PLACE IN REFURN 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.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c:/CI_RC/DateDqundd-p.15

 

SYNTHETIC UTILIZATION OF 1,3-DIOXONIUM CATION REARRANGEMENTS:
BROMINATION, AMIDATION AND SYNTHESIS OF THERAPEUTICALLY
IMPORTANT IMINOALDITOLS

By

Xuezheng Song

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
SYNTHETIC UTILIZATION OF 1,3-DIOXONIUM CATION REARRANGEMENTS:
BROMINATION, AMIDATION AND SYNTHESIS OF THERAPEUTICALLY
IMPORTANT IMINOALDITOLS
By

Xuezheng Song

In the attempt at N-bromosuccinimide (NBS) bromination of benzylidene acetal
protected gluconoheptonolactone derivative, a dioxonium cation rearrangement involving
neighboring pivaloyl group was discovered. The rearrangement mechanism was further
explored. Based on the experimental results, a new methodology of carbohydrate
structure chirality manipulation was developed. A series of synthetically useful
selectively brominated and protected D- and L-aldonolactone derivatives were
synthesized. This methodology was further applied in the synthesis of an orthogonally

protected bromobutanetriol as a chiral building block.

To introduce new functionalities in to the sugar structure, different inter- and
intramolecular nucleophiles were employed in the NBS-mediated dioxolonium ion
rearrangement. Intermolecular nucleophiles generally failed to compete with in-Situ
generated bromide ion, although in certain cases, the introduction of competitive
nucleophiles was shown. Cyano groups do not participate in the Hanessian-Hullar
reaction as intramolecular nucleophile to give expected lactam. The results were ascribed
to the better nucleophilicity of the bromide ion generated in-situ than that of the

competitive nucleophiles introduced inter- or intramolecularly.

Facilitated by dioxonium cation formation and rearrangement under strong acidic
conditions, selective bromination of aldonolactones was carried out to generate useful
bromosugars. The anomeric bromination of hexose pentapivaloates was achieved to
generate tetrapivaloyl glycosyl bromides as useful glycosyl donors. When acetonitrile
was employed as solvent and nucleophile, Ritter-type reactions occurred to afford several
glycosyl amides. Furthermore, a unique dioxonium cation assisted Ritter-type reaction
was discovered when glycerol dipivaloates were treated with strong acid in acetonitrile.
The stereoselectivity and regioselectivity of this new Ritter-type reaction were tested with
erythritol tripivaloate and 1,2,4-butanetriol dipivaloate as the substrates. This reaction
was employed to introduce amide groups to the primary position of aldonolactone to

generate N—contained sugar derivative.

A strategy for the preparation of carbohydrate derivatives bearing two leaving
groups for use in the preparation of azasugars by di-N-alkylation of amines was
developed. The target molecules were 1-deoxynojirimycin (DNJ) and l-
deoxymannojirimycin (DMJ). The selective acetal protections of dibromoaldonolactones
and dibromoalditols were explored. DNJ was synthesized from partially protected
dibromoalditol by di-N-alkylation with ammonium hydroxide. DMJ was synthesized

from D-rnannose using the same strategy.

ACKNOWLEDGEMENT

I would like to express my gratitude to my advisor Dr. Rawle I. Hollingsworth,
for his guidance and support during my four years of graduate studies. He not only
trained me as a synthetic carbohydrate chemist, but also taught me both scientific and
practical thinking. I also want to thank my guidance committee, Dr. James E. Jackson,

Dr. Jetze J. Tepe and Dr. Katharine C. Hunt, for their helpful advice on my thesis.

I would like to thank all the members in Hollingsworth’s group. We have created
a fi'iendly and free academic atmosphere, which is essential for creative work and happy

graduate life. I also benefited very much from all my friends in chemistry and life.

I am also thankful to the people of the facilities involved in this thesis, such as the

NMR and Mass spectrometry facilities for their help.

Finally, it is my pleasure to thank my family for their unconditional love and

support in these years. The thesis would not be possible without all those.

iv

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... viii
LIST OF FIGURES .......................................................................................................... ix
LIST OF SCHEMES ........................................................................................................ xi
Chapter 1 Literature Review: Synthesis and Structure-Activity Relationship of
Azasugars as Glycosidase Inhibitors ...................................................... l

. Introduction ................................................................................................ 1

1.2 Glycosidases and glycosyl transferases and their inhibitors in carbohydrate
metabolism ................................................................................................. l

1.3 Synthetic methodologies towards azasugars .............................................. 7
1.3.1 Synthesis towards azasugars from achiral or chirality-deficient molecule
..................................................................................................................... 8

