advantages and disadvantages to each of these approaches and the choice is most likely dictated by practical questions of whether it is easier to generate a series of mutants, or to synthesize a wide range of modified substrates and equivalently modified inhibitors. However, while both approaches are valid there is a slight disadvantage to the modified substrate/inhibitor approach. If the substitution introduces, for example, an extremely electronegative group into the substrate, and the reaction proceeds through a cationic transition state, the kcat/Km value may well be lowered through inductive effects or field effects from the substituent. Similarly, these effects will come into play on the uncatalyzed reaction, though not necessarily to the same degree. Such effects would not be operational upon the transition state analog binding, with the consequence of additional scatter in the plot. Fortunately these electronic effects tend to be rather small in comparison to the changes in binding affinity.

9.3

Glycosidase Mechanisms and Transition States

There are two fundamental mechanisms of enzymatic glycoside hydrolysis, differentiated by the stereochemical outcome of the reaction catalyzed. For reviews of mechanisms and structures of glycosidases, see the following [12-16]. Thus the glycoside is cleaved either with inversion or retention of anomeric configuration, the corresponding enzyme being referred to, respectively, as either an inverting or a retaining glycosidase. Abbreviated mechanisms for these two classes of enzyme are shown below in Figure 9-2, the example shown being that of an a-glucosidase. The active site of each enzyme class is shown as containing two key carboxylic acid residues. Obviously large numbers of other residues are important in binding and catalysis, but these have been omitted from this figure for clarity, and because their identities vary with each enzyme. The roles and relative placements of these two carboxylic acids are different in the two cases, as illustrated below. For the inverting glycosidases reaction involves a direct displacement of the sugar aglycone by water, with one carboxylic acid providing general acid catalytic assistance to aglycone departure and the other serving as a general base, deprotonating the water as it attacks. In this case the two carboxylic acids must be sufficiently separated to allow both the substrate and a water molecule to be placed

192

9 Potent Glycosidase Inhibitors: Transition State Mimics...

/ \ HO HO

OH

°v- ··

—-

HO

Figure 9-2. Mechanisms of retaining (top) and inverting (bottom) glycosidases, illustrated with an a-glucosidase.

between them. Measurements of this average separation (oxygens to oxygens) for inverting glycosidases reveal that, on average, the two groups are some 9.0 - 9.5 A apart [13, 17, 18]. For the retaining glycosidases the reaction involves a double displacement mechanism in which a covalent glycosyl-enzyme intermediate is formed and hydrolyzed with general acid/base catalytic assistance. Thus in the first step one of the carboxylic acids functions as a general acid catalyst, protonating the leaving group oxygen much as with the inverting enzymes, while the other carboxylate functions as a nucleophile, attacking at the sugar anomeric center and forming a covalent glycosyl-enzyme intermediate. In the second step a water molecule attacks at the anomeric center in a general base-catalyzed process to displace the sugar off the enzyme. This mechanism requires the two carboxylic acid residues to be much closer together, since the nucleophile attacks directly onto the sugar. Indeed, inspection of the structures revealed that the two carboxyl groups are, on average, only 5 A apart [13, 17, 18]. Interestingly, in this mechanism, the same carboxylic acid residue functions as both the general acid and the general base catalyst raising questions as to how this residue can function efficiently in both roles. Recent direct measurements of the p^a values of this residue both in the free enzyme and in the glycosyl-enzyme intermediate have provided an answer. In the free enzyme the pKa of the acid catalyst is above the pH optimum, thus the group is in the protonated state, ready to

9.3 Glycosidase Mechanisms and Transition States

193

deliver a proton. Upon formation of the glycosyl-enzyme intermediate the pKa of this group drops substantially; 2.5 units in the case of the Bacillus circulans xylanase [19]. This leaves the group in a deprotonated state, ready to function as the general base catalyst. These pKa shifts are in direct response to the changing charge on the catalytic nucleophlle. In the free enzyme the catalytic nucleophlle has the lower pKa, presumably as a consequence of its environment at the active site. The presence of this negative charge raises the pKa of the acid/base group from that normally expected for a carboxylic acid. However, formation of the covalent intermediate removes the charge from the catalytic nucleophlle, allowing the pKa of the acid/base residue to drop to a lower value. This leaves it deprotonated and therefore in the correct form to act as the general base catalyst. This pKa cycling therefore automatically adjusts the ionisation state of the acid/base catalyst to suit its role at each step of the mechanism and is an inherent component of the mechanism, being directly controlled by the charge on the nucleophlle. Distortion of the substrate has long been suggested to be an important component of catalysis by glycosidases, having been first suggested for lysozyme on the basis of model building studies on the early X-ray crystal structure [20-22]. This has been a matter of some debate ever since [14], but recent results appear to substantiate these findings both in lysozyme [23, 24], and in other enzyme systems [25, 26] in which /?-glycosidic bonds are cleaved. Such distortion presumably serves to destabilize the substrate ground state, assisting passage through the transition state to products. The glycosyl transfer steps occurring in each mechanism proceed through transition states with substantial oxocarbenium ion character, as shown generically in Figure 9-3. The positive charge on the anomeric carbon is shared with the endocyclic oxygen. This requires double bond character to be present in the ring between C-I and the endocyclic oxygen, as shown. Consequently the transition state structure of the sugar is distorted relative to the ground state, either in the form of a half chair, a boat or a skew boat, as shown in Figure 9-4.

Figure 9-3.

Generic glycosidase transition states. OH

^

V^

^ OH

^

•

· .yw( \/ O

Figure 9-4. Half chair, boat and skew boat conformations of o-glucopyranose.

194

9 Potent Glycosidase Inhibitors: Transition State Mimics...

Support for such a transition state structure comes from a range of studies, as well, of course, as by analogy with the non-enzymatic hydrolysis mechanism. Thus α-secondary deuterium kinetic isotope effects (aDKIE's) of &H/&D = 1.1 1.25, which provide insights into changes in hybridization at the sugar anomeric center between the ground and transition states, have provided solid evidence for sp2 hybridization, therefore oxocarbenium ion character, in the transition states of inverting enzymes (see for example [27, 28]). Similarly, positive «DKIE's for both steps in the retaining mechanism have provided support for such a transition state in those cases [14]. Indeed it is interesting to note that, at least for the hydrolysis of aryl /?-glycosides, the aDKIE's measured for the deglycosylation step tend to be larger than those for glycosylation, suggesting that there is more oxocarbenium ion character at the second than the first transition state for /?-glycosidases [29, 3O]. A second piece of evidence has been the discovery of linear free energy relationships between rates of enzyme-catalyzed and non-enzymatic hydrolysis for a series of deoxy- and deoxyfluoro-glycosides [31-33]. The principal consequence on spontaneous hydrolysis of substitution of the hydroxyl by the very electronegative fluorine or by the relatively electron-donating hydrogen substituent is an electronic (field or inductive) effect on the stability of the oxocarbenium ion [34-36]). Thus the presence of a correlation between the enzymatic and non-enzymatic rates, albeit obscured by binding effects, indicates that the enzymatic reaction also proceeds through an oxocarbenium ion-like transition state. A minimized variant of this approach has been applied to measure the positive charge present at the transition state [37]. This involves measuring the rates of glycosidase-catalyzed turnover of a pair of substrates differing in the electronic demand placed at the cationic reaction center by a pair of substituents of different electronegativity. The pair employed are the appropriate glycosyl fluoride for the enzyme, and the 1,1difluoro glycoside. The idea is that the additional fluorine at C-I replaces a relatively electropositive hydrogen, yet is not much larger. Thus the magnitude of the rate reduction should reflect the size of the charge [38]. Results are again completely consistent with the development of positive charge. The final evidence is, of course, the tight binding of supposed transition state analogs; the focus of this chapter.

9.4

Tight-Binding Glycosidase Inhibitors

This general area has been the subject of an excellent review [39], and of course is the focus of many of the chapters of this book. It will only, therefore, be reviewed in sufficient detail here to permit the reader to follow later arguments regarding transition state mimicry. Historically, some of the first tight-binding glycosidase inhibitors discovered were the glyconolactones [4O]. These were quite reasonably argued to function as transition state analog inhibitors by virtue of their resemblance, particularly in their less-contributory charge-separated resonance form,

9.4 Tight-Binding Glycosidase Inhibitors

195

OH

,OH

HO

Figure 9-5. Resonance forms of o-gluconolactone.

to the oxocarbenium ion, as shown below in Figure 9-5. Unfortunately their instability has limited their application. A second class of (sometimes) tight-binding and equally simple sugar derivatives found was that of the glycosylamines, and in some cases the glycosamines, as shown in Figure 9-6. In this case the good inhibition has been more reasonably ascribed to the presence of a protonated (or protonatable) amine moiety which interacts electrostatically with the carboxylate groups at the active site. The naturally occurring hydroxylated piperidines such as nojirimycin and 1-deoxynojirimycin (Figure 9-6) were soon added to this list. These compounds clearly resemble glucose very closely, differing only in the presence of an endocyclic nitrogen rather than oxygen. A wide range of hydroxylated piperidines, pyrrolidines and other related structures has since been synthesized and tested, as described elsewhere within this book. Particularly notable are the isofagomines, which have OH

OH

OH

NH2

β-D-Glucosylamine

U11

^Ai

D-Glucosamine

Nojirimycin

OH

OH

f CT

OH l-Deoxynojirimycin

HO

Isofagomine

OH

OH

?

!

Swainsonine

HO

OH

\/-

C astanospermine Figure 9-6. Structures of some naturally occurring glycosidase inhibitors.

196

9 Potent Glycosidase Inhibitors: Transition State Mimics...

been reviewed very recently [41]. An additional class of inhibitors, isolated primarily from plant sources, was that of the bicyclic indole alkaloids such as castanospermine and swainsonine (Figure 9-6). Please see other chapters in this book and [39] for leading references. These latter compounds are generally believed to bind in a similar mode to that of the substrate sugar. This is best seen for the bicyclic compounds when drawn in an analogous fashion to glucose, as shown in Figure 9-6. However opinion is somewhat divided regarding whether or not these are transition state analogs. Proponents of their transition state mimicry would argue that their tight binding is a consequence of a flattening of the structure around the 'anomeric centre' of the structure enforced by their bicyclic structure, coupled with their charge, making them oxocarbenium ion mimics. Others would argue that they are simply fortuitously hydroxylated point charges. Considerable effort has been invested into attempts to develop stable analogs of glyconolactones which bind as well as, or better than, the parent compound. These efforts have resulted in compounds such as the hydroximo-lactones and the N-phenylcarbamates shown below (Figure 9-7) which in many cases function as excellent inhibitors [42]. This class has been expanded to include derivatives with a nitrogen at the 5-position in the hope that this would provide both an sp2-hybridized 'anomeric' center and a positive charge, thus bind much more tightly. The OH

OH

HO HO

OH V XT ^O^_^N

Hydroximo-lactone

N-Phenylcarbamate

HO OH

NH2

Amidine

OH^NH-NH 2

Amidrazone

OH

Nojirimycin tetrazole

OH

HO HO

Hydroximolactam

Figure 9-7. Structures of some synthetic glycosidase inhibitors containing an sp2hybridized 'anomeric center': gluco configuration.

9.5 Probing Transition State Mimicry

197

parent compound, the D-g/wco-amidine shown below, was indeed found to bind tightly, but was quite hydrolytically unstable. More stable derivatives such as the amidrazone, the gluconohydroximolactam and the nojirimycin tetrazole were therefore prepared and tested. These were indeed more stable, and in general were quite good inhibitors. It was claimed that the amidines and amidrazones, although tightly binding, were actually less specific, inhibiting a range of glycosidases [43-45]. However, subsequent studies [46] revealed that, in some cases at least, this was not true, and that really the enzymes chosen themselves had broad specificity for a range of substrates. Thus the broad-ranging inhibitor specificity was simply a reflection of the wide-ranging substrate specificity of the enzymes tested. In addition, controversy still exists concerning the tautomeric form of these compounds in solution (endo- or exocyclic double bond), thus the nature of their inhibitory action [43-45, 47]. The presence of the electronegative substituent on the exocylic nitrogen not only renders the compounds hydrolytically stable, but also lowers the pKa of the amidine system, altering its capacity to bear a positive charge. This can have an effect on the pH-dependence of inhibition.

9.5 Probing Transition State Mimicry 9.5.1

Nojirimycin tetrazoles

The first published study of a correlation between substrate and inhibitor specificity as a probe of transition state mimicry in glycosidases was that on the glucoand majmo-nojirimycin tetrazoles with a series of glucosidases and mannosidases [46]. In this study a pair of substrates and inhibitors was studied with the corresponding pair of glycosidases from a variety of sources. Kinetic parameters were therefore determined for thep-nitrophenyl glycosides of D-mannose and D-glucose (a and β as appropriate) with the following glycosidases: α-mannosidases from jack bean and almonds, α-glucosidase from yeast, /?-mannosidase from snail and /?-glucosidase from Agrobacterium sp. Inhibition parameters (Ki values) were then determined for the Ό-gluco- and D-manno-tetrazoles. A plot of log (Km/kcat) for each of these substrates versus the logarithm of the K\ value for the tetrazole inhibitor of the corresponding configuration was linear with a slope of 0.96 and a correlation coefficient of ρ = 0.9; Figure 9-8. It is important to note that the plot of log Km for the substrate versus log K1 for the corresponding tetrazole was a scatter plot with a correlation coefficient of ρ = 0.2. These findings therefore clearly establish a strong correlation between inhibitor binding and transition state affinity, while also establishing the absence of any significant correlation of the inhibitor binding with ground state affinity, as estimated through ^T1n values. The tetrazoles can therefore be classified as transition state analogs, albeit imperfect ones given the fact that they are only binding some 103 - 104 fold more tightly than the corresponding sugar (glucose or man-

198

9 Potent Glycosidase Inhibitors: Transition State Mimics.

-2

__j -2

Q

2

4

Figure 9-8. Plot of log (Km/kcat) for hydrolysis of a series of p-nitrophenyl glycosides versus ^i values for the corresponding nojirimycin tetrazoles. Taken from [46].

nose). They are therefore binding in the correct mode at the positions substituted, but elsewhere are picking up only a part of the stabilizing interactions afforded the transition state. Quite likely the majority of this missing binding is related to the charge separation occurring at the transition state in the formation of the oxocarbenium ion upon loss of the aglycone and its attendant negative charge. While resonance forms can be drawn for the tetrazole which reflect the desired charge separation, they will not be major contributory forms thus cannot fully capitalize on this binding energy. Interestingly it appears that another important component of the good binding is the presence of hydrogen bonding interactions between the 'anomeric' nitrogen of g/wco-nojiritetrazole or the 'anomeric' oxygen of o-gluconolactone and an active site proton donor, believed to be the acid catalyst. In support of this hypothesis is the finding that the g/wco-triazole in which the 'anomeric' nitrogen has been replaced by a carbon, is a very weak inhibitor of the enzyme [48]. Additional evidence exists that these are binding as transition state analogs. Firstly a K{ value was determined for the gluco-nojintetrazole with the Glu358Asp mutant of the Agrobacterium sp. /?-glucosidase. Glu358 is the catalytic nucleophile which has been shown [49] to play no significant role in ground state binding, but a major role at the transition state through the stabilization of positive charge developed on the anomeric carbon. A K1 value of 200 μΜ was measured, indicating that binding is approximately 200-fold weaker as a consequence of moving the charge approximately 1A further away. This correlates to a rate reduction (kcai/Km for PNPGIc) of some 10,000-fold. Similar results were obtained for the binding of D-gluconolactone. However the binding of obvious ground state mimics such as /?-glucosyl benzene was weakened only 5 -fold. A second piece of, rather more circuitous, evidence concerns the inhibition of a mechanisticallyrelated glycosyl transferase, glycogen phosphorylase. This enzyme catalyzes the reversible transfer of a glucose moiety from the non-reducing terminus of glycogen to phosphate. The transition state of this reaction must therefore incorporate both the glucosyl cation analog and the phosphate moiety. The g/wco-nojiritetrazole was found to be a good inhibitor of this enzyme, but only when phosphate was also present. Indeed a three-dimensional structure of the complex of this

9.5 Probing Transition State Mimicry

199

enzyme with g/wco-nojiritetrazole and phosphate was determined, directly demonstrating the coincident binding of the two ligands [5O]. Importantly, the binding of glucose and phosphate are mutually exclusive with this enzyme as demonstrated kinetically and structurally [51-53]. Similar kinetic results had been obtained for the inhibition by D-gluconolactone [54]. Therefore the g/wc<9-nojiritetrazole and gluconolactone must be binding as transition state analogs to this enzyme.

9.5.2

Acarbose as a transition state analog?

Acarbose is a naturally occurring pseudo-tetrasaccharide inhibitor of a range of aglucosidases of various sorts [55]. This compound is currently in clinical use for suppression of post-prandial blood glucose levels, the mode of action involving inhibition of pancreatic α-amylase and, more importantly, the brush border a-glucosidases [56, 57]. As shown in Figure 9-9 acarbose consists of a hydroxylated aminocyclohexene moiety (valienamine) attached to the 4-position of a 6-deoxyglucose unit via an 'alpha-anomeric' linkage. This fragment is itself a(l-4) linked to a maltose moiety. Acarbose is a particularly potent inhibitor of glucoamylases, having a Ki value of 1.1 χ lQ-l2MfortheAspergillusnigerglucoamylase [58-6O]. Acarbose incorporates many of the elements one might want to design into a transition state analog inhibitor. The polyhydroxylated cyclohexene ring of the valienamine moiety mimics a flattened glucose residue. However, the double bond character in this case is at a position equivalent to the C-5-O-5 bond and not between the (sugar) O-5 and C-I atoms as in the oxocarbenium ion. In addition the exocyclic amine moiety linking it to the adjacent sugar places a positive charge in the active site right where charge would be generated upon protonation of the glycosidic bond and close to the location of such charge in the oxocarbenium ion. Two studies have now been performed to check whether or not acarbose indeed functions as a transition state analog [59, 61]. Interestingly both studies have been performed on enzymes for which X-ray crystallographic three-dimensional structures of complexes of the enzymes with acarbose have been determined [62-64]. Also of interest is the fact that one of the enzymes, the Family 15 Aspergillus niger glucoamylase [65], is an exo-acting inverting glucosidase while the other, the Family 13 Bacillus circulans cyclodextrin α-glucanotransferase, is an endo-acting retaining glycosyl transferase [63, 64]. CH2OH

HO

Figure 9-9. Structure of acarbose.

200

9 Potent Glycosidase Inhibitors: Transition State Mimics. -t

·<; ·· ^>^··

2

•

1

5 ο °

-2

*/^

~ /^**^ ^

A --+

I

-12

ι

ι

-10

Ι

ι

Ι

-8

-6 |0

9 Ki

ι

Ι

-4

ι

Ι

-2

ι

!

O

Figure 9-10. Plot of log(#m/&cat) for ltose vs log ^i for acarbose with ma glucoamylase mutants. Taken from [ 5 9 ] .

The approach followed in both cases involved the generation of a series of mutants of the enzyme in question, then measurement of kinetic parameters for the hydrolysis of a defined substrate by each of these mutants. The inhibition of each of these mutants by acarbose was then investigated, and K1 values determined. Values of K1 for acarbose so determined were correlated with substrate Km/kcat values for each of the mutants. The first such study was performed with \heAspergillus niger glucoamylase using maltose as substrate, and the results are shown in Figure 9-10. The correlation shown (slope = 0.38; ρ = 0.88) is real, indicating some degree of transition state mimicry. However, the slope is significantly less than unity, indicating that mimicry is far from perfect in these regions. This would suggest that, despite the remarkable affinity of this interaction, indeed the highest affinity of any carbohydrate/protein complex yet determined, the compound is not a true transition state analog for this inverting glycosidase. However, the correlation is much better than that seen with kCSLt, and no correlation was seen with Km. There is definitely, therefore, some resemblance to the transition state in its binding mode. A second study of acarbose as a transition state analog is that performed recently on the retaining cyclodextrin glucanotransferase (CGTase) from Bacillus circulans [61]. The normal function of CGTase is to convert starch into cyclodextrins, (cyclic malto-oligosaccharides), via a transglycosylation reaction. Recent studies have shown that glycosyl fluorides can function as good substrates for this enzyme, the activated fluoride leaving group serving as a sufficiently good leaving group to replace the several sugar moieties normally required on the reducing end side of the bond cleaved [66]. Use of a fluoride-sensitive electrode permits facile monitoring of these reactions. Indeed even α-glucosyl fluoride serves as a very good substrate, thus extended binding into the non-reducing (minus) sites is not essential when this excellent leaving group is present. (Note that we are using the subsite terminology recently suggested [67]). However, α-maltotriosyl fluoride is a 10fold better substrate, based upon kcat/Km values. Values of Jkcat, Km and kcat/Km were determined for the reaction of both α-glucosyl fluoride and a-maltotriosyl fluoride with each of 8 mutants and with the wild type enzyme. In addition, K1

201

9.5 Probing Transition State Mimicry 8

ο

6

6

1

I

I

2 0

^

2 σ) Λ ο Ο

D) _0

φ

*/^ /^·

O

-4

-4 -4

-2

* /χ"·· Jfr^*

_

^^/ I

ι

!

-4

O

• T/ //^

-λ

-2

-2

I

1B

14 -*

1

ι

-2

I

O

"OgK1

log K1

Figure 9-11. Plots of log (Km/kcat) for (A) α-glucosyl and (B) α-maltotriosyl fluorides versus log K{ for acarbose with CGTase. values were measured for the inhibition of each mutant by acarbose. Plots of log (Km/kcat) versus log K1 for acarbose are shown in Figure 9-11. As can be seen, a much better correlation was obtained with the data for a-glucosyl fluoride (slope = 2.2, ρ = 0.98) than with that for α-maltotriosyl fluoride (slope = 1.61,p = 0.90). This finding is quite consistent with the expected binding modes for these two substrates and that for acarbose. As is shown in Figure 9-12, the two substrates must bind in such a way that the fluoride moiety sits in the +1 site, and the sugars in the minus sites. Acarbose must bind with the valienamine moiety in the -1 site and the rest in the plus sites. HOH2Cv

HO" HQ ^^

HOH2C

HOH2Cx

0

\ OH

F

HOH2C

HOH2C

H3C

HOH2C

0

HOH2C

Figure 9-12. Binding modes for acarbose and malto-oligosaccharyl fluorides.

202

9 Potent Glycosidase Inhibitors: Transition State Mimics...

The only occupied subsite common to the substrates and to acarbose is the -1 site, thus effects of mutations upon interactions at that site will dominate the correlation. Effects will be particularly large if the mutation directly removes interactions at that site, though mutations at more remote sites that affect interactions will also be sensed. However remote mutations that do not affect interactions at that site will have no effect. Thus mutations affecting only interactions at the +1, +2 and +3 sites will affect the binding of acarbose, but will not necessarily affect Km/kcat values for the two substrates. Similarly, mutations affecting interactions in the -2 and -3 sites will affect Km/kcat values for α-maltotriosyl fluoride, but not values for acarbose. These site mismatches will result in scatter in the plots. However, since there is less of a mismatch for α-glucosyl fluoride, and since the interactions are likely optimized at the -1 site, the correlation in that case is much better. Acarbose therefore functions as a transition state analog for both an inverting and a retaining α-glucosidase/a-glucosyl transferase, though much more so in the latter case. This concurs with the generally held view that the transition states for inverting and retaining glycosidases are very similar, as measured through kinetic isotope effects [27, 28, 68], and as shown in several cases now by the conversion of a glycosidase from one mechanism to the other through mutation [18, 69]. This similarity of transition states also extends to the relationship between a- and/?-glycosidases. This is amply illustrated in Figure 9-8 since this correlation contains data from both classes of enzyme.

9.5.3

Castanospermine and deoxynojirimycin as transition state analogs?

Detailed kinetic studies have been performed on theAgrobacterium sp. /?-glucosidase with a wide range of substrate analogs, including an extensive series that is selectively epimerised, deoxygenated and fluorinated around the sugar ring. Not only steady state, but also pre-steady state kinetic analyses have been performed on this system [30, 31]. This collection of data afforded the opportunity to search for correlations between these kinetic parameters and K{ values for similarly modified inhibitors. Fortunately a wide range of these analogs was available through collaborations with a number of researchers who are named in the acknowledgements. Both castanospermine and deoxynojirimycin are good competitive inhibitors of this enzyme with K1 values of 2.8 and 12 μΜ respectively. Neither inhibitor shows signs of slow binding when assayed at 37 0C. However, when assayed at 25 0C, slow onset of inhibition was clearly observed for castanospermine. This is a commonly observed, but far from ubiquitous, phenomenon with tight binding inhibitors and has been variously ascribed to slow conformational changes in the protein required for the binding of the inhibitor, slow protonation of the inhibitor or of the substrate or several other possibilities [39]. However to avoid this problem in this 0 case all assays were performed at 37 C, the temperature at which the enzyme is normally assayed.

203

9.6 Conclusions A

O

1

*

2

ί-

* -

3

I

-

*

-4

I

2

I

-

1

1

.

I1

*

-4

•

2 ~

2~ ^

O)

B

O

0

1OgK1

*

1

1

ι

2

2

-

3

I

ι

-

I

2

ι

-

I

ι

1

I

0

ι

1

I

2

ι

3

log K,

Figure 9-13. Plots of log (Km/kcat) for 2,4-dinitrophenyl glycosides vs. log K\ for (A) deoxynojirimycins and (B) castanospermines with Agrobacterium sp. /?-glucosidase.

Plots of log (Km/kcai) versus log Ki for the castanospermines and the deoxynojirimycins are presented in Figure 9-13. It is quite apparent that, in contrast with the cases previously presented, no significant correlation exists. Neither does any significant correlation exist between the K1 values and values of kcat or Km (not shown). This clearly shows that neither of these classes of compound functions as a transition state analog, despite their high affinity binding. Rather, they appear to act as fortuitous inhibitors in which the charged ammonium moiety binds tightly to the predominantly anionic active site and the set of hydroxyl groups forms an array of hydrogen bonding partners which have evolved to pick up strong interactions with the enzyme active site. They may, or may not, bind in the same orientation as glucose, but they certainly do not bind in a manner which mimics the reaction transition state. Similar concerns about the binding of deoxynojirimycins and related compounds have been expressed previously [70], primarily on the basis of a lack of correlation between the pH dependences of inhibitor binding and of kinetic parameters (kc&t/Km) for substrate hydrolysis. These concerns have been further amplified since [39]. In addition, others have noted difficulties in correlating inhibitor modifications with activity in a predictable manner, based upon substrate configuration [71 ]. It therefore seems clear that these two classes of inhibitor, at least in terms of binding to Agrobacterium sp. /?-glucosidase, but also, by extension, to other glycosidases, do not function as transition state analogs.

9.6

Conclusions

The approach described provides a rigorous method to probe whether tight binding glycosidase inhibitors are indeed transition state mimics. The finding that the nojirimycin tetrazoles and acarbose are true transition state analogs of both inverting and retaining glycosidases, while the nojirimycins and castanospermines are not,

204

9 Potent Glycosidase Inhibitors: Transition State Mimics...

indicates that true transition state mimicry demands sp2 hybridization at the anomeric center. Formal charge is not essential, though analogous charge delocalization, as generated in some resonance forms, may well assist this process. The nojirimycins and castanospermines, and likely by extension the swainsonines and australines, are essentially the result of Nature's combinatorial library of polyhydroxylated cyclic amines. Attempts to take advantage of the 1016 fold affinity increases available to transition state analogs should therefore focus upon maintaining an sp2 geometry at the 'anomeric centre', optimizing charge separation with placement of positive charge at that position, and incorporating the natural aglycone.

Acknowledgements The authors thank the Protein Engineering Network of Centres of Excellence and the Natural Sciences and Engineering Research Council of Canada for financial support of this work. We also thank Drs. Fleet, Furneaux, Getman, Gravestock, and Barnes for their generous gifts of compounds.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

L. Pauling, Chem. Eng. News, 1946, 24, 1375. L. Pauling, Am. ScL, 1948, 36, 51-58. W. P. Jencks, Adv. Enzymol., 1975, 43, 219-410. J. L. Kurz, /. Am. Chem. Soc., 1963, 85, 987-991. R. Wolfenden, Nature, 1969, 223, 704-705. R. Wolfenden, Ace. Chem. Res., 1972, 5, 10. G. E. Lienhard, Annu. Rep. Med. Chem., 1972, 7, 249. G. E. Lienhard, Science, 1973, 180, 149-154. A. Radzicka, R. Wolfenden, Science, 1995, 267, 90-93. M. M. Mader, P. A. Bartlett, Chem. Revs., 1997, 97, 1281-1301. J. Frick, R. Wolfenden, in Design of Enzyme Inhibitors as Drugs (Ed.: M. Sandier, J. H. Smith), Oxford University Press, New York, 1989, p. 19. G. Davies, B. Henrissat, Structure, 1995, 3, 853-859. J. McCarter, S. G. Withers, Curr. Opin. Struct. Biol., 1994, 4, 885-892. M. L. Sinnott, Chem. Rev., 1990, 90, 1171-1202. A. White, D. R. Rose, Curr. Opin. Struct. Biol, 1997, 7, 645-651. G. Davies, M. L. Sinnott, S. G. Withers, in Comprehensive Biological Catalysis, Vol. 1, Ch. 3, (Ed.: M.L. Sinnott), Academic Press, New York, 1998. A. White, S. G. Withers, N. Gilkes, D. Rose, Biochemistry, 1994, 33, 14743-14749. Q. Wang, R. W. Graham, D. Trimbur, R. A. J. Warren, S. G. Withers, J. Am. Chem. Soc., 1994, 116, 11594-11595. L. P. Mclntosh, G. Hand, P. E. Johnson, M. Joshi, M. Korner, L. A. Plesniak, L. Ziser, W. W. Wakarchuk, S. G. Withers, Biochemistry, 1996, 35, 9958-9966.

References

205

[20] T. Imoto, L. Johnson, A. North, D. Phillips, J. Rupley, in The Enzymes (Ed.: P. Boyer), Academic Press, New York, 1972, pp. 666-668. [21] C. C. F. Blake, D. F. Koenig, G. A. Mair, A. C. T. North, D. C. Phillips, V. R. Sarma, Nature, 1965, 2(96, 757-761. [22] D. C. Phillips, Proc. Natl Acad. ScL U.S.A., 1967, 57, 484-495. [23] A. T. Hadfield, D. J. Harvey, D. B. Archer, D. A. MacKenzie, D. J. Jeenes, S. E. Radford, G. Lowe, C. M. Dobson, L. N. Johnson, /. MoL BioL, 1994, 243, 856-872. [24] N. C. J. Strynadka, M. N. G. James, /. MoL BioL, 1991, 220, 401-424. [25] G. Sulzenbacher, H. Driguez, B. Henrissat, M. Schulein, G. J. Davies, Biochemistry, 1996, 35, 15280-15287. [26] I. Tews, A. Perrakis, A. Oppenheim, Z. Dauter, K. S. Wilson, C. E. Vorgias, Nature Struct. BioL, 1996, 3, 638-648. [27] Y. Tanaka, W. Tao, J. S. Blanchard, E. J. Hehre, /. BioL Chem., 1994, 269, 3230632312. [28] H. Matsui, J. S. Blanchard, C. F. Brewer, E. J. Hehre, J. BioL Chem., 1989, 264, 87148716. [29] D. TuIl, S. G. Withers, Biochemistry, 1994, 33, 6363-6370. [30] J. B. Kempton, S. G. Withers, Biochemistry, 1992, 31, 9961-9969. [31] M. N. Namchuk, S. G. Withers, Biochemistry, 1995, 34, 16194-16202. [32] J. McCarter, M. Adam, S. G. Withers, Biochem. J., 1992, 286, 721-727. [33] I. P. Street, K. Rupitz, S. G. Withers, Biochemistry, 1989, 28, 1581-1587. [34] B. Capon, Chem. Revs., 1969, 69, 407-498. [35] S. G. Withers, M. D. Percival, I. P. Street, Carbohydr. Res., 1989, 187, 43-66. [36] S. G. Withers, D. J. MacLennan, I. P. Street, Carbohydr. Res., 1986, 154, 127-144. [37] I. P. Street, J. B. Kempton, S. G. Withers, Biochemistry, 1992, 31, 9970-9978. [38] A. Konstantinidis, M. L. Sinnott, Biochem. J., 1991, 279, 587-593. [39] G. Legler, Adv. Carbohydr. Chem. Biochem., 1990, 48, 319-385. [40] J. Conchie, G. A. Levvy, Biochem. J., 1957, 65, 389-392. [41] M. BoIs, Ace. Chem. Res., 1998, 31, 1-8. [42] A. Vasella, R Ermert, R. Hoos, A. B. Naughton, K. Rupitz, W. Thiel, M. Weber, W. Weber, S. G. Withers, in Proceedings of the Alfred Benzon Symposium 36 (Ed.: H. Clausen, K. Bock), Munksgaard, Copenhagen, Denmark, 1994, Synthesis and Evaluation of New Glycosidase Inhibitors, pp. 134-150. [43] B. Ganem, Ace. Chem. Res., 1996, 29, 340-347. [44] G. Papandreou, M. K. Tong, B. Ganem, J. Am. Chem. Soc., 1993, 775, 11682-11690. [45] M. K. Tong, G. Papandreou, B. Ganem, J. Am. Chem. Soc., 1990, 772, 6137-6139. [46] P. Ermert, A. Vasella, M. Weber, K. Rupitz, S. G. Withers, Carbohydr. Res., 1993, 250, 113-129. [47] R. Hoos, A. B. Naughton, W. Thiel, A. Vasella, W. Weber, K. Rupitz, S. G. Withers, HeIv. Chim. Acta, 1993, 76, 2666-2686. [48] T. D. Heightman, M. Locatelli, A. Vasella, HeIv. Chim. Acta, 1996, 79, 2190-2200. [49] S. G. Withers, K. Rupitz, D. Trimbur, R. A. J. Warren, Biochemistry, 1992, 31, 99799985. [50] E. P. Mitchell, S. G. Withers, P. Ermert, A. T. Vasella, E. F. Garman, N. G. Oikonomakos, L. N. Johnson, Biochemistry, 1996, 35, 7341-7355. [51] S. G. Withers, N. B. Madsen, B. D. Sykes, Biochemistry, 1982, 27, 6716-6722. [52] S. R. Sprang, E. J. Goldsmith, R. J. Fletterrick, S. G. Withers, N. B. Madsen, Biochemistry, 1982, 27, 5364-5371. [53] S. G. Withers, P. J. Kasvinsky, N. B. Madsen, B. D. Sykes, Biochemistry, 1979, 18, 5342-5348. [54] A. M. Gold, E. Legrand, G. R. Sanchez, /. BioL Chem., 1971, 246, 5700-5706. [55] E. Truscheit, W. Frommer, B. Junge, L. Muller, D. D. Schmidt, W. Wingender, Angew. Chem. Int. Ed. EngL, 1981, 20, 744-761. [56] R. Coniff, A. Krol, CZm. Therap., 1997, 79, 16-26. [57] S. P. Clissold, C. Edwards, Drugs, 1988, 35, 214-243.

206

9 Potent Glycosidase Inhibitors: Transition State Mimics...

[58] R Lalegerie, G. Legler, J. M. Yon, Biochimie, 1982, 64, 977-1000. [59] C. R. Berland, B. W. Sigurskjold, B. Stoffer, T. P. Frandsen, B. Svensson, Biochemistry, 1995,34, 10153-10161. [60] B. W. Sigurskjold, C. R. Berland, B. Svensson, Biochemistry, 1994, 33, 10191-10199. [61] R. Mosi, H. Sham, B. Dijkstra, S.G. Withers, Submitted, 1998. [62] A. E. Aleshin, L. M. Firsov, R. B. Honzatko, /. Biol Chem., 1994, 269, 15631-15639. [63] B. Strokopytov, D. Penninga, H. J. Rozeboom, K. H. KaIk, L. Dijkhuizen, B. Dijkstra, Biochemistry, 1995, 34, 2234-2240. [64] B. Strokopytov, R. M. A. Knegtel, D. Penninga, H. J. Rozeboom, K. H. KaIk, L. Dijkhuizen, B. W. Dijkstra, Biochemistry, 1996, 35, 4241-4249. [65] B. Svensson, M. S0gaard, J. BiotechnoL, 1993, 29, 1-37. [66] R. Mosi, S. He, J. Uitdehaag, B. W. Dijkstra, S. G. Withers, Biochemistry, 1997, 36, 9927-9934. [67] G. J. Davies, K. S. Wilson, B. Henrissat, Biochem. J., 1997, 327, 557-559. [68] E. J. Hehre, S. Kitahata, C. F. Brewer, J. Biol. Chem., 1986, 261, 2147-2153. [69] R. Kuroki, L. H. Weaver, B. W. Matthews, Nature Struct. Biol., 1995, 2, 1007-1011. [70] M. P. Dale, H. E. Ensley, K. Kern, K. A. R. Sastry, L. D. Byers, Biochemistry, 1985, 24, 3530-3539. [71] T. Kajimoto, K. K. C. Liu, R. L. Pederson, Z. Zhong, Y. Ichikawa, J. A. Porco, C.-H. Wong, J. Am. Chem. Soc., 1991, 113, 6187-6196.

10 Iminoalditols as Affinity Ligands for the Purification of Glycosidases ANNA DE RAADT, CHRISTIAN EKHART, GUNTER LEGLER and ARNOLD E. STUTZ

10.1 Introduction In order to understand how enzymes bind their substrates, mediate catalysis and transduce energy and information, it is desirable to obtain highly purified proteins to allow detailed chemical and physical investigations of enzyme structure. In addition, these purified biocatalysts can be exploited to synthesize compounds which otherwise are frequently cumbersome to prepare using solely chemical approaches. The separation and isolation of these proteins is a time-consuming task which can be achieved by three key approaches; namely electrophoresis, ultracentrifugation, and chromatography. More information about these methods can be found in many biochemistry text books [1]. In particular, affinity chromatography [2] is a powerful technique and generally applicable means of purifying enzymes. This approach takes advantage of the interaction between enzymes and their substrates or inhibitors. An appropriate substrate or inhibitor is covalently attached to an insoluble support. Subsequently, the protein mixture is added and the desired enzyme is then selectively bound to this molecule. Other, unwanted proteins remain largely unbound and can be removed by washing with buffer. The desired enzyme can be subsequently obtained by eluting the column with a different buffer, a solution of the enzyme substrate, or the free inhibitor. However, problems can be experienced with this method if the enzyme is complexed too strongly to the covalently bound affinity ligand, and more drastic conditions for enzyme elution - which can cause denaturation - are necessary. Initial attempts to isolate and purify glycosidases from a biological matrix date back to the early 1970s [3-11]. This is exemplified in one of the seminal papers of this period by Junowicz and Paris [4], who employed (4-amino)phenyl-/?-D-glucosaminides bound to Sepharose 4B with various spacer-arms for the purification of hexosaminidases. With the aid of a /?-D-glucopyranosiduronic acid derivative bound to the same solid support, an overall purification of a /?-D-glucuronidase of over 1000-fold with respect to the original tissue could be achieved. SubIminosugars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

208

10 Iminoalditols as Affinity

Ligands...

Solid support 1

H

strate-specific versus non-specific binding of glycosidases was investigated by Mega and Matsushima [11, 12]. These workers examined the binding properties of partially purified glycosidase mixtures from takadiastase as well as abalone liver. As ligands, immobilized (4-amino)phenyl-/^D-glucosaminide, bound by a 6-aminohexanoic acid spacer (Scheme 10-1), the corresponding /?-D-glucopyranoside as well as D-glucosamine with a spacer attached to the amino group at C-2 and related compounds were examined. In a similar line of research, Kaneki and Tanaka [13] investigated the adsorption behaviour of various /^-glycosidases with /?-D-glycosylamines immobilized on agarose via a 6-aminohexanoyl spacer-arm. They reported that all glycosidases probed were adsorbed when buffer concentrations remained as low as 20 mM, but the bioselectivities observed were not found to be consistent. As potential affinity ligand, /?-thioglycosides of D-galactose, D-glucose and Dmannose as well as an α-glycoside of L-fucose with (4-amino)thiophenol were synthesized [14-16]. The fucoside was found to exhibit a K{ of 710 μΜ with a clam α-L-fucosidase. Many of the early efforts reported have aimed at the preparation of suitable ligands for the affinity chromatography of N-acetylhexosaminidases and will be referred to in the corresponding section.

10.2

D-Glucosidases

10.2.1

Glucosidase I

During the course of their investigations into the biochemical properties of trimming glucosidases, Hettkamp et al. [17] found that glucosidase I, a highly specific enzyme involved in the processing of asparagine-linked oligosaccharides of glycoproteins cleaving terminal 1,2-linked α-D-glucopyranosides, is strongly inhibited by 1-deoxynojirimycin (1, 50% inhibition at 3 μΜ). Based on their discovery that N-dodecyl-1-deoxynojirimycin (2) is also a potent inhibitor of this enzyme (ICso 8 μΜ), these investigators designed N-(5-carboxypentyl)-l-deoxynojirimycin (3) as a potentially suitable ligand for its purification and separation from glucosidase II. Gratifyingly, this derivative inhibited glucosidase I as strongly as did the N-dodecyl compound 2. An affinity resin was prepared by attachment of compound 3 via the carboxyl group onto the terminal amino function of the C-6 spacer-arm of AH (amino-

70.2 D-Glucosidases

H0 OH

\ 2

Scheme 10.2

R-H R = (CH2)9CH3

3 R = (CH2)5C02H 4 R = (CH2)9CO2H

209

ι

NR

Η

5 R-H 6 R = (CH2)5C02H

hexyl)-Sepharose 4B with the aid of a water-soluble carbodiimide as the coupling reagent. Utilizing this approach, glucosidase I with a molecular mass of 85 kDa could be purified approximately 960-fold as compared with the crude microsome fraction for subsequent characterization [18], with an overall yield of 55 %. A purification procedure for glucosidase I from yeast was reported subsequently [19]. Applying the same protocol, glucosidase I from pig liver could be purified over 160-fold. The homogeneity of the preparation was further enhanced to 400-fold by adding a precipitation step with polyethylene glycol to the established procedure [2O].

10.2.2

Glucosylceramidase

Employing the same approach, a method for the affinity purification of human acid /^-glucosidase (glucosylceramidase, EC 3.2.1.45) was devised. Based on information obtained in the course of the characterization of glucosylceramidase from calf spleen [21], the normal human placental enzyme as well as the Gaucher diseaseassociated splenic glucosidase were isolated [22]. In this investigation aminohexyl-Sepharose-bound N-(9 -carboxynonyl)- and N-( 11 -carboxyundecyl)-1 deoxynojirimycin were employed as affinity ligands. Yields ranged from 10 to 20 %, and the reported degree of purification for the spleen enzyme was approximately 7500-fold.

10.2.3

Cytosolic /?-glucosidase

Cytosolic /^-glucosidase from calf liver was isolated by the same basic procedure, as described in the previous section, using N-(9-carboxynonyl)-!-deoxynojirimycin (^j 8.2 μΜ) attached to aminohexyl-Sepharose [23]. The length and lipophilicity of the alkyl chain of the N-substituent proved to be essential as preliminary results with shorter chains indicated K1 values in the mM range. These values are consequently too high for binding purposes. The purification obtained was about 350-fold with an overall yield of 37%. This enzyme had a molecular mass of 52.5 kDa and, interestingly, also exhibited /?-galactosidase activity.

210

10 Iminoalditols as Affinity Ligands...

10.2.4 Microsomal bile acid /?-glucosidase Recently, in the course of an investigation into the biological significance of bile acid 3 -0-glucosylation, a novel microsomal /?-glucosidase with a molecular mass of 100 kDa was purified to apparent homogeneity employing N-carboxynonyl-1-deoxynojirimycin (4) on AH-Sepharose 4B as the ligand [24]. The overall purification was reported to be about 73 000-fold. This enzyme was shown to be highly specific for bile acid 3-0-/?-D-glucoside because 6-0-/?-D-glucosides were found to be resistant to hydrolysis.

