This is to certify that the
dissertation entitled

THE DESIGN, SYNTHESIS AND EVALUATION OF NEW
CHEMICAL ENTITIES FOR THE DEVELOPMENT OF
DRUGS AGAINST CANCER, DIABETES, AUTOIMMUNE
DISORDERS AND INFECTIOUS DISEASES

presented by

Li Gao

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in / Chemistry

 

 

(ILL MLt ‘

Major Pr t ﬁssroESignature

I IL/os/I

Date

N.“

 

MSU is an Afﬁnnative Action/Equal Opportunity Institution

   

     
     

_ LIB 'ARY
M'ChIQan State
University

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

1:?

93.9%

DATE DUE

DATE DUE

 

a ZUUﬂ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65-p. 15

—-—_—

THE DESIGN, SYNTHESIS AND EVALUATION OF NEW CHEMICAL
ENTITIES FOR THE DEVELOPMENT OF DRUGS AGAINST CANCER,
DIABETES, AUTOIMMUN E DISORDERS AND INFECTIOUS DISEASES

By

Li Gao

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

2004

ABSTRACT

THE DESIGN, SYNTHESIS AND EVALUATION OF NEW CHEMICAL
ENTITIES FOR THE DEVELOPMENT OF DRUGS AGAINST CANCER,
DIABETES, AUTOIMMUNE DISORDERS AND INFECTIOUS DISEASES

By

Li Gao

The main objective of the work described in this dissertation was to develop strategies for
the design and synthesis of nitrogen heterocyclic systems for use as glycosidase and
glycosyltransferase inhibitors and as protein synthesis inhibitors. Potential applications
include use as antibacterial, antivirai, cancer, diabetes and autoimmune drugs. The
heterocyclic ring systems include bicyclic trihydroxy-Z-thiaquinolizidine ring systems,
bicyclic iminopentitols, 2.5-bis(hydroxymethyl)-3,4-dihydroxypyrrolidines, and
arylsubstituted 5-phenyl-thiomorpholine-3-carboxylic acids. The biological testing
showed the speciﬁcity of bicyclic trihydroxy-Z-thiaquinolizidine ring systems against
glycosidases, and they have some antibacterial activity. The bicyclic iminopentitols
containing the structural essence of a ribosyl cation in the form of a l,4-dideoxy-l,4-
iminopentitol can be used as scaffolds for the preparation of riboside hydrolase,
phosphorylase and transferase inhibitors. The arylsubstituted 5-phenyl-thiomorpholine-3-

carboxylic acid derivatives showed promising antibacterial activity. The ease of synthesis

of these heterocyclic ring systems provides access to the development of completely new

classes of glycosidase and glycosyltransferase inhibitors as well as new antibiotics.

ACKNOWLEDGEMENT

I would like to express my sincere gratitude to my academic advisor Dr. Rawle I.
Hollingsworth, for introducing me a whole new world of scientiﬁc frontiers including
carbohydrate chemistry and organic synthesis. His guidance and support are important
reasons for my progress. I am also grateful to my committee members, Dr. Babak Borhan,
Dr. James Jackson and Dr. Katharine Hunt for their inspiration, valued advice and help in
my research. I would like to thank Dr. Carol Mindock for the DHF-riboside assay, MIC
testing and Dr. Don Ward for x-ray crystallography analysis as well as group members in
Prof. Hollingsworth lab.

Finally, I would like to thank my family and my friends for their support and

understanding.

iv

TABLE OF CONTENTS

LIST OF TABLES ................................................................................ viii

LIST OF FIGURES ................................................................................ ix

Chapter 1. Glycosidase and Glycosyltransferase Inhibitors: A Discussion of
Mechanism, Mimics and Modes of Inhibition

1.1. lndroduction .................................................................................... 3
1.2. Glysosidase Structure and Mechanism .................................................... 6
1.3. Glycosyltransferases Structure and Mechanism ....................................... 17

1.4. Inhibitors of Glycosidic Bond

1 .4. l. Glycosidase Inhibitors ............................................................. 21
1.4.2. Glycosyltransferase Inhibitors .................................................... 50
1.5. References ..................................................................................... 56

Chapter 2. On the Design, Properties and Applications of Purine Nucleoside
Phosphorylase, Nucleoside Hydrolase and Phosphoribosyltransferase Inhibitors

2.1. lndroduction ................................................................................. 76
2.2. Structure-based Design of PNP Inhibitors ............................................. 82
2.3. Transition State Inhibitors ................................................................. 85
2.4. References ..................................................................................... 98

Chapter 3. Trihydroxy-Z-Thiaquinolizidine Derivatives as a New Class of Bicyclic
Glycosidase Inhibitors ,

3.1. Introduction .................................................................................. 106

3.1.1. Chemo-enzymatic Synthesis ..................................................... 107
3.1.2. Aminomercuration ................................................................ 108
3.1.3. Double Reductive Amination ................................................... 108
3.1.4. Intramolecular N-Alkylation .................................................... 109
3.1.5. Reductive Double-alkylation .................................................... 110
3.1.6. Triple Reductive Amination (TRA) ............................................ 111

3.2. Results and Discussions
3.2.1. Synthesis of Trihydroxy-Z-thiaquinolizidine Derivatives ................... 1 12

3.2.2. Inhibitory Activities of Trihydroxy-Z-thiaquinolizidine Derivatives against
Glycosidases .............................................................................. 120

3.2.3. Antibacterial Activities of Trihydroxy-Z-thiaquinolizidine Derivatives...132

3.3. Experimental Section
3.3.1. General Procedures ............................................................... 135
3.3.2. Inhibition Assay ................................................................... 135
3.3.3. MIC Testing Procedure ........................................................... 136
3.4. References .................................................................................... 144

Chapter 4. Design and Synthesis of Iminopentitol Scaffolds for the Preparation of
Riboside Hydrolase, Phosphorylase and Transferase Inhibitors

4.1. Introduction .................................................................................. 150
4.2. Results and Discussions
4.2.1. Synthesis of Iminopentitol Derivatives ........................................ 158

4.2.2. Inhibitory Activity of Iminopentitol Derivatives .............................. 164

vi

4.3. Experimental Section

4.3.1. General Procedures ............................................................... 168
4.3.2. Inhibition Assay ................................................................... 169
4.4. References .................................................................................... 178

Chapter 5. A Facile Approach to Chiral 2,5-Di-Substituted Pyrrolidines

5.1. Introduction ................................................................................. 182
5.2. Results and Discussions ................................................................... 190
5.3. Experimental Section ...................................................................... 195
5.4. References ................................................................................... 200

Chapter 6. 5-Substituted Thiomorpholine Carboxylic Acids: A Compound Class
with Promising Antibacterial Activity

6.1. Introduction ................................................................................. 205
6.2. Results and Discussions
6.2.1. Synthesis of S-Substituted thiomorpholine carboxylic acid derivatives...216

6.2.2. Antibacterial Activities of the Arylsubstituted 5-Phenyl-thiomorpholine-3-
carboxylic Acids ......................................................................... 219

6.3. Experimental Section

6.3.1. General Procedures ............................................................... 224
6.3.2. MIC Testing Procedure .......................................................... 224
6.3.3. Synthesis from L-cysteine ....................................................... 225
6.3.4. Synthesis from D-cysteine ...................................................... 227
6.4. References ................................................................................... 231

vii

LIST OF TABLES

Chapter 1
Table 1.1. Inhibition constant K, (,uM) of some natural inhibitors ............................ 24
Table 1.2. K, values in ,uM for compound 22-27 ................................................ 32
Table 1.3. K, values in ,uM for compound 28-34 ................................................ 33
Table 1.4. Inhibition constant K, in [1M for isofagomines .................................... 35
Table 1.5. Thermodynamic parameters for the binding of the two inhibitors to almond B-
glucosidase at pH 6.8 ............................................................................... 38
Table 1.6. Inhibition constant K, in nM .......................................................... 39
Table 1.7. K, values in pM ........................................................................ 43
Table 1.8. K, values in pM for compound 64-68 ............................................... 45
Table 1.9. K, values in ”M for compound 69-75 ...... . ........................................ 47
Chapter 2
Table 2.1. Inhibition constants for nucleohydrolase and HGPRT ........................... 93
Table 2.2. Inhibition constants .................................................................... 94
Chapter 3
Table 3.1. Inhibition constants (mM) for compounds 36a, 36b, 51a, 51b and 54 ....... 128
Table 3.2. Activities (K, mM) of the known aza-bicyclic systems and key monocyclic
systems ................................................................................ 131
Chapter 5
Table 5.1. Product ratios under different conditions ......................................... 194

viii

LIST OF FIGURES

Chapter 1

Figure 1.1. Glycoconjugate biosynthesis and cell surface recognition ........................ 4
Figure 1.2. The N-linked core oligosaccharide. The symbols for the different sugars are
used in the following ﬁgures ........................................................................ 5
Figure 1.3. Biosynthesis of the N-linked core oligosaccharide ................................ 6
Figure 1.4. General believed mechanism of glycosidase hydrolysis .......................... 8
Figure 1.5. General mechanism for inverting glycosidases .................................... 8
Figure 1.6. General mechanism for retaining glycosidases. ................................... 9

Figure 1.7. Stereographic "end-on" section through the electron density for the

unhydrolyzed substrate complex of CelSA ...................................................... 12
Figure 1.8. Schematic ﬁgure of the covalent 2-ﬂuoro-2-deoxy-cellotriosyl-enzyme
intermediate of CelSA active site .................................................................. 14
Figure 1.9. Structure of the covalent intermediate in the HEWL reaction. ................ 15

Figure 1.10 Active-site structure of Bacillus circulans xylanase trapped as a covalent 2-
ﬂuoroxylobiosyl intermediate (a) and implications for the transition state (b) ............. 17
Figure 1.11. Structures of common glycosyl donors. ......................................... 18

Figure 1.12. Ribbon diagrams of SpsA (PDB code 1QG8) from B. subtilis representing
GT-A superfamily and MurG from E. coli complexed to UDP-GlcNAc (PDB code

lNLM) representing the GT-B superfamily ..................................................... 20
Figure 1.13 Two possible reaction schemes for retaining glycosyltransfereases .......... 21
Figure 1.14. Schematic of the active site of glucoamylase II. ............................... 28
Figure 1.15. Stereoscopic images of the difference (Fo- c) electron density contoured at
the 30' level, and corresponding to castanospermine bound in the active site ............... 30
Figure 1.16. Proposed transition state structure for the a-galactosidase reaction ......... 31

Figure 1.17. Proposed favored binding interaction of DNJ in a-glucosidase (a) and
isofagomine in a B-glucosidase (b) .............................................................. 37

Figure 1.18. Proposed binding of 49 in the active site of a B-glucosidase (a), in this case
ﬁ'om white clover and a retaining oc-glucosidase (b) .......................................... 40

Figure 1.19. Schematic representation of the interactions observed between CelSA and
the cellobio-derived isofagomine (52) ........................................................... 41

Figure 1.20. Amide-iminol isomerization ....................................................... 49

Figure 1.21. Proposed reaction mechanism of a 1,3-GalT and possible interaction
between the l-N-iminosugar-based inhibitor and the enzyme ................................ 53

Figure 1.22. Structure of sialyl Lewis x. and the azatrisaccharide. tit-1,3-
ﬁtcosyltransferase V catalyze the addition of fucose to LacNAc or sialyl LacNAc ....... 54

Figure 1.23. Proposed model for the synergistic inhibition of a-l,3-fucosyltransferase V
by the combination of an aza sugar and GDP and the proposed mechanisms of hydrolysis

fucosylation by a- 1 ,3-fucosyltransferase V ..................................................... 55
Chapter 2

Figure 2.1. PNP catalyzed reaction ............................................................... 76
Figure 2.2. Nucleoside N-riboside hydrolases catalyzed reaction ........................... 78

Figure 2.3. Hypoxanthine-guanine phosphoribosyltransferases catalyzed reaction ....... 79

Figure 2.4. Substrate inosine, PNP transition state, nucleoside hydrolase and HGPRTase

transition states ...................................................................................... 81
Figure 2.5. PNP inhibitors ........................................................................ 83
Figure 2.6. Phosphate mimics for PNP inhibitors ............................................. 84
Figure 2.7. More hydrophobic mimics. ......................................................... 85
Figure 2.8. 11an family. ........................................................................ 87

Figure 2.9. The structures of bovine PNP with substrate analogues (inosine+SO4),
transition-state complex (Imrn-H+PO4) and products (hypoxanthine+ribose l-PO4) bound
at the catalytic sites ................................................................................. 89

Figure 2.10. Superpositions of the reactants (green), bound transition state analogues
(white with blue nitrogens and red oxygens), and products (gold) in the active sites of
bovine PNP (a) are compared with the superpositions of reactants, experimentally
determined transition states, and products (b) .................................................. 91

Figure 2.11. Conformations of free and bound immucillins are from the small molecule
crystal structure of ImmH (gold) and ImmH in PNP-han°PO4 (PDB 1880, white with
blue nitrogens and red oxygens) .................................................................. 92
Figure 2.12. Transition state structures for bovine and human PNP. ....................... 96

Figure 2.13. Second generation inhibitors for human purine nucleoside phosphorylase.97

Chapter 3

Figure 3.1. Chemo-enzymatic synthesis of 1 and 8 .......................................... 107
Figure 3.2. Aminomercuration .................................................................. 108
Figure 3.3. Double reductive amination ........................................................ 109
Figure 3.4. N-Intramolecular alkylation ....................................................... l 10
Figure 3.5. Reductive double-alkylation ....................................................... l l 1
Figure 3.6. Triple reductive amination ......................................................... 1 12
Figure 3.7. Synthesis of products with D-gluco and L-ido conﬁgurations. . . . . . . . . . . ......1 15
Figure 3.8. Synthesis of products with the D-manno and L-gulo conﬁgurations. . . . . ....1 17

Figure 3.9. X-ray structure of 7(S),8(R),9(R),l0(S)-t1ihydroxy-2-thiaquinolizidine 36a
and 7(R),8(R),9(R),10( S)-trihydroxy-2-thiaquinolizidine 51a showing the trans-type ring
junction and overall ﬂat geometry .............................................................. 118

Figure 3.10. Structure of oxycarbenium intermediate in the hydrolysis of glucosides from
an ab initio calculation showing equipotential contours .................................... l 19

Figure 3.11. Synthesis of 2,2-dioxy-7(S),8(R),9(R),l0(S)-tn'hydroxy-Z-thiaquinolizidin-

2-one ............................................................................................... 120
Figure 3.12. Enzyme inhibition scheme. .................................................... 121
Figure 3.13. Glucose oxidase/peroxidase enzyme mechanism ........................... 122

xi

Figure 3.14. Lineweaver-Burk plot for the determination of inhibition constant for
36a ................................................................................................... 123
Figure 3.15. Slope from figure 3.14 vs inhibitor concentration ............................ 124
Figure 3.16. Lineweaver-Burk plot for the determination of inhibition constant for
36a ................................................................................................... 125
Figure 3.17. slope from figure 3.15 vs inhibitor concentration ............................. 125
Figure 3.18. Lineweaver-Burk plot for the determination of inhibition constant for
36h ................................................................................................... 126

Figure 3.19. Lineweaver-Burk plot for the determination of inhibition constant for

54 ........................................................................ L ........................... 127
Figure 3.20. Stereoelectronic requirements for cleavage of a- and B-glycosidases ...... 129
Figure 3.21. Inhibitory activity against S. aureus 43300 .................................... 134
Figure 3.22. Inhibitory activity against E. faeralis 51299 ................................... 134
Figure 3.23. Inhibitory activity against E. faecalis 292129 ................................. 135
Chapter 4

Figure 4.1. Synthesis of 1,4-imino-D-ribitol 3 ............................................... 152
Figure 4.2. Synthesis of 1,4-dideoxy-1,4-imino-L-lyxitol ................................. 153
Figure 4.3. Synthesis of 1,5-dideoxy-l ,5-imino-D-ribitol ................................. 154
Figure 4.4. Synthesis of l,4-dideoxy~1,4-imino-D-ribitol ................................. 155
Figure 4.5. Synthesis of ImrnH ................................................................ 156
Figure 4.6. Architecture of glycosidase and transferase inhibitors base on
dihydroxyhexahydropyrrolothiazine scaffold ................................................. 158
Figure 4.7. Synthesis of compound 40 ........................................................ 160

xii

Figure 4.8. Synthesis of compound 41b ....................................................... 161
Figure 4.9. Synthesis of compound 42b ........................................................ 162

Figure 4.10 (a) Ortep drawing of X-ray structure of 41b. (b) Ortep drawing of X-ray
structure of 42b. (c) Ortep drawing of X-ray structure of 53 ............................... 164

Figure 4.11. (a) X-ray derived structure of 1,4-dideoxy-1,4-imino-1-(S)-(9-
deazahypoxanthin-9-Yl)-D- ribitol (b) Comparison of X-ray structure 42b and 1,4-

dideoxy- l ,4-imino-1 -(S)-(9-deazahypoxanthin-9-Y1)-D- ribitol ............................ 165
Figure 4.12. (a) Structure of 3',4'-dihydroxyﬂavone-4'-B-D-ribofuranoside sodium salt 54
and (b) the putative chelate 55 formed with iron after enzymatic hydrolysis. . . . . . ....166
Figure 4.13. Aabsorbance of the chelate 55 produced in the lysate assay with inhibitor 41
concentration from 0 ~ 15 ug/uL ................................................................ 168
Figure 4.14. Absorbance of the chelate 55 produced in the whole cell assay with inhibitor
41 concentration from 0 ~ 15 ug/uL ............................................................ 168
Chapter 5

Figure 5.1. Synthesis of DMDP ............................................................... 185

Figure 5.2. Synthesis of 2(R),5(S)-bis(hydroxymethyl)-3(R),4(R)-

dihydroxypyrrolidine .......... . ................................................................. 1 86
Figure 5.3. Synthesis from L-gulono-1,4-lactone. ......................................... 187
Figure 5.4. Synthesis of 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidines .............. 189
Figure 5.5. Synthesis of Iran ................................................................ 190
Figure 5.6. Synthesis of compound 17 ........................................................ 192
Figure 5.7. Synthesis of products 55 and 56 ................................................. 195
Chapter

Figure 6.1. Typical features for oxazolidinone antibacterial activity ..................... 210
Figure 6.2. Large-scale preparation of Linezolid ............................................ 212

xiii

Figure 6.3. Synthesis of DUP-721 ............................................................. 213

Figure 6.4. Synthesis of 5-trityloxymethyl-2-oxazolidinone .............................. 214
Figure 6.5. Synthesis of 5-hydroxymethyloxazolidinones .................................. 214
Figure 6.6. Synthesis of MCB 3570 ............................................................ 216
Figure 6.7. General structures of synthetic targets and known antibiotics ............... 218
Figure 6.8. Synthesis of arylsubstituted 5~phenyl-thiomorpholine-3-carboxylic acids
from L-cysteine .................................................................................... 219
Figure 6.9. Synthesis from D-cysteine ........................................................ 220
Figure 6.10. Inhibitory activity against S.aureus 43300 .................................... 221
Figure 6.11. Inhibitory activity against S.aureus 29213 .................................... 222

Figure 6.12. Inhibitory activity against E. faecalis 51299 for compound 35-41. . .. . ...222
Figure 6.13. Inhibitory activity against E. faecalis 29212 for compound 35-41. . . . . ...223
Figure 6.14. Inhibitory activity against E. faecalis 29212 for compound 43-49 ........ 224

Figure 6.15. Inhibitory activity against E. faecalis 51299 for compound 43-49. . . . . ...225

xiv

Chapter 1

Glycosidase and Glycosyltransferase Inhibitors:

A Discussion of Mechanism, Mimics and Modes of Inhibition

ABSTRACT

Carbohydrate processing is a central feature of the chemistry of life. Glycosidases and
glycosyltransferases are the critical enzymes that control the breakdown, processing and
synthesis of carbohydrates in biological systems. According to the stereochemistry
outcomes of the enzyme catalyzed reactions, two possible mechanisms have been
proposed as inversion and retention of anomeric conﬁguration. Mechanism studies
including x-ray structures of the enzyme-substrate complex, glycosyl-enzyme
intermediate, and enzyme-product complex, have provided us insight into the different
stages of the enzymatic reaction pathway. This insight is being used in the design of new
chemical entities that can inhibit or regulate the biochemical activity of speciﬁc enzymes
in a strategy for producing new drugs. Such drugs will ﬁnd applications across the entire
spectrum of therapeutics. In this chapter we review the progress made to date on the
identiﬁcation, design, and evaluation of natural and synthetic glycosidase and
glycosyltransferase inhibitors that mimic the possible transition states and intermediates

purported to be in the reaction pathway for these enzymes.

1.1. Introduction

Glycosidases (or glycosyl hydrolases) and glycosyltransferases are important enzymes
that catalyze the cleavage and formation of glycosidic linkages in biological systems.
They play crucial roles in the biosynthesis and degradation of a large number of
biological structures. These include polysacchrides, oligosaccharides, antibiotics,
glycolipids, glycoproteins, proteoglycans and peptidoglycans. They are also involved in a
number of other processes including the remodeling of cell wall, the damage and repair
of DNA by base excision or transfer, and other toxin mechanisms. Understanding their
roles in the formation of oligosaccharides and glycoconjugates is essential to modulate
the cellular interaction and to develop therapeutic agents.

Oligosaccharides are synthesized in step-wise fashion primarily in the endoplasmic
reticulum and Golgi apparatus (Figure 1.1)', a process that affords signiﬁcant product
diversityz. Exogenously supplied monosaccharides are taken up by cells and converted to
monosaccharide "building blocks" (typically nucleoside sugars) inside the cell. Several
steps of metabolic transformation might take place en route from an exogenous sugar to a
building block. The building blocks are imported into the secretory compartments where
they are assembled by glycosyltransferases into oligosaccharides bound to a protein (or
lipid) scaffold. Further trimming and elongation by glycosidases and transferases will
give the fully matured form of glycoconjugates. Upon reaching the cell surface, they can

serve as ligands for receptors on other cells or pathogens.

   

Extracellular mllleu

   

   

\1
Glycoconjugate assembly
Cyfml In the secretory compottrmnto

   

Figure 1.1. Glycoconjugate biosynthesis and cell surface recognition. (Images in this
thesis are presented in color)
The N-linked glycoproteins are belived to be involved in cell-cell adhesion,
diﬁ‘erentiation, recognition, regulation, modulation of protein receptors, and so on. They
present a great diversity in their structures, but all the different structures come from a
common precursor, the N-linked core oligosaccharide (Figure 1.2)3. During the synthesis
of N-linked glycans in mammalian cells, this core unit is assembled as a membrane-
bound dolichylpyrophosphate precursor in the endoplasmic reticulum (ER)“, and then
transferred to a growing, nascent polypeptide chain (Figure 1.3). After it is coupled to an

asparagine residue through an N-glycosidic bond catalyzed by an

I' 1 I.
O J

“ m- complex7, three terminal glucoses and one mannose are

trimmed away by glucosidase I, II and mannosidases“. When the glycoprotein moves to

the Golgi, further mannose trimming occurs, and new sugars are added during temiinal
glycosidation by transferases to produce complex N-linked glycans. In one pathway
shown in ﬁgure 1.3, GlcNAc, galactose, sialic acid, and fucose are added to the
glycoprotein and only ﬁve original sugars leﬁ. The ﬁnal oligosaccharide structures
depend on the polypeptides and on which processing glycosidases and transferases are

expressed in the cells.

Glucosldaae l —>

   

ER a—mannoaldm I

Glucoeldaee [I

A...“
O = munnooo

s N-ecotylglucoumlno
I

Aﬂucou
gm...
.=lhl|c.cid

Figure 1.2. The N-linked core oligosaccharide. The symbols for the different sugars are

used in the following ﬁgures.

     

Figure 1.3. Biosynthesis of the N-linked core oligosaccharide.

1.2. Glysosidase Structure and Mechanism

Glycosidases catalyze the hydrolysis of the glycosidic bond, which exists in a wide range
of biological systems. Although all the glycosidases catalyze the same reaction, the
hydrolysis of an acetal, the enzymes present an incredible structural diversity. Over 2000
glycosidases have been identiﬁed, and they are grouped to over 90 families according to
the amino acid sequence similarity. The number is growing steadily and updated
regularly and is available on the website http://afmbcnrs-
mrs.fr/~M/CAZY/index.html9. Since the ﬁrst glycosidase three-dimensional structure
of hen egg-white lysozyme (HEWL) was solved, there have been over 50 structure

representatives for these families available, revealing a large number of different folds.

Some glycosidases are almost entirely composed of [i-sheet, others being primarily a-
helical, and then a range of structures of mixed a-helix and B-sheet in between. This
structure diversity may in part be a reﬂection of the large variety of the different
substrates being cleaved.

The glycosidic bond is one of the most stable linkage within naturally occurring
biopolymers'o, and the great acceleration of the hydrolysis by glycosidases has made
them some of the most efﬁcient catalysts. Information of the mechanisms has been
obtained from kinetic studies, trapping of the intermediates, and ﬁom the three-
dimensional structures of such intermediates. The generally believed mechanism”12 of
glycosidic bond cleavage involves acid catalysis with protonation of the exocyclic
oxygen atom at the anomeric center to give a protonated glycoside (ﬁgure 1.4).
Subsequent cleavage of the C(l) — 0(1) bond leads to an oxocarbenium ion-like
transition state, in which the positive charge is delocalized over the ring oxygen and the
anomeric carbon. The oxocarbenium intermediate species then reacts with a molecule of
water to form the product. The oxocarbenium ion-like transition state has been conﬁrmed

'3’”, with KH/KD value greater than 1

by the a secondary deuterium kinetic isotope effects
(KH and K9 are rate constants for reactions of protium- and deuterium-labeled substrates).
Two possible mechanisms have been proposed according to the stereochemistry
outcomes of the enzyme catalyzed hydrolysis. Glycosidases that catalyze hydrolysis of
glycosides with inversion of anomeric conﬁguration are called inverting enzymes and

glycosidases that catalyze the hydrolysis with retention of conﬁguration are retaining

glycosidases.

OH

“go/$11? HO OS§HI:: e—b I'IO$_> H20 Hogﬂ‘.
OH OHR OH OH

1

3

4

Figure 1.4. General believed mechanism of glycosidase hydrolysis

The inverting glycosidases catalyze the reaction via a singe-displacement mechanism

involving an oxocarbenium ion-like transition state (figure 1.5). In the active site, the

substrate and a water molecule are bound between two carboxyl groups. One carboxyl

group acts as general acid to protonate the leaving group, while the other one

deprotonates the water molecule to facilitate the nucleophilic attack.

Figure 1.5. General mechanism for inverting glycosidases

II

 

 

II

I5

/

 

 

HO o
r'!
5+ \
\ r'o_H
war. : \‘\>."‘/ 6‘ /
A \
"_ I‘

 

 

Figure 1.6. General mechanism for retaining glycosidases.

For retaining glycosidases, Koshland16 ﬁrst pointed out a double-displacement

”“749 (ﬁgure 1.6) which involves an enzyme nucleophile. In the ﬁrst step,

mechanism
one of the carboxyl group protonates the glycosidic oxygen, and the other one acts as a

nucleophile to displace the aglycon through an oxocarbenium ion-like transition state to

form a covalent glycosyl-enzyme intermediate. In the second step the carboxylate group
deprotonates the incoming water molecule to generate a hydroxyl ion which attacks the
anomeric center of the intermediate. The net result is the retention of the anomeric
conﬁguration. In the mechanism, there are two transition states with substantial
oxocarbenium ion character and one glycosyl-enzyme intermediate. Kinetic studies,
trapping technology and the corresponding three-dimensional structures have provided
invaluable information about the nature of the transition state. The x-ray structures of the
enzyme-substrate complex (Michaelis complex), glycosyl-enzyme intermediate, and
enzyme-product complex, which are stable forms along the reaction coordinate, have
been solved for some enzymes. These give us insight into the different stages of the

enzymatic reaction pathway. Different structures such as natural substrateszo'zz,

nonhydrolyzable thiooligosaccharide substrate analogsn'zg, and 2-ﬂuoro-2-deoxy
oligosaccharide827'30’33 have been used to obtain the enzyme-unhydrolyzed substrate
complexes and the corresponding crystal structures. The introduction of the 2-ﬂuoro

34‘”. The rationale behind

group is a common strategy for trapping the enzyme complexes
is that the electron-withdrawing 2-ﬂuoro group destabilizes the positive charge that
develops at the transition state. It has also been proposed that replacement by ﬂurine also
removes an important interaction between the 2-hydroxyl group of the substrate and the
enzyme which is involved in the transition state stabilization36'38. The 2-OH is said to
contribute about 20 kJ/mol and 3, 4, 6 positions provide ~10 kJ/mol each for the binding
afﬁnity”. This is unlikely to be the real reason because the effect of ﬂuorine substitution

is very general for all classes of sugars regardless of ring size and the conﬁguration at the

2-position. The 6—2 effect should also be remembered. It is also possible that the 2-

10

oxygen stabilizes the positive charge at the anomeric position. In the absence of any
nucleophile, 1,2-anhydrides (epoxides) are formed by displacing good anomeric leaving
groups such as halides. From the structures of 3-D substrate-enzyme complexes, substrate
distortion is a common feature. Figure 1.730 is the unhydrolyzed substrate complex of a
B-glycosidase CelSA with 2,4-dinitropheny1 2-deoxy-2-ﬂuoro-B-D-cellobioside. The -l
subsite sugar conformation is a lS3 skew-boat conformation, away ﬁ'om the preferred 4C,
chair conformation. This conformation puts the leaving group in the pseudo-axial
orientation, and presents the anomeric carbon to the nucleophile in an appropriate
position for the in-line attack“). Other distortion of substrate in its ground state 4C,
conformation to 4E envelope or 'Ss skew boat conformers has been conﬁrmed by crystal
structure analysis of Michaelis complexes. These conformational changes direct the
scissile bond to an axial orientation, as required by the principle of stereoelectronic
control. However, this distortion is not universal. A number of glycoside complexes have

2"25'27‘28. Even with the same ligand

been reported without distortion
(thiopentasaccharide), different glycosidases (Cel7B4' and CelSA26) present different
binding modes (chair and skew-boat), but it is also possible these binding modes
represent different steps of the reaction pathway: from an initial binding with the -l
subsite sugar in the chair conformation to the distorted complex to facilitate the
nucleophilic attack. Nevertheless, there is not a ‘requirement’ for substrate distortion to
achieve the transition state with C5, 05, C1 and C2 coplanar and pseudo-axial orientation
for the leaving group. It should be pointed out that the 2,4-dinitrophenyl substituent on

the anomeric oxygen is a very powerful electron withdrawing group and the '83 skew-

boat conformation places the anomeric oxygen in the same relative orientation to the lone

ll

pair in the ring oxygen as is found in a-glycosides. This orientation is stabilized by the

“anomeric effect”.

OH

 

Figure 1.7. Stereographic "end-on" section through the electron density for the
unhydrolyzed substrate complex of CelSA. The enzymatic nucleophile, Glu 228, is
shown, as are the -1 subsite pyranoside ring atoms and the dinitrophenyl leaving group.
Substrate distortion to a skew-boat conformation permits nucleophilic attack "in-line"
with leaving group departure.

The existence of the covalent glycosyl-enzyme intermediate was ﬁrst illustrated by a
secondary deuterium kinetic isotope effects with substrate for which the second step is
rate determining. The KH/KD values are all greater than 113’42 (ranging from KH/KD = 1.25
for E. coli [3-galactosidase42 to KH/KD = 1.11 for both an Agrobacterium sp. [3-
glucosidasel4 and for Cex‘“), indicating the anomeric center undergoes rehybridization
ﬁ'om sp3 in the tetrahedral covalent intermediate to sp2 in the oxocabenium ion-like
transition state. Now, the covalent intermediate has been conﬁrmed by trapping the
intermediates and the corresponding x-ray crystal structures. The most common

strategy for trapping is to introduce an activated leaving group to the substrate and

incorporate ﬂuorine in the 2- or 5-position35’M. Good leaving groups can lead to the

rapid formation of the glycosyl-enzyme intermediate. The electron-withdrawing 2-ﬂuoro
group can slow down reactions for both steps of glycosylation and deglycosylation. The
net result is the breakdown of the intermediate is slowed to a greater extent than its
formation. Therefore, the intermediate accumulates, making the trapping possible. A
number of three-dimensional structures of such intermediate have been solved30‘38’47'53,
and they have unambiguously revealed the covalent linkage between the anomeric carbon
and the oxygen of the nucleophile. Figure 1.8 is the schematic ﬁgure of the covalent 2-
ﬂuoro-2-deoxy-cellotriosyl-enzyme intermediate of CelSA'". The sugar rings are in the
relaxed chair conformation without distortion, which is observed in most of the glycosyl-
enzyme intermediate. The nucleophile is Glu 228 and the acid/base catalyst is G1u139,
which is hydrogen bonding to a water molecule. This water molecule is 3.1 A away ﬁ'om
the anomeric carbon, and is well positioned for the nucleophilic attack. All the hydroxyl
groups are involved in the hydrogen bonding to the enzyme residues. The 6-OH is
relatively ﬂexible, and it has two possible conformations, with OH pointing out or
backwards. Comparing the two CelSA complexes with unhydrolyzed 2,4-dinitrophenyl 2-
deoxy-2-ﬂuoro-B-D-cellobioside and the covalent intermediate, we can see there is very
little change of the atomic movement except the ‘migration’ of the anomeric carbon
(from skew-boat to chair). A feature of these complexes is the static nature of the protein.
The protein enzyme residues remain almost the same position and conformation with the

exception of the nucleophile which rotate to avoid a steric clash with 2-F substituent in

the trapped intermediate.

13

 

Ser69

Figure 1.8. Schematic ﬁgure of the covalent 2-ﬂuoro—2-deoxy-cellotriosyl—enzyme
intermediate of CelSA active site. Experimentally detemrined H atoms are indicated in
blue, those with weak density in green and those for which no electron density was
observed in red.

The covalent glycosyl-enzyme intermediate was also obtained for hen egg-white
lysozyme (HEWL)5°, the ﬁrst glycosidase with 3D structure solved. The intermediate
between the HEWL (E35Q) mutant enzyme and substrate 2-acetamido-2-deoxy-B-D—
glucopyranosyl-(l -+4)—2-deoxy-2-fluoro-B—D-glucopyranosyl ﬂuoride (NAGZFGlcF) was
characterized by x—ray crystallography. The mutant enzyme in which the catalytic
acid/base carboxylate residue had been replaced by the corresponding amide is used to
reduce the rate of the reaction since no acid/base catalysts are available. Figure 1.9 is the

structure of the HEWL (E35Q) covalent glycosyl-enzyme formed by reaction with

NAGZFGlcF. It is clear that the covalent linkage (~1.4 A) is between or and o52 of

nucleophile Asp 52, and the -1 pyranose ring adopts an undistorted 4C, chair
conformation, as seen earlier and in the structures of most hexopyranose covalent
intermediates observed. This covalent glycosyl-enzyme intermediate puts an end to a
long-lived controversy and discards the ion-pair intermediate hypothesis found in all

textbooks.

 

Asp 52 Asp 52

Figure 1.9. Structure of the covalent intermediate in the HEWL reaction. Maximum
likelihood / aA—weighted 2Fo — Fc electron density for the covalent glycosyl-enzyme
intermediate of HEWL, contoured at 0.4 electrons per A.

In most trapped intermediate structures, the undistorted chair conformation is
observed27’38’4"48’5°’52’53, but in the x-ray structure of 2-deoxy-2-ﬂuoroxylobiosyl-ch
(BCX-ZFXb) intermediate“, a distortion of the sugar ring was observed51’54’”. One
xylose moiety remains the normal chair conﬁrmation, while the other one, through which
the sugar is linked to the enzyme via an ester bond, is distorted into a 2’5B boat
conformation (Figure 1.1011). This conformation of the ring allows its C5, 05, C1, and

C2 atoms to achieve a nearly planar geometry. It is reasonable to assume that the enzyme

may distort the substrate to a transition state conformation with C5, 05, C1, C2 coplanar,

15

and shorten the distance between the nuclophile and anomeric carbon to the length of a
glycosidic bond. Since the formed intermediate has this feature of coplanarity, the
formation and hydrolysis of this species should be considerably facilitated (Figure
1.10b). It should also be noted, however, that if the original 4C, conformation were
retained, there would be two unfavorable trans-annular (1,3) interactions between the
acyl group and the axial H-3 and H-5 protons. One of these is relieved in the 2'SB
conformation. Note also that this would not occur with a glucoside because the bulky
hydroxymethyl group would be forced to take the unfavorable “ﬂagpole” position. Such
distortion is not cormnon in the covalent glycosyl-enzyme intermediate, but it was also
observed in the covalent 2-ﬂuoromannotriosyl-mannanase intermediate, which adopts a
conformation close to the °S2 skew boat”. The same steric argument holds here again
except that the hydroxyrnethyl group on carbon 5 will clash with the ﬂuoro group. To
lessen this interaction, twist boat conformation is preferred. This is normally less stable
than the chair conformation and implies that the attack of the carboxylate group cannot
proceed if the ring is locked in a 4C, conformation. Together with structures of Michaelis
complexes, those glycosyl-enzyme intermediates provided important information about

transition state structures and reaction mechanism.

16

 

8+
.. ° ‘—>° O
HO H

 

 

Figure 1.10 Active-site structure of Bacillus circulans xylanase trapped as a covalent 2-

ﬂuoroxylobiosyl intermediate (a) and implications for the transition state (b).

For enzyme complexes with product, no signiﬁcant density could be observedzmo. This
presumably results from the absence of speciﬁc stabilizing interactions at that end of the
molecule. However, there is no evidence for any distortion of the reducing end ring away

from a chair conformation.

1.3. Glycosyltransferases Structure and Mechanism
Glycosyltransferases catalyze the transfer of sugars from activated sugar donors to other
molecules. The substrates for glycosyltransferases are relatively structurally complex.

The molecules to which these enzymes transfer sugars, the glycosyl acceptors, include all

17

categories of biopolymers-oligosaccharides, proteins, nucleic acids, and lipids-as well as
numerous natural products such as vancomycin, erythromycin, and daunomycin“.
Although the acceptor substrates of glycosyltransferases vary widely and include many
classes of oxygen and nitrogen nucleophiles, the donor substrates usually share common
features with sugars attached to a nucleoside diphosphate or monophosphate which is the

l eaving group. Figure 1.11 is some examples of common glycosyl donors.

OH OH
H
o
"L? HN Ho MN
0 0 A I o o A I
NHA II II o u 0" II II o n
O-P—O—P—O o o—e—o—e—
0' 0' 0' 0'
H on on
UDP-GlcNAc 5 UDP-Galactose 6
on
HO O
o /'~l "" II A
II II < A on P-O o N N m,
O—P-O—P— N M NH, 9“ $-
or o o "0 i 0
Act”! co, on OH
OH on "0
GDP-mannose 7 CMP-slalic acid 8

Figure 1.11. Structures of common glycosyl donors.

There is a myriad of glycosyltransferases all using similar activated sugars. Therefore, the
S'I><:ciﬁcity of these enzymes lies in the structures of the acceptors. Because of their
ELllrction and product structure diversity, glycosyltransferases display a high level of
diversity in terms of their primary sequences. There are currently >7000 known
glycosyltransferase sequences in the databank and they have been classiﬁed into dozens

0f different families. Although there are a large number of different sequences, only two

18

fundamental folds57’58 have been uncovered according to the available crystal structures
of glycosyltransferases, and sequence analysis has suggested that this situation will hold
true for a large number of the structurally uncharacterized families”. Until now, 21
glycosyltransferase structures from 14 different families have been reported, with 9
belonging to one superfamily, GT-A superfamily, and 8 to the other GT-B superfamily.
The GT-A superfamily includes most of the Glycosyltransferases found in Golgi
apparatus and endoplasmic reticulum. This superfamily, best represented by family GT-2
S psA ﬁ'om Bacillus subtilis (Figure 1.12)60’6', is characterized by a single Rossman fold
domain, strong dependence of metal ions (most commonly Mn2+) and sidechain residues
ﬁ-orn the protein. The metal ion is essential to catalysis, and it helps anchor the
pyrophosphoryl group of the sugar donor in the enzyme active site. The sidechain
residues are often called ‘DXD motif‘. The GT-B superfamily does not have such a motif
and no metal ion has been identiﬁed clearly in the 3D structures. A typical example for
GT-B superfamily is family 28 MurG from E. Coli. The crystal structure of MurG shows
two similar Rossman domains, and each of the Rossman fold all! open sheet structure is

Separated by a deep cleft in which the substrates bind62’63 (Figure 1.12).

SpsA MurG

 

19

Figure 1.12. Ribbon diagrams of SpsA (PDB code lQG8) from B. subtilis representing
GT-A superfamily and MurG from E. coli complexed to UDP-GlcNAc (PDB code

1 NLM) representing the GT-B superfamily. Residues important for substrate binding,
speciﬁcity, ion binding, and catalysis are shown in black. In SpsA the manganese ion is

shown as a dark gray sphere and in MurG substrate is shown in light gray.

Glycosyltransferase-mediated reactions are thought to proceed through an oxocarbenium-
ion-like transition state, similar to that proposed for glycosidase reactions““. The a-
secondary deuterium isotope effects on sucrose synthetasem, rabbit glycogen
synthetase“, B4Gal-Tl69, and a-l,3-fucosyltransferase V65 supported a mechanism in
which the anomeric carbon of the donor sugar possesses considerable sp2 character. An
ab initio computational study70 of the binding site has been performed supporting the
formation of an oxocarbenium species and characterizing the transition state structure.
Like the glycosidases, there are two basic mechanisms, an inverting and a retaining
pathway. The inverting glycosyltransferases catalyze reactions through a single
displacement with base activation of the acceptor. The retaining glycosyltransferases are
Shggested a double-displacement mechanism via a covalent glycosylenzyme
intermediate, but no evidence has been found for the covalent intermediate on retaining
u‘arisferases. The existence of this intermediate in transferases has been challenged by
Withers et. a1", and led to the revisit of the earlier suggestion of an ‘internal return’ (SNi)
11'lcechanismsn‘n (Figure 1.13), in which the departure of the leaving group and attack of

the nucleophile occur on the same (a) face of the glycoside, leading to a late

ONiocarbenium ion-like transition state.

20

Enzyme

A
Enzyme
O—UDP 0(N
O
PDU/ H/ \R
0 5” o
mlk -—> ‘23 —>
5-0"; “X5"
0 0
PD 00/ l_,/° \n PDU/ \n \
L ..

 

 

Figure 1.13 Two possible reaction schemes for retaining glycosyltransfereases. (A) A
base-catalyzed double-displacement mechanism via a covalent intermediate with the
enzyme. (B) An SNi-like mechanism involving a late oxocarbenium ion-like transition
state.
1 -4. Inhibitors of Glycosidic Bond
1 -4.]. Glycosidase Inhibitors
Glycosidase inhibitors are invaluable tools for mechanistic studies. We have seen 2-
ﬂlioro compounds as probes for the elucidation of intermediate and transition state
S‘l:]'uctures. Glycosidase inhibitors are also important targets for drug discovery. Several
eExtremely debilitating and eventually fatal diseases such as diabetes, arthritis and several
Ql-gan disorders are caused or mediated by carbohydrate processing. Inhibition of
itltestinal a-glucosidases can be used against diabetes by lowering the blood glucose

1evels. Inhibitors of glycoconjugate processing glycosidases can disrupt the biosynthesis

of glycoproteins and glycolipids and hence cell-cell or cell-virus recognition. Therefore,

21

 

 

means of inhibiting or regulating glycosidases and glycosyl transferases, especially by
small organic molecules, are much sought after in drug development.
A number of natural products which show glycosidase inhibition have been isolated from
plants and microorganisms. Most of them resemble the natural carbohydrate, but an
irnino group replaces the endocyclic oxygen atom. They are believed to be able to mimic
the ground state conformation and the positive charge character of the transition state
upon protonation of the basic nitrogen atom. These natural occurring sugar mimics are
classiﬁed into ﬁve structural classes": polyhydroxylated piperidines, pyrrolidines,
indolizidines, pyrrolizidines, and nortropanes. These natural occurring glycosidase
inhibitors have provided important information for potent inhibitor design and have
presented new challenges for the organic synthesis. A large number of newly designed
g1 ycosidase inhibitors have been based on the mimicry of the charge and shape of the
proposed transition state. We will discuss some of the natural glycosidase inhibitors and
the designed transition state analogues. Analogs that mimic the positive charge build-up
in transition states as well as analogues which do not mimic the charge of the proposed
hf‘ansition states will be discussed. The position for the positive charge build-up includes

elidocyclic oxygen (9), anomeric center (10), and exocyclic oxygen (11).

(Hon/K) (Hon/K). (Horn/ﬂ

H98
9 10 11

1.4.1.1. Inhibitors that mimic a positively charged endocyclic oxygen

22

 

The glycosyl oxocarbenium ion 9 is believed to be an important character of the
transition state of glycosidases, so mimics of the cation 9 are thus potentially transition

state analogs. The natural occurring nojirimycin family is an important class of this kind

 

of inhibitors.
OH OH OH OH
O NH NH "‘9‘.
HO HO HO HO
"7&3... @0. $$ ..
OH OH OH OH
Glucose l Nojirlmycln 12 Deoxynojlrlmycln 13 Me-Deoxynojlrlmyeln 14
"a
OH’ HO OH 2,
&
Nx ’NH
H0 30 N HO
HO
HO HO HO
Bu-Deoxynojlrlmycln 15 Castanospermine l6 Deoxymannonojlrimycln l7 Swalnsonlne l8

Nojirimycin 12 was discovered as the ﬁrst manual glucose mimics, with a nitrogen atom
in the place of the ring oxygen”, and it is a potent inhibitor of a- and B—glucosidases from
a variety of sources“. The more stable form l-deoxynojirimycin (DNJ)77 13 and l-
deoxymannojirimycin (DMJ)7M’0 17 were also isolated and showed potent inhibitory
activities against glycosidases. DNJ is a potent inhibitor of all kinds of glycosidases
(Table 1.1)”, but it is more selective to a-glucosidases and its inhibitory effect on
t3l'lammalian u-glucosidases opened the possibility of a therapeutic application for DNJ.

I‘iowever, despite the excellent a-glucosidase inhibitory activity of DNJ in vitro, its

efﬁcacy in vivo was only moderate“. Therefore, a large number of DNJ derivatives were

prepared in the hope of increasing the in vivo activity. Much of the work on chemical

23

modiﬁcation has been focused on N-substitution82’83. N-methyl-DNJ and N-butyl-DNJ
show better inhibition than DNJ, but longer alkyl chains do not show further increase in
activity84’85. Some natural bicyclic polyhydroxyheterocycles were also discovered and

“‘89 18 are typical

1' solated. The indolizidine castanosper'rrrinegf“87 16 and swainsonine
examples. These compounds have less obvious structure relationship to monosaccharide
but in each case the conﬁguration of hydroxyl groups on the ring can be compared to
those of sugars. Castanospennine can be considered a bicyclic derivative of DNJ, with an
ethylene brige between the hydroxyl group and the ring nitrogen. Castanosperrnine is an
excellent a-glucosidase inhibitor. Swainsonine is believed to be associated with DMJ,
and showed inhibition against a-mannosidases (Table 1.1). The inhibition spectra of
castanospermine, swainsonine and N-alkylation derivatives of DNJ showed they are more

potent or more speciﬁc glycosidase inhibitors than the corresponding deoxynojirimycins

and they have been targeted as anti-HIV, anticancer, antimicrobial etc. agents.