1.3.2 Synthesis of azasugars from sugars or other chirality-rich molecules ..... 18

1.3.3 Other synthetic methodologies towards azasugars .................................. 45

1.4 The structure-activity relationship (SAR) of azasugars as glycosidase
inhibitors and modified azasugars ........................................................... 48

1.4.1 Shape and charge of azasugar glycosidase inhibitors as transition state
analogues .................................................................................................. 49

1.4.2 Substitution effects of azasugars as glycosidase inhibitors ..................... 52

1.4.3 l-N-iminosugars as potent and selective inhibitors ................................. 54

1.4.4 Aromatic ring-fuzed azasugars as glycosidase inhibitors ........................ 58

1.5 Summary .................................................................................................. 6O
REFERENCE ............................................................................................................. 61
Chapter 2 NBS-Initiated Dioxonium Cation Rearrangement and Applications of
the Selective Bromination and Chirality Manipulation of
Carbohydrates ........................................................................................ 66

ABSTRACT ............................................................................................................... 66

2.1 Hanessian-Hullar reaction and dioxonium cation rearrangement ............ 66

2.2 Cascade dioxonium rearrangement of glucoheptonolactone benzylidene
derivative under Hanessian-Hullar reaction conditions ........................... 69

2.3 Selective protection, functionalization and chirality manipulation of sugar
lactones employing the NBS-initiated dioxonium cation
rearrangement .......................................................................................... 76

2.3.1 Selective 6-position bromination of protected D-gulono-y-lactone with
inversion at C-5 ........................................................................................ 76

2.3.2 Selective bromination and chirality manipulation of sugar lactone
derivatives with gluco- and galacto- configuration ................................. 79

2.4 C2 symmetric tartrate derivative desymmetrization using NBS-initiated
rearrangement .......................................................................................... 88

2.5 The exploration of NBS-initiated neighboring rearrangement of other

protecting group ....................................................................................... 90
2.6 Conclusion ............................................................................................... 91
EXPERIMENTAL ..................................................................................................... 92
REFERENCE ........................................................................................................... 106

Chapter 3 Exploration of Inter- and Intramolecular Competitive Nucleophiles in
the N-Bromosuccinimide-Mediated Dioxolonium Ion Rearrangement

for Introduction of New Functionalities ............................................ 107
ABSTRACT ............................................................................................................. 107
3. 1 Introduction ............................................................................................ 107
3.2 The base sensitivity of NBS-oxidative cleavage of protected benzylidene

sugar lactone .......................................................................................... 108

3.2.1 The effect of barium carbonate on the NBS reaction as an acid

scavenger ................................................................................................ 109

3.2.2 Pyridine assisted elimination in NBS treatment of 3,5-O-benzy1idene-

2,6,7-tri-O-t1imethylacetyl-D-glucoheptono-1 ,4-lactone l ................... l 14
3.3 The introduction of competitive nucleophiles into the NBS bromination of

3 ,5-O-benzy1idene-2,6,7-t1i-O-t1imethylacety1-D-glucoheptono- l ,4-

lactone l ................................................................................................. 116

3.3.1 The introduction of O- nucleophiles into the NBS oxidation ................ 116
3.3.2 The introduction of N-contained nucleophiles into the NBS oxidation

................................................................................................................. 121
3.4 The introduction of intramolecular competitive nucleophiles into the NBS

bromination of 3,5-O-benzylidene-2,6,7-tri-O-trimethylacetyl-D-

glucoheptono- 1 ,4-1actone ....................................................................... 128
3.5 Summary ................................................................................................ 129
EXPERIMENTAL ................................................................................................... 130
REFERENCE ........................................................................................................... 135

Chapter 4 Acid-Promoted Dioxonium Cation Facilitated Bromination and

Ritter-Type Reactions .......................................................................... 136
ABSTRACT ............................................................................................................. 136
4. 1 Introduction ............................................................................................ 1 36
4.2 Selective pivaloylation and bromination facilitated by dioxonium cation

Formation and Rearrangement ............................................................... 138

4.2.1 Synthesis of 2,6-dibromo-2,6-dideoxy-5-O—t1imethylacetyl-D-idono-1,4-

1actone .................................................................................................... 138

4.2.2 Synthesis of 6-bromo-6-deoxy-2,5-di-O—pivaloyl-D-galactono-1,4-1actone
................................................................................................................. 141
4.2.3 Synthesis of 2,6-dibromo-2,6-dideoxy-5-O-pivaloyl-D-manno-1,4-1actone

................................................................................................................. 143
4.3 Synthesis of chiral tetrasubstituted tetrahydrofuran (THF) ring from 2,6-

dibromo-Z,6-dideoxy-5-O-pivaloy1-D-idono-1,4-lactone by a weak-base

induced pivaloyl rearrangement ............................................................. 146

vi

4.4 Anomeric bromination and Ritter-type reaction of pentaacyl
monosacchrides ...................................................................................... 1 50