10.3

D-Mannosidases

D-Mannosidases of glycoprotein trimming have been the main targets of affinity purification in this group of enzymes. Following their successful purification procedure for trimming glucosidase I from calf liver [17], Schweden et al, for the first time, employed an immobilized form of 1-deoxymannojirimycin (5) for the isolation of a trimming mannosidase with a molecular mass of 56 kDa [25]. N-5-Carboxypentyl-l-deoxymannojirimycin (6) linked to aminohexyl-Sepharose 4B was used as the affinity resin. The affinity chromatographic step resulted in an ~ 120-fold increase of specific activity over crude microsomes whereas other D-mannosidases were found to be ten times less effectively concentrated. Further purification resulted in a 5-fold activity increase leading to an approximate overall 2200-fold purification factor for the Man9 mannosidase under consideration. Subsequently, a trimming Man9 mannosidase with a molecular mass of 49 kDa from pig liver was purified by the same method [26]. In this case, the affinity step increased the specific activity by a factor of 280. This allowed, after a total of five purification steps, a 16 000-fold enrichment of the final preparation as compared with crude microsomes, the overall recovery of this particular enzyme being 15-20 %. A Golgi-located endo-a-D-mannosidase of glycoprotein trimming with a molecular mass of around 60 kDa was isolated from rat liver [27]. In this case the investigators applied a different purification approach due to the strong inhibitory potency of the 3-O-glucosylated iminoalditol [28] which inherently resulted in the difficult recovery of the bound glycosidase. Carboxyoctyl 3-0-a-D-glucopyranosyl-a-D-mannopyranoside (7) was chosen as the affinity ligand. The reported four-step procedure resulted in a 70 000-fold overall purification. A Golgi Man9 trimming mannosidase with a molecular weight around 50 kDa was recently isolated from hen oviduct with the aid of a N-carboxypentyl-1-deoxymannojirimycin AH-Sepharose. This improved the enzyme purity by 680- to 1000-fold with a recovery of over 70% from the affinity chromatography step [29]. This enzyme was found to be capable of removing three mannosyl residues from the original substrate.

211

10.4 N-Acetyl^-D-hexosaminidases

/COOH

Scheme 10.3

10.4

O"

\^

\χ^

\^

^^

Af-Acetyl-/?-D-hexosaminidases

Early studies revealed that (4-aminobenzyl)thio N-acetylglucosaminide could be attached to Sepharose via a spacer-arm consisting of 1,6-diaminohexane and succinic acid [3O]. Employing this affinity resin, N-acetyl-/?-D-hexosaminidase from jack bean could be purified 150-fold with a recovery of 60-70%. Analogous observations were made for the corresponding enzyme from wheat-germ and similar levels of purification were found by other workers employing related ligands for N-acetylhexosaminidases from human skin fibroblasts [31], urine [32], and rat liver [4]. A somewhat higher yield was achieved when 2-acetamido-N-carboxypentyl-2-deoxy-D-glucopyranosylamine was used for the isolation of human placenta enzyme [33]. Surprisingly, in terms of enzyme specificity, Pokorny and Glaudemans found a 310-fold purification of N-acetyl-/?-D-hexosaminidase from bull epididymis with a resin obtained by coupling 2-acetamido-2-deoxy-D-mannono-l,4-lactone to Sepharose 4B via a benzidine spacer-arm [34]. The structure of the ligand is believed to be as depicted in Scheme 10-4. A similar degree of purification was reported for three isoenzymes isolated from human urine [35], albeit with low recovery from the resin. An example for the efficiency of iminoalditols as affinity ligands was reported for the affinity purification of the enzyme from bovine kidney [36]. Employing the N-carboxypentyl derivative 8 of l,2,5-trideoxy-2-acetamido-l,5-imino-D-glucitol (2-acetamido-l,2-dideoxynojirimycin, 9) bound to aminohexyl-Sepharose, two isoenzymes could be purified 6600- and 7300-fold, respectively. Due to the strong interaction of the ligand with the enzyme, the hexosaminidase had to be recovered under relatively harsh conditions by effecting partial (reversible) denaturation employing 4M urea solution. As with α-L-fucosidases, less powerful inhibitors than compound 8 could be envisaged to be more convenient ligands in terms of recovery and procedure simplicity.

Sepharose

Scheme 10.4

212

10 Iminoalditols as Affinity

NR \

ur^T^ -7 C 3 / NR/

H

NHAc

8 9

I OH HO

R = (CH2)5CO2H R =H

10.5

Ligands...

10 R = H 11 R = (CH2)5C02H

Scheme 10.5

0-L-Fucosidases

The significantly higher concentrations of L-fucose in cancer cell glycoproteins [31] as well as the related higher activity of a-L-fucosidases [38], in addition to the other important roles of L-fucose-containing glycoconjugates, have triggered considerable interest in the enzymes responsible for the metabolism of this sugar, such as fucosyltransferases and fucosidases. Simple L-fucose derivatives [3941], such as 1-thiofucopyranosides [42], have been employed initially as affinity ligands for the isolation of α-L-fucosidases from various sources, for example rat epididymis [42]. Fucose bound to Sepharose 4B, via a linker derived from epichlorohydrin, was employed for the preparation of an α-L-fucosidase from monkey brain. However, hexosaminidases could not be entirely removed by this approach [43]. In 1979, attempts at α-L-fucosidase isolation by affinity chromatography based on the immobilization of a fucose-containing pentasaccharide via a 1-Nderivative were reported [44]. The purification reported for this preparation was 150-fold. This was subsequently surpassed by a different approach leading to a purity which was a factor often higher [45]. Initial experiments aiming at applying iminocyclitols for this purpose were reported by Paulsen and Matzke. Following syntheses of l,5-dideoxy-l,5-imino-L-fucitol (1-deoxy-L-fuconojirimycin, 10) [46], a highly potent inhibitor of fucosidases exhibiting K1 values in the low nanomolar range, these workers designed an approach also allowing modifications at C6 of this compound [47]. In the course of their work they also prepared the N-carboxypentyl derivative (11) of compound 10. Unlike previous means of immobilization [17], these workers applied a method introduced by Pinto and Bundle [48] to convert the terminal carboxyl group via the hydrazide into the carboxylic azide which they could couple directly with amino-Sepharose. Because of the ex-

Sepharose

12

Scheme 10.6

10.6 Miscellaneous

213

(CHACO2H H

1--'-T--:: OH HO

Scheme 10.7

13

Λ.ΠΗ

\ H?/1 HO

14

tremely powerful interaction of compound 11 with the enzymes under consideration, recovery of the bound enzyme from this affinity ligand was found to be somewhat troublesome and led to reduced yields and/or partial denaturation of the protein due to the comparably harsh stripping conditions necessary. Subsequently, a range of less potent C-6-modified derivatives was synthesized in order to circumvent this problem [49]. Nevertheless, compound 10, containing amino-Sepharose resin, was successfully employed for the isolation of a novel α-L-fucosidase from almond [5O]. The affinity purification step turned out to be crucial to remove side activities from the preparation. The enzyme was purified 163 000-fold and could be shown to be a native dimer with a molecular mass of 54 kDa per subunit. For preparative purposes and order to obtain sufficient amounts of enzyme for α-L-fucosylations of various glycosyl acceptors, Svensson and Thiem took advantage of a different affinity ligand based on a C-glycosidically bound fucopyranosyl residue (12) [51]. The recovery of porcine liver fucosidase in this one-step procedure was reported to range between 65 and 80 % with a purification of 3700-fold. Recently, for the isolation of α-L-fucosidase from bovine kidney, a derivative of 10 was exploited which lacked the methyl group at C-5 [52]. This substituent is a crucial feature for the strong interaction of this inhibitor with the enzyme [52], Consequently, the 5-nor-derivative 13 exhibited a K1 value which was about three orders of magnitude higher than that of the parent compound (4.3 μΜ vs. 3 nM at pH 6) and, accordingly, more convenient for the recovery of resin-bound enzyme. Furthermore, l,5-dideoxy-l,5-imino-D-arabinitol (13) could be easily prepared from D-arabinose in a simple and high-yielding sequence. The two-step purification procedure allowed a 7400-fold purification and 63% recovery of enzyme.

10.6

Miscellaneous

Lysosomal /?-galactosidase from pig brain was purified on N-carboxynonyl-1deoxynojirimycin 4 [53]. For the first time, the N-carboxypentyl derivative (14) of 2,5-dideoxy-2,5-imino-D-mannitol was employed for the purification of invertase from yeast [54]. The decrease of inhibitory power upon N-alkylation was over 10-fold (^i B 3.5 μΜ for the parent compound, K^ 50 μΜ for the N-modified derivative). A satisfying purification factor of 250-fold was obtained and the enzyme could be recovered in 85 % yield.

214

10 Iminoalditols as Affinity

10.7

Ligands...

Conclusions

Iminoalditols are highly convenient tools for the isolation and purification of glycosidases by affinity chromatography methods. Depending on the class of enzymes addressed, a wide variety of structural types of inhibitors has become available. The purification of several interesting glycosidases, for example xylosidases and o-galactosidases, have still to be probed with this methodology and a range of potentially useful iminosugars remains to be tested as ligands. In some cases reported, such as with α-L-fucosidases and /?-D-hexosaminidases which are highly susceptible to inhibition by the class of sugar mimics under consideration, the substrate-analogous iminoalditols have frequently been found to adsorb to these enzymes too strongly in terms of their recovery from the affinity matrix. Clearly, this finding calls for the design and synthesis of less active, but highly selective derivatives and analogs so that this technique of enzyme purification may be made generally available for a broad range of applications in biochemistry and medicine, as well as in food and other biotechnologies.

Acknowledgments A. E. S. appreciates continuous support by the Austrian Fonds zur Forderung der Wissenschaftlichen Forschung as well as by the Jubilaumsfonds der Osterreichischen Nationalbank.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

L. Stryer, Biochemistry, 3rd ed, W. H. Freeman and Company, New York, 1985. J. H. Pazur, Adv. Carbohydr. Chem Biochem., 1981, 39, 405-447. E. Steers, Jr., P. Cuatrecasas, H. B. Pollard, J. Biol Chem., 1971, 246, 196-200. E. Junowicz, J. E. Paris, Biochim. Biophys. Acta, 1973, 321, 234-245. C. A. Mapes, C. C. Sweeley, /. Biol. Chem., 1973, 248, 2461-2470. R. Lotan, A. E. S. Gussin, H. Lis, N. Sharon, Biochem. Biophys. Res. Commun. 1973, 52, 656-662. J. H. Sharper, R. Barker, R. L. Hill, Anal. Biochem., 1973, 53, 564-570. J. I. Rood, R. G. Wilkinson, Biochim. Biophys. Acta, 1974, 334, 168-178. N. Harpaz, H. M. Flowers, N. Sharon, Biochim. Biophys. Acta, 1974, 341, 213-221. E. E. Grebner, I. Parikh, Biochim. Biophys. Acta, 1974, 350, 437-441. T. Mega, Y. Matsushima, /. Biochem., 1976, 79, 185-194. T. Mega, Y. Matsushima, /. Biochem., 1977, 81, 571-578. H. Kaneki, M. Tanaka, Chem. Pharm. Bull, 1982, 30, 1753-1759. R. H. Sha, O. P. Bahl, Carbohydr. Res., 1974, 32, 15-24. M. L. Chawla, O. P. Bahl, Carbohydr. Res., 1974, 32, 25-29.

References

215

[16] C. S. Jones, R. H. Shah, D. J. Kosman, O. P. Bahl, Carbohydr. Res., 1974, 36, 241-245. [17] H. Hettkamp, G. Legler, E. Bause, Eur. J. Biochem., 1984, 142, 85-89. [18] J. Schweden, C. Borgmann, G. Legler, E. Bause, Arch. Biochem. Biophys., 1986, 248, 335-340. [19] E. Bause, R. Erkens, J. Schweden, L. Jaenicke, FEBS Lett., 1986, 206, 208-212. [20] E. Bause, J. Schweden, A. Gross, B. Orthen, Eur. J. Biochem., 1989, 183, 661-669. [21] G. Legler, H. Liedtke, Biol Chem. Hoppe-Seyler, 1985, 366, 1113-1122. [22] K. M. Osiecki-Newman, D. Fabbro, T. Dinur, S. Boas, S. Gatt, G. Legler, R. J. Desnick, G. A. Grabowski, Enzyme, 1986, 35, 147-153. [23] G. Legler, E. Bieberich, Arch. Biochem. Biophys., 1988, 260, 427-436. [24] H. Matern, H. Heinemann, G. Legler, S. Matern, J. Biol. Chem., 1997, 272, 1126111267. [25] J. Schweden, G. Legler, E. Bause, Eur. J. Biochem., 1986, 757, 563-570. [26] J. Schweden, E. Bause, Biochem. J., 1989, 264, 347-355. [27] S. Hiraizumi, U. Spohr, R. G. Spiro, /. Biol. Chem., 1994, 269, 4697-4700. [28] W. A. Lubas, R. G. Spiro, J. Biol. Chem., 1987, 262, 3775-3781. [29] N. Hamagashira, H. Oku, T. Mega, S. Hase, J. Biochem., 1996, 779, 998-1003. [30] M. E. Rafestin, A. Obrenovitch, A. Oblin, M. Monsigny, FEBS Lett., 1974, 4O9 62-66. [31] G. Dawson, R. L. Propper, A. Dorfman, Biochem. Biophys. Res Commun., 1973, 54, 1107-1110. [32] E. E. Grebner, L Paritek, Biochim. Biophys. Acta, 1974, 350, 437-441. [33] B. Geiger, Y. Ben-Yoseph, R. Amon, FEBS Lett., 1974, 45, 276-281. [34] M. Pokorny, C. P. J. Glaudemans, FEBS Lett., 1975, 50, 66-69. [35] D. V. Marinkovic, J. N. Marinkovic, Biochem. Med., 1978, 20, 422-433. [36] G. Legler, E. Liillau, E. Kappes, F. Kastenholz, Biochim. Biophys. Acta, 1991, 7080, 89-95. [37] H. M. Flowers, Adv. Carbohydr. Chem. Biochem., 1981, 39, 279-345. [38] C. H. Bauer, P. Vischer, H.-J. Grunholz, W. Reutter, Cancer Res., 1977, 37, 1513-1518. [39] J. A. Alhadeff, A. L. Miller, H. Wenaas, T. Vedvick, J. S. O'Brien, /. Biol. Chem., 1975, 250,7106-7113. [40] D. J. Opheim, O. Touster, J. Biol. Chem., 1977, 252, 739-743. [41] J. A. Alhadeff, A. J. Janowsky, /. Neurochem., 1977, 28, 423-427. [42] R. S. Jain, R. L. Binder, A. L. Benshimol, C. A. Buck, L. Warren, /. Chromatogr., 1977, 739, 283-290. [43] T. Alam, A. S. Balasubramanian, Biochim. Biophys. Acta, 1979, 566, 327-334. [44] H. Yoshima, S. Takasaki, S. Ito-Mega, A. Kobata, Arch. Biochem. Biophys., 1979, 194, 394-398. [45] M. J. Imber, L. R. Glasgow, S. V. Pizzo, J. Biol. Chem., 1982, 257, 8205-8210. [46] G. W. J. Fleet, A. N. Shaw, S. V. Evans, L. E. Fellows, J. Chem. Soc., Chem. Commun., 1985, 841-842; G. W. J. Fleet, N. G. Ramsden, R. A. Dwek, T. W. Rademacher, L. E. Fellows, R. J. Nash, D. S. C. Green, B. Winchester, J. Chem. Soc., Chem. Commun., 1988, 483-485. [47] H. Paulsen, M. Matzke, Liebigs Ann. Chem., 1988, 1121-1126. [48] B. M. Pinto, D. R. Bundle, Carbohydr. Res., 1983, 724, 313-318 [49] H. Paulsen, M. Matzke, B. Orthen, R. Nuck, W. Reutter, Liebigs Ann. Chem., 1990, 953-963. [50] P. Scudder, D. C. A. Neville, T. D. Butters, G. W. J. Fleet, R. A. Dwek, T. W. Rademacher, G. S. Jacob, /. Biol. Chem., 1990, 265, 16472-16477. [51] S. C. T. Svensson, J. Thiem, Carbohydr. Res., 1990, 200, 391-402. [52] G. Legler, A. E. Stiitz, H. Immich, Carbohydr. Res., 1995, 272, 17-30. [53] S. Pohl, Ph. D. Thesis, University of Cologne, 1988; G. Scheffler, Thesis for the Diploma, University of Cologne, 1990. [54] G. Legler, A. Korth, A. Berger, C. Ekhart, G. Gradnig, A. E. Stutz, Carbohydr. Res., 1993, 250, 67-77.

11 Inhibitors of Glycoprotein Processing ALAN D. ELBEIN and RUSSELL J. MOLYNEUX

11.1

Introduction

Almost all polyhydroxy alkaloids with glycosidase inhibitory properties have been isolated and identified within the past two decades. The discovery of the indolizidine alkaloids swainsonine [1] and castanospermine [2], with their potent and specific inhibitory activities towards a-mannosidase and a- and/?-glucosidase, respectively, created a recognition that additional nitrogen-containing analogs of simple sugars might have similar properties and stimulated the search for new members of the class. As a result, more than 50 naturally-occurring members of the group have been discovered, almost doubling the number discussed in a previous review [3]. Many synthetic analogs have been prepared but this chapter will be restricted to those alkaloids isolated from natural sources, their glycosidase-inhibitory properties and consequent effects on glycoprotein processing.

11.2

Chemistry of Alkaloid Glycosidase Inhibitors

11.2.1 Structural classes The alkaloid glycosidase inhibitors discovered to date do not conform to a single structural class but do have several features in common, including two or more hydroxyl groups and a nitrogen atom, generally heterocyclic in character. A small group of glycosidase inhibitors isolated from micro-organisms also exists, which are structurally more closely related to aminosugars. However, it is possible to integrate the major class of heterocyclic compounds into structural groups based upon 5- and 6-membered rings, which may also be fused into bicyclic ring systems. Five different sub-classes can be defined, from the simple monocyclic examples to the more complex bicyclic rings, as follows: pyrrolidines, piperidines, pyrrolizidines, indolizidines, and nortropanes. Inrinosiigars as Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

77.2 Chemistry of Alkaloid Glycosidase Inhibitors

217

Alkaloids of the pyrrolidine class, with 5-membered rings, are exemplified by 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidine (DMDP) (1), which is fully (tetra-) substituted at all carbon atom ring positions [4]. Alkaloids with 6-membered rings of the piperidine class consist of nine members, of which 1-deoxynojirimycin (DNJ) (2) [5] is typical. The pyrrolizidine alkaloids that are inhibitors of glycosidases may be regarded in a formal structural sense as the result of fusion of two pyrrolidine ring systems, with the common nitrogen atom at the bridgehead. The tetra-substituted pyrrolizidines, australine (3) [6] and its epimers, are characteristic of this class. The indolizidine group may be visualized as a pyrrolidine ring fused with a piperidine ring, yielding a bicyclic 5/6 ring system, in an analogous manner to the pyrrolizidine alkaloids. Seven naturally-occurring members have been discovered, with the trihydroxylated alkaloid swainsonine (4) [1,7-9], and the tetrahydroxy alkaloid, castanospermine (5) being the most familiar. Several epimers of castanospermine have been isolated and identified. The polyhydroxy pyrrolidine, piperidine, pyrrolizidine and indolizidine groups have been known for some time but the nortropane group is a relatively new addi-

HO

ΛΧ

N^

(1)

(2)

ΊΐιΙΙΟΗ

(4)

HO,

HOT

(5)

^CH2OH

218

11 Inhibitors of Glycoprotein Processing

tion to the classes of alkaloids with glycosidase-inhibitory properties. The nortropane ring system can be visualized as a result of the fusion of a five-membered pyrrolidine ring with a six-membered piperidine ring, but in contrast to the indolizidines the fusion points are a- to the nitrogen atom of each monocyclic system. In spite of the recentness of its discovery the polyhydroxy nortropane group now consists of more individual alkaloids than any of the other classes, and the chemistry of these compounds has been the subject of a recent review [1O]. The alkaloids have been named calystegines after the source of the first member to be isolated, the bindweed Calystegia sepium [11, 12]. A consistent feature of all calystegines, in addition to the absence of N-methylation, is the presence of an α-OH group at the bridgehead juncture of the bicyclic ring system, i.e. an aminoketal functionality. Three subclasses have been defined, namely calystegines A, B, and C, each of which corresponds to tri-, tetra- and penta-hydroxylation, respectively. The tetrahydroxylated group is typified by the most widespread alkaloid, calystegine 62 (6) [12]. A few nitrogen-containing glycosidase inhibitors, although they are polyhydroxylated, do not fall readily within the above structural classifications. These include the amino-cyclopentanes, mannostatin A [13] and the much more complex glycosylated cyclic urea derivative, trehazolin [14], as well as kifunensine [15] and nagstatin [16], which may be regarded as highly modified piperidines. All of these compounds are metabolites isolated from various micro-organisms.

11.2.2

Occurrence and isolation from natural sources

11.2.2.1

Occurrence

The polyhydroxy alkaloid glycosidase inhibitors have been isolated primarily from plant sources, but also occur in micro-organisms and have occasionally been found in insects. Many of the earliest polyhydroxy alkaloids to be discovered, particularly the bicyclic pyrrolizidines and indolizidines, were found in the plant family Leguminosae. This apparent taxonomic relationship has now become far less secure with the isolation of the pyrrolizidine alkaloid, casuarine, from the Casuarinaceae and Myrtaceae [17], moreover, swainsonine (4) has recently been identified as a constituent of several Ipomoea species (Convolvulaceae), co-occurring with calystegines [18]. Similarly, the initial isolation of calystegines from the Convolvulaceae [11, 12, 19] has now been overshadowed by a much more widespread occurrence in the Solanaceae [20-26] and a limited presence in Moms species (Moraceae) [27, 28]. Certain individual alkaloids, predominantly DMDP and swainsonine, have a particularly widespread pattern of occurrence. Thus, DMDP (1) has been isolated from plants in the families Araceae, Campanulaceae, Euphorbiaceae, Hyacinthaceae, and Leguminosae [4, 29, 30, 31], as well as from the body of a lepidopteran (Urania fulgens) [30], and from a Streptomyces species [32]. Similarly, swainsonine (4) has also been discovered in two unrelated micro-organisms, Rhizoctonia

77.2 Chemistry of Alkaloid Glycosidase Inhibitors

219

leguminicola and Metarhizium anisopliae [8, 9], in addition to its quite widespread occurrence in plants [33]. It has been shown that the biosynthetic pathways to swainsonine in the Diablo locoweed, Astragalus oxyphysus, and in R. leguminicola are identical, implying either a direct or indirect relationship between plant and micro-organism [34]. Thus, the genetic ability to produce this alkaloid could have been transferred from one to the other in the course of evolution. Alternatively, micro-organisms capable of producing the alkaloid may have an endophytic association with the plants. The presence of a calystegine-catabolizing Rhizobium meliloti strain in roots of Calystegia sepium, but not within plants that do not produce calystegines, emphasizes the complexity of such interactions [35]. In contrast to the previous examples, castanospermine (5) and its epimers [2, 36-38] and the australine (3) class of alkaloids have so far been restricted to the monotypic Castanospermum australe and species of Alexa, which are closely related genera in the Leguminosae. It is apparent from these examples that no consistent conclusions can be drawn regarding the distribution of polyhydroxy alkaloids at the present time. It may be that these natural products are quite widely distributed, and many new sources will be discovered in the future. The comparative recency of their discovery relative to many other classes of alkaloids is probably a consequence of their cryptic nature, due to exceptional water solubility and relative insolubility in non-hydroxylic organic solvents [39]. The increasing number, regio- and stereochemical potential for structural variation, and significant biological properties of these glycosidase inhibitors will no doubt result in discovery of new members of the known classes. The recent identification of the nortropane group is also an indicator that new structural groups may yet remain to be discovered. 11.2.2.2 Isolation The hydrophilicity of the polyhydroxy alkaloids renders them incapable of being isolated by conventional extraction and purification methods which involve extraction into non-polar organic solvents and partitioning between aqueous acid and base. Ion-exchange chromatography is therefore generally employed for purification, following extraction from the natural source by water, methanol or ethanol, either alone or in various mixtures. Subsequent separation can be achieved by paper, column, or thin-layer chromatography. The alkaloids are particularly amenable to detection by thin-layer chromatography in association with specific spray reagents, gas chromatography with flame ionization or mass spectrometric detection, and by their glycosidase inhibitory properties. All of these techniques have recently been reviewed in detail [4O]. Structural determination places a particular reliance on nuclear magnetic resonance spectroscopy which generally permits establishment of the specific ring system present, the substitution pattern and relative stereochemistry of the hydroxyl groups. Mass spectrometry provides similar information, with the exception of stereochemistry. The isolation of increasing numbers of these alkaloids has furnished a spectroscopic database which renders the determination of structures increasingly facile. Determination of the absolute stereochemistry is dependent

220

11 Inhibitors of Glycoprotein Processing

upon X-ray crystallography, which can be used whenever well-refined crystal data can be obtained, either from the alkaloid itself or a crystalline derivative such as the hydrochloride salt. Alternatively, circular dichroism techniques may be applied, especially the benzoate chirality method [41]. Although this technique may have the most general utility, being independent of the physical state of the alkaloid, it has so far had only very limited application.

11.3 11.3.1

Glycosidase Inhibition Glycosidase inhibitory activity

The inhibitory activity of individual alkaloids may be remarkably specific, as with swainsonine (4), which inhibits only α-mannosidase and Golgi mannosidase II, or can be more general, showing a spectrum of activity against a series of glycosidases. Additionally, the potency may vary with the source of a particular enzyme, its purity, and the conditions, such as pH, under which the assay is performed. For these reasons the inhibitory properties of individual alkaloids are presented here only in a summary form (Table 11-1 and references 42-78). The inhibition of Nlinked glycoprotein processing by the most potent and specific of the alkaloids are discussed in detail in subsequent sections. Table 11-1.

Enzyme Inhibitory Activities of Polyhydroxy Alkaloids.

Alkaloid

Enzyme

Reference

Pyrrolidines CYB-3

α-glucosidase (weak)

[42]

6-Deoxy-DMDP

^-mannosidase

[6]

D-ABl

a-glucosidase a-D-arabinosidase

[42,43] [44]

1,4-Dideoxy-1,4-imino-D-ribitol

α-glucosidase (weak)

[45]

Nectrisine

a-glucosidase a-mannosidase

[46] [47]

W-Hydroxyethyl-2-hydroxymethyl3 -hy droxy pyrrolidine

undetermined

DMDP (1)

a- and /?-glucosidase /?-mannosidase invertase trehalase

[48-50] [51] [52, 53]

HomoDMDP

a- and /0-glucosidase

[48]

11.3 Glycosidase Inhibition Table 11-1.

221

(Continued).

Alkaloid

Enzyme

Reference

Piperidines 6-Deoxyfagomine

undetermined

Fagomine

/?-galactosidase α-glucosidase (weak)

[54] [42]

3-£/?/fagomine

/?-galactosidase

[54]

1-Deoxynojirimycin (DNJ) (1)

a- and /?-glucosidase invertase trehalase

[55]

W-Methyl-DNJ

a-glucosidase

[56]

1-Deoxymannojirimycin (DMJ)

a-mannosidase a-fucosidase

[57, 58] [57]

Nojirimycin

a- and /?-glucosidase

[59]

Mannojirimycin

a-mannosidase

[60]

Galactostatin

/^-galactosidase

[61]

a-Homonojirimycin

a-glucosidase

[42, 49, 50]

Australine (3)

amyloglucosidase

[6, 62]

Alexine

amyloglucosidase trehalase

[63]

1 -£/7/australine

amyloglucosidase a-glucosidase

[63]

3-E/?/australine

amyloglucosidase

[64]

7-Ep/australine

amyloglucosidase a-glucosidase

[63]

7a-E/?/alexaflorine

amyloglucosidase

[64]

Casuarine

undetermined

Pyrrolizidines

Indolizidines Lentiginosine

amyloglucosidase

[66]

2-E/?zlentiginosine

none

[66]

Swainsonine (4)

a-mannosidase

[67]

7-Deoxy-6-£p/castanospermine

amyloglucosidase

[36]

Castanospermine (5)

a- and /?-glucosidase

[68]

6-£/?/castanospermine

amyloglucosidase

[69]

6,7-Di£/?/castanospermine

amyloglucosidase /?-glucosidase

[78]

222

11 Inhibitors of Glycoprotein Processing

Table 11-1.

(Continued). Enzyme

Reference

Calystegine A3

/?-glucosidase trehalase

[19, 70]

Calystegine AS

none

[70]

Calystegine Ae

undetermined

Calystegine A7

trehalase

[20]

Calystegine BI

/?-galactosidase /?-glucosidase

[19, 70]

Calystegine B2 (6)

a-galactosidase /?-glucosidase trehalase

[71] [19, 70] [20]

Calystegine B3

/?-glucosidase (weak) trehalase

[20]

Calystegine 64

/?-glucosidase trehalase

[71]

Calystegine B5

undetermined

Af-Methylcalystegine B2

a-galactosidase trehalase

[20, 72]

Calystegine Ci

a-galactosidase /?-galactosidase /?-glucosidase trehalase

Calystegine C2

a-mannosidase

[20] [19, 20, 73] [19, 10, 73] [71] [74]

Af-Methylcalystegine Ci

a-galactosidase

[71]

Mannostatin A

a-mannosidase

[75]

Trehazolin

trehalase

[76]

Kifunensine

a-mannosidase

[77]

Nag statin

/?-Af-acetyl-glucosaminidase

[78]

Alkaloid norTropanes

Miscellaneous

11.3.2

Structure-activity relationships

Early approaches to correlation of structure of the polyhydroxy alkaloids with their glycosidase inhibitory properties appeared to indicate a rather straightforward relationship [3]. Swainsonine (4) was perceived as an aza-analog of D-mannopyranose, lacking the hydroxymethine group at C-4 but otherwise having the same relative

77.3 Glycosidase Inhibition

223

disposition of the remaining hydroxyl groups, which therefore accounted for its ability to inhibit a-mannosidase [67]. The structures of 1-deoxynojirimycin (1) and castanospermine (5) correlated even more closely, as monocyclic and bicyclic 'aza sugars', with that of glucose, and they inhibited glucosidases as expected. This naive approach had to be reconsidered with the isolation of 6 -ep/castanospermine which, in spite of its stereochemical similarity to mannose, failed to inhibit either a- or /?-mannosidase but instead proved to be an effective inhibitor of a-glucosidase, with a level of activity only slightly less than that of castanospermine [37]. Numerous additional examples of inhibitory specificities due to both naturally-occurring alkaloids and synthetic analogs have further undermined this empirical approach and it is obvious that structure-activity correlations can only be developed with the aid of sophisticated molecular modeling techniques. Molecular orbital calculations and molecular modeling have been applied to a series of known mannosidase inhibitors, and others which were expected to inhibit but failed to do so. The results showed that good inhibitors fit closely with a single low-energy conformer of the mannosyl cation and demonstrated that 6-^picastanospermine did not comply with the structural requirements [79, 8O]. The electronegative binding groups present in the inhibitor necessary for specificity and activity were established, as were those which were of little significance. Additional studies of this type should provide valuable information regarding the binding sites on the various enzymes, but the inhibition data available is compromised by the variability in enzymes and the conditions under which measurements have been made. A comprehensive screening program using standardized conditions would provide much more useful information for structure-activity correlations and consequently the design of specific and potent inhibitors. In the absence of comprehensive molecular modeling studies, the inhibition results obtained have been rationalized on the basis of generally accepted models for glycosidase inhibition. This approach has been developed most effectively for the calystegines, which provide a comprehensive series of structurally related natural polyhydroxy alkaloids. For /?-glucosidase inhibition, the model invokes the presence of two carboxylic acid groups at the active site of the enzyme, one responsible for generation, and the other for stabilization, of the glycosyl cation intermediate [21]. The essential requirement of equatorial hydroxyl groups at the 2- and 3 -positions is in accord with earlier studies of interaction of other inhibitors with /?-glucosidase. Thus the interaction of inhibitory calystegines with glycosidases can be envisioned as binding to the sites determining specificity and to the catalytic center, through specific hydroxyl groups and through the imino group. The mechanism of galactosidase inhibitory activity is less apparent. Calystegines BI, B2 and Ci are potent inhibitors of either a- or /?-galactosidase, yet calystegine BS, with a much closer configurational similarity to D-galactose than any of the former, has no inhibitory activity against these enzymes, an observation which is reminiscent of the situation with 6-£/?icastanospermine in the indolizidine alkaloid series. Obviously, a much larger set of natural or synthetic epimers, enantiomers and structural analogs is needed before a complete understanding of structure-activity relationships can be applied to prediction of inhibitory activity. Some progress in this direction has been made through a comparison of glycosi-

224

11 Inhibitors of Glycoprotein Processing

dase inhibition by synthetic analogs and derivatives of (+)-calystegine 62. The non-natural (-)-enantiomer showed no glycosidase inhibitory properties, whereas N-methylation of natural 62 suppressed inhibition of /?-glucosidase while activity towards α-galactosidase was retained [35].

11.3.3

Synthetic polyhydroxy alkaloids

In addition to the synthesis of known naturally-occurring alkaloids for the purpose of structural confirmation, many epimers, enantiomers and structural analogs have been prepared. The number of these synthetic alkaloids, particularly those related to swainsonine, castanospermine and australine, now approaches or perhaps exceeds those isolated from natural sources. It is probable that at least some of the synthetic compounds, especially epimers of known naturally occurring alkaloids, will subsequently be found to occur in nature. In addition, new structural classes have already been generated which might reasonably be expected to be biosynthesized by plants. Noteworthy among these are polyhydroxy quinolizidine alkaloids, consisting of two six-membered rings fused into a bicyclic system, which are ring-expanded homologs of the indolizidine alkaloids [81, 82], Although quinolizidine alkaloids are a well-established class of natural products, none have yet been isolated which bear more than two hydroxyl groups. This is probably a consequence of the high water solubility of polyhydroxylated derivatives which renders them unextractable into the non-hydroxylic solvents normally used for alkaloid purification. The combination of novel natural polyhydroxy alkaloids, together with synthetic analogs tailored to have specific structural features, will ultimately lead to a full comprehension of the interaction of these alkaloids with receptor sites on the enzyme which results in their glycosidase inhibitory properties.

11.4 11.4.1

Biological Activity of Glycosidase Inhibitors Mammalian toxicity

As might be expected from a class of compounds that inhibits glycosidases and consequently the fundamental cellular function of glycoprotein processing, the polyhydroxy alkaloids exhibit an exceptional diversity of biological activities. Discovery and isolation of many of the alkaloids has been a result of observations of the ultimate clinical effects which result from the consumption by animals of plants containing these bioactive compounds. Predominant among such examples is the occurrence of swainsonine (4) in Swainsona species (poison peas) of Australia [1] and Astragalus and Oxytropis species (locoweeds) of North America [33]. The potent α-mannosidase inhibitory activity of swainsonine disrupts glycoprotein

77.4 Biological Activity of Glycosidase Inhibitors

225

processing by mannosidase II in the Golgi, resulting in neuronal vacuolation due to abnormal storage of mannose-rich oligosaccharides, leading to the neurological damage so characteristic of the locoism syndrome. However, the clinical effects are not limited to the nervous system since emaciation, reproductive failure in both males and females, and congestive right-heart failure are also observed. Since the discovery of swainsonine as the causative agent, locoweed poisoning has now been established as a widespread phenomenon, with additional occurrences being reported from South America, and many parts of China and Tibet [83], Recently, swainsonine has been reported to co-occur with calystegines BI (6) and Ci in Ipomoea species of Australia which cause poisoning of sheep and cattle [18], and in 7. carnea, resulting in toxicity to goats in Mozambique. The clinical signs of poisoning are characterized by the expected neurological damage resulting from swainsonine ingestion but these are exacerbated by muscle-twitching, tremors and epileptiform seizures. Histological examination of tissues showed vacuolation of Purkinje cells in addition to swainsonine-induced cytoplasmic vacuolation of neurons and axonal dystrophy. The calystegines inhibit /?-glucosidase and α-galactosidase which would produce phenocopies of the genetic lysosomal storage defects Gaucher's disease and Fabry's disease, respectively, and the additional syndromes are significant indicators of the latter. In contrast to the above examples which exhibit a complexity of effects, the alkaloids concentrated in the chestnut-like seeds of Castanospermum australe (Black Bean), primarily castanospermine (5) and australine (3), together with several less potent epimers of both, produce gastrointestinal disturbances in livestock and humans but no discernable neurological damage [84]. This is consistent with the ability of the alkaloids to inhibit a- and /?-glucosidase, resulting in a syndrome phenotypic of the genetic defect, Pompe's disease. Although this relationship has not been directly established in field cases of poisoning, rodent feeding experiments with castanospermine resulted in vacuolation of hepatocytes and skeletal myocytes, and glycogen accumulation, consistent with Pompe's disease or type II glycogenesis [85]. Gastrointestinal problems and lethargy have also been observed in livestock grazing bluebells (Hyacinthoides non-scripta) in the U.K., and the recent demonstration of the presence of DMDP (1) and homoDMDP in this plant may account for the syndrome [48]. All of the above poisoning syndromes are relatively obvious once signs develop, although this may take several weeks of consumption of the plant because the alkaloids implicated often are present at very low levels. Nevertheless, they are potent inhibitors and it has been estimated that a swainsonine content of 0.001 % of the dry weight of the plant may be sufficient to induce locoism [83]. For those alkaloids which are less active or which are present at extremely low levels, it seems probable that the signs of poisoning would be subclinical, with no overt changes being apparent. In such cases, toxicity may only be manifested as minor digestive disturbances, failure to gain weight and other deviations from optimal health which could be attributed to stress or infectious diseases. The occurrence of various calystegines in human food plants from the family Solanaceae, such as potatoes, eggplant and peppers, could account for a variety of complaints, primarily gastrointestinal, reported in certain individuals consuming these vegetables [86].

226

11.4.2

11 Inhibitors of Glycoprotein Processing

Therapeutic activity

The capability of polyhydroxy alkaloids to disrupt the general cellular function of glycoprotein processing leads to the expectation that these compounds should have therapeutic potential for the treatment of various disease states. The significant mammalian toxicity of certain of the alkaloids is an obvious hindrance to their utility. However, this is frequently true of many drug candidates and it is not unreasonable to assume that an appropriate dose-response relationship could be achieved. Moreover, adverse effects, such as the neurological damage caused by swainsonine, often develop quite slowly and appear to be reversible if ingestion of the alkaloid is terminated, as would be the situation with most drug regimens. Investigation of the alkaloids for therapeutic potential has so far concentrated on three major disease states, i.e. for treatment of cancer and inhibition of metastasis, as anti-diabetic drugs, and for anti-viral activity. Swainsonine (4) has received particular attention as an anti-metastatic agent. In vivo experiments with mice have shown that pulmonary colonization is reduced by over 80 % if the animals are provided with drinking water containing 3 μg/mL of swainsonine for 24 hours prior to injection with B16-F10 murine melanoma cells [87]. This effect has been shown to be due to enhancement of natural killer Tcells and increased susceptibility of cancerous cells to their effect [88]. The pharmacokinetics of swainsonine in such experiments indicate that the levels of alkaloid and period of administration would not be sufficient to produce neurological damage [89]. It has been suggested that post-operative metastasis of tumor cells in humans could be suppressed by intravenous administration of the alkaloid prior to and following the surgery. Clinical trials in humans with very advanced malignancies showed that lysosomal α-mannosidases and Golgi mannosidase II were inhibited and some improvement in clinical status occurred [9O]. Castanospermine has also been reported to suppress metastasis in mice [91], but experiments with this alkaloid have not been as extensive as those with swainsonine. Castanospermine (5) and 1-deoxynojirimycin (2) have been shown to be capable of suppressing the infectivity of a number of retro viruses, including the human immunodeficiency virus (HIV) responsible for AIDS [92-95]. This effect is a consequence of inhibition of glycoprotein processing which results in changes in the structure of the glycoprotein coat of the virus. Cellular recognition of the host is thus prevented and syncytium formation is suppressed. In spite of this significant effect, both of these alkaloids suffer from the disadvantage that they are highly water-soluble and therefore excreted very rapidly. This defect has been overcome by derivatization to give 6-O-butyryl-castanospermine and N-butyl-deoxynojirimycin [96, 97], and both of these compounds have undergone clinical trials against AIDS in humans, either alone or in combination with AZT As might be expected, gastrointestinal disturbances have been reported as a significant side-effect. Another structural modification of 1-deoxynojirimycin, the N-hydroxyethyl derivative, miglitol, an inhibitor of α-glucosidase, has been clinically evaluated and released as an antidiabetic drug in insulin- and non insulin-dependent diabetes. The alkaloid was shown to potently inhibit glucose-induced insulin release and

77.5 Processing ofN-linked Oligosaccharides

227

also suppressed islet α-glucoside hydrolase activity, thus controlling postprandial glycemia [98]. The structurally related alkaloids, 2-0-a-D-galactopyranosyl-DNJ and fagomine, have also been shown to have antihypoglycemic activity in streptozocin-induced diabetic mice but have not been tested in humans [99]. The ability of polyhydroxy alkaloid glycosidase inhibitors to prevent cellular recognition has resulted in their evaluation for clinical situations where suppression of an immune response would be desirable, or for use against parasitic diseases. Thus, in vivo experiments have shown that castanospermine can be used as an immunosuppressive drug, promoting heart and renal allograft survival in rats [10O]. Parasitic diseases may also be controlled by altering cellular recognition processes. Castanospermine provides protection against cerebral malaria by preventing adhesion of Plasmodium falciparum to infected erythrocytes [101], while swainsonine inhibits the association of Trypanosoma cruzi, the causative agent of Chagas' disease, with host cells by formation of defective mannose-rich oligosaccharides on the cell surface [102]. There is no doubt that the polyhydroxy alkaloids have considerable potential for treatment of a variety of disease states in humans and animals. The primary challenge in introducing them as commercial drugs is to minimize their toxicity and enhance the specificity of their beneficial effects. Improvement of their pharmacokinetic properties should result in much lower dose rates being necessary so that undesirable side-effects are limited. Increased specificity of action can be achieved by preparation of synthetic derivatives and a comprehensive understanding of structure-activity relationships.

11.5 11.5.1

Processing of AMinked Oligosaccharides Introduction

Glycoproteins are widespread in nature, being found in all eucaryotic cells [103]. Recently, they have also been shown to be present in various archaebacteria as well as in some lower bacteria [104, 105]. In addition, it has become quite clear that specific carbohydrate structures on glycoproteins, glycolipids and proteoglycans are critically important as ligands in molecular recognition [106]. At least with regard to the N-linked glycoproteins, on which this review is focused, these molecules have been implicated in a number of important physiological functions, especially cell .cell recognition reactions involving such critical phenomenon as inflammation [116], pathogenesis [107], parasitism [108], development [109], cell adhesion [110], and symbiosis [111], to mention only a few. N-linked oligosaccharides are also key players in the targeting of lysosomal enzymes [112], in the uptake or removal of glycoproteins from the blood [113], in protein folding in the endoplasmic reticulum [114], and in many other physiological phenomena of potential significance [115, 116]. Although the carbohydrate

228

11 Inhibitors of Glycoprotein Processing

portion of the glycoprotein has not been shown to participate in every case of recognition, specific oligosaccharide structures are clearly central to many of these cases. Thus, inhibitors that block specific steps in the assembly of the various N-linked oligosaccharides and cause the formation of altered or immature oligosaccharide structures, should be valuable tools for examining the role of carbohydrates in glycoprotein function [117]. Fig. 11-1 shows three representative structures of the N-linked oligosaccharides. All of these oligosaccharides have the same core structure shown within the box, and are composed of a branched trimannose structure linked to a chitobiose disaccharide. The immature or initially synthesized oligosaccharide is the high mannose structure shown in (A), and this oligosaccharide is the biosynthetic precursor that gives rise to all of the other N-linked oligosaccharides. However high-mannose oligosaccharides are commonly found as components of glycoproteins from lower eucaryotes such as fungi and yeast. In additon, a small percentage of the N-linked oligosaccharides of animal cell surface proteins are of the high-mannose type. The middle structure (B) in Figure 11-1 is an example of one type of complex oligosaccharide that is frequently found in cell surface glycoproteins of higher eucaryotes, such as the low density lipoprotein receptor and many other membrane receptors. This particular structure is referred to as a biantennary complex chain,

Manal —-2Manal

X

6 Manal 3

R

Manal— 2Mana1x/

Manp1— 4GIcNACp 1-4GIcNAc- •As

3

Manal— 2Mana1— 2Manal/

A. High-Mannose Type

SA—Gal— GIcNAo- Manal R

Manp1- 4GIcNAcP 1-4GIcNAc- •As η

3

SA- GaI- GIcNAc- Manar^

B. Complex Type

Manal X

6 Manal

Manal/

6 3Μ3ηβ1- 4GlcNAcp1— 4GIcNAc- -Asn

SA- Gal— GIcNAc- Manal/

Figure 11-1. Structural Classes of N-Linked C. Hybrid Type

Oligosaccharides.