Table 1.1 Inhibition constant K, (uM) of some natural inhibitors

\

 

Enzyme and source 12 13 16 18
Q—glucosidases

Yeast 6.3 12.6 >1500

Rice 0.01 0.0l 0.015

Sucrase (Rabbit intestine) 0.13 0.032

Sucrase (Rat intestine) 0.00055

B-glucosidases
Aspergillus wentii 0.36 2.7 0.9

I‘

24

 

7 Sweet almonds 0.89 47 1.5
Calf liver (cytosol) - 210
Calf spleen (lysosomes) 4.5 180
Human placenta (lysosomes) 7
a-mannosidases
Jack beans 0.001
Almonds >1000
Rat liver (lysosomes) 0.07

mannosidase 1 AIR

Rat liver (Golgi ves.) >1000
B-mannosidases
Aspergillus wentii >2000

 

C astanospermine has been found to inhibit lysosomal a- and B-glucosidases and disturb
the lysosomal catabolism of glycogen90 and glycolipids”. These syndromes resemble the
genetic disorders Pompe's and Gaucher's disease, respectively. Both castanospermine and
DNJ inhibit processing a-glucosidases I and 119293, but castanospermine is more effective
against a-glucosidases I than DNJ“. Castanospermine and the N-alkylation derivatives
Me-DNJ have also been shown to exhibit antimetastatic activity by inhibiting platelet
aggregation of metastatic cells as well as reducing adhesion of tumor cells to the vascular

endothelium95’96. Furthermore, these compounds have been shown to inhibit cellular

tI‘ansformation by altering oncogene glycosylation”. For the inhibition of HIV,

castanospermine, DNJ, and synthetic N-butyl-DNJ (Bu-DNJ) are potent inhibitors of HIV

replication and HIV-mediated syncytium formation in vitro°8"0°. Bu-DNJ has been
developed as an anti-HIV agent. It has been reported that a major mechanism of action of
Bu-DNJ as an inhibitor of HIV replication is the impairment of viral entry at the level of
post-CD4 binding, due to an effect on viral envelope components'm. Bu-DNJ is also a
speciﬁc inhibitor of the glucosyltransferase-catalyzed biosynthesis of GlcCer, the ﬁrst
step in the biosynthetic pathway of GlcCer-based GSLsgg. This discovery constitutes
great progress toward the treatment of this group of severe diseases. DMJ and
swainsonine, which act on the later stages of N-linked oligosaccharide processing, have
no effect on the secretion of infectious virus'oo'm'. Swainsonine has no effect on Golgi a-
mannosidase 1'02, but it is a potent inhibitor (IC50=0.2 ,uM) of rat liver Golgi a-
mannosidase II, and swainsonine is the ﬁrst compound that was found to inhibit
glycoprotein processing. Swainsonine is also a potent inhibitor of lysosomal a-

mannosidasem

. The inhibition spectra of these azasugars revealed broad inhibition to all
kinds of glycosidases, and poor in vivo activity. In the therapeutic applications, potent
and speciﬁc inhibitor for one of the glycosidases could be more desirable. This requires
I'I'lore understanding of the enzyme properties and the action of the known inhibitors.
The three-dimensional structures of glycosidase complexes with DNJ and
eEitstanospermine have been solved and have revealed different binding modes. The
Structure for the complex")4 of DNJ with glucoamylase (Aspergillus awamori) has
provided information about the structural arrangement of functional groups and
interactions of the inhibitor in the active site. Glucoamylase catalyzes the removal of B-
D-glucose from the nonreducing ends of starch and other oligosaccharides, and it is an

105.l06

inverting glycosidase . Upon binding, the overall conformation of the protein in the

26

complex is essentially the same as the native structure at pH 6.0. The important residues
in the active site remain the same conformation and position, as seen earlier. There are
two molecules of DNJ bound in the active site with one associated with strong electron
density and representing the principal site of interaction. In the primary site, DNJ retains
the chair conformation (Figure 1.14) instead of half-chair which is the assumed transition
state conformation. In the active site of complex, 2-, 3-, 4- and 6- hydroxyl groups
occupy the approximate positions of water molecules in the free enzyme. Only one water
molecule (water 500) is present in the active site of the complex. This water molecule
forms hydrogen bonding to the oxygen of Glu400, 6-OH and ring nitrogen of DNJ, and it
is well oriented towards the C1 atom of the inhibitor (Figure 1.14). Therefore, it is
reasonable to assume that this water molecule is the nucleophile of the general base
catalyzed mechanism and that Glu400 is an important catalytic residue of the enzyme. In
the structure of 2-ﬂuoro-2-deoxy-cellotriosyl-enzyme intermediate of CelSA, the 6-OH
presents two possible conformations, but here the 6-OH is locked to only one
conformation: upwards relative to the base of the active site to hydrogen bond Asp 55
and water 500. Mutation of Asp 55 to Asn or 'l‘yr, and of Arg 54 to Lys or Tlrr leads to a

complete loss of activity, indicating the essential role of these hydroxyl groups.

27

Glu 400

 

 

Figure 1.14. Schematic of the active site of glucoamylase II. NOJl and N012 are l-

deoxynojirimycin at the primary and secondary site, respectively.

I07

The structure of the Exo-B-(l,3)-gluconase (Exg from Candida albicans) bound with

l08

castanospermine showed a different binding mode . Castanosperrnine is larger and less

ﬂexible than DNJ due to the fused 5-membered ring bridged by ethylene between the

9
86‘”) shows a

nitrogen and C6. The crystal structure of the unbounded castanospermine
chair conformation of the pyranose ring and an approximate envelope conformation of
the 5-membered ring. However, in the complex with Exg, it presents a twisted boat

conformation in which the positions of C3, C4, C5, C6 and N are almost superimposable

on the equivalent atoms in the chair form, but C2 and C1 are displaced upwards relative

28

to the base of the active site pocket (Figure 1.15). This conformation would facilitate the
departure of the leaving group and it can be considered a precursor to the transition state,
which is assumed to be a half-chair. The ring nitrogen forms close contact to the oxygen
of the nucleophile Glu229, consistent with an ionic interaction. The imino group also
shows good tetrahedron geometry, indicating protonation which would be expected at the
crystal pH of 6.0. Series of tight H-bonding contacts to the 2-, 3-, 4- hydroxyl groups lock
the castanospermine into place. Those hydroxyl groups also occupy the positions of water
molecules in the free enzyme. An important residue Glu 227 points centrally at the C3-
C4 bond so that each of its carboxylate oxygen atoms interacts with a hydroxy group. The
6-OH lies axially, and interacts with two water molecules. The protein again undergoes
very little change upon binding, except residue Asn 146 and Tyr 255, which are not
involved in the binding. Asn rotates away for the ligand in order to avoid a clash with C2
which has moved upwards to a twisted boat conformation. Because of the rigidity caused
by the 5-membered ring in castanospermine, Tyr 255 hydroxyl group moves 0.7 A away
to avoid the steric clash with C7. Again, conformations of protein, especially those of

important functional groups remain unchanged to achieve tight binding.

 

N146 E192

H135

Figure 1.15. Stereoscopic images of the difference (Fo- c) electron density contoured at
the 30 level, and corresponding to castanospermine bound in the active site. Surrounding
Exg residues are included with water molecules shown as bold circles.

Castemospermine and derivatives of deoxynojirimycin are considered selective inhibitors
for a-glucosidases, because they mimic the positive charge character of the ring oxygen
at transition state, which is believed to be an important feature of the a—glucosidase
transition state. However, it is questionable whether they are real transition state
analogues because they do not have the expected half-chair conﬁrmation. Therefore,
inhibitors that mimic both the positively charged ring oxygen and the half-chair
conﬁrmation have been developed.

The ﬁve-membered azasugars which is expected to be more ﬂexible than six-membered
rings were synthesized by C-H Wong et all"). The ﬁve-membered ring has more ﬂattened
chair conﬁrmation and the positive charge can be mimicked by the positive charged

imino group, so compound 19 showed better inhibition against a-glucosidase (K,= 2.8

30

pM'” for a-glucosidase) than DNJ and compound 20 showed strong inhibition to a-
galactosidase from coffee bean at pH 5.5 (K, = 0.05 ,uM). The x-ray crystal structurel '0 of
compound 19 indicated an envelop conformation of the ﬁve member ring. Figure 1.16
shows the proposed transition state for a-glucosidase and the positive charged compound
20. Therefore, ﬁve-membered azasugars are better transition state analogs than six-
membered ring azasugars. However, it is noted that compound 20 is a less potent
inhibitor to a-galactosidase compared to l-deoxy—D-galactonojirimycin 21 (K, = 1.6

nM)l 12, which has a chair conformation.
OH
on on on 0”
HO OH HO OH HO
OH
19 20 21

OH
OH

nu,+
HO
OH

20' H+

 

 

 

Figure 1.16. Proposed transition state structure for the u-galactosidase reaction.

Compound 20 is a mimic of the galactosyl cation.

31

Glycosidase inhibitors with spZ-hybridized anomeric carbon and ring nitrogen have been

designed to meet the ﬂattened chair conﬁrmation requirement. Series of glucoimidazole-

113-117 118-120 121

, glycotriazole -type inhibitors have been synthesized.

. glycopyrrole
However, most of them only showed weak or modest inhibition. The stronger inhibition
observed for compounds 28-32 (Table 1.3) was suggested to come from the interaction

of 2-NH3+ or ester function and the nucleophilic carboxylate of the enzymel '9.

Table 1.2. K, values in uM for compound 22-27.

 

 

R, R2 p-glu’ase' B—ngaseb
on
22 “q. u \ a, coocu3 H 6000 1150 (pH6.8)
23 ou— a, H coocrr3 300 410 (pH6.8)
24 coocrr3 coocrr, 25000 -
25 cnzon CHZOH r4000 -
on
N/ﬁ .
25 "Ono _N 0.7 new)
on
on
/N§N
27 no N
m/ .
-u
on

a Almond B-glucosidase. b Caldocellum saccharolyticum ﬁ-glucosidase.

 

Table 1.3. K, values in uM for compound 28-34.

 

 

OH R ﬁ-glucosaminidase‘ ﬁ-glu'ase"
N \ COzCH3 l9 -
"(I-l 28 Ac
O __
NHR 29 H - 12 (pH 6.8)
30 CF3CO 10 -
OH
H2. N \ 31 Ac 20 -
O __
NHFI 0026": 32 H ‘ 21(pH6.8)
33 CF3CO 13 -
OH
N\
u/ \ N
HO
@J
NHAc 34 0.2

3 Bovine kidney B-glucosaminidase. b Caldocellum saccharolyticum B-glucosidase.

 

122 4123

Compared to the imidazole analogue 26 and tetrazole analogs 27 & 3 , compounds
22-25 & 28-33 are relatively poor inhibitors, and they lack a basic nitrogen atom at the
glycosidic position. This might be essential to the strong inhibition and we will talk about
this kind of inhibitors later.

A class of ﬁve-membered imidazole compounds is also synthesized and tested (35-
37)121,124. Compound 36124 showed strong inhibition against a-mannosidase with ICso =

10 pM, K,= 5 pM), while other compounds are poor inhibitors. This is probably due to the

correct conﬁguration of hydroxyl groups in 35.

33

OH OH OH

. a I\\\‘“ \
\\ \\\“ \

«3 «3 " ~/\

HOlllmu N HOIIIum N "O N
\ \ \

35 36 37

A new class of glycosidase inhibitors which contain cyclic sulfonium ions is worth
mentioning. The natural products salacinol 38125 "26 and kotalanol 39m, the herbs used in
India transitional medicine for diabetes, are among the most potent glycosidase
inhibitors. The permanent positive charged sulfonium ion may contribute to the high

'28 of salacinol showed a unique spiro-

afﬁnity to the enzyme. The x-ray crystal structure
like conﬁguration of the inner salt comprised of l-deoxy-4-thioarabinofuranosyl
sulfonium cation and 1’-deoxyerythrosyl 3’-sulfate anion. This characteristic structure

may account for the strong inhibitory effect of salacinol.

on on on on on on
s+\/'\/\ 5+ on
HO HO

so, oso; on

OIIIIII

Salaclnol 38 Kotalanol 39

1.4.1.2. Inhibitors mimicking the charge build-up at the anomeric center
The positively charged anomeric carbncation 10 has not been paid much attention until
the mid-90s because its resonance form oxocarbenium ion 9 is believed to be a more

important feature of the transition state. In 1994, 8015 et al.'29"30 proposed that mimics of

34

the cation 10 could be transition state analogues. One way of incorporating a positive
charge at the anomeric center is to introduce a basic nitrogen to the anomeric position,
and replace the ring oxygen with methylene group in the corresponding sugar.
Isofagomine 41 (l-aza suagr), named after the natural product fagomine 40, was
synthesized. It has nitrogen at the anomeric position and glucose conﬁguration, while the

2-OH was removed because of the possible instability of the hemiaminalsl3 '.

OH OH on on
HO NH HO NH
HO HO NH HO NH w
' OH
OH
40 41 42 43
OH OH OH H
I
HO NH HO NH N OH
HO O
OH on on H
13 21 44

The inhibition studies showed that isofagomine 41 is a powerful glucosidase inhibitorI32

(Table 1.4), especially for B-glucosidase, with K, value 440 times lower than that
obtained for DNJ 13. Other l-aza sugars with different conﬁgurations were also
synthesized and tested by Ichikawa et a1.'33 They were found to be very potent and
speciﬁc against the corresponding B-glycosidases, with K, values in the nanomolar range

(Table 1.4).

Table 1.4. Inhibition constant K, in ,uM for isofagomines.

35

 

enzyme 41 13 42 21 43 44
(rt—glucosidasea 86 l 2.6
Isomaltasea 7.2 l 1

B-glucosidaseb 0.1 l 47
a-galactosidase° 50 0.0016
B-galactosidased 0, 3 28 8 1
B-galactosidasee 0,004
B-galactosidasef 0,2 1 25

u-fucosidaseg 4 0.029

a-fucosidaseh 6.4 0.0013

 

“ From yeast.7 From almonds. c From Green Coffee Bean. d From Saccharomyces
Fragilis. e From Aspargillus Oryzae. I From E. Cali. 8 From bovine kidney. h From
human placenta.

The galactose type l-aza sugar 42 is an extremely potent inhibitor of B-galactosidase,
with a K, value of 4.1 nM. The inhibition spectra of isofagomines are complementary to
those of DNJs, which show selective inhibition against corresponding a-glycosidases
(Table 1.4) by mimicking the positive charge developed at the ring oxygen. This
complementary inhibition proﬁle of DNJ and isofagomines is explained by the different
positions of the positive charge build-up of the transition state and the interaction of
inhibitors with the carboxylate groups of the enzymesl34'l36. For a-glucosidase, the
oxocarbenium ion 9 is believed to be an important state, and the catalytic nucleophile
carboxylate is located close to the [3 face of the pyranoside ring (Figure 1.17a), thus the
protonated nitrogen atom is well positioned to form a hydrogen bond or a salt bridge with

the anionic catalytic nucleophile. For B-glyosidases, 3 C-1 cation 10 is more important for

36

equatorial glycoside cleavage, with greater share of positive charge at anomeric carbon,
which is better suited to interact with the catalytic nucelophile of a B-glyosidase (Figure

1.171)).

e.

 

Figure 1.17. Pr0posed favored binding interaction of DNJ in a-glucosidase (a) and

isofagomine in a B-glucosidase (b).

The thermodynamic studies'37 of binding of isofagomine 41 and DNJ 13 to B-glucosidase
indicated very different binding modes for these two inhibitors (Table 1.5). The AH and
AS values were obtained by measuring K, at different temperatures and plotting —an, vs
l/T in a Van’t Hoff plot. The data show a negative enthalpy and small entropy for DNJ,
which indicates that the binding of DNJ to B-glucosidase is favorable in terms of bond
energies, while the entropy is relatively unimportant. This energy could be obtained from
forming a hydrogen bond between 2-OH and the enzyme as well as a strong hydrogen

bond or a salt bridge between the catalytic carboxylic acid group of the enzyme and ring

37

nitrogen. For isofagomine, the enthalpy is positive and the large entropy indicating the
binding of 41 is driven by the large increase of disorder, which can compensate for the

positive enthalpy.

Table 1.5. Thermodynamic parameters for the binding of the two inhibitors to almond B-

 

glucosidase at pH 6.8
K, (,uM, 25°C) AH (kJ/mol) AS (J/mol K)
OH
HO NH
110*
OH 26.3 -25.7 1.3
OH
HO
"° NH 0.27 58.6 323.8

 

A recent study of a new inhibitor noeuromycinI38 49, an analogue of isofagomine with
the 2-OH present, showed that the anomeric nitrogen can interact effectively with a-
glycosidases as well. This is contrary to the belief that anomeric positive charge is only
important to B-glycosidases. Noeuromycin (49) was tested for inhibition of glycosidases
(Table 1.6) and was found to be a remarkably strong glucosidase inhibitor. The K, values
were all in the nanomolar range and between 2 and 4000 times smaller than those of 41.
Evidently the incorporation of the 2-hydroxyl group increases inhibition profoundly. The
inhibitor 49 was also considerably more potent against glucosidases than the inhibitors
13, 21 and 19 resembling oxocarbenium ion intermediate 9 (Table 1.6). It is remarkable
that in contrast to 41 the inhibitor 49 inhibits both [3- and a-glucosidase strongly. The D-

galacto- (50) and L-fuco- (51) isomers of 49 were also glycosidase inhibitors in the

38

nanomolar range. The galactose analogue 50 is much more potent against a-galactosidase
than isogalactofagomine 42, while its inhibition of three B-galactosidases varied from 4
times more potent to 9 times less potent than 42. The L-fucose analogue 51 was 1000
times more powerful than isofucofagomine 43 against a-fucosidase. It is clear that the
incorporation of the 2-hydroxyl group contributes very tight binding to both 01- and B-
glycosidases. It was suggested that similar to the interaction between the anomeric

136,139

nitrogen and the nucleophilic carboxylate of B-glycosidases , a salt bridge also exists

between this nitrogen in 49, 50, 51 and the carboxylate of (it-glycosidases (Figure 1.18).

 

on on on
HO NH
..o& “0% ”9:;
NH NH OH
OH
41 42 43
on on on
HO NH
H
0" OH OH
49 50 51
Table 1.6. Inhibition constant K, in nM"
enzymes 41 49 42 50 43 51
a-glucosidaseb 86000 22 - - - -
isomaltaseb 7200 25 - - - -
B-glucosidasec l 10 69 - - - -
a-galactosidased - - 50000 742 - -
B-galactosidasee - - 328 91 - -
B-galactosidasef - - 4 35 - -

 

39

 

B-galactosidaseg - - 200 397 - -
a-fucosidaseh - - - - 4000 4.7
(r—fucosidasei - - - - 6400 3,2

 

“ -: inhibition not measured. D From yeast. 6 From almonds. d From Green Coffee Bean. 8
From Saccharomyces F ragilis. I From Aspargillus Oryzae. g From E. Coli.. h F rom bovine
kidney. ' From human placenta.

Glu453
o=<
0.
H\
o
A o 0" Enz
Gln o
N—H. <
H ‘0 HO // o
0 NH "0 N
/ \ H
/§N"H q HO
my ’11 g 0
Hall /0
"18137 GIN”?
”"182
a b

Figure 1.18. Proposed binding of 49 in the active site of a B—glucosidase (a), in this case

from white clover and a retaining a-glucosidase (b).

The importance of the charge-mimicry of transition state has been demonstrated by
azasugar analogs. Although it is assumed that protonated azasugars bind to the enzymes,
no direct evidence of the protonation state of azasugars has been obtained until recently.
The 1.0 A resolution three-dimensional structure of Ce15A from Bacillus agaradhaerens

in complex with compound 52 was solved by Davies al et '40.

40

OH

OH
"0 0 110 NH

OH
52
Figure 1.19 is the schematic representation of the interactions. It reveals that the imino
sugar is protonated within the active site. The carboxylate oxygen of Glu 228 is 2.6 A
form N1 of 52, indicating a close Coulombic interactions between the protonated
isofagomine and the negatively charged nucleophile. It is also noted that glucose and
isofagomine moieties are in undistorted 4C, (chair) conﬁrmations. This is the ﬁrst direct

observation of the protonation state of an azasugar glycosidase inhibitor upon binding.

 

 

 

0231101234
29A
OI-l “(mix/30A \ w _-
Wk N’H \3. 1A Tyr202
H9 \M
' ”.29
H," NH 26A Glu139
-o ’ -
\ I' (0 ~--
I )1 ND 0Y0
/
Tyr66 HIS‘IO‘I I'I \GIUZZB
L I L.-. J
-2 -1

Figure 1.19. Schematic representation of the interactions observed between CelSA and
the cellobio-derived isofagomine (52). All hydrogen atoms shown have been observed
experimentally. Only the -l subsite interactions are shown in detail. Distances around N]
are indicated.

1.4.1.3. Inhibitors which mimic the positively charged exocyclic oxygen

41

Since glycoside cleavage of an ordinary O-glycoside involves protonation of the
exocyclic oxygen of the glycosidic bond, an early transition state may have a
considerable charge at this atom. Thus, stable compounds that resemble the positively
charged ion 11, may be regarded as transition-state analogues. Compounds that have an
amine in place of the exocyclic oxygen are the most common transition-state analogs of
this kind.

Some natural inhibitors which contain an exocyclic nitrogen have been isolated. The
main classes of this type of compounds are the trehazolins 53, the allosamidines 54,
acarbose 55. The trehazolins exhibit powerful speciﬁc inhibition of various trehalases;

the allosamidines inhibit various chitinases, and acarbose is a very potent a-glucosidase

 
   

inhibitor'“.
OH
éio
HN HOﬁO’ﬁ/mo
7 OH OH I
0 OH
"”’OH
HO OH
Trehazoline 53 Allosamidin 54
OH
HO
HO H C
OH 3
HN O
HO OH
OH
O O
HO OH
OH
O O
Hm OH
OH

Acarbose 55

42

Glycosylamine are classical mimics of the protonated substrate. However, these

compounds are not stable and hydrolyze very quickly under assay conditions'”.

Nevertheless, a fucosylamine was synthesized and found to be a potent inhibitor (K, =

I43

0.75 ,uM) against a-L-fucosidase from bovine kidney . More stable forms have been

investigated, including glycosylamidines'“, C-(glycopyranosyl)methylamines”5"“,
amino-substituted (R/S)-phenylglucosides”mm"5°. No powerful inhibition has been

8'44 showed selective

observed except for compounds 56-60. Glycosylamidines 56-5
inhibition against glucosidases, especially B-glucosidases. Compounds 59147 and 60'49 are

quite potent and selective inhibitors against B-galactosidases and a-glucosidases.

Table 1.7. K, values in ,uM

 

Compounds 56-58 (x-glual a-glub [i-gluc [i-glud a-galc [i-galf

on
0
H190 0T‘Ph
0H NH; 21 >2500g 73 0.094 - -
Hon
HO O 0
011 mph - - 41 390 47 7.8
2
HaC
H o
morph
on "Hz. - - 1100 11 - -

' No inhibition at 2500 pM.“ Yeast a-glucosidase.b Aspergillus niger a-glucosidase.c
Almond B—glucosidase.d Aspergillus niger B-glucosidase." Aspergillus niger a-
galactosidasef E. coli [3_--galactosidase.’g Less than 50 % inhibition at 2500 ,uM.

 

43

OH OH

0 H o
"mecnzph 9'0
OH

OH
H2N Ph
59K.=2.3pM 60 K.=38pM
(E. coli lacZ ﬂ-galactosidase) (yeast a-glucosidase)

Another important class of compounds is aminocyclopentanes. They have a
stereochemistry similar to the stereochemistry of either gluco-, manno-, galacto-, fuco-,
or xylopyranose. Great progress has been made and some strong and speciﬁc inhibitors
have been synthesized. The amino(hydroxymethyl)cyclopentanetriols 61-63 with
different substituents R have shown powerful inhibition against mannosidases,
glucosidases, and galactosidases, with inhibition constants in the milimolar or

micromolar range. There is a review by Siriwardena et. all5 ' about this class of inhibitors.

 

61 62 63

1.4.1.4. Inhibitors that mimic positive charge in several places (Mimics of 9 & 11)

Both the endocyclic oxygen positive charge and the exocyclic positive charge have been
shown to be important features of glycosidase transition states. It is reasonable to assume
that stable compounds which mimic both the positively charged ion 9 and 11 are better

transition state analogs.

44

O W0
\\/
(Hot/g \ (H01n/ \ﬁ

HOB
9 11

Ganem Group ﬁrst synthesized amidine derivatives (64 and 65) of D-glucose‘sz‘lsz‘, which
combine the charge distribution at endo- & exocyclic oxygen and the desired
conformation of glucosyl cation 9. The amidine derivatives 64 and 65 have nitrogen in
place of both endo-and exocyclic oxygens, so the positive charge can be delocalized
among these two positions and anomeric center (66). The spz-hybridized anomeric carbon
provides the ﬂattened chair conformation of the assumed transition state. Inhibition
testing found that amidine 64 and 65 are potent, broad-spectrum inhibitors of
glycosidases (Table 1.8). It showed similar inhibition against B-glucosidase, B-
galactosidase and u-mannosidase, and the N, N-dimethyl derivative 65 is less potent than

64.

OH OH

HO N HO N
11

NH2 N/
OH on I

 

Table 1.8. K, values in pM for compound 64-68.

 

Enzymes [i-glucosidasea B-galactosidaseb a-mannosidaseC
64 8 8 9
65 83 83 ° 40
67 8.4 l 9 3. l

45

68 13.8 19 12

a Sweet almond B-glucosidase. b Bovine liver B-galactosidase ° Jack bean a-mannosidase

 

More stable forms amidrazone 67 and amidoxime 68 have been synthesized and
testedm‘lss. They all showed broad spectra inhibition of glycosidases. Despite the large
difference in the pKa values of amidine 64 (pKa = 10.4), amidrazone 67 (pKa = 8.7), and
amidoxime 68 (pKa = 5.6), their inhibition against B-glucosidase (sweet almond) is at the
same scale. This let Ganem et. a1 propose that it is the shape, not the charge of transition

state more important for binding to B-glucosidase.

OH OH

67 68

Another important class of inhibitors which mimic the ﬂattened chair conformation and

the positive charge at both positions consist of an azasugar fused with heterocyles such as

136,156. A

triazoles or imidazoles series of compounds can be obtained by introducing

different substituents at the heterocyles, and they are more stable than amidines and their

derivatives. The natural product nagstatinl5 7"”

(69) was shown to be an N-
acetylgalactosamine fused with a substituted imidazole, and it is a strong N-acetyl-B-

glucosaminidase inhibitor (Table 1.9). Other compounds which have different

46

98136,156,159

conﬁgurations and different substituents have been synthesized by Vasella and

”1'60"“ respectively. As can be seen in Table 1.9, the imidazoles are

Tatsutas’s groups
potent inhibitors of various glycosidases and more selective than the amidines and their
derivatives. Also, the 1,2,4-triazoles are potent and selective inhibitors though they are

not as strong as the imidazoles.

Table 1.9. K, values in pM for compound 69-75.

 

Enzymes a- [3- [3- B- a- B- NAc- NAc-
glua giub gluc gald manc manf B- u- -
glug galh

O
"ﬁ/l
11%" 69 >334 >334 0.013 63
1-1 N \
m 70 59 0.1 0.02

OH
o/égj
"0 —N 72 0.034

N \
"0 j 73 319 11 0.006 10

74 85 14

47

75 285 65 0.4

H
OH0

N \
"0 air—7. 76 0.5 0.008 >500 370 60 125

mICso value. aYeast a-glucosidase. bAlmond B-glucosidase. cCaldocellum saccharolyticum
B-glucosidase. dE.coli B-galactosidase. °Jack bean a-mannosidase. fSnail B-mannosidase.
gBovine kidney N-acetyl-B-glucosaminidase. hChicken liver N-acetﬂ-a-glucosaminidase

The selectivity and strength of inhibitors may be enhanced by introducing substituents
mimicking the aglycon. Introduction of the hydrophobic and ﬂexible phenethyl
substituent to the imidazole derivative 70 increased the inhibition of C. saccharolyticum

B-glucosidase, leading to the strongest known inhibitor of a B-glucosidase (K, = 0.11

nM)l65.

OH

Ph
—N

OH

77
1.4.1.5. Inhibitors which do not mimic charge
Because the proposed transition state has considerable sp2 hybridization with half-chair
conformation, a number of nonbasic analogs have been developed to mimic this feature.
8-A1donolactones and lactams have been found long time ago to be able to inhibit
glycosidases because they resemble the transition state geometrically. Recently
Nishimura’s group synthesized and studied all possible D-stereoisomers of hexono-l,5-
lactam and found D-manno isomer 78 was a good inhibitor of both a-mannosidase (jack

bean K, = 68 yM) and B-mannosidase (snail K, = 9 ,uM) as well as B-glucosidase, and B-

48

galactosidase. The gluco, manno, galacto, talo and ido isomers were able to inhibit

almond B-glucosidase'“. The D-glucosaminolactam 79 is also a good inhibitor of [i-

glucosidase (C. saccharomyces K, = 0.7 ,uM)'67.

OH 0"
Ho 03H" H30“ S& "E"
Ho 0 NH, 0
7s 79

A new type of lactams were introduced by Wither’s group in which the nitrogen is in the
anomeic position similar to the isofagomines and the carbonyl group in the 2-positionl68.
The rationale of this design is to incorporate the important 2-OH group to the l-N-
iminosugars and provide planarity for this molecule, if these lactams bind as the
protonated iminol tautomers. The D-galacto analogI69 (80) showed strong inhibition
against B-galactosidase from Aspergillus oryzae, and xylobiose analog inhibits Cex-
xylanase with K.- = 0.34 ,uM. However, the atomic-resolution structure of lactam with
xylanase XynlOA ﬁ'om Streptomyces lividans revealed that the lactam is bound to the
enzyme as the amide tautomer, which is consistent with the free energy cost of amide-

iminol isomerization (~45 kJ/mol) (Figure 1.20).

OH
OH

HO NH

80

49

HO
rim/0 —"—-— "time
OH HO
NH 01'" HO
O H

Figure 1.20. Amide-iminol isomerization

Other neutral compounds indcluding phosphonates, phosphonamidesI70 and
thioalditols|7l have been studied, but they were found to be poor inhibitors of

glycosidases.

1.4.2. Glycosyltransferase Inhibitors

Glycosyltransferases are essential in the oligosaccharide synthesis. Abnormal cell surface

glycosylation characterizes several cancers such as ovarian cancersm, breast cancerm’

175 6

, prostate cancer17 , and colon cancerm'I78

among a host of others such as
oropharyngeal cancer, leukemias and Hodgkin’s diseasem'lss. Many infectious diseases
are also mediated by glycosylation. Therefore, different strategies have been explored to
identify potent inhibitors of glycosyltransferases in order to modulate the formation of
oligosaccharides and glycoconjugates, study functions of these biopolymers, and develop

pharmaceutical agents. These strategies include design of analogs of donor molecules'sq‘

195 196-200 66.201 -204

, acceptor molecules transition state mimics and glycosyl-enzyme
intermediateszos. A few successful examples have been reported, but rational design of
transferase inhibitors still remain a difﬁcult task because of the intrinsic features of
glycosyltransferases: complex transition state with sugar donor, acceptor oligosaccharide
where the sequence and linkage are extremely critical; weak binding of the enzyme with

their natural substrates (usual Km values are in the mM range); and few structural data.

50

A remarkable donor analog of sialyltransferase has been reported by Schmidt et. a1'95 .

The donor substrate for sialyltransferase, independent of their source and their acceptor
speciﬁcity, is cytidine monophosphate N-acetylneuraminic acid (CMP-NeuSAc). By
incorporating a ﬂattened neuraminyl residue, phosphonate, and a cytidine
monophosphate (CMP) residue, a powerful sialyltransferase inhibitor (81) can be
obtained. The inhibition constant is in nanomolar range (40 nM). Other analogs with a
simple aryl or hetaryl moiety instead of the neuraminyl residue can also give potent
inhibitors“. Therefore, a ﬂat pyranosyl ring mimic, phosphonate and CMP residue are

important for a potent sialyltransferase inhibitor.

""2
\N AcHN ll
ﬂ ! “\\\\P\-O.NC’
__ _ N o HO __ —"
= o- ‘ J. '
HO : o H ‘a, ll 0
OO 9 o N
AcHN 2 OH O
HO OH OH
OH OH
CMP-NeuSAc 81

Because donor sugars are common substrates for glycosyltransferases, as mentioned
earlier, the speciﬁcity of transferases largely resides on the recognition of the acceptor
structures. One approach to speciﬁc glycosyltransferase inhibitors is to design the
acceptor analogs. A simple strategy is to remove the reactive hydroxyl group of a
glycosyltransferase oligosaccharide acceptor since presumably this hydroxyl group is not
essential to the binding of the acceptor. Palcic et 21] prepared eight acceptor analogs for
different glycosyltarnsferases‘96 where the reactive hydroxyl groups were replaced by

hydrogen. Four analogs showed competitive inhibition against the corresponding

51

glycosyltransferases, and the inhibition constants were in the range of the Km values, with
greater values. The other four analogs showed no inhibition, which is suggested to be
lack of the critical hydrogen bonding between this acceptor hydroxyl group and the
enzyme.

A peptide acceptor analog has been reported by Imperiali et. alZOO. The compound 82 is a
nanomolar inhibitor of the oligosaccharyl transferase (OT), which is acting on the
Asparagine-linked glycosylation. This class of cyclic peptide provided the ﬁrst example

of a readily available and adaptable family of potent protein glycosylation inhibitors.

3
o
0
HH
[1 0
HM
l H
HOWE,
+H3N
CH, O
82

In 1999, Ichikawa et. al205 reported a rationally designed a-l,3-ga1actosyltransferase (a
1,3-GalT) inhibitor 83. They proposed since a 1,3-Ga1T is a retaining tarnsferase, a
double displacement with formation of a glycosyl-enzyme intermediate is the possible
mechanism (Figure 1.21). Molecules with nitrogen at the anomeric position should be
able to mimic the glycosyl-enzyme intermediate by forming an inhibitor-enzyme
complex via a favorable electrostatic interaction. The pyrophosphate moiety and the
uridine residue are replaced with a vicinal diol and a 5’-thio-uridine respectively, to avoid
an unnecessary negative charge on the molecule and to increase the stability of the

designed inhibitor. Compound was found to be a potent and selective a 1,3-GalT

inhibitor with K, value 4.4 ,uM, but not of B 1,4-GalT which is an inverting tamsferase
and does not involve a glycosyl-enzyme intermediate with a nearby carboxylate. It is
noted that l-deoxygalactonojirimycin has no inhibition against a 1,3-GalT, nor [3 1,4-

GalT, indicating the importance of the position of the imino group.

OH Lo

HO .0 O
o (

"° 0 o "" I
on u 11 o N
(O-P—O—P—O I

1 1 O
O' O‘

/ Q °"
' OH - 7 OH
Ho I Ho I
o 0
Ho 0 _2/l » Ho '”0.
HH+
OHK

HO—Aeceptor _j

 

 

 

 

—

GlycosyI-enzyme complex Inhibitor-enzyme complex

OH OH
O OH OH OH
HOSL“V:H -Jg,,o 0’ 0
OH
034w”, o—§

0" NHAc
a 1,3-Gal epitope

Figure 1.21. Proposed reaction mechanism ofa 1,3-Ga1T and possible interaction

between the l-N-iminosugar-based inhibitor and the enzyme.

53

14‘207'209 in the biosynthesis of cell surface oligosaccharide

Fucosyltransferases are critica
since many of them are fucosylated and these fucose-containing structures accumulate in
a large variety of human canceresz'm' '. Inhibition of these enzymes is important to study
the oligosaccharide and to develop anti-inﬂammatory and anti-tumor agents. C-H Wong
et al."6 reported a synergistic inhibition of human a-l,3-fucosyltransferase V (Figure
1.22). The azatrisaccharide 85 which incorporate a fuconojirimycin moiety to the
acceptor substrate was synthesized to mimic the proposed transition state with substantial
oxocarbenium ion character. This azatrisaccharide alone cause moderate inhibition
against ﬁicosyltransferase with 5.7 mM. However, in the presence of 0.03 mM GDP, it
showed dramatically increased inhibition with ICso 0.031 mM. It was suggested that the
azatrisaccharide form a complex with GDP via electrostatic interactions between the
positively charged amino group and the negatively charged phosphate oxygen. This can

mimic the transition state of the enzymatic reaction (Figure 1.23).

54

 

 

 

 

 

HO
H30 0 0" 3.1,3.Fueosyltransferasev H 0H
s N
H 0" / NHAc OH
00 OH
OH 0 H 0" NHAc
HO OH O O
OH
L J
sialyl
Lewis x
84 85

Figure 1.22. Structure of sialyl Lewis x. and the azatrisaccharide. a-l,3-

fucosyltransferase V catalyze the addition of fucose to LacNAc or sialyl LacNAc.

 

 

PM“
I” “\ F.- — t
1 w
o§ \
/':'_°_/i’_o I;O;G| T T
°\
45’ ° \/P-O-/7-O o
5 OH OH 5'9 o
H 1 :
H 5 3 5 OH OH
Hac 5 OH
0" i
H o" NHAc 3
° :
O 0 ON Aeooptor-—-O-H-----:B/(
HO 0 5+
OH OH 1— _J

Figure 1.23. Proposed model for the synergistic inhibition of a—l,3-fucosyltransferase V
by the combination of an aza sugar and GDP and the proposed mechanisms of hydrolysis
fucosylation by 01- 1,3-fucosyltransferase V.

55

1.5. References
(1) Bertozzi, C. R.; Kiessling, L. L. Chemical glycobiology Science 2001, 291, 2357-
2364.

(2) Jenkins, N.; Parekh, R. B.; James, D. C. Getting the glycosylation right: Implications
for the biotechnology industry Nature Biotechnology 1996, 14, 975-981.

(3) Helenius, A.; Aebi, M. Intracellular functions of N-linked glycans Science 2001, 291,
2364-2369.

(4) Kornfeld, R.; Komfeld, S. Assembly of Asparagine-Linked Oligosaccharides Annual
Review of Biochemistry 1985, 54, 631-664.

(5) Gahmberg, C. G.; Tolvanen, M. Why mammalian cell surface proteins are
glycoproteins Trends in Biochemical Sciences 1996, 21, 308-311.

(6) Burda, P.; Aebi, M. The dolichol pathway of N-linked glycosylation Biochimica Et
Biophysica Acta-General Subjects 1999, 1426, 239-257.

(7) Silberstein, S.; Gilmore, R. Biochemistry, molecular biology, and genetics of the
oligosaccharyltransferase F aseb J. 1996, 10, 849-858.

(8) Moremen, K. W.; Trimble, R. B.; Herscovics, A. Glycosidases of the Asparagine-
Linked Oligosaccharide Processing Pathway Glycobiology 1994, 4, 113-125.

(9) Henrissat, B. B., A. Biochem. J. 1996, 316, 695-696.

(10) Wolfenden, R. L., X.; Young, G. Spontaneous hydrolysis of glycosides J. Am.
Chem. Soc. 1998, 120, 6814-6815.

(1 l) Sinnott, M. L. Catalytic Mechanisms of Enzymatic Glycosyl Transfer Chem. Rev.
1990, 90, 1171-1202.

(12) Legler, G. Glycoside hydrolases: Mechanistic information from studies with
reversible and irreversible inhibitors Adv. Carbohydr. Chem. Biochem. 1990, 48, 319-
385.

(13) Davies, G. S., M. L.; Withers, S. G. Glycosyl transfer. In Comprehensive Biological
Catalysis; Academic Press, 1998; Vol. 1, 119-208.

(14) Kernpton, J. B.; Withers, S. G. Mechanism of Agrobacterium Beta-Glucosidase -
Kinetic-Studies Biochemistry 1992, 31, 9961-9969.

56

(15) Zechel, D. L.; Withers, S. G. Glycosidase mechanisms: Anatomy of a ﬁnely tuned
catalyst Acc. Chem. Res. 2000, 33, 11-18.

(16) Koshland, D. B. Biol. Rev. 1953, 28, 416-436.

(17) McCarter, J. D.; Withers, S. G. Mechanisms of Enzymatic Glycoside Hydrolysis
Current Opinion in Structural Biology 1994, 4, 885-892.

(18) Wang, Q. P.; Graham, R. W.; Trimbur, D.; Warren, R. A. J.; Withers, S. G.
Changing Enzymatic-Reaction Mechanisms by Mutagenesis - Conversion of a Retaining
Glucosidase to an Inverting Enzyme J. Am. Chem. Soc. 1994, 116, 11594-11595.

(19) Davies, G.; Henrissat, B. Structures and Mechanisms of Glycosyl Hydrolases
Structure 1995, 3, 853-859.

(20) Tews, I., Perrakis, A., Oppenheim, A., Dauter, 2., Wilson, K. S. & Vorgias, C. E.
Bacterial chitobiase structure provides insight into catalytic mechanism and the basis of
Tay-Sachs disease Nature Struct. Biol. 1996, 3, 638-648.

(21) Varrot, A.; Schulein, M.; Davies, G. J. Insights into ligand-induced conformational
change in Ce15A from Bacillus agaradhaerens revealed by a catalytically active crystal
form J. Mol. Biol. 2000, 297, 819-828.

(22) D. M. A. Guerin, M.°B. L., M. Costabel, H. Souchon, V. Lamzin, P. Beguin and P.
M. Alzari Atomic (0.94 A) resolution structure of an inverting glycosidase in complex
with substrate J. Mol. Biol. 2002, 316, 1061-1069.

(23) Sulzenbacher, G.; Driguez, H.; Henrissat, B.; Schulein, M.; Davies, G. J. Structure
of the Fusarium oxysporum endoglucanase l with a nonhydrolyzable substrate analogue:
Substrate distortion gives rise to the preferred axial orientation for the leaving group

Biochemistry 1996, 35, 15280-15287.

(24) Sulzenbacher, G.; Schulein, M.; Davies, G. J. Structure of the endoglucanase I ﬁom
Fusarium oxysporum: Native, cellobiose, and 3,4-epoxybutyl beta-D-cellobioside-
inhibited forms, at 2.3 angstrom resolution Biochemistry 1997, 36, 5902-5911.

(25) J.-Y. Zou, G. J. K., J. S'tahlberg, H. Driguez, W. Nerinckx, M. Claeyssens, A.
Koivula, T. T. Teeri and T. A. Jones Crystallographic evidence for substrate ring

distortion and protein conformational changes during catalysis in cellobiohydrolase
Cel6A from Trichoderma reesei Structure 1999, 7, 1035-1045.

(26) Varrot, A.; Schulein, M.; Fruchard, S.; Driguez, H.; Davies, G. J. Atomic resolution
structure of endoglucanase Ce15A in complex with methyl 4,4(11),4(Ill),4(IV) tetrathio-
alpha-cellopentoside highlights the alternative binding modes targeted by substrate
mimics Acta Crystallographica Section D-Biological Crystallography 2001, 57, 1739-
1742.

57

(27) Juers, D. H.; Heightman, T. D.; Vasella, A.; McCarter, J. D.; Mackenzie, L.;
Withers, S. G.; Matthews, B. W. A structural view of the action of Escherichia coli (lacZ)
beta-galactosidase Biochemistry 2001, 40, 14781-14794.

(28) Hrrnova, M., Varghese, J. N., De Gori, R., Smith, B. J ., Driguez, H. & Fincher, G. B.
Catalytic mechanisms and reaction intermediates along the hydrolytic pathway of a plant
beta-D-glucan glucohydrolase Structure 2001, 9, 1005-1016.

(29) Varrot, A.; Frandsen, T. P.; Driguez, H.; Davies, G. J. Structure of the Humicola
insolens cellobiohydrolase Ce16A D416A mutant in complex with a non-hydrolysable
substrate analogue, methyl cellobiosy1-4-thio-beta-cellobioside, at 1.9 angstrom Acta
Crystallographica Section D—Biological Crystallography 2002, 58, 2201-2204.

(30) Davies, G. J .; Mackenzie, L.; Varrot, A.; Dauter, M.; Brzozowski, A. M.; Schulein,
M.; Withers, S. G. Snapshots along an enzymatic reaction coordinate: Analysis of a
retaining beta-glycoside hydrolase Biochemistry 1998, 37, 1 1707-1 1713.

(31) Varrot, A.; Davies, G. J. Direct experimental observation of the hydrogen-bonding
network of a glycosidase along its reaction coordinate revealed by atomic resolution
analyses of endoglucanase Ce15A Acta Crystallographica Section D-Biological
Crystallography 2003, 59, 447-452.

(32) Nashiru, 0.; Zechel, D. L.; Stoll, D.; Mohammadzadeh, T.; Warren, R. A. J.;
Withers, S. G. beta-Mannosynthase: Synthesis of beta-mannosides with a mutant beta-
mannosidase Angew. Chem, Int. Ed. Engl. 2001, 40, 417-420.

(33) Ducros, V. M. A.; Zechel, I). L.; Murshudov, G. N.; Gilbert, H. J.; Szabo, L.; Stoll,
D.; Withers, S. G.; Davies, G. J. Substrate distortion by a beta-mannanase: Snapshots of
the Michaelis and covalent-intermediate complexes suggest a B-2,B-5 conformation for
the transition state Angew. Chem, Int. Ed. Engl. 2002, 41, 2824-+.

(34) Withers, S. G.; Street, I. P.; Bird, P.; Dolphin, D. H. 2-Deoxy-2-Fluoroglucosides - a
Novel Class of Mechanism-Based Glucosidase Inhibitors J. Am. Chem. Soc. 1987, 109,
7530-7531.

(35) Withers, S. G.; Aebersold, R. Approaches to Labeling and Identiﬁcation of Active-
Site Residues in Glycosidases Protein Science 1995, 4, 361-372.

(36) Cleland, W. W. K., M.M. Low-barrier hydrogen bonds and enzymic catalysis.
Science 1994, 264, 1887-1890.

(37) Gerlt, J. A., Kreevoy, M.M., Cleland, W. & Frey, P.A. Understanding enzymic

catalysis: the importance of short, strong hydrogen bonds Chemistry & Biology 1997, 4,
259-267.

58

(38) Notenboom, V.; Birsan, C.; Nitz, M.; Rose, D. R.; Warren, R. A. J.; Withers, S. G.
Insights into transition state stabilization of the beta-1,4-glycosidase Cex by covalent
intermediate accumulation in active site mutants Nature Struct. Biol. 1998, 5, 812-818.

(39) Namchuk, M. N.; Withers,'S. G. Mechanism of Agrobacterium beta-glucosidase:
Kinetic analysis of the role of noncovalent enzyme/substrate interactions Biochemistry
1995, 34, 16194-16202.

(40) Vasella, A.; Davies, G. J .; Bohm, M. Glycosidase mechanisms Current Opinion in
Chemical Biology 2002, 6, 619-629.

(41) Mackenzie, L. F.; Sulzenbacher, G.; Divne, C.; Jones, T. A.; Woldike, H. F.;
Schulein, M.; Withers, S. G.; Davies, G. J. Crystal structure of the family 7
endoglucanase I (Ce17B) from Humicola insolens at 2.2 angstrom resolution and
identiﬁcation of the catalytic nucleophile by trapping of the covalent glycosyl-enzyme
intermediate Biochem. J. 1998, 335, 409-416.

(42) Sinnott, M. L.; Souchard, I. J. Mechanism of Action of Beta-Galactosidase - Effect
of Aglycone Nature and Alpha-Deuterium Substitution on Hydrolysis of Aryl
Galactosides Biochem. .1. 1973, 133, 89-98.

(43) Tull, D.; Withers, S. G. Mechanisms of Cellulases and Xylanases - a Detailed
Kinetic-Study of the Exo-Beta-l,4-Glycanase from Cellulomonas—Fimi Biochemistry
1994, 33, 6363-6370.

(44) McCarter, J. D.; Yeung, W.; Chow, J.; Dolphin, D.; Withers, S. G. Design and
synthesis of 2'-deoxy-2'-ﬂuorodisacchan'des as mechanism-based glycosidase inhibitors
that exploit aglycon speciﬁcityJ. Am. Chem. Soc. 1997, 119, 5792-5797.

(45) McCarter, J. D.; Withers, S. G. S-ﬂuoro glycosides: A new class of mechanism-
based inhibitors of both alpha- and beta-glucosidases J. Am. Chem. Soc. 1996, 118, 241-
242.

(46) Withers, S. G.; Rupitz, K.; Street, I. P. 2-Deoxy—2-F1uoro-D-Glycosyl Fluorides - a
New Class of Speciﬁc Mechanism-Based Glycosidase Inhibitors J. Biol. Chem. 1988,
263, 7929-7932.

(47) Burrneister, W. P.; Cottaz, S.; Driguez, H.; Iori, R.; Palrnieri, S.; Henrissat, B. The
crystal structures of Sinapis alba myrosinase and a covalent glycosyl-enzyme
intermediate provide insights into the substrate recognition and active-site machinery of
an S-glycosidase Structure 1997, 5, 663-675.

(48) Notenboom, V.; Birsan, C.; Warren, R. A. J.; Withers, S. 6.; Rose, D. R. Exploring

the cellulose/xylan speciﬁcity of the beta-l,4-glycanase Cex from Cellulomonas ﬁmi
through crystallography and mutation Biochemistry 1998, 37, 4751-4758.

59

(49) White, A.; Tull, D.; Johns, K.; Withers, S. G.; Rose, D. R. Crystallographic
observation of a covalent catalytic intermediate in a beta-glycosidase Nature Struct. Biol.
1996, 3, 149-154.

(50) Vocadlo, D. J.; Davies, G. J .; Laine, R.; Withers, S. G. Catalysis by hen egg-white
lysozyme proceeds via a covalent intermediate Nature 2001, 412, 835-838.

(51) Sidhu, G.; Withers, S. G.; Nguyen, N. T.; McIntosh, L. P.; Ziser, L.; Brayer, G. D.
Sugar ring distortion in the glycosyl-enzyme intermediate of a family G/ 11 xylanase
Biochemistry 1999, 38, 5346-5354.

(52) Burrneister, W. P.; Cottaz, S.; Rollin, P.; Vasella, A.; Henrissat, B. High resolution
x-ray crystallography shows that ascorbate is a cofactor for myrosinase and substitutes
for the function of the catalytic base .1. Biol. Chem. 2000, 275, 39385-39393.