77.5 Processing of N-linked Oligosaccharides

229

but other complex oligosaccharides may have three of the sialic acid-galactoseGIcNAc chains (triantennary chains), or four of these trisaccharide sequences (tetraantennary chains). The lower structure of (C) Figure 11-1 is a hybrid type of oligosaccharide that is produced by partial processing to the GIcNAc transferase I step, and then addition of various sugars to the 3-linked mannose branch. That is, hybrid structures are apparently the result of an absence of mannosidase II action, or activity. It is not clear whether hybrid structures are formed normally, but they are found in glycoproteins produced in individuals with HEMPAS disease, a condition where individuals lack mannosidase II activity. Hybrid structures can also be induced by treating cultured cells with swainsonine (4).

11.5.2

Biosynthesis of TV-linked oligosaccharides

The biosynthesis of the N-linked oligosaccharide chains involves two rather distinct series of reactions. The first of these pathways gives rise to the precursor, or immature, oligosaccharide which is then transferred cotranslationally to the protein chain, while the protein is being synthesized on membrane-bound polysomes [118], In contrast, the second series of reactions involves the modification of the oligosaccharide chain by the removal of some sugars and the addition of others, to give a large number of different oligosaccharide structures [119]. This first pathway requires the participation of a lipid carrier and the involvement of lipid-linked saccharide intermediates. The reactions leading to the production of the final lipidlinked oligosaccharide precursor are presented in Figure 11-2. -GDP-Man a 1,6 Man Trans.

a 1.3 Man Trans.

Man-GlcNAc-GlcNAc-PP-Dol 4

a 1,2 Man Trans,

Man-Man-Man

^^Man-GlcNAc-GlcNAc-PP-Dol

GDP-Man ^l

\^

Man-P-Dol

GlcNAc-GlcNAc-PP-Dol -GIc-P-DoI

A

Μα κ, ϋ,"ΜαΛ Man-Man-Man

GIcNAc-PP-DoI

^

IVIan-GlcNAc-GlcNAc-PP-Dol Glc-Glc-G!c-Man-Man-Man

\

Dol

x. <

Mevelonic

V

Dol-PP

HMG-CoA

"CORE" Glycoprotein

Figure 11-2.

Biosynthetic Assembly of the Core N-linked Oligosaccharide.

230

11 Inhibitors of Glycoprotein Processing

As shown in this Figure, the assembly of the N-linked oligosaccharide chain is initiated in the endoplasmic reticulum (ER) by the transfer of a GIcNAc-I-P from UDP-GIcNAc to dolichyl-P to form GlcNAc-PP-dolichol [12O]. A second GIcNAc is then added, also from UDP-GIcNAc, to produce GIcNAc^ 1,4GIcNAc-PP-dolichol [121]. Then, five mannose residues are added, all in α-linkages and all from GDP-mannose, to give the important intermediate, Man5(GIcNAc)2-PP-dolichol [122]. These first seven reactions are believed to occur on the cytosolic side of the ER membrane, since they involve nucleoside diphosphate sugars as the activated sugar donors, and these sugar donors are biosynthesized in the cytoplasm by soluble sugar nucleotide pyrophosphorylases. It seems likely, therefore, that the sugar acceptor, dolichyl-P, is initially oriented in the ER membrane in such a way that the phosphate group is oriented towards the cytoplasm, and is therefore able to accept sugars from the cytosol. After the addition of the first 7 sugars to give Mans(GIcNAc)2 -PP-dolichol, this lipid-linked oligosaccharide is believed to undergo a 'flip-flop' in the membrane so that the oligosaccharide chain now becomes inserted into the lumen of the ER [123]. The assembly of the oligosaccharide is completed by the addition of four more mannose residues and then three glucose units to give a GlCsMa^(GIcNAc)2PP-dolichol [124]. These next seven sugars (i.e., 4 mannose and 3 glucose units) are all added in the lumen of the ER, and are donated by the activated lipid precursors, mannosyl-P-dolichol and glucosyl-P-dolichol [125, 126]. These two sugar donors are synthesized from their sugar nucleotides, GDP-mannose and UDP-glucose, by transfer of the specific sugar to dolichyl-P [127]. The reactions for the synthesis of the activated lipid-linked monosaccharides are thought to occur on the cytosolic side of the ER membrane, and are catalyzed by the enzymes, dol-Pman synthase and dol-P-glc synthase [128, 129]. The final step in this pathway is the transfer of the GlCsMa^(GIcNAc)2 from its lipid carrier to specific asparagine residues on the polysome-bound protein, catalyzed by the enzyme oligosaccharyltransferase [130, 131]. The asparagine residue that acts as the acceptor of this oligosaccharide chain must be in the tripeptide consensus sequence, Asn-X-Ser(Thr), where X can be any amino acid except proline, but certain amino acids are favored over others [132]. In addition, the tripeptide sequence must be in a specific conformation, in order to be glycosylated [133]. Although all of the reactions in this pathway are known, it is still not clear how the pathway is regulated, nor which enzymes are under control.

11.5.3

Processing of AMinked oligosaccharides

After the oligosaccharide is transferred to protein, and while the protein chain is still being synthesized in the ER, the oligosaccharide begins to undergo a number of processing or trimming reactions. The initial reactions in this second pathway encompass the removal of 3 glucoses and up to 6 mannoses, but later processing reactions involve the addition of a number of other sugars, principally GIcNAc, galactose, neuraminic acid, L-fucose and possibly GaINAc [134]. The processing pathway is outlined in Figure 11-3.

77.5 Processing of N-linked Oligosaccharides

231

G3M9N2-PP-DoI

M-M M-MM-N-N-Asn G-G-G-M-M-M I GIcI GIcII

M-M-M'

I

I Man I M M '

M"

M-N-N-Asn I GIcNAc Trans. I

Hybrid Chai

"S

Man Il I M-N-N-Asn GIcNAc M' I

Figure 11-3. Processing Pathway of N-Linked Oligosaccharides.

SA-cai-GicNAc-rvi SA-Gai-GicNAc-M

| |n

The first processing step involves a membrane-bound glucosidase, called glucosidase I, which removes the outermost «1,2-linked glucose [135]. This enzyme is quite distinct from the common glycosidases, such as the lysosomal enzymes that are involved in the degradation of polysaccharides, glycolipids and other complex carbohydrates, since those enzymes usually have a pH optimum of around 5, whereas glucosidase I has a pH optimum near neutrality (6.4 to 6.8) [136]. In addition, the common glycosidases are only specific for the sugar to be removed and its anomeric configuration, but do not have strong specificity for the group to which this sugar is attached, nor the specific glycosidic linkage if that group is another sugar. Glucosidase I, on the other hand, will only cleave a non-reducing glucose that is attached in α 1,2-linkage to another glucose [137]. Thus, glucosidase I will not work with p-nitrophenyl-a-D-glucopyranoside. Finally, these different types of glucosidases can be distinguished by their location; the processing glucosidases are in the ER, while the other hydrolytic α-glucosidases are usually in the lysosomes or are secreted by lower eucaryotes. Glucosidase I is the enzyme that initiates the trimming or maturation of the Nlinked oligosaccharide chains and therefore may play a key role in controlling the rate of transport or exit of newly-formed glycoproteins from the ER to the Golgi apparatus. This enzyme has been purified from a number of sources, including calf [138] and porcine [139] liver, and bovine mammary glands [140], as well

232

11 Inhibitors of Glycoprotein Processing

as plants (mung bean seedlings) [137] and yeast (Saccharomyces cerevisiae) [141]. The pig liver glucosidase I was cloned from a human hippocampus cDNA library and expressed in COS 1 cells. The expressed enzyme had a molecular mass of 95 kDa and was degraded by endoglucosaminidase H (Endo H) to a 93 kDa form, indicating that the enzyme has a high-mannose oligosaccharide at the asparagine 655 glycosylation site [142]. The hydrophobicity profile of the enzyme, and the fact that trypsin treatment of microsomes released a 4 kDa fragment, support the view that the glucosidase I is a transmembrane glycoprotein containing a short cytoplasmic domain of about 37 amino acids, followed by a transmembrane domain and a large C-terminal catalytic domain on the luminal side of the ER membrane [142]. A second glucosidase, located in the lumen of the ER and called glucosidase II, removes the other two αϊ,3-linked glucoses to give a Man9(GlcNAc)2-protein. Interestingly, this enzyme removes the outermost a 1,3 -linked glucose quite rapidly (ti/2 = 5 min), whereas removal of the innermost αϊ,3-linked glucose is considerably slower (ti/2 = 20-30 min) [143]. Those earlier observations on the activity of this enzyme correlate well with the recently described role of this enzyme in protein folding. That is, a single α 1,3-linked glucose on the high-mannose chains functions as a recognition site to bind a chaperone to those proteins that are improperly folded or denatured, and that chaperone expedites or assists their proper folding. Thus, it has been shown that the ER contains a protein called calnexin that functions to help newly synthesized membrane proteins fold into their proper conformation, a step that is apparently necessary for many of these proteins to be transported to the Golgi apparatus at the proper rate [144]. Calnexin is a lectin that recognizes a single αϊ,3-linked glucose on the high mannose chains of unfolded or denatured proteins [145]. Since glucosidase II acts fairly slowly on the final α 1,3-linked glucose, there must be a time period when the glycoprotein has only a single glucose on its oligosaccharide. This glucose on the high-mannose chains of unfolded proteins is the recognition site for calnexin to bind to those proteins that have not yet assumed their proper conformation [146-148]. The ER also contains a safety mechanism to assure that unfolded or improperly folded glycoproteins can interact with this chaperone to obtain the conformation that is required for exit from the ER into the Golgi apparatus. In this regard, the ER contains an unusual glycosyltransferase that functions to transfer a glucose from UDP-glucose to high mannose chains on unfolded or denatured, but not on native, glycoproteins [149]. Once this glucose has been added, calnexin can recognize and assist this protein in its proper folding and transfer to the Golgi [15O]. As a result, a glycoprotein that has had all of its three glucoses removed by glucosidase I and II but has failed to fold into the proper conformation can be reglucosylated by this novel enzyme, and this signal then allows the protein another opportunity to interact with calnexin and fold properly. This mechanism, involving the removal of glucoses by the glucosidases and reglucosylation by the glucosyltransferase, is postulated to be part of a unique 'glycoprotein-specific folding and quality control mechanism' in the ER that allows this organelle to control and pass properly folded glycoproteins on to the next step in transport and processing.

77.5 Processing ofN-linked Oligosaccharides

233

Glucosidase II has a fairly high pH optimum of about 6.5 to 7.0, but also hydrolyzes /7-nitrophenyl-a-D-glucoside [151]. On the other hand, the enzyme does appear to be fairly specific for the α 1,3-linked glucose since hydrolysis of Glc2Man9(GlcNAc)2 is inhibited by nigerose, an αϊ,3-linked disaccharide of glucose, but not by the corresponding α 1,2-, α 1,4-, or α 1,6-linked disaccharides of glucose [152]. The enzyme from pig kidney was shown to have a subunit molecular mass of 100 kDa and to contain a high-mannose oligosaccharide [153], while the enzyme from mung bean seedlings had two 110 kDa subunits as well as high-mannose oligosaccharides [154]. On the other hand, in some other animal systems, glucosidase II subunits were reported to have molecular weights of 65 kDa [155, 156]. This enzyme was reported to be located in the rough and smooth ER of pig hepatocytes [157], but has also been localized in post-Golgi structures in tubular cells of pig kidney [158]. The processing glucosidases can best be assayed, in vitro, using the radiolabeled oligosaccharide substrates, [3H]-Glc3Man9GlcNAc and [3H]-Glc2Man9GlcNAc. These substrates are readily prepared in cultured animal cells infected with an enveloped virus, such as influenza virus, that has an N-linked glycoprotein coat. Thus, MDCK cells are infected with influenza virus, and progeny virus are produced in these cells in the presence of a glucosidase or mannosidase processing inhibitor to prevent the removal of those specific sugars [159]. For example, if the virus is grown in the presence of castanospermine (5), the oligosaccharide chains on its envelope glycoproteins will be mostly of the GlcsMan9(GlcNAc)2 structure, whereas if the virus is grown in the presence of deoxymannojirimycin (2, RI = /?-OH) or kifunensine, it would have mostly Man9(GlcNAc)2 structures [16O]. The oligosaccharides are radiolabeled by growing the virus in the presence of either [3H]galactose to label the three glucose residues of the oligosaccharides, or in [2-3H]mannose to label the 9 mannose units. The virus-infected MDCK cells are incubated for 40 hours to allow the virus to replicate and lyse the cells, and the virus particles are isolated from the culture medium by ultracentrifugation. The resulting viral pellet is treated exhaustively with pronase to digest the proteins, and the resulting glycopeptides are isolated by gel filtration and treated with Endo H, i.e., endoglucosaminidase H, to cleave the high-mannose and glucosecontaining high-mannose glycopetides [161]. The resulting oligosaccharides, having a single GIcNAc at the reducing end, are isolated by gel filtration on columns of Biogel P-4 [169], and used as substrates for the processing glycosidaces. Once the two glucosidases have removed all three glucoses from the N-linked oligosaccharide as shown in Fig. 11-3, a number of α-mannosidases can remove one or more of the four α 1,2-linked mannose residues to ultimately give a Mans(GIcN Ac)2 -protein (i.e., Manal,3(Manal,6)Manal,6[Manal,3]Man^l,4Glc/?1,4GIcNAc-protein) [162]. There are believed to be at least three different α 1,2mannosidases involved in the conversion of Man9(GlcNAc)2 to Man5(GlcNAc)2; an ER α-mannosidase, a Golgi Man9-mannosidase and a Golgi mannosidase I [162]. These enzymes differ in a number of properties including their substrate specificity, their sensitivity to various mannosidase inhibitors and their intracellular location. The ER mannosidase presumably removes only a single mannose to generate a unique and specific Mans(GlcNAc)2 structure. This enzyme is reported

234

11 Inhibitors of Glycoprotein Processing

to cleave the «1,2-mannosidic linkage in Man9(GlcNAc)2 that is normally resistant to hydrolysis by the Golgi Man9-mannosidase [163, 164]. However, a soluble form of the ER α-mannosidase has been shown to exhibit rather low specificity, in that it can release several different α 1,2-linked mannose residues from the Ma^GIcNAc substrate. These mannoses are removed in a random fashion so that three different MangGlcNAc structures are produced, as well as a number of ManvGlcNAc isomers [165]. The discrepancy in specificity between the ER mannosidase and the soluble mannosidase reported in these two studies may be due to the effects of the protein itself on substrate specificity, i.e., the ER α-mannosidase may act differently in its specificity on the free oligosaccharide, as compared to the proteinbound oligosaccharide. The Man9-mannosidase, at least the enzyme from pig liver, cleaves both free and peptide-bound Ma^(GIcNAc)2 to give a specific Man6(GlcNAc)2 isomer [165]. Thus, the ER mannosidase and the Man9-mannosidase may be complimentary to each other. Another α 1,2-mannosidase, isolated from rat liver Golgi and requiring Ca++, apparently cleaves each of the four al,2-mannoses in the Man9(GlcNAc)2 at a comparable rate, indicating that it alone could produce the Man5(GlcNAc)2 that is involved in the formation of complex types of oligosaccharides [166, 167]. The exact function of these different α-mannosidases is not currently known. The fact that each of these enzymes removes α 1,2-linkages, and that there is considerable redundancy in their action, indicates that each has a specific role in the processing, and perhaps the targeting pathway, and that they may function to produce oligosaccharides with specific signals for particular roles in the cell. In addition to these exo-al,2-mannosidases, some animal cells and tissues contain an endo-a 1,2-mannosidase that cleaves the glucose branch of the Glcs-iMan9(GlcNAc)2 between the two terminal mannoses to release a GlcsMan-, Glc2Man- or GlciMan- from the oligosaccharide and leave a Man8(GlcNAc)2-protein [168], This enzyme presumably prefers oligosaccharides with a single glucose on the high-mannose chain and may represent an alternate route to that utilizing glucosidase I and glucosidase IL Nevertheless, the specific role of this interesting enzyme in the processing pathway is still not clear; it may represent a new targeting route in some cells. After removal of the four α 1,2-linked mannose units, the Mans(GIcNAc)2-protein is a substrate for GIcNAc transferase I, a glycosyltransferase in the medial Golgi stacks, that transfers a GIcNAc from UDP-GIcNAc to the mannose on the α 1,3-branch to give GIcNAc-Mans(GIcNAc)2-protein [169, 17O]. This enzyme was purified to homogeneity from various sources and shown to be a type II integral membrane protein. The enzyme is specific for the Manal,3Man/?l,4GlcNAc arm of the N-glycan core, and transfers a GIcNAc in /71,2 -linkage to the terminal α 1,3-linked mannose [170, 171]. This reaction is necessary before mannosidase II can remove the α 1,3 and «1,6 mannoses from the Manal,6 arm to give the trimannose structure. The gene for this enzyme was disrupted by homologous recombination in embryonic stem cells and transmitted to the germ line. Mice lacking GIcNAc transferase I activity did not survive to term, and biochemical and morphological analysis of embryos showed that they were developmentally retarded especially in regard to neural tissue [172].

77.5 Processing ofN-linked Oligosaccharides

235

Once the GIcNAc has been added to the 3 -linked mannose, mannosidase II can remove the two mannoses that are linked to the α 1,6-linked mannose branch. The result of this reaction is a GlcNAc/?l,2Manal,3(Manal,6)Man/?l,4GlcNAc/?l,4GIcNAc-protein [173, 174]. Mannosidase II has been purified to homogeneity from rat liver [175] and mung bean seedlings [176], The animal enzyme and the plant enzyme had apparent molecular masses of about 125 kDa on SDS gels, and both enzymes appeared to be glycoproteins [130, 131]. However, the primary sequence of the murine mannosidase II derived from cloning studies predicted a MW of 132 kDa for the deglycosylated enzyme [177]. This discrepancy may be explained by anomalous migration on SDS-PAGE by the deglycosylated or glycosylated protein, since the glycosylated enzyme migrates as a 124 kDa protein. α-Mannosidase II activity has been demonstrated in all mammalian tissues that have been examined. However, the level of the enzyme is very low in brain [178]. Interestingly enough, this tissue has been found to have an alternate hydrolytic enzyme that has α 1,2-, α 1,3- and α 1,6-mannosidase activity and can cleave Man9(GlcNAc)2 down to Mans(GlcNAc)2 [179]. This enzyme is clearly distinct from mannosidase II in terms of its substrate specificity and its reaction to various mannosidase inhibitors. Its specific role in glycoprotein processing is still to be determined. A lack of mannosidase II has been observed in HEMPAS disease, a hereditary affliction that is characterized by altered expression of one or several of the glycoprotein processing enzymes [18O]. One form of the disease results from a deficiency in mRNA expression of «-mannosidase II. Lymphocytes derived from patients having this defect contain less than 10 % of control mannosidase II levels, and their glycoproteins contain mostly hybrid types of oligosaccharides [181]. The catalytic domain of the murine mannosidase II cDNA shows a considerable amount of similarity in sequence to the lysosomal α-mannosidase cloned from the slime mold, Dictyostelium discoideum [182]. Nevertheless, these two enzymes have considerable differences in pH optimum, substrate specificity, and localization within the cell. Based on the sequence similarity, it has been proposed that the two enzymes were derived from the duplication and divergence of a primordial α-mannosidase gene with later acquisition of localization information and substrate specificity. A lesser degree of sequence similarity was observed between murine α-mannosidase II and the endoplasmic reticulum α-mannosidase or its cytoplasmic homolog, or the yeast vacuolar α-mannosidase [183]. Following the action of the various glycosidases in the trimming part of the pathway, a number of glycosyltransferases act on the GIcN AcMans (GIcN Ac)2protein to produce the complex types of N-linked oligosaccharides. Thus, in the trans-Golgi apparatus, there are a number of GIcNAc transferases, galactosyltransferases, fucosyltransferases, and sialyltransferases, that can add these sugars to the N-linked chains to give a great diversity of complex chains, having biantennary, triantennary or tetraantennary structures. Many of these enzymes have been well characterized and a number of the genes for these important proteins have now been cloned [184]. Although there are not any good inhibitors of these enzymes currently available, the search for, or the chemical synthesis of, such compounds should be a rewarding future goal.

236

11.6 11.6.1

11 Inhibitors of GIy cop rote in Processing

Inhibitors of W-linked GIycoprotein Processing Introduction

As described earlier in this chapter, a number of low-molecular weight compounds have been isolated from natural sources, or synthesized chemically, that specifically inhibit the glycosidases in the trimming pathway. These inhibitors have become valuable tools to use in biological systems to determine the role of Nlinked oligosaccharide processing on the function of various membrane or secretory glycoproteins. These inhibitors are of especial interest since they are small molecules that are able to permeate most cells and therefore they can be used with intact cells and tissues to study 'in vivo' situations. In addition, these inhibitors have been very useful to distinguish the various processing enzymes from each other. The best example is shown in Table 11.2 where it can be seen that the many different α-mannosidases show very different sensitivities to the various mannosidase inhibitors [117, 16O]. The remaining sections of this chapter describe the biological activities of the different classes of alkaloid and alkaloidal-like compounds that function as inhibitors of N-linked oligosaccharide processing. A number of naturally occurring, sugar-like compounds, in which the ring oxygen is replaced by a nitrogen, have been isolated. These alkaloids have been discussed earlier in this chapter and have been shown to be potent inhibitors of various glycosidases. The nitrogen in the ring apparently mimics the catalytic intermediate in the reaction, i.e., an oxocarbenium intermediate, but these compounds are still specifically recognized and bound to the active site of a particular glycosidase because of their resemblance in chirality to specific sugars like o-glucose and D-mannose. Thus, they function as valuable inhibitors of glycosidases, such as the enzymes that are involved in glycoprotein processing. Furthermore, since these compounds are small molecules and uncharged at physiological pH, they readily permeate cells, and therefore can be used to great advantage in cell culture as well as in whole animals.

11.6.2

Glucosidase inhibitors

Castanospermine (5) is an indolizidine alkaloid that was first isolated from the seeds of the Australian tree Castanospermum australe [2]. The initial studies on the effect of this compound in biological systems demonstrated that it was a reasonably potent inhibitor of /?-glucosidase [68]. Later studies also showed that castanospermine inhibited a number of isolated α-glucosidases, including the glycoprotein processing enzymes, glucosidase I and glucosidase II, sucrase, maltase and lysosomal a-glucosidase [185]. Since this compound is such a potent inhibitor of intestinal maltase and sucrase, it prevents the degradation of the disaccharides sucrose and maltose, and therefore blocks the normal digestion of starch and sucrose. As a result, the seeds of Castanospermum australe are toxic to animals

11.6 Inhibitors ofN-linked Glycoprotein Processing

237

and cause severe diarrhea and other gastrointestinal upsets [84]. In addition, when castanospermine is fed to mice over a 4 or 5 day period, it inhibits the lysosomal aglucosidase and causes the accumulation of partially degraded glycogen particles within the lysosomes, i.e., a situation similar to that which occurs in Pompe's disease, a genetic disease where afflicted individuals are lacking the lysosomal a-glucosidase [186]. When various cultured animal cells are grown in the presence of castanospermine, the processing of the N-linked oligosaccharides is blocked at the first step (i.e., glucosidase I), and the asparagine-linked glycoproteins have mostly oligosaccharides with Glc3Man9_v(GlcNAc)2 structures [187]. However, in some cells there is an endomannosidase in the Golgi that can release a Glci_3«l,3Man from glucose-containing N-linked oligosaccharides [188]. Although the endomannosidase prefers the monoglucosylated oligosaccharide from which it releases the disaccharide, Glcal,3Man, it can apparently also cleave the oligosaccharide containing three glucose residues, although at a slower rate. Thus, cells that contain this enzyme may be able to get around a castanospermine block. As mentioned above, the role of the endomannosidase in glycoprotein processing is not yet understood. There are other glucosidase inhibitors that act at the level of glucosidase I and have similar effects to that of castanospermine, although they may have somewhat different levels of activity, or different specificities. These include 1-deoxynojirimycin (2), a polyhydroxylated piperidine analog that corresponds to o-glucopyranoside but with a nitrogen in the ring, which also inhibits a- and /?-glucosidases [189]. Another inhibitor is the pyrrolidine alkaloid, 2,5-dihydroxymethyl-3,4dihydroxypyrrolidine (DMDP) (1) [19O]. DMDP is much less effective then the above two inhibitors, suggesting that a six-membered ring structure is preferred for inhibitory activity. Nevertheless, DMDP does inhibit a- and /?-glucosidases [191]. The effect of preventing the removal of the glucose residues from the N-linked oligosaccharides on the targeting of the glycoproteins can be quite dramatic. Thus, when the hepatocyte cell line, Hep-G2, was incubated for various times in the presence of 1-deoxynojirimycin, the rate of secretion of the serum protein, «i-antitrypsin, was greatly diminished, whereas the rate of secretion of other serum N-linked glycoproteins, such as ceruloplasmin and the C3 component of complement, was only marginally affected [192]. Cell fractionation studies indicated that the antitrypsin had accumulated or was held up in the ER-Golgi compartment, suggesting that the presence of glucose on the oligosaccharides might retard the movement of those proteins from the ER to, or through, the Golgi apparatus. Similar results were obtained when the biosynthesis and targeting of the low density lipoprotein receptor of fibroblasts was examined in the presence of castanospermine. In these studies, it could be shown that cells grown in the presence of the inhibitor had only about one-half the number of receptor molecules at their cell surface, and therefore bound much less 125I-LDL. However, these inhibited cells still had the same total number of LDL receptor molecules in the cells. The missing receptor molecules were found to be located in the ER or Golgi, based on cell fractionation studies [193].

238

11 Inhibitors of Glycoprotein Processing

An interesting study was done in IM-9 lymphocytes where castanospermine was used to examine the role of oligosaccharide processing in the biosynthesis and targeting of the insulin receptor. Cells treated with castanospermine had a 50 % decrease in the number of insulin receptors at the cell surface, as demonstrated by the binding of 125I-insulin. The studies showed that removal of glucose residues from the N-linked glycoprotein was not necessary for the cleavage of the insulin proreceptor, i.e., for maturation of the receptor. However, as shown in other systems, the presence of glucose apparently slowed the transport of this glycoprotein out of the ER to the Golgi, resulting in a decrease in the number of receptor molecules at the cell surface [194]. In the case of the £2 glycoprotein of coronavirus, both castanospermine and deoxynojirimycin caused a significant drop by 2 logs in the formation of virus, and also a dramatic inhibition in the appearance of £2 glycoprotein at the cell surface. Significantly, the £2 that was formed in the presence of the glucosidase inhibitors was still acylated with fatty acid as was the control viral E2. However, the drug-induced £2 accumulated in an intracellular compartment that was not definitively identified, but was probably the ER [195]. Another study dealing with the sodium channel of rat brain neurons showed that addition of palmitic acid to this protein was not prevented by the processing inhibitors [196]. The sodium channel is composed of a and β subunits that form a complex during maturation of the channel. The α-subunit undergoes post-translational modification by the addition of a palmitate, and the incorporation of this fatty acid into the glycoproteins was prevented by tunicamycin, a glycosylation inhibitor that completely prevents formation of N-linked oligosaccharides. On the other hand, castanospermine prevented processing of the oligosaccharide chains and the addition of sialic acids, but had no effect on the addition of palmitic acid. This alkaloid also did not affect the covalent assembly of the a and β subunits or the biological function of the channel [196], Thus, the oligosaccharide is apparently necessary for palmitate addition, but the specific structure of the oligosaccharide (i.e., high-mannose or complex) is probably not critical for the addition of palmitate groups. GPl20 is the envelope protein of HIV, the AIDS associated virus, and this protein is a glycoprotein with many oligosaccharide chains. These oligosaccharides are involved in the recognition and mechanism of attachment of HIV to the CD4 receptor on T lymphocytes and other susceptible cells. GP120 interacts with target molecules on the susceptible cells to cause the fusion of the cells with the formation of syncytia, which are necessary for viral formation and infectivity. The glucosidase inhibitors, 1-deoxynojirimycin (DNJ) and castanospermine, caused a significant decrease in the formation of new virus and in syncytium formation [92, 94, 197]. These interesting results have led to the testing of these inhibitors in human clinical trials as potential anti-AIDS drugs. Although the results have not been published, one reported side effect in humans was the occurrence of diarrhea and other gastrointestinal problems in individuals taking these compounds. As shown in Table 11-1, there are a number of other compounds in addition to castanospermine and deoxynojirimycin (DNJ) that are also inhibitors of glycoprotein processing. One such compound is DMDP (1), which occurs in several differ-

77.6 Inhibitors of N-linked Glycoprotein Processing

239

ent plant families. When placed in a medium of cultured animal cells, DMDP inhibits the same step and gives the same oligosaccharide structure, i.e., Glc3Man9_7(GIcNAc)2, as does either castanospermine or DNJ [19O]. However, DMDP is much less effective than these other inhibitors and therefore considerably higher concentrations are required in the medium. The fact that a five-membered ring structure can show glycosidase activity against enzymes that act on hexopyranosides is significant and would certainly warrant modeling studies of this structure in comparison to the indolizidine and piperidine alkaloids. Several other unusual structures that show increased selectivity towards the two processing glucosidases, i.e., glucosidase I and glucosidase II, are discussed below. Australine (3) is a tetrahydroxy pyrrolizidine alkaloid that was found in the same seeds that contain castanospermine, namely Castanospermum australe [6]. However, australine is present in the seeds in much lower amounts than is castanospermine. This compound is a good inhibitor of fungal amyloglucosidase, but it also inhibits the processing glucosidase I. However, as opposed to the other glucosidase I inhibitors discussed above which are also fairly effective against glucosidase II, australine is a very poor inhibitor of glucosidase II [62]. Thus, australine is the first glucosidase inhibitor to distinguish between these two processing enzymes. Nevertheless, the major effect of australine in cell culture is to block glucosidase I and cause the accumulation of glycoproteins having GlcsMan9(GlcNAc)2 structures. Additional more specific compounds like australine, but with more potent activity, will be useful tools to help understand how specific inhibition of glucosidase I or glucosidase II affects cellular function. Another interesting glucosidase inhibitor is 2,6-diamino-2,6-imino-7-O(fl-O-glucopyYanosyl)-D-glycero-L-guloheptitol (MDL 25,637). This compound, referred to as MDL, was synthesized chemically to mimic a disaccharide that would function as a transition state analog of the intestinal enzyme, sucrase [108]. As anticipated, MDL did inhibit rat intestinal maltase, sucrase, isomaltase, glucoamylase, and trehalase when present in micromolar amounts. Most interesting was the observation that MDL also showed specificity for the glucosidases but in the opposite manner to that of australine. Thus, MDL was much more effective against glucosidase II than it was towards glucosidase I [199]. In cell culture, MDL was quite different from the other glucosidase inhibitors in that it caused the accumulation of glycoproteins having Glc2Man9(GlcNAc)2 structures. However, the overall effects of MDL on glycoprotein function in cell culture are likely to be very similar to those observed with castanospermine and other inhibitors of glucosidase I. It would be of great interest and importance to have an inhibitor that would only prevent removal of the last glucose and thus cause an accumulation of glycoproteins with GlciMan9(GlcNAc)2 structures. A compound named trehazolin was isolated as a trehalase inhibitor [200], but also inhibited removal of glucose from the processing glucosidases. This compound inhibited glucosidase I quite well, but was a very poor inhibitor of glucosidase II [201]. The isolation and demonstration that structures like australine, MDL or trehazolin do exist, and that these compounds have selective actions against the processing glucosidases should stimulate the search for more and better inhibitors. Such inhibitors will be useful tools for additional studies on the role of carbohy-

240

11 Inhibitors of Glycoprotein Processing

drate and especially of the glucose residues in the function and localization of Nlinked glycoproteins. In the last few years, it has become clear why and how inhibitors of glucosidase I cause many N-linked glycoproteins to accumulate in the ER. Helenius and coworkers, as well as other investigators [202], have elegantly shown that the ER has a 'protein correction and folding system' that helps newly synthesized ER proteins fold into the proper conformation that is necessary for transport to the Golgi apparatus. This system involves the action of a chaperone, L e., a protein that helps other proteins fold. The chaperone, named calnexin, is also a lectin that recognizes a monoglucosylated high-mannose oligosaccharide on the unfolded glycoprotein. In the presence of castanospermine or other glucosidase I inhibitors, the first glucose cannot be removed, and therefore the unfolded protein cannot be recognized by calnexin. As a result, it does not fold at the normal rate. Most proteins will fold on their own given enough time, but the folding of some may be very slow, and interaction with calnexin can help speed up this process. Thus, proteins like the LDL receptor, or the insulin receptor, or «i-antitrypsin, are transported to the Golgi at a much slower rate in the presence of glucosidase inhibitors because of the inability of calnexin to bind to the protein.

11.6.3

Mannosidase inhibitors

A number of α-mannosidase inhibitors have been identified from natural sources or synthesized chemically. In addition to their use as tools to examine the role of mannose oligosaccharides in the function of N-linked glycoproteins, they have also been valuable to distinguish the various α-mannosidase activities from each other. The first glycoprotein processing inhibitor to be reported was the indolizidine alkaloid swainsonine [1], an inhibitor of mannosidase II [203]. This compound was initially shown to be an inhibitor of the lysosomal α-mannosidase and when administered to animals caused symptoms like those of the lysosomal storage disease a-mannosidosis [67]. Thus, swainsonine was essentially the prototype which chemists could use to design other glycosidase inhibitors. Based on the structures of swainsonine, castanospermine and 1-deoxynojirimycin, it was proprosed that a useful glycosidase inhibitor should have the following characteristics: (i) A ring structure, probably of the pyranose type, with a nitrogen replacing the heterocyclic oxygen; (ii) A number of hydroxyl groups (minimum number necessary still not known); (iii) Stereochemistry of the hydroxyl groups should match that of the competing sugar. In this section, the mannosidase inhibitors will be discussed in the order in which they act in the glycoprotein processing pathway (Figure 11-2), rather then in order of their historical identification. Based on the fact that 1-deoxynojirimycin (DNJ) (2) was a good inhibitor of α-glucosidases, it was reasonable to assume that a related structure, but with man-

77.6 Inhibitors ofN-linked Glycoprotein Processing

241

nose chirality, would be an inhibitor of α-mannosidases. The 2-epimer of DNJ, namely 1-deoxymannojirimycin (DMJ) was synthesized chemically and was indeed found to be a potent inhibitor of the glycoprotein processing mannosidase I [204, 205]. Most interestingly, DMJ did not inhibit jack bean or lysosomal amannosidase, nor did it inhibit mannosidase II. Those observations on the selective specificity of DMJ demonstrate that it is dangerous to screen for new glycoprotein processing inhibitors using the commonly occurring aryl-glycosidases (i.e., a- and /?-glucosidase, galactosidase or mannosidase) to test for the inhibitory activity. That is, if the goal is to find a specific glycoprotein processing inhibitor, such as an inhibitor of ER α-mannosidase, then one would desire a specific inhibitor that does not work on Golgi mannosidase I or mannosidase II, or jack bean or lysosomal α-mannosidases. Therefore, if one used enzymes that hydrolyze aryl-mannosides (such as p-nitrophenyl-a-D-mannopyranoside) to screen for such a compound, the tests would likely be negative and a possible and perhaps important inhibitor would be discarded. In the period since deoxymannojirimycin was synthesized and shown to be a specific inhibitor of Golgi mannosidase I, a number of other neutral a-mannosidase activities have been reported in animal cells. These enzymes have all been discussed earlier although it is still not clear what role, if any, some of them play in the trimming of N-linked oligosaccharides. As also indicated earlier, these enzymes have different substrate specificities from mannosidase I, and thus many of them are resistant to inhibition by DMJ. As these new mannosidases are purified and separated from each other, and from other competing activities, and as rapid assays for measuring their activities become available, it will be easier to identify or synthesize specific new inhibitors for each of these enzymes. Nevertheless, at this time, a number of α-mannosidase inhibitors are known, and the activities of these various compounds on different α-mannosidases are shown in Table 11-2. In animal cells, DMJ inhibited the Golgi mannosidase IA/B and caused the accumulation of glycoproteins having a high-mannose oligosaccharide, mostly of the ManQ(GlcNAc)2 structure [206]. In contrast to the effect of the glucose analog, DNJ, which prevented the secretion of IgD and IgM by cells in culture, DMJ had no effect [207]. As suggested above, this effect of DNJ is due to the requirement for calnexin in protein folding, and its interaction with glucose. However, once the protein has folded and the glucoses are removed, the protein is treated normally as far as targeting, regardless of whether it has a high mannose or modified chain. In one interesting study, DMJ was used as a tool to determine whether glycoproteins were recycled through the Golgi during the endocytic process. In this experiment, membrane glycoproteins were synthesized in CHO cells in the presence of 3 DMJ to inhibit mannose trimming, together with [2- H]mannose to label the Nlinked glycoproteins. After an appropriate incubation, the medium was changed to remove inhibitor and label, and the cells were incubated for additional times. During this second period, the oligosaccharide structure of the transferrin receptor was determined under conditions where it would undergo endocytosis. Before the chase, the oligosaccharide structure of the transferrin receptor was of the high-

242

11 Inhibitors of Glycoprotein Processing

Table 11-2. Effect of Processing Inhibitors on Various a-Mannosidases. Alkaloid

Swainsonine Deoxymannojirimycin

Kifunensine

Mannostatin

Mannoamidrazone

Enzyme

?

0.5-1 μΜ

MQN-Man-Ase (ER)

5-7μΜ

?

?

?

Man-Ase IA (Golgi)

1-2μΜ

?

?

?

Man-Ase I (Mung Bean)

40-50 μΜ

0.02-0.05 μΜ -

4 μΜ

ER Man-Ase

Man-Ase II (Rat Liver)

0.2 μΜ

-

-

?

?

Man-Ase II (Mung Bean)

0.09 μΜ

-

-

0.09 μΜ

0.1 μΜ

mannose type, but during the chase period, a small percentage of the recycled receptor molecules underwent processing and gave complex types of structures. These studies indicated that some endocytosed glycoproteins do recycle through the Golgi compartments and may undergo oligosaccharide processing [208]. However, the number of glycoprotein molecules that were actually modified in this experiment was small, indicating that recycling through the Golgi is probably not a major route. UT-I cells were used to examine the role of the ER α-mannosidase in glycoprotein targeting and function. UT-I cells are cells that overexpress HMG CoA reductase, a glycoprotein enzyme that resides in the ER of the cell. The oligosaccharide chains of this protein are of the high mannose type and mostly Mang(GlcNAc)2 and Man6(GlcNAc)2 structures. Since previous studies had demonstrated that the ER mannosidase was not inhibited by DMJ, this inhibitor could be used to determine whether the initial trimming of mannose residues involved the ER mannosidase. In these studies, the HMG CoA reductase produced in the presence of DMJ had mostly Mang(GlcNAc)2 structures and the smaller oligosaccharides with fewer mannoses were not found, indicating that the ER enzyme was involved in the removal of the first mannose, but other mannoses were probably trimmed by DMJ-sensitive mannosidase(s) [209]. DIM (l,4-dideoxy-l,4-imino-D-mannitol) is another inhibitor that was synthesized from benzyl-a-D-mannopyranose and shown to be a good inhibitor of jack bean α-mannosidase [21O]. It also inhibited glycoprotein processing in cultured MDCK cells, and gave rise to glycoproteins having mostly Man9(GlcNAc)2 structures suggesting that it inhibited the Golgi α-mannosidase I [211]. In keeping with these observations, in vitro studies with a partially purified preparation of mannosidase I showed that DIM did inhibit release of 3H-mannose from 3H-Man9-

11.6 Inhibitors of N-linked Glycoprotein Processing

243

GIcNAc [178]. However, DIM is not nearly as effective an inhibitor of a-mannosidases as is either swainsonine or kifunensine (see below). On the other hand, DIM is of considerable interest as an inhibitor since: (i) It has a furanose rather than a pyranose ring structure, and (ii) It is synthesized chemically and therefore can be produced in large amounts and can readily be modified to produce various structural analogs. It is not clear whether this compound also inhibits the ER mannosidase since this activity was not readily detectable in MDCK cells. Kifunensine is an alkaloid produced by the actinomycete, Kitasatosporia kifunense, and it corresponds in structure to the cyclic oxamide derivative of 1amino-DMJ [15]. This alkaloid is a very weak inhibitor of jack bean a-mannosidase, as is DMJ, but it is a potent inhibitor of the Golgi mannosidase I (ICso = 2 to 5 χ 10"8M). This inhibition is almost 100 times greater than the inhibition of mannosidase I by DMJ. Interestingly, kifunensine had no effect on either the ER mannosidase or on mannosidase II [212]. Influenza virus-infected MDCK cells incubated in the presence of kifunensine produced influenza virus particles in which the envelope glycoproteins contained N-linked oligosaccharides mostly with Man9(GlcNAc)2 structures. This is the same effect as that seen in the presence of DMJ. However, kifunensine was much more effective in causing this change in structure, and only 1/50 as much of this inhibitor was needed as compared to DMJ [212]. A compound that mimics the mannopyranosyl cation, the intermediate proposed to be involved in the enzymatic hydrolysis of α-mannopyranosides, was synthesized chemically and named mannonolactam amidrazone [213]. This compound not only inhibited Golgi mannosidase I with an !€50 of 4 μΜ, and mannosidase II with an ICso of 100 nM, but was also a potent inhibitor of ER «-mannosidase (ICso of 1 mM) [214]. Furthermore, the compound also inhibited the aryl-a-mannosidase (ICso of 400 nM), and the aryl-/?-mannosidase (ICso of 150 μΜ), although it clearly preferred α-linkages. In cell culture studies, mannonolactam amidrazone gave rise to glycoproteins with the same type of high-mannose oligosaccharides as seen with DMJ and kifunensine. Thus, inhibition of Golgi mannosidase I (and/or ER mannosidase) appears to prevent trimming of most if not all mannose residues [214]. The designers of this compound hypothesize that the reason that mannonolactam amidrazone is so effective as a general mannosidase inhibitor is that it is the first analog of mannose that mimics the true half-chair conformation of the cationic intermediate that is believed to be involved in catalysis of the α-mannosides. Mannonolactam amidrazone should serve as a model for the synthesis of more specific mannosidase inhibitors. As mentioned earlier, the first processing inhibitor to be described was the indolizidine alkaloid, swainsonine (4) [I]. In early studies, swainsonine was added to the culture media of MDCK cells infected with influenza virus, and these cultures were labeled by the addition of [2-3H]mannose. This inhibitor caused a significant inhibition in the amount of mannose-labeled, Endo Η-resistant oligosaccharides (i.e., complex oligosaccharides), and a great increase in the amount of mannoselabeled Endo Η-sensitive structures. These latter oligosaccharides were shown to be hybrid types of oligosaccharides [215]. However, the change in the structure

244

11 Inhibitors of Glycoprotein Processing

of the viral oligosaccharides from complex to hybrid types did not affect the production, maturation or release of the influenza virus particles. These early studies did not identify the specific site of swainsonine inhibition, but later in vitro studies with the purified α-mannosidases demonstrated that swainsonine specifically inhibited mannosidase II, but and was inactive towards mannosidase I [216]. In keeping with this site of action, swainsonine caused the formation of hybrid structures when it was added to the medium of various cultured animal cells producing viral glycoproteins [217] or other membrane glycoproteins [218]. In most studies where swainsonine was used to determine the effect of changes in oligosaccharide structure on glycoprotein function, this inhibitor had little effect on functional aspects of the proteins in question, although it did cause alterations in structure to hybrid chains. Swainsonine did however, prevent the receptor-mediated uptake of mannose-terminated glycoproteins by macrophages. This inhibition was probably due to the formation of hybrid structures on the macrophage surface which could then react with and bind to the mannose receptors [219]. Swainsonine proved to be a valuable tool to determine the sequence of addition of certain sugars during the assembly of the N-linked oligosaccharides. Thus, the addition of L-fucose or sulfate to the influenza viral protein was examined in the presence of various processing inhibitors. When the glycoproteins were produced in the presence of castanospermine or DMJ, there was no [3H]fucose [220] or [35S]sulfate [221] associated with the glycoproteins, suggesting that fucose and sulfate were added after the mannosidase I processing step. However, in the presence of swainsonine, the glycoproteins contained both L-fucose and sulfate indicating that the transferases that added these groups worked after the GIcNAc transferase I processing step. These results agree with the reported acceptor oligosaccharide specificity, i.e., GlcNAc-Man5(GlcNAc)2, for the fucosyltransferase, and the sulfο transferase. In some studies, swainsonine did cause a loss in the function of specific proteins. Thus, glucocorticoid stimulation of resorptive cells, involving the attachment of osteoblasts to bone, was inhibited by swainsonine [222]. Treatment of either the parasite, Trypanosoma cruzi, or the macrophages with swainsonine inhibits the interaction of these cells with each other [102]. This alkaloid also caused a dramatic decline in the ability of B16 melanoma cells to colonize the lungs of experimental animals [223]. As a result of these and similar studies, swainsonine has been undergoing tests and consideration as a drug to treat certain types of cancers. These are only a few of the many studies that have been done with this interesting compound. Many of these other studies are summarized in a recent review [191]. Another inhibitor of mannosidase II, named mannostatin, was recently isolated from the fungus, Streptoverticillium verticillus [13]. This compound is of particular interest since it has a very unusual structure with an exocyclic nitrogen, a five-membered ring and a thiomethyl group. Nevertheless, it is still a glycosidase inhibitor. Mannostatin was found to be a potent inhibitor of jack bean a-mannosidase as well as mannosidase II (ICso = 100 nM). In cell culture studies, mannostatin caused the formation of the same types of hybrid oligosaccharides as are formed

References

245

in the presence of swainsonine [16O]. Interestingly, acetylation of the amino group of mannostatin resulted in loss of mannosidase activity. While this compound does not have any functional advantage over swainsonine as an inhibitor, it is of considerable interest since it adds a great deal of additional structural information to our understanding of the requirements necessary for a compound to be a glycosidase inhibitor.

References [11 S. M. Colegate, R R. Dorling, C. R. Huxtable, Aust. J. Chem., 1979, 32, 2257-2264. [2] L. D. Hohenschutz, E. A. Bell, P. J. Jewess, D. R Leworthy, R. J. Pryce, E. Arnold, J. Clardy, Phy to chemistry, 1981, 20, 811-814. [3] A. D. Elbein, R. J. Molyneux, in Alkaloids: Chemical and Biological Perspectives (Ed.: S. W. Pelletier), Wiley-Interscience, New York, 1987, vol. 5, p. 1-54. [4] A. Welter, J. Jadot, G. Dardenne, M. Marlier, J. Casimir, Phy to chemistry, 1976, 15, 747-749. [5] S. Murao, S. Miyata, Agric. Biol. Chem., 1980, 44, 219-221. [61 R. J. Molyneux, M. Benson, R. Y. Wong, J. E. Tropea, A. D. Elbein, /. Nat. Prod., 1988, 57, 1198-1206. [7] R. J. Molyneux, L. F. James, Science, 1982, 276, 190-191. [8] M. J. Schneider, F. S. Ungemach, H. P. Broquist, T. M. Harris, Tetrahedron, 1983, 39, 29-32. [9] M. Hino, O. Nakayama, Y. Tsurumi, K. Adachi, T. Shibata, H. Terano, M. Kohsaka, H. Aoki, H. Imanaka, J. Antibiot., 1985, 38, 926-935. [10] R. J. Molyneux, R. J. Nash, N. Asano, in Alkaloids: Chemical and Biological Perspectives (Ed.: S. W. Pelletier), Pergamon, Oxford, U.K., 1996, Vol. 11, p. 303-343. [11] D. Tepfer, A. Goldmann, N. Pamboukdjian, M. Maille, A. Lepingle, D. Chevalier, J. Denarie, C. Rosenberg, J. Bacteriol, 1988, 770, 1153-1161. [12] A. Goldmann, M-L. Milat, P-H. Ducrot, J-Y. Lallemand, M. Maille, A. Lepingle, I. Charpin, D. Tepfer, Phytochemistry, 1990, 29, 2125-2127. [13] T. Aoyagi, T. Yamamoto, K. Kojiri, H. Morishima, M. Nagai, M. Hamada, T. Takeuchi, H. Umezawa, J. Antibiot., 1989, 42, 883-889. [14] O. Ando, H. Satake, K. Itoi, A. Sato, M. Nakajima, S. Takahashi, H. Haruyama, Y. Ohkuma, T. Kinoshita, R. Enokita, /. Antibiot., 1991, 44, 1165-1168. [15] H. Kayakiri, S. Takase, T. Shibata, M. Okamoto, H. Terano, M. Hashimoto, T. Tada, S. Kota, /. Org. Chem., 1989, 54, 4015-1016. [16] T. Aoyagi, H. Suda, K. Uotani, F. Kojima, T. Aoyama, K. Horiguchi, M. Hamada, T. Takeuchi, J. Antibiot., 1992, 45, 1404-1408. [17] RJ. Nash, P. I. Thomas, R. D. Waigh, G. W. J. Fleet, M. R. Wormald, P. M. de Q. Lilley, D. J. Watkin, Tetrahedron Lett., 1994, 35, 7849-7852. [18] R. J. Molyneux, R. A. McKenzie, B. M. O'Sullivan, A. D. Elbein, J. Nat. Prod., 1995, 58, 878-886. [19] R. J. Molyneux, Y. T. Pan, A. Goldmann, D. A. Tepfer, A. D. Elbein, Arch. Biochem. Biophys.,1993, 304,81-88. [20] N. Asano, A. Kato, M. Miyauchi, H. Kizu, T. Tomimori, K. Matsui, R. J. Nash, R. J. Molyneux, Eur. J. Biochem., 1997, 248, 296-303. [21] N. Asano, A. Kato, K. Oseki, H. Kizu, K. Matsui, Eur. J. Biochem., 1995, 229, 369-376. [22] N. Asano, A. Kato, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, Carbohydr. Res., 1996, 293, 195-204.

246

11 Inhibitors of Glycoprotein Processing

[23] N. Asano, A. Kato, Y. Yokoyama, M. Miyauchi, M. Yamamoto, H. Kizu, K. Matsui, Carbohydr. Res., 1996, 284, 169-178. [24] A. Kato, N. Asano, H. Kizu, K. Matsui, S. Suzuki, M. Arisawa, Phytochemistry, 1997, 45, 425-429. [25] R. J. Nash, M. Rothschild, E. A. Porter, A. A. Watson, R. D. Waigh, P. G. Waterman, Phytochemistry, 1993, 34, 1281-1283. [26] B. Drager, A. van Almsick, G. Mrachatz, Planta Med., 1995, 61, 577-579. [27] N. Asano, K. Oseki, H. Kizu, K. Matsui, /. Med. Chem., 1994, 37, 3701-3706. [28] N. Asano, E. Tomioka, H. Kizu, K. Matsui, Carbohydr. Res., 1994, 253, 235-245. [29] J. V. Bring, G. C. Kite, R. J. Nash, T. Reynolds, BoL J. Linn. Soc., 1995, 777, 1-12. [30] G. C. Kite, J. M. Horn, J. T. Romeo, L. E. Fellows, D. C. Lees, A. M. Scofield, N. G. Smith, Phytochemistry, 1990, 29, 103-105. [31] S. V. Evans, L. E. Fellows, T. K. M. Shing, G. W. J. Fleet, Phytochemistry, 1985, 24, 1953-1955. [32] S. Watanabe, H. Kato, K. Nagayama, H. Abe, Biosci. Biotech. Biochem., 1995, 59, 936-937. [33] R. J. Molyneux, L. F. James, K. E. Panter, M. H. Ralphs, Phytochem. Anal, 1991, 2, 125-129. [34] C. M. Harris, B. C. Campbell, R. J. Molyneux, T. M. Harris, Tetrahedron Lett., 1988, 29, 4815-4818. [35] A. Goldmann, B. Message, D. A. Tepfer, R. J. Molyneux, O. Duclos, E-D. Boyer, A. D. Elbein, J. Nat. Prod., 1996, 59, 1137-1147. [36] R. J. Molyneux, J. E. Tropea, A. D. Elbein, J. Nat. Prod., 1990, 53, 609-614. [37] R. J. Molyneux, J. N. Roitman, G. Dunnheim, T. Szumilo, A. D. Elbein, Arch. Biochem. Biophys., 1986, 257, 450-457. [38] R. J. Molyneux, M. Benson, J. E. Tropea, Y. T. Pan, G. P. Kaushal, A. D. Elbein, Biochemistry, 1991, 30, 9981-9987. [39] R. J. Molyneux, in Bioactive Natural Products: Detection, Isolation and Structural Identification (Eds.: S. M. Colegate, R. J. Molyneux), CRC Press, Boca Raton, FL, 1993, p. 59-74. [40] R. J. Molyneux, in Methods in Plant Biochemistry: Volume 8 - Alkaloids and Sulphur Compounds (Ed.: P. G. Waterman), Academic Press, London, 1993, p. 511-530. [41] L. E. Fellows, E. A. Bell, D. G. Lynn, F. Pilkiewicz, I. Miura, K. Nakanishi, /. Chem. Soc., Chem. Commun., 1979, 977-978. [42] A. M. Scofield, L. E.Fellows, R. J. Nash, G. W. J. Fleet, Life ScL, 1986, 39, 645-650. [43] G. W. J. Fleet, S. J. Nicholas, P. W. Smith, S. V. Evans, L. E. Fellows R. J. Nash, Tetrahedron Lett., 1985, 26, 3127-3130. [44] M. T. H. Axamatwaty, G. W. J. Fleet, K. A. Hannah, S. K. Namgoong, M. L. Sinnott, Biochem. J., 1990, 266, 245-249. [45] B. Winchester, G. W. J. Fleet, Glycobiology, 1992, 2, 199-210. [46] T. Shibata, O. Nakayama, Y. Tsurumi, M. Okuhara, H. Terano, M. Kohsaka, J. Antibiot., 1988,47,296-301. [47] H. Kayakiri, K. Nakamura, S. Takase, H. Setoi, I. Uchida, H. Terano, M. Hashimoto, T. Tada, S. Koda, Chem. Pharm. Bull., 1991, 39, 2807-2812. [48] A. A. Watson, R. J. Nash, M. R. Wormald, D. J. Harvey, S. Dealler, E. Lees, N. Asano, H. Kizu, A. Kato, R. C. Griffiths, A. J. Cairns, G. W. J. Fleet, Phytochemistry, 1997, 46, 255-259. [49] A. M. Scofield, R Witham, R. J. Nash, G. C. Kite, L. E. Fellows, Comp. Biochem. Physiol., 1995, 112A, 187-196. [50] A. M. Scofield, P. Witham, R. J. Nash, G. C. Kite, L. E. Fellows, Comp. Biochem. Physiol., 1995, 112A, 197-205. [51] B. Winchester, Biochem. Soc. Trans., 1992, 142, 699-705. [52] S. V. Evans, L. E. Fellows, T. K. M. Shing, G. W. J. Fleet, Phytochemistry, 1985, 24, 1953-1955.

References

247

[53] S. Watanabe, H. Kato, K. Nagayama, H. Abe, Biosci. Biotech. Biochem., 1995, 59, 936-937. [54] N. Asano, K. Oseki, H. Kizu, K. Matsui, J. Med. Chem., 1994, 37, 3701-3706. [55] S. Murao, S. Miyata, Agric. Biol Chem., 1980, 44, 219-221. [56] H. Hettkamp, G. Legler, E. Bause, Eur. J. Biochem., 1984, 142, 85-90. [57] N. Ishida, K. Kumagai, T. Miida, T. Tsuruoka, Y. Yumoto, J. Antibiot., Ser. A, 1967, 20, 66-71. [58] J. Bischoff, R. Kornfeld, Biochem. Biophys. Res. Commun., 1984, 725, 324-331. [59] T. Niwa, S. Inouye, T. Tsuruoka, Y. Koaze, T. Niida, Agric. Biol. Chem., 1970, 4, 966-968. [60] T. Niwa, T. Tsuruoka, H. Goi, Y. Kodama, J. Itoh, S. Inouye, Y Yamada, T. Niida, M. Nobe, Y. Ogawa, /. Antibiot., 1984, 37, 1579-1586. [61] Y. Miyake, M. Ebata, Agric. Biol. Chem., 1988, 52, 661-666. [62] J. E. Tropea, R. J. Molyneux, G. P. Kaushal, Y T. Pan, M. Mitchell, A. D. Elbein, Biochemistry, 1989, 28, 2027-2034. [63] R. J. Nash, L. E. Fellows, J. V. Bring, G. W. J. Fleet, A. Girdhar, N. G. Ramsden, J. M. Peach, M. P. Hegarty, A. M. Scofield, Phytochemistry, 1990, 29, 111-114. [64] C. M. Harris, T. M. Harris, R. J. Molyneux, J. E. Tropea, A. D. Elbein, Tetrahedron Lett., 1989, 30, 5685-5688. [65] A. C. de S. Periera, M. A. C. Kaplan, J. G. S. Maia, O. R. Gottlieb, R. J. Nash, G. W. J. Fleet, L. Pearce, D. J., Watkin, A. M. Scofield, Tetrahedron, 1991, 47, 5637-5640. [66] I. Pastuszak, R. J. Molyneux, L. F. James, A. D. Elbein, Biochemistry, 1990, 29, 1886-1891. [67] P R. Dorling, C. R. Huxtable, S. M. Colegate, Biochem. J., 1980, 797, 649-651. [68] R. Saul, J. P Chambers, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys., 1983, 227,593-597. [69] R. J. Molyneux, J. N. Roitman, G. Dunnheim, T. Szumilo, A. D. Elbein, Arch. Biochem. Biophys., 1986, 257, 450-457. [70] N. Asano, A. Kato, K. Oseki, H. Kizu, K. Matsui, Eur. J. Biochem., 1995, 229, 369-376. [71] N. Asano, A. Kato, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, Carbohydr. Res., 1996, 293, 195-204. [72] A. Goldman, B. Message, D. A. Tepfer, R. J. Molyneux, O. Duclos, R-D. Boyer, A. D. Elbein, J. Nat. Prod., 1996, 59, 1137-1142. [73] N. Asano, K. Oseki, E. Tomioka, H. Kizu, K. Matsui, Carbohyd. Res., 1994, 259, 243-255. [74] A. Kato, N. Asano, H. Kizu, K. Matsui, S. Suzuki, M. Arisawa, Phytochemistry, 1997, 45, 425-429. [75] J. E. Tropea, G. P. Kaushal, I. Pastuszak, M. Mitchell, T. Aoyagi, R. J. Molyneux, A. D. Elbein, Biochemistry, 1990, 29, 10062-10069. [76] O. Ando, H. Satake, K. Itoi, A. Sato, M. Nakajima, S. Takahashi, H. Haruyama, Y. Ohkuma, T. Kinoshita, R. Enokita, J. Antibiot., 1991, 44, 1165-1168. [77] A. D. Elbein, J. E. Tropea, M. Mitchell, G. P. Kaushal, J. Biol. Chem., 1990, 265, 15599-15605. [78] T. Aoyagi, H. Suda, K. Uotani, F. Kojima, T. Aoyama, K. Horiguchi, M. Hamada, T. Takeuchi, J. Antibiot., 1992, 45, 1404-1408. [79] D. A. Winkler, G. Holan, J. Med. Chem., 1989, 32, 2084-2089. [80] D. A. Winkler, /. Med. Chem., 1996, 39, 4332-4334. [81] W. H. Pearson, E. J. Hembre, /. Org. Chem., 1996, 61, 5537-5545. [82] G. Gradnig, A. Berger, V. Grassberger, A. E. Stiitz, G. Legler, Tetrahedron Lett., 1991, 32, 4889-4892. [83] R. J. Molyneux, L. F. James, M. H. Ralphs, J. A. Pfister, K. E. Panter, R. J. Nash, in Poisonous Plants of the World: Agricultural, Phytochemical and Ecological Aspects (Eds.: S. M. Colegate, P. R. Dorling), CAB International, Wallingford, U.K., 1994, p. 107-112. [84] S. L. Everist, Poisonous Plants of Australia, Angus and Robertson, Sydney, 1974, p. 403-405.

248

11 Inhibitors of Glycoprotein Processing

[85] R. Saul, J. J. Ghidoni, R. J. Molyneux, A. D. Elbein, Proc. Nat. Acad. ScL USA, 1985, 82, 93-97. [86] N. Asano, A. Kato, K. Matsui, A. A. Watson, R. J. Nash, R. J. Molyneux, L. Hackett, J. Topping, B. Winchester, Glycobiology, 1997, 7, 1085-1088. [87] M. J. Humphries, K. Matsumoto, S. L.White, R. J. Molyneux, K. Olden, Clin. Exp. Metastasis, 1990, 8, 89-102. [88] M. J. Humphries, K. Matsumoto, S. L. White, R. J. Molyneux, K. Olden, Cancer Res., 1988, 48, 1410-1415. [89] D. Bowen, J. Adir, S. L. White, C. D. Bowen, K. Matsumoto, K. Olden, Anticancer Res., 1993, 13, 841-844. [90] P. E. Goss, J. Baptiste, B. Fernandes, M. Baker, J. W. Dennis, Cancer Res., 1994, 54, 1450-1457. [91] G. K. Ostrander, N. K. Scribner, L. R. Rohrschneider, Cancer Res., 1988, 48, 10911094. [92] R. A. Gruters, J. J. Neefjes, M. Tersmette, R. E. Y. de Goede, A. TuIp, H. G. Huisman, F. Miedema, H. L. Ploegh, Nature, 1987, 330, 74-77. [93] A. S. Tyms, E. M. Berrie, T. A. Ryder, R. J. Nash, M. R Hegarty, M. A. Mobberley, J. M. Davis, E. A. Bell, D. J. Jeffries, D. Taylor-Robinson, L. E. Fellows, Lancet, 1987, 1025-1026. [94] B. D. Walker, M. Kowalski, W. C. Goh, K. Kozarsky, M. Krieger, C. Rosen, L. Rohrschneider, W. A. Haseltine, J. Sodroski, Proc. Nat. Acad. ScL USA, 1987, 84, 8120-8124. [95] D. L. Taylor, L. E. Fellows, G. H. Farrar, R. J. Nash, D. Taylor-Robinson, M. A. Mobberley, T. A. Ryder, D. J. Jeffries, A. S. Tyms, Antiviral Res., 1988, 10, 11-26. [96] G. Bridges, S. P. Ahmed, M. S. Kang, R. J. Nash, E. A. Porter, A. S. Tyms, Glycobiology, 1995, 5, 243-247. [97] G. W. J. Fleet, A. Karpas, R. A. Dwek, L. E. Fellows, A. S. Tyms, S. Petursson, S. K. Namgoong, N. G. Ramsden, P. W. Smith, J. C. Son, F. Wilson, D. R. Witty, G. S. Jacob, T. W. Rademacher, FEBS Lett., 1988, 237, 128-132. [98] R. H. Taylor, H. M. Barker, E. A. Bowey, J. E. Canfield, Gut, 1986, 27, 1471-1478. [99] F.-J. Chen, N. Nakashima, I. Kimura, M. Kimura, N. Asano, S. Koya, Biol Pharm. Bull, 1995, 18, 1676-1680. [100] P. M. Grochowicz, A. D. Hibberd, Y. C. Smart, K. M. Bowen, D. A. Clark, W. B. Cowden, D. O. Willenborg, Transpl. Immunol, 1996, 4, 275-285. [101] P. S. Wright, D. E. Cross-Doersen, K. K. Schroeder, T. L. Bowlin, P. P. McCann, A. J. Bitonti, Biochem. Pharmacol, 1991, 41, 1855-1861. [102] F. Villalta, F. Kierszenbaum, MoI. Biochem. Parasitol, 1985, 16, 1-10. [103] A. Gottschalk, Glycoproteins: Their Composition, Structure and Function, Elsevier, New York, 1972. [104] M. F. Mescher, U. Hansen, J. L. Strominger, /. Biol Chem., 1976, 257, 7289-7294. [105] F. A. Troy, Glycobiology, 1992, 1, 5-23. [106] T. A.Yednock, S. D. Rosen, Adv. Immunol, 1989, 44, 313-378. [107] T. Feizi, Nature, 1985, 314, 53-57. [108] P. A. Haynes, M. A. J. Ferguson, G. A. M. Cross, Glycobiology, 1996, 6, 869-878. [109] H. Leffler, F. R. Masiarz, S. H. Barondes, Biochemistry, 1989, 28, 9222-9229. [110] Q. Zhou, R. D. Cummings, in Cell Surface Carbohydrates and Cell Development (Ed.: M. Fukuda), CRC Press, Boca Raton, FL, 1992, p. 99. [Ill] A. Kobata, Eur. J. Biochem., 1992, 209, 483-501. [112] S. Kornfeld, J. Clin. Invest., 1986, 77, 1-6. [113] G. Ashwell, J. Harford, Annu. Rev. Biochem., 1982, 57, 531-554. [114] C. Hammond, I. Braakman, A. Helenius, Proc. Natl Acad. ScL, USA 1994, 91, 913-917. [115] K. Olden, J. B. Parent, S. L. White, Biochim. Biophys. Acta, 1982, 650, 209-232. [116] R. U. Margolis, R. K. Margolis, Neurobiology of Glycoconjugates, Plenum Press, New York, 1989.

References [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157]

249

A. D. Elbein, FASEB J., 1991, 5, 3055-3063. R. J. Staneloni, L. F. Leloir, Crit. Rev. Biochem., 1982, 12, 289-326. R. Kornfeld, S. Kornfeld, Annu. Rev. Biochem., 1985, 54, 631-664. G. P. Kaushal, A. D. Elbein, J. Biol. Chem., 1985, 260, 16303-16309. L. F. Leloir, R. J. Staneloni, H. Carminatti, N. H. Behrens, Biochem. Biophys. Res. Commun., 1973, 52, 1285-1292. C. B. Sharma, G. P. Kaushal, Y. T. Pan, A. D. Elbein, Biochemistry, 1990, 29, 89018907. C. B. Hirschberg, M. D. Snider, Annu. Rev. Biochem., 1987, 56, 63-87. C. M. D'Souza, C. B. Sharma, A. D. Elbein, Anal Biochem., 1992, 203, 211-217. J. S. Rush, J. G. Shelling, N. S. Zingg, P. H. Ray, C. J. Waechter, /. Biol Chem., 1993, 268, 10109-13117. W. J. Jankowski, T. Mankowski, T. Chojnacki, Biochim. Biophys. Ada, 1974, 337, 153-162. J. R Spencer, A. D. Elbein, Proc. Natl Acad. ScL USA, 1981, 77, 2524-2527. A. Haselbeck, W. Tanner, Proc. Natl. Acad. ScL USA, 1982, 79, 1520-1524. P. Gold, M. Green, J. Biol. Chem., 1983, 258, 12967-12975. D. J. Kelleher, G. Krebach, R. Gilmore, Cell, 1992, 69, 55-65. V. Kumar, F. S. Heinemann, J. Ozols, Biochem. Biophys. Res. Commun., 1995, 36, 817-826. E. Bause, W. Breur, S. Peters, Biochem. J., 1995, 372, 979-985. D. K. Struck, W. J. Lennarz, in The Biochemistry of Glycoproteins and Glycolipids (Ed.: W. J. Lennarz), Plenum Press, New York, 1980, p. 35. M. D. Snider, in Biology of Carbohydrates (Eds.: V. Ginsburg, P. W. Robbins), Wiley and Sons, New York, 1984, Vol. 2, pp. 163-198. S. Turco, B. Stetson, P. W. Robbins, Proc. Natl. Acad. ScL, USA 1977, 74, 4411-4414. J. Schweden, C. Borgmann, G. Legler, E. Bause, Arch. Biochem. Biophys., 1986, 248, 335-340. T. Szumilo, G. P. Kaushal, A. D. Elbein, Arch. Biochem. Biophys., 1986, 247, 261-271. H. Hettkamp, G. Legler, E. Bause, Eur. J. Biochem., 1984, 742, 85-90. E. Bause, J. Schweden, A. Gross, B. Orthen, Eur. J. Biochem., 1989, 183, 661-669. K. Shailubhai, M. A. Pratta, I. K. Vijay, Biochem. J., 1987, 247, 555-562. E. Bause, R. Erkens, J. Schweden, L. Jaenicke, FEBS Lett., 1986, 206, 208-213. B. Kalz-Fuller, E. Bieberich, E. Bause, Eur. J. Biochem., 1995, 257, 344-351. S. C. Hubbard, P. W. Robbins, /. Biol Chem., 1979, 254, 4568-4579. C. Hammond, I. Braakman, A. Helenius, Proc. Natl. Acad. ScL USA, 1994, 97, 913-917. I. Braakman, K. Hoover-Litty, R. Wagner, A. Helenius, J. Cell Biol, 1991, 114, 401-411. F. E. Ware, A. Vassilakos, P. A. Peterson, M. R. Jackson, M. A. Lehrman, D. B. Williams, /. Biol Chem., 1995, 270, 4697-4704. R. G. Spiro, Q. Zhu, V. Bhoyroo, H. D. Soling, J. Biol. Chem., 1996, 277, 1158811594. A. DeSilva, I. Braakman, A. Helenius, J. Cell Biol., 1993, 120, 647-655. M. Sousa, M. A. Ferrero-Garcia, A. Parodi, Biochemistry, 1992, 31, 97-105. P. Choudhury, Y. Liu, R. J. Bick, R. N. Sifers, /. Biol Chem., 1997, 272, 13446-13451. D. M. Burns, O. Touster, /. Biol Chem., 1982, 257, 9991-10000. G. P. Kaushal, I. Pastuszak, K. Hatanaka, A. D. Elbein, /. Biol Chem., 1990, 265, 16271-16279. D. Brada, U. C. Dubach, Eur. J. Biochem., 1984, 747, 149-156. G. P. Kaushal, Y. Zeng, A. D. Elbein, J. Biol. Chem., 1993, 268, 14536-14542. S. Saxena, K. Shailubhai, B. D. Yu, I. K. Vijay, Biochem. J., 1987, 247, 563-570. F. Maartiniuk, A. Ellenbogen, R. Hirschhorn, /. Biol Chem., 1985, 260, 1238-1242. J. M. Lucocq, D. Brada, J. Roth, J. Cell Biol., 1986, 102, 2137-2146.

250 [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] [174] [175] [176] [177] [178] [179] [180] [181] [182] [183] [184] [185] [186] [187] [188] [189] [190] [191] [192] [193] [194] [195] [196] [197] [198]

11 Inhibitors of Glycoprotein Processing D. Brada, D. Kerjaschki, J. Roth, /. Cell BioL, 1990, 110, 309-318. T. Szumilo, A. D. Elbein, Anal. Biochem., 1985, 757, 32-40. A. D. Elbein, Annu. Rev. Biochem., 1987, 56, 497-534. F. Maley, R. B. Trimble, A. L. Tarentino, T. H. Plummer, Jr., Anal Biochem., 1989, 180, 195-204. P. F. Daniel, B. Winchester, C. D. Warren, Glycobiology, 1994, 4, 551-566. F. Lipari, A. Herscovics, J. BioL Chem., 1996, 277, 27615-27622. J. Bischoff, R. Kornfeld, J. BioL Chem., 1983, 258, 7907-7910. J. Schweden, E. Bause, Eur. J. Biochem., 1986, 757, 563-570. I. Tabas, S. Kornfeld, /. BioL Chem., 1979, 254, 11655-11663. W. T. Forsee, J. S. Schutzbach, J. BioL Chem., 1981, 256, 6577-6582. W A. Lubas, R. G. Spiro, J. BioL Chem., 1987, 262, 3775-3781. H. Schachter, Biochem. Cell BioL, 1986, 64, 163-181. M. Metzler, A. Gertz, M. Sarkar, H. Schachter, J. W. Schrader, J. D. Marth, EMBO J., 1994, 13, 2056-2065. N. Harpaz, H. Schachter, J. BioL Chem., 1980, 255, 4885-4893. B. Yip, S.-H. Chen, H. Mulder, J. W. M. Hoppener, H. Schachter, Biochem. J., 1997, 527, 465-479. E. Ioffe, R Stanley, Proc. Natl. Acad. ScL USA, 1994, 97, 728-732. D. P. R. Tulsiani, S. C. Hubbard, R W. Robbins, O. Touster, /. BioL Chem., 1982, 257, 3660-3668. K. W Moreman, O. Touster, /. BioL Chem., 1986, 267, 10945-10951. G. P. Kaushal, T. Szumilo, I. Pastuszak, A. D. Elbein, Biochemistry, 1990, 29, 21682176. K. W. Moreman, P. W Robbins, J. Cell BioL, 1991, 775, 1521-1534. L. S. Wilkerson, O. Touster, Arch. Biochem. Biophys., 1993, 303, 238-245. D. R. R Tulsiani, O. Touster, J. BioL Chem., 1985, 260, 13081-13087. M. N. Fukuda, K. A. Masri, A. Dell, L. Luzzatto, K. W. Moreman, Proc. Natl. Acad. ScL USA, 1990, 87, 7443-7447. M. N. Fukuda, Glycobiology, 1990, 7, 9-15. J. Schatzle, J. Bush, J. Cardelli, J. BioL Chem., 1992, 267, 4000-4007. K. W. Moreman, R. B. Trimble, A. Herscovics, Glycobiology, 1994, 4, 113-125. R. Kleene, E. G. Berger, Biochim. Biophys. Ada, 1993, 7754, 283-325. Y. T. Pan, J. Ghidoni, A. D. Elbein, Arch. Biochem. Biophys., 1992, 303, 134-144. P. Baudhuin, H. G. Hers, H. Loeb, Lab. Invest., 1964, 13, 1139-1145. Y. T. Pan, H. Hori, R. Saul, B. A. Sanford, R. J. Molyneux, A. D. Elbein, Biochemistry, 1983, 22, 3975-3984. S. E. Moore, R. G. Spiro, J. BioL Chem., 1990, 265, 13104-13112. D. D. Schmidt, W. Frommer, L. Miiller, E. Truscheit, Naturwissenschaften, 1979, 66, 584-585. A. D. Elbein, M. Mitchell, B. A. Sanford, L. E. Fellows, S. V. Evans, /. BioL Chem., 1984, 259, 12409-12413. A. D. Elbein, in Cell Surface and Extracellular Glycoconjugates (Eds.: R. P. Meecham, D. Roberts), Academic Press, NY, 1993, p. 119-180. H. Lodish, N. Kong, J. Cell BioL, 1984, 98, 1720-1729. E. H. Edwards, E. A. Sprague, J. L. Kelley, J. J. Kerbacher, C. J. Schwartz, A. D. Elbein, Biochemistry, 1989, 28, 7679-7687. R. F. Arakai, J. A. Hedo, E. Collier, P. Gordon, /. BioL Chem., 1987, 262, 1188611892. R. Repp, T. Tamura, C. B. Boschek, H. Wege, R. T. Schwarz, H. Niemann, J. BioL Chem., 1985, 260, 15873-15879. J. W. Schmidt, W. A. Catterall, J. BioL Chem., 1987, 262, 13713-13723. D. C. Montefiori, W. E. Robinson, Jr., W. M. Mitchell, Proc. Natl. Acad. ScL USA, 1988, 85, 9248-9252. P. Liu, J. Org. Chem., 1987, 52, 4717-4721.

References

251

[199] G. P. Kaushal, Y. T. Pan, J. E. Tropea, M. Mitchell, R Liu, A. D. Elbein, /. Biol Chem., 1988, 263, 17278-17283. [200] C. Uchida, H. Kitahashi, S. Watanabe, S. Ogawa, J. Chem. Soc., Perkin Trans. /, 1995, 1707-1717. [201] Y.-C. Zeng, A. D. Elbein, 1998, manuscript in preparation. [202] A. Helenius, MoL Biol Cell 1994, 5, 253-265. [203] A. D. Elbein, R. SoIf, K. Vosbeck, P. R. Dorling, Proc. Natl Acad. ScL USA, 1981, 78, 7393-7397. [204] G. Legler, E. Julich, Carbohydr. Res., 1984, 128, 61-72. [205] V. Fuhrmann, E. Bause, G. Legler, H. Ploegh, Nature, 1984, 307, 755-758. [206] A. D. Elbein, G. Legler, A. Tlusty, W. McDowell, R. Schwarz, Arch. Biochem. Biophys., 1984, 235, 579-588. [207] V. K. Gross, T. R. Tran-Thi, W. McDowell, R. T. Schwarz, Eur. J. Biochem., 1985,150, 41-46. [208] M. D. Snider, O. C. Rogers, J. Cell Biol., 1986, 103, 265-275. [209] J. Bischoff, L. Liscum, R. Kornfeld, /. Biol. Chem., 1986, 261, 4766-4774. [210] G. W. J. Fleet, P. W. Smith, S. V. Evans, and L. E. Fellows, /. Chem. Soc., Chem. Commun., 1984, 1240-1241. [211] G. Palamarczyk, M. Mitchell, P. W. Smith, G. W. J. Fleet, A. D. Elbein, Arch. Biochem. Biophys., 1985, 243, 35-45. [212] A. D. Elbein, J. E. Tropea, M. Mitchell, G. P. Kaushal, J. Biol. Chem., 1990, 265, 15599-15605. [213] B. Ganem, G. Papandreou, /. Am. Chem. Soc., 1991, 113, 8984-8985. [214] Y. T. Pan, G. P Kaushal, B. Ganem, A. D. Elbein, /. Biol. Chem., 1992, 267, 83138318. [215] A. D. Elbein, K. Vosbeck, P R. Dorling, M. Horisberger, /. Biol. Chem., 1982, 257, 1573-1576. [216] D. P. R. Tulsiani, H. P. Broquist, L. F. James, O. Touster, Arch. Biochem. Biophys., 1984, 232, 76-85. [217] M. S. Kang, A. D. Elbein, J. Virol., 1983, 46, 60-69. [218] R. G. Arumughan, M. L. Tanzer, /. Biol. Chem., 1983, 258, 11883-11889. [219] K.-N. Chung, V. L. Sheperd, P. Stahl, /. Biol. Chem., 1984, 259, 14637-14641. [220] P. Schwartz, A. D. Elbein, J. Biol. Chem., 1985, 260, 14452-14458. [221] R. Merkle, A. D. Elbein, A. Heifetz, /. Biol. Chem., 1985, 260, 1083-1089. [222] Z. Bar-Scavit, A. J. Kahn, L. E. Pegg, R. R. Stone, S. L. Teitelbaum, J. CUn. Invest., 1984, 73, 1277-1283. [223] M. T. Humphries, K. Matsumoto, S. L.White, K. Olden, Proc. Natl. Acad. ScL USA, 1986, 83, 1752-1756.

Tables of Glycosidase Inhibitors with Nitrogen in the Sugar Ring and Their Inhibitory Activities CHRISTIAN W. EKHART, MARTIN H. FECHTER, PHILIPP HADWIGER, EVA MLAKER, ARNOLD E. STUTZ, ANDREAS TAUSS, and TANJA M. WRODNIGG

Inrinosiigars as Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

254

1

Tables of Glycosidase Inhibitors

Introduction

When initial plans for this book were made, the publisher as well as quite a few of the other authors suggested that it should be accompanied by a table of the compounds concerned as well as their glycosidase inhibitory activities. Not being able to fully appreciate the problems involved in this task, we have immediately agreed to prepare such a compilation. The following tables now include data of more than 400 compounds and their activities with a total of approximately 200 different glycoside hydrolases that have been extracted from the literature available in Graz by early 1998. Inhibitors taken into consideration have (somewhat arbitrarily and with a few exceptions) been more or less limited by the authors to compounds exhibiting inhibitory activities of Ki < 1000 μΜ, ICso < 1000 μΜ, and inhibition > 80% at 1 mM, respectively, as reported in the respective original literature. Furthermore, it was attempted to provide useful additional information concerning the availability of NMR spectral data as well as other features such as the access to the compound in question (fermentation/isolation, de novo, chiral pool, chemo-enzymatic). The more than 300 references are confined to publications including inhibitory data and have been listed by their year of appearance. This limitation results merely from the fact that for quite a few key compounds a number of different syntheses have been published, the careful reviewing of which would have gone far beyond the scope of this account. About twenty late entries have been added to the end of this list. For easier comparison, in a few selected cases, the inhibitory concentrations have been converted from mg/mL into μΜ by the reviewers. The applied nomenclature has remained a topic of serious discussions and is now largely based on carbohydrate/iminoalditol rules. In a few cases this approach is (more than) slightly forced and could not be maintained throughout the entire table due to the large diversity of structural types of compounds under consideration. In most cases, a second or generic name is given, as used by the authors of the original papers. For many compounds not even the original publication provides any kind of name and it would have exceeded the scope of our efforts to address this problem. Due to these foreseeable shortcomings, the structures given will be the main guidances for data retrieval. For convenience, these have been put into smaller groups of compounds with similar structural features mainly based on the types of ring systems (mono- or bicyclic), the ring size as well as the total number of carbons. We deeply apologize for any inconvenience caused by the inevitable omittances and mistakes or "typos" in the tables as well as in the included references which, despite our best efforts, will be detected by the reader. At this point, it should also be mentioned that in some of these cases we will have to "pass the blame" on to the authors of the original contributions which have not always made it easy to retrieve valid information. This is mainly the case and is also an inherent problem regarding the respective conditions employed in screening procedures such as pH and, occasionally, even the enzyme probed or its origin.

Tables of Glycosidase Inhibitors

255

Compiling this table has been the major pastime of the listed authors during 1997 and early 1998 and we do hope that it will be a useful source of rapid information and guiding references to other workers in the field.

Legend

_t

,

Structure

Name synonym(s) (if available) means of access (g* novo, cmral pool, chemo« enzymatic, modification of natural product, natural

!*i^M»Mi>h

Oofnpomd: number in table X-Ray;Ref of X-ray data (if available) PK, (if available) 1,5-Dideoxy4,5-imino-D-glueit0l l-DeoxynQjirimycin chiral poo! chemical modification of natural EC:3.2,120,xeesr product natural product, tmi cortex, bnchocarpus Ki: 25±4 {6.8} /2297f Ki: 25 (6,0) f-?96) ICw: 330 sericeus, streptomyces lavendutae (opt) imi K<: 14.6 (6.5)/f5a;, ICw: n$v#ml Compound; c
n.g., not given or not available; opt, pH optimum of enzyme; Ac, acetyl; Ph, phenyl

256

Tables of Glycosidase Inhibitors

Contents 1 2.1 2.2 2.3 2.3.1 2.3.2 2.4 3.1 3.1.1 3.1.2 3.1.3 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.2.5.1 3.2.5.2 3.2.5.3 3.2.5.4 3.2.6 3.2.7 3.2.7.1 3.2.7.2 3.2.7.3 3.2.8 3.2.9 3.2.10 3.3 3.3.1 3.3.2 3.4 4 5 5.1 5.2 5.3 5.3

Four-Membered Ring Five-Membered Rings, C4 Five-Membered Rings, C5 Five-Membered Rings, C6 1,4-Iminoalditols and Relatives 2,5-Iminoalditols and Relatives Five-Membered Rings, C7 Six-Membered Rings, C5 arabino-, ribo-, xylo-Configurations Azafagomines Pyrimidines Six-Membered Rings, C6 allo-, altro-Configurations D-galacto-Configuration L-fuco-, L-galacto-Configurations GlucNAc-Mimics 1-Deoxynojirimycin and Analogs N-Modified O-Glycosylated N-Modified and O-Glycosylated Other Modifications L-gluco-, L-gulo-, L-ido-Configurations 1-Deoxymannojirimycin and Analogs N-Modified O-Glycosylated or Ether Other Modifications talo-, rhamno-, D-gulo-Configurations Isofagomine and Analogs Siastatins and Analogs Six-Membered Rings, C7 1,5-Iminoalditols 2,6-Iminoalditols Six-Membered Rings, C8 and Higher Seven-Membered Rings Nagstatins and Other C-l/N-5 Containing Heterocycles Heterocycles Containing One Nitrogen Atom, NI Heterocycles Containing Two Nitrogen Atoms, N2 Heterocycles Containing Three Nitrogen Atoms, N3 Heterocycles Containing Four Nitrogen Atoms, N4

258 258 258 265 265 268 274 275 275 278 279 282 282 283 288 290 292 296 310 312 317 326 327 330 330 333 337 338 342 347 347 347 351 353 356 356 357 359 360

Tables of Glycosidase Inhibitors 6 7 7.1 7.2 8 9

Pyrrolizidines Indolizidines Castanospermine and Analogs Lentiginosines, Swainsonines and Analogs Quinolizidines Calystegines

257 362 364 366 373 377 378

258

Tables of Glycosidase Inhibitors

a H

a a a HCL H^ -H CNJ

po

Ill ili r$* ±-

--L

O)

CO

H

CO

o

I

.• _

'

v—-? ^

'

o

o 01

01

Iminosngars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

[HSSBRRHREH! EC: 3.2.1 .21 . bitter almonds iemulsin) ICso: 200 (4.8) /026/ RREBRiERBEHu mouse out diaestive ICso: 13 (n.g.) /040/ WfRpBlEPflTRBH^^ EC: 3.2.1 .10, mouse gut digestive ICso: 4.0 (n.g.) /040/ HIBBBIlSBHiHRE^ 3.2.1.10. rat intestine ICso: 5.8 (5.8) /205/ [1 75/ IHiJiHR-ISi EC: 3.2.1 .28, porcine kidney ICso: 4.8 (5.8) [1 75/ SHlEIEHciEC: 3.2.1.28, rat intestine ICso: 25 (5.8) /205/ [1 75/

IBBiEIEBBEtBEEi EC: 3.2.1 .23, bovine liver cytosolic ICso: 1 000 (6.8) [1 75/

ICso: 260 (5.8) /205/ /175/

IHBBIEtSBEIBEBEIIEBEEBI EC: 3.2.1.23, rat intestine

Ιβ-D-fructofuranosidase (invertase, sucrasell-MlflEll 26, rat intestine ICso: 16 (5.8) /205/

HBBHiBBHBEHi EC: 3.2.1.24, rat liver lysosomal ICso: 1 10 (4.5) /205/ fi 75/

HBIBHBiBHRHS!! EC: 3.2.1 .24. rat liver αοΐαί Il Ki: 35 (5.5), ICso: 46 (5.5) /205/ [175].

ICso: 53 (5.5) /775/

HBBEBBBEiBEEB EC: 3.2.1 .24, rat liver golgi I

IBiK^.^ffl!BHSrEC: 3.2.1.24. rat eoididvmis IC50: 84(5.2)/) 75/

260

Tables of Glycosidase Inhibitors

OH

NHCI

f OH

OH

J--?

H3C*

φ

COPh

H3C*- ( OH

£)

CH2Ph

OH

J--7

HO

'*— /

A JO

1

Ίφ

HO

HO

ΒΗιΒΠΗΒΒΗΪιΕΗϊΙ EC: 3 2 1 20. yeast iCso: 0.08 (n.g.) /305/.

Compound: a012 Synthetic/NMR: [210]

chiral pool chemical

4-Amino-4.5-dideoxy-L-arabinose hydrochloride

chiral pool chemical Compound: a011 Synthetic/NMR: [210]

N-Benzoyl-1 ,4,5-trideoxy-1 ,4iminoL-arabinitol

Compound: aOW Synthetic/NMR: 1210]

chiral pool chemical

Ix-L-rhamnosidase (narinqinase)R3 3.2.1 .40, penicillium Ki: 0.14 (5.0) [2101

Ix-L-rhamnosidase inarinqinasejRa 3.2.1.40, penicillium Ki: 46 (5.0) [21 0].

decumbens

decumbens

N-Benzyl-1 ,4,5-trideoxy-1 ,4-imino- lx-L-rhamnosidase (narinqinaseiRfl 3.2.1 .40, peniciHium decumbens Ki:11.5(5.0)/2tOJ. L-arabinitol

Compound: a009 Synthetic/NMR: [210]

chiral pool chemical

KJ: 5.5 (5.0) [21 0].

f«BJ,Mul,f,HI^WJI,MJI,MI,^4! EC:3.2.1 .40, penicillium decumbens

arabinitol

1 Δ «LTriHAftvy.l A.iminn.1 .

bC: 3.2.1 .10, mouse out diaestive

Bltffi|;fl»|p9!HflBff!li^^ IC50: 0.066 (n.g.) /040/

chiral pool chemical Compound· aflflfl Synthetic/NMR: /026/ /040]

IC50: 0.24 (n.g.) /04OJ.

1 ,4-Dideoxy-1 ,4-imino-L-arabinitol JBiBIBBBHf^H:! EC: 3 2.1.20. yeast ICuo:10(6.8)/026J. (2S,3S,4S)-2-Hydroxymethyl-3,4BISBHiHBHS! mouse put digestive dihydroxypyrrolidine; LAB1

Compound: a007 Synthetic/NMR: /3057, /059]

nectria ludda (fungus), hyacinthoides non-scripta

f2R,3f?,4Rj-3,4-cf^yflfra-2-/7yQfroxymef/7y/-2-Hpyrrole-3,4-diol; Nectrisine; FR900483

4-Amino-4-deoxy-D-arabinose

262

Tables of Glycosidase Inhibitors

Γ

HO

HO-ι

HO

OH

ijy

(M

OH NH2

·#

πΛ1

Synthetic/NMR: /276]

Pfunnniinri* a01Q v^uiiipuuiiu. αϊ/ / {7

4-(4-Amino-4-deoxy-B-Dribofuranosyl)aniline Ki:12±1(7.5)f276/.