(53) Sulzenbacher, G.; Mackenzie, L. F.; Wilson, K. S.; Withers, S. G.; Dupont, C.;
Davies, G. J. The crystal structure of a 2-ﬂuorocellotriosyl complex of the Streptomyces
lividans endoglucanase Ce1B2 at 1.2 angstrom resolution Biochemistry 1999, 38, 4826-
4833.

(54) Sabini, E.; Wilson, K. S.; Danielsen, S.; Schulein, M.; Davies, G. J. Oligosaccharide
binding to family 11 xylanases: both covalent intermediate and mutant product
complexes display B-2,B-5 conformations at the active centre Acta Crystallographica
Section D-Biological Crystallography 2001, 5 7, 1344-1347.

(55) Sabini, E.; Sulzenbacher, G.; Dauter, M.; Dauter, Z.; Jorgensen, P. L.; Schulein, M.;
Dupont, C.; Davies, G. J.; Wilson, K. S. Catalysis and speciﬁcity in enzymatic glycoside
hydrolysis: a B-2,B-5 conformation for the glycosyl-enzyme intermediate revealed by the
structure of the Bacillus agaradhaerens family 11 xylanase Chemistry & Biology 1999, 6,
483-492.

(56) Weymouth-Wilson, A. C. Nat. Prod. Rep. 1997, 14, 99-110.

(57) Unligil, U. M.; Rini, J. M. Glycosyltransferase structure and mechanism Current
Opinion in Structural Biology 2000, 10, 510-517.

(58) Bourne, Y.; Henrissat, B. Glycoside hydrolases and glycosyltransferases: families
and functional modules Current Opinion in Structural Biology 2001, 11, 593-600.

(59) Wrabl, J. O.; Grishin, N. V. Homology between O-linked GlcNAc transferases and
proteins of the glycogen phosphorylase superfamily J. Mol. Biol. 2001, 314, 365-374.

(60) Chamock, S. J.; Davies, G. J. Structure of the nucleotide-diphospho-sugar

transferase, SpsA from Bacillus subtilis, in native and nucleotide-complexed forms
Biochemistry 1999, 38, 6380-6385.

60

(61) Tarbouriech, N.; Chamock, S. J .; Davies, G. J. Three-dimensional structures of the
Mn and Mg dTDP complexes of the family GT-2 glycosyltransferase SpsA: A
comparison with related NDP-sugar glycosyltransferases J. Mol. Biol. 2001, 314, 655-
661.

(62) Ha, S.; Walker, D.; Shi, Y. G.; Walker, S. The 1.9 angstrom crystal structure of
Escherichia coli MurG, a membrane-associated glycosyltransferase involved in
peptidoglycan biosynthesis Protein Science 2000, 9, 1045-1052.

(63) Hu, Y. N.; Chen, L.; Ha, S.; Gross, B.; Falcone, B.; Walker, _D.; Mokhtarzadeh, M.;
Walker, S. Crystal structure of the MurG : UDP-GlcNAc complex reveals common
structural principles of a superfamily of glycosyltransferases Proc. Natl. Acad. Sci. U. S.
A. 2003, 100, 845-849.

(64) Davies G, S. M. L. a. W. S. G. Glycosyl transfer. In: Sinnott ML (Ed)
Comprehensive Biological Catalysis: A Mechanistic Reference; San Diego: Academic
Press, 1998; Vol. 1, 119-209.

(65) Murray, B. W.; Wittmann, V.; Burkart, M. D.; Hung, S. C.; Wong, C. H. Mechanism
of human alpha-1,3-fucosyltransferase V: Glycosidic cleavage occurs prior to
nucleophilic attack Biochemistry 1997, 36, 823-831.

(66) Qiao, L.; Murray, B. W.; Shimazaki, M.; Schultz, J.; Wong, C. H. Synergistic
inhibition of human alpha-l,3-ﬁ1cosyltransferase V J. Am. Chem. Soc. 1996, 118, 7653-
7662.

(67) Singh, A. N.; Hester, L. 8.; Raushel, F. M. Examination of the Mechanism of
Sucrose Synthetase by Positional Isotope Exchange J. Biol. Chem. 1987, 262, 2554-2557.

(68) Kim, S. C.; Singh, A. N.; Raushel, F. M. The Mechanism of Glycogen-Synthetase as
Determined by Deuterium-Isotope Effects and Positional Isotope Exchange Experiments
J. Biol. Chem. 1988, 263, 10151-10154.

(69) Kim, S. G; Singh, A. N.; Raushel, F. M. Analysis of the Galactosyltransferase
Reaction by Positional Isotope Exchange and Secondary Deuterium-Isotope Effects
Archives of Biochemistry and Biophysics 1988, 26 7, 54-59.

(70) Tvaroska, 1.; Andre, 1.; Carver, J. P. Ab initio molecular orbital study of the catalytic
mechanism of glycosyltransferases: Description of reaction pathways and determination

of transition-state structures for inverting N-acetylglucosaminyltransferases .1. Am. Chem.
Soc. 2000, 122, 8762-8776.

(71) Withers, S. G.; Wakarchuk, W. W.; Strynadka, N. C. J. One step closer to a sweet
conclusion Chemistry & Biology 2002, 9, 1270-1273.

61

(72) Sinnott, M. L.; Jencks, W. P. Solvolysis of D-Glucopyranosyl Derivatives in
Mixtures of Ethanol and 2,2,2-Triﬂuoroethanol J. Am. Chem. Soc. 1980, 102, 2026-2032.

(73) Klein, H. W.; Irn, M. J.; Palm, D. Mechanism of the Phosphorylase Reaction -
Utilization of D-Gluco-Hept-l-Enitol in the Absence of Primer Eur. J. Biochem. 1986,
157,107-114.

(74) Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Sugar-mimic glycosidase
inhibitors: natural occurrence, biological activity and prospects for therapeutic
application Tetrahedron Assym. 2000, 11, 1645-1680.

(75) Inoue, S. T., T.; Niida, T. J. Antibiot. 1966, 19, 288.
(76) Inoue, S. T., T.; Ito, T.; Niida, T. Tetrahedron 1968, 24, 2125.

(77) Ezure, Y.; Maruo, S.; Miyazaki, K.; Kawamata, M. Moranoline (l-
Deoxynojirimycin) Fermentation and Its hnprovement Agricultural and Biological
Chemistry 1985, 49, 1119-1125.

(78) Fellows, L. E.; Bell, E. A.; Lynn, D. G.; Pilkiewicz, F.; Miura, 1.; Nakanishi, K.
Isolation and Structure of an Unusual Cyclic Amino Alditol from a Legume Journal of
the Chemical Society-Chemical Communications 1979, 977-978.

(79) Kite, G. C.; Fellows, L. E.; Fleet, G. W. J.; Liu, P. S.; Scoﬁeld, A. M.; Smith, N. G.
Alpha-Homonojirimycin 2,6-Dideoxy-2,6-Imino-D-Glycero-L-Gulo-Heptitol from
Omphalea-Diandra L - Isolation and Glucosidase Inhibition Tetrahedron Lett. 1988, 29,
6483-6486.

(80) Molyneux, R. J.; Pan, Y. T.; Tropea, J. E.; Elbein, A. D.; Lawyer, C. H.; Hughes, D.
J.; Fleet, G. W. J. 2-Hydroxymethyl-3,4-Dihydroxy-6-Methylpyrrolidine (6-Deoxy-
Dmdp), an Alkaloid Beta-Mannosidase Inhibitor from Seeds of Angylocalyx-Pynaertii J.
Nat. Prod. 1993, 56, 1356-1364.

(81) Junge, B. M., M.; Stoltefuss, J. In Handbook of Experimental Pharmacology New
York, 1996; Vol. 119,411-482.

(82) Fleet, G. W. J.; Karpas, A.; Dwek, R. A.; Fellows, L. E.; Tyms, A. S.; Petursson, S.;
Namgoong, S. K.; Ramsden, N. G.; Smith, P. W.; Son, J. C.; Wilson, F.; Witty, D. R.;
Jacob, G. S.; Rademacher, T. W. Inhibition of Hiv Replication by Amino-Sugar
Derivatives Febs Letters 1988, 23 7, 128-132.

(83) Karpas, A.; Fleet, G. W. J.; Dwek, R. A.; Petursson, S.; Namgoong, S. K.; Ramsden,
N. G.; Jacob, G. S.; Rademacher, T. W. Aminosugar Derivatives as Potential Anti-
Human Immunodeﬁciency Virus Agents Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 9229-
9233.

62

(84) Schweden, J.; Borgmann, C.; Legler, G.; Bause, E. Characterization of Calf Liver
Glucosidase-I and Its Inhibition by Basic Sugar Analogs Archives of Biochemistry and
Biophysics 1986, 248, 335-340.

(85) Vandenbroek, L.; Verrnaas, D. J.; Heskamp, B. M.; Vanboeckel, C. A. A.; Tan, M.;
Bolscher, J. G. M.; Ploegh, H. L.; Vankemenade, F. J.; Degoede, R. E. Y.; Miedema, F.
Chemical Modiﬁcation of Azasugars, Inhibitors of N-Glycoprotein-Processing
Glycosidases and of Hiv-I Infection - Review and Structure-Activity-Relationships
Recueil Des T ravaux Chimiques Des Pays-Bas-Journal of the Royal Netherlands
Chemical Society 1993, 112, 82-94’.

(86) Hohenschutz, L. D.; Bell, E. A.; Jewess, P. J.; Leworthy, D. P.; Pryce, R. J.; Arnold,
E.; Clardy, J. Castanospermine, a 1,6,7,8-Tetrahydroxyoctahydroindolizine Alkaloid,
from Seeds of Castanosperrnum-Australe Phytochemistry 1981, 20, 811-814.

(87) Nash, R. J.; Fellows, L. E.; Dring, J. V.; Stirton, C. H.; Carter, D.; Hegarty, M. P.;
Bell, E. A. Castanospermine in Alexa Species Phytochemistry 1988, 27, 1403-1404.

(88) Colegate, S. M.; Dorling, P. R.; Huxtable, C. R. Spectroscopic Investigation of
Swainsonine - Alpha-Mannosidase Inhibitor Isolated from Swainsona-Canescens
Australian Journal ofChemistry 1979, 32, 2257-2264.

(89) Molyneux, R. J.; James, L. F. Loco Intoxication - Indolizidine Alkaloids of Spotted
Locoweed (Astragalus-Lentiginosus) Science 1982, 216, 190-191.

(90) Saul, R.; Ghidoni, J. J.; Molyneux, R. J.; Elbein, A. D. Castanospermine Inhibits
Alpha-Glucosidase Activities and Alters Glycogen Distribution in Animals Proc. Natl.
Acad. Sci. U. S. A. 1985, 82, 93-97.

(91) Cenci di Bello, I. M., D.; Nash, R. J.; Winchester, B. In Lipid Storage Disorders;
Plenum: New York, 1988,635-641.

(92) Saunier, B.; Kilker, R. D.; Tkacz, J. S.; Quaroni, A.; Herscovics, A. Inhibition of N-
Linked Complex Oligosaccharide Formation by l-Deoxynojirimycin, an Inhibitor of
Processing Glucosidases J. Biol. Chem. 1982, 257, 4155-4161.

(93) Zeng, Y. C.; Pan, Y. T.; Asano, N.; Nash, R. J.; Elbein, A. D. Homonojirimycin and
N-methyl-homonojirimycin inhibit N-linked oligosaccharide processing Glycobiology
1997, 7, 297-304.

(94) Szumilo, T.; Kaushal, G. P.; Elbein, A. D. Puriﬁcation and Properties of
Glucosidase-I from Mung Bean Seedlings Archives of Biochemistry and Biophysics 1986,
247, 261-271.

(95) Humphries, M. J.; Matsumoto, K.; White, S. L.; Olden, K. Inhibition of
Experimental Metastasis by Castanospermine in Mice - Blockage of 2 Distinct Stages of

Tumor Colonization by Oligosaccharide Processing Inhibitors Cancer Research 1986,
46, 5215-5222.

(96) Spearman, M. A.; Ballon, B. C.; Gerrard, J. M.; Greenberg, A. H.; Wright, J. A. The
Inhibition of Platelet-Aggregation of Metastatic H-Ras-Transformed 10tl/2 Fibroblasts
with Castanospermine, an N-Linked Glycoprotein Processing Inhibitor Cancer Letters
1991, 60, 185-191.

(97) Hadwiger, A.; Niemann, H.; Kabisch, A.; Bauer, H.; Tamura, T. Appropriate
Glycosylation of the Fms Gene-Product Is a Prerequisite for Its Transforming Potency
Embo Journal 1986, 5, 689-694.

(98) Fischer, P.; Collin, M.; Karlsson, G. B.; James, W.; Butters, T. D.; Davis, S. J.;
Gordon, S.; Dwek, R. A.; Platt, F. M. The Alpha-Glucosidase Inhibitor N-
Butyldeoxynojirimycin Inhibits Human-Immunodeﬁciency—Virus Entry at the Level of
Post-Cd4-Binding Journal of Virology 1995, 69, 5791-5797.

(99) Platt, F. M.; Neises, G. R.; Dwek, R. A.; Butters, T. D. N-Butyldeoxynojirimycin Is
a Novel Inhibitor of Glycolipid Biosynthesis J. Biol. Chem. 1994, 269, 8362-8365.

(100) Pu, Y. K.; Hart, T. K.; Jonak. Z. L.; Bugelski, P. J. Physicochemical Dissociation
of Cd4-Mediated Syncytium Formation and Shedding of Human-Immunodeﬁciency-
Virus Type-l Gp120 Journal of Virology 1993, 67, 3818-3825.

(101) Sattentau, Q. J.; Zollapazner, S.; Poignard, P. Epitope Exposure on Functional,
Oligomeric Hiv-l Gp41 Molecules Virology 1995, 206, 713-717.

(102) Tulsiani, D. R. P.; Harris, T. M.; Touster, O. Swainsonine Inhibits the Biosynthesis
of Complex Glycoproteins by Inhibition of Golgi Mannosidase-Ii J. Biol. Chem. 1982,
25 7, 7936-7939.

(103) Dorling, P. R.; Huxtable, C. R.; Colegate, S. M. Inhibition of Lysosomal Alpha-
Mannosidase by Swainsonine, an Indolizidine Alkaloid Isolated from Swainsona-
Canescens Biochem. J. 1980, 191, 649-651.

(104) Harris, E. M. S.; Aleshin, A. E.; Firsov, L. M.; Honzatko, R. B. Reﬁned Structure
for the Complex of l-Deoxynojirimycin with Glucoamylase from Aspergillus-Awarnori
Var X100 to 2.4-Angstrom Resolution Biochemistry 1993, 32, 1618-1626.

(105) Weill, C. E.; Burch, R. J.; Van Dyk, J. W. Cereal Chem. 1954, 31, 150-158.

(106) Manjunath, P.; Shenoy, B. C.; Raghavendra Rao, M. R. J. Appl. Biochem 1983, 5,
235-260.

64

(107) Chambers, R. S. B., M.J.; Cannon, R.D.; Carne, A.; Emerson, G.W.; Sullivan,
P.A. An exo-B-(1,3)-glucanase of Candida albicans: puriﬁcation of the enzyme and
molecular cloning of the gene J. Gen. Microbiol. 1993, 139, 325-334.

(108) Sue M. Cutﬁeldl, G. J. D., Garib Murshudov2, Bryan F. Anderson3, Peter C. E.
Moody2, Patrick A. Sullivanl, 3 and John F. Cutﬁeld The structure of the exo- -(l,3)-
glucanase from Candida albicans in native and bound forms: relationship between a
pocket and groove in family 5 glycosyl hydrolases] J. Mol. Biol. 1999, 294, 771-783.

(109) Hempel, A.; Carnerman, N.; Mastropaolo, D.; Camerrnan, A. Glucosidase
Inhibitors - Structures of Deoxynojirimycin and Castanospermine J. Med. Chem. 1993,
36, 4082-4086.

(110) Wang, Y. F.; Takaoka, Y.; Wong, C. H. Remarkable Stereoselectivity in the
Inhibition of Alpha-Galactosidase from Coffee Bean by a New Polyhydroxypyrrolidine
Inhibitor Angew. Chem, Int. Ed. Engl. 1994, 33, 1242-1244.

(111) Liu, K. K. C.; Kajimoto, T.; Chen, L. R.; Zhong, Z. Y.; Ichikawa, Y.; Wong, C. H.
Use of Dihydroxyacetone Phosphate Dependent Aldolases in the Synthesis of
Deoxyazasugars J. Org. Chem. 1991, 56, 6280-6289.

(112) Legler, G.; Pohl, S. Synthesis of 5-Amino-5-Deoxy-D-Galactopyranose and 1,5-
Dideoxy-1,5-Imino-D-Ga1actitol, and Their Inhibition of Alpha-and Beta-D-
Galactosidases Carbohydr. Res. 1986, 155, 119-129.

(113) Frankowski, A.; Seliga, C.; Bur, D.; Streith, J. Irnidazole-Analogs of 6-
Epicastanosperrnine and of 3,7a-Diepia1exine Helv. Chim. Acta 1991, 74, 934-940.

(114) Frankowski, A.; Deredas, D.; Lenouen, D.; Tschamber, T.; Streith, J. On the Way
to Glycoprocessing Inhibitors - Synthesis of an Imidazo-L-Xylo-Piperidinose Derivative
Helv. Chim. Acta 1995, 78, 1837-1842.

(115) Frankowski, A.; Deredas, D.: Streith, J.; Tschamber, T. Synthesis of an imidazolo-
L-lyxo-piperidinose derivative Tetrahedron 1998, 54, 9033-9042.

(116) Streith, J.; Boiron, A.; Frankowski, A.; Lenouen, D.; Rudyk, H.; Tschamber, T. On
the Way to Glycoprocessing Inhibitors - a General One-Pot Synthesis of Imidazolosugars
Synthesis-Stuttgart 1995, 944-946.

(117) Tschamber, T.; Siendt, H.; Boiron, A.; Gessier, F .; Deredas, D.; Frankowski, A.;
Picasso, S.; Steiner, H.; Aubertin, A. M.; Streith, J. Stereoisomeric imidazolo-pentoses -
Synthesis, chiroptical properties, and evaluation as glycosidase inhibitors European J.
Org. Chem. 2001, 1335-1347.

65

(118) Granier, T.; Gaiser, F.; Hinterrnann, L.; Vasella, A. Synthesis via a carbohydrate-
derived Munchnone of pyrrolopyridines (indolizines) and imidazopyridines, and their
evaluation as inhibitors of B-D-Glucosidases Helv. Chim. Acta 1997, 80, 1443-1456.

(119) Lillelund, V. H.; Jensen, H. H.; Liang, X. F.; Bols, M. Recent developments of
transition-state analogue glycosidase inhibitors of non-natural product origin Chem. Rev.
2002, 102, 515-553.

(120) Panday, N.; Granier, T.; Vasella, A. Synthesis and evaluation of indolizine-type
inhibitors of N-acetyl-beta-D-glucosaminidases Helv. Chim. Acta 1998, 81, 475-490.

(121) Krulle, T.; delaFuente, C.; Pickering, L.; Aplin, R. T.; Tsitsanou, K. E.; Zographos,
S. E.; Oikonomakos, N. G.; Nash, R. J.; Griffiths, R. C.; Fleet, G. W. J. Triazole
carboxylic acids as anionic sugar mimics? Inhibition of glycogen phosphorylase by a D-
glucotriazole carboxylate Tetrahedron Assym. 1997, 8, 3807-3820.

(122) Tatsuta, K.; Miura, S.; Ohta, S.; Gunji, H. Total Syntheses of De-Branched
Nagstatin and Its Analogs Having Glycosidase Inhibiting Activities Tetrahedron Lett.
1995, 36, 1085-1088.

(123) Vonhoff, S.; Heightman, T. D.; Vasella, A. Inhibition of glycosidases by lactam
oximes: Inﬂuence of the aglycon in disaccharide analogues Helv. Chim. Acta 1998, 81 ,
1710-1725.

(124) Streith, J.; Rudyk, H.; Tschamber, T.; Tarnus, C.; Strehler, C.; Deredas, D.;
Frankowski, A. The synthesis of imidazol sugars which mimic cyclic carboxonium ions
formed during the glycosidase-catalysed hydrolysis of oligo- and polysaccharides
European J. Org. Chem. 1999, 893-898.

(125) Yuasa, H.; Takada, J.; Hashimoto, H. Synthesis of salacinol Tetrahedron Lett.
2000, 41, 6615-6618.

(126) Ghavarni, A.; Johnston, B. D.; Pinto, B. M. A new class of glycosidase inhibitor:
Synthesis of salacinol and its stereoisomers J. Org. Chem. 2001, 66, 2312-2317.

(127) Yoshikawa, M.; Murakami, T.; Yashiro, K.; Matsuda, H. Kotalanol, a potent alpha-
glucosidase inhibitor with thiosugar sulfonium sulfate structure, from antidiabetic

ayurvedic medicine Salacia reticulata Chemical & Pharmaceutical Bulletin 1998, 46,
1339-1340.

(128) Yoshikawa M, M. T., Shimada H, Matsuda H, Yarnahara J, Tanabe G, Muraoka O
Salacinol, potent antidiabetic principle with unique thiosugar sulfonium sulfate structure

from the ayurvedic traditional medicine Salacia reticulata in Sri Lanka and India
Tetrahedron Lett. 1997, 38, 8367-8370.

66

(129) Jespersen, T. M.; Dong, W. L.; Sierks, M. R.; Skrydstrup, T.; Lundt, 1.; Bols, M.
Isofagomine, a Potent, New Glycosidase Inhibitor Angew. Chem, Int. Ed. Engl. in
English 1994, 33, 1778-1779.

(130) Jespersen, T. M.; Bols, M.; Sierks, M. R.; Skrydstrup, T. Synthesis of Isofagomine,
a Novel Glycosidase Inhibitor Tetrahedron 1994, 50, 13449-13460.

(131) Paulsen, H. Liebigs Ann. Chim. 1965, 683, 187-198.

(132) Stutz, A. E. Iminosugars As Glycosidase Inhibitors: Nojirimycin and Beyond;
Wiley-VCH: Weinheim, 1999.

(133) Ichikawa, Y.; Igarashi, Y.; Ichikawa, M.; Suhara, Y. I-N-iminosugars: Potent and
selective inhibitors of beta-glycosidases J. Am. Chem. Soc. 1998, 120, 3007-3018.

(134) Bols, M. l-aza sugars, apparent transition state analogues of equatorial glycoside
formation/cleavage Acc. Chem. Res. 1998, 31, 1-8.

(135) Dong, W. L.; Jespersen, T.; Bols, M.; Skrydstrup, T.; Sierks, M. R. Evaluation of
isofagomine and its derivatives as potent glycosidase inhibitors Biochemistry 1996, 35,
2788-2795.

(136) Heightman, T. D.; Vasella, A. T. Recent insights into inhibition, structure, and
mechanism of conﬁguration-retaining glycosidases Angew. Chem, Int. Ed. Engl. 1999,
38, 750-770.

(137) Bulow, A.; Plesner, I. W.; Bols, M. A large difference in the thermodynamics of
binding of isofagomine and l-deosynojirimycin to beta-glucosidase J. Am. Chem. Soc.
2000, 122, 8567-8568.

(138) Liu, H. Z.; Liang, X. F .; Sohoel, H.; Bulow, A.; Bols, M. Noeuromycin, a glycosyl
cation mimic that strongly inhibits glycosidases J. Am. Chem. Soc. 2001, 123, 5116-5117.

(139) Williams, S. J .; Hoos, R.; Withers, S. G. Nanomolar versus millimolar inhibition by
xylobiose-derived azasugars: Signiﬁcant differences between two structurally distinct
xylanases J. Am. Chem. Soc. 2000, 122, 2223-2235.

(140) Varrot, A.; Tarling, C. A.; Macdonald, J. M.; Stick, R. V.; Zechel, D. L.; Withers,
S. G.; Davies, G. J. Direct observation of the protonation state of an imino sugar
glycosidase inhibitor upon binding J. Am. Chem Soc. 2003, 125, 7496-7497.

(141) Truscheit, E.; Frommer, W.; Junge, B.; Muller, L.; Schmidt, D. D.; Wingender, W.

Chemistry and Biochemistry of Microbial Alpha-Glucosidase Inhibitors Angew. Chem,
Int. Ed. Engl. in English 1981, 20, 744-761.

67

(142) Legler, G.; Herrchen, M. N-Substituted D-Galactosylamines as Probes for the
Active-Site of Beta-D-Galactosidase from Escherichia-Coli Carbohydr. Res. 1983, 116,
95-103. ‘

(143) Legler, G.; Stutz, A. E.; Immich, H. Synthesis of 1,5-Dideoxy-1,5-Imino-D-
Arabinitol (S-nor-L-Fuco-l-Deoxynojirimycin) and Its Application for the Afﬁnity
Puriﬁcation and Characterization of Alpha-L-Fucosidase Carbohydr. Res. 1995, 272, 17-
30.

(144) Guo, W. F .; Hiratake, J.; Ogawa, K.; Yarnamoto, M.; Ma, S. J.; Sakata, K. beta-D-
glycosylamidines: Potent, selective, and easily accessible beta-glycosidase inhibitors
Bioorg. & Med. Chem. Lett. 2001, l 1, 467-470.

(145) Maity, S. K.; Dutta, S. K.; Banerjee, A. K.; Achari, B.; Singh, M. Design and
Synthesis of Mannose Analogs as Inhibitors of Alpha-Mannosidase Tetrahedron 1994,
50, 6965-6974.

(146) Lai, W.; Martin, O. R. Synthesis and Inhibition Studies of C-(D-
Glycopyranosyl)Methy1amines Carbohydr. Res. 1993, 250, 185-193.

(147) Bemiller, J. N.; Gilson, R. J.; Myers, R. W.; Santoro, M. M.; Yadav, M. P. N-
Substituted (Beta-D-Ga1actopyranosylmethyl) Amines as Reversible Inhibitors of Beta-
D-Galactosidase Carbohydr. Res. 1993, 250, 93-100.

(148) Bemiller, J. N.; Yadav, M. P.; Kalabokis, V. N.; Myers, R. W. N-Substituted (Beta-
D-Galactopyranosylmethyl)Amines, and C-Beta-D-Galactopyranosylforrnamides, and
Related-Compounds Carbohydr. Res. 1990, 200, Ill-126.

(149) Dietrich, H.; Schmidt, R. R. Aminosubstituted Alpha-D-Glucosylmethylbenzenes
(Benzyl Alpha-C-Glucosides) and an N-(C-Alpha-D-Glucosylmethy)Aniline
(Anilinomethyl Alpha-C-Glucoside) - Novel Alpha-D-Glucosidase Inhibitors Carbohydr.
Res. 1993, 250, 161-176.

(150) Schmidt, R. R.; Dietrich, H. Amino-Substituted Beta-Benzyl-C-Glycosides - Novel
Beta-Glycosidase Inhibitors Angew. Chem, Int. Ed. Engl. in English 1991, 30, 1328-
1329.

(151) Berecibar, A.; Grandjean, C.; Siriwardena, A. Synthesis and biological activity of
natural aminocyclopentitol glycosidase inhibitors: Mannostatins, trehazolin, allosamidins,
and their analogues Chem. Rev. 1999, 99, 779-844.

(152) Tong, M. K.; Papandreou, G.; Ganem, B. Potent, Broad-Spectrum Inhibition of

Glycosidases by an Amidine Derivative of D-Glucose J. Am. Chem. Soc. 1990, 112,
6137-6139.

68

(153) Ganem, B. Inhibitors of carbohydrate-processing enzymes: Design and synthesis of
sugar-shaped heterocycles Acc. Chem. Res. 1996, 29, 340-347.

(154) Ganem, B.; Papandreou, G. Mimicking the Glucosidase Transition-State - Shape
Charge Considerations .1. Am. Chem. Soc. 1991, 113, 8984-8985.

(155) Pan, Y. T.; Kaushal, G. P.; Papandreou, G.; Ganem, B.; Elbein, A. D. D-
Mannonolactam Amidrazone - a New Mannosidase Inhibitor That Also Inhibits the
Endoplasmic-Reticulum or Cytoplasmic Alpha-Mannosidase J. Biol. Chem. 1992, 267,
8313-8318.

(156) Ermert, P.; Vasella, A. Synthesis of a Glucose-Derived Tetrazole as a New Beta-
Glucosidase Inhibitor - a New Synthesis of l-Deoxynojirimycin Helv. Chim. Acta 1991,
74, 2043-2053.

(157) Aoyagi, T.; Suda, H.; Uotani, K.; Kojima, F.; Aoyama, T.; Horiguchi, K.; Hamada,
M.; Takeuchi, T. Nagstatin, a New Inhibitor of N-Acety1-Beta-D-Glucosaminidase,
Produced by Streptomyces-Amakusaensis Mg846-Ff3 - Taxonomy, Production, Isolation,
Physicochemical Properties and Biological-Activities Journal of Antibiotics 1992, 45,
1404-1408. .

(158) Aoyama, T.; Naganawa, H.; Suda, H.; Uotani, K.; Aoyagi, T.; Takeuchi, T. The
Structure of Nagstatin, a New Inhibitor of N-Acetyl-Beta-D-Glucosaminidase Journal of
Antibiotics 1992, 45, 1557-1558.

(159) Panday, N.; Vasella, A. Synthesis of glucose- and mannose-derived N-acetylarnino
imidazopyridines and their evaluation as inhibitors of glycosidases Synthesis-Stuttgart
1999, 1459-1468.

(160) Tatsuta, K. Total synthesis and chemical design of useful glycosidase inhibitors
Pure and Applied Chemistry 1996, 68, 1341-1346.

(161) Tatsuta, K.; Miura, S. Total Synthesis of Nagstatin, an N-Acetyl-Beta-D-
Glucosaminidase Inhibitor Tetrahedron Lett. 1995, 36, 6721-6724.

(162) Tatsuta, K.; Miura, S.; Gunji, H.: The total synthesis of a glycosidase inhibitor,
nagstatin Bulletin of the Chemical Society of Japan 1997, 70, 427-436.

(163) Tatsuta, K.; Miura, S.; Ohta, S.; Gunji, H. Syntheses and Glycosidase Inhibiting
Activities of Nagstatin Analogs Journal of Antibiotics 1995, 48, 286-288.

(164) Tatsuta, K.; Ikeda, Y.; Miura, S. Synthesis and glycosidase inhibitory activities of
nagstatin triazole analogs Journal of Antibiotics 1996, 49, 836-838.

(165) Panday, N.; Canac, Y.; Vasella, A. Very strong inhibition of glucosidases by C(2)-
substituted tetrahydroimidazopyridines Helv. Chim. Acta 2000, 83, 58-79.

69

(166) Nishimura, Y.; Adachi, H.; Satoh, T.; Shitara, E.; Nakamura, H.; Kojima, F.;
Takeuchi, T. All eight stereoisomeric D-glyconic-delta-lactams: Synthesis,
conformational analysis, and evaluation as glycosidase inhibitors J. Org. Chem. 2000, 65,
4871-4882.

(167) Panday, N.; Meyyappan, M.; Vasella, A. A comparison of glucose- and
glucosarnine-related inhibitors: Probing the interaction of the 2-hydroxy group with
retaining beta-glucosidases Helv. Chim. Acta 2000, 83, 513-538.

(168) Williams, S. J.; Notenboom, V.; Wicki, J.; Rose, D. R.; Withers, S. G. A new,
simple, high-afﬁnity glycosidase inhibitor: Analysis of binding through X-ray
crystallography, mutagenesis, and kinetic analysis Journal of the American Chemical
Society 2000, 122, 4229-4230.

(169) Sohoel, H.; Liang, X. F .; Bols, M. Isogalactofagomine lactam. A neutral nanomolar
galactosidase inhibitorJ. Chem. Soc., Perkin Trans. 1 2001, 1584-1585.

(170) Darrow, J. W.; Drueckhammer, D. G. A cyclic phosphonamidate analogue of
glucose as a selective inhibitor of inverting glycosidases Bioorganic & Medicinal
Chemistry 1996, 4, 1341-1348.

(171) LeMerrer, Y.; Fuzier, M.; Dosbaa, 1.; Foglietti, M. J.; Depezay, J. C. Synthesis of
thiosugars as weak inhibitors of glycosidases Tetrahedron 1997, 53, 16731-16746.

(172) Turner, G. A.; Goodarzi, M. T.; Thompson, S. Glycosylation of alpha-l-proteinase
inhibitor and haptoglobin in ovarian cancer: evidence for two different mechanisms
Glycoconjugate Journal 1995, 12, 211-218.

( 173) Couldrey, C.; Green, J. E. Metastases: the glycan connection Breast Cancer
Research 2000, 2, 321-323.

(174) Schumacher, U.; Mohamed, M.; Mitchell, B. S. Differential expression of
carbohydrate residues in human breast and colon cancer cell lines grown in vitro and in
vivo in SCID mice Cancer Journal 1996, 9, 247-254.

(175) Touitou, I.; Garcia, M.; Westley, B.; Capony, F.; Rochefort, H. Effect of
tunicamycin and endoglycosidase H and F on the estrogen regulated 52000-Mr protein
secreted by breast cancer cells Biochimie 1985, 67, 1257-1266.

(176) Rostenberg, 1.; Guizar-Vazquez, J.; Suarez, P.; Rico, R.; Nungaray, L.; Dominguez,
C. Distinct glycosylation of serum proteins in patients with cancer Journal of the
National Cancer Institute 1978, 60, 83-87.

(177) McGuckin, M. A.; Devine, P. L.; Ward, B. G. Early steps in the biosynthesis of

MUCZ epithelial mucin in colon cancer cells Biochemistry and Cell Biology 1996, 74,
87-93.

70

(178) Schumacher, U.; Adam, B; Flavell, D. J.; Boehm, D.; Brooks, S. A.; Leathem, A.
J. Glycosylation patterns of the human colon cancer cell line HT-29 detected by Helix
pomatia agglutinin and other lectins in culture, in primary tumors and in metastases in
SCID mice Clinical & Experimental Metastasis 1994, 12, 398-404.

(179) Hakomori, S. Glycosylation deﬁning cancer malignancy: New wine in an old bottle
Proceedings of the National Academy of Sciences of the United States of America 2002,
99, 10231-10233.

(180) Scanlin, T. F.; Glick, M. C. Terminal glycosylation and disease: inﬂuence on
cancer and cystic ﬁbrosis Glycoconjugate Journal 2001, l 7, 617-626.

(181) Dennis, J. W.; Granovsky, M. Protein glycosylation and cancer Carbohydrates in
Chemistry and Biology 2000, 4, 923-943.

(182) Catterall, J. B.; Jones, L. M. H.; Turner, G. A. Membrane protein glycosylation and
CD44 content in the adhesion of human ovarian cancer cells to hyaluronan Clinical &
Experimental Metastasis 2000, 1 7, 583-591.

(183) Dennis, J. W. G., M.; Warren, C. E. Glycoprotein glycosylation and cancer
progression Biochimica et Biophysica Acta 1999, 14 73, 21-34.

(184) Kim, Y. J .; Varki, A. Perspectives on the signiﬁcance of altered glycosylation of
glycoproteins in cancer Glycoconjugate Journal 1997, 14, 569-576.

(185) Dall'Olio, F. Protein glycosylation in cancer biology: an overview Clinical
Molecular Pathology 1996, 49, M126-M135.

(186) Taylor-Papadimitriou, J.; Epenetos, A. A. Exploiting altered glycosylation patterns
in cancer: progress and challenges in diagnosis and therapy Trends in Biotechnology
1994, 12, 227-233.

(187) Hakomori, S. Aberrant glycosylation in cancer cell membranes as focused on
glycolipids: overview and perspectives Cancer Research 1985, 45, 2405-2414.

(188) Narayanan, S. Sialic acid as a tumor marker Annals of Clinical and Laboratory
Science 1994, 24, 376-384.

(189) Korytnyk, W.; Angelino, N.; Klohs, W.; Bemacki, R. Cmp and Cmp-Sugar
Analogs as Inhibitors of Sialic-Acid Incorporation into Glycoconjugates European J.
Med. Chem. 1980, 15, 77-84.

(190) Camarasa, M. J.; Femandezresa, P.; Garcialopez, M. T.; Delasheras, F. G.;
Mendezcastrillon, P. P.; Alarcon, B.; Carrasco, L. Uridine 5'-Diphosphate Glucose
Analogs - Inhibitors of Protein Glycosylation That Show Antiviral Activity J. Med.
Chem. 1985, 28, 40-46.

71

(191) Vagheﬁ, M. M.; Bemacki, R. J.; Dalley, N. K.; Wilson, B. E.; Robins, R. K.
Synthesis of Glycopyranosylphosphonate Analogs of Certain Natural Nucleoside
Diphosphate Sugars as Potential Inhibitors of Glycosyltransferases J. Med. Chem. 1987,
30, 1383-1391.

(192) Cai, S.; Stroud, M. R.; Hakomori, S.; Toyokuni, T. Synthesis of Carbocyclic
Analogs of Guanosine 5'-(Beta-L-Fucopyranosyl Diphosphate) (de-Fucose) as Potential
Inhibitors of Fucosyltransferases J. Org. Chem. 1992, 5 7, 6693-6696.

(193) Endo, T.; Kajihara, Y.; Kodama, H.; Hashimoto, H. Novel aspects of interaction
between UDP-Gal and GlcNAc beta-1,4-galactosyltransferase: Transferability and
remarkable inhibitory activity of UDP-(mono-O-methylated gal), UDP-Fuc and UDP-
Man Bioorganic & Medicinal Chemistry 1996, 4, 1939-1948.

(194) Wang, R.; Steensma, D. H.; Takaoka, Y.; Yun, J. W.; Kajimoto, T.; Wong, C. H. A
search for pyrophosphate mimics for the development of substrates and inhibitors of
glycosyltransferases Bioorganic & Medicinal Chemistry 1997, 5, 661-672.

(195) Muller, B.; Schaub, C.; Schmidt, R. R. Efﬁcient sialyltransferase inhibitors based
on transition-state analogues of the sialyl donor Angew. Chem, Int. Ed. Engl. 1998, 3 7,
2893-2897.

(196) Hindsgaul, O.; Kaur, K. J.; Srivastava, G.; Blaszczykthurin, M.; Crawley, S. C.;
Heerze, L. D.; Palcic, M. M. Evaluation of Deoxygenated Oligosaccharide Acceptor
Analogs as Specific Inhibitors of Glycosyltransferases J. Biol. Chem. 1991, 266, 17858-
17862.

(197) Lowary, T. L.; Hindsgaul, 0. Recognition of Synthetic Deoxy and Deoxyﬂuoro
Analogs of the Acceptor Alpha-L-Fucp-(l- 2)-Beta-D-Galp-or by the Blood-Group-a and
Blood-Group-B Gene-Speciﬁed Glycosyltransferases Carbohydr. Res. 1993, 249, 163-
195.

(198) Kajihara, Y.; Kodama, H.; Wakabayashi, T.; Sato, K.; Hashimoto, H.
Characterization of Inhibitory Activities and Binding Mode of Synthetic 6‘-Modiﬁed
Methyl N-Acetyl-Beta-Lactosarninide toward Rat-Liver Cmp-Deuteriurn-
Neu5acdeuterium-Galactoside-(2- 6)-Alpha-Deuterium-Sialyltransferase Carbohydr. Res.
1993, 247, 179-193.

(199) Bause, E. Active-Site-Directed Inhibition of Asparagine N-Glycosyltransferases
with Epoxy-Peptide Derivatives Biochem. J. 1983, 209, 323-330.

(200) Hendrickson, T. L.; Spencer, J. R.; Kato, M.; Imperiali, B. Design and evaluation
of potent inhibitors of asparagine-linked protein glycosylation J. Am. Chem. Soc. 1996,
118, 7636-7637.

(201) Platt, F. M.; Neises, G. R.; Reinkensmeier, G.; Townsend, M. J.; Perry, V. H.;
Proia, R. L.; Winchester, B.; Dwek, R. A.; Butters, T. D. Prevention of lysosomal storage
in Tay-Sachs mice treated with N-butyldeoxynojirimycin Science 1997, 276, 428-431.

(202) Inokuchi, J. I.; Radin, N. S. Preparation of the Active Isomer of 1-Pheny1-2- -
Decanoylamino-3-Morpholino-1-Propanol, Inhibitor of Murine Glucocerebroside
Synthetase Journal of Lipid Research 1987, 28, 565-571.

(203) Palcic, M. M.; Heerze, L. D.; Srivastava, O. P.; Hindsgaul, O. A Bisubstrate
Analog Inhibitor for Alpha-(l- 2)-Fucosyltransferase J. Biol. Chem. 1989, 264, 17174-
17181.

(204) Hashimoto, H.; Endo, T.; Kajihara, Y. Synthesis of the ﬁrst tricomponent
bisubstrate analogue that exhibits potent inhibition against GlcNAc:beta-l,4-
galactosyltransferaseJ. Org. Chem. 1997, 62, 1914-1915.

(205) Kim, Y. J.; Ichikawa, M.; Ichikawa, Y. A rationally designed inhibitor of alpha-
1,3-galactosyltransferaseJ. Am. Chem. Soc. 1999, 121, 5829-5830.

(206) Schaub, C.; Muller, B.; Schmidt, R. R. Sialyltransferase inhibitors based on CMP-
quinic acid European J. Org. Chem. 2000, 1745-1758.

(207) Natsuka, S.; Lowe, J. B. Enzymes Involved in Mammalian Oligosaccharide
Biosynthesis Current Opinion in Structural Biology 1994, 4, 683-691.

(208) Holmes, E. H.; Ostrander, G. K.; Hakomori, S. Biosynthesis of the SialyI-Lex
Determinant Carried by Type-2 Chain Glycosphingolipids (Iv3neuacii13fucnlc4,
Vi3neuacv3fucnlc6, and Vi3neuaciii3v3fuc2nlc6) in Human-Lung Carcinoma Pc9 Cells
J. Biol. Chem. 1986, 261, 3737-3743.

(209) Hanisch, F. G.; Mitsakos, A.; Schroten, H.; Uhlenbruck, G. Biosynthesis of
Cancer-Associated Sialyl-X Antigen by a (l- 3)-Alpha-L-Fucosy1transferase of Human
Amniotic-Fluid Carbohydr. Res. 1988, 178, 23-28.

(210) Hakomori, S.; Nudelman, E.; Levery, S. B.; Kannagi, R. Novel Fucolipids
Accumulating in Human Adenocarcinoma .2. Glycolipids with Difucosylated or
Trifucosylated Type-2 Chain J. Biol. Chem. 1984, 259, 4672-4680.

(211) Feizi, T. Demonstration by Monoclonal-Antibodies That Carbohydrate Structures

of Glycoproteins and Glycolipids Are Onco-Developmental Antigens Nature 1985, 314,
53-57.

73

Chapter 2

On the Design, Properties and Applications of Purine
Nucleoside Phosphorylase, Nucleoside Hydrolase and
Phosphoribosyltransferase Inhibitors

74

ABSTRACT

Purine nucleoside phosphorylase, nucleoside hydrolase and phosphoribosyltransferase are
important enzymes that are involved in the salvage and regulation of purine and
pyrimidine pools. They are all basically N-glycanases like glycosidases and glycosyl
transferases. Kinetic isotope effect studies support the conclusion that oxocarbenium ions
are features of the transition states for all these enzymes. A large amount of work has
been devoted to development of effective inhibitors against these enzymes. The potential
use of such inhibitors includes the development of antiviral, cancer and autoimmune
drugs as well as drugs against certain neurological disorders. In this chapter, the
structure-based design of inhibitors and transition state-based analogs of these three

enzyme classes and related enzymes are reviewed

75

2.1. lndroduction

Purine nucleoside phosphorylases (PNP) are the enzymes that catalyze the cleavage
(phosphorolysis) of the glycosidic bond (C-l’ to N-9) of ribo- and 2’-deoxyribo-
nucleoside to form the purine base and a-D-(deoxy)riboside-1-phosphatel. The main
substrates of PNP are inosine (figure 2.1), guanosine and 2’-deoxyguanosine. PNP is an
important enzyme in the base salvage pathway. It has been discovered that lack of PNP
can result in human T-cell immunodeﬁciency”. Because PNP is the only enzyme that
degrade 2-deoxyguanosine, if it is inhibited, 2-deoxyguanosine can be transported and
phosphorylated by deoxycytidine kinase to deoxyguanosine triphosphate (dGTP)“. The
dGTP can accumulate in the blood because of the high level of deoxycytidine kinase, and
low activity of nucleotidases, that degrade dGT P. High levels of dGTP in T-cells can
cause inhibition [of ribonucleotide reductase, which is responsible for the cellular

formation of dCDP and dUDP. This prevents DNA synthesis for T-cell proliferations'lo.

HQ -, N. O HO.
\/o"ﬁ".~;~/ PNP \; 0‘ ,
-1” ‘ N (,NH > ,’ 'OP032' + Hypoxanthine
HO ,I HO
0. .0
[P.f
-O O‘
lnosi ne Ribose-1 phosphate

Figure 2.1. PNP catalyzed reaction

76

Undesirable T-cell activation and proliferation result in psoriasis, rheumatoid arthritis, T-
cell leukemias and lymphomas, transplant tissue rejection, and other type IV autoimmune
disorders. Consequently PNP inhibition has become a very important target for the
development of drugs for a variety of diseases including cancer and autoimmune
diseasesI "'4.
Nucleoside N-riboside hydrolases (nucleoside hydrolases) are enzymes that catalyze the
hydrolysis of the N-ribosidic bond of purine and purimidine nucleoside. Figure 2.2
shows the hydrolysis of inosine catalyzed by nucleoside hydrolase. Nucleoside
hydrolases are found in protozoan parasites and bacteria (rihA, rihB and rihC genes in E.
coli)”, but not in mammalian cells”. Nucleoside hydrolases play important roles in
protozoa because they provide free bases to DNA and RNA synthesis for these
organisms”. The protozoa have no ability for de novo purine biosyntehsis, therefore, the
required bases are completely obtained from the host by a base salvage pathway.
Nucleoside hydrolases are essential for the base recovery, and they have been found in
the purine salvage pathway of the trypanosomes including Crithidiafasciculate'g"9. Their
speciﬁc role in the protozoa made them an important target for antibiotic design. The role
of nucleoside hydrolases in bacteria is not well understood but they could be important

during infection because the organisms could use the host nucleosides to make its own

genetic material.

77

Ho ,.
\.o\_
\ /\
HO oH f:

H20

lnosine

r N ,
(A

‘H/I
l

N
‘3 ‘.,O
N, NH

HO
Nucleoside
0 ‘OH
hydrolase b \1 f + Hypoxanthine
HO OH
Ribose

Figure 2.2. Nucleoside N-riboside hydrolases catalyzed reaction

Hypoxanthine-guanine phosphoribosyltransferases (HGPRTases) catalyze the transfer of
5’-phosphoribose from 5’-phosphoribosylpyrophosphate to the purine base hypoxanthine
to form a nucleoside 5’-phosphate (inosine monophosphate) and pyrophosphate (ﬁgure
2.3). HGPRTases are manganese dependent enzymes and they are critical in both the
base de novo and salvage pathway to obtain hypoxanthine, guanosin, or xanthine for the
synthesis of DNA and RNA in protozoa parasites. Human HGPRTase does not use

xanthine as substrate. Deﬁciency of human HGPRTase results in Lesch-Nyhan

syndrome. Hyperuricemia and neurological disorders are features of this disease”.

Because of their key roles in the nucleotide metabolism, they are also targeted as a drug

strategy in antimicrobial, antiparasite and anticancer therapeutics.