ΙΙΙΗΊΙΙΜΜ»Η[ιΓϊ1!?7»Ι~"1ΡΤ:~3 crithidia fasciculata Ki:0.030±0.002(7.5)i276;.

|oligo-1,6-glucosidase (isomaltasei^»»ieili[tlBB)?S!f^^^^^^^^^^M 1C·»: 17 (5.8) /f 75/. 4-Bromo-1 -(4-amino-4-deoxy-p-D- IHBillMfJ«THf«niTO [TORT^ crithidia fasciculata Ki: 0.028±0.004 (7.5) [276]. ribofuranosyl)benzene 1 IBBBllHBHlfBBRBBIHISffi trypanosoma brucei brucei Pnrnnniinri VsUlllpUUIIU. aCHR QvJlQ Ki:113±6(7.5)£76/. Synthetic/NMR: [276]

Ki:4.5±0.4(7.5)/276/

Ιβ-D-alucosidase (cellohiase^ Ι?3ΚΙΪΙΤ·ΒΗβΒ?^^^^^^^^^^^^^1 ICso: 230 (5.8) [1 75]. [flgffilflljji^ trypanosomd brucei brucei Κι:44±4(7.5)/276/.

fSBBHBSBHff^^ EC: 3.2.1 .23, rat intestine ICso: 17 (5.8) [1 75]. ItBBHHBiHBBI EC: 3.2.1 .23, bovine liver cytosolic ICso: 580 (6.8) [1 75/.

raifli'iMil'iftUN^.L^ EC: 3.2.1 .24. rat liver Ivsosomal ICso: 1000 (4.5) ^75J.

to

U)

o>

264

Tables of Glycosidase Inhibitors

a

H co.

6

•A _

ι

S! •g

is:·

" I~o S>.

1! ι! S >»

s

HO

HO

HO

HO

HC,

CH2Ph

HH

chiral pool chemical Compound: a027 Synthetic/NMR: /0427

1,4-Dideoxy-1,4-imino-D-glucitol hydrochloride

chiral pool chemical Compound: a026 Synthetic/NMR: [1477, [1607

imino-D-galactitol

N-Benzyl-DIA chiral pool chemical Compound: a025 Synthetic/NMR: /0877 2-Acetamido-1,2,4,-trideoxy-1,4-

N-Benzyl-1,4-dideoxy-1,4-imino-Lallitol hydrochloride

DlA chiral pool chemical Compound: a024 Synthetic/NMR: [087], 1101]

hydrochloride

1,4-Dideoxy-1,4-imino-L-allitol

1,4-lminoalditols and Relatives

2.3.1

HCl

Five-Membered Rings, Ce

2.3

Ki: 700 (5.2) [042/ BBBHBBHBBBI EC: 3.2.1.21, sweet almonds K1:125 (5.2) [042].

I BEEEEmEEC: 3.2.1.20, yeast

\\S^SE\MMMm^: 3.2.1.30, human liver, neutral Ki: 100 (4.0), ICso: 250 (4.0) [147]. Ιι.Ρ4ΐ»η«Μΐι^:€*Β«ιιιΐΓΐΐ[«ΒΜ«ΐϊΐ^χ·κΒΐιιιιιι»ΒΜ^.ίΐ' EC: 3.2.1.52, human liver lysosomal Ki: a: 220 (4.4), β: 18 (4.4) [16O]. BBaaaiBiBiBialAliftUMIMU EC: 3.2.1.53, human liver, neutral Ki: 200 (4.0), ICso: 600 (4.0) [147].

I EC: 3.2.1.51, human liver, neutral Ki: 50 (5.5) /087/.

BSEHmMEEEC: 3.2.1.24, human liver lysosomal Ki: 120 (4.0) /fOf7, K: 170 (4.0) [087]. IKiBHiBiHBHSiI EC: 3.2.1 .24, human liver, qolqi Il IC: 91% 1mM (5.75) [087].

266

Tables of Glycosidase Inhibitors

9

H 3

u

(HCI)2

CT

~|

HO — ι

O

L\ TOHHO)

N

Γ

vj_j/

HH «HHfe

HO

Η,Ν

liuBliBBBHBBii EC: 3.2.1.24. human liver, neutral IC: 91% 1mM (6.5) [117], IC50: 300 (6.5) /762/

K1: 1000 (n.g.) /097/

Η3Ρ|Π3ΒΕΕ3:| EC: 3.2.1 .21 r sweet almonds

ΠίΒΒΠΐΒΗ EC: 3.2.1 .24, jack beans Ki: 0.5 (5.0), ICso: 0.6 (5.0) /027/

1,4-Dideoxy- N-methyl-1,4-iminoD-mannitol N-Methyl-DIM Compound: a034 Synthetic/NMR: [11 7]

Compound* a033 Synthetic/NMR: [162]

chiral ΠΠΛ! phpmiofll

1

1 ^ IA (\\ MR )!

UBBBBBBHBHS EC: 3.2.1 .24. human liver, qolqi Il IC: 91% 1mM (5.75) [117], IC50: 60 (5.75) /762].

ίΓ· 1ΠΠΟ/ Ί-ηΛΛ (A (\\ H 171 If,· 1 ^ M ft\ HtRI li.·

JBBBBiIBHBSHS bL: ^.2.1 .24, human liver lysosomal

ΙΒ·.ΙΙ/;Ι>!ΓιΜ·|[ΐΗ:Β EC: 3.2.1.55r human liver, neutral IC: 91% 1mM (n.g.) /736/

IC5o:13-106(n.g.)/779/

ΒΒιΒίΊΜΊΙΊ?»ΗΤ?Π"Η EC: 3.2.1.24. jack beans

Μι1ϋΜΊΙ'ιΜΊΕΕΗ5 EC: 3.2.1 .24, human liver, neutral ICs§: 25 (6.5) /762/ !Hili'iM'il'iBHt»g".S EC: 3.2.1.24, human liver, neutral IC: 45% 1mM (6,5) /777/

IKiBiBBBHIBB! tC: 3.2.1 .24r human liver, qolqi Il IP«*· *¥) ^ 7M M ROl

1 -4,6-Trideoxy-6-fluoro-1 ,4-imino- !Bii"iM'iUM3f»Et:!g EC: 3.2.1.24. human liver lysosomal Ki: 1.5 (4.0) /762/ D-mannitol

Compound: a032 Synthetic/NMR· [027] [097] [136]

fhiral nnnl rhpmipal

6-Amino-1 ,4,6-trideoxy-1 ,4-iminoD-mannitol dihydrochlortde chiral pool chemical Compound: a031 Synthetic/NMR: /179] 1 ,4,6-Trideoxy-1 ,4-imino-Dmannitol

κ-> <>j

i-

~*J·

^i

1

U)

ISO

K)

nbered Rin^

268

Tables of Glycosidase Inhibitors

S

_CD

.~gi CO

CSl CO CNI

pNHAc

V* OH \^

Η

W

Η

Γ

Π H? °

ΗΟ

Γ

tj

H3C „ ρΟΗ

HO-,

OH

Γ

KJ

°-i H r-°H

H

1 -Acetamido-1 ,2,5-trideoxy-2,5imino-D-glucitol chiral pool chemical chiral pool chemoenzymatic Compound: a041 Synthetic/NMR: [1641 [154] 2,5,6-Trideoxy-2,5-imino-D-glucitol de novo chemo-enzymatic Compound: a042 Synthetic/NMR: [148] 1,2,5-Trideoxy-2,5-imino-D-glucitol de novo chemo-enzymatic Compound: a043 Synthetic/NMR: [148]

2,5-Dideoxy-2,5-imino-D-glucitol (2R,3R,4R,5S> 2,5-Dihydroxymethyl-3,4dihydroxy-pyrrolidine; 2, 5-Dideoxy-2, 5-iminoL-gulitol chiral pool chemo-enzymatic, de novo chemoenzymatic Compound: a040 Synthetic/NMR: [301], [291], [119]

ΙΒΒίΙΜ·1.ΜΗ53 EC: 3.2.1 .51 r bovine kidney Ki:8(5.5)ff48y.

!HBffAHBEEES EC: 3.2.1.51, /bov/ne kidney K1: 4 (5.5) /f 48|

IJBiEKiilflllFtNl^.W.W»!^ EC: 3.2.1 .30, jack beans Ki: 3.6 (6.5) [154]. i^iTiaiBigfflBflSffiffiffiEffi EC: 3.2.1 .30, />ov//?e kidney Ki: 68.6 (6.5) /f 54], Ki: 39,3±0.51 (4.4) /f 64].

ΙΗΒΒΠΒΒΗΙΜΒ EC: 3.2.1 .21 , sweet almonds Ki: 19 (6.5) [1 19], Id: 52 (5.0) /3Of/ Ki: 19 (6.8) [291].

|Β3ΡΙΠ35[Η EC: 3.2.1 .21 , bovine kidney Ki: 500 (5.0) /3Of].

[HSSl^fflBBB EC: 3.2.1 .21 , aspergillus wentii Ki: 1150 (5.0) [301].

Ki: 450 (5.0), 330 (6.0), 74 (7.0) [301].

K,: 80 (6.5) /3Of; ||'.-D-fructofuranosidase (inverlase. sucrase) f?*fldkJUW?ffliB^^^^^^^^M

ΒϊΒΠΙΒΒΗΙΠίΒΗ EC: 3.2.1.20, yeast

K8: 2.8 (6.5) [11 9], K1: 2,8 (6.8) [291].

IBBlllHBSBHSil EC: 3.2.1.20, yeast brewer's

iBifftl&HgHEEHg EC: 3.2.1 .22, green coffee beans K1: 50 (6.5) [119], Id: 50 (6.8) [291].

NJ

K)

Oo

S'

J50

Si

^rr^ ^?

1

*· S ^;

^l < r^

Oo

270

Tables of Glycosidase Inhibitors

1

NH(CH2)3NH2

Hm Vj

Γ

I

OH

1— 'CH,

Π Η

ΗΟ

OH

r

r\"y( H^I

PLNHAC OH

"Ί Η

ΗΟ

r HO/I

°— ] H N Lx \

H

1 -(3-Aminopropyl)-amino-1 ,2,5trideoxy-2,5,-imino-D-mannitol chiral pool chemical Compound: a046 Synthetic/NMR: /299J1 [278] 1 -2,5-Trideoxy-2,5-imino-Dmannitol de novo chemo-enzymatic Compound: a047 Synthetic/NMR: /f 48]

Synthetic/NMR: [1541 [278], [299]

_chemical __.—_--_-

rhiral iwil phprnn-pnTvinstip Hiiral nnnl

1-Acetamido-1,2 5-trideoxy-2,5imino-D-mannitol

TBBWtM-WRT-E! EC: 3.2.1 .51 r bov/ne Wci/ie/ Kc 22 (5.5) 048/

UBiJ'lUIildEES EC: 3.2.1 .21 r aqrobactenum faecalis Ki: 15 (7.0) /299/

Ik'P^B'fl'll'ffPHBOFBf! EC: 3.2.1 .30. bov/ne /rWney Ki: 9.8 (6.5)/f54|

EC: 2.7.1. 90, (n.g) IC:48%50ppm(n.g.)/f69;. HHiHBBI EC: 3.2.1.28, corynebacterium sp. ICso: 0.35 (n.g.) /209/ il/jiBKHJ EC: 3.2.1 .28, plutella xylostella ICso: 10 (n.g.) [209/ ^JiHIHJ EC: 3.2.1 .28, porcine kidney ICso: 200 (5.8) [175]. UfRHISBiEC: 3.2.1.28, rat intestine ICso: 360 (5.8) [205J, /)75/ U^uBtBiff^HM^IiiiriWBCB EC· 3 2 1 30 /ac/c beans K: 1.9 (6.5) 054/. UJSJ^^J^ bC: 3.2.1 .21 r aqrobactenum taecahs K8: 25 (7.0) /299/.

|PvroDhosDhate-D-fructose-6-DhosDhate-1-ohosDhotransferase (PFP)^^^^^^H

1C»: 91 (5.8) /205/, ρ 75;.

Ki: 1.7 (5.0) [3Of/, K: 7,8 (6.8) /29fJ, Ki: 160 (5.0) /273/. ΙέΒ!4^!>Ηί^.Μ^1Μί;Γιίΐ!φΒ!^-!·Η!>ίΚ;ΙΙ EC: 3.2.1 .37. asperaillus niaer ICso: 250 (5.0) [024].

ltflig«1l"f^"«"?f»n"ra EC: λ.2.1 .21 r sweef almonds

S"

bO ^ h— k

O

OQ

^

.? Oi ^ M S

51

U.

272

Tables of Glycosidase Inhibitors

CH, I

ι— NH2

Π

ΗΟ

V^IOH OH

TnVn

Ν

H3<W HO OH

H

O0H3

\OHHO_

OH

ΤΝΗδ> Y_Jfl V— -'"—OH

HO-,

2,5-Dideoxy N-methyl-2,5-iminoD-mannitol 2, 5-Dideoxy-2, 5-(methyliminiumyl)-D-mannitol modification of natural product Compound* 3053 Synthetic/NMR: [2057 1,2,5-Trideoxy-2,5-imino-D-talitol (2R,3R,4S,5R)-5-Hydmxymethyl-3,4dihydroxy-2-methylpyrrolidine chiral pool chemical Compound: a054 Synthetic/NMR: /275] 1 -Amino-1 ,2,5,6-tetradeoxy-2,5imino-L-talitol chiral pool chemical Compound: a055 Synthetic/NMR: [210] 5-Amino-5-deoxy-D-fructose 2-N-Dehydm-2,5~dideoxy-2,5-imino-Dmannitol de novo chemo-enzymatic Compound: a056 Synthetic/NMR: [291] EC: 3.2.1.26. rat intestine

HBPEBHE EC: 3-2.1 .20, yeas/ brewer's Ki:2.6±0.2(6.8)/29fJ.

Ιϋ·Ι^Μ!;Μ^>ΗΤ»ΕΕΒ EC: 3.2.1.23. e. co// Ki:276±51(6.8)/291].

IHUiHBBHfflBB EC: 3.2.1 .21 T sweef almonds Ki:13±0.7(6.8)f29f/ BBHB8H8H.B EC: 3.2.1 .51 r bov/ne epididymis Ki:381±71(6.0)/29y.

HiJJiEIS]JEHESE EC: 3.2.1 .24, jack beans Ki:17±1(6.8)/29fJ.

JBKflfAHBRcB EC: 3.2.1 .51 , bovine kidney Ki: 1.9 (5.5) PiOJ.

IBBUBiHBHSi EC: 3.2.1 .51 , bovine kidney ICso: 9 (5.5) /275|

UBiim.hliM5fBET.fg EC: 3. 2. 1.24, /ac/f heans ICso: 53 (4.5) /275J.

ICso: 190 (6.5) /205/.

BfflBHilBiHBliffi tC: :^^.1 'M1 rat liver soluble

ItiiBfffjHfflTlffiHiniHiB^^ ICso: 620 (5.8) [205].

ΗίΙ'ΙΙΙΙΨΗΕΕΠ? EC: 321 .20, rat liver lysosomal ICso: 470 (6.8) /2057.

274

Tables of Glycosidase Inhibitors

3.1 Six-Membered Rings, C5

o 0)

S

X CO

CO

CO

Inrinosiigars as Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

275

276

Tables of Glycosidase Inhibitors

3.1 Six-Membered Rings, €5

277

278

Tables of Glycosidase Inhibitors

3.1.3

NH

'

NH +

^=^ / \

v^~ °

2

N 2

\V""\_/~~

HoLfl ^ \_

r-OH

H(N-H

M

/1V-Y-^-

1 V V \_/~~" ~\_/

'

ΝΗ +

/R-V=V

fv_/~»~\-P

1

H

H

/ Vr |v_/~~

1

Pyrimidines

de novo chemical Compound: b018 Synthetic/NMR: [216] (4R,5S)-5-Hydroxy-4-(hydroxymethyl)-2-(p-aminophenyl)amino1 ,4,5,6-tetrahydropyrimidine de novo chemical Compound: b019 Synthetic/NMR: [216] (4R,5R)-5-Hydroxy-4-(hydroxymethyl)-2-(p-aminophenyl)amino1 ,4,5,6-tetrahydropyrimidine de novo chemical Compound: b020 Synthetic/NMR: [216] (4R,5S)-5-Hydroxy-4-(hydroxymethyl)-2-(p-nitrophenyl)amino1 ,4,5,6-tetrahydropyrimidine de novo chemical Compound: b021 Synthetic/NMR: [216] (4R,5R)-5-Hydroxy-4-(hydroxymethyl)-2-(p-nitrophenyl)aminoI A f\ R.tptr^hwHrnnwriiniHinp de novo chemical Compound: b022 Synthetic/NMR: [216]

(4R,5S)-5-Hydroxy-4(hydroxymethyl)-2-methylamino1 ,4,5,6-tetrahydropyrimidine

l\i. oUU± I UU (\>.O) [ει O].

if · Iflftj-lfifi /β fl\ F)IRl

luiffilllPAHBBfB EC: 3.2.1. 20r yeast Ki:6000±1000(6.8)/2f6/.

Hiil!IBRHi!Effi EC: 3.2.1 .21 . sweet almonds Κι:30±10(6.8)/2ί6/.

IBiIjllifJ>H!«R:gl EC: 3.2.1 .20r yeast Ki:2000±1000(6.8)/2f6/.

kflifl.'MET4?«CT»ET"fl EC: 3.2.1.23. e. coll Ki:80±10(6.8)py6J.

EC:

miff'THIiHEEE 3.2.1 .21 T sweet almonds Ki:2000±1000(6.8)/2f6J.

ItIiItIIIfAHBB^I EC: 3.2.1 .21 r sweet almonds Ki:1800±200(6.8)i2f6J.

ro

^l

C<5

OQ

S'

a (^^ 5-X *2w ^

I

R-

Cn

I— ι

Oj

280

Tables of Glycosidase Inhibitors

H

I

π

h =N _y K~ \_ \_r/ ~°

j—OH

H

Vl?

^

ύΗ

"Λ.

^-fi

ΠΗ

f\.^y^

H<

de novo chemical compound: oc/zy Synthetic/NMR: /256]

2-[N-(Benzyloxy)amino]4,5-dihydroxy-6-(hydroxymethyl)tetrahydro-1 ,3-pyrimidine

α,β-D-glucopyranose de novo chemical Compound: Jb027 Synthetic/NMR:/2)6/ 2-(Benzylamino)-4,5-dihydroxy-6(hydroxymethyl)-tetrahydro-l ,3pyrimidine de novo chemical Compound: 6028 Synthetic/NMR: /256/

tptrahvHrnnvrimiHin-9-vl1aiT.inn-

4-Deoxy-4-[(4R,5R)-5-hydroxy-4(hydroxymethyl)-l ,4,5,6-

Ki: 13.2 (7.5) /256/

IHRMTHfflmemPMI co. ο ο Λ oo

ΊΙ

ΒΒιΙ·ΜΕΜΜ3ΒΕΗ:! EC: 3.2.1 .22. qreen coffee beans Ki: 6.4±0.9 (7.5), 15.2±2.8 (6.0) /256/ |iiiffiM!:W>HBEEa EC: 3.2.1 .23, aspergillus niger Ki: 24.7 (7.5) /256/.

^tf^HifiEIFffl EC: 3-2.1 .22, Qreen coffee beans Ki: 527 (7.5), 151 (8.8), 28.1 (10.5), 6.3 (12.0), 4.9 (13.0) /256/. IH3PBBS5SE3T! EC: 3.2.1 .23t aspergillus niger K: 62 (7.5) /256/.

Κι:10000±2000(6.8)/2ί6/. IBiBHHSuHIiHSJI EC.· 3.2 1.23, e. coil Κι:5500±1000(6.8)/2ί6|

ItliftlHfiBHBtcgB EC: 3.2.1 .21 , sweet almonds

OQ

'

I

282

Tables of Glycosidase Inhibitors

O

cE

X CO

01

cr>

Iminosngars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. StiHz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

3.2.2

HO

OH

H

UA sL)

r-OH

r

H3»|B^EEffi EC: 3.2.1.21 . calf liver cytosolic Ki: 570 (7.0) /067/.

1 ,5-Dideoxy-1 ,5-imino-D-galactitol THiWrCTuJjyjQE^^ EC: 3.2.1 .22 qreen coffee beans Ki: 0.016 (n.g.) [227], Ki: 0.0016 (6.0) [21O]. 1 -Deoxygalactonojirimycin, Galactostatin Compound: c003 I /-D-cialactosidase (ceramide trihexosidasel-MRfOT 22, human placenta Synthetic/NMR: [067], [227], [187] ICso: 0.004 (n.g.) /227/.

D-galacto-Configuration

ΚΒΒΗΗΠ8ΗΒΗ3Ι EC: 3.2.1 .23r rat liver lysosomal IC5o:740(opt)/287/.

iHiJJMEMRHffET.5 EC: 3.2.1.23. bovine liver Ki: 1.5 (4.0) /775J1IC50: 3 (opt) [287].

OQ

δ?S'

I

284

Tables of Glycosidase Inhibitors

UJ LU UJ H. UJ S _. ·πΐ-^· κπ^· κη ^~> KtI

UJ . UJ ν.τιΓ?-Γ·τπ

1-Deoxy-D-galactonojirimycinuronicacid chiral pool chemical Compound: c007 ynthetic/NMR:

galactonic acid

2,6-Dideoxy-2,6-imino-L-

N-Heptyl-1-deoxygalactonojirimycin Compound: c006 Synthetic/NMR: [043], [067]

1,5-Dideoxy-N-heptyl-1,5-iminoD-galactitol

N-Butyl-1 -deoxygalactonojirimycin modification of natural product Compound: c005 Synthetic/NMR: [187]

N-Butyl-1,5-dideoxy-1,5-imino-Dgalactitol -1,2-qlucosidase i

EC: 3.2.1.23, e. co//

iC: 3.2.1.21, calf liver cytosolic

Ki: 0.75 (6.6) /096/

| HppSEmEp EC: 3.2.1.22, green coffee beans

Κ: 0.6 (7.0) /067/

K1:1.6 (7.2)/043/

BiBBlHBBHBISHiEC: 3.2.1.22, e. co//

K8:0.26 (7.0)/043/

Ki: 47 (5.0) /043/

IHBBIBBIiBBIJHJi EC: 3.2.1.21, sweet almonds

JMIPBMImWHaBI raf ^er microsomes ICw: 2130 (n.g.)/787/ I EC: 3.2.1.22, green coffee beans Ki: 3.2 (6.0)/043/

submaxillary glands IC50:41.38 (n.g.)/W/

IC5o:2130(n.g.)/W/

286

CM CO

Tables of Glycosidase Inhibitors

CM CO

. CM CO

. CN4 CO

. CM CO

UJ^-UJcoUJ

. CNJ CO

.UJ

. CM CO

.

LJJ ^,

. CM CO

^^ CM CO

UU ^-

LU

. CM CO

. CM CO

.UJcB1UJ

. CM CO

. CM CO

^^ C^ CNj COSTCO

CM..

· CM CO

UJcoUJco'UJcoUJ^^UJ

J-(J

OH

Λ

J-N

L

\| /H

KOH

HO

pOH

OH

I

Μ

2

Mil

^y-NH2

^aT V-MH

HOJ-N

pOH

OH

I

V-T

w)=°

HO

ι

1-Deoxygalactonojirimycinamidrazone chiral pool chemical Compound: c012 Synthetic/NMR: [310]

1 -Amino-1 ,5-dideoxy-1 ,5-imino-Dgalactitol amidrazone

1-Deoxygalactonojirimycinamidine chiral pool chemical Compound :c011 Synthetic/NMR: [310]

1 -Amino-1 ,5-dideoxy-1 ,5-imino-Dgalactitol amidine

D-Galactonojirimycinlactam de novo chemo-enzymatic Compound: c010 Synthetic/NMR: /065]

5-Amino-5-deoxy-Dgalactonolactam

BiilJfclfcMJ^HgEgB EC: 3.2.1.22. qmen coffee beans Ki:8.3±0.4(6.6)/3f07,/227;. BBBIBIBHTfBH!! EC: 3.2.1 .21 , sweet almonds Ki:2.4±0.7(5.6)/3fOy. [EBBHHSBHIIHB EC: 3.2.1.23,fcov/neliver Ki: 6.5±0.1 (7.0) /3f0/, Ki: 6.5±0.1 (4.5) [227].

BBiItKIBBR1HgB-B EC: 3.2.1.22r green coffee beans Ki:8.5±0.4(6.6)/3fOJ,/227J. IBiBIBBiISIIIBS EC: 3.2. 1 .21 r sweet almonds Ki:25±5(5.6)/3fO;.

EC:

ΙΙΜ'^ΙΙίΙϋίΒΕΕΕ 3.2.1 .23. peniciHium multicolor Κι:13.1±0.8(6.0)/0β5/

IC50: 0.164 (7 .2) [166]. BBBfItIBHBBBS EC: 3.2.1 .21 . asperqillus wentii Ki:400(5.0)/043J. IBiBIBHBIBu EC: 3.2.1 .21 T sweet almonds Ki: 22.4 (5.0) /043/

IBiBHHSHSIIHS! EC: 3.2.1 .23T saccharomyces fraqilis

IC50: 0.944 (4.5) [1 66].

IC50: 1.17 (4.5) [166]. ^PfffSSPlfSl^^C: 3.2.1. 23r rat spleen

IMiI»MP;WRHgg".a EC: 3.2.1 .23. rat liver lysosomal

288

Tables of Glycosidase Inhibitors

H

HO

OH

β> N£_y Γ

COPh

(CH2JtCOOCH3


φ

1 ,5-Dideoxy-N-methyl-1 ,5-imino-Lfucitol N-Methyl-1-deoxy-L-fuconojirimycin chiral pool chemical Compound: c016 Synthetic/NMR: /089] N-(5-Carboxypentyl)-1 ,5-dideoxy1,5-imino-L-fucitol N-(5-Carboxypentyl)-1-deoxy-Lfuconojirirnycin chiral pool chemical Compound: c017 Synthetic/NMR: [054] 1,5-Dideoxy-N-(5methoxycarbonylpentyl)-1 ,5imino-L-fucitol N-(5-Methoxycarbonylpentyl)-1-deoxy-Lfuconojirimycin chiral pool chemical Compound: c018 Synthetic/NMR: [054], [106], [055] N-Benzoyl-1 ,5-dideoxy-1 ,5-iminoL-fucitol N-Benzoyl-1-deoxy-L-fuconojirimycin Compound: c019 Synthetic/NMR: /0547, [313] 5-Amino-5-deoxy-L-f ucono-1 ,5lactam L-Fuconic-S-lactam chiral pool chemical Compound: c020 Synthetic/NMR: [31 2]

bovine epididymis

EC:

epididymis

Ki: 400 (5.5) [312].

IRHIIRRBI-RiEC: 3.2.1.51, human liver, neutral

bovine kidney

UBBilMtHtiEEEIEC: 3.2.1.51. bovine ICso: 0.31 (6.0) /089J.

K1: 300 (5.5) [31 2]. '

epididymis

epididymis 3-2.1 .51 bovine .

ICso: 0.48 (6.0) [089].

SMllH'HI'Jii=]

ΒΒΒ1ΙΜ·Η[»Ε".ΕΙ EC: 3.2.1 .51 .bovine ICso: 0.45 (6.0) [089].

K1: 1.1 (4.5) [f 07/

IiBIHiHBIBI EC: 3.2.1 .51charonia . lampas

Ki: 0.05 (5.5) ^ 07/.

ICso: 0.475 (6.0) [089]. URIfRRBEUEC: 3.2.1.51. human liver, neutral

N) OO

290

Tables of Glycosidase Inhibitors

I

Ii (O O

ο

δ ** Ol

NHAc

τ

HO — ι

NHAc

NHAc

p^/^OH

>

Η0

EC:

8

r

HII

2-AcetamidQ-5-3mmo-? ^-dide^yy- 1 IJl^lliffllFAlMUMK5ffi EC: 3.2.1.30, bovine kidney D-glucono-1,5-lactam 2-Acetamidonojirimycinlactam modification of natural product Compound: c027 Synthetic/NMR: /124/

^^^^^^^^^^^^^^^^^^^•iN-Ac-n-D-Qiucosamimd K1: 1.2 (4.0) /074/.

1

UUBRfSHJ^B^EH 3-2.130. bovine kidney 1 ,2,5-Trideoxy-2-N-benzamidoKi: 100 (4.3) /264/. 1,5-imino-D-glucitol 2-N-Benzamido-1,2-dideoxynolirimycin chiral pool chemical Compound: c024 Synthetic/NMR: /264/ (1H,«C) ["RiBtBifltliTBBKTiliiTfifB ?H3 EC: 3.2.1. 30 bovine kidney 2-Acetamido-1 ,2,5-trideoxy-N,NK : 4.0 (4.25) [124], K : 0.56 (6.25) [124]. dimethyl-1 ,5-imino-D-glucitol iodide 2-Acetamido-N,N-dimethyl-1,2dideoxynojirimycin modification of natural product Compound: c025 1 Synthetic/NMR: /124/ ( H) 2-Acetamido-5-amino-2,5-dideoxy- IJR!B8ililHim»l-f:1'i'iiri!t«gHg EC: 3.2.1 .30. jack beans Ki: 0.001 2 (5.0) /074/. D-glucose fTKfefflinMiTwrtSnmiii! S3 EC: 3.2.1.30, bovine kidney 2"AcetamidO'2-deoxynojinmycin 0.003 (4.3) /074/. Ki: 0.002 (4.25)^24/, K: chiral pool chemical Compound* c026 Ιι.ΒίΜ·Β·-ΡΙ1ιΐΜ^Πιιιιικ SlEC: 3.2.130, helix pomatia Ki: 0.5 (5.0) /074/. Synthetic/NMR: [124], [074] S EC: 3.2.1.30. aspergillus niqer

K)

S.

INJ

292

Tables of Glycosidase Inhibitors

ISHBEB EC: 3.2.1 .20, mouse small intestine ICso: 0.14-0.074 (6.0) /064/ IBHIHSS EC: 3.2.1 .20, rabbit small intestine ICso: 1.0 (6.0) /080/ ICso: 0.23 (6.0) /064/ IC50: 0.17Mg/mL (6.0) /064/ ISHIHSB EC: 3.2. 1 .20, rat small intestine ICso: 0.63-0.17(6.0) /064/ IBHIHSS EC: 3.2.1 .20, rhesus monkey small intestine ICso: 0.1 (6.0) /064/ HBBHSSSiHIBHHHSSSBSB calf pancreas microsomes ICso: 2 (6.8) [01 2/ ISIHfHS EC: 3.2.1 .48, beagle dog small intestine ICso: 0.12 (6.0) /064/ ISIHHSS EC: 3.2.1 .48, guinea pig intestinal ICso: 0.57 (6.0) [219]. ISBSfHS EC: 3.2.1 .48, mouse small intestine ICso: 0.14-0.098 (6.0) /064/ HIHfUB EC: 3.2.1 .48, rabbit small intestine ICso: 0.41 (6.0) /080/ ICso: 0.16 (6.0) /064/ ICso: 0.067Mg/mL (6.0) /064/ Ki: 0.115±0.005 (5.85) /009/ Ki: 0.032±0.004 (6.8) /009/ Ki: 0.022±0.001 (7.45) /009/ SBSfHS EC: 3.2.1 .48, rat small intestine ICso: 0.12-0.066 (6.0) /064/ Ki: 0.024 (6.9) /274/ 3IHfBB EC: 3.2. 1 .48, rhesus monkey small intestine ICso: 0.12 (6.0) /064/ UHiHISSS EC: 3.2.1.28, porcine kidney ICso: 41 (opt) [185]. BHHHHSSSBHSS rabbit skeletal muscle ICso: 1.7 (6.4) /085/ ICso: 1.8 (6.0) /085/ 1C»: 1.4 (6.5) /085/ IC50: 2.2 (7.0) /085/ ICso: 2.5 (7.5) /085/ HBIuHSSSBHSS EC: 3.2.1 .20. bacillus steamthermoohHus Ki: 0.44 (6.8) /273/ IiUiIHSSSBHSS EC: 3.2.1.20, calf liver microsomes Ki:2(n.g.)/020/

ξ?

£

U)

to

OS

OQ

S*

oS

c-Membei

294

Tables of Glycosidase Inhibitors

lift^lbUitkl^lfnBTAIEi^^^ftklj^ EC: 3.2.1 .37, asoeraillus niaer ICso: 400 (5.0) /024/

IHBBIIBBHBHSB EC: 3.2.1 .21T sweet almonds K1: 47±7 (6.8) /229J1 Ki: 1700 (5.0) /273/ Ki: 300 (5.0), 38 (6.0) /296/ IC50: 200 (opt) [185]t ICso: 24.OMgMt (5.0) [139], Ki: 76 (5.6) [112], IC50: 272Mg/mL (4.5) /064/ K1: 47 (n.g.) /020/ Ki: 29 (6.2) /023/ K: 18 (5.6) /3fO/ IC50: 81 (4.8) /024/

IHBBIBHHIiBffl EC: 3.2.1 .21 . helix pomatia Ki:60(n.g.)/020/

IBEHHHH EC: 3.2.1 .21 r calf spleen lysosomal K1: 390 (4.6) /033/.

I lifliltlUUM^lJEC: 3.2.1.21. calf liver cvtosolic K8: 210 (7.0) [067].

296

Tables of Glycosidase Inhibitors

IcH3

I ^i

OH

/^TC/ 0'"7

Γ°ΗΟΗ

OH

Jw

C-C2H5

OH

HONL/

"Ι

I— OH 3 X ) N-CH3 /ΓΛ4 *\

Synthetic/NMR: /082], [067], [033]

Pnmnoiinri* rfiW

1 ,5-Dideoxy-N-propyl-1 ,5-imino-Dglucitol N-Propyl- 1 -deoxynojirimycin modification of natural product

r

^


^o

^ l·^

OQ IHiBlItHiHBlSHI EC: 3.2.1 .21 . calf spleen lysosomal Ks: 210 (4.6) /033/

VA. So

S

^

<3-j

3

?

k>

Uj

fAVJMif4!4Jyhf •ILVMiTA^i^l'iI^j^^Jrg ECi 2.4.1 .19, bddllus stesrothermophilus ICso: 290Mg/mL (5.5) /082|

ICso: 6OMgMiL (5.7) /082|

ltiifl»lll[«?»HNrT^EC: 3.2.1.21. calf liver cvtosolic Ki: 3500 (7.0) [067].

ICso: 700Mg/mL (5.5) /082/.

BTBBBMfBiEElRg^^ EC: 3.2.1 .13. rhizopus niveus ICso: 6.9Mg/mL (5.7) [064], ICso: 25Mg/mL (4.5) /0647, ICso: SOpg/mL (5.7) /082/

50

r

WIIMtHNkI-Ml calf liver microsomes 1 ,5-Dideoxy-N,N-dimethyl-1 ,5IC50: 0.4 (n.g.) [021], Ki: 0.4 (6.8) [048]. imino-D-glucitol BIIBBHBEHiI pig liver crude microsomes N,N-Dimethyl-1-deoxynojirimycin Ki: 0.4 (6.4) [086]. modification of natural product, chiral pool chemical Compound: c030 Synthetic/NMR: [123], [086], [021], [048] 1 ,5-Dideoxy-N-methyl-1 ,5-imino-D- BBiIJIIiMtHgEHg EC: 3.2.1.20. yeast Κι:>10000(6.5)/ί22| glucitol-N-oxide fBT!PIlH?5»l!ffl EC: 3.2. 1 .21 sweet almonds N-Methyl-1-deoxynojirimycin-N-oxide Κι:80(6.5)Ρ22/. chiral pool chemo-enzymatic Compound: c031 13 Synthetic/NMR: /f 22J(H C) 1 ,5-Dideoxy-N-ethyM ,5-imino-D- llflifWHAHEEEB EC: 3.2. 1 .21 sweet almonds IC : 395Mg/mL (4.5) [064]. glucitol 8IHHB EC: 3.2.1 .48, rabbit small intestine N-Ethyl-1-deoxynojirimycin ICso: O.ISpg/mL (6.0) [064]. modification of natural product BflffiBS EC: 3.2.1 .20, rabbit small intestine Compound: c032 1C»: 0.42Mg/mL (6.0) [064]. Synthetic/NMR: /064/ [082]

298

Tables of Glycosidase Inhibitors

Compound: c040 Synthetic/NMR: /064/ [082/

N-Hexyl-1-deoxynojirimycin ptoluenesulfonate modification of natural product

1,5-Dideoxy-N-hexyl-1,5-imino-Dglucitol p-toluenesulfonate

Compound: c039 Synthetic/NMR: /033/ /048/ /067]

N-Hexyl-1-deoxynojirimycin modification of natural product

1,5-Dideoxy-N-hexyl-1,5-imino-Dglucitol

N-Pentyl-1-deoxynojirimycin ptoluenesulfonate Compound: c038 Synthetic/NMR: /082/ [151]

1,5-Dideoxy-N-pentyl-1,5-imino-Dglucitol p-toluenesulfonate

Compound: c037 Synthetic/NMR: /064]

N-lsobutyl-1-deoxynojirimycin ptoluenesulfonate modification of natural product

N-lsobutyl-1 ,5-dideoxy-1 ,5-iminoD-glucitol p-toluenesulfonate

£: 2.4.1.19, bacillus stearothermophilus

ICso: >1000Mg/mL (5.5)/082/

I EC: 3.2.1.21, sweet almonds IC50:99Mg/mL (4.5)/064/ HIHHSS EC: 3.2.1.48, rabbit small intestine ICso: 0.29Mg/mL (6.0) /064/ 3.2.1.20, rabbit small intestine ICso: 5.5Mg/ml_ (6.0)/064/ B EC: 3.2.1.13, rhizopus niveus ICso: 145Mg/mL (5.7) /064/ ICso: 3.2Mg/mL (4.5) /064/ ICso: 120Mg/mL (5.7) /082/

Ki: 0.13 (6.8)/048/ EC: 3.2.1.21, calf spleen lysosomal Ki: 69 (4.6) /033/

calf liver micmsomes

K,: 200 (7.0) /067/

I EC: 3.2.1.21, calf liver cytosolic

ICso: >1000Mg/mL (5.5) /082/

ICso: 108Mg/ml (5.7) /082/

lalucoamvlase (exo-1 4-.

glucoamylase (exo-1 4ICso: 200Mg/mL (5.7) [064], IC: 16UMg/mL (4.5) [064].

j EC: 3.2.1 .48, rabbit small intestine IC50: 1.3Mg/mL (6.0) [064]. BHUSHS EC: 3.2.1 .20, rabbit small intestine ICee:4.9Mg/mL (6.0) /064/.

K)

^ Co

300

Tables of Glycosidase Inhibitors

ΤοβΗ

f

f

/^12H26

TosH Η0(T / I^

I

OH

H$-f

CX*-

HO^

/^""^TosH

ΗΟ

Ν

ΛπΚ

^™ITw»^[»Kfl*iflM[ll»fi»^

IQQPIQQS^Q EC: 3.2.1.21, calf spleen lysosomal Κ: 0.19 (4.6) /033/.

[«IfflJHSiHH calf liver microsomes ICso: >500(n.g.) /027/.

HBBBBEBI calf liver microsomes IC5o:8(n.g.)/02f/.

IMlgWIM.»[*feH5EC: 3.2.1.21r calf liver cytosnlir. Ki: 3.8 (7.0) /067/.

HIMifcHj EC: 3.2. 1 .48, rabbit small intestine IC5o:0.40Mg/mL(6.0)/064/. ϋΙΚΠΟΜ EC: 3.2.1 .20, rabbit small intestine 1C»: 4.4Mg/mL (6.0) /064/ NHIM»!:h'iMkUJ^>lifilliI>lll[j«UgggHI EC: 3.2.1.13, rhizopus niveus ICso: 145Mg/mL (5.7) /064/, IC: 165pg/mL (4.5) /064/.

j^X^iiMiL^iaibfliR EC: 3.2.1.13. rhizopus niveus ICso: 68Mg/mL (5.7) /064/, IC50: 44Mg/mL (4.5) /064/

HiHifefcH EC: 3.2.1 .48, rabbit small intestine ICso: 0.31 Mg/mL (6.0) [064], IBHnUSB EC: 3.2.1 .20, rabbit small intestine ICso: 3.2Mg/mL (6.0) /064/.

modification of natural product Compound: c048 Synthetic/NMR: [064]

EC: 3.2.1.45, human placenta Ki: 0.0076 (6.45), IC50: 23 (6.5) [1 09]. 1 ,5-Dideoxy-N-dodecyl-1 ,5-imino- MiHEHJ EC: 3.2.1 .48, rabbit small intestine ICso: 0.81 pg/mL (6.0) /064/. D-glucitol p-toluenesulfonate ffiHBHSI EC: 3.2. 1 .20, rabbit small intestine N-Dodecyl-1 -deoxynojirimycin pICso: 1.8MgAnL (6.0) /064/ toluenesulfonate

N-Decyl-1 ,5-dideoxy-1 ,5-imino-Dglucitol p-toluenesulfonate N-Decyl-1 -deoxynojirimycin ptoluenesulfonate modification of natural product Compound: c045 Synthetic/NMR: /0647 1 ,5-Dideoxy-N-undecyl-1 ,5-iminoD-glucitol p-toluenesulfonate N-Undecyl-1 -deoxynojirimycin ptoluenesulfonate modification of natural product Compound: c046 Synthetic/NMR: /064] 1 ,5-Dideoxy-N-dodecyl-1 ,5-iminoD-glucitol N-Dodecyl-1 -deoxynojirimycin chiral pool chemical, modification of natural product Compound: c047 Synthetic/NMR: /027/ /033/ [109], /067/

O

UJ

£

^a

Oq

S'

>3

S-X

^ S

3

O.

a;

on ?'

k>

Uj

302

Tables of Glycosidase Inhibitors

C2H4O-

N-Cinnamyl-1 -deoxynojirimycin hydrochloride Compound: c060 Synthetic/NMR: [151], [082]

N-Cinnamyl-1,5-dideoxy-1,5imino-D-glucitol hydrochloride

Synthetic/NMR:/f 5^/082J

ICso: SOpg/mL (5.5) /082/

ICso: 190Mg/mL (5.5) /082J.

· >

,np97 W

ια»:8θμ9/πΐ (5.7) [082].

9

/Γτ1(5 5

1,5-imino-D.glucitol

·

ir ICso 630μ

ICso: 200Mg/mL (5.7) /082/

ICso: 50Mg/mL (5.5) /082/

N-(4-Phenoxybutyl)-1-deoxynojirimycin Compound: c059

N-(3-Phenoxypropyl)-1-deoxynojirimycin p-fo/i/enestv/fonaie Compound: c058 Synthetic/NMR: [151], [082]

1,5-Dideoxy-N-(3-phenoxypropyl)1,5-imino-D-glucitol p-toluenesulfonate

N-(2-Phenoxyethyl)-1-deoxynojirimycin Compound: c057 Synthetic/NMR: [151], [082]

1,5-Dideoxy-N-(2-phenoxyethy I)1,5-imino-D-glucitol ICse: 300g/mL (5.5) [082J.

EC: 2.4.1.19, bacillus stearothermophilus

EC: 3.2.1.13, rhizopus niveus

I EC: 2.4.1.19, bacillus stearothermophilus

EC: 3.2.1.13, rhizopus niveus

': 2.4.1.19, bacillus stearothermophilus

lEC: 3.2.1.13, rhizopus niveus

I EC: 2.4.1.19, bac///us stearothermophilus

B EC: 3.2.1.13, rhizopus niveus

1 EC: 3.2.1.21, calf/fvercKtoso/fc Ki: 1700 (7.0) 067/.

N-(N'-Butylcarboxyamidopentyl)· 1 ,5-dideoxy-1 ,5-imino-D-glucitol

N-(N'-Butylcarboxyamidopentyl)-1deoxynojirimycin chiral pool chemical Compound: c056 Synthetic/NMR: [067/

, rhizopus niveus \ EC: 2.4.1.19, bacillus stearothermophilus

ICso: 30OMgMiL (5.5) [082].

alucoamvlase (exo-1 A-( ICso: 176Mg/mL (5.7) [082].

N-(3-Phenylpropyl)-1-deoxynojirimycin Compound: c055 Synthetic/NMR: [151], [082]

1,5-Dideoxy-N-(3-phenylpropyl)1,5-imino-D-glucitol

OQ

y

Cfc

^

§

304

Tables of Glycosidase Inhibitors

1

™

HO—,

HO

(CH2)4Si(CH3)3

/

AH

H2)3Si(CH3)2Ph

/(C

\

^CH2)6Si(CH3)3

OH

(CH2)4Si(CH3)2PH

HO^Y OH

HO—ι

HO- 1

1 ,5-Dideoxy-N-(dimethylphenylsilyl)propyl1,5-imino-D-glucitol N-(Dimethylphenylsilyl)propyl-1deoxynojirimycin modification of natural product Compound: c065 Synthetic/NMR: /274] 1 ,5-Dideoxy-N-(trimethylsilyl)butyl-1 ,5-imino-D-glucitol N-(Trimethylsilyl)butyl-1-deoxynojirimycin modification of natural product Compound: c066 Synthetic/NMR: [274] 1,5-Dideoxy-N(dimethylphenylsily l)buty 1-1 ,5imino-D-glucitol N-(Dimethylphenylsilyl)butyl-1deoxynojirimycin modification of natural product Compound: c067 Synthetic/NMR: [274] 1,5-Dideoxy-N(trimethylsilyl)pentyl-l, 5-imino-Dglucitol N-(Trimethylsilyl)pentyl-1-deoxynojirimycin modification of natural product Compound: c068 Synthetic/NMR: [274] HIHfeHj EC: 3.2.1 .48, rat small intestine K«: 0.033 (6.9) /274/ fHSBHBSHli EC: 3.2.1 .10, rat small intestine Ki: 0.016 (6.9) /274/. BinEBEIiiBlBSB EC: 3.2. 1 .13, rat small intestine Kr O 07 (n g ) /274/

MlMfeHd EC: 3.2.1 .48, rat small intestine Ki: 0.039 (6.9) /274/. BSBiEPHcI EC: 3.2.1 .10, rat small intestine Ki: 0.018 (6.9) /274/.

Ki: 0.12 (n.g.) /274/.

HIHifeHd EC: 3.2.1 .48, rat small intestine Ki: 0.016(6.9) /274/. IHiBBnH-BEC: 3.2.1.10, rat small intestine Ki: 0.04 (6.9) /274/.

MlMifeHJ EC: 3.2.1 .48, rat small intestine Ki: 0.017 (6.9) /274/ BBBHffiBB EC: 3.2.1 .10, rat small intestine K,: 0.027 (6.9) /274/ [•fflSHSHEfflEC: 3.2.1.13, rat small intestine Ki: 0.01 (n.g.) /274/

306

Tables of Glycosidase Inhibitors

/H\

AH

Si(CH3)3

r~°^C2H40-^^-COOCH3

^i

°

H

Si(CH3J3

>=s/^Si(CH3)3

IsT/

°~l

H

OH

Λ>Γ\

HO —ι f y

HO^-f OH

faY^OV,

HO- 1

N-[(4-Methoxycabonyl)-2-phenoxyethyl]-1deoxynojirimycin Compound: c077 Synthetic/NMR: /0857

1 ,5-Dideoxy-N-[(4-methoxycarbonyl)-2-phenoxyethylj1, 5-imino-D-glucitol

N-(4-Trimethylsilyl)phenyl-1-deoxynojirimycin modification of natural product Compound: c076 Synthetic/NMR: [274;

1,5-Dideoxy-N-(4-trimethylsilyl)phenyl-1 ,5-imino-D-glucitol

N-(3-Trimethylsilyl)phenyl-1-deoxynojirimycin modification of natural product Compound: c075 Synthetic/NMR: /274/

1,5-Dideoxy-N-(3-trimethylsilyl)phenyl-1 ,5-imino-D-glucitol

N-(2-Trimethylsilyl)phenyl-1-deoxynojirimycin modification of natural product Compound: c074 Synthetic/NMR: /274]

1,5-Dideoxy-N-(2-trimethylsilyl)phenyl-1 ,5-imino-D-glucitol

1,5-Dideoxy-N-(trimethylsilyl)-Eprop-2-enyl-1,5-imino-D-glucitol N-(Trimethylsilyl)-E-prop-2-enyl1-deoxynojirimycin modification of natural product Compound: c073 Synthetic/NMR: /274]

Πίί'ΙβΗϋΕΕ^ EC: 3.2.1. 20r human liver lysosomal Ki: 2.6 (n.g.) [274]. SBHSH9 EC: 3.2.1 .48, rat small intestine Ki: 0.064 (6.9) [274]. IBiBHBSBH EC: 3.2.1 .10, rat small intestine K,: 0.83 (6.9) /274J. !SnBBHSBH-B EC: 3.2.1.13, rat small intestine Ki: 0.028 (n.g.) /274/ ra>iffj.["fAT3TrET"fl rabbit skeletal muscle IC50: 7. 1(6.4) /085/

aiMifeHj EC: 3.2.1 .48, rat small intestine Ki: 0.055 (6.9) [274]. IBSBiHffiSI EC: 3.2.1 .10, rat small intestine Ki: 11. 5 (6.9) [274]. cnnTOSmMBKiSS EC: 3.2.1 . 13, rat small intestine Kr.1(n.g.)[27-i;.

K,:10(n.g.)/274;.

WWBR-WIUW!! FP · ^ 9 1 1^ rat email /n/oc/ino

HIHifefctj EC: 3.2.1 .48, rat small intestine Ki: 8.5 (6.9) /274/. BSBHffiBB EC: 3.2. 1.10, rat small intestine Ki: 4 (6.9) [274].

MiMfeHJ EC: 3.2.1 .48, rat small intestine K1: 0.22 (6.9) [274].

308

Tables of Glycosidase Inhibitors

/

^"""T

HO- 1

HO

OH

0

^-C9H19

H

000

S^C5HlO

Η<Λ— [

N-(5-Carboxypentyl)-1 ,5-dideoxyN-methyl-1 ,5-imino-D-glucitol N-(5-Carboxypentyl)-N-methyl-1deoxynojirimycin modification of natural product Compound: c081 Synthetic/NMR:023; N-Decanoyl-1 ,5-dideoxy-1 ,5-iminoD-glucitol N-Decanoyl-1 -deoxynojirimycin modification of natural product Compound: c082 Synthetic/NMR:033;,/048; MIINMMUJI calf liver microsomes Ki: 70.0 (6.8) [048]. IBKIBHSHEBB EC: 3.2.1 .21 , calf spleen lysosomal Ki: 65 (4.6) /033/

WllM»HNF.H;ll pig liver crude microsomes Ki: 1.4 (6.5) 023J

O

U)

I

1

I

310

Tables of Glycosidase Inhibitors

Iminosngars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

1,5-Dideoxy-4-0-(a-D-glucopyranosyl)-1,5-imino-D-glucitol 4-0-a-D-Glucopyranosyl-1-deoxynojirimycin modification of natural product Compound: c087 Synthetic/NMR: [185], [080], [151], [082]

1,5-Dideoxy-4-0-(p-D-glucopyranosyl)-1,5-imino-D-glucitol 4-0-J3-D-Glucopyranosyl-1-deoxynojirimycin modification of natural product Compound: c086 Synthetic/NMR: [185] im

EC: 3.2.1.20, r/ce

[ ICso: 46Mg/mL (5.5) /082/

^

IC5o:80(opt)ff85| EC: 3.2.1.20, rice ICso: 0.61 (opt) [185/ :: 3.2.1.21, caldocellum saccharolyticum IC5o:50(opt)/f85;. EC: 3.2.1.20, rat liver lysosomal ICso: 440 (opt) /Ϊ 85/ 3IBBJH3 EC: 3.2.1.48, rafcb/f sma// /nfesi/ne iC5o:22(6.0)/080/ IBHffiBB EC: 3.2.1.20, rabM small intestine ICso: 610 (6.0) /080/ Icvclodextrin alvcosvltransferase. 4-< EC: 2.4.1.19, bacillus sp. alkalophilic |IC5o:50Mg/mL(4.6)ft5f/

I EC: 3.2.1 .21 r sweet almonds

ICso: 560 (opt) [185/ : 3.2.1.28, porcine kidney ICso: 600 (opt) [185/ IuBIIIBiHBEHi EC: 3.2.1.20, raf liver lysosomal ICso: 1000 (opt) /f 85/

[fflSHHfflSffl EC: 3.2.1 .21 , caldocellum saccharolyticum

ICso:22(opt)

δ.S'

OQ &

1

^

312

Tables of Glycosidase Inhibitors

α-D-gluc-O

CH2Ph

OH

|

|

/ — HTosH

—I

α-D-gluc-O

IT /

α-D-gluc-O

HO

f

I

J— N TosH

irVOSH

α-D-gluc-O

HO

glucopyranosyl-N-hexyl-1,5-iminoD-glucitol p-toluenesulfonate 4-O-a-D-Glucopyranosyl-N-hexyldeoxynojirimycin p-toluenesulfonate Compound: c103 Synthetic/NMR: [082] N-Benzyl-1 ,5-dideoxy-4-0-a-Dglucopyranosyl-1 5-imino-Dglucitol p-toluenesulfondte N-Benzyl-4-0-a-D-glucopyranosyl-1deoxynojirimycin p-toluenesulfonate .. ΐ ;, ι . , chiral pool chemical Compound: c088 Synthetic/NMR: /08OJ(1H1^C), [151], [082]

N-(n-Butyl)-1,5-dideoxy-4-0-ct-Dglucopyr3nosyl-1 5-imino-D· ylucitol p-toluencsulfonBte N-(n-Butyl)-4-0-a-D-glucopyranosyl-1deoxynojirimycinp-toluenesulfonate chiral pool chemical, chiral pool chemoenzymatic Compound: c101 1 13 Synthetic/NMR: [080] ( H1 C), [1511 1082] 1 5-Dideoxy-4-0-a«Dglucopyranosyl-N-pentyl-1 ,5imino-D-glucitol ptoluenesulfonate 4-O-a-D-Glucopyranosyl-N-pentyldeoxynolirimycinp-toluenesulfonate Compound: c102 Synthetic/NMR: [082] 1 5-Dideoxy-4O-oc-D-

EC: 2.4.1.19, bacillus sp. alkalophilic ICso: 105Mg/mL (4.6) [151]. -»------— -^—-— ---—" —a·—-»-— 1 ^πΓ.ΤϊΠΠ^ΙιΙ.|^^1^Ι.Η^^^ bC: 2.4.1 .19. baaHus steamthermoohilus ICso: SOOpg/mL (5.5) [082].

ktlHfcHj EC: 3.2.1.48, rabbit small intestine ICso:23(6.0)i080J.

hffiHtMAaMUKIfflrftMmifeliHgggS EC: 2.4.1 .19. bacillus stearothermoohilus ICso: 180Mg/mL (5.5) /082/.

WAlXiWUlItWmWMt^GnRR EC: 2.4.1 .19. bacillus stearothermophilus ICso: 260Mg/mL (5.5) [082].

HIHfeHj EC: 3.2.1 .48, rabbit small intestine IC50: 49 (6.0) /080/ Icvclodextrin cilvcosvltransferase, 4-( /-(1 ,4-(x-alucano)-transferase^^^^^^^^l EC: 2.4.1.19, bacillus sp. alkalophilic IC5o:23Mg/mL(4.6)ff5f;. Β^ΜΜίΕκιΙιΙ riMffiHffiiffiiiHtcfpS! bC: 2.4.1 .19. bacillus stearothermophiius ICso: 260Mg/mL (5.5) [082].

U)

I— -

U)

is^

Cn

OQ

S"

^H

3.2 Six-Memberec

314

Tables of Glycosidase Inhibitors

si!!!

OH

J-N

—I

aDgtaJ^V OH

Ho

^,JV

)— H TosH

a-D-gluo-O

f~\

C2H4OH

F,H,—

f£)

J-X

HO- 1

Compound: c093 Synthetic/NMR: /0807(1rVC), [151], [082] 1 ,5-Dideoxy-4-0-a-Dglucopyranosyl-N-(2-hydroxyethyl)-1 ,5-imino-D-glucitol 4-0-a-D-Glucopyranosyl-N-(2-hydroxyethyl)1-deoxynojirimycin chiral pool chemo-enzymatic Compound: c094 Synthetic/NMR: [151], [082]

nn 7UiYi Q tip

1 ,5-Dideoxy-4-0-a-D-glucopyranosyl-N-(3-phenyl-propyl)1,5-imino-D-glucitol 4-0-a-D-Glucopyranosyl-N-(3-phenylpropyl)1-deoxynojirimycin chiral pool chemical, chiral pool chemoenzymatic Compound: c092 1 3 Synthetic/NMR: /0807( H1I C), [151], [082] N-Cinnamyl-1 ,5-dideoxy-4-0-a-Dglucopyranosyl-1 ,5-imino-Dglucitol p-toluenesulfonate N-Cinnamyl-4-0-a-D-glucopyranosyl-1deoxynojirimycin p-toluenesulfonate chiral pool chemical, chiral pool chemoEC: 2.4.1 .19, bacillus stearothermophilus

Icvclodextrin qlvcosvltransferase. 4-
ICso:38Mg/mL(5.5)/082J.

BTHfff?f3T"O?Ttro";fl.T"rTi^^

EC: 2.4.1.19, bacillus sp. alkalophilic IC5o:13Mg/mL(4.6)/f5ij.

UHUUl EC: 3.2.1 .48, rabbit small intestine IC5o:16(6.0)/080/ BHtHB EC: 3.2.1 .20, rabbit small intestine IC5o:460(6.0)/080j.

IC50: 290pg/mL (5.5) /082/

EC: 2.4.1.19, bacillus sp. alkalophilic ICso:63Mg/mL(4.6)/f5f;.

kllHfcHJ EC: 3.2.1 .48, rabbit small intestine IC5o:16(6.0)/080|

δS'

I^3 I

^

KJ ^

316

Tables of Glycosidase Inhibitors

%-

HO»"™-'

3.2.5.4 Other Modifications

3-0-a-D-Glucopyranosyl- 1 , 2dideoxymannojirimycin modification of natural product Compound: c105 Synthetic/NMR: [161]

glucopyranosyl)-1 ,5-imino-Darabino-hexitol hydrochloride

1,2,5.Τπ^Αηγγ-3.0.(ΓΥ.η.

natural product Compound: c104 1 Synthetic/NMR: /775/ /287J( H^C)

Fagomine, 1,2-Dideoxynojirimycin

1 ,2,5-Trideoxy-1 ,5-iminoD-arabino-hexitol

Ι4ιϊ·!·ΒΒ·1ιΗίΙΒΕΒ3 rat liver QO!QJ membrane 1C»: 300 (7.0) [161].

IBJBBIBBBHBBIB EC: 3.2.1.23r bovine liver ICso: 38 (opt) /287/

ItMJMMJfcHBBHg EC: 3.2.1.23. bovine liver cvtosolic ICso: 38 (6.8) /775/

yppfX^C^fHiffSP^^^ EC: 3 2.1 23 rat intestine ICso: 15 (5.8) /775/ IC: 15 (opt) /287/

BBBIBBHBBfB EC: 3.2.1 .51 r bovine epididymis ICso: 140 (5.2) [175].

HBiBBIiIiIiHBEEB EC: 3.2.1 .24. rat liver qolqi Il ICso: 1000 (5.5) [1 75/

BBBIBSiHBBHS EC: 3.2.1.20. rice ICso: 320 (opt) /287/

H3J!Jl5^ES 3.2.1 .20, rat liver endoplasmatic reticulum Il ICso: 840 (6.8) [1 75].

EC:

UMWiMtHKBEH EC: 3.2.1.20, bovine liver lysosomal ICso: 520 (4.0) [175].

ICso: 820 (5.8) [1 75/ IC: 46 (opt) /287/

HBBBlHBHBBBIi EC: 3.2.1 .22, qreen coffee beans ICso: 56 (opt) [287].

lolicio-1 ,6-alucosidase fiso ma ltase)l^9kNllbJK!IAI^//i!9··!^·^···· 1C»: 460 (5.8) [1 75].

£

s O5

S ^

O0

318

Tables of Glycosidase Inhibitors

"

Q

OH

HO^" "" ι

/—

Ν

COOH

Hi1*—/

n

/biT Λ j\» /

,-P

IBHBEB EC: 3.2.1 .20, guinea pig intestinal 1C»: 12 (6.0) [219].

IBHIHS:! EC: 3.2.1.20, guinea mouse intestinal IC5o:15(6.0)/2f9/

BIBBSH3 EC: 3.2.1 .48, guinea pig intestinal 1C»: 8.8 (6.0) [21 9].

IffliJflHHIiH EBT-B EC: 3.2.1 .21 , bovine kidney Ki: 130 (5.0), 29 (6.0) /296/. SBBIlIBBHEUWS EC: 3.2.1.2O1 rice Ki: 0.35 (6.0) /296/

Baphiaracemosa, chiral pool chemical Compound: c113 Synthetic/NMR: [240], [022], [273] (1H113C), [242;

2,6-Dideoxy-2,6-imino-L-gulonic acid 1-Deoxynojirimycinuronic acid

chiral pool chemical Compound: c112 Synthetic/NMR: [295] (1H113C)

1 ,5,6-Trideoxy-6,6-difluoro-1 ,5imino-D-glucitol 1 , 6-Dideoxy-6, 6-difluoronojirimycin

chiral pool chemical Compound: d 11 Synthetic/NMR: [127]

IBiIJIIIMIIfcnigggB EC: 3.2.1 .31. human liver, neutral K.: 80 (4.0) /240/ Ki: 80 (4.75), 1C»: 260 (4.75) /022/ BBCTHBBHSS human liver, neutral KSD: 35 (3.5) /022/.

IHiISIIfSl^BEHE^: 3.2.1.31, bovine liver IC:89.4%0.56mM(5.0)/242/

1ΒιΙ«1ΙΙΜ·ΗΕΕΠ3 EC: 3.2.1 .20r bacillus stearothermophilus Ki: 83 (6.8) /273/

H3PIEE3J2SEC: 3.2.1.21r sweet almonds K8: 8700 (5.0) /295/

IBiIWHiM-WEHg EC: 3.2.1 .20, yeasf Ki: 7500 (6.7) /295/

Ki: 0.4 (6.0) /296/

Id: 600 (6.0)Ρ9β/ IBGMEffiH bC: 3.2.1 .21 r asperqillus wentii Ki: 250 (5.0) /296/

1 ,5,6-Trideoxy-6-fluoro-1 ,5-imino- ffBifl WIMtMBEHg EC: 3.2.1.20, yeast Ki: 19 (6.0) /296/ D-glucitol EC: Β0ΡΙΠΗΗ 3.2.1 .21 , sweet almonds 1 , 6-Dideoxy-6-fluoronojirimycin

pKa: 5.75

^ I

S

U)

b

O3

OQ

S"

5n

^

«1

E? i>^ 2

k>

Uj

320

Tables of Glycosidase Inhibitors

HNH2

ι

/

OH

OH

OH

β=°

J\l

L-HNHCH2Ph

HO^Y OH

J

Ki: 40 (4.5 -7.0) /227/

ICso: 1500 (5.2) [01 B].

IBiBIIIHiHBPI EC: 3.2.1 .21 T bovine liver cytosolic K1: 2.6 (7.0) PM].

IBiBIItHiHBIJEH EC: 3.2.1 .21 r tricnoderma ICso: 1000 (n.g.) [018].

BBBiIBHSfRSI:! EC: 3.2.1 .21 , apricot ICso: 120 (n.g.) [018].

IBJBUtHHBIiHiI EC: 3.2.1 .21 , bovine liver lysosomal Ki: 120 (5.0) PM].

[018], [023], 1242]

δ-Gluconolactam chiral pool chemical, modification of natural product Compound: c120 Synthetic/NMR: /294/, [266], [122], [021],

BM»'lim>HEEEg EC: 3.2.1.20. yeast Ki: 1100 (6.8) PW/. IH33HS5BB8 EC: 3.2.1 .21 , sweet almonds Kt: 125 (6.8) [294], Kc 37 (6.2) /023/, IC: 94.1% 1.OmM (5.0) /242/ IHSilHHiiBIH EC: 3.2.1 .21 , agmbacterium faecalis Ki: 5.2 (7.0) /294/ ΜιΙίΊΜΊίιΜΙΤΠ:!?·! EC: 3.2.1.24, rat epididymis

Ki:20±5(5.6)/ff2J.

liii-HIBftHBEEH EC: 3.2.1 .21 , sweet almonds

5-Amino-5-deoxy-D-glucono-1 ,5lactam

N-Benzyl- 1 -β-ammo- 1 -deoxynojirimycin modification of natural product Compound: d 1 9 Synthetic/NMR: [11 2]

5-Amino-N-benzyl-5-deoxy-p-Dglucopyranosylamine

1-j3-Amino-1-deoxynojirimycin modification of natural product Compound: d 1 6 Synthetic/NMR: [1121 [310], [227]

S-Amino-S-deoxy-p-Dglucopyranosylamine

ΗΐιιΐΗ»ΒΐπΜΒΐ^«ι:^^κΒΐΒΐη?ΠΐΜ«ΗΗΒ^3ι EC: 3.2.1.13, rhizopus niveus ICso: 200Mg/mL (5.7) [064]. lfflifl«HIM«Hf«EHg EC! .17.1 .21 , sweet almonds K,:40±3(5.6)fff2/ IHSflBRHBFffi EC: 3.2. 1 .21 , sweet almonds Ki: 40±3 (5.6) [31 0].

OJ K)

322

Tables of Glycosidase Inhibitors

0

\

/

Vv

r-OH

OH

HO^"T

(&»)-{{ NH2

/

Γ,Η

OH

β-"' """

1-Deoxynojirimycinhydroxyamidine chiral pool chemical Compound* c128 Synthetic/NMR: [310]

1 -Amino-1 ,5-dideoxy-1 ,5-imino-Dglucitol hydroxyamidine

1-Deoxynojirimycinamidrazone chirai pool chemical Compound: c127 1 13 Synthetic/NMR: PtOJ( H1 C)

1 -Amino-1 ,5-dideoxy-1 ,5-imino-Dglucitol amidrazone

N-Dodecyl-D-gluconamidine chiral pool chemical Compound: c126 Synthetic/NMR: [265], [266]

νίηβ liver

cytosonc

Ki: 19 (7.0) /265/. IQQfUQQSQ EC: 3.2.1 .21 . bovine liver lysosomal Ki: 3.3 (5.0) /265/. IBSBUBBBBHSI EC: 3.2.1 .21 . sweef almonds Kr: 8.4±0.9 (5.6) [31 0], Ki: 4.7 (5.0) /265/ iBti^'iki'il'if{HBE".S EC: 3.2.1 .24, jack beans Ki: 12 (4.5) /294/. BBBIBBKBISHS EC: 3.2.1.20. yeast K1: 2.9 (6.8) /294/. IBBBIIIHiHBHiIi tt,: 3.2. 1 ./I . sweet almonds K,: 13±3 (5.6) [310], to: 16 (6.8) /294/ BBBUTIBHBHB EC: 3.2.1 .21 , aqmbacterium faecalis Ki: 0.6 (7.0) /294/. HfflHSSBS EC: 3.2.1.21. caldocellum saccharolyticum KJ: 3.3 (6.8) /294/

· °°

EC: 3 2 1 21

H3PIBSEB!

Bliii'iMiliMy?R"ra EC: 3.2.1 .24, jack beans Ki: 3.1±0.6 (5.0)/3fO/. ΠΚΙΒΒΗΒΜ EC: 3.2.1 .51 . bovine kidney K8: 0.82 (6.0) /265/. BffSBHBffiHBEBS bt: O.Z.I M, bovine liver Ki:19±1(7.0)pfO/. BBBHtBBHffiHS EC: 3.2.1 .21 . asperqillus wentii K1: 0.031 (5.0) /265/.

[Π»ϊ·ΙΗϊΗίίΗ? EC: 3.2.1 .21 , bovine liver cytosolic Ki: 0.00020 (7.0) /265J, Ki: 0.0005±0.00002 (7.0) [266].

r

1 -Amino-1 -N-dodecyl-1 ,5-dideoxy-IJMfllffiBHBBHg EC: 3.2.1 .21 bovine liver lysosomal Kr. 0.0007 (5.0) /265J, Kr. 0.0025±0.00002 (5.0) [266/ 1,5-imino-D-glucitol amidine

Ks: 0.13 (7.0) [265], Kr. 0.5±0.15 (7.0) /266/.

MifflUMMBHSig EC: 3 9 1.21 , bovine liver cytosolic

U) K) U)

P

OQ Cn

S'

TQ

1

C5r^ ^ •^ r^ 5X

S

« i-

Oo

Uj

i^j

324

Tables of Glycosidase Inhibitors

OH

\ Q /~~

COOH

Nl

j/ 0

COO-Na+

/°H Nl

COOH

flJi-Rll^JJJJJEEEE EC: 32131, bovine liver Ki: 0.039 (5.2) [240/

IliiftlllWIiMHEEEB EC: 3.2.1 .31 . bovine liver IC: 98.5% 0.1 mM (5.0) /242J IC: 99.6% 1.OmM (5.0) /242/

5-Amino-5-deoxy-D-glucaro-1 ,5lactam sodium salt

Sod/u/T)-D-g/ucaro-<$-/acfarn Compound: d 35 Synthetic/NMR: /242] 5-Amino-2,5-dideoxy-D-glucaro1,5-lactam Compound: c136 Synthetic/NMR: [240]

IfflBl'llBiHEESE M^t almonds Ki: 1200 (5.0) /002/

5-Amino-5-deoxy-D-glucaro-1 ,5lactam Compound: c134 Synthetic/NMR: [002]

U) K)

ι

i

9"

bo

326

Tables of Glycosidase Inhibitors

Sy

3.2 Six-Membered Rings, C6

Iminosugars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

327

328

Tables of Glycosidase Inhibitors

ι to

OO

OQ

3? S'

C£

BBBUTBBHBHSS EC: 3.2.1 .21 , sweet almonds K1: 300±90 (6.8) /229/ K1: 1400 (5.0) /273/ Ks: 5300 (5.0) /020/ IC50: 7.3 (4.8) /Of4/ BBiBHiIBHBHSS EC: 3.2.1.25, aspergillus wentii Ki: 4600 (4.5) /020/

Oo

IBKTBBHBHS EC: 3.2.1 .51 , human placenta KJ: 64 (5.0), 9.5 (6.0), 6.6 (7.0) [21 5/

BKIBBHBBB EC: 3.2.1.51, human /ή/er, neutral KJ: 5.0 (5.5) /077/ IC: 92% 1mM (5.5) [311], Ki: 5.0 (5.5) [137], K1: 5 (n.g.) /2Of/ KK 0.05 (5.5) [107].

IBKBBHBH-IJI EC: 3.2.1 .51 r human liver lysosomal ICM:5(n.g.)/f48/

IfKTBBBHSS EC: 3.2.1 .51 r bovine kidney Ki: 30±10 (5.5) /307/ K1: 0.13 (5.5) /273/ Ki: 30±10 (5.5) [144].

UBIBBHBEB! EC: 3.2.1 .51 , bovine epididymis IC: 97% 1.25mM (6.7) /302/ IC50: 26 (5.2) [1751 ICso: 22 (6.5) /024/

HBBEBIBHBHSi EC: 3.2.1 .24, sweet almonds Ki: 110 (5.0) /020/

IBInHBBHBHSS EC: 3.2.1 .24. rat liver lysosomal 1C»: 570 (4.5) [1 75/

SBinffiSffiPffl EC: 3.2.1 .24, rat liver golgi Il KJ: 410 (5.5), IC50: 500 (5.5) [1 75].

ililnMil'iM-fffRJB EC: 3.2.1 .24, rat liver qolqi I IC50: 25 (5.5) /775/

330

cr>

Tables of Glycosidase Inhibitors

/OH

HO

l\j)

\

/\HO\^ 1

/OCHA

H3CO / \ H O \ ^ '

HO^-j)

4*

1 5-Dideoxv-3-O-ir/-D-

modification of natural product Compound: d016 Synthetic/NMR: [161]

3-0-(6-0-Methyl-a-D-glucopyranosyl)-1deoxymannojirimycin hydrochloride

1 5-Dideoxy-3-0-(6-0-rnethyl-a-Dglucopyranosyl)-1 ,5-imino-Dmannitol hydrochloride

modification of natural product Compound: d015 Synthetic/NMR: [161]

3-0-(3-0-Methyl-a-D-glucopyranosyl)-1deoxymannojirimycin hydrochloride

1 5-Dideoxy-3-0-(3-0-methyl-a-Dglucopyranosyl)-1 ,5-imino-Dmannitol hydrochloride

modification of natural product Compound: d014 Synthetic/NMR: [161]

3-0-(2-0-Methyl-a-D-glucopyranosyl)-1deoxymannojirimycin hydrochloride

1,5-Dideoxy-3-0-(2-0-methyl-a-Dglucopyranosyl)-1 ,5-imino-Dmannitol hydrochloride

modification of natural product Compound: d013 13 Synthetic/NMR: [161], [143] (H C)

3-O-a-D-Glucopyranosyl- 1 deoxymannoiirimycin

glucopyranosyl)-1 ,5-imino-Dmannitol hydrochloride

Wil«!«gtffliflnM'iriM-1iVM"f3 rat liver qolqi membrane ICso: 400 (7.0) [1 61].

IC50: 350 (7 .0) [1 61].

BBEBUIiKiEHBBHBBT-B rat liver qolqi membrane

!SBBBiiBiBffinn!Sgt>B":g rat liver golgi membrane ICso: 4.4 (7.0) [161].

[JjjfBBiffifffffi rat liver golgi membrane IC50: 1 .7 (7.0) [1 61], ICs0: 5.6 (7.0) [143].

h—'

U)

^ U)

S OQ tn

50 r»«.

U

Hi ·>{

c^>

ξ

s·

Co

U)

k)

332

Tables of Glycosidase Inhibitors

icJ/ H0

Π

HO^—-^

o

3.2.7.3 Other Modifications

/OHHO\

1 -2-5-Trideoxy-2-fluoro-1 ,5-iminoD-mannitol 1,2-Dideoxy-2-fluoromannojirimycin chiral pool chemo-enzymatic Compound: d023 Synthetic/NMR: /Ϊ22] (1H113C)

1 ,5,6-Trideoxy-1 ,5-imino1,5-D-mannitol 1 -Deoxy-D-rhamnonojirimycin Compound: d022 Syntte1\cimR:[307],[148].

1 ,5-Dideoxy-3-0-(cc-D-mannopyranosyl)-1 ,5-imino-D-mannitol 3-O-a-D-Mannopyranosyl- 1 deoxymannojirimycin modification of natural product Compound: d021 1 13 Synthetic/NMR: ff 43j ( H1 C)

lliiftlilM»HEEEE! EC: 3.2.1 .21 1 sweet almonds Ki: >1000 (6.5)/f22/.

IC5o:70(n.g.)/f48;. ΙΒι1ΊΊΜ;ΐΊΜ-
ΙΒΠΒΙΒ8ΗΒΒ.Β EC: 3.2.1 .51 r human liver lysosomal

HBiIBBHBEEB EC: 3.2.1 .51 , human liver, neutral KJ: 70 (n.g.), IC: 90% 1.OmM (n.g.) [201].

Κι:11±1(5.5)/3077· HBBBBBHBB138 EC: 3.2.1.24. human liver lysosomal IC50: 1000 (n.g.)/f48/.

!BB/KAHNRSg EC: 3.2.1 .51 . bovine kidney

ICso:25.1 (7. Q) [143].

Μ·Μ··ι···Μ·1ΙΒΗΒΒΕΒί rat liver golgi membrane

'

U)

U)

U)

OQ

1

I

334

Tables of Glycosidase Inhibitors

J

NH+

HO

Bn

HpN OCH3

Amidinepseudodimannoside chiral pool chemical, de novo chemical Compound: d029 Synthetic/NMR: [202J(1H113C), [191]

Synthetic/NMR: [174J(WC), [191], [202]

Benzylmannoamidine chiral pool chemical, de novo chemical

1 -(N-Benzyl)amino-l ,5-dideoxy1,5-imino-D-mannitol amidine

1-Amino-1-deoxy-mannojirimycin amidine chiral pool chemical Compound: d027 Synthetic/NMR: [140]

IHBBEBBBHBBHi EC: 3.2.1.25, sna/7 achatina Ki:120(4.0)/202J,/i9fJ.

κ7Ϊ2θ"(47θ)/ί9ί].

Ks: 0.55 (4.5) [19I]1 Ki: 0.55 (4.5) [174]. EC:3.2.1.21,sweefa/monds KJ: 5 (5.6) [191], Ki: 25 (5.6) [174]. _J EC: 3.2.1.30, jack beans Ki: 6 (4.0)/WJ. EC: 3.2.1.25, snail achatina Ki: 6.0 (4.0) [174], [191]. I EC: 3.2.1.24,/adf teens Ki: 2.6 (4.5) [191], K: 2.6 (4.5) /202J, Ki: 2.6 (4.5) /202J. EC:3.2.1.21,sweefa/monds Ki: 100 (5.6)/WA Ki: 100 (5.6)/202/.

TBJJffllBBHffiBTECr3.2.1.24. iack beans

ICso: 0.4 (5.0)fMQJ.

MDCK-ceiis

MDCK-cells IC50: 3-4 (5.0) /NOJ.

1C»: 0.09-0.1 (5.0) [14O].

IC50:150 (5.0) [14O]. IiJBtBBBHBEHlI mung bean golgi IC5o:4(5.0)/NQJ.

EC: 3.2.1 .25, aspergillus niger

1 -Amino-1,5-dideoxy-1,5-imino-D- MiUBBUBBHU bL: 3.2.1.24. iack beans ICs0: 0.4 (5.0) [14O]. mannitol amidine

i-D-qlucosidase EC:32A21,trichoderma ICso: > 10000 (η.g.)/OfSJ.

U) U)

u> KJ

336

Tables of Glycosidase Inhibitors

3.2.8

:-L-rhamnosidase (naringinase) EC: 3.2.1.40, penicillium decumbens K8:288 (5.0) /302/, K: 490 (5.0) [21O].

ΤΒΒΙΙΒΒΒ!Ι1Ι!ΜΓΕ(..73.2.1.24. human liver Ivsosomal IBJBHIB8HI8BH8 EC: 3.2.1.21. human liver, cytosolic

D-Rhamnono-1 ,5-lactam

D-d-Rhamnonolactam Compound: d036 Synthetic/NMR: [148]

1-Deoxy-D-talonojirimycin Compound: c/037 Synthetic/NMR: [307]

1 ,5-Dideoxy-1 ,5-imino-D-talitol

Κι:53±6(5.5)/307/

i-mannosidase EC: 3.2.1.24, human liver, neutral ICso:100(n.g.)/20f;. EC:3.2.1.51,/)ow?e/c/c(ney

imMiiBiiHmrfBEMiBBBBiiiBMai EC: 3.2.1.40, penicillium Ki: 220 (4.0) [1 52].

L-S-Rhamnonolactam Compound: d035 Synthetic/NMR: [152]

L-RhamnOnO-1,5-laCtam

EC: 3.2.1.22, aspergillus niger IC: 100% 1.25mM (4.5) /302/ Ki: 8.2 (4.5) [302]. EC:3.2.1.40,pen/c////um Ki: 2730 (4.0) /152;.

Ki: 469 (6.7) /302|

[152]

i:8.2 (n.g.)/302/ BBIBJ1SBH.IS EC: 3.2.1.51, bovine kidney Kt: 900000±500000 (5.5) /307/, K8:213 (n.g.) /302/. BBIBSUBBH EC: 3.2.1.51, bovine epididymis

; EC: 3.2.1 .40, Denicillium decumbens

SffiSHSlSSffl EC: 3.2.1.22, green coffee /jeans

K1: 1.0 (4.O)1IC50: 5 (4.0) [238].

Synthetic/NMR: [307], [148], /302/(1H113C),

Compound: d034

chiral pool chemo-enzymatic, chiral pool chemical

L-1-Deoxyrhamnojirimycin

1,5-Dideoxy-1,5-imino-L-rhamnitol

Compound: 4033 1 13 Synthetic/NMR: [238J( H1 C)

5-ep/-L-Deoxyrf?amno/7f/myc/n chiral pool chemical

1,5,6-Trideoxy-1,5-imino-D-gulitol

talo-, rhamno-, D-gulo-Configurations

338

Ii

Tables of Glycosidase Inhibitors

CH3 1+

Λ

OH

HO*-"TO— CH3

uA

HO

\

°\

^H

COOH

OH

HO^-^Q-CH3

/

A

?)°

Η

V^ fc^

Ί

Ο — CH,

OH

HO ^^Τ

^Γ^Γ h\

V* (Q

HO

Nk

HO^CH3 )—0

\-y faj

V

1/

°—1 ^-N+/

Η

.,Λ

HO

2-C-Carboxy-1 ,2,5-trideoxy-1 ,5imino-D-xylitol lsofagominuronic acid chiral pool chemical Compound: d044 Synthetic/NMR: /240]

hydroxymethylpiperidinium]-a-Dglucoheptopyranoside chiral pool chemical Compound: d042 Synthetic/NMR: /229/, [182] Methyl-6,7-dideoxy-7[(1 S,3R,4R,5R)-3,4-dihydroxy-5hy droxy methyl pi peridiny I]-Gt-Dglucoheptopyranoside N-oxide chiral pool chemical Compound: d043 Synthetic/NMR: [229]

Hihx/HroYv-'?·

piperidinium]-a-D-glucoheptopyranoside chiral pool chemical Compound: d041 Synthetic/NMR: /229] Methyl-6,7-dideoxy-7[(1 S,3R,4R,5R)-N-methy 1-3,4-

dihvrirnYv-S-hvHrnyunnAthvl·

Methyl-6,7-dideoxy-7[(1 R1SR1AR1SR)-N-ItIeIHyI-S^-

ItfliB.IIIMHM.It.KT^ EC: 3.2.1 .31 . bovine liver Ki: 0.79 (5.0) /24O].

^JH^^[P^^| EQ· 32110 veast Κί:19±2(6.8)/229/. BHIHBH8BEH8 EC: 3.2.1 .13, asperqillus awamori Ki:0.24±0.04(4.5)/229].

ltBBHBBHBHB EC: 3.2.1 .21 , sweet almonds Κι:0.38±0.02(6.8)/229].

BlifiliffiRHBR^ EC: 3.2.1.20, yeast Kr.70±8(6.8)/229].

Kr. 190±20 (6.8) /22P], κι: 103 (6.8) /782]. IPBBHnHEEB EC: 3.2.1 .13, asperqillus awamori Ki:160±20(4.5)/22P/.

IBiBIIIHiHIBHi EC: 3.2. 1.21, sweet almonds Ki:510±30(6.8)/229/.

Mifl'i'.M'i.'iM3T«B"B EC: 3.2.1 .24r jack beans Ki:400±50(4.5)/229].

IBUJUBBHffBB EC: 3.2. 1 21 , sweet almonds Ki:150±10(6.8)/229]. BIfBSBBHEH;) EC: 3.2.1 .13, asperqillus awamori Ki:94±10(4.5)/229].

IBif»1IIBB":ft«R:S EC: λ 9 1 ?0 yeast Ki:280±60(6.8)/220].

340

Tables of Glycosidase Inhibitors

H

OH

<^-f ^1h°

rf> «Μ

1 ,3,5-Trideoxy-2-C-hydroxymethyl-1,5-iminoD-erythro-pentitol chiral pool chemical Compound: d050 13 Synthetic/NMR: /2847 ( C) 5-Amino-3,5-dideoxy-2-C-hydroxymethyl-D-erythropentono1,5-lactam chiral pool chemical Compound: d051 13 Synthetic/NMR: /284J(H C) ffl!i»ll!H»Hl«E3 EC: 3.2.1. 20r yeasf K.: 2300 (6.8) /284/

lffl!flill!H»HI»Effl EC: 39 1 ?1 ,
U) -Ρ--

I

342

CO

Tables of Glycosidase Inhibitors

O

H3C-S

H3CS

H

NHAc

I/I

r

LJ

'NHAc

H

^ NH Ac

\m

'NHAc

koHHO\

1

^

/OHHO\

H

\^NHAC

^

KOH HO\

H2N- -

^1

\l

/OHHtA

N3— J7-H

(2R,3R,4S,5R)-2-Acetamido-5azidomethyl-piperidine chiral pool chemicdi Compound: d056 Synthetic/NMR: /222] (2R 3R 48 5R)-2-Acetamido-5aminomethyl-3,4-dihydroxypiperidine chiral pool chemical Compound: d057 Synthetic/NMR: /222] (2R,3R,4S,5R)-2-Acetamido-5hydroxymethyl-3,4-dihydroxypiperidine chiral pool chemical Compound: d058 Synthetic/NMR: /2227 (2R,3R,4S,5S)-2-Acetamido-5methylthiomethyl-3,4-dihydroxypiperidine chiral pool chemical Compound: d059 Synthetic/NMR: [222/ (2R,3R,4S,5S)-2-Acetamido-5methylsulfinylmethyl-3,4dihydroxy-piperidine chiral pool chemical Compound: d060 Synthetic/NMR: /2227 IJiiliflliiltklkMM-lili'iiriffgR;!:! EC: 3.2.1 .49, chicken liver IC so: 224 (n.g.) f222|

IC so: 85.4 (n.g.) /2227-

IJB!i!atliflgHm>H;h'illilf»BH;l EC: 3.2.1.30. bovine epididymis IC so: 1.2 (n.g.) /222|

υϋι!ιΒΜ·Β1ΙΙΜ·ΙΊϊΙίΊΙΙί1Μ."^! EC: 3.2.1.30, bovine epididymis IC so: 2.06 (n.g.) /222| fflgWBBHEBffiS^^ EC: 3.2.1.49, chicken liver IC so: 1.32 (n.g.) /2227.

f7RT!WBiF?nTO~»~5O~ffipT^ EC· 3 2 1 49 chicken liver ICso:187(n.g.)i2227.

l!Bt8Biiiljjg|g|^^ EC: 3.2.1.30. bovine eoididvmis IC so: 196 (n.g.) /222/ k'P..!*gCTii»MM*iM^li'il|'tf^!"^ EC: 3.2.1.49, chicken liver IC so: 2.62: /222|

U) 4--U)

£

&

OQ

S' ^J

^

i

S O^

I

S'

Cn

^j ^o

344

Tables of Glycosidase Inhibitors

H NHAc

OH

""--· γ NHAc

H

\ J\

O Il

HNHCOCF3

Γ) k /

H0JLJyY

HO

OH OH

V \

\z/

HOOC

HOOC

NHAo

JL/Λ ~~^\j/i

H0

r bovine liverlysosomal

ICso: 250 (6.0) /245/

|ΙΜ·'Ι1!Ι_1!1ΜΪΕΕΞ!! EC: 3.2.1 .31 , bovine liverlysosomal IC50: 1.6 (5.0) /245/

|fflif*lllH!ItliEEffi ICso: (5.0) /245/

EC: 3 2 1 31

1C»: 15.5 (n.g.)/f767.

ΙΒΕΒΊΗΒΕΕ EC: 3.2.1 .31 . bovine liver IC5o:112(n.g.)/f76/.

flHiIJIIIM.UIikM:! EC: 3 ? 1 .20, yeast IC5o:80(n.g.)/)767.

chiral pool chemical Compound: d070 Synthetic/NMR. [132]

1 ,4,5-trideoxy-(3-L-lyxosylamine 4,5-Diepisiastatin B streptomyces nobilis SANK60192 Compound: d069 Synthetic/NMR: /2457, /2437(1H113C) (3S,4S,5S,6S)-6-Trifluoracetamido- !HiinBHiHBIiTgS EC: 3.2.1.20T yeast ICso: 18.5 (n.g.)/f3273,4,5-trihydroxy-3-piperidine IliifWWIIiMffEEEE! EC: 3.2.1 .31 . bovine liver rarhnvx/lir ^rid If* · Λ OQ in \ M 907 so. ( .g.;/ y.