78

Hypoxanthine ; ' N.“ O

HN- «:>
NH 0. ,O'
.. PTo : ”"2". 0 N 9"~--
0 1 ,O‘ ,0 , HGPRTase N '5’ 1
\Pﬁﬂ p’0 p0 ‘ ' r 1. / 1 NH
A. “")/‘ \‘ H "l g ', N-~:,-’
~, o 0 HO OH
HO OH
Phosphonbosylpyrophosphate 1nosine-5'-phosphate (1MP)

Figure 2.3. Hypoxanthine-guanine phosphoribosyltransferases catalyzed reaction

The three kinds of enzymes are all basically N-glycanases like glycosidases and
glycosyltransferases, and they are involved in the salvage and regulation of purine and
pyrimidine pools'z. They also are involved in base excision repair and in the inactivation

2'22. Therefore, a large amount of work has been devoted to studies of

of ribosomes
enzyme mechanisms, transition states to seek powerful inhibitors for those enzymes.
Vern. L. Schramm etc. have developed methods for the determination of the transition
state structures by kinetic isotope effect studies at multiple positions in the substrate

molecule”26

. The methods include measuring kinetic isotope effects at every position in
a substrate molecule that might be perturbed at the transition state, computing a truncated
transition state structure with bond lengths and angles matching the isotope effects, and
then optimizing the structure for the transition state. The results showed oxocarbenium
ions are features for all these enzymes. The difference exists in the degree of the
oxocarbenium ion formation and the nucleophilic displacement”. PNP and HGPRTase
catalyze reactions with inversion of conﬁguration at the anomeric center. According to

28-31

Schramm , the transition state of PNP has bond order 0.4 to the leaving group, but

79

bond order < 0.04 to the incoming nucleophile. In nucleoside hydrolasel6’32‘33, the
oxocarbenium ion is nearly fully developed, with C—l ’ rehybridized almost completely to
sp2, and the attack of the water nucleophile is late well behind N-ribosidic bond breaking
with approximately 0.03 bond order from the attacking oxygen to Cl’. An SNI type
reaction is more likely. Schramm proposes that for all these enzymes, the nucleophilic
displacement occurs relatively late in the reaction coordinate. As we will see later, this
conﬂicts with the ﬁnding that placing a nitrogen atom at the anomeric position in
transition state inhibitors leads to better binding and better inhibition. According to the
Schramm’s model: the low bond order to the anionic nucleophile with signiﬁcant bond
order to the leaving group causes the ribosyl group to be electron deﬁcient, with partial
oxocarbenium character; the N-7 of the purine ring in these transition states is protonated
or hydrogen-bonded to assist the departure of the leaving group; The pKa of this group is
elevated toward 7 or above as the ribosidic bond is broken. Figure 2.4 illustrates the
substrate inosine, PNP transition state, nucleoside hydrolase and HGPRTase transition
states according to Schramm. The PNP transition state is reached when the Cl ’-N9 bond
distance is elongated to 1.77A from 1.44A in inosine, with 3.0A from the anomeric
carbon, and the phosphate anion is proposed to play an important electrostatic role in the

transition state stabilization.

80

1 l
‘ ,0. N. /
5' 4' )‘I. TRY/7 1
3' J N _, gNH

lnosine

ribosyl ring
packer 03' earn

 

PNP transition state

V. leaving group activated
nearly full oxocarbenium H i by "7 Habitation
ion charge and 900m WY 1,111. 3'

\ \‘
\ ‘N9

s
u
1

weak bonds to LG
and nucleophile
i

/

,. "/‘

 

Hi0
\\
specific enzyme 1
activated water

N ucleoside hydrolases trasltion state

0

H 1

N r.

f, N"

N ,. \\ N

oaPo”\,. 0?
Ho OH 01,100 ":2

0‘. \P/o
Mg - ‘O

HGPRTase transition state

Figure 2.4. Substrate inosine, PN P transition state, nucleoside hydrolase and HGPRTase

transition states.

81

2.2. Structure-based Design of PNP inhibitors

Structure-based design and structure-activity relationships have been used for the PNP
inhibitor design34'39. It is based on the three-dimensional structure of the native PNP and
computer-assisted molecular modeling. The catalytic site of PNP containing substrates or
product analogs or weak inhibitors was sequentially ﬁlled with newly designed analogs
followed by subsequent structural and kinetic characterization and reﬁnement of catalytic
site contacts. A number of 9-substituted deazaguanine derivatives were designed to
search for the potent inhibitors with favorable interaction with PNP active site including
purine binding site, the hydrophobic pocket and the phosphate binding site. The N-9 was
replaced with carbon because no interaction was identiﬁed between N-9 and PNP
residue, and 9-deazainosine itself binds 15-20 times more tightly than inosine“). A
linkage to carbon instead of nitrogen is desirable because such a linkage cannot be
cleaved by N-glycosidases of any sort. Compounds with substitution at the 9-position by
aromatic, alicyclic and heteroalicyclic groups attached through a methylene group were
found to be good inhibitors34 (Figure 2.5) even if the substituents bore no similarity to a
sugar residue and an equivalent of a positively charged glycosyl ring oxygen or anomeric
carbon was not present. Aryl groups such as thienyl. furanyl, pyridyl and phenyl provided
potent inhibitors. The 3- and 4—substituted phenyl group gave low leo values in the
nanomolar range, but compounds with 2-chloro and 2-hydroxylphenyl groups are poor
inbitors. This is probably due to the displacement of the 9-deazaguanine moiety from its
optimal binding position resulting in disturbing of the noncovalent interactions.
Cyclopentyl, cyclohexyl, tetrahydrothienyl, dithianyl derivatives which have greater

hydrophobicity also showed equal potency when assayed in 1 mM phosphate. They all

82

have favorable interaction with the purine and hydrophobic binding sites of PNP, and

they can penetrate the cells more easily.

9 o
H H
HN" . > N HN T N»
H2N./*‘:::N " H2" ,7 N /« l
4
s
7
s
O O
.i H .g H .
HN” \3~N_ HN' f“"3 HN‘
H2N N HZN’ N HzN”
1C” (11M)
(1 mM P0.) = 29 25
9 10

leo (I'M)

R 5 in 1 mM 1’04,
3,4-dichlorophenyl 17
3—chlorophenyl 20
4-iodophenyl 23
3-ﬂuorophenyl 24
4-chlorophenyl 25
phenyl Sl
3-methylphenyl S7
3-methoxylphenyl 82
O O

H i H
\ ~N HN/ \ 7 r N)
. ”I L3; /
N . HzN
,. S. ,, S
11 \, ‘ ' s
1 l 1 8
11 12

Figure 2.5. PNP inhibitors

To increase the binding afﬁnity of the inhibitors, functional groups which can interact

with the PNP phosphate binding site can be introduced. A potent PNP inhibitor, acyclovir

diphosphate“ (ACVdir) (Figure 2.6) was found to be a bisubstrate inhibitor binding in

83

the purine binding site and in the phosphate binding site with some interaction of the
ACV side chain with the hydrophobic pocket”. Some 9-substituted guanines were
investigated as phosphate mimics with the functional end being sulfonic acid, carboxylic
acid, sulfonamide, carboxamide and nitrile”. The result showed that the sulfonic acid and
the carboxylic acid end groups interact signiﬁcantly with the phosphate binding site, but
both of them bind more weakly than the phosphate (Figure 2.6). The sulfonamide,
carboxamide, and the nitrile have no interaction with the phosphate binding site. It was
suggested that the only phosphate mimics that bind like phosphate are themselves ionic,

probably with limited ability to penetrate cell membranes.

ICs» (FM)
0 p g X ‘ f lmM Po4
HN ”‘7'“ M.;, HN' ”P r N 13 P0311; 0.044
H2", N N Hail/”N . "I 14 803‘] 0.18
HO 3 30.1,“...«0» \\ 15 SOZNHZ 100
O. o i” 16 coon 8.0
2:120:13 iphmphate ”X 117 CONH2 200

Figure 2.6. Phosphate mimics for PNP inhibitors

More hydrophobic mimics were developed to improve the cell penetration and oral
activity (Figure 2.7). Compound 18 was found to be a good inhibitor with IC50 35 nM.
Derivatives with different substituents were studied and compound (8)-20 with 3-
chlorophenyl and -CH2COOH group has IC50 of 5.9 nM, whereas the corresponding (R)-

isomer was 30-fold less potent36‘37.

84

ICso (11M)

0 0 no. ‘ x 1mm P04
HN\ N H” »\ N 19 H 20
H,N”“‘N”"N Han/4:1." N (S)-20 cnzcozn 5.9
.x (R)-20 cnzcozn 160
esp/7° ~ 21 CHICN 11
Ho 0” (:1 22 cnzcnzo 25
18 23 CHICOZMe 85

Figure 2.7. More hydrophobic mimics.

2.3. Transition State Inhibitor

The transition state studies have guided the transition state analog inhibitor design.
Chemically stable compounds with the same molecular shape and volume, incorporating
the key transition state features should provide inhibitors with high affinity to the target
enzymes. Because of the oxocarbenium ion feature of the transition state, l,4-dideoxy-
1,4-imino-D-ribitol compounds which can be largely protonated at pH 7 (pKal = 65)”,
should be potent transition state analog inhibitors of both PNP and nucleoside hydrolases.

‘3 . . . .
”‘44“, am1drazone 2547, and nicotinamide-

Compounds such as phenyl iminoribitol 24
based analog 2646 were studied and revealed potent (nM) inhibitors of the Crithidia

fasciculata nucleoside hydrolase such as p-bromo or p-aminophenyl iminoribitol.

85

HQ H HO H HQ H [N
,N.\ R /N. ' ,N, <5; ,7: 7
\g f ‘ \ y "N ~an \\ 2/ 1;" CON"?
‘1 ., I, '1, ,‘ \ .
no“ ”on no" “on no“ "on
R. = substituted phenyl R2 = H, C6H5, p-I-C5H4,
P’NOZ'C6H4
24 25 26
HOVH 8' HOK H 2 HO\ H " 1 HQ H 2
\i X/ _ \L Y/ . \\ “II/ p "Hz \\ i/ ..
Ho" ’OH no ”on HO“ ‘OH HO" "on
K, values (nM) for Crithidiafasciculam nucleoside hydrolase
28 30 51 1100
27 28 29 30

Further investigation of the substituted iminoribitol led to the conclusion that for more
potent inhibition of PNP, hydrolases, and transferaes, the R1 group needs to better mimic
the purine base of the substrate. Therefore, a family of Immucillins (1mm) was designed
and synthesized speciﬁcally to mimic the transition state structures. ImmH (31, (lS')-1-(9-
deazahypoxanthin-Q-ylH,4-dideoxy-l,4-imino-D-ribitol) and Iran (32, (18)-1-(9-
deazaguanin-9-yl)-l,4-dideoxy-l,4-imino-D-ribitol) were developed ﬁrst”5 '. In the
structure of ImmH, The ring oxygen is replaced by imino group, the same strategy in the
glycosidase inhibitor design. It mimics the partial positive charge of the
ribooxocarbenium transition state upon protonation. The C-C ribosidic bond instead of C-
N bond would mimic the lengthened C-N linkage in transition state and provide chemical

stability against breaking. The N7 of ImmH has greater pKa (9.8), and can be easily

86

protonated, similar to inosine at transition state and unlike N7 in the substrate (pKa=l.2).
These features cause immucillins in the picomolar range inhibitors (Figure 2.8), and

immucillin-H is currently in the phase I/II clinic trials for the treatment of T-cell

leukemia.

H H

N 1° ,N. o ,N. 9

Ho » ---.NH Ho H 1 ‘7, Ho H 2,», ~ 1.

I ‘1 NH \ ‘\ NH
\/‘o\¢N 1’ \’ N\/ \ \ ._ _ \K I N / /
\ / N \ / N \ / N \
\ ‘ I” , \ j, 77 "Hz
HO OH HO“ OH HO OH

lnosine ImmH 31 lmmG 32

K, = 23 pM (bovine PNP)
K. = 72 pM (Human PNP)

K. = 30 pM (bovine PNP)
K, = 29 pM (Human PNP)

Figure 2.8. ImmH family.

The crystal structure of PNP complex with immucillin-H and P04 has been solved“.
Together with the PNP-substrate (Michaelis complex )52 and PNP-product structures”,
they have provided important information about the structural details of the catalytic
mechanism. Figure 2.9 is the noncovalent interactions in the active sites of bovine PNP
complexes. The structure of the PNP-ImmH-PO4 resembles the structures of
PNP°inosine°SO4 (reactant) and PNP-hypoxanthine'ribose l-PO4 (products), but
PNP-ImmH-PO4 complex involves closer contacts between the protein and 11an
(Figure 2.9). Five direct hydrogen bonds to the purine base and two additional hydrogen

bonds in a proton-transfer bridge characterize the PNP interactions in these complexes.

Six of these seven contacts are shortened in the PNP'ImmH-PO4 complex compared to the

87

Michaelis complex. The ribose and phosphate moieties also form closer interactions with
the enzyme. This results in high afﬁnity of ImmH and decreased protein dynamics. The
distance between N9 of inosine or hypoxanthine to 04 of the phosphate group, called
reaction coordinate distance, is 4.8131 in PNP-ImmH-PO4 complex, while they are 5.6 and
5.2.31 in the Michaelis and product complex, respectively. This shorter distance indicates
tight binding between PNP and the inhibitor. It also ruled out an SNl mechanism in which
approximately 6A is required for the formation of a fully developed oxocarbenium ion
transition state while the shorter distances are associated with signiﬁcant bond orders to
leaving and/or attacking groups. There is no direct interaction ﬁ'om protein to stabilize
the partial positive charge developed at transition state, but neighboring groups play
important roles in the transition state stabilization. The distance between 04 of the
phosphate and N4’ of ribosyl group is 2.8/31 in PNP-Iran°PO4 complex, which is
signiﬁcantly shorter than the 04-04’ sulfate interaction in the PNP-substrate structure.
This stabilization of transition state provided by neighboring group phosphate ion is
different from the glycosidases and glycosyltransferases, in which a pair of enzymatic

carboxylates stabilizes the oxocarbenium ion transition state.

88

a) PNP-hosine-SO4
|248-N
243M 2.9/

"2 0);”- .-O
’ 2.7 . E201

  
  

 
           

\ N\\ 3" l '
3.4 ‘s ,’ 291’
3‘3 \‘.\ \ ,”’:I"~/ :: I ’ ' I
\N No . 3.3 NH 3.2
"W / \ / new» 2
" 1 6H. 2.7 N N 0"
; s 2 §\
3.. g gong-«nus 2.8 o gnaw-M219
,0 a =5 : . . =§ :
2.8,” \\ $534.1 £2.82. {.0'8220 2.8/4’ ‘\ §:§32:”2.q,0-3220
/ 2. (g? ",/ ,I’ 2 .535 ',
~‘= 2 --------- Y88-O *- 6“: ------ Ow.-3~‘
"8'0 u/ 0&0" \7. '-~Ho-v129 2.9/04.; 2‘ 2} '-~Ho-Y129
,I \‘ I; \ \\
”382.5291 '3\’:’ (i 833.03%». .1 Q?\’:' 9
:2?" 211/ 12.9 10‘. ‘N“"'A"6 :2?" 370 z’ ‘12.: 3.0“.“ ‘N’Aﬂ‘
/ ~, 333 , ‘1.
R84“ N-H86 N RM ”4"“ N833
NH1 N52 um NE:
c) PNP-Hx-R-l-PO4
l246-N
3.2!
2.3.9“

243M

 

1
z I
I, ‘ ' ,’
Y’s-O I”; ‘ 02 ...... W‘.‘~z‘5
2.9.” \ / 2.7 2;; “HO-Y129

833-03“? «'9: \qs\29 i
’ Ill-Has "'333

R84-N
NH1 NE2

89

Figure 2.9. The structures of bovine PNP with substrate analogues (inosine+SO4),
transition-state complex (lmm-H+PO4) and products (hypoxanthine+ribose l-PO4) bound
at the catalytic sites. Hydrogen bond distances are shown in angstroms. Red indicates
bonds that shorten signiﬁcantly and blue indicates bonds that lengthen signiﬁcantly in the
conversion of (a) to (b), and of (b) to (c).

One important feature of these complexes is the immobilization of the nucleophile
phosphate. It forms 10 relatively well positioned H-bonds from main chain atoms, side
chain atoms, a water molecule, and the 2’- and 3’-hydroxyl groups of the ribosyl group.
Therefore, the nucleophile is ‘ﬁxed’ in the binding site and the C1 ’ of ribosyl group has
to ‘migrate’ towards the nucleophile oxygen (04 of phosphate). Comparison of
PNP°inosine-SO4 to PNP-ImmH-PO4 revealed that the enzyme brings the iminoribitol C1 ’
0.9 A nearer to 04 of phosphate in the PNP-ImmH-PO4 complex (Figure 2.10:!) Thus,
formation of the transition state involves the electrostatically driven migration of the
ribooxocarbenium carbon from the purine ring towards the phosphate. The transition state
is reached when the Cl ’-N9 bond has lengthened from 1.44 to 1.77 (Figure 2.1%), while
the 5’-hydroxyl of the ribosyl group remain immobile. This migration of the anomeric
carbon also occurs in the glycosidases and glycosyltrnasferases. It reﬂects the ability of

enzymes to accomplish atomic motion of reactive groups in the protected environment of

the catalytic site3 '53.

90

 

Figure 2.10. Superpositions of the reactants (green), bound transition state analogues
(white with blue nitrogens and red oxygens), and products (gold) in the active sites of
bovine PNP (a) are compared with the superpositions of reactants, experimentally
determined transition states, and products (b). Distances (in angstroms) shown in (b) are
deduced from transition state analyses for PNP.

To favor the ribosyl oxocarbenium ion formation, there is a signiﬁcant distortion of the
ImmH geometry (Figure 2.11). The conformation of the iminoribitol moiety changes
ﬁom a C2’-endo to a C4’-endo on formation of the complex. The N9-Cl’-CZ’-H2’
dihedral angle is near 0° in the bound ImmH, which is consistent with the favored
transition state conformation with orbital overlap between the breaking N9-Cl ’ bond and
C2’-HZ’ bond of inosine. This angle is 42° in the unbounded ImmH. The 5’-hydroxyl is
anti to the ribosyl group in the free conformation, but adopts a syn conformation when
bound to PNP. This conformation place the 5’-oxygen 2.7 A above 04’ or N4’, with 04
of phosphate 2.8 A below the oxocarbenium oxygen. This ribosyl ring oxygen (04’) is

sandwiched by two electron-rich oxygens, favoring release of electron ﬁ'om 04’,

cleavage of the C1’-N9 bond, and formation of the oxocarbenium ion transition state.

91

bound /free ImmH

a b
l

M) .
A/

35187 75

Figure 2.11. Conforrnations of free and bound immucillins are from the small molecule
crystal structure of ImmH (gold) and ImmH in PNP-ImmH-PO4 (PDB 1B80, white with
blue nitrogens and red oxygens). Molecular overlap in (a) is C3' and C4' and in (b) is the
purine rings.

  

' .I ' L L 'L h n 27,54-56

f r J

and ribosyl hydrolases57'58

The purine and pyu'
adopt the similar migrating oxocarbenium mechanism. The pyrophosphate in
[L [L 'L J" “ as“ is immobilized by two bidentate magnesium ions in a scaffold
of 12 noncovalent interactions. It is important to note that the scenarios outlined above
are rationalizations and inferences based on the X-ray crystal structures which themselves
cannot provide direct proof of the pathways leading up to the structures. Detailed
molecular dynamics analyses that can support these inferences are still outside of
computational reach because of the signiﬁcant length scales and time scales involved.

Analogs of ImmH have also been developed to seek for powerful inhibitors of PNP,
hydrolases, and other ribosyltransferases. Iran 31 and 11an 32 are very potent
inhibitors of PNP. 11an 31, ImmG 32, and ImmA 33 are inhibitors of nucleoside
hydrolase“’”. 5’-phosphate ImmH 34 and ImmG 35 are the most potent inhibitors known

of several purine ,L L 'L " “ 6° (table 2.1).

r ""J

 

92

H
ii R‘ ,,/N.\
HQ a "207”. u “x...
\ N / ‘ I," \/'N\ / g .1/
, N \x / N “\
5 Hz ' R3
HO" OH HO OH
31 ImmH R,=0H, R2=H 34 ImmHP R3=H
32 ImmG R,=0H, 11,:an 35 ImmGP R3=NH2

33 ImmA R1=NH2, R2=H

Table 2.1. Inhibition constants for nucleohydrolase and HGPRT

 

Ki (nM) IUa IAGF Human HGPRT Malaria] HGPRT
ImmH 42 24
ImmG 84 l 10
ImmA 7 0.9
ImmHP 1 .8 1.0
ImmGP 4.6 14

 

a [U = inosine-uridine nucleoside hydrolase from Crithidia fasciculate. b IAG = inosine-

adenosine-guanosine nucleoside hydrolase from T rypanosoma brucei.

Because early studies on transition state analogs of AMP-N-ribosylhydrolase revealed
that formycin 5’-phosphate 36 is a powerful inhibitor6"62, 8-aza-immucillins 37-39 were
synthesized to study the effect of the extra nitrogen on the inhibition by modifying the
pKal of the 7-NH residue”. The result showed that 8-aza-immucillins are powerful
inhibitors for mammalian PNPs, comparable to the parent immucillins (table 2.2).

Immucillin 38 was inactive against human and bovine PNP. Perturbation of N-7

93

interactions with target enzymes by the extra 8-nitrogen has little effect on PNPs but

cause weaker inhibition on nucleosides hydrolases.

 

ﬂ 9 H o 11 NH: H
Ho 0 J \ KN" Ho H N: ; {NH HO. H N\’ » 4." HO H N" f-CONHa
V 7 ‘ ‘1 N\ «f? . ,N. ‘ . ,N. ‘ 7\
\f / N 7' \\ 47') N7 \Y / N' “ \ 7) NH,
HO" ’OH no“ ”on HO“ ”OH no“ "on
36 37 3s 39
Table 2.2. Inhibition constants
K (nM) Human PNP Bovine PNP IU IAG
ImmH 0.072 0.023
37 0.104 0.042 1600 13900
38 150 14000
39 0.096 0.060 4800 3700

 

The structure-activity relationship studies of immucillin analogs have been carried out to
obtain more potent inhibitors“. Modiﬁcations at the 2’-, 3’- or 5’-positions of the
azasugar moiety or at the 6-, 7-, or 8-positions of the deazapurine, as well as methylene-
bridged analogs resulted in poorer inhibitors than the parent immucillins for human PNP.
This reﬂects the close mimic of the transition state features by immucillins.

Immucillins have been established to be the most potent PNP inhibitors reported. [1an
is a 23 pM inhibitor for bovine PNP but a 2~3 fold weaker inhibitor for the human PNP.
Because of the different binding afﬁnity of ImmH with human and bovine PNPs48’65, the

transition state structure of human PNP was examined. Despite an 87% amino acid

94

identity between these two enzymes“, transition state of human PNP showed the more
dissociated nature between the ribosyl group and the leaving purine, with bond length
greater than 2.5131, and the necleophile is also farther for human PNP than for bovine
PNP. These led to a further developed oxocarbenium ion transition state with more
positive charge character at Cl’. Figure 2.12 is the transition state structures for bovine
and human PNP. This has guided the design of the second generation of PNP inhibitors“
68, which are speciﬁcally targeted to human PNP. Figure 2.13 is the second generation
inhibitor DADME (4’-deaza-l ’-aza-2’-deoxy-l ’-(9—methy1ene))-immucillin-H, DADME-
immucillin-G, and 8-aza-DADME-immucillin-H. 1n the structure of DADME-
immucillin-H, the nitrogen (immino group) is moved from the 4’ position to the
corresponding Cl’, which mimics the more positive charged anomeric position in the
transition state. The methylene bridge between the deaza purine and ribosyl analogues
groups of DADME-immucillin-H separates the two parts with 2.5 A in distance, which
makes the inhibitor similar to the more dissociated transition state. The inhibition studies
showed greater binding affinities of DADME-immucillins to the human PNP (Figure
2.13), and they are the most powerﬁil PNP inhibitors yet described. Comparing the
structures of PNP transition state, ImmH, and DADME-ImmH, we can conclude that
closer match to the transition state in both electrostatics and geometry are required to
achieve high affinity to the enzyme. Although those chemically stable analogs are
imperfect mimics of the enzyme transition state because of their covalent nature,
strategies based on the understanding of the transition state nature can be developed to

approach the goal of the transition state inhibitors.

95

lnosine

EHM

 

q I20
.7; P\
-0” 0“

bovine PNP transition state

 

human PNP transition state

H O
N-

"O W. '\
N V

/
. 1':
\i K53 "

-"
.
\ .
' NH
1

HO OH

immuclilln-H
Kn. bovine = 23 PM
Kc, tun-I = 56 PM

DADMe-immucillin-H
Kc. iii-nan 3 16 D"

Figure 2.12. Transition state structures for bovine and human PNP. The distances of 1.5
A for immucillin-H and 2.5 A for DADMe—immucillin-H refer to the linear distance
between the deazapurine ring and C- 1' or N-l' of the ribosyl analogues.

96

~2::
1

HO " N"

‘\ / M” ‘NH,
\/ IINH‘
HO.

DADMe—Immuclllin-G
Kd = 7 MI

40

H 9
,N ..
HO .1" NH
Ho.
DADMe-lmmucillln—H
Kd = 16 pH
41

HO gr‘
I >

\ / 'l 7 \ ‘ ‘ ., / N
‘ NH+

HO

B-ua-DADMe-lmm-H
K, = 2.0 nM

42

Figure 2.13. Second generation inhibitors for human purine nucleoside phosphorylase

97

2.4. References

(l) Stoeckler, J. D.; Cambor, C.; Parks, R. E. Human Erythrocytic Purine Nucleoside
Phosphorylase - Reaction with Sugar-Modiﬁed Nucleoside Substrates Biochemistry
1980, 19, 102-107.

(2) Hershﬁeld, M. S., Mitchell, B. S. In The Metabolic Basis of Inherited Disease; 7th
ed.; Scriber, C. R., Beaudet, A. L., Sly, W. S., Valle, D., Ed.; McGraw-Hill: New York,
1995, pp 1725-1768.

(3) Markert, L. M. Immunodeﬁc. Rev. 1991, 3, 45-81.

(4) Krenitsky, T. A.; Tuttle, J. V.; Koszalka, G. W.; Chen, 1. S.; Beacham, L. M.;
Rideout, J. L.; Elion, G. B. Deoxycytidine Kinase from Calf Thymus - Substrate and
Inhibitor Speciﬁcity Journal of Biological Chemistry 1976, 251 , 4055—4061 .

(5) Carson, D. A., Kaye, J., Seegmiller, J. E. Lymphospeciﬁc toxicity in adenosine
deaminase deﬁciency and purine nucleoside phosphorylase deﬁciency: possible role of
nucleoside kinase(s) Proc. Natl. Acad. Sci. U. S. A. 1977, 74, 5677-5681.

(6) Mitchell, B. S.; Mejias, 13.; Daddona, P. E.; Kelley, W. N. Purinogenic
Immunodeﬁciency Diseases - Selective Toxicity of Deoxyribonucleosides for T-Cells
Proc. Natl. Acad. Sci. U. S. A. 1978, 75, 5011-5014.

(7) Ullman, B.; Gudas, L. J .; Cliﬁ, S. M.; Martin, D. W. Isolation and Characterization of
Purine-Nucleoside Phosphorylase-Deﬁcient T-Lymphoma Cells and Secondary Mutants
with Altered Ribonucleotide Reductase - Genetic Model for Immunodeﬁciency Disease
Proc. Natl. Acad. Sci. U. S. A. 1979, 76, 1074-1078.

(8) Thelander, L., Reichard, P. Reduction of ribonucleotides Annu. Rev. Biochem. 1979,
48, 133-158.

(9) Snyder, F. F ., Jenuth, J. P., Mably, E. R., Mangat, R. K. Point mutations at the purine
nucleoside phosphorylase locus impair thymocyte differentiation in the mouse Proc. Natl.
Acad. Sci. USA 1997, 94, 2522-2527.

(10) Boehncke, W. H.; Gilbertsen, R. B.; Hemmer, J .; Sterry, W. Evidence for a Pathway
Independent from 2'—Deoxyguanosine and Reversible by Il-2 by Which Purine
Nucleoside Phosphorylase Inhibitors Block T-Cell Proliferation Scandinavian Journal of
Immunology 1994, 39, 327-332.

(11) Montgomery, J. A. Purine Nucleoside Phosphorylase - a Target for Drug Design
Medicinal Research Reviews 1993, 13, 209-228.

98

(12) Morris, P. E., Montgomery, J. A. Inhibitors of the enzyme purine nucleoside
phosphorylase Exp. Opin. T herapeut. Pats. 1998, 8, 283-299.

(13) Parsons, W. H. In Annual Reports in Medicinal Chemistry, Vol 29, 1994; Vol. 29, pp
175-184. 7

(14) Sircar, J. C.; Gilbertsen, R. B. Purine nucleoside phosphorylase (PNP) inhibitors:
Potentially selective immunosuppressive agents Drugs of the Future 1988, 13, 653-668.

(15) Petersen, C.; Moller, L. B. The RihA, RihB, and RihC ribonucleoside hydrolases of
Escherichia coli - Substrate speciﬁcity, gene expression and regulation Journal of
Biological Chemistry 2001, 276, 884-894.

(16) Gopaul, D. N.; Meyer, S. L.; Degano, M.; Sacchettini, J. C.; Schramm, V. L.
lnosine-uridine nucleoside hydrolase from Crithidia fasciculata. Genetic characterization,
crystallization, and identiﬁcation of histidine 241 as a catalytic site residue Biochemistry
1996, 35, 5963-5970.

(17) Hammond, D. J.; Gutteridge, W. E. Purine and pyrimidine metabolism in the
trypanosomatidae Mol. Biochem. Parasitol. 1984, 13, 243-261.

(18) Parkin, D. W.; Horenstein, B. A.; Abdulah, D. R.; Estupinan, B.; Schramm, V. L.
Nucleoside Hydrolase from Crithidia-Fasciculata - Metabolic Role, Puriﬁcation,
Speciﬁcity, and Kinetic Mechanism J. Biol. Chem. 1991, 266, 20658-20665.

(19) Estupinan, B.; Schramm, V. L. Guanosine-lnosine-Preferring Nucleoside N-
Glycohydrolase from Crithidia-Fasciculata J Biol. Chem. 1994, 269, 23068—23073.

(20) Sculley, D. G.; Dawson, P. A.; Emmerson, B. T.; Gordon, R. B. A Review of the
Molecular-Basis of Hypoxanthine-Guanine Phosphoribosyltransferase (Hprt) Deﬁciency
Human Genetics 1992, 90, 195-207.

(21) Fong, W. P.; Wong, R. N. S.; Go, T. T. M.; Yeung, H. W. Enzymatic-Properties of
Ribosome-Inactivating Proteins (Rips) and Related Toxins Life Sciences 1991, 49, 1859-
1869.

(22) Barbieri, L.; Battelli, M. G.; Stirpe, F. Ribosome-Inactivating Proteins from Plants
Biochimica Et Biophysica Acta 1993, 1154, 237-282.

(23) Rodgers, J.; Femec, D. A.; Schowen, R. L. Isotopic Mapping of Transition-State
Structural Features Associated with Enzymic Catalylis of Methyl Transfer J. Am. Chem.
Soc. 1982, 104, 3263-3268.

(24) Cleland, W. W. The use of isotope effects to determine transition-state structure for
enzymic reactions Methods Enzymol. 1982, 87, 625-641.

99

(25) Northrop, D. B. Steady-state analysis of kinetic isotope effects in enzymic reactions
Biochemistry 1975, 14, 2644-2651.

(26) Schramm, V. L. Enzymatic transition states and transition state analog design
Annual Review of Biochemistry 1998, 6 7, 693-720.

(27) Tao, W.; Grubmeyer, C.; Blanchard, J. S. Transition state structure of Salmonella
typhimurium orotate phosphoribosyltransferase Biochemistry 1996, 35 , 14-21.

(28) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - lnosine Hydrolysis,
Tight-Binding of the H ypoxanthine Intermediate, and 3rd-the-Sites Reactivity
Biochemistry 1992, 31, 5964-5973. ‘

(29) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - Catalytic
Mechanism and Transition-State Analysis of the Arsenolysis Reaction Biochemistry
1993, 32, 13212-13219.

(30) Kline, P. C.; Schramm, V. L. Pre-Steady-State Transition-State Analysis of the
Hydrolytic Reaction Catalyzed by Purine Nucleoside Phosphorylase Biochemistry 1995,
34, 1153-1162.

(31) Fedorov, A.; Shi, W.; Kicska, G.; Fedorov, E.; Tyler, P. C.; Fumeaux, R. H.;
Hanson, J. C.; Gainsford, G. J.; Larese, J. Z.; Schramm, V. L.; Almo, S. C. Transition
state structure of purine nucleoside phosphorylase and principles of atomic motion in
enzymatic catalysis Biochemistry 2001, 40, 853-860.

(32) Mazzella, L. J.; Parkin, D. W.; Tyler, P. C.; Fumeaux, R. H.; Schramm, V. L.
Mechanistic diagnoses of N-ribohydrolases and purine nucleoside phosphorylase J. Am.
Chem. Soc. 1996, 118, 2111-2112.

(33) Parkin, D. W.; Limberg, 0.; Tyler, P. C.; Fumeaux, R. H.; Chen, X. Y.; Schramm,
V. L. lsozyme-speciﬁc transition state inhibitors for the trypanosomal nucleoside
hydrolases Biochemistry 1997, 36, 3528-3534.

(34) Montgomery, J. A.; Niwas, 8.; Rose, J. D.; Secrist, J. A.; Babu, Y. S.; Bugg, C. E.;
Erion, M. D.; Guida, W. C.; Ealick, S. E. Structure-Based Design of Inhibitors of Purine
Nucleoside Phosphorylase .1. 9-(Arylmethyl) Derivatives of 9-Deazaguanine J. Med.
Chem. 1993, 36, 55-69.

(35) Secrist, J. A.; Niwas. 8.; Rose, J. D.; Babu, Y. S.; Bugg, C. E.; Erion, M. D.; Guida,
W. C.; Ealick, S. E.; Montgomery, J. A. Structure-Based Design of Inhibitors of Purine
Nucleoside Phosphorylase .2. 9-A1icyclic and 9-Heteroalicyclic Derivatives of 9-
DeazaguanineJ. Med. Chem. 1993, 36, 1847-1854.

(36) Erion, M. D.; Niwas, 8.; Rose, J. D.; Ananthan, S.; Allen, M.; Secrist, J. A.; Babu,
Y. S.; Bugg, C. E.; Guida, W. C.; Ealick, S. E.; Montgomery, J. A. Structure-Based

100

Design of Inhibitors of Purine Nucleoside Phosphorylase .3. 9-Ary1methyl Derivatives of
9-Deazaguanine Substituted on the Methylene Group J. Med. Chem. 1993, 36, 3771-
3783.

(37) Brion, M. D.; Niwas, 8.; Rose, J. D.; Ananthan, S.; Allen, M.; Secrist, J. A.; Babu,
Y. S.; Bugg, C. E.; Guida, W. C.; Ealick, S. E.; Montgomery, J. A. Structure-Based
Design of Inhibitors of Purine Nucleoside Phosphorylase .3. 9-Arylmethy1 Derivatives of
9-Deazaguanine Substituted on the Methylene Group (Vol 36, Pg 3773, 1993) J. Med.
Chem. 1994, 37, 1034-1034.

(38) Guida, W. C.; Elliott, R. D.; Thomas, H. J .; Secrist, J. A.; Babu, Y. S.; Bugg, C. E.;
Erion, M. D.; Ealick, S. E.; Montgomery, J. A. Structure-Based Design of Inhibitors of
Purine Nucleoside Phosphorylase .4. A Study of Phosphate Mimics J. Med. Chem. 1994,
37, 1109-1114.

(39) Farutin, V.; Masterson, L.; Andricopulo, A. D.; Cheng, J. M.; Riley, B.; Hakimi, R.;
Frazer, J. W.; Cordes, E. H. Structure-activity relationships for a class of inhibitors of
purine nucleoside phosphorylase J. Med. Chem. 1999. 42, 2422-2431.

(40) Stoeckler, J. D. R., J. B.; Parks, R. E., Jr.; Chu, M.-Y.; Lim,; M.-L.; Ren, W.-Y. K.,
R. S., Inhibitors of Purine Nucleoside Phosphorylase: Effecta of 9-Deazapurine
Ribonucleorrides and Synthesis of 5’-Deoxy-5’-iodo-s-deazainosine Cancer Res. 1986,
46, 1774-1778.

(41) Tuttle, J. V. K., T. A. Effects of Acyclovir and Its Metabolites on Purine Nucleoside
Phosphorylase J. Biol. Chem. 1984, 269, 40664069.

(42) Ealick, E. B., Y. S.; Bugg, C. E.; Erion, M. D.; Guida, W. C.;; Montgomery, J. A. S.,
J. A.; . 111. Application of Crystallographic and Modeling Methods in the Design of
Purine Nucleoside Phosphorylase Inhibitors Proc. Natl. Acad. Sci. USA. 1991, 88,
11540-11644.

(43) Horenstein, B. A.; Schramm, V. L. Electronic Nature of the Transition-State for
Nucleoside Hydrolase - a Blueprint for Inhibitor Design Biochemistry 1993, 32, 7089-
7097.

(44) Horenstein, B. A.; Zabinski, R. F.; Schramm, V. L. A New Class of C-Nucleoside
Analogs - 1-(S)-Ary1-1,4-Dideoxy-1,4-Imino-D-Ribitols, Transition-State Analog
Inhibitors of Nucleoside Hydrolase Tetrahedron Lett. 1993, 34, 7213-7216.

(45) Fumeaux, R. H.; Limberg, 0.; Tyler, P. C.; Schramm, V. L. Synthesis of transition
state inhibitors for N-riboside hydrolases and transferases Tetrahedron 1997, 53, 2915-
2930.

(46) Fumeaux, R. H.; Schramm, V. L.; Tyler, P. C. Transition state analogue inhibitors of
protozoan nucleoside hydrolases Bioorganic & Medicinal Chemistry 1999, 7, 2599-2606.

101

(47) Boutellier, M.; Ganem, B.; Horenstein, B. A.; Schramm, V. L. Design and Synthesis
of D-Ribofuranoamidrazones as Inhibitors of Nucleoside Processing Enzymes Synlett
1995, 510-512.

(48) Miles, R. W.; Tyler, P. C.; Fumeaux, R. H.; Bagdassarian, C. K.; Schramm, V. L.
One-third-the-sites transition-state inhibitors for purine nucleoside phosphorylase
Biochemistry 1998, 37, 8615-8621.

(49) Evans, G. B.; Fumeaux, R. H.; Gainsford, G. J.; Schramm, V. L.; Tyler, P. C.
Synthesis of transition state analogue inhibitors for purine nucleoside phosphorylase and
N-riboside hydrolases Tetrahedron 2000, 56, 3053-3062.

(50) Schramm, V. L. Development of transition state analogues of purine nucleoside
phosphorylase as anti-T-cell agents Biochimica Et Biophysica Acta-Molecular Basis of
Disease 2002, 1587, 107-117.

(51) Schramm, V. L. Enzymatic transition state poise and transition state analogues Acc.
Chem. Res. 2003, 36, 588-596.

(52) Mao, C.; Cook, W. J.; Zhou, M.; Federov, A. A.; Almo, S. C.; Ealick, S. E. Calf
spleen purine nucleoside phosphorylase complexed with substrates and substrate
analogues Biochemistry 1998, 3 7, 7135-7146.

(53) Schramm, V. L.; Shi, W. X. Atomic motion in enzymatic reaction coordinates
Current Opinion in Structural Biology 2001, 11, 657-665.

(54) Shi, W. X.; Li, C. M.; Tyler, P. C.; Fumeaux, R. H.; Cahill, S. M.; Girvin, M. E.;
Grubmeyer, C.; Schramm, V. L.; Almo, S. C.“ The 2.0 angstrom structure of malarial
purine phosphoribosyltransferase in complex with a transition-state analogue inhibitor
Biochemistry 1999, 38, 9872-9880.

(55) Shi, W. X.; Li, C. M.; Tyler, P. C.; Fumeaux, R. H.; Grubmeyer, C.; Schramm, V.
L.; Almo, S. C. The 2.0 angstrom structure of human hypoxanthineguanine
phosphoribosyltransferase in complex with a transition-state analog inhibitor Nature
Struct. Biol. 1999, 6, 588-593.

(56) Heroux, A.; White, B. L.; Ross, L. J.; Davis, R. L.; Borhani, D. W. Crystal structure
of Toxoplasma gondii hypoxanthine-guanine phosphoribosyltransferase with XMP,
pyrophosphate, and two Mg2+ ions bound: Insights into the catalytic mechanism
Biochemistry 1999, 38, 14495-14506.

(57) Horenstein, B. A.; Parkin, D. W.; Estupinan, B.; Schramm, V. L. Transition-State

Analysis of Nucleoside Hydrolase from Crithidia-Fasciculata Biochemistry 1991, 30,
10788-10795.

102

(58) Degano, M.; Almo, S. C.; Sacchettini, J. C.; Schramm, V. L. Trypanosomal
nucleoside hydrolase. A novel mechanism from the structure with a transition-state
inhibitor Biochemistry 1998, 3 7, 6277-6285.

(59) Miles, R. W.; Tyler, P. C.; Evans, G. B.; Fumeaux, R. H.; Parkin, D. W.; Schramm,
V. L. Irninoribitol transition state analogue inhibitors of protozoan nucleoside hydrolases
Biochemistry 1999, 38, 13147-13154.

(60) Li, C. M.; Tyler, P. C.; Fumeaux, R. H.; Kicska, G.; Xu, Y. M.; Grubmeyer, C.;
Girvin, M. E.; Schramm, V. L. Transition-state analogs as inhibitors of human and
malarial hypoxanthine-guanine phosphoribosyltransferases Nature Struct. Biol. 1999, 6,
582-587.

(61) Dewolf, W. E.; Fullin, F. A.; Schramm, V. L. Catalytic Site of Amp Nucleosidase -
Substrate-Speciﬁcity and Ph Effects with Amp and Fonnycin 5'-Po4 J. Biol. Chem. 1979,
254, 868-875.

(62) Leung, H. B.; Schramm, V. L. Adenylate Degradation in Escherichia-Coli - the Role
of Amp Nucleosidase and Properties of the Puriﬁed Enzyme J. Biol. Chem. 1980, 255,
867-874.

(63) Evans, G. B.; Fumeaux, R. H.; Gainsford, G. J.; Hanson, J. C.; Kicska, G. A.; Sauve,
A. A.; Schramm, V. L.; Tyler, P. C. 8-aza-immucillins as transition-state analogue
inhibitors of purine nucleoside phosphorylase and nucleoside hydrolases J. Med. Chem.
2003, 46, 155-160.

(64) Evans, G. B.; Fumeaux, R. H.; Lewandowicz, A.; Schramm, V. L.; Tyler, P. C.
Exploring structure-activity relationships of transition state analogues of human purine
nucleoside phosphorylase J. Med. Chem. 2003, 46, 3412-3423.

(65) Kicska, G. A.; Tyler, P. C.; Evans, G. B.; Fumeaux, R. H.; Kim, K.; Schramm, V. L.
Transition state analogue inhibitors of purine nucleoside phosphorylase from
Plasmodium falciparum J. Biol. Chem. 2002, 277, 3219-3225.

(66) Lewandowicz, A.; Tyler, P. C.; Evans, G. B.; Fumeaux, R. H.; Schramm, V. L.
Achieving the ultimate physiological goal in transition state analogue inhibitors for
purine nucleoside phosphorylase J. Biol. Chem. 2003, 278, 31465-31468.

(67) Evans, G. B.; Fumeaux, R. H.; Lewandowicz, A.; Schramm, V. L.; Tyler, P. C.
Synthesis of second-generation transition state analogues of human purine nucleoside
phosphorylaseJ. Med. Chem. 2003, 46, 5271-5276.

(68) Evans, G. B.; Fumeaux, R. H.; Tyler, P. C.; Schramm, V. L. Synthesis of a transition

state analogue inhibitor of purine nucleoside phosphorylase via the Mannich reaction
Org. Lett. 2003, 5, 3639-3640.

103

Chapter 3

Trihydroxy-Z-Thiaquinolizidine Derivatives as a New Class of

Bicyclic Glycosidase Inhibitors

104

ABSTRACT

Trihydroxy-2-thiaquinolizidines, a new class of bicyclic dideoxy-iminohexitol
glycosidase inhibitor derivatives with nominally the D-gluco, L-ido, D-manno and L-gulo
conﬁgurations were synthesized. X-ray analyses indicated that the preferred
conformation for D-gluco and D-manno derivatives was a ﬂat trans- fused system.
Unlike deoxynojirimycin, compound with D-gluco conﬁguration was selective for a-
glucosidases (yeast and rice) and showed no inhibitory activity towards B-glucosidase
(almond), a-galactosidase (green coffee beans), a-galactosidase (E-coli) and a-
mannosidase (jack bean), while L-ido derivative was speciﬁc for B-glucosidase (almond).

Possible reasons for the speciﬁcity will be discussed.

105

3.1. Introduction

Glycosidase inhibitors such as deoxynojirimycin 1 and castanospermine 2, generally
known as iminoalditols or azasugars and their derivatives have great promise for
numerous medical applications"3 ranging from diabetes“5 and other metabolic disorders
through antimicrobials7'9, cancer"), autoimmune diseases”"4, and neurological
disorders”. Despite this, they have not realized their full clinical potential. This is largely
because of a lack of commercially viable syntheses and difﬁculty in preparing a

comprehensive palette of variant structures. Many of the possible drug candidates are

available in only small exploratory amounts.

     

deoxynojirimycin castanospermine Swainsonine

Azasugars are carbohydrate pyranose or furanose analogs in which the ring oxygen atom
is replaced with a nitrogen atom. This presents a signiﬁcant synthetic challenge to
organic chemists. Bicyclic systems such as castanospermine (2) and swainsonine (3) are
especially interesting because of their potency as glycosidase inhibitors. Their structural
rigidity and the added interaction of the second ring might be contributing factors to this.
This second ring increases the difﬁculty of the synthesis. A number of chemical and

16-31

enzymatic syntheses of azasugars have been reported in recent years . Several

strategies have been used to introduce a basic nitrogen atom in the ring of

106

deoxynojirimycin and its derivatives. These strategies include chemo-enzymatic method,
arninomercuration, double-reductive amination, N-alkylation, reductive double-alkylation

and triple reductive amination. Some representative methods are described below:

3. l . 1. Chemo-enzymatic Synthesis

The chemo-enzymatic approach has been extensively developed by C.-H. Wongzz‘zs. It is
generally a two-step process involving an enzymatic aldol condensation and a catalytic
intramolecular reductive amination. The asymmetric synthesis of 1 and 8 is shown in
figure 3.1. However, this method of synthesis involving enzymes is not a general method

and it is limited by scale-up problems.

 

 

Et 9H
+ :
magi/on + E N "" = 03" '
3 2) PD? aldolase ; "3
OH OH OH
4 s 6: R 7:8
C
I) phosphatase H
- H - " " + H ' H
2) H2, Pd/C HO HO
OH
, 1:L2~l:¢l

overall yield 48% ~ 64%

Figure 3.1. Chemo-enzymatic synthesis of l and 8

107

3. 1.2. Aminomercuration26

The intramolecular aminomercuration is illustrated in figure 3.2 for the case of D-
glucose. A one-pot, reductive ring opening and reductive amination afforded an
aminoalkene. When this key intermediate was reacted with mercuric triﬂuoroacetate in
anhydrous THF, a 3:2 mixture of bromomercurials 11 and 12.was isolated. The major
cyclization product 11 could be transformed to l-deoxynojirimycin 1 by reductive
oxygenation (NaBH4-DMF-Oz) and hydrogenolytic deprotection. In this synthesis, only
one of the bromomercurials (11) can give the desired product, which results in a low

yield of deoxynojirimycin l. The mercury toxicity is also a problem.

Zn moms", BnO ”“3“ agrococrp,“
Bus“, 91% :Tm
an”) 03:!

10

r) NaBHrDMF-O;
2) up I’d/C

Figure 3.2. Aminomercuration

3.1.3. Double Reductive Amination27’28
The double reductive amination of dicarbonyl sugars with primary amines and sodium
cyanoborohydride in methanol allowed for the syntheses of l-deoxynojirimycin, 1-

deoxyrnannojirimycin and N-alkylated derivatives (figure 3.3).

108

 

 

 

 

CH20H
H——OH
BuZSnO °=‘ o 303” - 5° HO——H
MeOH ...umo 2 H——OH
Brz, crlzcr2 70% =0
00C, 480/0 ”00 ‘
HO LH20H
l4
1 15
R H
RNH2 pro/”4,," N
# + a
NaBH3CN .
MeOH ”0W "”’OH
OH
17
RN "2 yield ratio
PhZC EN"; 70% 95 5
PhZCHCHzNH2 73% 95 5
C4H9NH2 55% 95 5
CnstNHZ 2.7./o 95 5

Figure 3.3. Double reductive amination

3.1.4. Intramolecular N-Alkylation30

In this approach, the cyclization to furnish the ring nitrogen was achieved by a direct
intramolecular N-alkylation of a primary amino group by mesyl displacement in a
partially protected hexitol prepared in several steps ﬁom diethyl tartrate. This approach

for the synthesis of l in enantiomerically pure form is illustrated in figure 3.4.

l09

N, (X
"(g—(OH 1.2. "(k/$496 ”W's MOMWO)$

czusooc cooczu5

2) none: uoud o’
Diethyl L-tartrote 19 20 68%
‘8 1)TBAF
2) use:
on one» N3 o>(
HO N" -H—C' mono "" 1)__H2. MC uou “
HO 2) TEA/Moon , I, OM:
on yo ""0" mono o’
1 90% 22 92% 21 89%

Figure 3.4. N-Intramolecular alkylation

3.1 .5. Reductive Double-alkylation3|

A reductive double-cyclization has been employed in the synthesis of analogs of the
mannosidase inhibitor swainsonine (3). The synthesis of the ring-expanded derivatives
(30, 31) of swainsonine started from D-arabinose. Protection of hydroxyl groups, Wittig
reaction and Mitsunobu reaction afforded the azide 26. Epoxidation with mCPBA yielded
two diastereomeric epoxides. The epoxide group and a chlorinated carbon were the two
nucleophilic centers involved. Upon reduction, primary amines were generated, which
.were cyclized to give quinolizidines 30 and 31 (ﬁgure 3.5). This synthesis yields the
bicyclic azasugars, however, the diastereoselectivity is not high and it still suffers ﬁ'om

the drawback that separation of the diastereoisomers has to be performed (28 and 29).

110

 

30 31

Figure 3.5. Reductive double-alkylation. i, BrPh3P(CH2)4Cl, KN(TMS)2, 71%; ii, HN3,
PPh3, DEAD, 84%; iii, mCPBA, 88%; iv, a, 112, Pd/C, EtOH; b, K2CO3, EtOH, reﬂux; 0,
separate diasteromers; v, H2 (45psl), Pd/C, HCl, MeOH, 99%.

3.1.6. Triple Reductive Amination (TRA)29

A triple reductive amination involving three carbonyl groups and an amine was used for
the synthesis of the bicyclic aza sugars castanospermine 2 (ﬁgure 3.6) and swainsonine 3.
It was suggested that initial amination occurred at one of the carbonyl groups, and the
resulting carbinolamine underwent sequential intramolecular reactions with the remaining

carbonyl groups at a faster rate than competing intermolecular processes. The TRA

approach is efﬁcient for the synthesis of compounds with bicyclic indolizidine

lll

framework, but the preparation of the precursor with aldehyde and ketone functionalities

is not an easy exercise.

   
 

Figure 3.6. Triple reductive amination i, (a) Swem oxidation; (b) 03, CH2C12, -78°C then
Ph3P; (c) THF-9M HCl 74% for 3 steps; ii, NH4HC02, NaCNBH3, MeOH, 78%; iii, 10%
Pd-C, MeOH-HCOOH, 80%.

3.2. Results and Discussions

3.2.1. Synthesis of Trihydroxy-Z-thiaquinolizidine Derivatives

We explored the possibility of synthesizing bicyclic analogs of iminohexitols such as
deoxynojirimycin and related compounds in which 0-6 were replaced by a sulfur atom.
We also envisaged bridging the 6-position to the ring nitrogen with a two-carbon
fragment to form a trihydroxy-Z-thiaquinolizidine ring system thus increasing rigidity and
increasing lipophilicity. Such systems have never been reported before but hold great

promise because the formation of a carbon-carbon bond is circumvented as in 2, 3 and 30.

The presence of sulfur (closely related to oxygen) at a position that is normally

112

oxygenated is also a decided advantage. If such systems could be reached using a general
' strategy, analogs with differing conﬁgurations at the various carbon centers and differing
substitution patterns could be made, increasing the chance of obtaining compounds with
useﬁrl therapeutic potential. A total of four such systems (36a, 36b, 51a, and 51b) were

prepared corresponding to the D-gluco, L-ido, D-manno, L-gulo conﬁgurations.

 

Figure 3.7 shows the synthesis for the “D-gluco” and “L-ido” compounds. It starts from
the B-methyl-D-glucopyranoside. Triphenylphosphine and carbon tetrabromide
selectively converted the primary hydroxyl group to a bromo group. The hydroxyl groups
at C(2), C(3), C(4) positions were protected by acylation using pivaloyl chloride
(trimethylacetyl chloride). The selection of protecting group was based on the tolerance
of the chromium oxidation and the subsequent reactions. Normal protecting groups such
as acetals, being attacked, methyl ether and benzyl ethers are unsuitable under this
condition. Methyl and benzyl ethers have been reported to be attacked by chromium
trioxide in acetic acid”. Ester groups are stable as the protecting group. Pivaloyl groups
were selected over acetyl groups because of the partial deacylation of acetates by

aminoethanethiol, resulting in difficulty in puriﬁcation and low yield. Oxidation of the

113

peracylated 6-bromo-6-deoxy-glycoside 39 with chromium trioxide in acetic acid
afforded a 5-ulosonic acid ester 40. The oxidation of acetylated B-glycopyranosides by
chromium trioxide has been reported by S. J. Angyal and K. James33 to afford 5-keto
esters, independent of the conﬁgurations on C2, C3, or C4. The oxidation is speciﬁc to B-
glycopyranosides, while a-glycopyranosides are not attacked. Treatment of 40 with 2-
arninoethanethiol yielded the aminal 41 directly which underwent reduction by
NaBCNH3 and cyclization to the desired lactam 42a and the L-ido isomer 42b. The
recyclization to form lactams occurred spontaneously over one day. On adding sodium
carbonate, the rate was accelerated of lactam formation. There was some variability in the
actual amount formed ranging from traces to 2.521 in favor of the D-gluco analog.
Reduction of the lactams with borane yielded clean amine products, with protecting
groups intact. The crude product underwent deacylation by methoxide to afford the D-

gluco and L-ido compounds.

114

OH Br Br

HO 0 I _ Ho 0 rr FWD 0
HO OMB HO 0M0 PlVO 0M0

OH OH OPIV
37 38 39

 

 

 

     
   

Br
L“. Plvo /0 o .31.,
PlvO
OPiV 0MB
40
42b
Vii viii
42a 35'
Vii viii
——>
42b 43b 36b

115

Figure 3.7. Synthesis of products with D-gluco and L-ido conﬁgurations. i. Ph3P, CBr4,
pyridine, 85%; ii. PivCl, pyridine, 84%; iii. Cr03, AczO, HOAc, 97%; iv. HS(CH2)2NH2,
CH3OH; v. NaCNBH3, CH3OH; vi. NazCO3, CHC13, 74% for 3 steps; vii. BH3-THF; viii.
NaOCH3, CH3OH, 71% for 36a and 75% for 36b for 2 steps.

Products with the D—manno and L-gulo conﬁgurations (51a and 51b) were also obtained
in the same fashion starting from B-methyl-D-mannoside (ﬁgure 3.8). The D-manno
lactam 49a and L-gulo lactam 49b was obtained in 4:1 ratio in favor of the D-manno
lactam. Reduction of lactams and deprotection yielded products 51a and 51b in good
yields. Besides the bromo function, the leaving group at the C(6) position can also be
mesyl and other groups that are stable under chromiun oxidation conditions. Therefore,
this synthetic strategy is a powerful method to synthesize bicyclic azasugars with
different conﬁgurations starting from methyl B-D—glycopyranosides. The ease of

synthesis can lead to the generation of a large family of bicyclic azasugars, with

possibilities of identifying active compounds against glycosidases.

116

HO 0"" HO Pervo

 

 

 

Br
Plv
111 PlvO , :0
PIVO
47 — as — H N
PIVOl'" O
PIVO OPlv
49b

 

Figure 3.8. Synthesis of products with the D-manno and L-gulo conﬁgurations. i. Ph3P,
CBr4, pyridine, 84%; ii. PivCl, pyridine, 80%; iii. Cr03, AczO, HOAc, 91%; iv
HS(CH2)2NH2, CH3OH; v. NaCNBH3, CH3OH, 81% for 2 steps; vi. BH3-THF; vii.
NaOCHg, CH3OH, 73% for 51a and 71% for 51b for 2 steps.

117

 

 

Figure 3. 9. X-ray structure of 7(.S'),8(R),9(R),10(S)-tr..L.,1.-.“, 2 ‘L ' " T 36a
and 7(R),8(R),9(R),10(S)-u'l.., --... ., 2 5111 showing the trans-type ring
junction and overall ﬂat geometry.

Products 36a and Sla were recrystalized in water to give colorless crystals. X-ray
crystallography analysis (ﬁgure 3.9) indicated two six-membered rings with relaxed
chair conformation for both D-gluco and D-manno products. A trans-diequatorial type
fusion between the rings gives the molecules an overall ﬂat geometry. The expected

intermediate oxocarbenium species is very ﬂat because of the double bond character

between the ring oxygen and C-1 (Figure 3.10).

118

 

Figure 3.10. Structure of oxycarbenium intermediate in the hydrolysis of glucosides from
an ab initio calculation showing equipotential contours.

The sulfur atom at 6—position can be derivatized easily to give the corresponding sulfone
or sulfoxide. From lactam 42a, mCPBA oxidation afforded the clean product 52 and after
reduction of the lactam and deprotection, the sulfone can be obtained in 84% yield from

the lactam 42a without puriﬁcation of the intermediates (figure 3.11).

119

  
 

  
 

PiVOIIIIIIo- O PivOllllln-

PivO ‘0in PM) ”0in

 

Figure 3.11. Synthesis of 2,2-dioxy-7(.S'),8(R),9(R),10(S)-trihydroxy-2-thiaquinolizidin-
2-one. i, mCPBA, CHzClz; ii, BH3, T‘HF; iii, NaOMe, MeOH, 84% yield overall.

3.2.2. Inhibitory Activities of Trihydroxy-2-thiaquinolizidine Derivatives against
Glycosidases

The inhibitory activity is evaluated by the dissociation constant K, of the inhibitor. The
enzyme inhibition is described by the Michaelis - Menten rate law. The scheme that

corresponds to competitive inhibition is ﬁgure 3.1234.

120

 

 

+
I
l K.
El
Figure 3.12. Enzyme inhibition scheme.

The dissociation constant for the inhibitor is

_ [El [IL
Ki ‘ [E11 eq. 1

 

By steady-state assumption, the resulting Michaelis-Menten equation is

 

Vmu [S]
V = [I] where KM = M
Km(1+?i‘)+lsl k1

eq. 2
A lineweaver - Burk plot can be used to determine the inhibition constant K). Taking the

reciprocal of equation 2, we obtain

1_KM

 

[I] 1 +1

( _ _

+
V VIIIIX lg [S] vaX eq. 3

The slope of the line will be

K I
M(1+ ll

vmax Ki )

Slope =
eq. 4

Thus, K, can be determined by plotting the slope values vs. [ I ]. The resulting secondary
plot will have Y-axis intercept of KM/Vmax and a slope of KM/Vmax Ki. The value of K, is

the intercept /slope of this plot.

121

Enzymes were assayed according to standard procedures35 by following the hydrolysis of
nitrophenyl glycosides spectrophotometrically except a-glucosidase (rice). A series of
enzyme assays using a-glucosidase (yeast), a-glucosidase (rice), B-glucosidase (almond),
err-galactosidase (green coffee beans), B-galactosidase (E. Coli), a-mannosidase (jack
beans) were carried out to study the inhibitory activities of compound 36a, 36b, 51a, 51b
and 54. The corresponding p-nitrophenyl glycopyranosides were used as substrates for a-
glucosidase (yeast), B-glucosidase, a-galactosidase, B-galactosidase, a-mannosidase.
Each assay was performed in a phosphate or an acetate buffer at the optimal pH for each
enzyme. Inhibition studies were performed by adding the inhibitor to a ﬁnal
concentration of 0.05 mM to llmM to the respective buffer solutions along with enzyme.
The solutions were incubated at 37°C before adding substrates to the reactions. The
release of product, p-nitrophenyl alcohol was monitored at 400nm. For a-glucosidase
(rice), maltose was used as the substrate, and the assay was based on the glucose
oxidase/peroxidase enzyme procedure (ﬁgure 3.13). The glucose released from maltose
can be oxidized by glucose oxidase to generate D-gluconic acid and hydrogen peroxide.
Under the catalysis of peroxidase, hydrogen peroxide will react with dianisidine to give
the oxidized form with brown color. The absorbance of the solution was determined at

500 nm for oxidized o-dianisidine.

Glucose
Oxidase
D-glucose + H 20 + 02 _—_’ D-gluconic Acid + H202
Peroxidase

H202 + o—Reduced Dianisidine -———> Oxidized o-Dianisidine (Brown)

Figure 3.13. glucose oxidase/peroxidase enzyme mechanism

 

 

/
/
100 - fl" [1] = 6.3 mM
/.
,/
‘ /
/
,-/
80 - /-
/’
,/
/
,/
/.

60- /
>° /
‘ /
'- ,/ . [I] = 2.1 mM

/ ,.-"/
40. / /./' ’
/ , ,
/’ D //./’
D ,// ///’II
// _,.-£/
/ ./
20. 42““ .1.- [I] = 0.6 rnM
.# L..—xx"
./ ., ,/-/':_"_w,,_.r~I-""'
42 1:1—‘””'""- [I] = 0 mM
0 2 4 6 8 10
1 [[A]

Figure 3.14. Lineweaver-Burk plot for the determination of inhibition constant for 36a

Figure 3.14 is the double reciprocal plot of initial rates and substrate concentrations (l/vo

vs 1/[A]) for inhibitor concentrations of 0 mM, 0.6 mM, 2.1 mM, and 6.3mM for

7(S),8(R),9(R),l0(S)-trihydroxy-2-thiaquinolizidine (36a) using yeast a-glucosidase and

p-nitrophenylglucoside as substrate. Figure 3.15 is the plot of slope from ﬁgure 3.14 vs

inhibitor concentration [1] (mM).

123

10-

Q
l

Slope of reciprocal plot

 

 

 

Figure 3.15. Slope from ﬁgure 3.14 vs inhibitor concentration

So the K, value for (Jr-glucosidase (yeast) is

K, = intercept / slope = 0.485 / 1.46 = 0.33 mM
Figure 3.16 and 3.17 are similar plot for 3621 where the enzyme is rice a-glucosidase and
the rate is monitored by a coupled enzyme reaction in which freed glucose from maltose
is oxidized to gluconic acid by glucose oxidase. The concentrations of compound 36a are

0, 0.42 and 8.87 mM, and the inhibition constant is 0.9 mM.

124

 

 

1 l
5) ﬂ [r]=aa'rmu
l/"
4 /
4. x/
//
' ,/
/
3- /
o /
z .
F
2 ' [I]=0.42mM
- r --.-—---/i._,/-/
1 II I] f_,,’-’ ’::: ............ "H
1,“ j :/-—-I”/
d ywyﬂJI/I/“W
O'I'I'I'U'I'U'I'I'I
0 1 2 3 4 5 6 7 8 9

Figure 3.16. Lineweaver-Burk plot for the determination of inhibition constant for 36a

1.2 -
1.0 q l§=0.9 mM
0.8 .
0.6 -
0.4 .

0.2-
1/

0.0

Slope of reciprocal plot

 

O 2 4

['1 (MM)

03-!
on
..A
0

Figure 3.17. slope from figure 3.15 vs inhibitor concentration

For B—glucosidase, a-galactosidase, B-galactosidase, a-mannosidase, no signiﬁcant
inhibition was observed for 36a at this inhibitor concentration range. Compound 36b was
tested against a-glucosidase (yeast), B-glucosidase and a-mannosidase. It showed
opposite inhibition pattern to compound 36a. It only inhibits B-glucosidase with K, value
1 mM without inhibiting (Jr-glucosidase and (rt-mannosidase. Figure 3.18 is the double
reciprocal plot of initial rates and substrate concentrations using yeast a-glucosidase and

p-nitrophenylglucoside as substrate.

     

 

250- [11:5me
200.:
' [I]=2.74II'IM
150. /
'20 //
'- ‘ / [l]=1.37mM
100‘l /
/// m=o.eemu
/ /
’4"
5°“ / [I]=OmM
0' I r I I I I I

 

1I[A]

Figure 3.18. Lineweaver-Burk plot for the determination of inhibition constant for 36h

Compound 51a and 51b were also tested for a-mannosidase (jack beans) and no
inhibition was observed. Compound 54 was tested against a-glucosidase (yeast) and

some inhibition was observed with K; value 1.32 mM (Figure 3.19).

3.4 '-

 

 

3.2 .. /y
d I /
3.0. /’ m=1.097 mM
////
2.8 -1 ’ ./
/-
o /’ ,-”'/.’
'- 2,5 . / ’l _ "I' I
- [Ir-o
2.4 - f "’ ”
X.
2.2- 23'
//,';r
II,
2.0 v I y I v I v I W ' v ' I ' T I I
o 1 2 3 4 5 6 7 8 9
148]

Figure 3.19. Lineweaver-Burk plot for the determination of inhibition constant for 54

Table 3.1 is the inhibition constants (mM) for compounds 36a, 36b, 51a, 51b and 54.
The results indicate that compound 36a is a speciﬁc inhibitor of a-glucosidase. This is
consistent with the study that deoxynojirimycin type inhibitors with nitrogen atom at the
ring oxygen position are more selective for a-glucosidase26’36'38. According to the
stereoelectronic requirements, in a-glycosidases, the positively charged leaving group
and the lone pair of the ring oxygen are positioned antiperiplanar and cooperatively
facilitate the glycosidic bond cleavage. Thus, oxocarbonium ion 56 can be formed
directly, with the positive charge development at the endocyclic oxygen (ﬁgure 3.20).
Under the assumption that deoxynojirimycin type inhibitors including compound 36a
were protonated in the active site, those results suggest that an oxacarbonium ion 56 is
important for enzymes catalyzing axial glycoside cleavage (rt-glucosidase). For B-

glycosidases, the glycosidic bond cleavage cannot receive aid form the lone pair of the

ring oxygen, so the ring could ﬂip to a boat (58) or other conformations (59) to facilitate
the bond cleavage. Substrate distortion is generally the case for B-glycosidases.
Compound 36a with its rigid bicyclic structure, is locked in its trans-fused chair
conformation, so it cannot ﬂip or change to other conformations. Therefore, it showed no
inhibition against B—glucosidase. In contrast, compound 36b showed speciﬁc inhibition
against B-glucosidase while no inhibition for a-glucosidase. This is probably because of
the change of the conformation caused by the inverted stereochemistry of the second ring.
With the axial position of the second ring, conformation 61 is not stable for compound
36b. This whole molecule may ﬂip to conformation 62 with the second ring in equatorial
position while the hydroxyl groups in axial position. This is the kind of conformation
which puts the B-glycosidic bond in axial position to facilitate the cleavage by

stereoelectronic control.

OH OH

36:! 61 62

Table 3.1. Inhibition constants (mM) for compounds 36a, 36b, 51a, 51b and 54

 

enzymes 36a 36b 51a 51b 54
(at-glucosidase (yeast) 0.33 ni nd nd 1.32
(rt-glucosidase (rice) 0.9 nd nd nd nd

 

128

 

 

 

 

 

 

 

 

 

B-glucosidase (almond) ni 1.0 nd nd ni
a-galactosidase ni nd nd nd nd
(green coffee beans)
B-galactosidase (E. Coli) ni nd nd nd nd
(rt-mannosidase (jack beans) ni nd ni nr nd
ni, no inhibition observed in this concentration range
nd, not determined
0 F Gall
W W
(H0)../ -—- moxﬂ
HOR Q n
+ L, + d
55 56
W
"’ ("Ola/‘AR
' as H ' " 0"
by 60
was
(H0) /
N g _ n .1
AA/OR —"" 53
‘"°’" \U
57 k HR

P

Figure 3.20. Stereoelectronic requirements for cleavage of u- and B-glycosidases.

The activities and speciﬁcities of the known aza-bicyclic systems and key monocyclic

systems are shown in table 3.2. Compared to deoxynojirimycin, compound 36a is a

129

relatively weak inhibitor, but it showed speciﬁcity for a-glucosidase. No inhibitory
activity towards B—glucosidase (almond), a-galactosidase (green coffee beans), or-
galactosidase (E-coli) and a-mannosidase (jack bean) was observed. Compared to 6-
deoxy derivatives 63 and 64, compound 36a showed superior activity against a-
glucosidase. One of the major problems with the use of iminosugars and their derivatives
as inhibitors is the lack of specificity. Hence compound 1 in table 3.2 has low K, values
but shows poor speciﬁcity. Compound 36a showed very speciﬁc activity against a-
glucosidase, and 36b showed speciﬁcity against B-glucosidase. The speciﬁcity comes
from the structural rigidity, which prevents 36a ﬁ'om distortion to boat or other
conformations that are important for B-glucosidase. Compound 36b could ﬂip to other
conformations to put the second ring in equatorial position and favor B-glucosidase
binding. Compound 51a showed no activity against o—mannosidase (jack beans). This is
probably because of the rigid structure and the second 6-member ring, which interfere the
hydrogen bonding between 2-hydroxyl group and the enzyme.

Castanospermine is one of the most active bicyclic systems. It showed poor activity
against yeast a-glucosidase but strongly inhibited the rice enzyme. However, it showed
non-selectivity by inhibiting almond B-glucosidase. It has relatively ﬂexible structure
compared to 36a and 36b, and it presented a twisted boat conformation of the 6-member
ring when bounded to Exo-B—(1,3)-g1uconase39. This distortion can not be made for
compound 36a which has a trans-diequatorial type fusion between the rings. Compound
30, which is the slightly ring-expanded version of the potent a-mannosidase inhibitor
swainsonine (3) showed complete loss of inhibition of a-mannosidase (table 3.2). As a

general rule, decalin-type bicyclic systems show much reduced or no inhibitory activity

130

compared to their acyclic or octahydroindene-type analogs. Therefore, one important
conclusion that can be made is that structural ﬂexibility leads to nonspeciﬁc inhibitory
activity. Monocyclic systems generally showed poor speciﬁcity by inhibiting both a- and
B-glycosidsaes because of their ﬂexible structures. Castanospermine (2) and the
thiaquinolizidine described here (36a) are the most impressive of the reported bicyclic
aza-type iminosugar derivatives with a nitrogen atom at the ring junction.
Thiaquinolizidine 363 was superior against and selective for a-glucosidases compared to
Castanospermine. Thiaquinolizidines therefore represent a signiﬁcant advancement in

this area.

Table 3.2. Activities (Ki mM) of the known aza-bicyclic systems and key monocyclic

 

 

 

systems
Enzymes W T
Hoﬂﬂ Ho "" “ﬂ”.
0H HOHO "H m OH
OH OH
122 634° 64 6522
a-glucosidase
Yeast 8.67*10'6 ni ni [4'] 3.69‘104
1.56x10'31221 ni [‘2]
Rice ni
Rat intestinal ni
maltase
Rat intestinal ni
isomaltase
Rat intestinal ni
sucrase .
Type I (calf 1*10'° 7.0210'8
liver)
B-glucosidase
Sweet almond 1.8*10‘5 7.8X10‘4 [32] 25*10'3 [4” 4.3*10‘5

[(350 1.4't10'3
I421

 

131

 

 

 

 

 

 

 

 

 

 

 

 

Human liver 97% inhibition
at conc.1mM
with substrate
0.5mM [‘3]
a-D-
mannosidase
Jack bean 4* 104 ni [4']
ni [42]
“-
galactosidase
Green coffee ni [4']
bean ICso 4.44*10"’
[421
B-D-
galactosidase
Jack bean ni ni [4']
row 3.6"‘10*5
[421
Table 3.2. continued
Enzymes
s
0H
36a 2‘“ 66“5 303'
a-glucosidase
Yeast 3.3*10“‘ >1500uM IC50>2000
Rice 9*10“ 1.51108
Rat intestinal 5.5* 10' '0
sucrase
B-glucosidase
Sweet almond ni l.5"'10'6 ni IC50>2000
Human liver
a-D-
mannosidase
Jack bean ni IC50>2000
a-
galactosidase
Green coffee ni ni
bean

 

ni, no inhibition observed.

132

3.2.3. Antibacterial Activities of Trihydroxy-Z-thiaquinolizidine Derivatives

Compounds 36a, 51a and 54 were tested against ﬁve gram-positives strains S. aureus
43300, S. aureus 29213, E. faecalis 51299, E. faecalis 29212, B. subtilis PY79 and six
gram-negative strains E.aerogenes 49469, E. coli 25922, E.coli DH5 alpha, E. cloacea
49141, P. aeruginosa 27853, Salmonella sp.35664. Among the 11 strains tested, three of
them showed inhibition of growth by these compounds, and they are all gram-positive
strains. All three compounds showed inhibition against S. aureus 43300 (ﬁgure 3.21).
Compound 51a and 54 showed inhibitions against E. faecalis 51299 (ﬁgure 3.22), while
only compound 54 has inhibitory activity against E. faecalis 2921 (ﬁgure 3.23). The
inhibition pattern all showed the initial drop of the reading, followed by slow decrease.
Compound 54 is the most active in the three compounds, and both 54 and 51a can inhibit
E. faecalis 51299 very well. At concentration 0.49 ug/mL, both compounds have over
50% inhibition of the bacterial growth. The 50% inhibition of the other two strains can be
reached at 125 ug/mL or 250 ug/mL. Compared to compound 36a, 54 has a sulfone
functional group, which is a very important factor for its antibacterial activity. Further

modiﬁcation should lead to more active compounds against bacteria.

S. aureus 43300

 

 

 

0.01
0.84
0.4-
o
8 0.3-
( A____-—--"’\ L‘ A A 7t ;
' ' véL/ﬁ
0.2+
. W88
01 *cupdﬂa
' -o-cnpd84
o I I I I I I f I T I I

 

280 128 82.8 31.28 18.83 7.81 3.81 1.88 0.88 0.48 0.0
ugllni uglml uglmi uglrni uglmi uglmi uglmi uglmi uglml uglmi uglml

compd cone.

Figure 3.21. Inhibitory activity against S. aureus 43300

E. faecalis 81288 against compda.

0.14 -

0.12 a

0.1 -

A800

008‘
C

 

 

0.04 ‘

+cmpo 81:
0.02 1 -—cu1po u

 

 

o I I I I I I I I T I I
280 128 82.8 31.28 18.83 7.81 3.81 1.88 0.88 0.48 0.0
uglrni uglml uglmi uglrni uglml uglrni ugli'nl uglmi ugIml uglmi uglrnl
compd cone.

Figure 3.22. Inhibitory activity against E. faecalis 51299

134

E. faecalis 28212 against compound 54

0.18 a
0.15 1
0.14 'l
0.12 n
g 0.1 a v
< 0.05 . W7

0.06 ~

0
1

—o— cmpd 54
0.04 .

0.02 '-

 

250 125 82.5 31.25 15.83 7.81 3.81 1.85 0.88 0.48 0.0
uglmi uglmi uglmi uglml uglmi uglmi uglml uglrni rig/mi uglmi uglml

cone.

Figure 3.23. Inhibitory activity against E. faecalis 292129

3.3. Experimental Section

3.3.1. General Procedures

Melting points were measured. on a Ficher-Johns melting point apparatus. Optical
rotations were measured (1 = 589 nm) at room temperature using a Jasco P-IOlO
polarimeter. IR spectra were recorded on a FT-IR instrument. The 'H (and '3C) NMR
spectra were recorded at 500 (125.5) MHz on a Varian VXR spectrometer. The HRMS
FAB mass spectra were obtained using a Jeol HX-l 10 double-focusing mass

spectrometer operating in positive ion mode.

3.3.2. Inhibition Assay
Inhibitory potency was determined by spectrophotometrically measuring the residual

hydrolytic activities of the glycosidases against the corresponding nitrophenyl a- or B-D-

135

glucopyranoside except (Jr-glucosidase (rice). The glycosidases used were (rt-glucosidase
(yeast), a-glucosidase (rice), B-glucosidase (almond), a-galactosidase (green coffee
beans), B-galactosidase (E. coli) and a-mannosidase (jack beans). All enzymes were
purchased from Sigma. (it-Glucosidase, B-glucosidase assays were performed in 50 mM
phosphate buffer, pH 6.8 at 37°C. a-Galactosidase assays were performed in 50 mM
phosphate buffer, pH 7.3 at 37°C. B-Galactosidase assays were performed in 50 mM
phosphate buffer, pH 5.0 at 37°C. a-Mannosidase assay was performed in 50 mM acetate
buffer, pH 4.5 at 37°C. Inhibition studies were performed by adding the inhibitor to a
ﬁnal concentration of 0.05 mM to 11 mM to the respective buffer solutions along with
enzyme. The solutions were incubated at 37°C before adding substrates to the reactions.
The absorbance of the resulting mixture was determined at 400 nm (for p-nitrophenol).
For (rt-glucosidase (rice), maltose was used as the substrate, and the assay was based on
the glucose oxidase/peroxidase enzyme procedure. The assay was performed in sodium
acetate buffer at pH 4.0 at 37°C. The inhibitor was added to a ﬁnal concentration of 0.4
mM and 8.9 mM to the substrate solution. The enzyme was added to the solution at 37°C,
and the reaction was stopped aﬁer 10 and 30mins by dilute perchloric acid solution.
Pipette the glucose oxidase/peroxidase solution into the reaction mixture, and incubate at
37°C for 30mins. The absorbance of the solution was determined at 500 nm for oxidized

o-dianisidine.

3.3.3. MIC Testing Procedure
Compounds were assayed according to the standard MIC testing procedure for

antimicrobials. Inhibitory potency was determined by spectrophotometrically measuring

136

the growth of the bacteria at 600 nm. The organisms used were ﬁve gram-positives
strains S. aureus 43300, S. aureus 29213, E. faecalis 51299, E. faecalis 29212, B. subtilis
PY79 and six gram-negative strains E.aerogenes 49469, E. coli 25922, E. coli DH5 alpha,
E. cloacea 49141, P. aeruginosa 27853, Salmonella sp.35664. Serial dilutions are made
of the compounds in bacterial growth media M-H media. The concentration range from
250ug/mL, to 0.49ug/mL, Opg/mL. Bacteria were grown to reach 1.5><10° cfu/mL and
then added to the dilutions of the compounds. The solutions were incubated at 37°C and

the grth of the bacteria was monitored at 2, 5, l4, l8, and 24 hours.

Methyl 6-Bromo-6-deoxy-B-D-glucopyranoside (38) To a stirred solution of methyl [3-
D-glucopyranoside 37 (10.15g, 50mmol) in anhydrous pyridine (300mL) at 0°C were
added triphenylphosphine (26.2g, lOOmmol) and carbon tetrabromide (24.87g, 75mmol).
The resulting mixture was protected from moisture and stirred at 0 °C for ten minutes,
then was allowed to warm to 65 °C and was stirred for 4 hours. Methanol (lOmL) was
added to decompose any excess of reagent. The solvent was removed by evaporation and
the residue was puriﬁed by column chromatography (CHzClz, followed by 20:1
CHzClz/MeOH). Crystallization from methanol--hexanes afforded white crystalline solid

(10.96g, 85%), mp. 139-140 °c, lit. mp. 154°C, [(11120D = -27.6° (c 0.22, H20).

Methyl 6-Bromo-6-deoxy-2,3,4-tri-0-pivaloyl-B-D—glucopyranoside (39) Pivaloylation
of 38 (6.73g, 26mmol) by trimethylacetyl chloride (28.8mL, 32.4mmol) in pyridine
(300mL) at room temperature for 2 days afforded an white solid 8 (11.2g, 84%), mp.

109-110°C, [11120.3 = -2.4° (c 0.31, CHC13). IR (CH3CI) vm 2971.7, 1745.6, 1140.9 cm".

137

lH NMR (500MHz, CDC13) 8 5.29 (1H, t, J=9.5 Hz), 4.99 (2H, t, J=9.7 Hz), 4.42 (1H, d,
J=8.0 Hz), 3.70 (1H, m), 3.50 (3H, s), 3.39-3.12 (2H, m), 1.14 (9H, s), 1.13 (9H, s), 1.08
(9H, s); 13’C NMR (125.5MHz, CDC13) 8 177.14, 176.63, 176.51, 101.36, 73.7, 71.9, 71.2,
70.8, 57.1, 38.8, 38.7, 30.6 ppm; HR-FABMS (M + H+) Calcd. 509.1750, found

509.1736.

6-Bromo-6-deoxy-2,3,4-tri-O-pivaloyl-D-xylo-S-ulosonicacid methylester (40). To a
solution of 39 (1g, 1.96mmol) in acetic acid (100m1.) and acetic anhydride (lOmL),
chromium trioxide (1.18g, 11.8mmol) was added and the suspension was stirred at room
temperature for 3 hours. The mixture was then poured slowly into cold water (500mL).
The water was extracted 5 times with CHzClz and the combined organic phase was
washed with brine, saturated sodium bicarbonate and dried (N32SO4), concentrated. The
resulting residue was passed through a small pad of silica gel to give 9 as a colorless oil
(lg, 97%), [a]2°p = +36.5° (c 0.12 CHC13), IR (CHzClz) vmax 2975.85, 1743.63, 1132.00
cm". '11 NMR (500MHz, CDC13) 5 5.72, (1H, t, J=4.8 Hz), 5.57 (1H, d, J=4.5 Hz), 5.23
(1H, d, J=5.0 Hz), 4.12 (1H, d, J=14.0 Hz), 4.01 (1H, d, J=13.5 Hz), 3.72 (3H, s), 1.25
(9H, s), 1.21 (9H, s), 1.18 (9H, 3); '3C NMR (125.5MHz, CDC13) 6 194.5, 177.1, 176.9,
176.8, 167.1, 72.9, 70.2, 69.4, 52.7, 38.9, 38.8, 38.7, 31.6, 27.0, 26.9 ppm; HRFABMS

(M + W) Calcd. 523.1543, found 523.1530.
7(S),8(R),9(R),10(S)-Trihydroxy-2-thiaquinolizidin-6-one (42a) and

7(S),8(R),9(R),10(R)—Trihydroxy-2-thiaquinolizidin-6-one (42b) A solution of 40 (7g,

13.4mmol) and IIS(CH2)2NH2 (1.24g, 16.1mmol) in methanol (250 mL) was stirred at

138

room temperature for one hour, followed by addition of sodium cyanoboron hydride
(1.26g, 20mmol). The reaction mixture was stirred overnight and sodium carbonate was
added to facilitate the lactam cyclization. Aﬁer stirred for several hours, the suspension
was ﬁltered and acetic acid (2mL) was added and concentrated. The residue was puriﬁed
by column chromatography (10:1 Hexanes/Acetone) to yield two lactam diastereomers
42a and 42b (4.64g, 73.6%), the ratio is 2.5 : l.

Lactam 42a (3.31 g, 52.6%) was given as a white solid, mp. 188-190 °C, [a]2°p = +12.6°
(c 0.1 CHClg), IR (CHCI3) vm, 1744.54, 1685.34 cm“. 'H NMR (500MHz, CDC13) 5
5.53 (1H, t, J=10.5 Hz), 5.30 (1H, d, J=l 1.0 Hz), 5.19 (1H, dd, J=10.5, 8 Hz), 4.94 (1H,
dt, J=13.5, 3 Hz), 3.53 (1H, ddd, J=9.8, 8.4, 3.4 Hz), 2.87 (1H, td, J=14.3, 2.5 Hz), 2.66
(1H, td, J=13.0, 3.0 Hz), 2.61-2.49 (3H. m): 1.21 (9H, s), 1.17 (9H, s), 1.12 (9H, 5); '3C
NMR (125.5MHz, CDC13) 8 177.4, 177.1, 176.6, 164.3, 70.4, 69.4, 67.8, 60.2, 44.7, 38.9,
38.7, 31.8, 27.1, 26.6 ppm. HRFABMA (M + H“) calcd. 472.2369, found 472.2379.
Lactam 42h (1.33g, 21.0%) was given as a white solid, mp. 179-181°C. 1R (CH2C12)vma,,
1741.07, 1679.15, 1137.70 cm"; 'H NMR (500MHz, CDC13) 8 5.76 (1H, t, J=10.3 Hz),
5.27 (1H, dd, J=11.5, 6.3 Hz), 4.78 (1H, m), 4.02 (1H, ddd, J=1 1.8, 6.3, 2.0 Hz), 3.07
(1H, t, 12.3 Hz), 2.98-2.88 (3H, m), 2.50 (1H, d, J=13.0 Hz), 2.35 (1H, m), 1.20 (9H, s),
1.17 (9H, s), 1.14 (9H, s); 13C NMR (125.5MHz, CDC13) 8 177.42, 176.82, 163.67, 67.66,
67.03, 59.23, 47.05, 38.90, 38.69, 27.74, 27.11, 27.06, 26.96, 26.34 ppm. FABMS (M +

H“) calcd. 472.2369, found 472.2.

7(S),8(R),9(R),1 0(S)-Trihydroxy-Z-thiaquinolizidine (363)

139

A solution of lactam 42a (2g, 4.24mmol) and BH3-THF (20mL, 1.5M) in anhydrous THF

(30mL) was reﬂuxed for 4 hours and the TLC and NMR showed the completion of the
reduction. The solvent was removed and methanol was added and concentrated for 3

times. The residue was dissolved in methanol (30mL), followed by addition of NaOMe
(0.15g, 2.8mmol). The reaction was stirred for 8 hours and concentrated. The residue was
applied to an ion exchange column (Dowex 50WX8-400, 30g), which was washed with
water (50 mL) and eluted with NH40H (50mL). The elution was concentrated and
puriﬁed by column chromatography ( 15:1 CH2C12/MeOH) to afford a white solid (0.62g,
71% ), mp. 235-237 °C; [@1200 = + 20.2° (c 0.06 H20); IR (KBr) v“m 3355.78, 3275.61

cm"; 'H NMR (500MHz, cock) 8 3.50 (1H, ddd, J=11.0, 9.1, 4.9 Hz), 3.25 (1H, t,
J=9.3 Hz), 3.12 (1H, dt, J=12.5, 3.0 Hz), 3.06 (1H, t, J=9.5 Hz), 2.93 (1H, dt, J=14.0, 2.5

Hz), 2.84 (1H, dd, J=11.5, 5.0 Hz), 2.75 (1H, td, J=13.0, 3.0 Hz), 2.52 (1H, m), 2.45 (1H,
t, J=12.3 Hz), 2.43 (1H, m), 2.19 (1H, t, J=l 1.3 Hz), 2.13 (1H, td, J=10.0, 2.5 Hz); l3C
NMR (125.5MHz, CDCl3) 8 77.9, 74.1, 68.5, 65.6, 59.4, 55.7, 29.3, 26.3 ppm. HR-

FABMS (M + H+) calcd. 206.0851, found 206.0849.

7(S),8(R),9(R),10(R)-Trihydroxy-2-thiaquinolizidine (361)) was obtained by the same
method as 36a from lactam 428 (75% from 42b). 'H NMR (500MHz, 020) 8 3.51-2.42
(2H, m), 3.06 (2H, d, J=10.0 Hz), 2.90 (4H, dd, J=26.0, 12.5 Hz), 2.82 (2H, t, J=10.5 Hz),
2.60 (2H, dd, 1:120, 4 Hz). 2.20 (2H, d, .I=14.0 Hz). HR-FABMS (M + H‘) calcd.

206.0851, found 206.0849.

Manno-derivatives were obtained in the same fashion as gluco-derivatives.

140

Methyl 6-Bromo-6-deoxy-B—D-mannopyranoside (45) was given as white solid (84%).
lH NMR (500MHz, D20) 8 3.91 (1H, d, J=2 Hz), 3.75 (1H, d, J=2.5 Hz), 3.73 (1H, d,
J=2.5 Hz), 3.571 (2H, dd, J=11.8, 5.8 Hz), 3.569 (1H, d, J=1.5 Hz), 3.46 (3H, s), 3.43

(1H, m); 13c NMR (125.5MHz, 1520) 8 101.41, 74.89, 72.83, 70.47, 68.96, 57.24, 32.95

Methyl 6-Bromo-6-deoxy-2,3,4-tri-O-pivaloyl-B-D-mannopyranoside (46) was
obtained as white solid (80%). 'H NMR (500MHz, CDCl3) 8 5.41 (1H, dd, J=3, 0.75 Hz),
5.24 (1H, t, J=10 Hz), 5.08 (1H, dd, J=10, 3 Hz), 4.57 (1H, d, J=1 Hz), 3.68 (1H, m),
3.49 (3H, s), 3.47-3.39 (2H. m), 1.24 (9H, s), 1.15 (9H, s), 1.09 (9H, s); l3C NMR
(125.5MHz, CDCl3) 8 177.28, 177.22, 176.80, 99.74, 73.82, 70.85, 68.31, 68.21, 57.18,
39.06, 38.84, 38.75, 31.08, 27.16, 27.04, 27.02 ppm. HRFABMA (M + H+) Calcd.

509.1750, found 509.1748.

Methyl 6-Bromo-2,3,4-tri-0-pivaloyI-S-keto-ester (47) was obtained as colorless oil
(91%). lH NMR (500MHz, CDC13) 8 5.73 (1H, dd, J=8.5, 3 Hz), 5.71 (1H, d, J=2.5 Hz),
5.01 (1H, d, J=9 Hz), 4.13 (ll-I, d, J=14 Hz), 4.00 (1H, d, J=13.5 Hz) 3.69 (3H, s), 1.24
(9H, s), 1.21 (9H, s), 1.15 (9H, s); ”C NMR (125.5MHz, CDCl3) 8 194.78, 176.90,
176.85, 176.59, 167.76, 73.16, 68.99. 68.85, 52.75, 38.91, 38.81, 38.69, 38.91, 38.81,
38.69, 31.38, 26.96, 26.86, 26.84 ppm. HRFABMA (M -'r H+) Calcd. 523.1543, found

523.1545.

141

Lactam 49a was obtained as white solid (65%). 1H NMR (500MHz, CDCl3) 8 5.67 (1H,
d, J=3.5 Hz), 5.39 (1H, m), 5.01 (1H, m), 4.99 (1H, t, J=3 Hz), 3.66 (1H, dt, J=11.5, 2
Hz), 2.99 (1H, m), 2.85 (1H, td, J=13, 2 Hz), 2.74 (1H, td, J=13, 2.5 Hz), 2.59 (1H, dt,
J=13.5, 2 Hz), 2.46 (1H, m), 1.22 (9H, s), 1.21 (9H, s), 1.20 (9H, s); 13C NMR
(125.5MHz, CDCl3) 8 176.94, 176.86, 176.36, 163.27, 69.43, 68.18, 66.51, 62.75, 46.54,
38.91, 38.85, 38.81, 31.06, 27.18, 27.08, 26.98, 26.66 ppm. HRFABMA (M + Hi) calcd.
472.2369, found 472.2366.

Lactam 49b was obtained as white solid (16%). 'H NMR (500MHz, cock) 8 5.65 (1H,
d, J=2.5 Hz), 5.34 (1H, t, J=5 Hz), 5.31 (1H, m), 5.04 (1H, dt, J=8.5, 3 Hz), 3.77 (1H, m),
2.79 (1H, td, J=13.5, 2 Hz), 2.68 (2H, m), 2.53 (1H, m), 2.39 (1H, dt, J=13.5, 2 Hz), 1.25
(9H, s), 1.20 (9H, s), 1.18 (9H, s); l3C NMR (125.5MHz, CDCl3) 8 177.12, 176.66,
176.23, 165.11, 67.13, 66.90, 66.34, 58.60, 44.57, 39.14, 38.94, 38.85, 28.78, 27.19,

27.11, 27.04 ppm. HRFABMA (M + 11+) calcd. 472.2369, found 472.2367.

7(R),8(R),9(R),10(S)-Trihydroxy-2-thiaquinolizidine (51a) was obtained as white solid
(73% from 498). lH NMR. (500MHz, D20) 8 3.93 (1H, m), 3.43 (1H, dd, J=10, 3.5 Hz),
3.35 (1H, t, J=9.5 Hz), 3.06 (1H, dt, J=12.5. 3 Hz), 2.90 (1H, dt, J=8.5, 2.5 Hz), 2.81-
2.74 (2H, m), 2.51-2.44 (2H, m), 2.39-2.32 (2H, m), 2.05 (1H, t, J=9 Hz); 13C NMR
(125.5MHz, D20) 8 73.90, 71.15, 67.48, 66.24, 59.37, 55.87, 28.77, 26.02 ppm.

HRFABMA (M + W) calcd. 206.0851, found 206.0851.

7(R),8(R),9(R),10(R)-Trihydroxy-2-thiaquinolizidine (51b) was obtained by the same

way as 5111 (71% from 49b). 'H NMR (500MHz, 020) 8 3.97 (1H, m), 3.87 (1H, s), 3.69

142

(1H, d, J=2.0 Hz), 3.10 (1H, d, ./=12.5 Hz), 2.84 (2H, dd, J=13.5, 11.0 Hz), 2.76 (1H, m),
2.61 (2H, m), 2.44 (2H, t, J=11.0 Hz), 2.36 (1H, d, J=14.0 Hz). HRFABMA (M + H*)

calcd. 206.0851, found 206.0850.

2,2-dioxy-7(S),8(R),9(R),10(R)-Trihydroxy-2-thiaquinolizidin-Z-one (sulfone 52)

Lactam 42a (0.5g, 1.1mmol) was dissolved in 30mL dichloromethane, and cooled to 0°C.
SmL dichloromethane solution of m-chloroperbenzoic acid (0.65g, 70% w/w) was added
to the reaction in several portions. After stirred at room temperature for half an hour,
saturated sodium bicarbonate solution was added to neutralize the reaction. The product
was extracted by dichloromethane for 3 times and the combined organic phase was
washed with saturated sodium bicarbonate, brine and dried (Na2804). The resulting
sulfone 2 was obtained as white solid and was used to the next step without puriﬁcation.
lH NMR (500MHz, CDCl3) 8 5.59 (1H, t, J=10 Hz), 5.41 (1H, d, J=10.5 Hz), 5.28 (1H,
m), 4.99 (1H, dt, J=10.5, 3.5 Hz), 3.89 (1H, m), 3.23-3.01 (5H, m), 1.19 (9H, s), 1.17 (9H,
s), 1.10 (9H, s); 13C NMR (125.5MHz, CDCl3) 8 177.54, 177.09, 176.35, 164.22, 69.13,

68.97, 67.53, 57.06, 54.58, 50.03, 40.41, 38.81, 38.79, 38.68, 27.03, 26.97 ppm.

2,2-Dioxy-7(S’),8(R),9(R),10(S)-Trihydroxy-2-thiaquinolizidin-Z-one (sulfone S4)

A solution of sulfone 52 (0.40g, 0.08nunol) and BH3-THF (5.0mL, 7.5mmol) in
anhydrous THF (lOmL) was reﬂuxed for 8 hours and NMR showed the completion of the
reduction. The solvent was removed and methanol was added and concentrated for 3
times. The residue was dissolved in methanol (ISmL), followed by addition of NaOMe

(0.10g, 1.8mmol). The reaction was stirred for overnight and concentrated. The residue

143

was applied to an ion exchange column (Dowex 50WX8-400, 2g), which was washed
with water (30 mL) and eluted with 2N NH40H (30mL). The elution was concentrated to
afford a white solid (0.15g, 84%). lH NMR (500MHz, D20) 8 3.56 (1H, m), 3.53 (1H, m),
3.30-3.23 (4H, m), 3.14 (1H, t, J=9.5 Hz), 3.01 (2H, m), 2.85 (1H, m), 2.58 (1H, m), 2.27
(1H, t, J=11.5 Hz); 13C NMR (125.5MHz, D20) 8 76.88, 73.11, 68.50, 63.03, 57.99,

52.25, 52.06, 49.98 ppm. HRFABMA (M + W) calcd. 238.0748, found 238.0748.

144

3.4. References

(1) Vandenbroek, L.; Vermaas, D. J.; Heskamp, B. M.; Vanboeckel, C. A. A.; Tan, M.;
Bolscher, J. G. M.; Ploegh, H. L.; Vankemenade, F. J.; Degoede, R. E. Y.; Miedema, F.
Chemical Modiﬁcation of Azasugars, Inhibitors of N-Glycoprotein-Processing
Glycosidases and of Hiv-1 Infection - Review and Structure-Activity-Relationships
Recueil Des Travaux Chimiques Des Pays-Bas-Journal of the Royal Netherlands
Chemical Society 1993, 112, 82-94.

(2) Scheen, A. J. Drug treatment of non-insulin-dependent diabetes mellitus in the 19905
- Achievements and future developments Drugs 1997, 54, 355-368.

(3) Witczak, Z. J. In Carbohydrates as New and Old Targets for Future Drug Design in
Carbohydrates in Drug Design; Witczak, Z. J ., Ed.; Marcel Dekker Inc.: New York, 1997,

pl.

(4) Anzeveno, P. B.; Creerner, L. J.; Daniel, J. K.; King, C. H. R.; Liu, P. S. A Facile,
Practical Synthesis of 2,6-Dideoxy-2,6-Irnino-7-0-Beta-D-G1ucopyranosy1-D-Glycero-L-
Gulo-Heptitol (Mdl 25,637) J. Org. Chem. 1989, 54, 2539-2542.

(5) Platt, F. M.; Neises, G. R.; Reinkensmeier, G.; Townsend, M. J.; Perry, V. H.; Proia,
R. L.; Winchester, B.; Dwek, R. A.; Butters, T. D. Prevention of lysosomal storage in
Tay-Sachs mice treated with N-butyldeoxynojirimycin Science 1997, 276, 428-431.

(6) Truscheit, E.; Frommer, W.; Junge, B.; Muller, L.; Schmidt, D. D.; Wingender, W.
Chemistry and Biochemistry of Microbial Alpha-Glucosidase Inhibitors Angew. Chem,
Int. Ed. Engl. in English 1981, 20, 744-761.

(7) Karpas, A.; Fleet, G. W. J.; Dwek, R. A.; Petursson, S.; Namgoong, S. K.; Ramsden,
N. G.; Jacob, G. S.; Rademacher, T. W. Aminosugar Derivatives as Potential Anti-
Human Immunodeﬁciency Virus Agents Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 9229-
9233.