Compound: d068 Synthetic/NMR: /2457, /2437 (1H113C) 1 -N-Acetvl-5-aminn-4-C-carhoy v-

1 -N-Acetyl-5-amino4-C-carboxy-1 ,4,5-trideoxyβ-L-arabinosylamine 4-Episiastatin B streptomyces nobilis SANK60192 Compound: d067 1 13 Synthetic/NMR: /2457, /2437( H1 C) 1 -N-Acetyl-5-amino-4-C-carboxy1,4,5-trideoxy-a-L-ribitosylamine 6-Episiastatin B streptomyces nobilis SANK60192

chiral pool chemical Compound: d066 Synthetic/NMR: [131]

OQ

C^

346

Tables of Glycosidase Inhibitors

3.3 Six-Membered Rings, Cj

la

8 3 of

t* Q)

1 6 -<M —-

s


l"t. ._ o

O L

o)

£

1$J '

~

- &i o&j ^6^

~J

'

-52 o

J2 3

E O

'?=

I

I -C |

I9

I T

"C

ci of

X CO

CO CO

-r

CO CO

C\l CO CO

Iminosngars (is Glycosulase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. StiHz C'opyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

347

348

Tables of Glycosidase Inhibitors

y ~1

HO

hp£p|

H

N?HH(^

/Ά

OH OH

1 ι

K\

H

/OH

HCI

OH

OH

H

chiral pool chemical Compound: e013 Synthetic/NMR: /2387 (1H^3C)

β-L-Homofuconojirimycin, 2, 6, 7-Trideoxy-2, 6imino-L-glycero-D-manno-heptitol chiral pool chemical Compound: e012 Synthetic/NMR: /077J(U13C), /255J(U13C) 1 ,2,6-Trideoxy-2,6-imino-Dglycero-L-manno-heptitol J3-homO'L-Rhamnojirimycin, 2,6, 7-Trideoxy2,6-imino-L-glycero-L-galacto-heptitol

pyranosyl-2,6-imino-D-glycero-Lgulo-heptitol hydrochloride 6-0-J3-Glucopyranosyl-a-homonojirimycin hydrochloride chiral pool chemical Compound: e010 Synthetic/NMR: [315] 2,6-Dideoxy-2,6-imino-L-glyceroD-talo-heptitol 6-epi-Homomannojirimycin chiral pool chemo-enzymatic Compound: e011 Synthetic/NMR: [137], [076], [189] \ ,2,6-Trideoxy-2,6-imino-Dglycero-D-galacto-heptitol

2 β-ΠΐΗρπγν.7-Ο-β-πΙιΐΓη-

chiral pool chemo-enzymatic Compound: e009 Synthetic/NMR: [104], 3 2 1 26 rat intestine

ΙΙΒΒΙΙ^ΓιΤΓιΒΗΒΒΤίΒΙΙΠΠΙΠΤΠΠ^Η^ EC· 3 2 1 40 penicillium rier.umhpns IC5o:730(4.0)f238/.

ΗΠΒΙΒΒΗΒΕΕΒ EC: 3.2.1 .51 , charonia lampas Ki: 0.14 (4.5) [107].

ΒΗΠΒΒΗΒΗΒ! EC: 3.2.1.51T human liver, neutral Ki: 0.01 (5.5) [077], Ki: 0.01 (5.5) [107].

BiBWiM-WEHg EC: 3.2.1 .51 r bovine epididymis Ki:0.02(6.0)/f07y.

!HBrtFAHRRSf;! EC: 3.2.1 .51 r human liver, neutral K8: 4.5 (5.5) [288], IC: 93% 1mM (5.5) β11].

UflpB^ffBBnnnfftHttFfffnfff!^^ KJ: 2.0 (n.g.) [31 S].

350

Tables of Glycosidase Inhibitors

3.4

iv

HO

I

HO ]

* L— OH

/Ph H0\

/-A

HO

niW/LoH

Η

M k ?Λ

I^

J

/CjH^ \

Hu

β-1-C-Phenyl-deoxymannojirimycin, 2, 6dideoxy-2,6-imino-6-C-phenyl-L-glycero-Dmanno-hexitol chiral pool chemical Compound: e021 Synthetic/NMR: /0777, [107]

1 ,5-Dideoxy-1 -C-phenyl-1 ,5-iminoD-glycero-D-galacto-hexitol

Synthetic/NMR: [077]

PnmnminH· &09CI

j3-1-C-Ethyl-deoxymannojirimycin chiral pool chemical

2,6,7-Trideoxy-2,6-imino-Lglycero-D-manno-octitol

6-C-Ethyl-1-deoxy-Lfuconojirimycinhydrochloride chiral pool chemical Compound: e01 9 Synthetic/NMR: [089]

1 ,5,6,7,8-Pentadeoxy-1 ,5-imino-Lgalacto-octitol hydrochloride

Six-Membered Rings, Ce and Higher

2, N-Dehydro-2,6-imino-2,6,7-trideoxy-Lglycero-D-manno-heptitol,(6S)-6-amino-6deoxy-6-C-methyl-L-tagatose chiral pool chemical Compound: e018 Synthetic/NMR:/29JJ

6-Amino-6,7-dideoxy-L-mannohept-2-ulose

IiBBIiHBBHJ! EC : 3.2. 1 .5 1 , chamnia lampas Ki: 2000 (4.5) [1 07/.

IBBBHIIiBHi EC: 3.2.1.51, human liver, neutral K1: 1.0 (5.5) [077/,K1: 1(5.5) [107].

BBBJiMtHKHSI EC: 3.2.1 .51 r bovine epididymis Ki: 5.5 (6.0) [1 07].

HKIBiHBBS tC: Λ.Ζ1 .M t chamnia lampas ft: 0.58 (4.5) [1 07].

HKIHiHBHJi EC: 3.2.1.51, human liver, neutral Ki: 0.07 (5.5) [077], Ki: 0.07 (5.5) [107].

BBBBfAT-IBRcS EC: 3.2.1. 51T bovine epididymis Ki: 2 (6.0) [1 07].

BBB^IfAHItR-B EC: 3.2.1.51 r bovine epididymis ICso: 2.35 (6.0) [089].

I EC: 3.2.1.51, bovine epididymis Ki:0.0069±0.0020(6.8)/29V.

U)

352

Tables of Glycosidase Inhibitors

4 Seven-Membered Rings

Iminosugars (is Glycosiflase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. StiHz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

353

354

Tables of Glycosidase Inhibitors

\···ΌΗ

NH2

HO

CH2Ph

OH

--ρ·-

S

HO — J

HO

(3R,4R,5R,6R)-N-Benzyl-3,4,5,6tetrahydroxyazepane N-Benzyl-1, 6-dideoxy-1, 6-imino-D-mannitol chiral pool chemical Compound: f01 6 Synthetic/NMR: /22OJ, /225] X-Ray: /22OJ

Compound: f01 5 Synthetic/NMR:f)79/ Ki: 23.4±3.8 (6.0) [257], [220], [225].

ΈΚ-^^^ΗΒΒΒΓΕΟ: 3.2.1 .24, /ac/f beans 4S-Amino-(3S,5S,6S)IC5o:>850(n.g.)[t79/. trihydroxyazepene 4-Amino-1, 4, 6-trideoxy- 1, 6-imino-D-mannitol chirai pool chemical

I

364

Tables of Glycosidase Inhibitors

Iminosugars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz C'opyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

Γ

Γ

ι

9

H

T

(_|

I

H

^

1

OH

\ OH

OH

\ OH

OH

(3S)-(Hydroxymethyl)-swainsonine chiral pool chemical Compound: f0902 Synthetic/NMR: /283/

(1S,2R,3S,8R,8aR)-3(Hydroxy methyl)- 1,2,8trihydroxyindolizidine

(3R)-(Hydroxymethyl)-swainsonine chiral pool chemical Compound: /0892 Synthetic/NMR: /2837

(Hydroxymethyl)-1,2,8trihydroxyindolizidine

(1S 2R 3R 8R 8aR)-3-

JBiIiWWNHREEg EC: 3.2.1 .24, jack beans ICso: 45 (5) /283/

IC50: 1.2 (5) [283/

I ·ΐβ·*ΜΜΛ«·ΜΑ! EC· 3 9 1 94 lack beans

366

Tables of Glycosidase Inhibitors

7.1 Castanospermine and Analogs

367

368

Tables of Glycosidase Inhibitors

7.1 Castanospermine and Analogs

369

370

Tables of Glycosidase Inhibitors

7.1 Castanospermine and Analogs

311

372

Tables of Glycosidase Inhibitors

7.2

OH

/

1

I)

OH

OH

OH

HKLI>

H

k^N^/

Γ

xxlX

Ci>"

u

H

P

H

(1 S,2S,8aS)-1 ,2Dihydroxyindolizidine Lentlginosine chiral pool chemical, astralagus lentiginosus [091] Compound: /0742 Synthetic/NMR: [091], [211], [226] (1 R^R1SaR)-I ,?nihyHrnvyinrinliTiHinA (+Lentiginosine chiral pool chemical Compound: /0752 Synthetic/NMR: [211] (1R,2S,8aR)-1,2Dihydroxyindolizidine ),8a-D/-ep/-/enf/g/nos/ne de novo chemical, chiral pool chemical Compound: /0762 Synthetic/NMR: [019]

Lentiginosines, Swainsonines and Analogs

!Hili'ikhi'it«HttB.B EC: 3.2.1.24, human acidic (tysosomal) ICso: 7500 (4) [019].

K

RnngBmmgSEmRT^ EC: 3.2.1 .3. aspera/7/us n/aer »= ^O (5), IC50: 110 (5) /2f ft IC50: 108 (5) [226]. I UMIi.H-IIIH*llftffl EC: 3.2.1 .3, rhizopus mold Ki:98(5)/2fftlC5o:180(n.g.)

I ϋϋ-ϋδΙΙ-Ε-ΕΙϊ" EC: 3.2.1 .3, asperg/te n/ger IC50: 32 (5) [111], Ki: 10 (5) [211], IC50: 5 (5) [091], Ki: 2.0 (5), IC50: 2.7 (5) /2f ft /226/. I ^MIIIHMftffl EC: 3.2.1 .3.rfwzoousmo/cf Ki: 3.0 (5), Cso: 3.1 (5) Pf ft /226J.

'

LO
O

Co

S'

>5

ο*

374

Tables of Glycosidase Inhibitors

|

Η<

?

H

PH

V-OH

PH

Η

OH

Η

Γ y OH

H

T

H

H

/H

H

OH

PH

CJIy"

?

Η

k^N-^y

r^H-A Γ T) OH

?

Η

(JO °

ί

[

I

2, 8-Di-epi-swainsonine modification of natural product Compound: /0832 Synthetic/NMR: /050], [130]

(1S,2S,8S,8aR)-1,2,8Trihydroxyindolizidine

(1R,2R,8S,8aR)-1,2,8Trihydroxyindolizidine 1 , 8-Di-epi-swainsonine chiral pool chemical Compound: /0822 Synthetic/NMR: [041]

(1S,2S,8R,8aR)-1,2,8Trihydroxyindolizidine 2-Epi-swainsonine chiral pool chemical Compound: /0792 Synthetic/NMR: /050] (1S,2R,8S,8aR)-1,2,8Trihydroxyindolizidine 8-Ep/-swa/nson/ne chiral pool chemical Compound: /0802 Synthetic/NMR: /0-ffJ (1S,2R,8R,8aS)-1,2,8Trihydroxyindolizidine 8a-Epi-swainsonine chiral pool chemical Compound: /0872 Synthetic/NMR; [035], [052]

EC:

UiiIWIMtHtiEHg EC: 3.2.1 .21 . bitter almonds (emulsin) K1: 500 (5), 1C»: 150-200 (5) /050/.

1C»: 10 (4) /306| ili-l^'iM'il'iM^ffrag EC: 3.2.1.24, human liver Ivsosomal Ki: 75 (4), IC: 93% 1mM (4) /088/. IUlBBBBlIBHBI EC: 3.2.1 .24r human acidic (lysosomal) Ki: 75 (4), /088/, IC: 93% 1mM (4) /035/. HBBHiBHJBISEB EC: 3.2.1 .24. human liver, qolqi Il IC: 96% 1mM (5.75) /088/. ΜιΒ.'.ΜΊ.ίΝΗΤίΕ"^ EC: 3.2. 1 .24r rat liver lysosomal 1C»: 50 (4) /306|

BBiflV.M'i.'if.Hfreng EC: 3.2.1 .24, rat liver lysosomal 1C»: 50 (4) /306/.

EEHEHlE^E 3.2.1 .3. aspemillus niaer Ks: 50 (5.0), ICso: 50 (5.0) [050], Kc 50 (5.0) 1226]. filffl?fflTB!5?B3;|I mwng bean seedlings 1C»: 50 (6.5) /05OJ.

U)
Co

OQ

G ^ ?Γ

^ S

Is.

a s

^n

S <^>

S ^

I

On

δ

S O Co S

OQ

"-K

t> <^ S

>) to

376

Tables of Glycosidase Inhibitors

0-

O

^

-?•

0

rn

O CL

^ >

0.

CL

S. rn

§

&

Q3 s.i p> asx2

^ 2 ao to

< &

3|

^ 5-

O)

<-<. :?. ?

',/i

o c> *- -" 2~

O

B ~ 3 -4
V °L <S

-!

E P s" 5 € sz o£ 5

8

OH

OH

^^

^-^

?H ?H >\f^\

*— OH

HO

OH

H0*^\^[x^

-AN^,

HO

s^»^ H0IT)

HO^

H

XT^i ^/N^ HO

OH OH : H f HO^ ^^ 1 >s.

HO

HoXXj

HOx^J^X

HO

HO' ^^

T T^^J

HO

Quinolizidines

(1R,2R,3S,9R,9aR)-1 ,2,3,9Tetrahydroxyquinolizidine 1-Epi-homocastanospermine chiral pool chemical Compound: f0923 Synthetic/NMR: 1300] (1 R,2R,3R,9R,9aR)-1 ,2,3,9Tetrahydroxyquinoiizidine chiral pool chemical Compound: TO3 Synthetic/NMR: flSOj, /2507 (1 R,2R,3R,9R,9aS)-1 ,2,3,9Tetrahydroxyquinolizidine chiral pool chemical Compound: f0943 Synthetic/NMR: [250] (1S,2R,3S,4R,9R,9aR)-4Hydroxymethylquinolizidine1,2,3,9-tetraol chiral pool chemical Compound: f0953 Synthetic/NMR: [316]

Qwnthotir/NMR H-/D/

(1R,2R,3S,9S,9aR)-1 ,2,3,9Tetrahydroxyquinolizidine Homocastanospermine chiral pool chemical Compound' f0913

niB»U!IiLlIi£Hl EC: 3.2.1 .20. p/q /f/dne|/ IC5o:0.3(n.g.)/3f6/.

JBBIINM-WET3": EC: 3.2.1 .51 1 bov/ne epididymis IC5o:360(5)f2507.

IBil^iM'iliM-ltn.^ EC: 3.2.1 .24. human acidic (lysosomal) Id: 8.5 (4.5) flSOJ.

roiii'iMil'iNHfin^ EC: 3.2.1 .24, /ac/f beans Ki:360(5.5)/300/.

IJiifflfflBHBH tC! ^.^.1 ^4T fack beans Ki: 520 (5.5) [300].

ItfliflJIIIfJ.Mf.Rng EC: 3.2. 1 .21 , sweef a/monofs K.:390(5)/300/. [ffllEIHSBSffl EC: 3.2.1.21, aspergillus wentii Ki:8(5)/300;.

00

U)
s* 5

S"

M

S* S-

O g

378

Tables of Glycosidase Inhibitors

Iminosugars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. StiHz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

4-0-j3-D-gal-calystegine 82 modification of natural product Compound: f1014 Synthetic/NMR: [297]

ICso: 880 (opt) /297/

EC:

3.2.1.23. bovine liver

EC: 3.2.1.21, caldocellum saccharolyticum K,: 0.55 (opt), ICso: 2.4 (opt) [213], /262/ BHiHBB EC: 3.2.1.28, porcine kidney Ki: 5.3 (opt), ICso: 10 (opt) [213], Ki: 5.3 (6.5), ICso: 10 (6.5) /262/ IBiBHBlaRHBRlug EC: 3.2.1.22, green coffee beans ICso: 26 (opt), 80 (opt) /297/

ICso: 75(opt) [213], ICso: 70 (5) /262/

Ki: 2.3(opt), ICso: 3.9 (opt) f2i 3/ /262/ Ki: 7 (n.g.) /248/ ΙΗϋΒΙΒΒΒΒΗΒ EC: 3.2.1 .21 . sweet almonds Ki: 15 (5.0) [1491 Ki: 1.2(opt), IC50: 2.6 (opt) [213], /262/ Ki: 4.2 (n.g.) [248]. EC: 3.2.1 .20. rice

IJJiIJMM4fcHBBH9 EC: 3.2.1.23, bovine liver Ki: 46 (opt), ICso: 240 (opt) [213], Ki: 45(opt) /262/

EC: 3.2.1 .22. bovine liver Ki: 45 (opt), ICso: 240 (opt) /262/

ICso: 7.8 (opt) [213].

Ki: 0.86 (opt), ICso: 1.9 (opt) [213], [262].

Calystegine 82 chiral pool chemical, solanum tuberosum, s. dulcamara, s. melongena[149], calystegia sepium [095], physalis alkekengi [213], hyoscyamusniger[262]. Compound: ft 004 Synthetic/NMR: /095/ [213]

MiBEIBBgHEEag EC: 3.2.1.22, green coffee beans

(1 R,2S,3R,4S,5R)-1,2,3,4-

.: 3.2.1.20, rice Κι: 0.9±0.1 (5.0), ICso: 1.9 (5.0) /297/ EC: 3.2.1.21, caldocellum saccharolyticum ICso: 300 (opt)/297/

EC: 3.2.1 .21T sweet almonds

Tetrahydroxy-8-azabicyclo [3.2.1 ]octane

3-0-j3-D-glc-calystegine Bi natural product, modification of natural product, nicandra physalodes [039] Compound: /0994 Synthetic/NMR: /297/

380

Tables of Glycosidase Inhibitors

! EC: 3.2.123. rat intestine

ΒΒΙΒΪΕΙΕΤΒΒΗΒΕ^ EC: 3.2.1.22, aspergillus niger ^68(0Pt) ICso: 100(opt)/262/ IB8IWHM EC: 3.2.1.21, sweet almonds K,: 14 (opt), ICso: 34 (opt)/262J.

Γ3 2 llOCtane CalystegineN1

natural product, hyoscyamus niger [262] — Compound: ft 064 Synthetic/NMR: /262]

UHiHIJSi EC: 3.2.1 .28, porcine kidney K1: 62 (6.5), ICso: 100 (6.5) /262/

IHiBIIBBH!iH.I3 EC: 3.2.1.21, caldocellum saccharolyticum Ki: 5.5 (opt), ICso: 14 (opt) /262/.

Mi-BHEBRHfiEEB EC: 3.2.1.22, green coffee beans

trihydroxy-8-azabicyclo

KK 75(opt) jCso.260 (opt) /262/

3HiHHSi EC: 3.2.1 .28, porcine kidney 1C»: 270 (opt) /2f 3/.

IBiBIIIBiHBHUi EC: 3.2.1.21, caldocellum saccharolyticum Ki: 0.29 (opt), ICso: 0.86 (opt) [21 3].

BilJIIIIJMMJl! EC: 3.2.1.20, rice ICso: 420 (opt) [213].

Ki: 0.45 (opt), ICso: 0.82 (opt) [213].

>-galactosidase (lac IC5o:0.38(opt)/2f3|

Ki: 14 (opt), ICso: 440 (opt) /2)3/

I EC: 3.2.1.22, aspergillus niger

IBBH!!HB8HBHB! EC: 3.2.1.23, bovine liver K1: 3.6 (opt), IC50: 16 (opt) [21 3].

I EC: 3.2.1.22, green coffee beans K1:90 (opt), ICso: 360 (opt) [213].

1 R-Amino-(2S,3R-4S,5R)-2,3,4-

Compound: ft 054 Synthetic/NMR: [213]

Calystegine Ci natural product, physalis alkekengi [213]

(1R,2S,3R,4S,5R,6S)-1,2,3,4,6Pentahydroxy-8-azabicyclo [3.2.1 Joctane

U) OO

382

Tables of Glycosidase Inhibitors

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]

A. L. Tarentino, F. Maley, J. Biol Chem., 1974, 249, 811-817 G. Legler, F. Witassek, Hoppe-Seylers' Z. Physiol Chem., 1974, 355, 617-625 L. E. Fellows, J. Chem. Soc., Chem. Commun., 1979, 977-978 S. M. Colegate, P. R. Dorling, C. R. Huxtable, Aust. J. Chem., 1979, 32, 2257-2264 B. W. Skelton, A. H. White, Aust. J. Chem., 1980, 33, 435-439 R R. Dorling, C. R. Huxtable, S . M. Colegate, Biochem. J., 1980, 191, 649-651 L. D. Hohenschutz, E. A. Bell, R J. Jewess, D. R Leworthy, R. J. Pryce, E. Arnold, J. Clardy, Phy to chemistry, 1981, 20, 811-814 G. Kinast, M. Schedel, Angew. Chem., 1981, 93, 799-800 G. Hanozet, H-R Pircher, R Vanni, B. Oesch, G. Semenza, /. Biol. Chem., 1981, 256, 3703-3711 A. C. Weymouth-Wilson, Natural Product Reports, 1982, 99-110 R. J. Molyneux, L. F. James, Science, 1982, 216, 190-191 B. Saunier, R. D. Kilker, J. S. Tkacz, A. Quaroni, A. Herscovics, J. Biol. Chem., 1982, 257, 14155-14161 K. Chotai, C. Jennings, B. Winchester, P. Dorling, J. Cell. Biochem., 1983, 21, 107 S. V. Evans, L. E. Fellows, E. A. Bell, Phy to chemistry, 1983, 22, 768-770 M. J. Schneider, F. S. Ungemach, H. R Broquist, T. M. Harris, Tetrahedron, 1983, 39, 29-32 G. W. J. Fleet, R W. Smith, S. V. Evans, L. E. Fellows, /. Chem. Soc., Chem. Commun., 1984, 1240-1241 G. W. J. Fleet, M. J. Gough, R W. Smith, Tetrahedron Lett., 1984, 25, 1853-1856 T. Niwa, T. Tsuruoka, H. Goi, Y. Kodama, J. Itoh, S. Inouye, Y. Yamada, T. Niida, M. Nobe, Y. Ogawa, J. Antibiot., 1984, 37, 1579-1586 S. M. Colegate, R. R. Dorling, C. R. Huxtable, Aust. J. Chem., 1984, 37, 1503-1509 G. Legler, E. Jiilich, Carbohydr. Res., 1984, 128, 61-72 H. Hettkamp, G. Legler, E. Bause, Eur. J. Biochem., 1984, 142, 85-90 I. Cenci di Bello, P. Dorling, L. Fellows, B. Winchester, FEES Lett., 1984, 176, 61-64 M. R Dale, H. E. Ensley, K. Kern, K. A. R. Sastry, L. D. Byers, Biochemistry, 1985, 24, 3530-3539 S. V. Evans, L. E. Fellows, T. K. M. Shing, G. W. J. Fleet, Phy to chemistry, 1985, 24, 1953-1955 R. J. Nash, E. A. Bell, J. M. Williams, Phytochemistry, 1985, 24, 1620-1622 G. W J. Fleet, S. J. Nicholas, R W Smith, S. V. Evans, L. E. Fellows, R. J. Nash, Tetrahedron Lett., 1985, 26, 3127-3130 M. J. Eis, C. J. Rule, B. A. Wurzburg, B. Ganem, Tetrahedron Lett., 1985, 26, 53975398 C. J. Rule, B. A. Wurzburg, B. Ganem, Tetrahedron Lett., 1985, 26, 5379-5380 M. Hino, O. Nakayama, Y. Tsurumi, K. Adachi, H. Terano, M. Kohsaka, H. Aoki, J. Antibiot., 1985, 38, 926-935 R J. Card, W D. Hitz, /. Org. Chem., 1985, 50, 891-893 D. R. R Tulsiani, H. R Broquist, O. Touster, Arch. Biochem. Biophys., 1985, 236, 427-434 G. Palamarczyk, M. Mitchell, R W. Smith, G. W. J. Fleet, A. D. Elbein, Arch. Biochem. Biophys., 1985,243,35-45 G. Legler, H. Liedtke, Biol Chem. Hoppe-Seyler, 1985, 366, 1113-1122 G. W. J. Fleet, R W. Smith, R. J. Nash, L. E. Fellows, R. B. Parekh, T. W. Rademacher, Chem. Lett., 1986, 1051-1054 K.-I. Tadano, Y. Hotta, M. Morita, T. Suami, B. Winchester, I. Cenci di Bello, Chem. Lett., 1986, 2105-2108 Y. Imura, Y. Hotta, C. Fukabori, K. Tadano, T. Suami, J. Carbohyd. Chem., 1986, 5, 147-152

Iminosiigars as Glycosulase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

Tables of Glycosidase Inhibitors

383

[37] G. W. J. Fleet, K. James, R. J. Lunn, Tetrahedron Lett., 1986, 27, 3053-3056 [38] B. P. Bashyal, H.-F. Chow, G. W. J. Fleet, Tetrahedron Lett., 1986, 27, 3205-3208 [39] R. C. Griffiths, A. A. Watson, H. Kizu, N. Asano, H. J. Sharp, M. G. Jones, M. R. Wormald, G. W. J. Fleet, R. J. Nash, Tetrahedron Lett., 1986, 37, 3207-3208 [40] G. W. J. Fleet, P. W. Smith, Tetrahedron, 1986, 42, 5685-5692 [41] K.-I. Tadano, Y. Imura, Y. Hotta, C. Fukabori, T. Suami, Bull. Chem. Soc. Jpn., 1986, 59, 3885-3892 [42] J. Kuszmann, J. Kiss, Carbohydr. Res., 1986, 153, 45-53 [43] G. Legler, S. Pohl, Carbohydr. Res., 1986, 155, 119-129 [44] J. Schweden, G. Legler, E. Bause, Eur. J. Biochem., 1986, 157, 563-570 [45] G. Trugnan, M. Rousset, A. Zweibaum, FEBS, 1986, 199, 28-32 [46] E. Bause, R. Erkens, J. Schweden, L. Jaenicke, FEBS Lett., 1986, 206, 208-212 [47] T. Szumilo, G. R Kausal, A. D. Elbein, Arch. Biochem. Biophys., 1986, 247, 261-271 [48] J. Schweden, C. Borgmann, G. Legler, E. Bause, Arch. Biochem. Biophys., 1986, 248, 335-340 [49] Y. Miyake, M. Ebata, J. Antibiot., 1987, 1, 122-123 [50] A. D. Elbein, T. Szumilo, B. A. Sanford, K. B. Sharpless, C. Adams, Biochemistry, 1987, 26, 2502-2510 [51] Ellmers, Biochem. Pharmacol, 1987, 36, 2381-2385 [52] K.-I. Tadano, Y. Hotta, M. Morita, T. Suami, B. Winchester, I. Cenci di Bello, Bull Chem. Soc. Jpn., 1987, 60, 1403-1404 [53] H. J. G. Broxterman, G. A. van der Marel, J. J. Neefjes, H. L. Ploegh, J. H. van Boom, Reel Trav. Chim. Pays-Bas, 1987, 106, 571-576 [54] H. Paulsen, M. Matzke, Liebigs Ann. Chem., 1988, 1112-1121 [55] H. Paulsen, M. Matzke, Liebigs Ann. Chem., 1988, 1121-1126 [56] R. J. Nash, L. E. Fellows, J. V. Dring, C. H. Stirton, D. Carter, H. P. Hegarty, E. A. Bell, Phytochemistry, 1988, 27, 1403-1404 [57] G. W. J. Fleet, N. G. Ramsden, R. J. Molyneux, G. S. Jacob, Tetrahedron Lett., 1988, 29, 3603-3606 [58] R. J. Nash, L. E. Fellows, J. V. Dring, G. W. J. Fleet, A. E. Derome, T. A. Hamor, A. M. Scofield, D. J. Watkin, Tetrahedron Lett., 1988, 29, 2487-2490 [59] T. Shibata, O. Nakayama, Y. Tsurumi, M. Okuhara, H. Terano, M. Kohsaka, /. Antibiotics, 1988, 41, 296-301 [60] D. Hendry, L. Hough, A. C. Richardson, Tetrahedron, 1988, 44, 6143-6152 [61] G. W. J. Fleet, J. C. Son, D. St. C. Green, I. Cenci di Bello, B. Winchester, Tetrahedron, 1988, 44, 2649-2655 [62] R. J. Nash, L. E. Fellows, A. C. Plant, G. W. J. Fleet, A. E. Derome, P. D. Baird, M. P. Hegarty, A. M. Scofield, Tetrahedron, 1988, 44, 5959-5964 [63] R. J. Molyneux, M. Benson, R. Y. Wong, J. E. Tropea, A. D. Elbein, J. Nat. Prod., 1988, 51, 1198-1206 [64] Y. Yoshikuni, AgHc. Biol Chem., 1988, 52, 121-128 [65] Y. Miyake, M. Ebata, AgHc. Biol. Chem., 1988, 52, 1649-1654 [66] K.-I. Tadano, M. Morita, Y. Hotta, S. Ogawa, B. Winchester, I. Cenci di Bello, J. Org. Chem., 1988, 53, 5209-5215 [67] G. Legler, E. Bieberich, Arch. Biochem. Biophys., 1988, 260, 427-436 [68] R. R. Schmidt, J. Michel, E. Riicker, Liebigs. Ann. Chem., 1989, 423-428 [69] H. Hashimoto, M. Hayakawa, Chem. Lett., 1989, 1881-1884 [70] G. W. J. Fleet, C. J. Mathews, J. A. Seijas, M. P. V. Tato, D. J. Brown, J. Chem. Soc., Perkin Trans. I, 1989, 1065-1066 [71] G. W. J. Fleet, C. J. Mathews, J. A. Seijas, M. P. V. Tato, D. J. Brown, J. Chem. Soc., Perkin Trans. I, 1989, 1067-1068 [72] G. W. J. Fleet, S. Petursson, A. L. Campbell, R. A. Mueller, J. R. Behling, K. A. Babiak, J. S. Ng, M. G. Scaros, /. Chem. Soc., Perkin Trans. I, 1989, 665-666 [73] N. Chida, Y. Furuno, S. Ogawa, / Chem. Soc., Chem. Commun., 1989, 1230-1231 [74] E. Kappes, G. Legler, J. Carbohydr. Chem., 1989, 8, 371-388

384

Tables of Glycosidase Inhibitors

[75] J. E. Tropea, R. J. Molyneux, G. R Kaushal, Y. T. Pan, M. Mitchell, A. D. Elbein, Biochemistry, 1989, 28, 2027-2034 [76] I. Bruce, G. W. J. Fleet, I. Cenci di Bello, B. Winchester, Tetrahedron Lett., 1989, 30, 7257-7260 [77] G. W. J. Fleet, S. K. Namgoong, C. Barker, S. Baines, G. S. Jacob, B. Winchester, Tetrahedron Lett., 1989, 30, 4439-4442 [78] C. M. Harris, T. M. Harris, R. J. Molyneux, J. E. Tropea, A. D. Elbein, Tetrahedron Lett., 1989, 30, 5685-5688 [79] N. M. Carpenter, G. W. J. Fleet, I. Cenci di Bello, B. Winchester, L. E. Fellows, R. J. Nash, Tetrahedron Lett., 1989, 30, 7261-7264 [80] Y. Yoshikuni, Y. Ezure, T. Seto, K. Mori, M. Watanabe, H. Enomoto, Chem. Pharm. Bull, 1989, 37, 106-109 [81] T. Furumoto, N. Asano, Y Kameda, K. Matsui, J. Antibiot., 1989, 42, 1302-1303 [82] Y Ezure, S. Maruo, N. Ojima, K. Konno, H. Yamashita, K. Miyazaki, T. Seto, N. Yamada, M. Sugiyama, Agric. Biol. Chem., 1989, 53, 61-68 [83] H. Kayakiri, S. Takase, T. Shibata, M. Okamoto, H. Terano, M. Hashimoto, T. Tada, S. Koda, /. Org. Chem., 1989, 54, 4015-4016 [84] Y Auberson, R Vogel, Angew. Chem., 1989, 101, 1554-1555 [85] M. Bollen, W. Stalmans, Eur. J. Biochem., 1989, 181, 775-780 [86] E. Bause, J. Schweden, A. Gross, B. Orthen, Eur. J. Biochem., 1989, 183, 661-669 [87] S. Al Daher, G. Fleet, S. K. Namgoong, B. Winchester, Biochem. J., 1989, 258, 613-615 [88] I. Cenci di Bello, G. W J. Fleet, S. K. Namgoong, K.-L Tadano, B. Winchester, Biochem. J., 1989, 259, 855-861 [89] H. Paulsen, M. Matzke, R. Nuck, B. Orthen, W. Reutter, Liebigs Ann. Chem., 1990, 953-963 [90] S. Aoyagi, S. Fujimaki, C. Kibayashi, J. Chem. Soc., Chem. Commun., 1990, 14571459 [91] L Pastuszak, R. J. Molyneux, L. F. James, A. D. Elbein, Biochemistry, 1990, 29, 18861891 [92] R. J. Nash, L. E. Fellows, J. V. Bring, G. W. J. Fleet, A. Girdhar, N. G. Ramsden, J. M. Peach, M. P. Hegarty, A. M Scofield, Phy to chemistry, 1990, 29, 111-114 [93] A. M. Scofield, J. T. Rossiter, P. Witham, G. C. Kite, R. J. Nash, L. E. Fellows, Phytochemistry, 1990, 29, 107-109 [94] R. C. Bernotas, G. Papandreou, J. Urbach, B. Ganem, Tetrahedron Lett., 1990, 31, 3393-3396 [95] P H. Ducrot, J. Y. Lallemand, Tetrahedron Lett., 1990, 31, 3879-3882 [96] M. K. Tong, E. M. Blumenthal, B. Ganem, Tetrahedron Lett., 1990, 31, 1683-1684 [97] G. Legler, Adv. Carbohydr. Chem. Biochem., 1990, 48, 319-384 [98] R. J. Molyneux, J. E. Tropea, A. D. Elbein, J. Nat. Prod., 1990, 53, 609-614 [99] S. A. Miller, A. R. Chamberlin, J. Am. Chem. Soc., 1990, 112, 8100-8112 [100] M. K. Tong, G. Papandreou, B. Ganem, J. Am. Chem. Soc., 1990, 112, 6137-6139 [101] W. E Collin, G. W. J. Fleet, M. Haraldsson, I. Cenci di Bello, B. Winchester, Carbohydr. Res., 1990, 202, 105-116 [102] F. R. Heiker, A. M. Schueller, Carbohydr. Res., 1990, 203, 314-318 [103] G. W. J. Fleet, N. G. Ramsden, R. J. Nash, L. E. Fellows, G. S. Jacob, R. J. Molyneux, I. Cenci di Bello, B. Winchester, Carbohydr. Res., 1990, 205, 269-282 [104] G. R Kaushal, I. Pastuszak, K-L Hatanaka, A. D. Elbein, J. Biol. Chem., 1990, 265, 16271-16279 [105] A. D. Elbein, J. E. Tropea, M. Mitchell, G. R Kaushal, J. Biol. Chem., 1990, 265, 15599-15605 [106] R Scudder, D. C. A. Neville, T. D. Butters, G. W. J. Fleet, R. A. Dwek, T. W. Rademacher, G. S. Jacob, J. Biol. Chem., 1990, 265, 16472-16477 [107] B. Winchester, C. Barker, S. Baines, G. S. Jacob, S. K. Namgoong, G. Fleet, Biochem. J., 1990,265,277-282

Tables of Glycosidase Inhibitors

385

[108] B. G. Winchester, I. Cenci di Bello, A. C. Richardson, R. J. Nash, L. E. Fellows, N. G. Ramsden, G. W. J. Fleet, Biochem. J., 1990, 269, 227-231 [109] P. Greenberg, A. H. Merrill, D. C. Liotta, G. A. Grabowski, Biochim. Biophys. Acta, 1990, 1039, 12-20 [110] K. Tatsuta, Y. Niwata, K. Umezawa, K. Toshima, M. Nadata, /. Antibiot., 1991, 912-914 [111] R. J. Molyneux, Y. T. Pan, J. E. Tropea, M. Benson, G. R Kaushal, A. D. Elbein, Biochemistry, 1991, 30, 9981-9987 [112] H. Yoon, S. B. King, B. Ganem, Tetrahedron Lett., 1991, 32, 7199-7202 [113] J. Wagner, P. Vogel, Tetrahedron Lett., 1991, 32, 3169-3170 [114] P. S. Liu, M. S. Kang, P. S. Sunkara, Tetrahedron Lett., 1991, 32, 719-720 [115] J. Wagner, P. Vogel, Tetrahedron, 1991, 47, 9641-9658 [116] L. J. Liotta, J. Lee, B. Ganem, Tetrahedron, 1991, 47, 2433-2447 [117] A. J. Fairbanks, G. W J. Fleet, A. H. Jones, I. Bruce, S. Al Daher, I. Cenci di Bello, B. Winchester, Tetrahedron, 1991, 47, 131-138 [118] A. C. de S. Pereira, M. A. C. Kaplan, J. G . S. Maia, O. R. Gottlieb, R. J. Nash, G. W. J. Fleet, L. Pearce, D. J. Watkin, A. M. Scofield, Tetrahedron, 1991, 47, 5637-5640 [119] K. K.-C. Liu, T. Kajimoto, L. Chen, Z. Zhong, Y. Ichikawa, C.-H. Wong, /. Org. Chem., 1991, 56, 6280-6289 [120] J.-L. Reymond, A. A. Pinkerton, P. Vogel, J. Org. Chem., 1991, 56, 2128-2135 [121] A. Frankowski, C. Seliga, D. Bur, J. Streith, HeIv. Chim. Acta, 1991, 74, 934-940 [122] T. Kajimoto, K. K-C. Liu, R. L. Pederson, Z. Zhomg, Y. Ichikawa, J. A. Porco, C-H. Wong, /. Am. Chem. Soc., 1991, 113, 6187-6196 [123] E. Bause, A. Gross, J. Schweden, FEBS Lett., 1991, 278, 167-170 [124] G. Legler, E. Liillau, E. Kappes, F. Kastenholz, Biochim. Biophys. Acta, 1991, 1080, 89-95 [125] S. E Moss, S. L. Vallance, /. Chem. Soc., Perkin Trans. I, 1992, 1959-1967 [126] S. Takahashi, H. Kuzuhara, Chem. Lett., 1992, 21-24 [127] A. Berger, K. Dax, G. Gradnig, V. Grassberger, A. E. Stiitz, M. Ungerank, Bioorg. Med. Chem. Lett, 1992, 2, 27-32 [128] O. Duclos, A. Dureault, J. C. Depezay, Tetrahedron Lett., 1992, 33, 1059-1062 [129] Y. Chen, P. Vogel, Tetrahedron Lett., 1992, 33, 4917-4920 [130] P. Herczegh, I. Kovacs, L. Szilagyi, M. Zsely, F. Sztaricskai, Tetrahedron Lett., 1992, 33,3133-3136 [131] Y. Nishimura, S. Kondo, T. Takeuchi, T. Kudo, /. Antibiot., 1992, 45, 954-962 [132] Y. Nishimura, S. Kondo, T. Takeuchi, T. Kudo, J. Antibiot., 1992, 45, 963-970 [133] T. Aoyagi, H. Suda, K. Uotani, F. Kojima, T. Aoyama, K. Horiguchi, M. Hamada, T. Takeuchi, J. Antibiot., 1992, 45, 1404-1408 [134] T. Aoyama, H. Naganawa, H. Suda, K. Uotani, T. Aoyagi, T. Takeuchi, J. Antibiot., 1992, 45, 1557-1558 [135] K. Burgess, I. Henderson, Tetrahedron, 1992, 48, 4045-4066 [136] A. J. Fairbanks, N. C. Carpenter, G. W. J. Fleet, N. G. Ramsden, I. Cenci di Bello, B. G. Winchester, S. S. Al-Daher, G. Nagahashi, Tetrahedron, 1992, 48, 3365-3376 [137] I. Bruce, G. W. J. Fleet, I. Cenci di Bello, B. Winchester, Tetrahedron, 1992, 48, 1019110200 [138] K. Adelhorst, G. M. Whitesides, Carbohydr. Res., 1992, 232, 183-187 [139] N. Chida, Y. Furuno, H. Ikemoto, S. Ogawa, Carbohydr. Res., 1992, 237, 185-194 [140] Y-T. Pan, G. R Kaushal, G. Papandreou, B. Ganem, A. D. Elbein, J. Biol Chem., 1992, 267, 8313-8318 [141] S. F. Moss, R. Southgate, J. Chem. Soc., Perkin Trans. I, 1993, 1787-1794 [142] S. Picasso, Y. Chen, R Vogel, Carbohydr. Lett, 1993, 1, 1-8 [143] H. Ardron, T. D. Butters, F. M. Platt, M. R. Wormald, R. A. Dwek, G. W. J. Fleet, G. S. Jacob, Tetrahedron Assym., 1993, 4, 2011-2024 [144] G. C. Look, C. H. Fotsch, C-H. Wong, Ace. Chem. Res., 1993, 26, 182-190 [145] R. H. Furneaux, R C. Tyler, L. A. Whitehouse, Tetrahedron Lett., 1993, 34, 3609-3612

386

Tables of Glycosidase Inhibitors

[146] S. Knapp, Y. H. Choe, E. Reilly, Tetrahedron Lett., 1993, 34, 4443-4446 [147] R. H. Furneaux, G. P. Lynch, G. Way, B. Winchester, Tetrahedron Lett., 1993, 34, 3477-3480 [148] Y.-F. Wang, D. R Dumas, C.-H. Wong, Tetrahedron Lett., 1993, 34, 403-406 [149] R. J. Nash, M. Rothschild, E. A. Porter, A. A. Watson, R. D. Waigh, P. G. Waterman, Phytochemistry, 1993, 34, 1281-1283 [150] G. Rassu, G. Casiraghi, L. Pinna, P Spanu, F. Ulgheri, M. Cornia, F. Zanardi, Tetrahedron, 1993, 49, 6627-6636 [151] S. Maruo, Y. Kyotani, H. Yamamoto, K. Miyazaki, H. Ogawa, T. Sakai, M. Kojima, Y. Ezure, BioscL Biotech. Biochem., 1993, 57, 1294-1298 [152] S. Washiyama, A. Kamiya, S. Esaki, A. Tanaka, N. Sugiyama, S. Kamiya, BioscL Biotech. Biochem., 1993, 57, 847-849 [153] S. F. Martin, H.-J. Chen, C.-P Yang, J. Org. Chem., 1993, 58, 2867-2873 [154] Y Takaoka, T. Kajimoto, C.-H. Wong, J. Org. Chem., 1993, 58, 4809-4812 [155] P Zhou, H. Mohd. Salleh, J. F. Honek, J. Org. Chem., 1993, 58, 264-266 [156] R. Hoos, A. B. Naughton, W. Thiel, A. Vasella, W. Weber, HeIv. Chim. Acta, 1993, 76, 2666-2686 [157] C-K. Lee, H. Jiang, L. L. Koh, Y Xu, Carbohydr. Res., 1983, 239, 309-315 [158] P A. Fowler, A. H. Haines, R. J. K. Taylor, E. J. T. Chrystal, M. B. Gravestock, Carbohydr. Res., 1993, 246, 377-381 [159] P Ermert, A. Vasella, M. Weber, K. Rupitz, S. G. Withers, Carbohydr. Res., 1993, 250, 113-128 [160] B. Liessem, A. Giannis, K. Sandhoff, M. Nieger, Carbohydr. Res., 1993, 250, 19-30 [161] S. Hiraizumi, U. Spohr, R. G. Spiro, J. Biol. Chem., 1993, 268, 9927-9935 [162] B. Winchester, S. Al Daher, N. C. Carpenter, I. Cenci di Bello, S. S. Choi, A. J. Fairbanks, G. W. J. Fleet, Biochem. J., 1993, 290, 743-749 [163] R. J. Molyneux, Y. T. Pan, A. Goldman, D. A. Topfer, A. D. Elbein, Arch. Biochem. Biophys., 1993,304, 81-88 [164] M. H. M. G. Schumacher-Wandersleb, St. Petersen, M. G. Peter, Liebigs Ann. Chem., 1994, 555-561 [165] K. E. Holt, F. J. Leeper, S. Handa, J. Chem. Soc., Perkin Trans. I, 1994, 231-234 [166] Y. Miyake, M. Ebata, Agric. Biol. Chem., 1988, 52, 153-158 [167] J. M. J. Tronchet, M. Balkadjian, F. Barbalat-Rey, G. Zosimo-Landolfo, I. Komaromi, J. Chem. Res., 1994, 20-21 [168] A. R. Beacham, K. Biggadike, H. E. Taylor, L. Hackett, B. G. Winchester, J. Chem. Soc., Chem. Commun., 1994, 2001-2002 [169] M. S. Chorghade, C. T. Cseke, P S. Liu, Tetrahedron Assym., 1994, 5, 2251-2254 [170] A. E. McCaig, B. Chomier, R. H. Wightman, J. Carbohydr. Chem., 1994, 13(3), 397-407 [171] K. Suzuki, H. Hashimoto, Tetrahedron Lett., 1994, 35, 4119-4122 [172] D. J. A. Schedler, B. R. Bowen, B. Ganem, Tetrahedron Lett., 1994, 35, 3845-3848 [173] C. H. Fotsch, C.-H. Wong, Tetrahedron Lett., 1994, 35, 3481-3484 [174] Y. Bleriot, A. Genre-Grandperre, C. Tellier, Tetrahedron Lett., 1994, 35, 1867-1870 [175] N. Asano, K. Oseki, H. Kizu, K. Matsui, J. Med. Chem., 1994, 37, 3701-3706 [176] Y. Nishimura, S. Kondo, T. Takeuchi, T. Kudo, J. Antibiot., 1994, 47, 101-107 [177] T. Tschamber, F. Backenstrass, M. Neuburger, M. Zehnder, J. Streith, Tetrahedron, 1994,50, 1135-1152 [178] P. Herczegh, I. Kovacs, L. Szilagyi, F. Sztaricskai, Tetrahedron, 1994, 50, 1367113686 [179] R. A. Farr, A. K. Holland, E. W. Huber, N. P. Peet, P. M. Weintraub, Tetrahedron, 1994, 50, 1033-1044 [180] I. Lundt, R. Madsen, S. Al Daher, B. Winchester, Tetrahedron, 1994, 50, 7513-7520 [181] Y. Konishi, A. Okamoto, J. Takahashi, M. Aitani, N. Nakatani, BioscL Biotech. Biochem., 1994, 58, 135-139