(8) Taylor, D. L.; Sunkara, P. S.; Liu, P. S.; Kang, M. S.; Bowlin, T. L.; Tyms, A. S. 6-0-
Butanoylcastanosperrnine (Mdl 28,574) Inhibits Glycoprotein Processing and the Growth
of Hiv Aids 1991, 5, 693-698.

(9) Fleet, G. W. J.; Karpas, A.; Dwek, R. A.; Fellows, L. E.; Tyms, A. S.; Petursson, S.;
Namgoong, S. K.; Ramsden, N. G.; Smith, P. W.; Son, J. C.; Wilson, F.; Witty, D. R.;
Jacob, G. S.; Rademacher, T. W. hrhibition of Hiv Replication by Amino-Sugar
Derivatives F ebs Letters 1988, 23 7, 128-132.

(10) Goss, P. E.; Baker, M. A.; Carver, J. P._: Dennis, J. W. Inhibitors of carbohydrate
processing: A new class of anticancer agents Clinical Cancer Research 1995, I, 935-944.

145

(11) Kino, T.; lnamura, N.; Nakahara, K.; Kiyoto, S.; Goto, T.; Terano, H.; Kohsaka, M.;
Aoki, H.; Imanaka, H. Studies of an Immunomodulator, Swainsonine .2. Effect of
Swainsonine on Mouse Immunodeﬁcient System and Experimental Murine Tumor
Journal of Antibiotics 1985, 38, 936-940.

( 12) Elbein, A. D. Glycosidase Inhibitors - Inhibitors of N-Linked Oligosaccharide
Processing Faseb J. 1991, 5, 3055-3063.

(13) Das, P. C.; Roberts, J. D.; White, S. L.; Olden, K. Activation of Resident Tissue-
Speciﬁc Macrophages by Swainsonine Oncology Research 1995, 7, 425-433.

(14) Elbein, A. D. M., R. J. In Iminosugars as Glycosidase Inhibitors; Stutz, A. E., Ed.;
Wiley-VCH: Weinheim, 1999, pp 216-251 and references therein.

(15) Molyneux, R. J .; McKenzie, R. A.; Osullivan, B. M.; Elbein, A. D. Identiﬁcation of
the Glycosidase Inhibitors Swainsonine and Calystegine B-2 in Weir Vine (Ipomoea Sp
Q6 Aff Calobra ) and Correlation with Toxicity J. Nat. Prod. -Lloydia 1995, 58, 878-886.

(16) Ermert, P.; Vasella, A. Synthesis of a Glucose-Derived Tetrazole as a New Beta-
Glucosidase Inhibitor - a New Synthesis of l-Deoxynojirimycin Helv. Chim. Acta 1991,
74, 2043-2053.

(17) Brandstetter, T. W.; Davis, 8.; Hyett, D.; Smith, C.; Hackett, L.; Winchester, B. G.;
Fleet, G. W. J. Tetrazoles of Manno-Pyranoses and Rhamno-Pyranoses - Inhibition of
Glycosidases by Tetrazoles and Other Mannose Mimics Tetrahedron Lett. 1995, 36,
7511-7514.

(18) Igarashi, Y. 1., M.; Ichikawa, Y. Synthesis of a new inhibitor of -fucosidase Bioorg.
Med. Chem. Lett. 1996, 6, 553-558.

(19) Legler, G.; Stutz, A. E.; Immich, H. Synthesis of l,5-Dideoxy-1,5-Imino-D-
Arabinitol (5-nor-L-Fuco-l-Deoxynojirimycin) and Its Application for the Afﬁnity
Puriﬁcation and Characterization of Alpha-L-Fucosidase Carbohydr. Res. 1995, 272, 17-
30.

(20) Takahashi, S. K., H. Syntheses of L-fucopyranose and its homologs with ring
heteroatoms other than oxygen. Stereocontrolled conversion of a common D-
arabinofuranoside intermediate Chem. Lett. 1992, 21-24.

(21) Defoin, A. S., H.; Streith, J. Synthesis of 1,6-dideoxynojirimycin, 1,6-dideoxy-D-
allo-nojirimycin, and 1,6-dideoxy-D-gum-nojirimycin via asymmetric hetero-Diels-Alder

reactions Helv. Chim. Acta. 1996, 79, 560-567.

(22) Kajimoto, T.; Liu, K. K. C.; Pede'rson, R. L.; Zhong, Z. Y.; Ichikawa, Y.; Porco, J.
A.; Wong, C. H. Enzyme-Catalyzed Aldol Condensation for Asymmetric-Synthesis of

146

Azasugars - Synthesis, Evaluation, and Modeling of Glycosidase Inhibitors J. Am. Chem.
Soc. 1991, 113, 6187-6196.

(23) Kajimoto, T.; Chen, L. R.; Liu, K. K. C.; Wong, C. H. Palladium-Mediated
Stereocontrolled Reductive Amination of Azido Sugars Prepared from Enzymatic Aldol
Condensation - a General-Approach to the Synthesis of Deoxy Aza Sugars J. Am. Chem.
Soc. 1991, 113, 6678-6680.

(24) Wong, C. H.; Ichikawa, Y.; Krach, T.; Gautheronlenarvor, C.; Dumas, D. P.; Look,
G. C. Probing the Acceptor Speciﬁcity of Beta-l,4-Galactosyltransferase for the
Development of Enzymatic-Synthesis of Novel Oligosaccharides J. Am. Chem. Soc. 1991,
113, 8137-8145.

(25) Liu, K. K. C.; Kajimoto, T.; Chen, L. R.; Zhong, Z. Y.; Ichikawa, Y.; Wong, C. H.
Use of Dihydroxyacetone Phosphate Dependent Aldolases in the Synthesis of
Deoxyazasugars J. Org. Chem. 1991, 56, 6280-6289.

(26) Ganem, B. Inhibitors of carbohydrate-processing enzymes: Design and synthesis of
sugar-shaped heterocycles Acc. Chem. Res. 1996, 29, 340-347.

(27) Baxter, E. W.; Reitz, A. B. Expeditions Synthesis of Azasugars by the Double
Reductive Amination of Dicarbonyl Sugars J. Org. Chem. 1994, 59, 3175-3185.

(28) Dhavale, D. D.; Saha, N. N.; Desai, V. N. A stereoselective synthesis of 1,6-
dideoxynojirimycin by double-reductive amination of dicarbonyl sugar J. Org. Chem.
1997, 62, 7482-7484.

(29) Zhao, H.; Hans, S.; Cheng, X. H.; Mootoo, D. R. Allylated monosaccharides as
precursors in triple reductive amination strategies: Synthesis of castanospermine and
swainsonineJ. Org. Chem. 2001, 66, 1761-1767.

(30) Iida, H.; Yamazaki, N.; Kibayashi, C. Total Synthesis of (+)-Nojirimycin and (+)-1-
Deoxynojirimycin J. Org. Chem. 1987, 52, 3337-3342.

(31) Pearson, W. H.; Hembre, E. J. Synthesis of tetrahydroxyquinolizidines: Ring-
expanded analogs of the mannosidase inhibitor swainsonine J. Org. Chem. 1996, 61,
5537-5545.

(32) Angyal, S. J. J., K. Oxidative demethylation with chromium trioxide in acetic acid
Carbohydr. Res. 1970, 12, 147-149.

(33) Angyal, S. J.; James, K. Oxidation of Carbohydrates with Chromium Trioxide in
Acetic Acid .1. Glycosides Australian Journal of Chemistry 1970, 23, 1209-&.

(34) Mathews, C. K., Van Holde K..E. In Biochemistry; Benjamin/Cummings Pub. Co.,
Inc.: Menlo Park, Calif, 1996., pp 375-388.

147

(35) Halvorson, H. 0. Methods Enzymol. 1966, 8, 55.

(36) Gijsen, H. J. M.; Qiao, L.; Fitz, W.; Wong, C. H. Recent advances in the
chemoenzyrnatic synthesis of carbohydrates and carbohydrate mimetics Chem. Rev. 1996,
96, 443-473.

(37) Wong, C. H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Enzymes in Organic-
Synthesis - Application to the Problems of Carbohydrate-Recognition .1 Angew. Chem,
Int. Ed. Engl. 1995, 34, 412-432.

(38) Wong, C. H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Enzymes in Organic-
Synthesis - Application to the Problems of Carbohydrate-Recognition .2 Angew. Chem,
Int. Ed. Engl. 1995, 34, 521-546.

(39) Cutﬁeld, S. M.; Davies, G. J.; Murshudov, G.; Anderson, B. F.; Moody, P. C. E.;
Sullivan, P. A.; Cutﬁeld, J. F. The structure of the exo- -(1,3)-glucanase from Candida
albicans in native and bound forms: relationship between a pocket and groove in family 5
glycosyl hydrolases] J. Mol. Biol. 1999, 294, 771-783.

(40) Yasuda, K. K., H.; Yamashita, T.; Kameda, Y.; Kato, A.; Nash, R. J.; Fleet, G. W. J.;
Molyneux, R. J .; Asano, N. New Sugar-Mimic Alkaloids from the Pods of Angylocalyx
pynaertii J. Nat. Prod. 2002, 65, 198-202.

(41) Bemotas, R. C. P., G.; Urbach, J.; 0anem, B. A new family of ﬁve-carbon
iminoalditols which are potent glycosidase inhibitors Tetrahedron Lett. 1991, 31, 3393-
3396.

(42) Patil, N. T. J., S.; Sabharwal, S. G.; Dhavale, D. D. l-Aza-sugars from -glucose.
Preparation of 1-deoxy-5-dehydroxymethyl-nojirimycin, its analogues and evaluation of
glycosidase inhibitory activity Bioorg. Med. Chem. 2002, 10, 2155-2160.

(43) Godskesen, M. L., 1.; Madsen, R.; Winchester, B. Deoxyiminoalditols from
aldonolactones — V. Preparation of the four stereoisomers of 1,5-dideoxy-1,5-
iminopentitols. Evaluation of these iminopentitols and three 1,5-dideoxy-1,5-
iminoheptitols as glycosidase inhibitors Bioorg. Med. Chem. 1996, 6, 1857-1865.

(44) Legler, G. Glycoside hydrolases: Mechanistic information from studies with
reversible and irreversible inhibitors Adv. Carbohydr. Chem. Biochem. 1990, 48, 319-385.

(45) van Hooft, P. A. V. L., R. E. J. N.; van der Mare], G. A.; van Boeckel, C. A. A.; van

Boom, J. H. Stereoselective Transformations on D-Glucose-Derived Eight-Membered
Ring Carbocycles Org. Lett. 2001, 3, 731-733.

148

Chapter 4

Design and Synthesis of Iminopentitol Scaffolds for the
Preparation of Riboside Hydrolase, Phosphorylase and

Transferase Inhibitors

149

ABSTRACT

Two bicyclic molecular scaffolds containing the structural essence of a ribosyl cation in
the form of a l,4-dideoxy-l,4-iminopentitol (or pentose “aza-sugar”) have been prepared.
These molecules, with a carboxymethyl substituted dihydroxyhexahydro-pyrrolothiazine
or pyrrolobenzothiazine contain provision to facilitate the attachment of analogs of
departing and incoming groups at the anomeric position with the correct orientation. This
would allow the facile preparation of glycosyl transferase and glycosidase inhibitors
covering several families of enzymes that are important drug targets for a variety of

disease treatment applications.

150

4.1. Introduction

Iminopentitols, especially those with the D-ribo conﬁguration, are extremely valuable for
the development of glycosidase and glycosyltransferase inhibitors across the entire
spectrum of possible applications. They are literal mimetics of furanosidic oxocarbenium
ion species and they closely mimic the transition states of pyranosidic species. Glycosidic
linkages in nucleosides and nucleotides are involved in all kinds of important processes.
The ribooxocarbenium species are key features for N-glycanases including nucleoside

3, hypoxanthine-guanine phosphoribosyltransferases4 (HGPRTases) and

hydrolases"
purine nucleoside phosphorylasess'8 (PNP). Examples of ribo iminopentitols that have

been successfully used as inhibitors are immucillin—H (”9,10 and its derivatives.

H H
H i 7) -. H ‘11 ,2 ‘\
,N, ~ N" ,N, “ .- . .33" NH
K, \I/ \». \/ \/ \ :7“ (I
\ / N \ l N \
Ask ‘ i‘ "Hz
HO OH HO ’OH
ImmH 1. ImmG 2

Naturally occurring iminopentitols, that have the D-ribo (3) as well as L and D-arabino

conﬁgurations (4 and 5 respectively) have been isolated from several sourcesI "'4.

151

0'" H OH H OH H
, ,N.‘ /N. .
‘7 W‘ ‘2 K1 2
Ho OI-I HoI on HO‘ OH
3 4 5
D-ribo-l,4-dideoxy- L-arabino-l,4—dideoxy- D-arabino-l,4-dideoxy-
1,4-iminopentitol l,4-1minopent1tol l,4-iminopent1tol

Some synthetic schemes have been developed for the preparation of the D-ribo
iminopentitols. Examples are described below:

George W.J. Fleetl5 ﬁrst reported the synthesis of 1,4-imino-D-ribitol 3 from sugar
lactone 6. D-gulonolactone was converted to pyrrolidine by reduction of the lactone to
diol, followed by recyclization with benzylamine. Periodate oxidation of the diol 10 and

reduction of the resulting aldehyde afforded the imino-D-ribitol (ﬁgure 4.1).

 

 

 

OH OH: .
° i °“ U”
8H 8H —-"‘ o><o .19... °><6 4;.
——0H __Q __
——0H —-o>< _ >< X
6 7

-" 219-10

10 11 12 3

0""
a:

I
0.811..

 

Figure 4.1. Synthesis of l,4-imino-D-ribitol 3.i ,acetone, dimethoxypropane, 85%;ii

152

LAH, 87%; iii, Methanesulphonyl chloride; iv, BnNH2, 77%; v, 80% acetic acid, 93%;
vi, a, H104; b, NaBH4, 71%; vii, H+, 86%; viii, H2, 78%.

In the synthesis of l,4-dideoxy-1,4-imino-L-lyxitol (ﬁgure 4.2) and 1,5-dideoxy-1,5-
imino-D-ribitol (ﬁgure.4.3) reported by K. H. Park'°, a partially protected chiral 3,4,5-
trihydroxy-l-pentene (13) was used as the starting material. The diastereoselective
epoxidation of this chiral synthon afforded the key intermediate (15). Conversion of the
epoxide to an amino alcohol via an azide, hydroxyl activation and cyclization yielded the

desired products.

"a
OMOM 5mg"
13 i4 15
MOMC; puou HQ, .011
0M8 OMOM . 3. . .e' '- 4'
OM VII VIII
TBS 2 N3 , N M.,/0733 ' Z N xmy/OH
.-. H
8M0M H
16 17 18

Figure 4.2. Synthesis of 1,4-dideoxy-1,4-imino-L-lyxitol. i, mCPBA, CH2CI2; ii, NaN3,
NH4C1, MeOH/H20 (82'1), reﬂux, 12 h; iii, MOMCl. N -ethyldiisopropylamine, CH2C12, 0
°C - rt, 12 h; iv, 70% ACOH, rt, 8 h; v, TBDMSCl, Imidazole, DMF, rt, 10 min; vi, MsCl,
Et3N, THF, 0 °C, 10 min; vii, H2, 10% Pd/C, MeOH, 0.5 N NaOH, rt, 10 h; viii, Dowex
50W-X8, MeOH, reﬂux, 12 h.

153

an;

H i ouou
0H ouou H MON! 11 Q’s. .8“
HOMNa —l_-. M8 N3 _‘
i i a
anon oven
19 20 21
1111
9111011
“$630110"
N
H
22

Figure 4.3. Synthesis of 1,5-dideoxy-l,5-imino-D-ribitol. i, MsCl, Et3N, CH2Cl2, -40 °C,
20 min; vii, H2, Pd/C, MeOH, 0.5 N NaOH, rt, 9 h; viii, Dowex 50W-X8, MeOH, reﬂux,

12 h.

In the synthesis reported by Hassner, The asymmetric synthesis of l,4-dideoxy-l,4-

imino-D-ribitol was achieved utilizing the stereoselective addition of allylphenylsulfone

to a chiral N-sulﬁnyl-Z-furfuryl imine and ring-closing metathesisl7 (ﬁgure 4.4).

154

 

p-Toi N Fur-2
Fur-2 H2N Fur-2
23 24 25
\ $02911 iv \ v Q v1
_> —> N Fur-2 _"
\/\
”2" my, i-iN Fur-2 A”
26 28
>< _.><
H9, .9" 0. .9 9. .9 H9, .9"
’ viill
Ci'
Fur-2000140 000809 N
I 1 "2’
80¢ Boc Boc 0"
29 30 31 32

Figure 4.4. Synthesis of 1,4-dideoxy-1,4-imino-D-ribitol. i, allylphenylsulfone, LDA,
THF, -100 °C; ii, TFA, MeOH, 0°C; iii, t-BuOH, THF, 0°C; iv, (a) allylbromide, K2CO3;
(b) Sml2-THF, HMPA, -20°C; v, (a) N-TBCBT; (b) [C12(Pcy3)2Ru=CHPh]; vi, Cat. Os04,
NMO, t-BuOH; vii, (a) 2,2-DMP, Cat.p-TSA; (b) Cat. Ru02, NaIO4; (c) CH2N2; viii,
DIBAL-H, -78°C; viiii, (a) 80% TFA, (b) aq HCl.

The synthesis of han and its derivatives started from imine 331°, derived from 1,4-
dideoxy-l,4—imino-D-ribitol. Assembling of the 9-deazapurine on the preformed aza-C-

glycoside afforded the desired product (ﬁgure 4.5).

155

 

 

Figure 4.5. Synthesis of ImmH. i, a, NCS, pentane; b, LiTMP, -78°C; ii, a, CNCH2MgX
or CNCH2Li; b, (Boc)20; iii, a, t-BuOCH(NMe2)2, DMF, 70°C; b, THF, HOAc, H2O; iv,
H2NCH2C00Et.HCl, NaOAc, MeOH; v, ClC02Bn, DBU, CH2Cl2, reﬂux; vi, a, H2,
Pc/C, EtOH; b, H2NCH=NH.H0Ac, EtOH, reﬂux; vii, TFA.

Imino sugars hold tremendous promise as well as tremendous challenges for drug
development targeting glycosyl transferases and glycosidases. Unfortunately, none of the

ones obtainable from natural sources are readily functionalizable in ways that can

augment their activity or enhance or alter their speciﬁcity. There is therefore a great need

156

for scaffolds and the synthetic methodology that allow the development of drugs that
target these enzymes. There is no general strategy for the preparation of imino-sugar
based transition state nucleoside sugar mimetics. Aromatic groups, carboxylic acid,
alcohol or amino groups strategically placed at the anomeric position are not present in
the naturally occurring or synthetic compounds that have been described in the literature.
Those functional groups allow modiﬁcations and functionalizations that alter binding or
speciﬁcity. None of the naturally occurring or synthetic candidates contain ring systems
in positions that allow the modiﬁcation of the solubility, polarity, dipole moment,
polarizability or general steric bulk or shape of candidates without affecting the
recognition role of the hydroxyl groups. These considerations are all important in tuning
the biological properties of drugs. Therefore, our strategy for developing inhibitors to
enzymes that process ribosides through the formation of an oxocarbenium intermediate is
to tether the aglycon analogue structures to the irninoalditol moiety so that the two are in
a relative orientation reﬂecting bond making and breaking at the transition state (ﬁgure
4.1). Structures that integrate the key structural features of the ribosyl cation and have
provision for the easy integration of the glycosidically linked components were therefore
developed. To this end, synthetic strategies aimed at preparing the hexahydro-
pyrrolothiazines 40 and 4111 were designed. The ester function can be readily transformed
to a free acid making the coupling to a variety of aglycon analogues via ester, amide or
other linkages possible. Alternatively the carboxylate group can be converted to an
alcohol function thus expanding the number of linkage possibilities. The phenyl ring can
be easily modiﬁed to tether aglycon analogs, the length and type of tether can be adjusted

to provide the ﬂexibility. The aromatic moiety also provides aspect of the environment of

157

purine base in PNP and related enzymes. A related compound with the targeted structure
4211 was prepared to facilitate conformational studies through crystallography and to

allow evaluation of the ring fusion on the biological activity of the iminopentitol

    
   
 
    
   

component.
Scaffold ,f ~
“1" Accepto
Scaffold/0’7 ~ , analo-
.\/\9\ aglycon
anal
H O OH or diphosphate
HO OH

Donor

Glycosidase design Glycosyl transferase design

Figure 4.6. Architecture of glycosidase and transferase inhibitors based on

611,:. ' ',:.-,,. ' scaffold.

J

COOC H
8N 3 SO

 

158

4.2. Results and Discussions

4.2.1. Synthesis of Iminopentitol Derivatives

The synthesis of compound 40 starting from D-ribose 43 is illustrated in ﬁgure 4.7.
Treatment of D-ribose with methanol and an acid catalyst yielded the methyl glycoside
44 which was converted to the 5-bromo-5-deoxy riboside 45 by treatment with carbon
tetrabromide and triphenylphosphine. Compound 45 was transformed to its 2, 3-
dipivaloyl derivative 46 oxidation of which with chromium trioxide yielded the 5-bromo-
5-deoxy-4-ulosonic acid methyl ester 47. Chromium trioxide can oxidize both a- and 13-
protected ribosides to keto esters”. Therefore, both isomers can be used as the starting
materials unlike the case with the pyranosides where only the B-isomer can be oxidized.
Reaction of 47 with L-cysteine methyl ester hydrochloride in mild base followed by
reduction with cyanoborohydride yielded the lactam 48 and uncyclized product 49. The

lactam 48 was readily reduced with borane and deacylated to the desired compound 40.

159

H0
H6 6H

43

Br O
1.7%

and 691v

46

HO
., 137......
H6 6H

44

0
..iv_. EMS—3”“ v .

PIVO 691v

47

Br
J. @0011: 1.

H6 6H
45
NcoocH, °°°°H°
N o NHo
H‘" + H 5 .3 OCH:
P1110“ bptv ”'0 0"“
4s 49
lvl
N
H“. n -
H6 15H
40

Figure 4.7. Synthesis of compound 40. i. HCl/MeOH; ii. Ph3P, CBr4, pyridine, 86.7% for
2 steps; iii PivCl, pyridine, 91.1%; iv. Cr03, AC2O, HOAc, 96.6%; v. (1) cysteinmethyl
ester hydrochloride, NaHC03, CH3OH, (2) NaCNBH3, CH30H; (3) Na2C03, CH2C12,
32% for 3 steps; vi. (1) BH3-THF; (2) Na0CH3, CH30H, 81% for 2 steps.

The synthesis of product 41a was attempted by treating the 5-bromo-5-deoxy-4-ulosonic

acid methyl ester 47 with aminothiophenol followed by the reduction, cyclization and

deprotection. Lactams 50a and 50b and some uncyclized product were formed with the

L-lyxo conﬁguration lactam 50b as the major product (ﬁgure 4.8). This was indicated by

a small ~3Hz coupling constant of the three protons of the aza-ribo moiety.

160

O 0
Br
woe": i S 1 S 1
5" N t N
PlvO Plv \w o o
H H

47 50a 501)

ON"

Figure 4.8. Synthesis of compound 41b. i. (1) 2-aminothiophenol, CH30H; (2)
NaCNBH3, CH30H; (3) Na2CO3, CH2C12, 67% for 2 steps; ii. (1) BH3-THF; (2)
NaOCH3, CH3OH, 90.9% for 2 steps.

A similar scheme (ﬁgure 4.9) was used in an attempt to prepare compound 42a except
that aminoethanethiol was used instead of cysteinernethyl ester. In this case two other
products were obtained namely the compound with the alternative conﬁguration of the
carbon at the ring junction (52b) and an elimination product 53. Conditions under which
the elimination product was not formed were developed. Basic conditions were avoided
by adding sodium cyanoborohydride / triﬂoroacetic acid at the same time as the
aminoethanethiol. Only lactam 52a and 52b were formed with the desired compound

being the minor component (20% yield for lactam 52a) based on the coupling constants

of the three protons of the azasugar ring. The formation of the elimination product was

161

suppressed the sooner sodium cyanoborohydride was added after mixing compound 47

and the aminoethanethiol.

O
”vac“.

9M5 5m

47

 

 

NH 0
ll
HO : ', OCH3 ———>
Pivo‘ bPiv
s/\
N N
50° + \ °
H . . - -.
Plvd ’oPIv Plvd bPIv
52b 53

1m

Figure 4.9. Synthesis of compound 42b. i. HS(CH2)2NH2, CH30H; ii. NaCNBH3,
CF3COOH, 57% for 2 steps; iii. (1) BH3-Tl-IF; (2) NaOCH3, CH30H, 85% for 2 steps.

The crystal structures of compounds 41b, 42b and 53 were obtained (figures 4.108, b,

and c). This allowed the deﬁnitive identiﬁcation of the correct stereochemistry of the

intermediate lactams. The X-ray structure of compound 53 conﬁrmed the presence of a

double bond. It also indicated that the ethylene fragment bridging the two heteroatoms in

53 could take either of two possible orientations. These structures are invaluable from a

computation and modeling perspective. They provided information on the geometry

162

about the ring junction, and allowed more than one geometrical perspective on how
attaching an aglycon analogue to 40 and 41 (a or b) would satisfy the geometric
constraints of an actual substrate. They also served as benchmarks in the evaluation how
reliably semi-empirical methods and molecular mechanics methods described the
geometry and conformational preferences of this class of molecules. Although in one
case the desired D-ribo analog was not obtained as the major product, the evaluation of
geometric constraints on the placement of the aglycon moiety was still favorable based

on the x—ray analysis and computational studies.

 

163

 

b c

Figure 4.10 (a) Ortep drawing of X-ray structure of 41b. (b) Ortep drawing of X-ray
structure of 42b. (c) Ortep drawing of X-ray structure of 53.

As reported earlier, the structures of the PNP imino-sugar inhibitor Imm-H (l,4-dideoxy-
l,4-imino-1-(S)-(9-deazahypoxanthin-9-Yl)-D-ribitol)8 and a formycin-A derivative (1-
C-(7-amino-1H-pyrazolo[4,3-d]pyrimidin-3-yl)-l,4-anhydro—D-r’ibitol)19 bound to the
active site of the enzyme have been determined by X-ray crystallography. These
structures were used as starting points for determining the proper orientation of the
aglycon and deciding the structural and geometric requirements of the tether and the
aglycon analogue. Figures 4.11s shows the crystal structure of ImmH and ﬁgure 4.11b
shows the superimposition of the crystal structures of 42b and ImmH. The asterisk
indicates the hydrogen atom that is replaced by a carboxymethyl group in compound 40

or a phenyl group in 41 (a or b). Molecular mechanics models based on these X-ray

structures indicate that the choice of cysteine isomer (L or D) could be used to position
the aglycon analog at the appropriate position above (as is the case with PNP inhibitors)
or below the plane of the iminopentitol ring respectively. The phenyl group of 41 (a or b)
can also be attached to a variety of aglycons at the proper orientation. The sulfur atom in
42b is at the same position as the 5-OH of the inhibitors shown in ﬁgures 4.11. This
provides another possibility for modifying this structure at a strategic point by oxidizing
the sulfur to a sulfoxide or a sulfone. The negative charge of the sulfone oxygen can
stabilize some transition states where the 5-position of the sugar ring might be

phosphorylated.

 

Figure 4.11. (a) X-ray derived structure of l,4-dideoxy-1,4-imino-1-(S)-(9-
deazahypoxanthin-9-Yl)-D— ribitol (b) Comparison of X-ray structure 42b and 1,4-
dideoxy-1,4-imino-1-(S)-(9-deazahypoxanthin-9-Yl)-D- ribitol.

4.2.2. Inhibitory Activity of Iminopentitol Derivatives

Compounds 40, 41b and 42b were tested against bacterial produced B-ribosidase. The

mechanism of action of B—ribosidase also features the ribooxocarbenium ion transition

state, similar to riboside hydrolase, phosphorylase and related enzymes. In this assay,

165

 

3’,4’-dihydroxyﬂavone-4’-B-D-riboﬁ.1ranoside 54 (UHF-riboside) was used as the
substrate”. This chromogenic substrate can be hydrolyzed by B-ribosidase-producing
bacteria to release 3',4'-dihydroxyﬂavone, which forms a chelate with iron (55). The
assay was performed in phosphate buffered saline (PBS) at pH 7.0. P. aeruginosa 27853
and Salmonella sp. 35664 strains were used to produce B-ribosidase. Both whole cell and
lysed strain were tested. Inhibition studies were performed by adding the inhibitor to a
ﬁnal concentration of 0.2ug/uL to l5 ug/uL to the respective buffer solutions along with
DHF-riboside and ferric ammonium. citrate. Each strain was grown and added to each
buﬁer solution containing substrate and inhibitor. The solutions were incubated at 37°C

and the absorbance of this complex 55 was determined at 387 nm.

r

Riboae O x ;.,
0Na 0'

55

Figure 4.12. (a) Structure of 3',4'-dihydroxyﬂavone-4'-B-D-ribofuranoside sodium salt 54
and (b) the putative chelate 55 formed with iron aﬁer enzymatic hydrolysis.

Figure 4.13 is the absorbance of the chelate 55 produced in the lysate assay with
inhibitor 41b concentration from 0 ~ 15 ug/uL and ﬁgure 4.14 is the similar plot for the
whole cell assay. The results showed compound 41b has good inhibitory activity against

B-ribosidase produced from strain. About 70% inhibition at concentration l5 ug/uL

I66

 

(67mM) at 24 hours for both whole cell and lysate was observed. At concentration 7.5
ug/uL, about 50% inhibition was observed. Compound 40 and 42b have no signiﬁcant
inhibition for the same strain. These results suggest that phenyl group in compound 41b
represents an important factor for the activity. The phenyl ring increases the
hydrophobicity of the inhibitor, which enhance the afﬁnity towards enzyme. The
increased hydrophobicity also makes it easy to penetrate the cell membrane. The same
inhibition observed for compound 41b in the lystate and whole cell experiments
demonstrates the easy uptake through the cell envelope. It is also important to note that
with increasing concentration of the inhibitor, lysate experiment showed inhibition
gradually, while the whole cell experiment showed inhibition only aﬁer certain inhibitor
concentration. It suggests that the inhibitor needs to reach some concentration in the cell
to start inhibition.

The activity of compound 41b against ribosidase demonstrated that these iminopentitol
scaffolds can provide key features of the ribosyl cation, which represents an important
state of a variety of nucleoside related enzymes. These scaffolds are also easy to couple
with all kinds of aglycon analogs, such as bases and phosphate groups. It opens up

possibilities to the preparation of a wide range of biological important compounds.

167

 

A for lysate

 

 

 

1.2 -‘ 0.5
"__,.’:’/1‘ 0.9
,-"’ , ' .
1.0 - t 0.2
‘ ° ‘ - e 0
/V
0.8 - //
V‘ 1.9
'1 ,./
// ‘
0.6 -w / .. A —/""" 3.8
‘ . K.;/”II /~I . wwﬂmg
0,4 _ / /,/‘ff’ . i// 7.5 uglul
0.2 e 15 uglul
t
0.0 4
’0‘2 T ‘ T ' r r T r I r I v
0 5 10 15 20 25

time (hr)

Figure 4.13. Aabsorbance of the chelate 55 produced in the lysate assay with inhibitor
41b concentration from 0 ~ 15 ug/uL

A for whole cell

 

 

 

1.2 -
0.9
r ”’ o 2
1.0 - ,. - - .
//, :_:V7:/:
0.8 - / / 1 .9
0.6 - A __ _____ A 3.8
. ‘1 If ’ I. if .. w-m-«w O
0.4 - ' / 1 7 5
. I
0.2 _ 15ug/ul
0.0 1
l '— T I r T 1 l i
0 5 1o 15 20 25
time (hr)

168

Figure 4.14. Absorbance of the chelate 55 produced in the whole cell assay with inhibitor
41b concentration from 0 ~ 15 ug/uL

In conclusion, novel hetero bicyclic aza-sugars 40 and 41b have been designed and
synthesized. The relatively easy synthesis constitutes a general approach to bicyclic
structures with either natural or unnatural stereochemistry. This strategy provides an easy
access to preparation of iminopentitol scaffolds for ribosidase, phosphorylase and related
enzyme inhibitors. With an iminopentitol moiety, an ester function or phenyl group and a
heteroatom, sulfur, a variety of aza-sugar with different aglycon and their analogues can
be synthesized. Those functional groups make the preparation of compound libraries with
different aglycon analogues possible.

Based on computational and modeling studies, the conformation of the iminopentitol ring
in these bicyclic systems and the general presentation of the hydroxyl groups and the
additional ring residue match closely the structures of the known inhibitors. The
inhibition studies show the activity of this kind of compounds against B-ribosidase, and
indicate the importance of the ribooxocarbenium ion mimicry and hydrophobicity for the
inhibition. These iminopentitol scaffolds therefore provide key features of the ribosyl
cation and easy integration of an aglycon component. The possibility of obtaining the L-
lyxo conformation and the fact that such derivative is still active against ribosidase really

expand the scope of this approach.

4.3. Experimental Section

4.3.1 . General Procedures

169

IR spectra were recorded on a Nicolet 710 FT-IR instrument. The 1H (and '3C) NMR
spectra were recorded at 500 MHz (125.5 MHz for 13-C) on a Varian VXR spectrometer.
The HRMS FAB mass spectra were obtained using a Jeol HX-l 10 double-focusing mass

spectrometer operating in positive ion mode.

4.3.2. Inhibition Assay

Inhibitory potency of 40, 41b and 42b were determined by spectrophotometrically
measuring the residual hydrolytic activities of the bacterial produced B-ribosidase. The
bacteria used to produce B-ribosidase were P. aeruginosa 27853 and Salmonella
sp.35664. both whole cell and lysed strains. 3’,4’-Dihydroxyﬂavone-4’-B—D-
ribofuranoside 54 (DHF-riboside) was used as substrate”. The assay was performed in a
PBS buffer at pH 7.0, 37°C. Serial dilutions were made of the inhibitors to a ﬁnal
concentration of 0.2pg/ul. to lSug/uL in each assay solution. Each strain was grown to
mid exponential phase and removed from the media and rediluted in PBS buffer.
Inhibition studies were performed by adding each strain to the respective buffer solutions
along with DHF-riboside, ferric ammonium citrate and inhibitors. For compound 41b,
10% DMSO solution was used to increase the solubility. The solutions were incubated at
37°C and the absorbance of the complex formed between released 3',4'-dihydroxyﬂavone
and iron was determined at 387 nm at time 15min, 30min, 45min, 60min, 75min, 90min,

105min, 2hours, 2.5hours, 3.0hours, 4.0hours, 4.5hours, Shours, 20hours, and 24hours.

Methyl S-Bromo-S-deoxy-D-ribofuranoside (45) A solution of D-ribose 43 (15g,

0.1mol) and hydrochloric acid (lSmL) in methanol (lSOOmL) was stirred at room

170

temperature for 20 hours, followed by addition of sodium carbonate (5g, 0.06mol). The
reaction mixture was stirred for 2 hours and two thirds of the solvent was removed at a
temperature of less than 30°C. The mixture was ﬁltered and concentrated. The resulting
methyl D-riboside 44 was obtained as yellow oil and was used without further
puriﬁcation. To a stirred solution of 44 (5g, 30rnrnol) in anhydrous pyridine (300mL) at
0°C were added triphenylphosphine (16.0g, 60mmol) and carbon tetrabromide (18.0g,
50mmol). The resulting mixture was protected from moisture and stirred at 0°C for ten
minutes. The mixture was then allowed to warm to 65°C and was stirred for 6 hours.
Methanol (lOmL) was added to decompose any excess reagent. The solvent was removed
by evaporation and the residue was dissolved in dichloromethane and was then passed
though a pad of silica gel. Dichloromethane was removed by evaporation and the
resulting residue was dissolved in toluene and the product was extracted with water for
several times. Removal of water afforded the expected mixture of methyl S-Bromo-S-
deoxy-a- and B-D-riboﬁiranoside 45 (6.0g, 86.7%, (1:0 = 3:1). IR (CHCl3) vmax 3387.18,
2933.49, 1121.60, 1088.32, 1025.91 cm". The mixture was used in the next step without
separating the anomer: Methyl S-Bromo-S-deoxy-a-D-ribofuranoside IH NMR
(500MHz, D20) 6 4.86, (1H, s), 4.19-4.113 (2H, m), 4.02 (1H, d, J=5 Hz), 3.64 ( 1H, dd,
J=11, 4.3 Hz), 3.53 (1H, dd, J=1 1.5, 6.3 Hz), 3.35 (3H, s); l3C NMR (125.5MHz, D20)
108.14, 81.73, 74.62, 73.12, 55.44, 34.22 ppm. Methyl S-Bromo-S-deoxy-li-D-
ribofuranoside IH NMR (500MHz, D20) 8 4.97, (l H, d, J=4.5 Hz), 4.22 (1H, m), 3.93
(1H, m), 4.15-4.11 (1H, m), 4.01 (1H, m), 3.57 (2H. m), 3.38 (3H, 3); '3C NMR

(125.5MHz, D20) 103.68, 82.98, 71.68, 70.98, 55.77, 33.45 ppm.

171

 

Methyl S-Bromo-S-deoxy-Z,3-di-0-trimethylacetyl-D-ribofuranoside (46)
Pivaloylation of 45 (2.0g, 8.8mmol) by trimethylacetyl chloride (4.34mL, 35.0mmol) in
pyridine (1 SOmL) at room temperature for 2 days afforded a white solid 46 (3.2g, 91.1%),
IR (CHC13) vm 2972.45, 1739.35, 1162.70, 1139.15 cm'l. The mixture was used in the
oxidation step without separating the anomer: Methyl 5-Bromo-5-deoxy-2,3-di-0-
trimethylacetyl-a-D-ribofuranoside 1H NMR (500MHz, CDCl3) 5 5.25, (1H, m), 5.20
(1H, d, J=4.5 Hz), 4.82 (1H, s), 4.28, (1H, m), 3.47 (2H, t, J=6.5 Hz), 3.37 (3H, s), 1.19
(9H, s), 1.16 (9H, s); 13C NMR (125.5MHz, CDCl3) 177.12, 176.93, 106.24, 80.04,
74.90, 74.05, 55.21, 38.82, 38.67, 33.52, 27.05 ppm. Methyl 5-Bromo-5-deoxy—2,3-di-
0-trimethylacetyl-B-D-ribofuranoside lH NMR (500MHz, CDCl3) 8 5.17, (1H, d, J=4
Hz), 5.04 (1H, m), 4.90 (1H, m), 4.22 (1H, m), 3.64 (1H, dd, J=11, 4 Hz), 3.55 (1H, dd,
J=11, 4.5 Hz), 3.35 (3H, s), 1.19(91-1, s), 1.18 (9H, s); l3c NMR(125.5MHz,CDC13)

177.76, 177.47, 101.02, 80.46, 74.05, 71.96, 55.59, 38.82, 38.66, 33.06, 27.05 ppm.

Methyl S-Bromo—S-deoxy-2,3-di-0-trimethylacetyl-D-erythro-4-pentulosonic acid
methyl ester (47) Chromium trioxide (2.43g, 24.3mmol) was added to a solution of 46
(3.2g, 8.1mmol) in acetic acid (100mL) and acetic anhydride (lOmL). The suspension
was stirred at room temperature for 2 hours. The mixture was then poured slowly into
cold water (500mL). The water was extracted 5 times with CHzClz and the combined
organic phase was washed with brine, saturated sodium bicarbonate and dried (NaZSO4).
The solution was passed through a small pad of silica gel to remove coloration due to
chromium salts and dried to give 47 as a colorless oil (3.2g, 96.6%). IR (CHC13) vm

2975.59, 1745.62, 1125.06 cm'l. lH NMR (500MHz, CDCl3) 5 5.86, (1H, d, J=3.0 Hz),

 

5.70 (ll-I, d, J=2.5 Hz), 4.08 (1H, d, J=l3 Hz), 4.01 (1H, d, J=12.5Hz), 3.76 (3H, s), 1.24
(9H, s), 1.20 (9H, s); 13C NMR (125.5MHz, CDCl3) 5 195.60, 176.62, 176.36, 166.72,
74.88, 71.54, 52.86, 38.83, 38.71, 31.28, 26.87 ppm; HRFABMA (M + H+) Calcd.

409.0862, found 409.0863.

7(S),8(R),9(S)-bis(trimethylacetoxy)-6-oxo-hexahydro-pyrrolo-[1,4]-thiazine-4-
carboxylic acid methyl ester (48) A solution of 47 (1 .Og, 2.4mmol), sodium bicarbonate
(0.25g, 2.4mmol) and cysteinmethyl ester hydrochloride (0.5g, 2.9mmol) in methanol
(100 mL) was stirred at room temperature for one hour, followed by addition of sodium
cyanoboronhydride (0.23g, 3.7mmol). The reaction mixture was stirred overnight and
concentrated. The resulting residue was dissolved in dichloromethane and sodium
carbonate was added to facilitate the lactam cyclization. After stirred for overnight, the
suspension was ﬁltered and the dichloromethane solution was washed by brine and dried
(NaZSO4), concentrated. The residue was puriﬁed by column chromatography (6:1
Hexanes/Ethyl acetate) to yield lactam 48 (0.3g, 32.0%) and the uncyclized product 49
(0.5g, 53.0%).

Lactam 48 was given as a white solid. IR (CHC13) vmax 1745.80, 1728.75, 1161.29 cm‘
'.'H NMR (500MHz, CDCl3) 5 5.61 (1H, m), 5.58 (1H, d, J=5.5 Hz), 4.13 (1H, dt,
J=11.5, 3.5 Hz), 3.78 (3H, s), 3.13 (1H, dt, J=l4, 2.3 Hz), 2.88 (1H, dd, J=13.8, 4.3 Hz),
2.69 (1H, dd, J=13.5, 11.5 Hz), 2.43 (1H, dt, J=13.5, 2.3 Hz), 1.23 (9H, s), 1.23 (9H, s);
13C NMR (125.5MHz, CDCl3) 5 176.94, 176.89, 169.25, 168.55, 69.05, 68.07, 55.07,
52.88, 52.33, 39.11, 38.82, 28.59, 27.21, 27.07, 26.63 ppm; HRFABMA (M + H+) Calcd.

416.1743, found 416.1742.

173

Compound 49 was given as a white solid. IR (CHC13) vm 1740.01, 1736.18, 1155.51
cm“. 'H NMR(500MH2,CDC13)5 5.38 (1H, d, J=7.5 Hz), 5.16 (1H, dd, J=9, 3 Hz),
3.71 (3H, s), 3.70 (3H, s), 3.36 (1H, m), 2.64 (1H, m), 251-2.31 (4H, m), 1.22 (9H, s),
1.15 (9H, s); '3(: NMR (125.5MHz, CDC13) 8 177.14, 176.93, 171.43, 167.36, 73.79,

70.30, 59.34, 56.00, 52.32, 52.18, 38.86, 38.75, 29.40, 28.42, 27.00. 26.97 ppm;

7(S),8(R),9(S)-dihydroxy-hexahydro-pyrrolo—l1,4]-thiazine-4-carboxylic acid methyl
ester (40) A solution of lactam 48 (0.16g, 0.42mmol) and BH3-THF (3.5mL, 5.3mmol) in
anhydrous THF (lOmL) was reﬂuxed for 8 hours and NMR showed the completion of the
reduction. The solvent was removed and methanol was added and concentrated for 3
times. The residue was dissolved in methanol (20mL), followed by addition of NaOMe
(0.06g, 1.1mmol). The reaction was stirred for overnight and concentrated. The residue
was applied to an ion exchange column (Dowex 50WX8-400, 2g), which was washed
with water (30 mL) and eluted with 2N NH4OH (30mL). The elution was concentrated
and puriﬁed by column chromatography (15:1 CH2C12/MCOH) to afford a white solid
(0.08g, 81%); lH NMR (500MHz, CDC13) 8 4.34-4.30 (1H, m), 3.98 (1H, m), 3.69 (3H,
s), 3.43 (1H, d, J=11 Hz), 3.23 (1H, d, J=9.3 Hz), 2.98 (1H, dd, J=l4, 4 Hz), 2.84 (1H,
dd, J=13.8, 3.3 Hz), 2.79-2.74 (1H, m), 2.71 (1H, d, J=13 Hz), 2.43 (1H, d, J=13 Hz); l3C
NMR (125.5MHz, CDC13) 6 173.53, 72.50, 68.62, 59.11, 58.60, 55.42, 52.29, 28.46,

26.47; HRFABMA (M + H“) Calcd. 234.0800, found 234.0799.

Lactam 50a and 50b A solution of 47 (2.0g, 4.9mmol) and 2-aminothiophenol (0.73g,

5.9mmol) in methanol (150 mL) was stirred at room temperature for one hour, followed

174

by addition of sodium cyanoboronhydride (0.56g, 7.3mmol). The reaction mixture was
stirred overnight and concentrated. The resulting residue was dissolved in
dichloromethane and sodium carbonate was added to facilitate the lactam cyclization.
After stirred for overnight, the suspension was ﬁltered and the dichloromethane solution
was washed by brine and dried (Na2S04), concentrated. The residue was puriﬁed by
column chromatography (10:1 Hexanes/Ethyl acetate) to yield two lactam diastereomers
Product 50a (0.13g, 9.6%) was given as a white solid. IR (CHCl3) vm 2973.71, 1738.05,
1145.55 cm". IH NMR (500MHz, CDCl3) 8 8.06 (1H, d, J=8 Hz), 7.15-7.06 (3H, m),
5.63 (1H, t, J=7.5 Hz), 5.51 (1H, d, J=7 Hz), 4.34 (1H, m), 3.17 (1H, t, J=12 Hz), 2.96
(1H, dd, J=13, 2.5 Hz), 1.27 (9H, s), 1.25 (9H, 8); '3C NMR (125.5MHz, CDC13) 8
177.42, 177.22, 165.28, 132.20, 126.11, 125.75, 124.76, 123.42, 72.97, 71.27, 54.66,
38.87, 38.74, 27.10, 26.95 ppm; HRFABMA (M + H+) Calcd. 406.1689, found
406.1689.

Product 50b (0.71 g, 57.4%) was given as a white solid. IR (CHC13) v“rm 2972.76,
1742.61, 1155.27 cm'l.; lH NMR (500MHz, CDCl3) 8 8.72 (1H, dd, J=8.5, 1 Hz), 7.14
(1H, dd, J=8, 1.5 Hz), 7.10 (1H, td, J=8, 1.5 Hz), 6.99 (1H, td, J=7.5, 1.5 Hz), 5.72 (1H,
t, J=4.8 Hz), 5.65 (1H, d, J=5.5 Hz), 4.32 (1H, m), 3.14 (1H, dd, J=13, 10.5 Hz), 2.88
(1H, dd, J=13, 2.5 Hz), 1.24 (9H, s), 1.21 (9H, 8); '3C NMR (125.5MHz, CDCl3) 8
177.14, 176.94, 166.79, 133.69, 127.32, 125.76, 124.51, 121.12, 120.41, 68.96, 67.89,
57.38, 39.18, 38.87, 27.16, 27.08, 25.01; HRFABMA (M + H+) Calcd. 406.1689, found

406.1691.

175

1(S) 2(R), 3(S)-dihydroxy-2,3,3a,4-tetrahydro-lH-pyrrolo-[2,l-c][1,4]benzothiazine
(41b) A solution of lactam 50b (0.40g, 1.0mmol) and BH3-THF (3.3mL, 5.0mmol) in
anhydrous THF (lOmL) was reﬂuxed for 4 hours and the TLC and NMR showed the
completion of the reduction. The solvent was removed and methanol was added and
concentrated for 3 times. The residue was dissolved in methanol (20mL), followed by
addition of NaOMe (0.20g, 3.8mmol). The reaction was stirred for 8 hours and
concentrated. The residue was puriﬁed by column chromatography (15:1 CH2C12fMeOH)
to afford a white solid (0.2g, 90.9%). IR (CHC13) vm, 3113.94, 1124.86 cm". 'H NMR
(500MHz, CD3OD) 8 6.93 (2H, d, J=7.5 Hz), 6.48 (1H, td, J=7.5, 1 Hz), 6.41 (1H, d, J=8
Hz), 4.37 (1H, m), 4.11 (1H, t, J=3.8 Hz), 3.80 (1H, dt, J=10.5, 3 Hz), 3.51 (1H, t, J=8.5
Hz), 3.15 (1H, t, J=8.5 Hz), 2.96 (1H, m), 2.85 (1H, dd, J=12.5, 3 Hz); 13C NMR
(125.5MHz, CD30D) 8 146.55, 130.93, 129.84, 119.67, 119.18, 115.14, 77.05, 74.25,

65.67, 55.75, 27.99 ppm; HRFABMA (M + H*) Calcd. 224.0745, found 224.0745.

7(S),8(R),9(S)-hexahydro-bis-trimethylacetoxy-1H-pyrrolo-[l,4]-thiazin-6-one (52a)
and 7(S),8(R),9(R)-hexahydro-bis-trimethylacetoxy-1H-pyrrolo-[l,4]-thiazin-6—one
(52b): A solution of 47 (0.8g, 2.0mmol), HS(CHz)2NH2 (0.18g, 2.3mmol) and NaCNBH3
(0.18g, 2.9mmol) in methanol (50 mL) was stirred at room temperature for 24 hours and
then concentrated. The resulting residue was dissolved in dichloromethane and washed
by brine, dried (NaZSO4) and concentrated. Compound 52a and 52b were isolated by
column chromatography (4:1 Hexanes/Ethyl acetate).

Product 52a (0.12g, 20%) was obtained as a white solid. IR (CHC13) vmax 1739.70,

1717.38, 1160.52, 1136.98 cm". lH NMR (500MHz, CDCl3) 8 5.34 (1H, d, J=7 Hz),

176

5.09 (1H, dd, J=7, 2.3 Hz), 4.44 (1H, dt, J=13.5, 2.8 Hz), 3.60 (1H, dt, J=11, 2.5 Hz),
3.02 (1H, m), 2.75 (1H, d, J=13 Hz), 2.62-2.51 (2H, m), 2.42 (1H, m), 1.20 (9H, s), 1.16
(9H, s); l3C NMR (125.5MHz, CDCl3) 8 177.41, 177.06, 166.64, 70.12, 68.15, 62.27,
42.24, 38.83, 38.76, 30.10, 27.11, 26.97, 26.41 ppm; HRFABMA (M + H) Calcd.
358.1689, found 358.1688. Product 52b (0.40g, 57%) was obtained as a white solid. IR
(CHC13) vw 1739.61, 1705.31, 1141.03 cm". ‘H NMR (500MHz, CDCl3) 8 5.50 (1H, t,
J=5 Hz), 5.43 (1H, dd, J=6, 1.5 Hz), 4.43 (1H, dt, J=13.5, 2.8 Hz), 3.88 (1H, m), 2.97
(1H, m), 2.72-2.64 (2H, m), 2.52 (1H, m), 2.40 (1H, dt, J=13.5, 2.5 Hz), 1.21 (9H, s),
1.20 (9H, s); 13C NMR (125.5MHz, C1103) 8 177.03, 176.89, 167.72, 68.78, 67.47,
57.69, 42.23, 39.08, 38.82, 27.31, 27.20, 27.12, 26.97 ppm; HRFABMA (M + 11+) Calcd.
358.1689, found 358.1685. The formation of the elimination product 53 was favored by
delaying the addition of sodium cyanoboronhydride. A solution of 47 (1.9g, 4.6mmol)
and HS(CH2)2NH2 (0.43g, 5.6rmnol) in methanol (150 mL) was stirred at room
temperature for one hour, followed by addition of sodium cyanoboronhydride (0.44g,
7.0mmol). The reaction mixture was stirred overnight and concentrated. The resulting
residue was dissolved in dichloromethane and sodium carbonate was added to facilitate
the lactam cyclization. Aﬁer stirred for several hours, the suspension was ﬁltered and the
dichloromethane solution was washed by brine and dried (Nasta), concentrated. Three
products (19, 24.1%; 20, 8.4%; 21, 26.5%) were isolated by column chromatography (4:1
Hexanes/Ethyl acetate). Product 53 (0.44g, 26.5%) was obtained as a yellow solid. IR
(CHC13) vmax 1741.94, 1153.58 cm". lH NMR (500MHz, CDCl3) 8 5.83 (1H, d, J=6.5
Hz), 5.72 (1H, s), 5.47 (1H, d, J=-'6.5 Hz), 4.20 (1H, m), 3.67 (1H, m), 2.94-2.89 (2H, m),

1.26 (9H, s), 1.18 (9H, s); ”C. NMR (125.5MHz. CDCl3) 8 177.37, 177.08, 167.65,

177

 

130.16, 100.02, 67.98, 67.26, 39.91, 38.91, 38.90, 27.14, 26.98, 24.07 ppm; HRFABMA

(M+) Calcd. 355.1454, found 355.1451.

7(S),8(R),9(S)-dihydroxy-hexahydro-lH-pyrrolo-[l,4]-thiazine (42b) A solution of
lactam 52b (0.24g, 0.67mmol) and BH3-THF (2.2mL, 3.4mmol) in anhydrous THF
(lOmL) was reﬂuxed for 3 hours and the TLC and NMR showed the completion of the
reduction. The solvent was removed and methanol was added and concentrated for 3
times. The residue was dissolved in methanol (20mL), followed by addition of NaOMe
(0.10g, 1.9mmol). The reaction was stirred for 8 hours and concentrated. The residue was
applied to an ion exchange column (Dowex 50WX8-400, 5g), which was washed with
water (30 mL) and eluted with 2N NHaOH (30mL). The elution was concentrated to
afford a white solid (0.1g, 85.0%). iH NMR (500MHz, D20) 8 4.23 (1H, m), 4.07 (1H,
m), 3.22 (1H, dt, J=12, 2.5 Hz), 2.86-2.74 (3H, m), 2.52 (1H, dt, J=13.5, 3 Hz), 2.49 (1H,
m), 2.44 (1H, m), 2.33-2.25 (2H, m); 13C NMR (125.5MHz, D20) 8 71.92, 67.35, 67.02,

60.90, 53.86, 26.41, 26.28 ppm; HRFABMA (M + H+) Calcd. 176.0745, found 176.0745.

178

4.4. References

(1) Gopaul, D. N.; Meyer, S. L.; Degano, M.; Sacchettini, J. C.; Schramm, V. L. lnosine-
uridine nucleoside hydrolase from Crithidia fasciculata. Genetic characterization,
crystallization, and identiﬁcation of histidine 241 as a catalytic site residue Biochemistry
1996, 35, 5963-5970.

(2) Mazzella, L. J .; Parkin, D. W.; Tyler, P. C.; Fumeaux, R. H.; Schramm, V. L.
Mechanistic diagnoses of N-ribohydrolases and purine nucleoside phosphorylase J. Am.
Chem. Soc. 1996, 118, 2111-2112.

(3) Parkin, D. W.; Limberg, G.; Tyler, P. C.; Fumeaux, R. H.; Chen, X. Y.; Schramm, V.
L. Isozyme-speciﬁc transition state inhibitors for the trypanosomal nucleoside hydrolases
Biochemistry 1997, 36, 3528-3534.

(4) Tao, W., Grubmeyer, C., Blanchard, J. S. Transition State Structure of Sahnonella
typhimurium Orotate Phosphoribosyltransferase Biochemistry 1996, 35, 14-21.

(5) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - lnosine Hydrolysis,
Tight-Binding of the Hypoxanthine Intermediate, and 3rd-the-Sites Reactivity
Biochemistry 1992, 31, 5964-5973.

(6) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - Catalytic
Mechanism and Transition-State Analysis of the Arsenolysis Reaction Biochemistry
1993, 32, 13212-13219.

(7) Kline, P. C.; Schramm, V. L. Pre-Steady-State Transition-State Analysis of the
Hydrolytic Reaction Catalyzed by Purine Nucleoside Phosphorylase Biochemistry 1995,
34, 1153-1162.

(8) F edorov, A.; Shi, W.; Kicska, G., Fedorov, 13.; Tyler, P. C.; Fumeaux, R. H.; Hanson,
J. C.; Gainsford, G. J.; Larese, J. Z.; Schramm, V. L.; Almo, S. C. Transition state
structure of purine nucleoside phosphorylase and principles of atomic motion in
enzymatic catalysis Biochemistry 2001, 40, 853-860.

(9) Miles, R. W.; Tyler, P. C.; Fumeaux, R. H.; Bagdassarian, C. K.; Schramm, V. L.
One-third-the-sites transition-state inhibitors for purine nucleoside phosphorylase
Biochemistry 1998, 3 7, 8615-8621.

(10) Evans, G. B.; Fumeaux, R. H.; Gainsford, G. J.; Schramm, V. L.; Tyler, P. C.

Synthesis of transition state analogue inhibitors for purine nucleoside phosphorylase and
N-riboside hydrolases Tetrahedron 2000, 56, 3053-3062.

179

 

(11)Yasuda, K.; Kizu, H.; Yamashita, T.; Kameda, Y.; Kato, A.; Nash, R. J .; Fleet, G.
W. J .; Molyneux, R. J .; Asano, N. New sugar-mimic alkaloids from the pods of
Angylocalyx pynaertii J. Nat. Prod. 2002, 65, 198-202.

(12) Asano, N.; Yamashita, T.; Yasuda, K.; Ikeda, K.; Kizu, H.; Kameda, Y.; Kato, A.;
Nash, R. J.; Lee, H. S.; Ryu, K. S. Polyhydroxylated alkaloids isolated from mulberry
trees (Morus alba L.) and silkworms (Bombyx mori L.) Journal of Agricultural and Food
Chemistry 2001, 49, 4208-4213.

(13) Asano, N.; Yasuda, K.; Kizu, H.; Kato, A.; Fan, J. Q.; Nash, R. J.; Fleet, G. W. J.;
Molyneux, R. J. Novel alpha-L-fucosidase inhibitors from the bark of Angylocalyx
pynaertii (Leguminosae) Eur. J. Biochem. 2001, 268, 35-41.

(14) Fleet, G. W. J .; Nicholas, S. J.; Smith, P. W.; Evans, S. V.; Fellows, L. E.; Nash, R.
J. Potent Competitive-Inhibition of Alpha-Galactosidase and Alpha-Glucosidase Activity
by l,4-Dideoxy-l,4-Iminopentitols - Syntheses of l,4-Dideoxy-1,4-Irnino-D-Lyxitol and
of Both Enantiomers of 1,4—Dideoxy-l,4-Iminoarabinitol Tetrahedron Lett. 1985, 26,
3127-3130.

(15) Fleet, G. W. J .; Son, J. C. Polyhydroxylated Pyrrolidines from Sugar Lactones -
Synthesis of 1,4-Dideoxy-l,4-Imino-D-Glucitol from D-Galactonolactone and Syntheses
of 1,4-Dideoxy-1,4-Imino-D-Allitol, 1,4-Dideoxy-1,4-Imino-D-Ribitol, and (2s,3r,4s)-
3,4-Dihydroxyproline from D-Gulonolactone Tetrahedron 1988, 44, 2637-2647.

(16) Jang, K. C.; Jeong, I. Y.; Yang, M. S.; Choi, S. U.; Park, K. H. Stereoselective
synthesis of 1,4-dideoxy-l,4-imino-L-lyxitol and l,5-dideoxy-1,5-imino-D-ribitol
Heterocycles 2000, 53, 887—896.

(17) Kumareswaran, R.; Hassner, A. Stereochernistry 93. Asymmetric synthesis of 1,4-
dideoxy-l,4-imino-D-ribitol via stereoselective addition of allylphenylsulfone to an aryl
N-sulﬁnylimine Tetrahedron Assym. 2002, 12, 3409-3415.

(18) Angyal, S. J .; James, K. Oxidation of Carbohydrates with Chromium Trioxide in
Acetic Acid .1. Glycosides Australian Journal of Chemistry 1970, 23, 1209-&.

(19) Koellner, G. B., A.; Wielgus-Kutrowska, B.; Luic, M.; Steiner, T.; Saenger, W.;
Stepinski, J. , 351 Open and closed conformation of the E. coli purine nucleoside

phosphorylase active center and implications for the catalytic mechanism J. Mol. Biol.
2002, 315, 351-371.

(20) Butterworth, L. A., J. D.; Davies, 6.; Burton, M.; Reed, R. H.; Gould, F. K.

Evaluation of novel B-ribosidase substrates for the differentiation of Gram-negative
bacteria Journal of Applied Microbiology 2004, 96, 170-176.

180

 

Chapter 5

A Facile Approach to Chiral 2,5-Di-Substituted Pyrrolidines

181

ABSTRACT

Chiral 2,5-di-substituted pyrrolidines, so called homoazasugars or aza-C—glycosides, are
an important class of molecules because of their inhibitory activity against glycosidases
and their stability towards chemical and enzymatic degradation. Synthesis of this type of
molecules is normally difﬁcult because of the multiple chiral centers and the nitrogen
atom in the ring. Here we describe a new strategy for the synthesis of 2,5-
dihydroxymethy1-3,4-dihydroxypyrrolidines from different methyl 2-acetamido-3,4,6-tri-
O-acetyl-2-deoxy-B-D-pyranosides. It is the most efﬁcient route demonstrated to date and

provides an easy access to more complicated homoazasugars.

182

5.1. Introduction

Iminopentitols (ﬁve membered ring azasugars) have become an important class of
molecules as glycosidase inhibitors. They display envelope conformations which are
more ﬂexible conformations than those shown by their six membered ring counterparts,
so they can mimic the proposed half chair and related transition states more closely. Of
the azasugars developed, those having a hydroxymethyl group or a polyhydroxylated
carbon chain linked to the anomeric carbon, so called homoazasugars or aza-C-glycosides
(2,5-disubstituted pyrrolidines) are especially interesting because they retain the same or
higher biological activity of the parent azasugars. They present the possibility of
preparing glycoside analogs which are not suscepitible to cleavage because a carbon to
carbon bond now replaces the O-glycoside linkage. Natural homoazasugars have been
isolated and synthesized. Typical examples are 2,5-dideoxy-2,5-imino-D-mannitol 1
(DMDP), australine 2, alexine 3 and hyacinthacine C. 4. They have either cis or trans
hydroxymethyl substituents at C-2 and C-5, combined with either cis or trans hydroxyl

groups at C-3 and C-4, and they all have potent activity against glycosidases.

183

 

Because of its biological importance, several synthetic approaches have been developed
for DMDP 1”. The ﬁrst synthesis was reported by Fleet in 1985L8 who employed D-
glucose as the starting material (ﬁgure 5.1). Protection with acetone followed by O-
benzylation at the 3-position gave the protected sugar 5. The 5,6-O—isopropylidene group
was replaced with a carbonate residue to yield 6. Acidic cleavage of the 1,2-O-
isopropylidene acetal in methanol furnished the a,B-methyl-glucosides. The primary
hydroxyl group of the (rt-isomer was protected and the 2-position was triﬂated.
Subsequent displacement of the triﬂate group at C-2 with azide gave the key intermediate
7. Compound 7 was reduced to a primary amine which underwent cyclization to C-5 via a
double displacement at that center in which an intermediate epoxide was formed. There
was overall retention of conﬁguration at C-5. Subsequent deprotection converted 10 into
DMDP 1. This synthesis is relatively long and only the a-isomer from the mixture after

methanolysis of the 1,2-acetal can be used for the subsequent reactions.

184

 

 

O 0—
>< °=<
i 0 o 11 ° 0 in
D-glucose __> 0311 __, OBn _,
O O
0% OK

 

5 6
H0— 820 O
110‘“ 0 M80 0
1V N v 0
<01 am "am —> an . ——> oenu.
Due 0110 Otto
7 8 9
OBn H
HO N .,o“\Ol-l
i, M90 0’ 611on L s
N H6 0H
82
10 1

Figure 5.1. Synthesis of DMDP. i, (a) acetone, W; (b) BnCl, NaH; ii, (a) 0.5% HCl in
MeOH. room temp, 12 h; then (MeO)2CO, NaOMe, reﬂux; iii, (a) Dowex 50W-X8,
MeOH; (b) szO (1.1 eq) in CHzClz, -20°C, 20 min, then NaNg, (3 equiv) dimethyl
forrnamide, 50 oC, 2 d; (c) MeOH with a trace of NaOMe. room temp; iv, (a) BzCl; (b)
MsCl; v, NaOMe (2 equiv), DMF, 50°C, 5 h; vi, palladium black, H2, EtOH; remove
catalyst, then 50 °C overnight; PhCHzOCOCl, ether, H20 containing NaHCO3; vii,
CF3COOH - H20 (1:1), room temp, 1 h; then NaBH4 in EtOH - H20; ix, palladium
hydroxide, H2, EtOH.

Derivatives of DMDP have been synthesized and tested. Wongg‘IO reported the synthesis
of 2(R),5(S)-bis(hydroxymethyl)-3(R),4(R)-dihydroxypyrrolidine 17, the 5-epimer of

DMDP 1 (ﬁgure 5.2). This compound was synthesized via a thermodynamically

185

controlled fructose-1,6-diphosphate (FDP) aldolase reaction with racemic 2-azido-3-
hydroxypropanal and dihydroxyacetone phosphate (DHAP). The azide group was then
reduced and the ketone group underwent an intramolecular reductive amination. In the
aldol reaction two diasteriomeric products are formed but they can be interconverted
through enolization and only one product (the thermodynamically favorable one) is
eventually formed. This compound exhibits a broad spectrum of inhibition against
glycosidases with K; in the micromolar range. However, this method of synthesis

involving enzymes is not a general method and it is limited by scale-up problems.

OH

0 .
HO\/=’\/0H J... ”OJ—\IOH 4L. "0N0" 111

N3

11 12 13

O

M IV on OH HO
H OH —> 2- V m VI
0 UFO —'> O OH ———>
N3 Nam 3 N3 OH :
HO HO

Hos OH

OH

14 15 16 17

Figure 5.2. Synthesis of 2(R),5(S)-bis(hydroxymethyl)-3(R),4(R)-dihydroxypyrrolidine. i,
MCPBA, CH3CN, 92%; ii, NaN3, NHaCl, MeOH/HZO, reﬂux, 90%; iii, NaIOa, 0°C, 5min;
iv, DHAP/FDP aldolase; v, acid phosphatase, pH 4.7, 37 °C, 78% from 11; vi, Hz/Pd,
50psi, 97%.

Another synthetic approachll (ﬁgure 5.3) to compound 17 starts from 2-O-triﬂate 18,

derived from L-gulono-1,4-lactone'2. The nitrogen was introduced to the molecule via

azide displacement of a triﬂate. A protected lactone intermediate 22 bearing a suitably

186

located triﬂate group and in which the azido group has been reduced to an amino group
was eventually formed after several steps. Ring opening of the lactone with base allowed
the intramolecular displacement of the triﬂate group by the amino group to form the
nitrogen heterocycle. Subsequent reduction and deprotection afforded the desired product

17.

 

 

no OH . "39., orsr]
I
. . °>< ° 7<
o
18 19 20
N 0 $1
34". [1
iii iv
—’ on —’
o
engorsu 01120180
21 22
[$110 $0” H0 @011
v “ Vi n
M00 M cuzorsr] "OH: H (21-12011
H
23 17

Figure 5.3. Synthesis from L-gulono-1,4-lactone. i, 0.97 eq. NaN3, DMF, 2.5 h, 75%; ii,
TBDMSOTf, pyridine, CH2C12, 85%; iii, (a) AcOH / H2O (4:1), 99% (b) TBDMSCl,
imidazole, DMF, 76% (c)Tf2O, pyridine, CH2C12, 95%; iv, H2, Pd-black, EtOAc, 99%; v,
NaOAc, MeOH, 74%; vi, (a) LiHBEl3, T HF; (b) 1% HCl in MeOH, 90%.

A more general strategy for the synthesis of 2,5-dihydroxymethyl-3,4-

dihydroxypyrrolidines using a thiazole-based aminohomologation protocol of furanoses

187

was reported by Dondoni’s groupms. This method (ﬁgure 5.4) starts ﬁom 2,3,5-tri-O-
benzyl-furanose 24. The hydroxylamine 25 was obtained as a single a-isomer by treating
the furanose with N-benzylhydroxylamine. The open-chain form (26) of the
hydroxylamine was reacted at -70°C with 2-lithiothiazole to afford the product as a
mixture of diastereoisomers. Subsequent reductive dehydroxylation using a Zn-Cu
couple'6 and cyclization yielded the product 31. Reduction and deprotection can afford
the ﬁnal homoazasugar 33. Other 2,5-Dihydroxymethyl-3,4-dihydroxypyrrolidines with
different conﬁgurations (1, 34, and 35) were also obtained by the same fashion starting
from different furanoses. This strategy is more general than other methods, however, it
still suffers ﬁ'om the drawback of the separation of the diastereoisomers, such as 29 and

30, and the use of lithium reagent is not desirable.

188

 

CE‘LI OH gen

27 8110 5 , Th . 29
——> g in
5811 N(or-r)en ' +
(11' =3 9:1 75%: H 93"
28 Bn 7 Th
53a 141-1311
30
.\“‘\Q 3110/ "9 ..e'h‘CHZOHH / ’4 MM
111
"65“ 1311635:
31 32 33
H H
/” N “ I N \\
HO 6' w“ \o Ho/’6" ..“\\\ \0H
s 2% 5
H6 6” Hx‘ HO OH
34 35

Figure 5.4. Synthesis of 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidines. Th = 2-
thiazolyl. Reagents: i, BnNHOH, 88%; ii, (AcO)2Cu, Zn, 8% for 29, 78% for 30; iii, Tf20,
pyridine, 65%; iv, TfOMe; then NaBH4; then AgNO; in MeCN-H2O, 54%; v, H2, 20%
Pd(OH)2/C; then Dowex (OH'), 86%.

The 2,5-Dihydroxymethyl-3,4-d1hydroxypyrrolidines are not only potent glycosidase

inhibitors, they are also important precursors for the synthesis of more complicated aza—

189

C-glycosides, such as inhibitors of nucleoside hydrolasesn‘w, hypoxanthine-guanine
phosphoribosyltransferases20 (HGPRTases) and Purine nucleoside phosphorylases”24
(PNP). In the synthesis of PNP inhibitor immucillin-H (42)25‘26 and its derivatives, 1,4-
imino-D-ribitol was synthesized ﬁrst following Fleet’s method”, and the extra carbon
was installed later by addition of CNCH2Li to the imine. Assembling of the 9-

deazapurine on the preformed aza-C-glycoside afforded the desired product (ﬁgure 5.5).

TBD
N
.5. E. —*_> i
‘ ' ‘a

I
§

21

 

5‘ i 0
—-OH HO OH x
36 37 38
TBDM TBDM ?°°
/N N cmcu
\
11 m
—’ -' a __’ s" 9.
55 3, § ‘5
39 4o

2
O
z
<
z:

0 0 e a
x HO OH
41 42

Figure 5.5. Synthesis of ImmH. i, (a) TBDMSCl; (b) Acetone, H+; ii, (a) NCS, pentane;
(b) LiTMP, -78°C; iii, (a) CNCH2MgX or CNCH2Li; (b) (Boc)20; iv, (a) t-
BuOCH(NMe2)2, DMF, 70°C; (b) THF, HOAc, H2O; v, (a) H2NCH2COOEt.HCl, NaOAc,

190

MeOH; (b) ClC02Bn, DBU, CH2C12, reﬂux; (c) H2, Pc/C, EtOH; (d) H2NCH=NH.H0Ac,
EtOH, reﬂux; (e) TF A.

5.2. Results and Discussions

Our research goal is to develop a simple and general strategy to synthesize
homoazasuagrs, especially the 2,5-dihydroxymethy1-3,4-dihydroxypyrrolidines. The
desired approach is to use natural carbohydrates as the starting material. Natural
glycosarnines were chosen because of their intrinsic chirality and the available acetarnido
group which could provide an amino group that is normally obtained by displacement of
a leaving group with azide and reduction. Starting from protected glycosamines, a
chromium oxidation-borane reduction strategy was developed and dihydroxypyrrolidines

17 and 34 and 43 were synthesized quickly and efﬁciently.

I7 34 43

The syntheses of the 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidines start from methyl
2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-B-D-pyranosides. The D-gluco derivatives are
readily available from a variety of highly abundant sources such as chitin. The
commercially available methyl 2-acetamido-3,4,6-tri-0-acety1-2-deoxy-B-D-
glucopyranoside 44 was converted to pyrrolidine 17 in just two steps (ﬁgure 5.6).
Chromium oxidation of glucopyranoside 44 followed by borane reduction yielded a

single isomer 17. The oxidation of acetylated B-glycopyranosides by chromium trioxide

191

has been reported by S. J. Angyal and K. James28 to afford 5-keto esters, independent of
the conﬁgurations on 02, C-3, or C-4 and we have used this method in previous
syntheses. In the present synthesis, the oxidation gave the S-keto ester 46 as an
intermediate. Deacylation of the 2-acetamido group to form a free amino group
proceeded oxidation. Some of the N-deacetylated intermediate 45 was actually isolated
and the proton NMR spectrum showed signals for three O-acetyl groups. After it is
formed, the 5-keto function can cyclize to the 2-amino group to form a hemiaminal. The
hemiaminal can be acetylated to give the product 47. This product was obtained as a
mixture of diasteroisomers with 3:2 ratio. Both diasteroisomers can be reduced by borane
to yield the ﬁnal 2,5-dihydroxymethyl-3,4-dihydroxypyrrolidines 17 with overall 78%

yield.

 

 

 

OAc OAc
AcO 0 Ac ’0 0
MO 0109 —"’” M
NH, NH, cm
45 46

   

47 17

192

Figure 5.6. Synthesis of compound 17. i, 003, HOAc, AC2O; ii, BH3, THF, 78% for two

steps.

Compounds 34 and 43 were obtained from methyl 2-acetamido-3,4,6-tri-0-acety1-2-
deoxy-B-D-allopyranoside. This starting material was synthesized from its gluco-isomer
(ﬁgure 5.7) through an oxidation-reduction sequence to invert the stereochemistry at the
C-3 position. The starting glucopyranoside 44 was deacetylated ﬁrst. Treatment of the
triol with benzaldehyde dimethyl acetal gave the 4,6-aceta1 49 with the 3-OH free. This
hydroxyl group can be oxidized to ketone by chromium trioxide-pyridine system29 or
dimethyl sulfoxide (DMSO)-acetic anhydride (AC20). Both oxidations worked well in
this synthesis, but the chromium oxidation was more efﬁcient. In less than 10 minutes,
the reaction was done and the yield was over 90%. The DMSO-AC2O oxidation took
longer time and some competing reactions to give small amounts of byproducts were
observed. Reductions of the ketone to give 3-axia1 hydroxyl group of 51 were tried under
different conditions. Products with axial and equatorial products are possible. The desired
product has the axial orientation. Table 5.1 shows the ratio of products with equatorial
versus axial hydroxyl group. Sodium borohydride and lithium borohydride gave the
equatorial hydroxyl as the major product, while cyanoborohydride gave the opposite
result. The reason for this is that for a smaller borohydride, it tends to attack from the
axial position, while cyanoborohydride tends to attack from the equatorial position.
Another possible reason for cyanoborohydride is it can form some complex with oxygen
at 3-position and nitrogen at 2-position to form a ﬁve membered ring, so it locks the

oxygen at axial position. The reduction result also showed that at lower temperature,

193

more equatorial product 49 can be obtained, and solvent has little effect on the ratio of
the products. This is probably because lower temperature favors the equatorial attack and

the formation of the complex is not favored.

Table 5.1. Product ratios under different conditions.

 

 

Conditions equatorial : axial (49 : 51)
NaBHa/MeOH, room temp. 3 : 1
NaBHa/MeOH, -10°C ~ 0°C 4 : 1
NaBH4/20%[CH3(CH2)3]NHS04 3 : 1
MeOH/Propanol, room temp.

NaBH4/20%[CH3(CH2)3]NHSO4 2.5 : 1
MeOH/Propanol, -lO°C ~ 0°C

NaBHa/ [CH3(CH2)3]NH804 7 : 1
Propanol, -78°C

LiBHa/MeOH, room temp. 5 : 1
NaBCNH3fMeOH, room temp. 1 : 4
NaBCNH3/MeOH, -10°C ~ 0°C 1 : 2
NaBCNH3/Propanol, room temp. l : 3.5

 

With compound 51 in hand, methyl 2-acetamido-3,4,6-t1i-0-acetyl-2-deoxy-B-D-
allopyranoside was obtained by acid deprotection and acetylation. Subsequent chromium
oxidation afforded the desired ﬁve membered ring product 54 as an imine form instead of
the hemiaminal. Borane reduction of this imine yielded two isomers with hydroxymethyl
group up (43) or down (34) at 5-position. Acetylation of the mixture and separation by

chromatography yielded products 55 and 56.

194

 

MR/g/OM.‘ ——-> 'I'O':iI-;&/OMQIII/v —> OH.

NHAc "HM
44 4s 49
OH
Fri/Va Pit/T Ho 0
0 one
A 0M0 i. m1”; 0M0 .l. H NHAc
0 MIN: 0"
so 51 52
0
Ac Ac /N 0”.
——’Vi Ac 0M0 —"v" .. -. _.v111
i 2
OAc "HA“ Act? OAc
53 54
H
HO/II’ yo“ ”V”
villi
+ —>
HO OH HO OH
34 43
Ac Ac
AGO/4’ N OAc A N OAc
+
AcO OAc AeO 0A1:
55 56

Figure 5.7. Synthesis of products 55 and 56. i, K2CO3, MeOH; ii, Benzaldehyde
dimethyl acetal, H2804, DMF, 74% for two steps; iii, CrO3-pyridine, CH2C12, 91% or
DMSO-AC2O, 82%; iv, NaCNBl—l3, MeOH, 76%; v, 5% HCl, MeOH; vi, A020, pyridine,
75% for two steps; vii, Cr03, HOAc, AC2O; viii, BH3-THF; viiii, AC2O, pyridine, 69% for

This synthetic approach we describe here is a quick and easy way to get 2,5-

dihydroxymethyl-3,4-dihydroxypyrrolidines from different methyl 2-acetamido-3,4,6-tri-

195

0-acety1-2-deoxy-B-D-pyranosides. It is the most efﬁcient route demonstrated to date.
Here we take advantages of the multiple chiral centers of carbohydrate and the amino
group of the glycosamines to form homoazasugars which are normally difﬁcult to
synthesize. More complicated homoazasugars should be synthesized through this strategy
and more glycosidase inhibitors should be obtained. In this approach, different oxidations
and reductions have been applied at different stages. Chromium oxidation showed its
efﬁciency in the ketone and ketoester formations. Different borohydrides are useful in
reducing the 3-ulose to invert the stereochemistry of the hydroxyl group to yield the

desired product.

5.3. Experimental Section

General Procedures: Melting points were measured on a Ficher-Johns melting point
apparatus. Optical rotations were measured (7» = 589 nm) at room temperature using a
Jasco P-lOlO polarimeter. The 'H (and '3C) NMR spectra were recorded at 500 (125.5)
MHz on a Varian VXR spectrometer. The HRMS FAB mass spectra were obtained using

a Jeol HX-l 10 double-focusing mass spectrometer operating in positive ion mode.

2(R),5(S)-Bis(hydroxymethy1)-3(R),4(R)-dihydroxypyrrolidine (17) To a solution of
methyl 2-acetomido-3,4,6-tri-0-acety1-2-deoxy-B-D-glucopyranoside 44 (0.5g, 1.38mmol)
in acetic acid (50mL) and acetic anhydride (5an), chromium trioxide (0.42g, 4.2mmol)
was added and the suspension was stirred at room temperature for 5 hours. The mixture
was then poured slowly into cold water (300mL). The water was extracted 5 times with

CH2C12 and the combined organic phase was washed with brine, saturated sodium

196

bicarbonate and dried (N a2S04), concentrated. The resulting residue was passed through
a small pad of silica gel to give 47 as a colorless oil (0.44g, 95%). The mixture was used
in the next step without separating the diasteroeoisomers.

A solution of 47 (0.1 g, 0.3mmol) and BH3-THF (10mL, 1.5M) in anhydrous THF (lOmL)
was reﬂuxed for 4 hours and the TLC and NMR showed the completion of the reduction.
The solvent was removed and methanol was added and concentrated for 3 times. The
residue was puriﬁed by column chromatography (6:1 CH2Cl2/Me0H) to afford a
colorless oil (0.04g, 82%). 1H NMR (500MHz, CD3OD) 8 3.98 (1H, dd, J=6.0, 4.0 Hz),
3.95 (1H, t, J=4.0 Hz), 3.76-3.59 (4H, m), 3.06 (1H, dd, J=10.0, 4.5 Hz), 2.72 (1H, dd,
J=7.5, 4.0 Hz); 13C NMR (125.5MHz, CD3OD) 8 77.07, 76.71, 69.84, 65.12, 61.09,

60. l 6ppm.

Methyl 2-acetamido-4,6-benzylidine-2-deoxy-B-D-glucopyranoside (49) To a solution
of methyl 2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-B-D-glucopyranoside 44 (lg,
2.77mmol) in 20mL methanol and 0.5mL water was added potassium carbonate (2.5g,
l8mmol), and the suspension was stirred at room temperature for 12 hours. The mixture
was then diluted with 50mL methanol and 25mL water and was passed through a mixed
bed column (Dowex MR-3). The elution was concentrated and dissolved in 200mL
anhydrous DMF. Benzaldehyde dimethyl acetal (0.15g, 1.0mmol) and 30 1.1L of sulﬁrric
acid were added and the solution was stirred at room temperature for 16 hours. The
mixture was then poured into ice containing 2.5g of potassium carbonate and stirred until
all the ice melted. The mixture was ﬁltered, washed with water and dried to yield white

solid (0.66g, 74%). IH NMR (500MHz, DMSO) 8 7.79 (1H, d, J=9.0 Hz), 7.45-7.43 (2H,

197

m), 7.37-7.36 (3H, m). 5.59 (1H, s), 5.24 (1H, d, J=5.5 Hz), 4.38 (1H, d, J=8.5 Hz), 4.20
(1H, q, J=5.0 Hz), 3.72 (1H, t, J=10.0 Hz), 3.57 (1H, m), 3.73 (1H, m), 3.53 (1H, t, J=9.0
Hz), 3.41 (1H, t, J=9.0 Hz), 3.33 (3H, s), 1.81 (3H, s); 13c NMR (125.5MHz, DMSO) 8
169.13, 137.74, 128.77, 127.95, 126.30, 102.36, 100.64, 81.27, 70.52, 67.86, 65.95, 55.93,

23.04 ppm. HR-FABMS (M+H+) Calcd. 324.1446, found 324.1447.

Ketone 50: Chromium oxidation: Chromium trioxide (1.24g, 12.4mmol) was added to a
stirred solution of pyridine (1.96g, 24.8mmol) in 20mL dry dichloromethane. The
mixture was stirred for 15 min and 49 (1g, 3.1mmol) in 10mL dichloromethane was
added with stirring at room temperature. Acetic anhydride (1.26g, 12.4mmol) wad added
immediately and the reaction was stirred for another 10 min. The reaction mixture was
diluted with ethyl acetate and passed through a short column of silica gel. The product
was eluted with ethyl acetate and acetone, concentrated to yield a white solid (0.9g, 91%).
'H NMR (500MHz, DMSO) 8 7.44-7.39 (5H, m), 5.68 (1H, s), 4.68 (2H, m), 4.47 (1H, t,
J=8.0 Hz), 4.38 (1H, q, J=5.0 Hz), 3.89 (1H, t, J=10.0 Hz), 3.64 (1H, m), 3.43 (3H, s),
1.89 (3H, s); l3C NMR (125.5MHz, DMSO) 8 169.29, 137.05, 129.08, 128.13, 126.25,
103.46, 100.39, 80.91, 68.18, 65.65, 60.71, 56.41, 22.46 ppm. HR-FABMS (M+H‘)
Calcd. 322.1291, found 322.1292.

MDSO-AC2O oxidation: To a solution of 49 (0.5g, 1.6mmol) in 10mL DMSO was added
acetic anhydride (0.95g, 9.6mmol) and the solution was stirred at room temperature for
12 hours. The solution was then poured into ice water containing 3g potassium carbonate.

The mixture was ﬁltered, washed with water and dried to give a white solid (0.41 g, 82%).

198

Methyl 2-acetamido—4,6—benzylidine-2deoxy—ﬂ-D—allopyranoside (51) Sodium
cyanoborohydride (0.39g, 6.2mmol) was added to a mixture of 50 (lg, 3.1mmol) in
300mL methanol. The reaction was stirred at room temperature for 2 hours and
concentrated. The two diastereoisomers were separated by ﬂash chromatography (1:1
ethyl acetate/hexanes) and 51 was obtained as white solid (0.76g, 76%). lH NMR
(500MHz, DMSO) 8 7.44-7.40 (5H, m), 5.62 (1H, s), 5.40 (1H, d, J=4.0 Hz), 4.55 (1H, d,
J=8.5 Hz), 4.23 (1H, q, J=5.0 Hz), 3.94 (1H, m), 3.85 (1H, m), 3.73 (1H, m), 3.68 (1H, t,
J=10.0 Hz), 3.28 (3H, s), 1.82 (3H, s); I3C NMR (125.5MHz, DMSO) 8 169.29, 28.76,
127.94, 126.34, 124.88, 100.61, 100.26, 78.53, 68.30, 67.21, 62.79, 56.10, 22.68 ppm.

HR-FABMS (M+H*) Calcd. 324.1446, found 324.1445.

Methyl 2-acetamido-3,4,6-tri-0-acetyl-2-deoxy-li-D-allopyranoside (53) To a solution
of 51 (0.5 g, 1.6mmol) in 100ml. methanol was added 5mL hydrochloric acid and solution
was stirred at 5°C for 1 hour. Potassium carbonated (5g) was added and stirred for 10min,
and the mixture was ﬁltered and the solution was concentrated. The residue was
dissolved in 20mL pyridine and 15mL acetic anhydride. The solution was stirred at room
temperature for 2 hours and concentrated. The residue was puriﬁed by ﬂash
chromatography (4:1 ethyl acetate/hexanes) to give a colorless oil (0.42g, 75%). 1H NMR
(500MHz, CDCl3) 8 5.58 (1H, t, J=3.0 Hz), 4.99 (1H, dd, J=9.5, 3.0 Hz), 4.60 (1H, d,
J=8.0 Hz), 4.29 (1H, m), 4.23 (2H, d, J=4.0 Hz), 4.03 (1H, m), 3.49 (3H, s), 2.16 (3H, s),
2.09 (3H, s), 1.994 (3H, 8), 1.991 (311, s); 13C NMR (125.5MHz, CDCl3) 8 170.76,
169.66, 169.65, 169.06, 100.32, 70.50, 69.22, 66.83, 62.45, 56.35, 50.19, 23.28, 20.79,

20.74, 20.54 ppm. HR-FABMS (M+-H*)Ca1od. 362.1451, found 362.1450.

199

54, 34 and 43 were synthesized following the same method as 47 and 17.
2(S),5(S)-Bis(acetoxymethyl)-3(R),4(S)-diacetoxypyrrolidine (55) and meso-
bis(acetoxymethyl)-3,4-diacetoxypyrrolidine (56) The mixture of 34 and 43 were
acetylated and separated by ﬂash chromatography (2:1 ethyl acetate/hexanes) to give 55
(0.05g, 47%) and 56 (0.02g, 22%).

55: lH NMR (500MHz, CDCl3) 8 5.43 (1H, m), 5.39 (1H, m), 4.49 (1H, dd, J=1 1.5, 4.5
Hz), 4.43 (1H, m), 4.38 (1H, m), 4.30 (1H, dd, J=11.0, 4.0 Hz), 4.26 (1H, m), 4.21 (1H,
dd, J=10.0, 3.0 Hz), 2.16 (3H, s), 2.12 (3H, s), 2.09 (3H, s), 2.07 (3H, s), 2.04 (3H, s);
13C NMR (125.5MHz, CDCl3) 8 170.62, 170.13, 169.70, 169.47, 73.15, 69.87, 63.36,
61.10, 61.08, 56.86, 20.80. 20.76, 20.63, 20.44 ppm. HR-FABMS (M+H+) Calcd.
374.1451, found 374.1450.

56: 'H NMR (500MHz, CDCl3) 8 5.18 (1H, d, J=4.0 Hz), 4.05 (1H, d, J=2.5 Hz), 4.04
(1H, d, J=1.5 Hz), 3.16 (1H, m), 2.08 (3H, s), 2.07 (3H, s); l3C NMR (125.5MHz, CDCl3)

8 171.05, 170.86, 72.60, 63.84, 63.38, 20.84, 20.72 ppm.

200

 

5.4. References

(1) Fleet, G. W. J.; Smith, P. W. Enantiospeciﬁc Syntheses of Deoxymannojirimycin,
Fagomine and 2r,5r-Dihydroxymethyl-3r,4r-Dihydroxypyrrolidine from D-Glucose
Tetrahedron Lett. 1985, 26, 1469-1472.

(2) Card, P. J.; Hitz, W. D. Synthesis of 2(R),5(R)-Bis(Hydroxymethyl)-3(R),4(R)-
Dihydroxypyrrolidine - a Novel Glycosidase Inhibitor Journal of Organic Chemistry
1985, 50, 891-893.

(3) Reitz, A. B.; Baxter, E. W. Pyrrolidine and Piperidine Arninosugars from Dicarbonyl
Sugars in One-Step - Concise Synthesis of l-Deoxynojirimycin Tetrahedron Letters 1990,
31 , 6777-6780.

(4) Dureault, A.; Portal, M.; Depezay, J. C. Enantiospeciﬁc Syntheses of 2,5-Dideoxy-
2,5-Imino-D-Mannitol and 2,5-Dideoxy-2,5-Irnino-L-Iditol from D-Mannitol Synlett
1991, 225-226.

(5) Hung, R. R.; Straub, J. A.; Whitesides, G. M. Alpha-Amino Aldehyde Equivalents as
Substrates for Rabbit Muscle Aldolase - Synthesis of l,4-Dideoxy-D-Arabinitol and
2(R),5(R)-Bis(Hydroxymethyl)-3(R),4(R)-Dihydroxypyrrolidine Journal of Organic
Chemistry 1991, 56, 3849-3855.

(6) Legler, G.; Korth, A.; Berger, A.; Ekhart, C.; Gradnig, G.; Stutz, A. E. 2,5-Dideoxy-
2,5-Irnino-D-Mannitol and -D-Glucitol - 2-Step Bioorganic Syntheses from 5-Azido-5-
Deoxy-D-Glucofuranose and -L-Idofuranose - Evaluation as Glucosidase Inhibitors and
Application in Affinity Puriﬁcation and Characterization of Invertase ﬁ'om Yeast
Carbohydrate Research 1993, 250, 67-77.

(7) Baxter, E. W.; Reitz, A. B. Expeditions Synthesis of Azasugars by the Double
Reductive Amination of Dicarbonyl Sugars Journal of Organic Chemistry 1994, 59,
3175-3185.

(8) Fleet, G. W. J.; Smith, P. W. Methyl 2-Azido-3-O-Benzyl-2-Deoxy-Alpha-D-
Mannofuranoside as a Divergent Intermediate for the Synthesis of Polyhydroxylated
Piperidines and Pyrrolidines - Synthesis of 2,5-Dideoxy-2,5-Imino-D-Mannitol 2r,5r-
Dihydroxyrnethyl-3r,4r-Dihydroxypyrrolidine Tetrahedron 1987, 43, 971-978.

(9) Liu, K. K. C.; Kajimoto, T.; Chen, L. R.; Zhong, Z. Y.; Ichikawa, Y.; Wong, C. H.
Use of Dihydroxyacetone Phosphate Dependent Aldolases in the Synthesis of
Deoxyazasugars J. Org. Chem. 1991, 56, 6280-6289.

(10) Wang, Y. F.; Takaoka, Y.; Wong, C. H. Remarkable Stereoselectivity in the

Inhibition of Alpha-Galactosidase from Coffee Bean by a New Polyhydroxypyrrolidine
Inhibitor Angew. Chem, Int. Ed. Engl. 1994, 33, 1242-1244.

201

 

(11) Long, D. D.; Frederiksen, S. M.; Marquess, D. 0.; Lane, A. L.; Watkin, D. J.;
Winkler, D. A.; Fleet, G. W. J. Potential intermediates for incorporation of
polyhydroxylated prolines into combinatorial libraries Tetrahedron Lett. 1998, 39, 6091-
6094.

(12) Hubschwerlen, C. A Convenient Synthesis of L-(S)-Glycera1dehyde Acetonide from
L-Ascorbic-Acid Synthesis-Stuttgart 1986, 962-964.

(13) Dondoni, A.; Giovannini, P. P.; Perrone, D. New synthesis of pyrrolidine
homoazasugars via aminohomologation of furanoses and their use for the stereoselective
synthesis of aza-C-disaccharides Journal of Organic Chemistry 2002, 6 7, 7203-7214.

(14) Dondoni, A.; Giovannini, P. P.; Marra, A. Convergent synthesis of pyrrolidine-based
(l -> 6)- and (1 -> 5)-aza-C-disaccharides Tetrahedron Letters 2000, 41, 6195-6199.

(15) Dondoni, A.; Perrone, D. New entry to pyrrolidine homoazasugars: conversion of D-
arabinose into 2,5-anhydro-2,5-imino-D-glucitol via aminohomologation Tetrahedron
Letters 1999, 40, 9375-9378.

(16) Dondoni, A.; Perrone, D.; Rinaldi, M. Grignard addition to aldonitrones.
Stereochemical aspects and application to the synthesis of C-2-symmetiic diamino
alcohols and diamino diols Journal of Organic Chemistry 1998, 63, 9252-9264.

(17) Gopaul, D. N.; Meyer, S. L.; Degano, M.; Sacchettini, J. C.; Schramm, V. L.
lnosine-uridine nucleoside hydrolase from Crithidia fasciculata. Genetic characterization,
crystallization, and identiﬁcation of histidine 241 as a catalytic site residue Biochemistry
1996, 35, 5963-5970.

(18) Mazzella, L. J.; Parkin, D. W.; Tyler, P. C.; Fumeaux, R. H.; Schramm, V. L.
Mechanistic diagnoses of N-ribohydrolases and purine nucleoside phosphorylase J. Am.
Chem. Soc. 1996, 118, 2111-2112.

(19) Parkin, D. W.; Limberg, G.; Tyler, P. C.; Fumeaux, R. H.; Chen, X. Y.; Schramm, V.
L. Isozyrne-speciﬁc transition state inhibitors for the trypanosomal nucleoside hydrolases
Biochemistry 1997, 36, 3528-3534.

(20) Tao, W., Grubmeyer, C., Blanchard, J. S. Transition State Structure of Salmonella
typhimurium Orotate Phosphoribosyltransferase Biochemistry 1996, 35, 14-21.

(21) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - lnosine Hydrolysis,
Tight-Binding of the Hypoxanthine Intermediate, and 3rd-the-Sites Reactivity
Biochemistry 1992, 31 , 5964-5973.

(22) Kline, P. C.; Schramm, V. L. Purine Nucleoside Phosphorylase - Catalytic

Mechanism and Transition-State Analysis of the Arsenolysis Reaction Biochemistry 1993,
32,13212-13219.

202

 

(23) Kline, P. C.; Schramm, V. L. Pre-Steady-State Transition-State Analysis of the
Hydrolytic Reaction Catalyzed by Purine Nucleoside Phosphorylase Biochemistry 1995,
34, 1153-1162.

(24) Fedorov, A.; Shi, W.; Kicska, G.; Fedorov, E.; Tyler, P. C.; Fumeaux, R. H.; Hanson,
J. C.; Gainsford, G. J.; Larese, IJ. Z.; Schramm, V. L.; Almo, S. C. Transition state
structure of purine nucleoside phosphorylase and principles of atomic motion in
enzymatic catalysis Biochemistry'2001, 40, 853-860.

(25) Miles, R. W.; Tyler, P. C.; Fumeaux, R. H.; Bagdassarian, C. K.; Schramm, V. L.
One-third-the-sites transition-state inhibitors for purine nucleoside phosphorylase
Biochemistry 1998', 37, 8615-8621.

(26) Evans, G. B.; Fumeaux, R. H.; Gainsford, G. J.; Schramm, V. L.; Tyler, P. C.
Synthesis of transition state analogue inhibitors for purine nucleoside phosphorylase and
N-riboside hydrolases Tetrahedron 2000, 56, 3053-3062.

(27) Fleet, G. W. J.; Son, J. C. Polyhydroxylated Pyrrolidines from Sugar Lactones -
Synthesis of l,4-Dideoxy-l,4-Imino-D-Glucitol from D-Galactonolactone and Syntheses
of 1,4-Dideoxy-l,4-hnino-D-Allitol, 1,4-Dideoxy-1,4-Imino-D-Ribitol, and (2s,3r,4s)-
3,4-Dihydroxyproline from D-Gulonolactone Tetrahedron 1988, 44, 2637-2647.

(28) Angyal, S. J .; James, K. Oxidation of Carbohydrates with Chromium Trioxide in
Acetic Acid .1. Glycosides Australian Journal of Chemistry 1970, 23, 1209-&.

(29) Garegg, P. J.; Samuelsson, B. Oxidation of Primary and Secondary Alcohols in
Partially Protected Sugars with Chromium Trioxide-Pyridine Complex in Presence of
Acetic-Anhydride Carbohydrate Research 1978, 6 7, 267-270.