Tables of Glycosidase Inhibitors

387

[182] T. M. Jespersen, W. Dong, M. R. Sierks, T. Skrydstrup, I. Lundt, M. BoIs, Angew. Chem., 1994, 106, 1858-1860 [183] Y.-F. Wang, Y. Takaoka, C.-H. Wong, Angew. Chem., 1994, 106, 1343-1345 [184] W. Zou, W A. Szarek, Carbohydr. Res., 1994, 254, 25-33 [185] N. Asano, K. Oseki, E. Kaneko, K. Matsui, Carbohydr. Res., 1994, 258, 255-266 [186] N. Asano, K. Oseki, E. Tomioka, H. Kizu, K. Matsui, Carbohydr. Res., 1994, 259, 243-255 [187] F. M. Platt, G. R. Neises, G. B. Karlsson, R. A. Dwek, T. D. Butters, /. Biol. Chem., 1994,269,27108-27114 [188] T. Oishi, T. Iwakuma, M. Hirama, S. Ito, Synlett, 1995, 404-406 [189] A. Defoin, H. Sarazin, J. Streith, Synlett, 1995, 1187-1188 [190] M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem., 1995, 3, 161-165 [191] Y. Bleriot, T. Dintinger, A. Genre-Grandpierre, M. Padrines, C. Tellier, Bioorg. Med. Chem. Lett., 1995, 5, 2655-2660 [192] K. Tatsua, M. Kitagawa, Tetrahedron Lett., 1995, 36, 6717-6720 [193] O. R. Martin, L. Liu, F. Xie, Tetrahedron Lett., 1995, 36, 4027-4030 [194] O. R. Martin, O. M. Saavedra, Tetrahedron Lett., 1995, 36, 799-802 [195] M. Ichikawa, Y. Igarashi. Y. Ichikawa, Tetrahedron Lett., 1995, 36, 1767-1770 [196] B. Davis, T. W. Brandstetter, C. Smith, L. Hackett, B. G. Winchester, G. W. J. Fleet, Tetrahedron Lett., 1995, 36, 7507-7510 [197] A. W.-Y Chan, B. Ganem, Tetrahedron Lett., 1995, 36, 811-814 [198] K. Tatsuta, S. Miura, Tetrahedron Lett., 1995, 36, 6721-6724 [199] G. Mikkelsen, T. V. Christensen, M. BoIs, I. Lundt, Tetrahedron Lett., 1995, 36, 65416544 [200] K. Tatsuta, S. Miura, S. Ohta, H. Gunji, Tetrahedron Lett., 1995, 36, 1085-1088 [201] T. W. Brandstetter, B. Davis, D. Hyett, C. Smith, L. Hackett, B. G. Winchester, G. W. J. Fleet, Tetrahedron Lett., 1995, 36, 7511-7514 [202] Y. Bleriot, T. Dintinger, N. Guillo, C. Tellier, Tetrahedron Lett., 1995, 36, 5175-5178 [203] Y. Ichikawa, Y. Igarashi, Tetrahedron Lett., 1995, 36, 4585-4586 [204] C. R. Johnson, A. Golebiowski, H. Sundram, M. W. Miller, R. L. Dwaihy, Tetrahedron Lett., 1995, 36, 653-654 [205] N. Asano, H. Kizu, K. Oseki, E. Tomioka K. Matsui, M. Okamoto, M. Baba, J. Med. Chem., 1995, 38, 2349-2356 [206] Y. Suhara, K. Achiwa, Chem. Pharm. Bull., 1995, 43(3), 414-420 [207] K. Tatsuta, S. Miura, S. Ohta, H. Gunji, /. Antibiot., 1995, 48, 286-288 [208] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, Tetrahedron, 1995, 51, 12611-12630 [209] S. Watanabe, H. Kato, K. Nagayama, H. Abe, Biosci. Biotech. Biochem., 1995, 59, 936-937 [210] C-H. Wong, L. Provencher, J. A. Porco, S-H. Jung, Y-F. Wang, L. Chen, R. Wang, D. H. Steensma, J. Org. Chem., 1995, 60, 1492-1501 [211] A. Brandi, S. Cicchi, F. M. Cordero, R. Frignoli, A. Goti, S. Picasso, P. Vogel, /. Org. Chem., 1995, 60, 6806-6812 [212] T. D. Heightman, P. Ermert, D. Klein, A. Vasella, HeIv. Chim. Acta., 1995, 78, 514-532 [213] N. Asano, A. Kato, K. Oseki, H. Kizu, K. Matsui, Eur. J. Biochem., 1995, 229, 369-376 [214] J. Lehmann, B. Rob, Carbohydr. Res., 1995, 272, C11-C13 [215] G. Legler, A. E. Stutz, H. Immich, Carbohydr. Res., 1995, 272, 17-30 [216] J. Lehmann, B. Rob, H.-A. Wagenknecht, Carbohydr. Res., 1995, 278, 167-180 [217] D.-K. Kim, G. Kim, Y.-W. Kim, /. Chem. Soc., Perkin Trans. I, 1996, 803-808 [218] H.-D. Stachel, K. Zeitler, S. Dick, Liebigs Ann., 1996, 103-107 [219] C-K. Lee, H. Jiang, A. Linden, A. Scofield, Carbohydr. Lett., 1996, 1, 417-423 [220] X. Qian, F. Moris-Varas, M. C. Fitzgerald, C.-H. Wong, Bioorg. Med. Chem., 1996, 4, 2055-2069 [221] M. Godskesen, I. Lundt, R. Madsen, B. Winchester, Bioorg. Med. Chem., 1996, 4, 1857-1865

388

Tables of Glycosidase Inhibitors

[222] T. Satoh, Y. Nishimura, S. Kondo, T. Takeuchi, T. Kudo, Bioorg. Med. Chem., 1996, 4, 91-96 [223] Y. Igarashi, M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem. Lett., 1996, 6, 553-558 [224] D. Damour, M. Barreau, J.-C. Blanchard, M.-C. Burgevin, A. Doble, F. Herman, G. Pantel, E. James-Surcouf, M. Vuilhorgne, S. Mignani, L. Poitout, Y. Le Merrer, J.-C. Depezay, Bioorg. Med. Chem. Lett., 1996, 6, 1667-1672 [225] X. Qian, F. Moris-Varas, C.-H. Wong, Bioorg. Med. Chem. Lett., 1996, 6, 1117-1122 [226] A. Goti, F. Cardona, A. Brandi, S. Picasso, P. Vogel, Tetrahedron Assym., 1996, 7, 1659-1674 [227] B. Ganem, Ace. Chem. Res., 1996, 29, 340-347 [228] E. P. Mitchell, S. G. Withers, P. Ermert, A. T. Vasella, E. F. Garman, N. G. Oikonomakos, L. N. Johnson, Biochemistry, 1996, 35, 7341-7355 [229] W. Dong, T. Jespersone, M. BoIs, T. Skrydstrup, M. R. Sierks, Biochemistry, 1996, 35, 2788-2795 [230] L. Sun, P. Li, D. W. Landry, K. Zhao, Tetrahedron Lett., 1996, 37, 1547-1550 [231] L. Poitout, Y. Le Merrer, J.-C. Depezay, Tetrahedron Lett., 1996, 37, 1609-1612 [232] E. Frerot, C. Marquis, P. Vogel, Tetrahedron Lett., 1996, 37, 2023-2026 [233] M. P. Persson, W. M. Butt, M. Jorgensen, P. Christensen, L. T. Hansen, M. BoIs, Tetrahedron Lett., 1996, 37, 2097-2100 [234] O. R. Martin, L. Liu, F. Yang, Tetrahedron Lett., 1996, 37, 1991-1994 [235] K. Suzuki, T. Fujii, K.-I. Sato, H. Hashimoto, Tetrahedron Lett., 1996, 37, 5921-5924 [236] A. A. Bell, L. Pickering, A. A. Watson, R. J. Nash, R. C. Griffiths, M. G. Jones, G. W. J. Fleet, Tetrahedron Lett., 1996, 37, 8561-8564 [237] B. Davis, A. A. Bell, R. J. Nash, A. A. Watson, R. C. Griffiths, M. G. Jones, C. Smith, G. W. J. Fleet, Tetrahedron Lett., 1996, 37, 8565-8568 [238] J. P. Shilvock, J. R. Wheatley, B. Davis, R. J. Nash, R. C. Griffiths, M. G. Jones, M. Muller, S. Crook, D. J. Watkin, C. Smith, G. S. Besra, P. J. Brennan, G. W. J. Fleet, Tetrahedron Lett., 1996, 37, 8569-8572 [239] A. Baudat, P. Vogel, Tetrahedron Lett., 1996, 37, 483-484 [240] Y. Igarashi, M. Ichikawa, Y. Ichikawa, Tetrahedron Lett., 1996, 37, 2707-2708 [241] J. J. Mclntyre, A. T. Bull, A. W. Bunch, Biotechnol. Bioeng., 1996, 49, 412-420 [242] T. Tsuruoka, H. Fukuyasu, M. Ishi, T. Usui, S. Shibahara, S. Inouye, J. Antibiot., 1996, 49, 155-161 [243] T. Takatsu, M. Takahashi, Y. Kawase, R. Enokita, T. Okazaki, H. Matsukawa, K. Ogawa, Y. Sakaida, T. Kagasaki, T. Kinoshita, M. Nakajima, K. Tanzawa, J. Antibiot., 1996, 49, 54-60 [244] T. Satoh, Y. Nishimura, S. Kondo, T. Takeuchi, /. Antibiot., 1996, 49, 321-325 [245] Y. Kawase, M. Takahashi, T. Takatsu, M. Arai, M. Nakajima, K. Tanzawa, J. Antibiot., 1996, 49, 61-64 [246] K. Tatsuta, Y. Ikeda, S. Miura, /. Antibiot., 1996, 49, 836-838 [247] L. A. G. M. van den Broek, Tetrahedron, 1996, 52, 4467-4478 [248] A. Goldmann, B. Message, D. Tapfer, R. J. Molyneux, O. Duclos, R-D. Boyer, Y. T. Pan, A. D. Elbein, /. Nat. Prod., 1996, 59, 1137-1142 [249] W. H. Pearson, E. J. Hembre, J. Org. Chem., 1996, 61, 5546-5556 [250] W. H. Pearson, E. J. Hembre, J. Org. Chem., 1996, 61, 5537-5545 [251] O. M. Saavedra, O. R. Martin, /. Org. Chem., 1996, 61, 6987-6993 [252] F. L. van Delft, M. de Kort, G. van. der Marel, J. H. van Boom, J. Org. Chem., 1996, 61, 1883-1885 [253] K. Tatsuta, Pure Appl. Chem., 1996, 68, 1341-1346 [254] T. D. Heightman, M. Locatelli, A. Vasella, HeIv. Chim. Acta., 1996, 79, 2190-2200 [255] L. Qiao, B. W. Murray, M. Shimazaki, J. Schultz, C.-H. Wong, J. Am. Chem. Soc., 1996, 118, 7653-7662 [256] J.-H. Jeong, B. W. Murray, S. Takayama, C.-H. Wong, J. Am. Chem. Soc., 1996, 118, 4227-4234 [257] F. Moris-Varas, X.-H. Qian, C.-H. Wong, /. Am. Chem. Soc., 1996, 118, 7647-7652

Tables of Glycosidase Inhibitors

389

[258] L. J. Mazzella, D. W. Parkin, R C. Tyler, R. H. Furneaux, V. L. Schramm, /. Am. Chem. Soc., 1996,118,2111-2112 [259] Y.-F. Liao, A. LaI, K. W. Moremen, /. Biol Chem., 1996, 271, 348-358 [260] A. Baudat, S. Picasso, R Vogel, Carbohydr. Res., 1996, 281, 277-284 [261] V. Zsoldos-Mady, I. Pinter, R Sandor, A. Messmer, Carbohydr. Res., 1996, 281, 321-326 [262] N. Asano, A. Kato, Y. Yokoyama, M. Miyauchi, M. Yamamoto, H. Kizu, K. Matsui, Carbohydr. Res., 1996, 284, 169-178 [263] T. Satoh, Y Nishimura, S. Kondo, T. Takeuchi, Carbohydr. Res., 1996, 286, 173-178 [264] G. Gradnig, G. Legler, A. E. Stiitz, Carbohydr. Res., 1996, 287, 49-57 [265] G. Legler. M.-T, Finken, Carbohydr. Res., 1996, 292, 103-115 [266] G. Legler. M.-T, Finken, S. Felsch, Carbohydr. Res., 1996, 292, 91-101 [267] N. Asano, A. Kato, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, Carbohydr. Res., 1996, 293, 195-204 [268] S. Weng, R. G. Spiro, Arch. Biochem. Biophys., 1996, 325, 113-123 [269] C. J. Vaughan, M. B. Murphy, B. M. Buckley, Lancet, 1996, 348, 1079-1082 [270] B. Kramer, T. Franz, S. Picasso, P. Pruschek, V. Jager, Synlett, 1997, 295-297 [271] C. Thomassigny, K. Bennis, J. Gelas, Synthesis, 1997, 2, 191-194 [272] M. BoIs, R. G. Hazell, I. B. Thomsen, Chem. Eur. J., 1997, 3, 940-947 [273] Y L. Merrer, L. Poitout, J-C. Depezay, I. Dosbaa, S. Geoffrey, M-J. Foglietti, Bioorg. Med. Chem., 1997, 5, 519-533 [274] B. Lesur, J-B. Ducep, M-N. Lalloz, A. Ehrhard, C. Danzin, Bioorg. Med. Chem. Lett., 1997, 7, 355-360 [275] A. Defoin, T. Sifferlen, I. Dosbaa, M. J. Foglietti, Tetrahedron Assym., 1997, 8, 363-366 [276] D. W. Parkin, G. Limberg, P. C. Tyler, R. H. Furneaux, X. Y. Chen, V. L. Schramm, Biochemistry, 1997, 36, 3528-3534 [277] R. E. Lee, M. D. Smith, R. J. Nash, R. C. Griffiths, M. McNeil, R. K. Grewal, W. Yan, G. S. Besra, P. J. Brennan, G. W. J. Fleet, Tetrahedron Lett., 1997, 38, 6733-6736 [278] T. M. Wrodnigg, A. E. Stutz, S. G. Withers, Tetrahedron Lett., 1997, 38, 5463-5466 [279] A. A. Watson, R. J. Nash, M. R. Wormald, D. J. Harvey, St. Dealler, E. Lees, N. Asano, H. Kizu, A. Kato, R. C. Griffiths, A. J. Cairns, G. W. J. Fleet, Phytochemistry, 1997, 46, 255-259 [280] R. H. Furneaux, G. J. Gainsford, J. M. Mason, P. C. Tyler, Tetrahedron, 1997, 53, 245-268 [281] A. Defoin, H. Sarazin, J. Streith, Tetrahedron, 1997, 53, 13769-13782 [282] A. Defoin, H. Sarazin, J. Streith, Tetrahedron, 1997, 53, 13783-13796 [283] E. J. Hembre, W H. Pearson, Tetrahedron, 1997, 53, 11021-11032 [284] A. Hansen, T. M. Tagmose, M. BoIs, Tetrahedron, 1997, 53, 697-706 [285] A. Lohse, M. BoIs, Tetrahedron, 1997, 53, 6917-6924 [286] N. Asano, M. Nishida, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, J. Nat. Prod., 1997, 60, 98-101 [287] A. Kato, N. Asano, H. Kizu, K. Matsui, A. A. Watson, R. J. Nash, J. Nat. Prod., 1997, 60, 312-314 [288] I. B. Parr, B. A. Horenstein, J. Org. Chem., 1997, 62, 7489-7494 [289] T. Granier, F. Gaiser, L. Hintermann, A. Vasella, HeIv. Chim. Acta., 1997, 80, 14431456 [290] T. Granier, N. Panday, A. Vasella, HeIv. Chim. Acta., 1997, 80, 979-987 [291] S. Takayama, R. Martin, J. Wu, K. Laslo, G. Siuzdak, C.-H. Wong, J. Am. Chem. Soc., 1997, 119, 8146-8151 [292] L. Deng, O. D. Scharer, G. L. Verdine, J. Am. Chem. Soc., 1997, 119, 7865-7866 [293] R. J. Molyneux, J. N. Roitman, G. Dunnheim, T. Szumilo, A. D. Elbein, Arch. Biochem. Biophys., 1997, 251, 450-457 [294] R. Hoos, A. Vasella, K. Rupitz, S. G. Withers, Carbohydr. Res., 1997, 298, 291-298 [295] M. A. Szarek, X. Wu, W. A. Szarek, Carbohydr. Lett, 1997, 299, 165-170

390

Tables of Glycosidase Inhibitors

[296] S. M. Andersen, M. Ebner, C. W. Ekhart, G. Gradnig, G. Legler, I. Lundt, A. E. Stiitz, S. G. Withers, T. Wrodnigg, Carbohydr. Res., 1997, 301, 155-166 [297] N. Asano, A. Kato, H. Kizu, K. Matsui, R. C. Griffiths, M. G. Jones, A. A. Watson, R. J. Nash, Carbohydr. Res., 1997, 304, 173-178 [298] G. Limberg, I. Lundt, J. Zavilla, preprint, 1998 [299] T. M. Wrodnigg, A. E. Stiitz, S. G. Withers, unpublished results, 1998 [300] G. Legler, A. E. Stiitz, unpublished results, 1998 [301] G. Legler, A. Korth, A. Berger, C. W. Ekhart, G. Gradnig, A. E. Stiitz, Carbohydr. Res., 1993, 250, 67-77 [302] R Zhou, H. M. Salleh, R C. M. Chan, G. Lajoie, J. F. Honek, R T. Chandra, O. R Ward, Carbohydr. Res., 1993,239, 155-166 [303] R. C. Bernotas, Carbohydr. Res., 1987, 167, 312-316 [304] G. C. Kite, L. E. Fellows, G. W. J. Fleet, R S. Liu, A. M. Scofield, N. G. Smith, Tetrahedron Lett., 1988, 29, 6483-6486 [305] J. F. Witte, R. W McClard, Tetrahedron Lett., 1991, 32, 3927-3930 [306] D. A. Winkler, G. Holan, J. Med. Chem., 1989, 32, 2084-2089 [307] D. R Dumas, T. Kajimoto, K. K.-C. Liu, C.-H. Wong, D. B. Berkowitz, S. J. Danishefsky, Bioorg. Med. Chem. Lett., 1992, 2, 33-36 [308] R M. Myerscough, A. J. Faibanks, A. H. Jones, I. Bruce, S. S. Choi, G. W. J. Fleet, S. S. Al-Daher, L Ceni di Bello, B. Winchester, Tetrahedron, 1992, 48, 10177-10190 [309] S. Inouye, T. Tsuruoka, T. Ito, T. Niida, Tetrahedron, 1968, 24, 2125-2144 [310] G. Papandreou, M. K. Tong, B. Ganem, J. Am. Chem. Soc., 1993,115, 11682-11690 [311] I. Bruce, G. W. J. Fleet, I. Cenci di Bello, B. Winchester, Tetrahedron, 1992, 48, 1019110200 [312] G. W J. Fleet, N. G. Ramsden, R. A. Dwek, T. W. Rademacher, L. E. Fellows, R. J. Nash, D. St. C. Green, B. Winchester, J. Chem. Soc., Chem. Commun., 1988, 483-485 [313] G. W J. Fleet, A. N. Shaw, S. V. Evans, L. E. Fellows, J. Chem. Soc., Chem. Commun., 1985, 841-842 [314] R. C. Bernotas, M. A. Pezzone, B. Ganem, Carbohydr. Res., 1987, 167, 305-311 [315] R S. Liu, J. Org. Chem., 1987, 52, 4717-4721 [316] R S. Liu, R. S. Rogers, M. S. Kang, R S. Sunkara, Tetrahedron Lett., 1991, 32, 58535856 [317] R Ermert, A. Vasella, HeIv. Chim. Acta, 1991, 74, 2043-2053 [318] Saul, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys., 1984, 230, 668-675 [319] C. Danzin, A. Ehrhard, Arch. Biochem. Biophys., 1987, 257, 472-475 [320] R. Saul, J. R Chambers, R. J. Molyneux, A. D. Elbein, Arch. Biochem. Biophys., 1983, 227, 593-597 [321] R S. Sunkara, T. L. Bowlin, R S. Lin, A. Sjoerdsma, Biochem. Biophys. Res. Comm., 1987,148, 206-210 [322] D. R. R Tulsiani, T. M. Harris, O. Touster, J. Biol. Chem., 1982, 257, 7936-7939

Index

General Acquired Immune Deficiency Syndrom 20 Affinity chromatography 207 ff Affinity ligand 207 ff Aglycon - binding site, interaction with inhibitors 56 AIDS 20,226 Aldolase - Iminosugar synthesis 84 Alkaloids 216ff - Glycosidase inhibitors 216 ff Amination - Anomeric centre 72 -of the Chain 76 - Reductive 80 - "True" sugars 72 Anomeric -Configuration 191 -Position 112 Anti-cancer 22,125,150 Anti-diabetic 125 Anti-inflammatory 125, 151 Anti-malarial 125 Anti-viral 20 f, 125, 149ff Arthritis 151 Asymmetrisation, enzymatic 86 Baeyer-Villiger see Oxidation Biochemical evaluation 104 Biological activity 224 ff

Cell adhesion 227 Chagas' desease 227 Chemotaxonomy 15ff Cycloaddition - Iminosugar synthesis 85 Cyclodextrin ct-Glucanotransferase - Bacillus circulans 199 Diabetes mellitus 4, 5 Dreiding model 167,173 Effects - on insects 22 ff Electronic effects 191 Enzymatic - Asymmetrisation 86 - Hydroxylation 87 Feeding behavior 23 f Fortuitous - Binders 188 ff -Inhibitors 203 Free energy - Relationships, linear 194 ct-L-Fucosidase - Affinity chromatography 212 f o-Galactosidase -Inhibitors 159 General acid/base 91 Glucoamylase - Aspergillus niger 199 Glucose isomerase 76 Glucosidase - CC-D-, Yeast 197 - P-D-, Agrobacterium sp. 197, 202

Iminosugars us Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

392

Index

- Affinity chromatography 208 ff -Inhibitors 168 - Purification 208 ff Glycogen phosphorylase 198 Glycoprotein - Processing 19, 216, 227 — Inhibitors 216 ff, 236 ff of Glucosidases 236 ff of Mannosidases 240 ff Glycosidase - Affinity chromatography 207 ff -Ground state 193 - Inhibition 216 ff - Inhibitor --Alkaloid 216ff — Biological activity 224 ff --Isolation 218 --Natural sources 8ff, 218ff — Occurrence 218 — Therapeutic activity 226 f Castanospermine 226 Swainsonine 226 — Tight binding 194 - Mechanism, general 34 ff, 19 Iff - Mechanism, double displacement 192 - Inverting 35, 37, 191, 200, 202 - Purification by Affinity Chromatography 207 -p^a cycling 193 - Rate enhancement factor 33, 36 - Retaining 34, 191, 202 - Slow inhibition 63 - Transition state 101 ff — Analogs 188 Acarbose 199 --Mimics 188 ff — Mimicry 189,197 — Models 34 f — Structure 193 Boat 193 Half chair 193 Skew boat 193 — see also Transition state Glycosyl — Enzyme intermediate 192 - - Oxocarbenium ion 32, 37, 38, 44, 46 Hexosaminidase, N-Acetyl-β-ο- Affinity chromatography 211 History 1 HIV 149

- see also Human immunodeficiency virus Human immunodeficiency virus 20 f, 226 - see also HIV Hydrogen bonds - Inhibition enhancement at active site 46 f, 59 ff Hydroxylation - Enzymatic 87 - Sharpless Osmylation 87 Iminosugar - Preparation or Synthesis 68 f --Aldolases 84 — Aldonolactones 93 ff - - Cycloaddition 85 ff --Cyclic Dienes 86 ff Immunosuppressant 125 Inflammation 227 Inhibitor - Fortuitous 203 - o-Galactosidase 159 - D-Glucosidase 168ff — Structure-activity relationships 177 ff - Isolation 219 - o-Mannosidase 162 — Structure-activity relationships 165ff - Tight binding 194 — see also Glycosidase Inversion - pyramidal 191 — Ring nitrogen castanospermine 181 Inverting 191,200,202 Ion-pair - Formation in actives site —with basic inhibitors 42 — with cationic inhibitors 42 f —with Glycosyl oxocarbenium ion 37, 60 Isotope effects - α-Secondary kinetic deuterium 194 kcat 190 ff, 200, 203 Rvalue 190ff Km 190 ff, 200, 203 Lysosomal disease 21 Lysozyme 193

Chemical compounds and substances Mannopyranosyl oxocarbonium ion 164ff Mannosidase OC-D-, - from Almonds 197 - from Jack bean 197 β-D-, -from Snail 197 - Affinity chromatography 210 ff -Inhibitors 162 - Inhibition of oc-L-fucosidases 167 - Structure-activity relationships 165ff Mechanism, see Glycosidase Model -Dreiding 167,173 Moreton Bay chestnut - see Castanospermum australe Mutation 190,202

Structure-activity relationships 157 ff, 222 ff - o-Mannosidase inhibitors 165 ff - o-Glucosidase inhibitors 177 ff Thermodynamic box 189 Transition state - see also Glycosidase - 'Activated complex' 189 -Analogs 188 ff - Mimics 188 ff -Mimicry 189,197 - Structure 193 -Theory 189 - Transmission coefficient 189 Transition state resemblance - of Inhibitors 61 ff Wolfenden

Nematicidal 25 f Nucleophile 192 Oligosaccharides - N-Linked

--Biosynthesis 229

- - Complex type 228 f - - High-mannose type 228 - - Hybride type 229 --Processing 227, 230ff --Structural classes 228 f Osmylation -Sharpless 82 Oxidation - Baeyer-Villiger 89 Oxocarbenium ion 193,197 Parasitism 227 Pathogenesis 227 Plant growth regulatory 26 Pompe disease 147f Resolution, kinetic 85 Retaining 191, 202 Retention 191 Ring closure reactions 94 ff, 120 Sharpless - Epoxidation 126 - Osmylation - - see Osmylation S tereoselection -Acyclic 81

393

188

Xylanase - Bacillus circulans 193

Chemical compounds and

^ SllbstailCeS

Acarbose 3,146 - Glycosidase inhibition 37 - as a Transition state analog 199 L-Alanine 82 Alditols -Dibromo 94, 98 ff - Conversion into iminosugars 79 - Dihalogenated 80 Aldonolactones 93 Alexine 13, 18, 24 Allitol - l,5-dideoxy-l,5-imino-D- 159 Aminosugars - Iminosugars from 71 Arabinitol - l,4-dideoxy-l,4-imino-D- 164, 175 — see also D-AB1 - l,5-dideoxy-l,5-imino-D- 163 Ascorbic acid - 2-O-cc-D-glucopyranosyl 147 Australine 13, 16, 18, 24, 175, 217 - Glycosidase inhibition 55

394

Index

Azasugars - Five-membered, glycosidase inhibition 50 ff - N-Oxides of, Glycosidase inhibition 42 f - Seven-membered 53 - Six-membered — see Nojirimycin, 1-deoxy 1-Azasugars - Glycosidase inhibition 4Of - see also Isoiminosugars Azepanes, tetrahydroxy 162, 164, 176 Azidodeoxy lactones see Lactones AZT 149 - Bromodeoxy- 93,120 Bromodeoxyaldonolactones see Aldonolactones Calystegines 14 f, 18, 177, 218 - Glycosidase inhibition 56 Castanospermine 4, 5, 12, 14, 18, 20, 24 f, 172 f, 196, 217 - Analogs 125 ff - Biological Activity 144 ff - Glycosidase inhibition 55 - Ring nitrogen pyramidal inversion 181 - Stereoisomers 126 ff - Stucture-activity relationships 145 ff, 112 f, 181 - Synthesis 125 ff - 6-O-Butanoyl- 20 -Epi 88, 126 ff - 3 -(Hydroxymethyl)- 6 -epi- 71 Casuarine 13, 15, 18 CYB-3 Uf, 14, 18 Cyclitols 89 D-ABl 11,15,18,24 - see also Arabinitol, .,4 -dideoxy-1,4 -imino-DDeoxy noj irimy cin 1- 168ff, 195 - Glycosylated derivatives 169 f - Modifications 170 ff - seco-Derivatives 172 - see also Glucitol, .,5 -dideoxy-1,5 -imino-DDiene, cyclic - Iminosugar synthesis 86 ff Dimethyl L-tartrate 88 DMDP 11 ff, 15 ff, 22 ff, 147, 217, 225, 237,239

- see also Mannitol, 2,5 -dideoxy-2,5 -imino-oD-RBl 11 - see also Ribitol, 1,4 -dideoxy-1,4 -imino-DEpoxides 94 ff Fagomine 9f, 117 'Flap-up' o-mannopyranosyl oxocarbonium ion 164 ff Fuconojirimycin, 1-Deoxy-L- 119, 163, 167 L-Fucose -Analog 71 oc-L-Fucosidase 71, 163 - Inhibition by D-mannosidase inhibitors 167 Galactitol - 1,5-Dideoxy-1,5-imino-D- 159f — see also Galactonojirimycin - 2,5-Dideoxy-2,5-imino-o- 160 Galactonojirimycin - 1-Deoxy 159 D-Galactosamine 8,71 Galactose - 5-Amino-5-deoxy 159 — see also Galactostatin Galactostatin 9 - see also Galactose, 5 -Amino-5 -deoxy-D- Deoxy- 119 Glucitol - 1,5-Anhydro-D- 2 - 1,5-Dideoxy-1,5-imino-D- 168 Glucono-l,5-lactone-o- 198 o-Glucosamine 8 Glucosimine 9 α-Glucosyl fluoride 200 L-Glyceraldehyde 83 Glyconamidine 196 - Glycosidase inhibition 47 Glyconamidrazones 196 - Glycosidase inhibition 47, 60 Glycono-l,5-lactam oximes 196 Glycono-1,5 -lactams - see also Hexonolactams - Glycosidase inhibition 44 f Glycono-1,5-lactone oximes 196 - N-Phenylcarbamates 196

Chemical compounds and substances Glycono-1,5 -lactons - Glycosidase inhibition 44 f Glyconolactones 194 - Reduction 80 - Synthesis of iminosugars 75 Glycoside - Deoxy 194 -Deoxyfluoro 194 -1,1 -Difluoro 194 Glycosyl fluorides 194, 200 Glycosylamines 72, 195 - Glycosidase inhibition 38 ff, 42 - N-Alkyl derivatives 39,57 Glycosylmethylamines - Glycosidase inhibition 40 Glycotetrahydropyrido[l,2d]imidazoles - see Nojiritriazoles Glycotetrahydropyrido[ 1,2d]tetrazoles - see Nojiritetrazoles Glycotetrahydropyrido[l,2d]triazoles - see Nojiritriazoles Guanidines -Cyclic — see Guanidinosugars Hexitol - 5 -Amino-5 -deoxy-1,5 -anhydro 2 — see also 1-Deoxynojirimycin, 1,5 -Dideoxy-1,5 -iminohexitol - 1,2,5 -Trideoxy-1,5 -immo-O-ribo 16Of Hexitols - Biochemical Evaluation — 1,4-Dideoxy-1,4-imino 104f — 1,5-Dideoxy-1,5-imino 106 ff --2,5-Dideoxy-2,5-imino 104 ff Hexonolactams 44 f - 5-Amino-5-deoxy-D- and derivatives — see also Glycono-1,5-lactams and derivatives Hexose - 5-Amino-5-deoxy- 38, 60 — see also Nojirimycin Homomannonojirimycin 9 f Homonojirimycin 9 f, 15 ff, 21 ff Hy droximolactams - see Glycono-l,5-lactam oximes Hydroximolactams - see Glycono-l,5-lactam oximes Hydroximolactones - see Glycono-1,5-lactone oximes

395

Iditol -1,4 -Dideoxy-1,4 -imino-L- 161 - 1,6 -Dideoxy-1,4 -imino-L- 176 Iminoalditols - as Affinity Ligands 207 ff - Purification of Glycosidases 207 ff - Structure-activity relationships 157ff Iminoamides 94 ff Iminosugars - from Alditols 80 - from Aminosugars 71 - from Dibromoalditols 98 ff - from Iminoamides 95 ff - from Lactams 100 ff - Structure-activity relationships 157 ff Indolizidine 217 - Pentahydroxy 132 ff Isofagomine 113, 117 ff, 176, 195 -6-Deoxy 120ff - Glycosidase inhibition 40 -N-Substituted 57f - Epimers 160 Isoiminosugars 112 ff - see also 1-Azasugars, Isofagomine, Siastatin Kifunensine 243 Lactams 94, 100ff, 114, 120 Lactones - Azidodeoxy- 94,103,114 Lentiginosine 12 Lyxitol - 1,4-Dideoxy-1,4-imino-D- 16Of, 163 - 1,5 -Dideoxy-1,5 -imino-D— see also Arabinitol, 1,5-Dideoxy-1,5-imino-Dcc-Maltotriosyl fluoride 200 Mannitol - 1,4 -Dideoxy-1,4 -imino-D- 163 -1,5 -Dideoxy-1,5 -imino-D— see also Mannonojiriymycin, 1 -Deoxy - 1,5-Dideoxy-1,5-imino-L- 160 - 2,5-Dideoxy-2,5-imino-D- 160, 173 — see also DMDP - 1,4,6-Trideoxy-1,4-imino-D- 165 Mannonojirimycin 162 - Biosynthesis 69

396

Index

- 1-Deoxy 9f, 16, 18, 20, 22, 118 — see also Mannitol, 1,5 -dideoxy-1,5 -imino Mannopyranosyl - Oxocarbonium ion,'flap-up' 164 ff Miglitol 146, 169, 226 Moranoline 14

Tetrahydroxyazepanes 162, 164, 176 Tetritols, Imino - Biochemical evaluation 104 f Thiazole 81

NANA - Aldolase 71 Nojiriimidazoles - GIycosidase inhibition 38,47 Nojirimycin 9 ff, 15, 18, 168, 195 -B 9f, 18 - - see also Mannonojirimycin, 1-deoxy 3 - 1-Deoxy- 118, 168, 195 - - Glycosidase inhibition 37, 39 - - N-Substituted 39, 57 - 1-Deoxygalacto 159 - Glycosidase inhibition 38,60 Nojiritetrazoles 196ff - Glycosidase inhibition 46 f Nojiritriazoles - Glycosidase inhibition 46 f rcorTropane 217

Xylopyranose - 5 -Acetamido-5 -deoxy-D- 1 -5-Amino-D- 2 -5-Thio-D- 1

Oxiranes - see Epoxides Piperidines 195,217 Pyrrolidines 195,217 Pyrrolizidine 217 L-Quebrachitol 89 Quinolizidine - Tetrahydroxy 134f Ribitol - 1,4 -Dideoxy-1,4 -imino-D- 161 — see also D-RB1 L-Serinal 81 D-Serine 83 SiastatinB 113, 114ff Sorbitol -2,6-Anhydro-L- 2 Sucrose - 6,6'-Dichloro-6,6'-dideoxy- 76 Swainsonine 5, 11, 18, 20ff, 25f, 164, 196,217 - Glycosidase inhibition 55

Validamycin A 4, 5 Valienamine 3, 5, 199, 201

Genera and Aglaodorum 16 Aglaonema 9, 15 f Agrobacterium sp. - β-Glucosidase 197,202 Alcides metaurus 23 15,218 Aiexa - grandiflora 13 -leiopetala 13,125 Anchomanes 16 Angylocalix 16 - boutiqueanus 11 Arachniodes standishii 11 Aspergillus -niger 159, 162 - - Glucoamylase 199 - oryzae 160 -wentii 181 Astragalus 21, 164, 224 - lentiginosus 12,21 - oxyphysus 219 Bacillus 9 -circulans 193 _ _ Cyclodextrin cc-glucanotransferase 199 Callosobruchus maculatus 22 f Calystegia sepium 14, 218 f Castanospermum 16 _ australe 11 f, 16, 125, 136, 146, 218 Casuarina equisetifolia 13 Cytomegalovims 20 Derris 16 -elliptica 11 Duboisa leichhardtii 14

Genera and species Endospermum

Fagopyrum esculentum 9

Physalis alkekengi 14 f Plasmodium falciparum 227 Pseudohydrosme 16

Globodera 25 Gluconobacter suboxidans 169

Rhizobium meliloti 219 Rhizoctonia leguminicola 12, 218 f

Hyacinthus orientalis 17 Hyoscyamus niger 15

Schistocerca gregaria Forsk 147 Scopolia japonica 14 Spodoptera littoralis 23 ff Streptomyces 9, 218 - lavandulae 9 - lydicus 9, 159 - nojiriensis 9 - roseochromogenes 9 Streptoverticillium verticillus 244 Swainsona 12, 21, 224 - canescens 12,21,164 Syzygiumjambolana 17

Ipomoea

16,23

218

Kitasatosporia kifunense

243

Locusta migratoria L. 147 Lonchocarpus 11, 16 f, 22 - costaricensis 9, 162 - sericeus 9,162 Lyssa macleayi 23 Metarhizium anisopliae 219 Moloney murine leukemia virus 20 9, 15, 168, 218 11,14 Nephthytis

16

Omphalea 22, 25 -diandra 9,15,16,22 - queenslandiae 23 Oxytropis 21,224 - kansuensis 21 - ochrocephala 21

Trypanosoma cruzi 227, 244 Urania fulgens

16, 22 f, 25, 218

Xanthocercis zambesiaca 15 f Xiphinema diversicaudatum 25

397

rf.

r~~^

-«r

Hi

-^.

<

o

PJ

C:

C/2

EL m

Q.

g

S

1

tO

S5 3

o p> S'

5

i-I

a? 1 O >

jj-

*<

§ € oI Z. *<

I

HO

HO

1

:

1

1

: <5H

^^^

HO

HO

HO

y—

CH OH

v— CO CH

>v C02CH3

^^

HO ,.. X^N--^

Compound: /Of 90 Synthetic/NMR: [289; X-Ray: /2897

chiral pool chemical

(5R,6R,7S,8S)-5,6-7,8-Tetrahydro6,7,8-trihydroxyindolizine-1,2,5trimethanol chiral pool chemical Compound: /Of 70 Synthetic/NMR: [289] Methyl-(5R,6R,7S,8S)-5,6,7,8tetrahydro-6,7,8-trihydroxy-5(hydroxy methy l)indolizine-1 carboxylate chiral pool chemical Compound: /Of 80 Synthetic/NMR: [289] X-Ray: [289] Methyl-(5R,6R,7S,8S)-5,6,7,8tetrahydro-6,7,8-trihydroxy-5(hydroxymethyl)indolizJne-2carboxylate liiiflWIliM-iBEHg EC: 3.2.1 .21 r sweet almonds Ki: 6000 (6.8) [289].

liliff»lllf»M-1f?ra EC: 3.2. 1 .21 . sweet almonds Ki: 300 (6.8) /289|

liiifjIllfAMfi.^ EC: 3.2. 1 .21 . sweet almonds to: 14000 (6.8) [289].

Heterocycies Containing One Nitrogen Atom. Ni

5.1

HO.

Nagstatins and Other C-1/N-5 Containing Heterocycles

5

Bicyclic Systems

^ ^

5.2

HO

OH

°*CH'

C

I / —C02ch3

-XS^

chiral pool chemical Compound: /0200 Synthetic/NMR:/28d;

Dimethyl-(5R,6R,7S,8S)-5,6,7,8tetrahydro-6, 7,8-trihydroxy-5(hydroxymethyl)indolizine-l ,2dicarboxylate

.

Ki: 0.004 (7) [290], IC50:0.008 (n.g.) [2071 [253].

IC50:0.006 (5.0)/207/,/253|

,,^,30,^^

IBtBUBiHIiEBi EC: 3.2.1 .21 , sweet almonds IC50: 320 (5.0) /207/

"-"U"-*

PVUW**

IC5o:10(n.g.)i207| IBJBHBSBHBHB EC: 3.2.1.23, e. con

3EC:3.2.1.30,c^/c/cen//Vef

IgiiTOB'iMIMVnaiiUliaiBa EC: 3.2.1.30, bovine kidney IC5o:60(n.g.)i207/ ..M.^.»M,>»-..--J EC: 3.2.1.21, sweet almonds ICso: 0.5 (5.0) [207]. KBiBBBBIiHBHJB EC: 3.2.1.25, snail ICso:37(n.g.)f207J.

gEC: 3.2.1.30, chicken liver 11^ IBWRIRHPHRRFH EC: 3.2.1.23, e. con

IMiIiIiItAHf(JEEg EC: 3.2.1 .21 T sweet almonds K1: 25000 (5.8) /289/

chiral pool chemical Compound: /0210 Synthetic/NMR: [207J1 [200]

(5R,6S,7S,8S)-8-Acetamido5,6,7,8-tetrahydro-5(hydroxymethyl)-imidazo[1,2-a]

^ — chiral pool chemical Compound: /D2fO Synthetic/NMR:(207;,[200J

5-(hydroxymethyi)-imidazo[i,2-a] Dvridine-5,6,7-triol

Heterocycles Containing Two Nitrogen Atoms,

I

HO

'

S"

3

358

Tables of Glycosidase Inhibitors

5.3 Heterocycles Containing Three Nitrogen Atom, N3

359

360

Tables of Glycosidase Inhibitors

5.4 Heterocycles Containing Four Nitrogen Atom, N^

361

362

Tables of Glycosidase Inhibitors

CO

Iminosngars (is Glycosidase Inhibitors: Nojirimycin and Beyond. Edited by Arnold E. Stiitz Copyright © 1999 Wiley-VCH Verlag GmbH ISBN: 3-527-29544-5

H

OH

H

H

OH

OH

HO — "

c

»-

HO

HO —»

HO

HO — *

HO---/ τ y

HO

(1S,2R,3R,7S,7aR)-3Hydroxymethyl-1,2,7trihydroxypyrrolizidine 1-Epi-australine; 1, 7a-Di-epi-alexine natural product, castanospermum australe [092], [078] Compound: /04f 1 Synthetic/NMR: [078],[092] X-Ray:/092/ (1R,2R,3S,7S,7aR)-3Hydroxymethyl-1 ,2,7trihydroxypyrrolizidine 3-Epi-australine; 3, 7a-Di-epi-alexine chiral pool chemical, castanospermum australe [062] Compound: /0421 Synthetic/NMR: /062] X-Ray:/062j (1 R,2R,3R,7R,7aR)-3Hydroxymethyl-1,2,7trihydroxypyrrolozidine 7-Epi-australine; 7, 7a-Di-epi-alexine chiral pool chemical, a/exa leiopetala [092] Compound: /0431 Synthetic/NMR: [092] ΒΙΒΠΒΒΙΙΙΜΦΙΙ·Η^ EC: 3.2.1 .3, aspergillus niger ICso: 0.13 (6) /092/ [JEHBEfflEE EC: 3.2.1 .20, mouse gut digestive ICso: 16 (n.g.) /092/ IBiBIIBBHBffiB EC: 3.2.1 .21 T mouse gut digestive ICso: 230 (n.g.) /092/

homogenates ICso: 210 (6) /062/.

||')-D-fructofuranosiclase (invertase, s u c ra s e If=WMkJ IWHJTiMJJ.!JWt'^fflfi^^^M

Mi'iWH»lllM«H.«feHJ EC: 3.2.1 .3, aspergillus niger ICso: 2.1 (6) [092].

ICso: 26 (n.g.) /092/ IC50: 1.5 (6) /078/, IC50: 95 (n.g.) /092/

BSBIIIBiHItEHS EC: 3.2.1 .20, mouse qut digestive

BBBIIBBJBiHB EC: 3.2.12O1 yeast ICso: 270 (5) [078].

Mi'iLUMflllg&H.iliHj EC: 3.2.1.3, aspergillus niqer ICso: 1.5 (6) [092], IC50: 26 (5) [111], [226].

6 Pyrrolizidines