 

Chapter 6

5-Substituted Thiomorpholine Carboxylic Acids: A Compound

Class with Promising Antibacterial Activity

204

ABSTRACT

Because of the development of bacterial resistance, new classes of antibacterial agents
are in great need. Nitrogen heterocyclic systems with aromatic substituents are typical
features of the known antibiotics.The oxazolidinones and the quinolones are two
examples of such drugs. The oxazolidinones, with their novel structures and unique
mechanism, have been developed as a newest type of antibiotics based on structure-
activity relationship studies. A series of enantiopure arylsubstituted 5-phenyl-
thiomorpholine-3-carboxylic acid derivatives were synthesized from L- and D-cysteine in
just one step. These molecules have key structural features of known antibacterial drugs.
Biological testing was performed on these compounds and the results showed promising
antibacterial activities. Compounds with electron-withdrawing 2-NO2 substituent are the
most active among the series. These compounds can be a good starting point for
development of a new class of antibacterial agents and provide the opportunity for a

completely new drug class.

205

6.1. Introduction

The growing incidence of bacterial resistance to antibiotics has been a serious problem in
recent years. Drug resistance, especially multiple drug resistance, has become a primary
health concem'. There is a great need for new class of antibiotics that are more potent
and less prone to resistance development. Heterocyclic systems with nitrogens in rings
and aromatic substituents are often good antibiotics. Two common such types of
antibiotics are quinolones and oxazolidinones. Typical examples of quinolones are
Nalidixic acid, Ciproﬂoxacin 1 and Levoﬂoxacin. Ciproﬂoxacin, which has nitrogen
heterocyles, ﬂuoro substituent and other ﬁrnctionalities, is the best known drug in this
class. The mechanism of action of quinolone antibiotics is tied to the activity of two
enzymes, DNA gyrase and DNA topoisomerase IV. Both enzymes are essential enzymes
for bacterial growth. Ciproﬂoxacin targets these two enzymes and eliminates their
functions. It also actively poisons cells by trapping these two enzymes on DNA as
drug/enzyme/DNA complexes in which double-strand DNA breaks. Ciproﬂoxacin has
been used as an antibiotic for a long time, but its effectiveness has been compromised by

bacterial resistance.

F COOH

|
0 2

Ciproﬂoxacin 1

206

 

The oxazolidinones are a new class of antibacterial agents. They represent a new
development in the effective treatment of Gram-positive bacterial infections including
those caused by strains resistant to other antibiotics. Linezolid (Zyvox, Pharmacia
Corporation, Peapack, NJ), the ﬁrst marketed member of this class, is the ﬁrst
antibacterial drug in the last 30 years. It showed good efﬁcacy with an impressive
antibacterial spectrum. This includes activity against gram-positive organisms
Staphylococcus aureus, Streptococcus pneumoniae, and Enterococcus spp, which pose a
considerable threat to health. They also showed activity against vancomycin-resistant
Enterococcus faecium, and E. faecium related bacterernia3. This is related to their unique
mechanism of action. It has been generally accepted that oxazolidinones are protein
synthesis inhibitors“, and they interact with ribosomes. However, there are still some
debates as to their mode of action. Different results have been reported about more
detailed mechanism. In one study, oxazolidinones were found to bind to the SOS
ribosomal subunits, and to inhibit N-formylmethionyl-tRNA binding in a cell-free system
using puriﬁed ribosomes from E. Cold. It was then determined that linezolid inhibits
protein synthesis at the initiation step by preventing binding of the N-forrnylmethionyl-
tRNA to the 708 ribosome7. This is different from other antibiotics such as the
macrolides and streptogramins that act on the elongation step of protein synthesis.
However, another study revealed that oxazolidinones do not interfere with translation
initiation during mRNA binding or during formation of 30S pre-initiation complexess. Its
mode of action targets at a late step during initiation. 1t inhibits the puromycin-mediated
release of 35S-formylmethionine from 70S initiation complexes in a dose-dependent

manner. A more recent study suggested that oxazolidinones inhibit bacterial protein

207

 

biosynthesis by interfering with the binding of initiator fMet-tRNAiM“ to the ribosomal

peptidyltransferase P-siteg.

.QQJO
F kiNl-l

)=O
Linezolid 2
Oxazolidinones have been developed through structure-activity relationship. The core
structure of oxazolidinones is a simple heterocyclic system that is not found frequently in
drug structures. The 3-, 4-, 5-substitutions can lead to all kinds of derivatives of
oxazolidinones. The structural simplicity and possible variations allow the performance
of comprehensive structure-activity relationship studies. Such structure—activity

relationship studiesIO led to the development of the ﬁrst commercial drug of this kind,

linezolid, and will guide additional oxazolidinones marketed in the future.

/l2k X = a heteroatom (often 0 or N)

R R = alkyl or aryl
XR

Studies from a group at Du Pont were the ﬁrst in recent times to highlight the potential
importance of oxazolidinone compounds. There were two main candidates, Dup-721 and

Dup-105”"2, which originated from a series of racemic 5-halomethyl-3-phenyl-2-

208

oxazolidinones. Compound 3 is one example of this class”. Further modiﬁcations led to
the enantio pure analog 4, which showed some in vitro and in vivo activity”. From a
study of a series of compounds, some structure—activity rules began to appear. Activity
was determined to be a function of the size and nature of the 5-substituent, the nature of

the 3-substituent, and the conﬁguration at the 5-position.

(0)..

O J°L ° )1
S N O S N 0 =3
/ Lh. H2", Lh.
CI OH
3 4

OH >—Nj\o 0°91 Hie
LLNH / O Lt1er
r=° re

DUP-721 5 DUP—IOS 6

This ﬁnding is illustrated in Figure 6.1. If the molecule is oriented with the
oxazolidinone group to the left and the carbonyl group pointing up then the active
molecule is the isomer in which the substituent at the 5-position is coming forward
(proximal) and not the one in which it is receding (distal)'5. In the case of linezolid this
corresponds to the (S)-isomer. It is important to note that if the acetarnido group in
linezolid were changed to an acetoxy or thioacetarnido group, the Cahn—Ingold—Prelog
designation would be (R), hence the distal and proximal orientations are used here. The 4-
substitution on the ring has no signiﬁcant effect to the antibacterial activity, and usually it

has deleterious or indifferent effect. Another structure-activity feature of oxazolidinones

209

was that the 3-substituent was invariably a phenyl group. It was observed that molecules
containing 3-aryl substituents with 2,4 and 2,5 disubstituents had weak or no antibacterial
activity. 3,4-Disubstituted or 4-monosubstituted compounds had good activity provided

l'°. It was also observed that conjugated electron-withdrawing

the 3-substituent was smal
substituents in the 4-position of the phenyl group also had increased activity compared to

nonconjugated analog of similar lipophilicity.

N-Aryl group required for activity

/

/ O
{l
\ N 2

3 '0 Proximal

rue

C-5 acylaminomethyl group
essential for good activity

A
—> 1mm "'
2:

Figure 6.1. Typical features for oxazolidinone antibacterial activity.

Therefore, further modiﬁcations are focused on the 3-substututed N-aryl group. Three
subclasses of oxazolidinone analogs have been developed by Pharrnacia:
piperazinylphenyloxazolidinones (for example, PNU-97665)”, indolinyloxazolidinones
(for example, PNU-97456)’8 and the troponylphenyloxazolidinones (for example, PNU-
97786)'9’2°. Piperazine derivatives were selected as the candidates because of the
excellent in vitro and in vivo activity, as well as ease of synthesis. It was found that

electron withdrawing ﬂuorine atom at the meta position of the phenyl ring enhanced the

210

antibacterial activity and other desirable properties. These results led to the
monoﬂurophenyl congener PNU-100592 (eperazolid)2"22, thiomorpholine derivative

PNU-10048923 and morpholine analog PNU-100766, which became known as

linezolidmz.
O
HO O
m 52 1
.EcA. ..Qio 5}" .
F WNH WNH
7=° )=°
PUN-97665 PUN-97456
H3CHN i
o O O " °
,. LL

N3=O

FUN-97 786

Synthetic access to oxazolidinones has become a matter of high priority. Synthetic

“’26 (transition metal and enzymatiC),

strategies including asymmetric catalysis
biotransformation”, and chiral pool approachesm’29 have been reported. The introduction
of the chiral center at S-position of the ring is an important and general feature for the
development of the synthetic strategies. Methods in which a chiral 3-carbon fragment or
synthon is coupled to an ativated nitrogen compound dominate the literature on
oxazolidinone synthesis.

In the large-scale preparation of Linezolid (ﬁgure 6.2), a 3-carbon synthon l-chloro-2,3-

dihydroxypropane was reacted with a carbamate 9 to form a 3-aryl-5-hydroxymethyl-2-

211

oxazolidinone 10. The primary hydroxyl group is then transformed to an acetarnido group
by reacting with meta-nitrophenylsulphonyl chloride, and amonia, followed by

acetylation to form the 5-acetamidomethyl group”.

.. on i
F—QNosz —\=’—> O CNEQNHZEBZC—l> K2C03 0, NQNH OBn
2. PdlC, H2 90%
overall F

 

F
7 9 Ho \H

r. cr\)“\/OH
2. KOIBII
3. LDA

I.NH4OH,

MeOH 45°C

2. Ac20, 15th, NsCl O/ \N

Linezolid ‘—0 INKS 5".overall NL—koﬂ
85% overall
0N8

Figure 6.2. Large-scale preparation of Linezolid

A different approach was used in the preparation of DUP-721 (ﬁgure 6.3). In this
synthesis, an isocyanate 12 was reacted with glycidyl butyrate 13 in the presence of tri-n-
butylphosphine oxide to yield the N-arylated oxazolidinone nucleus directly. The
hydroxymethyl group is then deacylated and transformed to the acetarnidomethyl group
by the same process involving mesylation, displacement by azide, reduction, and

acylation described earlier.

212

O UBr,n-Bu3l’0 i
)—ONCO 4. NOW Xylene reﬂux 0 N O
O O
12 l3 14 L“ V

l NaOMe

 

0 j]: 1. M50 0 o
)O‘NLi/ﬂ <2.1mm3 Mia
1:: 3:23?” ‘ C on
our-721 15

Figure 6.3. Synthesis of DUP-721

The introduction of commercially important methods for preparing optically pure 3,4-
dihydroxybutyric acids and various 3- and 4-carbon derivatives3l'33 has opened up the
way to the preparation of optically pure oxazolidinones. The 3-hydroxy-y-butyrolactone
16 can be easily converted to enantiopure 3,4-dihydroxybutyramides 17. In one method”,
the 4-trityloxy derivative of this butyramide was subject to Hoffmann rearrangement to
form 5-trityloxymethyl-2-oxazolidinone 19 (ﬁgure 6.4). In another method”, protection
of enantiopure 3,4-dihydroxybutyramides with alkyl or arylboronic acids followed by
Hoffmann rearrangement yielded the free 5~hydroxymethyloxazolidinones 22 directly
(ﬁgure 6.5). These are new intermediates that can be used for the quick and efﬁcient
synthesis of oxazolidinone families. This is an important new development since
enantiopure 3,4-dihydroxybutyramides are available in only two steps from starch,

lactose, maltose, and hemicelluloses.

2l3

 

.1.il.n- 1.! ." .

 

(Jrl (3
O _N_"3’_"_22, NH, 1'91"» TrO\/'\/u\ oer/ou-
NI-I2
HO HO
H0
16 17 13

NH

V C NH ‘:
TfOW N], J H Tr0\)\/ TrO\/\/

I9 20

 

 

Figure 6.4. Synthesis of 5-trityloxymethy1-2-oxazolidinone

H0
l6 17 21

 

 

22 23

Figure 6.5. Synthesis of 5-hydroxymethyloxazolidinones

214

Oxazolidinones have showed promising potential to bacterial resistance with their novel
structure and unique mechanism. However, incidences of resistance to linezolid have
been observed in clinical isolates of Staphylococcus aureus and Enterococcus 3p”. A
promising approach to developing new antibiotics is to combine two pharmacophores of
two different classes of antibiotics in one molecule. This offers the possibility to
overcome the current resistance by targeting two different active sites. A series of
oxazolidinone-quinolone hybrids has been synthesized and studied37’38. These hybrids
that combine features of linezolid and ciproﬂoxacin may act on DNA replication and
protein synthesis simultaneously, and offer the opportunity to achieve a broader spectrum
towards resistance development. These two systems were connected through a variety of
spacers and the structure-activity relationships were studied with variations of spacer,

oxazolidinone substructures and quinolone substructures.

O
o F \ COOH
H 0% —X I / I
R,’"\,“..-K/N‘<\:)—spacor V 'f
R:

oxazolidinone-quinolone hybrids

The strategy for the synthesis of these hybrids was ﬁrst to link the oxazolidinone to the
central spacer, followed by coupling to the quinolone intermediates. A typical example is
illustrated in ﬁgure 6.6 for the synthesis of MCB 3570. The synthesis of oxazolidinones

followed the general scheme of linezolid synthesis.

215

M: n—s
o o
25 26
o
\ coon
I I
F u N N
0
° KEV WC» A
III Oi"
LJ
H \e‘
N..._\

MCB 3570

Figure 6.6. Synthesis of MCB 3570. i, spacer as O-mesylate, DMF, 70°C, 12h, 75%; ii,
H2, Pd/C MeOH, EtOAc, 2h, RT, 100%; iii, N-methylpyrrolidinone, 1-cyclopropyl-7-
chloro-6-ﬂuoro-I,4-dihydro-4-oxo-1,8-naphthyridine carboxylic acid (commercial),
iPrzEtN, 80°C, 3h, 80%.

The biological testing results showed that substituent variations on the quinolone and

oxazolidinone moieties had moderate effect on the antibacterial activity. However,

different spacers have dramatic effect on the activity as to acting primarily as a quinolone

216

or as an oxazolidinone, or retaining both activities. Hybrids containing 6-membered ring
spacers normally have weaker activity than 4- or 5-membered ring spacers, and 5-
membered ring spacers give the most potent antibacterial activity. Compounds with 3-
hydroxymethylpyrrolidinyl spacers displayed the most potent and balanced dual mode of
action. MCB 2936 and MCB 2900 which contain the two diastereomeric 3-
hydroxypyrrolidinyl spacers showed different antibacterial activity: The quinolone
character dominates in the (S)-diastereoisomer (MCB 2900) while the oxazolidinone
character dominates in the (R)-diastereoisomer (MCB 2936). MCB 3570, MCB 3650,
MCB 2900 and MCB 3355 were highly active against linezolide as well as ciproﬂoxacin

resistant Gram-positive strains (MICs $0.25 ug/ml).

O
o F \ COOH
. -—-- m
WQW “I
F

Spacers
O /°/,h'/\ \o/llh' \O’E /O
‘C a,
MCB 2936 MCB 2900 MCB 3570 MCB 3650 MCB 3355

6.2. Results and Discussions

6.2.1. Synthesis of 5-Substituted thiomorpholine carboxylic acid derivatives

Our goal is to synthesize heterocyclic compounds incorporating features of known
antibiotics, especially oxazolidinones and ciproﬂoxacin, by a simple synthetic strategy

that readily allows preparation of compound libraries. Biological testing of those

217

compounds should provide information for the development of structure-activity
relationships and provide a starting point for further modiﬁcation of the structures to
develop compounds with good antibacterial activity. Therefore, compounds with
substituted phenyl group and nitrogen, sulfur heterocyclic structure are designed and
synthesized (ﬁgure 6.7). They have some features of the known antibiotics linezolid and
ciproﬂoxacin. With amino, sulfur, and carboxylate functional groups, these compounds
can be ﬁtrther modiﬁed to generate a large number of library molecules for structure-
activity relationship study. One important advantage of these structures is their easy
synthesis with stereospeciﬁc character. All those compounds were synthesized in one

step and were obtained as single isomers in very high yields.

COOH sCOOl'l

OCN‘FQ—NEQNH pmﬁom
70

H A

Linezolid Ciproﬂoncln
Figure 6.7. General structures of synthetic targets and known antibiotics

Figure 6.8 shows the general procedure for the preparation of arylsubstituted 5-phenyl—
thiomorpholine-3-carboxylic acids from L-cysteine and a phenacyl bromide. The

enantiomers of these compounds can also be synthesized from D-cysteine (ﬁgure 6.9). In

218

this synthesis, the active a-bromo group is easily displaced with sulﬁir, and the
intramolecular reductive amination leads to the heterocyclic ring. The stereoselectivity is
induced by the chiral cycteine moiety. Because of this, the reduction by sodium
cyanoborohydride proceeds stereoselectively with hydride addition from the less

hindered side.

0 COOH
3" NH: NaHC HN
\ as NaCNgli
l + argon/mo \ S
/ / COOH l
R be teine 85-94‘Vo / /
_ Y5
34 Fl
27: H R =
28: 2'-ocn3
29: 4'-OCH3 35: H
30: 2'-No2 36: 2'-0CH3
31: 3'-N02 37: 4'-OCH3
32: 4'-N02 38: 2'-No2
33: 4'-Br 39: 3'-N02
4o: «v-No2
41: 4'-Br

Figure 6.8. Synthesis of arylsubstituted S-phenyl-thiomorpholine-B-carboxylic acids

from L-cysteine.

219

O E
Br NH ,ch-Izo muco3 HN
\ HS/\/ 2 NaCNBHi /\‘
l + g CHJOHIHZO \ we S
/ / COOI-I I
88~94%
R R = D-cysteine.HCl.H20 / /
42 R
27: H R =
28: 2'-OCH3
29: 4'-OCH3 43: H
30: 2'-N02 44: 2'-OCH3
31: 3'-N02 45: 4'-OCH3
32: 4'-N02 46: 2'-NOz
33: 4'-Br 47; 3'-N02
48: 4'-N02
49: 4'-Br

Figure 6.9. Synthesis from D-cysteine

6.2.2. Antibacterial Activities of the Arylsubstituted 5-Phenyl-thiomorpholine—3-
carboxylic Acids

The arylsubstituted 5-phenyl-thiomorpholine-3-carboxylic acids were tested for their
antibacterial activity against ﬁve gram-positive bacterial strains S. aureus 43300, S.
aureus 29213, E. faecalis 51299, E. faecalis 29212, B. subtilis FY 79 and six gram-
negative strains E.aerogenes 49469, E. coli 25922, E.coli DH5 alpha, E. cloacea 49141,
P. aeruginosa 27853, Salmonella sp.35664. These compounds were assayed according to
the standard MIC testing procedure for antimicrobials. The minimum inhibitory
concentration (MIC) of an antibacterial is deﬁned as the maximum dilution of the product
that will still inhibit the growth ofa test microorganism. Serial dilutions were made of the
inhibitors in bacterial growth media. The test organisms are then added to the dilutions of

the products, incubated, and scored for growth. The gram-negative strains only showed

220

weak inhibition by these compounds, but they inhibited gram-positive strains very well.
Among the ﬁve gram-positive strains, S. aureus 43300, S. aureus 29213, E. faecalis
51299 and E. faecalis 29212 were inhibited (Figure 6.10-6.13). All seven compounds
derived ﬁ'om L-cysteine showed inhibition against E. faecalis 51299 and E. faecalis
29212 at the initial dilute concentration, and remain the same kind of inhibition even with
the increasing concentrations of the inhibitors. Compound 38 which has the electron-
withdrawing 2-nitro substituent showed good activity against S. aureus 43300 and S.
aureus 29213. It reached 77% inhibition at the concentration of 250 ug/mL against S.

aureus 43300, and 57% inhibition against S. aureus 29213 at this concentration.

S. aureus 43300 against cmpds 24 hr

0.45 .
0.4 -
0.35 -
0.3 -
025 -

A600

0.2 ..

0.15 J
+2400:

0.1 .
0.05 4

o I I I I I I I j I I I
250 125 62.5 31.25 15.63 7.81 3.91 1.95 0.98 0.49 0.0
uglml uglml uglml uglml uglml uglml uglml uglml uglml uglml uglml

COMPC C006.

 

 

Figure 6.10. Inhibitory activity against S.aureus 43300

A000

A000

S. aureus 20213-24 hrs

 

 

 

 

 

 

 

0.5 -
0.45 -
0.4 5
0.35 -
0.3 -
0.25 -
0.2 1
015 .1 +2-802
0.1 l
0.05 .
o I I I I f I I I T I j
250 125 52.5 31.25 15.53 7.51 3.01 1.05 0.05 0.45 0.0
"WWWWWWWWWWW
compd cone.
Figure 6.11. Inhibitory activity against S.aureus 29213
E. meals 51250
0.35 -
0.3 -
o
0.25 1
0.2 - '
? 3 3 ‘ //
0.15 < l ; J: - v5“ ’2...” +11
/ . ..... _. "Lg—.2." +2-055-
o 1 d T m t 4'“
' +2402
+3'JIO2
0.05 -‘ +t-N02
-+—4’-Ir
o I I f I I I I 1 I I I

 

250 125 52.5 31.25 15.53 7.51 3.01 1.05 0.05 0.45 0.0
uglml uglml uglml uglml uglml uglml uglml uglml uglml uglml uglml
compd cone.

Figure 6.12. Inhibitory activity against E. faecalis 51299 for compound 35-41

222

E. faecalis 20212-24 hrs

 

 

 

0.3 -
0.201
0.24 /‘
e . _ .
8 0.151
( A ‘ ¢ : “4'" *H
m "’ +24»-
0-1 ‘ -.V-.. a 420m
*2”
0.05 . +345“
+4505“
-1—4'-Ir
o r I i I I I 1

 

250 125 52.5 31.25 15.53 7.51 3.51 1.55 0.55 0.45 0.0
uglml uglml uglrnl uglml uglml uglml uglml uglml uglml uglml uglml

compd cone.

Figure 6.13. Inhibitory activity against E. faecalis 29212for compound 35-41

Figure 6.14 and ﬁgure 6.15 show the inhibition activity of compounds 43-49. Compared
to compounds derived from L-cysteine, 43-49 are less active. They only showed
moderate inhibition against E faecalis 51299 and E faecalis 29212. Compound 46 with
2-N02 is still the most active derivative, with 58% inhibition at concentration 250ug/mL
and 72% inhibition at this concentration.

In conclusion, arylsubstituted 5-phenyl-thiomorpholine—3-carboxylic acids can be easily
synthesized in a stereoselective manner and provide easy access to compounds for the
library screen approach. The preliminary biological testing of the fourteen compounds
has showed antibacterial activity against some gram-positive strains and compounds
which are derived from L-cysteine generally have better antibacterial activity than their
enantiomers. Compounds with electron-withdrawing 2-N02 substituent are the most

active among the series. It can also been easily modiﬁed to many different types of

223

groups including ﬂuoro group. These compounds can be a good starting point for

development of a new class of antibacterial agents and provides the opportunity for a

 

 

completely new drug class.
E. faecalis 25212-24 hours
0.25 -
0.2 .
0.15 - e : -
9 H .2 «I l/
g ---r‘-—'" '— -- W“— A
“ ’5'“ - -*—-‘= f +11
0.1 l " ‘9‘ +20".
‘ 14°05.
+2410:
0'“ ' alt-$1102
4.4-«02
—+—4'-Ir
o I I I I I I I r If I j

 

250 125 52.5 31.25 15.53 7.51 3.51 1.55 0.55 0.45 0.0
uglml uglml ugllnl uglml uglml uglml uglml uglrnl uglml uglml uglml
compd cone.

Figure 6.14. Inhibitory activity against E. faecalis 29212 for compound 43-49

224

E. lamb 51205-24 hours

 

  
 

 

 

 

0.3 -
0.25 4
0.2 -
O
8 0.15 .
< +11
+2-011-
0.1 1 1- 11011.
+2410:
0.05 ‘ +5410:
+0410:
+155:
0 T I r 1 1 1 . . r .
250 125 52.5 31.25 15.53 7.51 3.51 1.55 0.55 0.45 0.0
ugknlughrﬂughnluglmlughrlughnlughrﬂughrlughrﬂughnlugm
compd cone.
Figure 6.15. Inhibitory activity against E. faecalis 51299 for compound 43-49
6.3. Experimental Section

6.3.1. General Procedures

Optical rotations were measured (2. = 589 nm) at room temperature using a Jasco P-lOlO
polarimeter. IR spectra were recorded on a FT-IR instrument. The 1H (and 13C) NMR
spectra were recorded at 500 (125.5) MHz on a Varian VXR spectrometer. The HRMS
FAB mass spectra were obtained using a Jeol HX-l 10 double-focusing mass

spectrometer operating in positive ion mode.

6.3.2. MIC Testing Procedure
Compounds were assayed according to the standard MIC testing procedure for

antimicrobials. Inhibitory potency was determined by spectrophotometrically measuring

225

 

the growth of the bacteria at 600 nm. The organisms used were ﬁve gram-positives
strains S. aureus 43300, S. aureus 29213, E. faecalis 51299, E. faecalis 29212, B. subtilis
FY 79 and six gram-negative strains E. aerogenes 49469, E. coli 25922, E. coli DH5 alpha,
E. cloacea 49141, P. aeruginosa 27853, Salmonella sp.35664. Serial dilutions are made
of the compounds in bacterial growth media M-H media. The concentration range from
250pg/mL, to 0.49ug/mL, Oug/mL. Bacteria were grown to reach 1.5><106 cﬁI/mL and
then added to the dilutions of the compounds. The solutions were incubated at 37°C and

the growth of the bacteria was monitored at 2, 5, 14, 18, and 24 hours.

6.3.3. Synthesis from L—cysteine

To a stirred solution of sodium bicarbonate (0.84 g, lOmmol) in 2:1 methanol/water (60
mL) was added L-Cysteine 34 (0.61 g, 5mmol) at room temperature. To the mixture, was
added the substituted phenacyl bromide followed by addition of sodium
cyanoborohydride (0.47g, 7.5mmol). The reaction mixture was stirred for 2 hours after
which NMR spectroscopy indicated the completion of the reaction. The solvent was
removed by evaporation and the residue was applied to an ion exchange column (Dowex
50WX8-400, 10g), which was washed with water (100 mL) and eluted with ammonium
hydroxide (100mL). The eluate was concentrated to yield a crystalline residue of

arylsubstituted 5-pheny1-thiomorpholine-3-carboxylic acids.
-H (35): was obtained as white solid (1.05g, 94%). lH NMR (500MHz, CD3OD) 5 7.50

(2H, d, J=7.0 Hz), 7.42-7.36 (3H, m), 4.29 (1H, dd, J=1 1.5, 2.5 Hz), 3.75 (1H, dd, J=11.0,

3.0 Hz), 3.13-2.96 (3H, m), 2 64 (1H, d, J=14.0 Hz); I3C NMR (125.5MHz, CD3OD) a

226

173.84, 140.61, 130.09, 129.91, 128.13, 64.32, 64.00, 32.42, 28.86 ppm. HR-EIMS (M+)

Calcd. 223.0667, found 223.0664. [a]20|) = -325 ° (c 0.1 1N HCl).

2-OMe (36): was obtained as white solid (1.19g, 94%). IH NMR (500MHz, D20) 6 7.32
(1H, t, J=8.0 Hz), 7.26 (1H, d, J=7.5 Hz), 6.99 (1H, d, J=8.5 Hz), 6.96 (1H, t, J=8.0 Hz),
4.45 (1H, dd, J=11.0, 2.0 Hz), 3.78 (3H, s), 3.72 (1H, dd, J=11.5, 3.0 Hz), 3.02-2.84 (3H,
m), 2.56 (1H, d, J=14.0 Hz); I3C NMR (125.5MHz, CD3OD) 8 174.17, 156.54, 130.70,
127.45, 125.95, 121.47, 112.08, 62.30, 57.99, 55.86, 29.21, 27.60 ppm. HR-EIMS (M+)
Calcd. 253.0773, found 253.0775. [(11209 = 230 ° (c 0.2 1N HCI).

4-OMe (37): was obtained as light yellow solid (1.1g, 87%). 1H NMR (500MHz, CD3OD)
8 7.40 (2H, d, J=8.5 Hz), 6.95 (2H, d, J=9.0 Hz), 4.23 ( 1H, dd, J=1 1.5, 2.5 Hz), 3.79 ( 1H,
s), 3.72 (1H, dd, J=1 1.5, 3.0 Hz), 3.10-3.03 (2H, m), 2.95 (1H, m), 2.63 (1H, dt, J=14.0,
2.0 Hz); l3C NMR (125.5MHz, CD3OD) 8 177.86, 158.74, 135.18, 128.23, 114.54, 63.04,
62.07, 55.75, 32.36, 28.77 ppm. HR-EIMS (M+) Calcd. 253.0773, found 253.0770. [a]2°D

= -26.3 ° (c 0.1 1N HCl).

2-N02 (38): was obtained as yellow solid (1.15g, 86%). IH NMR (500MHz, D20) 8 7.93
(1H, d, J=8.5 Hz), 7.70 (2H, m), 7.51 (1H, m), 4.54 (1H, dd, J=11.0, 2.0 Hz), 3.65 (1H,
dd, J=11.0, 3.0 Hz), 2.96-2.80 (3H, m), 2.71 (1H, d, J=14.0 Hz); 13C NMR (125.5MHz,
CD3OD) 8 176.08, 148.42, 134.66, 129.73, 128.00, 125.35, 63.01, 57.99, 30.96, 28.23
ppm. HR-FABMS (M+H+)Ca1cd. 269.0596, found 269.0595. [(11200 = -66.7 ° (c 0.1 1N

HCl).

3-N02 (39): was obtained as light yellow solid (1.2g, 90%). lH NMR (500MHz, CD3OD)
5 8.40 (1H, s), 8.19 (1H, dd, J=8.0, 2.0 Hz), 7.89 (1H, d, J=7.5 Hz), 7.64 (1H, t, J=8.0
Hz), 4.34 (1H, dd, J=1 1.0, 2.0 Hz), 3.68 (1H, dd, J=11.0, 2.0 Hz), 2.99-2.86 (3H, m),
2.62 (1H, d, J=13.5 Hz); 13C NMR (125.5MHz, CD30D) 5 175.54, 149.84, 144.60,
134.64, 131.17, 128.18, 122.95, 64.11, 63.57, 33.21, 29.55 ppm. HR-EIMS (M+) Calcd.

268.0518, found 268.0513. [n19D = 477 ° (0 0.15 IN HCl).

4-N02 (40): was obtained as light yellow solid (1.24g, 93%). 1H NMR (500MHz, D20) 5
7.99 (2H, d, J=8.5 Hz), 7.44 (2H, d, J=8.5 Hz), 3.97 (1H, d, J=1 1.0 Hz), 3.46 (1H, dd,
J=1 1.0, 2.5 Hz), 2.78 (1H, d, J=13.5 Hz), 2.67-2.56 (2H, m), 2.37 (1H, d, J=13.0 Hz); 13C
NMR (125.5MHz, D20) 5 178.06, 150.34, 147.01, 127.81, 124.21, 62.70, 62.03, 32.33,
28.82 ppm. HR-FABMS (M+H+) Calcd. 269.0596, found 269.0595. [a]209 = -18.40 (c 0.1

1N HCl).

4-Br (41): was obtained as light red solid (1.283, 85%). IH NMR (500MHz, CD3OD) 5
7.57 (2H, d, J=8.5 Hz), 7.43 (2H, d, J=8.5 Hz), 4.33 (1H, dd, J=11.5, 2.0 Hz), 3.77 (1H,
dd, J=11.5, 2.5 Hz), 3.10-2.95 (3H, m), 2.69 (1H, d, J=14.0, Hz); 13C NMR (125.5MHz,
CD3OD) 5 174.06, 138.08, 132.36, 129.00, 122.83, 62.44, 62.08, 31.05, 27.64 ppm. HR-

EIMS (1W) Calcd. 300.9772, found 300.9777. [(11209 = -234 ° (c 0.1 1N HCl).

6.3.4. Synthesis from D-cysteine
To a stirred solution of sodium bicarbonate (0.084 g, lmmol) in 2:1 methanol/water (15

mL) was added D-Cysteine 42 (0.08g, 0.5mmol) at room temperature. To the mixture,

was added the substituted phenacyl bromide (0.5 mmol) followed by addition of sodium
cyanoborohydride (0.047 g, 0.75mmol). The reaction mixture was stirred for 2 hours after
which time NMR spectroscopy indicated the completion of the reaction. The solvent was
removed by evaporation and the residue was applied to an ion exchange column (Dowex
50WX8-400, 10g), which was washed with water (20 mL) and eluted with ammonium
hydroxide (20mL). The elute was concentrated to yield a crystalline residue of

arylsubstituted 5-phenyl-thiomorpholine-3-carboxylic acids.

-H (43): was obtained as white solid (0.10g, 91%). lH NMR (500MHz, CD3OD) 5 7.50
(2H, dd, J=8.0, 1.5 Hz), 7.45-7.39 (3H, m), 4.37 (1H, dd, J=11.5, 2.5 Hz), 3.80 (1H, dd,
J=11.5, 3.0 Hz), 3.13-3.00 (3H, m), 2.73 (1H, dd, J=14.5, 2.0 Hz); 13C NMR (125.5MHz,
D20) 5 173.82, 138.18, 129.44, 129.41, 127.36, 127.00, 62.49, 62.27, 30.43, 27.27 ppm.

HR-FABMS (M+H+) Calcd. 224.0745, found 224.0744. [(11200 = + 32.9 ° (c 0.1 1N HCl).

2-OMe (44): was obtained as white solid (0.118g, 93%). 'H NMR (500MHz, 1320) s 7.29
(1H, t, J=8.0 Hz), 7.23 (1H, d, =7.5 Hz), 6.97 (1H, d, J=8.5 Hz), 6.93 (1H, t, J=7.5 Hz),
4.47 (1H, d, J=1 1.0 Hz), 3.75 (3H, s), 3.73 (1H, dd, J=1 1.0, 3.0 Hz), 3.02-2.82 (3H, m),
2.58 (1H, d, J=14.5 Hz); l3C NMR (125.5MHz, 1320) 5 173.88, 156.47, 130.73, 127.39,
125.56, 121.39, 112.00, 62.19, 57.99, 55.75, 29.04, 27.43 ppm. HR-FABMS (M+H+)

Calcd. 254.0851, found 254.0851. [612% -- +237 ° (c 0.1 1NHC1).

4-OMe (45): was obtained as white solid (0.119g, 94%). IH NMR (500MHz, D20) 5 7.26

(2H, d, J=8.5 Hz), 6.88 (2H, d, J=8.0 Hz), 4.21 (1H, d, J=1 1.0 Hz), 3.75 (l H, dd, J=12.0,

229

3.0 Hz), 3.67 (1H, s), 3.03-2.84 (3H, m), 2.52 (1H, d, J=14.0 Hz); 13C NMR (125.5MHz,
020) 8 173.03, 159.62, 129.93, 128.70, 114.76, 62.22, 61.98, 55.56, 29.97, 27.00 ppm.

HR-FABMS (M+H+) Calcd. 254.0851, found 254.0850. [(11209 = +27.8 ° (c 0.1 1N HCl).

2-N02 (46): was obtained as yellow oil (0.118g, 88%). 'H NMR (500MHz, D20) 5 7.89
(1H, d, J=8.5 Hz), 7.65 (2H, m), 7.46 (1H, m), 4.54 (1H, d, J=10.5 Hz), 3.64 (1H, dd,
J=1 1.5, 3.0 Hz), 2.93-2.77 (3H, m), 2.67 (1H, dd, J=13.5, 2.0 Hz); 13C NMR (125.5MHz,
CD3OD) s 177.96, 150.38, 138.18, 134.43, 129.77, 129.44, 125.06, 64.72, 59.47, 33.75,
30.42 ppm. HR-FABMS (M+H*) Calcd. 269.0596, found 269.0595. [(11200 = +68.4 ° (c

0.1 1N HCl).

3-N02 (47): was obtained as white solid (0.12g, 90%). 'H NMR(500M1-Iz, 1320) 5 8.16
(1H, s), 8.04 (1H, d, J=8.0 Hz), 7.69 (1H, d, J-—7.0 Hz), 7.49 (1H, t, J=8.0 Hz), 4.19 (1H,
d, J=11.5 Hz), 3.59 (1H, d, J=11.0 Hz), 2.84-2.71 (3H, m), 2.55 (1H, d, J=13.5 Hz); 13C
NMR (125.5MHz, CD3OD) s 175.85, 148.19, 142.15, 133.72, 130.39, 123.63, 121.96,
62.53, 61.66, 31.17, 27.99 ppm. HR-FABMS (M+H*) Calcd. 269.0596, found 269.0597.

[camp = +19.5 ° (c 0.15 IN HCl).

4-N02 (48): was obtained as light yellow solid (0.119g, 89%). lH NMR (500MHz, D20)
5 8.13 (2H, d, J=8.5 Hz), 7.54 (2H, d, J=8.5 Hz), 4.33 (1H, d, J=11.0 Hz), 3.69 (1H, dd,
J=1 1.5, 2.5 Hz), 2.93-2.79 (3H, m), 2.64 (1H, d, J=13.5 Hz); l3C NMR (125.5MHz, D20)
5 174.68, 147.77, 146.32, 128.12, 124.42, 62.29, 61.79, 30.72, 27.60 ppm. HR-FABMS

(M+H*) Calcd. 269.0596, found 269.0595. [er-0., = + 19.4° (c 0.1 1N HCI).

230

4-Br (49): was obtained as white solid (0.137g, 91%). 1H NMR (500MHz, D20) 5 7.29
(2H, d, J=8.0 Hz ), 7.10 (2H, d, J=7.5 Hz), 3.99 (1H, d, J=11.0 Hz), 3.51 (1H, d, J=10.5
Hz), 2.83-2.71 (3H, m), 2.77 (1H, d, J=13.5, Hz); 13C NMR (125.5MHz, D20) 8 174.51,
138.44, 132.25, 128.88, 122.57, 62.37, 61.99, 31.06, 27.66 ppm. HR-FABMS (M+H+)

Calcd. 301.9850, found 301.9849. [111300 = +239 ° (c 0.1 1N HCl).

231

6.4. References

(1) Ford, C. W.; Hamel, J. C.; Stapert, D.; Moerman, J. K.; Hutchinson, D. K.; Barbachyn,
M. R.; Zurenko, G. E. Oxazolidinones: New antibacterial agents Trends in Microbiology
1997, 5, 196-200.

(2) Pfaller, M. A.; Jones, R. N.; Doem, G. V.; Kugler, G. Bacterial pathogens isolated
from patients with bloodstream infection: Frequencies of occurrence and antimicrobial
susceptibility patterns from the SENTRY antimicrobial surveillance program (United
States and Canada, 1997) Antimicrobial Agents and Chemotherapy 1998, 42, 1762-1770.

(3) Rybak, M. J.; Cappelletty, D. M.; Moldovan, T.; Aeschlimann, J. R.; Kaatz, G. W.
Comparative in vitro activities‘and postantibiotic effects of the oxazolidinone compounds
eperezolid (PNU-100592) and linezolid (PNU-100766) versus vancomycin against
Staphylococcus aureus, coagulase-negative staphylococci, Enterococcus faecalis, and
Enterococcus faecium Antimicrobial Agents and Chemotherapy 1998, 42, 721-724.

(4) Shinabarger D., M. K., Murray R., Lin A., Melchior E., Swaney S., Dunyak D.,
Demyan W. and Buysse J. Mechanism of action of oxazolidinones: effects of linezolid
and eperezolid on translation reactions. Antimicrob Agents Chemother 1997, 41, 2132-

2136.

(5) Lin A., M. R., Vidmar T. and Marotti K. The oxazolidinone eperezolid binds to the
50S ribosomal subunit and competes with binding of chloramphenicol and lincomycin
Antimicrob Agents Chemother 1997, 41 , 2127-2131.

(6) Swaney S., A. H., Ganoza M. and Shinabarger D. The oxazolidinone linezolid inhibits
initiation of protein synthesis in bacteria. Antimicrob Agents Chemother 1998, 42, 3251-
3255.

(7) Zurenko, G. E. G., J. K.; Shinabarger, D. L.; Aristoff, P. A.; Ford, C. W.; Tarpley,
W.G. Oxazolidinones: a new class of antibacterials Current Opinion in Pharmacology
2001, 1, 470-476.

(8) Burghardt, H.; Schimz, K. L.; Muller, M. On the target of a novel class of antibiotics,
oxazolidinones, active against multidrug-resistant Gram-positive bacteria F ebs Letters
1998, 425, 40-44.

(9) Patel, U.; Yan, Y. F.; Hobbs, F. W.; Kaczmarczyk, J.; Slee, A. M.; Pompliano, D. L.;
Kurilla, M. G.; Bobkova, E. V. Oxazolidinones mechanism of action: Inhibition of the
ﬁrst peptide bond formation J. Biol. Chem. 2001, 276, 37199-37205.

(10) Barbachyn, M. R.; Ford, C. W. Oxazolidinone structure-activity relationships
leading to linezolid Angew. Chem, Int. Ed. Engl. 2003, 42, 2010-2023.

(11) Gregory, W. A.; Brittelli. D. R.; Wang, C. L. J.; Wuonola, M. A.; McRipley, R. J.;
Eustice, D. C.; Eberly, V. S.;: Bartholomew, P. T.; Slee, A. M.; Forbes, M. Antibacterials
- Synthesis and Structure Activity Studies of 3-Aryl-2-Oxooxazolidines .1. The B-Group
J. Med. Chem. 1989, 32, 1673-1681.

(12) Riedl, B.; Endermann, R. Recent developments with oxazolidinone antibiotics
Expert Opinion on Therapeutic Patents 1999, 9, 625-633.

(13) R. B. Fugitt, R. w. L. D. US 4128654 1978.
(14) (DuPont), w. A. G. US 4461773 1984.

(15) Eustice, D. C.; Brittelli, D. R.; Feldman, P. A.; Brown, L. J.; Borkowski, J. J.; Slee,
A. M. An Automated Pulse Labeling Method for Structure Activity Relationship Studies
with Antibacterial Oxazolidinones Drugs under Experimental and Clinical Research
1990, 16, 149-155.

(16) Park, C. H.; Brittelli, D. R.; Wang, C. L. J.; Marsh, F. D.; Gregory, W. A.; Wuonola,
M. A.; McRipley, R. J.; Eberly, V. S.; Slee, A. M.; Forbes, M. Antibacterials - Synthesis
and Structure Activity Studies of 3-Ary1-2-Oxooxazolidines .4. Multiply-Substituted Aryl
Derivatives J. Med. Chem. 1992, 35, 1156-1165.

(17) D. K. Hutchinson, M. R. B., S. J. Brickner, R. B. Gammill, M. V. Patel (Upjohn) US
5547950 1996.

(18) (Upjohn), S. J. B. US 5164510 1992.

(19) Barbachyn, M. R.; Toops, D. S.; Ulanowicz, D. A.; Grega, K. C.; Brickner, S. J.;
Ford, C. W.; Zurenko, G. E.; Hamel, J. C.; Schaadt, R. D.; Stapert, D.; Yagi, B. H.;
Buysse, J. M.; Demyan, W. F .; Kilburn, J. 0.; Glickman, S. E. Synthesis and antibacterial
activity of new tropone-substituted phenyloxazolidinone antibacterial agents .1.
Identiﬁcation of leads and importance of the tropone substitution pattern Bioorg. & Med.
Chem. Lett. 1996, 6, 1003-1008.

(20) Barbachyn, M. R.; Toops, D. S.; Grega, K. C.; Hendges, S. K.; Ford, C. W.; Zurenko,
G. E.; Hamel, J. C.; Schaadt, R. D.; Stapert, D.; Yagi, B. H.; Buysse, J. M.; Demyan, W.
F .; Kilburn, J. 0.; Glickman, S. E. Synthesis and antibacterial activity of new tropone-
substituted phenyloxazolidinone antibacterial agents .2. Modiﬁcation of the phenyl ring -
The potentiating effect of ﬂuorine substitution on in vivo activity Bioorg. & Med. Chem.
Lett. 1996, 6, 1009-1014.

(21) M. R. Barbachyn, S. J. B., D. K. Hutchinson (Upjohn) US 5688792 1997.
(22) Brickner, S. J .; Hutchinson, D. K.; Barbachyn, M. R.; Manninen, P. R.; Ulanowicz,

D. A.; Garmon, S. A.; Grega, K. C.; Hendges, S. K.; Toops, D. S.; Ford, C. W.; Zurenko,
G. E. Synthesis and antibacterial activity of U-100592 and U-100766, two oxazolidinone

233

 

antibacterial agents for the potential treatment of multidrug-resistant Gram-positive
bacterial infections J. Med. Chem. 1996,39, 673-679.

(23) Barbachyn, M. R.; Hutchinson, D. K.; Brickner, S. J.; Cynamon, M. H.; Kilburn, J.
0.; Klemens, S. P.; Glickman, S. E.; Grega, K. C.; Hendges, S. K.; Toops, D. S.; Ford, C.
W.; Zurenko, G. E. Identiﬁcation of a novel oxazolidinone (U-100480) with potent
antimycobacterial activity J. Med. Chem. 1996, 39, 680-685.

(24) Schaus, S. E.; Jacobsen, E. N. Dynamic kinetic resolution of epichlorohydrin via
enantioselective catalytic ring opening with TMSN(3). Practical synthesis of aryl
oxazolidinone antibacterial agents Tetrahedron Lett. 1996. 37, 7937-7940.

(25) Cardillo, G.; Orena, M.; Sandri, S.; Tomasini, C. An Efﬁcient Synthesis of (R)-(+)-
Propranolol and (S)-(-)-Propranolol from Resolved 5-Iodomethyloxazolidin-2-Ones
Tetrahedron 1987, 43, 2505-2512.

(26) Sugiyama, S.; Morishita, K.; Ishii, K. A three-step synthesis of optically active 5-
halomethy1-2-oxazolidinones; Asymmetric desymmetrization of prochiral 1,3-dihalo-2-
propyl carbamates Heterocycles 2001, 55, 353-364.

(27) Garcia-Urdiales, E.; Rebolledo, F.; Gotor, V. Enzymatic ammonolysis of ethyl (+/-)-
4-chloro-3-hydroxybutanoate. Chemoenzymatic syntheses of both enantiomers of
pyrrolidin-3-ol and 5-(chloromethyl)-1,3-oxazolidin-2-one Tetrahedron Assym. 1999, 10,
721-726.

(28) Lohray, B. B.; Baskaran, S.; Rao, B. S.; Reddy, B. Y.; Rao, 1. N. A short synthesis of
oxazolidinone derivatives linezolid and eperezolid : A new class of antibacterials
Tetrahedron Lett. 1999, 40, 4855-4856.

(29) Danielmeier, K.; Steckhan, E. Efﬁcient Pathways to (R)-5-Hydroxymethyl-2-
Oxazolidinone and (S)-5-Hydroxymethylr2-Oxazolidinone and Some Derivatives
Tetrahedron Assym. 1995,6, 1181-1190.

(30) Pearlman, B. A. P., W. R.; Barbachyn, M. R.; Manninen, P.R.; Toops, D. S.; Houser,
D. J .; Fleck, T. J. (Upjohn) US$837870 1998.

(31) Wang, G. J.; Hollingsworth, R. 1. Synthetic routes to L-carnitine and L—gamma-
amino-beta-hydroxybutyric acid from (S)-3-hydroxybutyrolactone by functional group
priority switching Tetrahedron Assym. 1999, [0, 1895-1901.

(32) Hollingsworth, R. l. Taming carbohydrate complexity: A facile, high-yield route to
chiral 2,3-dihydroxybutanoic acids and 4-hydroxytetrahydrofuran-2-ones with very high
optical purity from pentose sugars J. Org. Chem. 1999, 64, 7633-7634.

(33) Wang, G. J.; Hollingsworth, R. I. Direct conversion of (S)-3-hydroxy-gamma-
butyrolactone to chiral three-carbon building blocks J. Org. Chem. 1999, 64, 1036-1038.

(34) Wang, G. J.; Hollingsworth, R. I. A simple three-step method for preparing
homochiral 5-trity1oxymethyl-Z-oxazolidinones from optically active 3-hydroxy-gamma—
butyrolactones Tetrahedron Assym. 2000, 11, 4429-4432.

(35) Hollingsworth, R. I. W., G. US Pat. 6,288,238 2001.

(36) Tsiodras, S.; Gold, H. S.; Sakoulas, G.; Eliopoulos, G. M.; Wennersten, C.;
Venkataraman, L.; Moellering, R. C.; Ferraro, M. J. Linezolid resistance in a clinical
isolate of Staphylococcus aureus Lancet 2001, 358, 207-208.

(37) Hubschwerlen, C.; Specklin, J. L.; Sigwalt, C.; Schroeder, S.; Locher, H. H. Design,
synthesis and biological evaluation of oxazolidinone-quinolone hybrids Bioorganic &
Medicinal Chemistry 2003, 11, 2313-2319.

(38) Hubschwerlen, C.; Specklin, J. L.; Baeschlin, D. K.; Borer, Y.; Haefeli, S.; Sigwalt,
C.; Schroeder, S.; Locher, H. H. Structure-activity relationship in the oxazolidinone-
quinolone hybrid series: Inﬂuence of the central spacer on the antibacterial activity and
the mode of action Bioorganic & Medicinal Chemistry Letters 2003, 13, 4229-4233.

235

111111111111111111111)

3 129