4.4.1 Synthesis of glucose pentapivaloate and its anomeric isomerization under
acidic condition ...................................................................................... 151

4.4.2 Synthesis of tetrapivaloyl glycosyl bromides with gluco- and manna-
configurations ........................................................................................ 1 54

4.4.3 Synthesis of glycosyl amide by dioxonium cation facilitated anomeric
Ritter-type reaction ................................................................................ 156

4.5 Dioxonium cation in protected acyclic polyols: stereoselective Ritter-type
reactions ................................................................................................. 158

4.5.1 t-Butyl-l,3-dioxonium cation as an electrophile: synthesis of N-acyl-
aminopropanediol .................................................................................. 1 58

4.5.2 The stereoselectivity of t-buty1-1,3-dioxonium cation facilitated Ritter-
type reaction: Synthesis of protected chiral bromobutanetriol and N-
acetylaminobutanetriol ........................................................................... 160

4.5.3 Dioxonium cation facilitated functional group transformation of (S)-1,2,4-
butanetriol: more mechanism and applications ...................................... 163

4.6 Synthetic application of dioxonium cation facilitated Ritter-type reaction:
synthesis of protected 6-aminoga1actonolactone ................................... 167
EXPERIMENTAL ................................................................................................... 171
REFERENCE ........................................................................................................... 188
Chapter 5 Synthesis Efforts towards l-Deoxynojirimycin and 1-
Deoxymannojirimycin ......................................................................... 1 90
ABSTRACT ............................................................................................................. 190
5.1 Some explorations of 1,5-iminosugar syntheses .................................... 190
5.2 The isopropylidene protection of dibromoaldonolactones and
dibromoalditol ........................................................................................ 1 96

5.2.1 Isopropylidene protection of dibromoaldonolactone ............................. 198
5.2.2 Isopropylidene protection of dibromoalditol ......................................... 201

5.3 The 1,2-diacetal protection of dibromoaldonolactones and
dibromoalditols ...................................................................................... 203

5.3.1 1,2-diaceta1 protection of dibromoaldonolactones ................................. 204
5.3.2 1,2-diacetal protection of dibromoalditols ............................................. 206

5.4 Synthesis of l-deoxynojirimycin (DNJ) ................................................ 209
5.5 Synthesis of l-deoxymannojirimycin (DMJ) ......................................... 212
5.6 Summary and Future Work .................................................................... 215
EXPERIMENTAL ................................................................................................... 2 1 9
REFERENCE ........................................................................................................... 238
APPENDIX .................................................................................................................... 239

vii

LIST OF TABLES

Chapter 1

Table 1.1 Several classes of inhibitors and their targets ....................................... 4
Table 1.2 Several plant alkaloids as glycosylation inhibitors ................................. 6
Table 1.3 Inhibition Potencies of Various Azasugars (uM) ................................. 50
Table 1.4 Inhibitory Potencies of 1-N-Iminosugars .......................................... 56

Table 1.5 ICso values [uM] of the annulated imidazoles with gluco-, manno- and galacto-
configurations .......................................................................... 59

viii

LIST OF FIGURES

Chapter 1
Figure 1.1 Biosynthesis of the N-linked core oligosaccharide ................................ 2
Figure 1.2 General mechanisms for inverting (a) and retaining (b) glycosidases .......... 3

Figure 1.3 The structure and charge resemblence of azasugar and the oxocarbenium
transition state .......................................................................... 7

Figure 1.4 Several important azasugars as synthetic target of glycosidase inhibitors 7
Figure 1.5 Gluconolactone as glucosidase inhibitor .......................................... 49

Figure 1.6 Preferred conformation of deoxynojirimycin, N-alkyl-deoxynojirimycin and
castanosperrnine ...................................................................... 53

Figure 1.7 Schematic representation of the interactions observed between CelSA and the
cellobio-derived isofagomine ....................................................... 57

Figure 1.8 Some azasugars filsed with A) tetrazoles; B) imidazoles; C) pyrroles; D)
1,2,3-t1iazoles ......................................................................... 58

Figure 1.9 Proposed direction of protonation A) of a glucoside perpendicular to the plane
of the ring and B) and C) of gluco- tetrazole and a glucoside 1n the plane of the
ring ..................................................................................... 60

Chapter 2

Figure 2.1 The 1H NMR spectrum of 3-O-Benzoyl-7-bromo-7-deoxy-2,5,6-t1i-O-
pivaloyl-D-glycero-D-gulo-heptono-1 ,4-lactone (1 1) ........................... 74

Figure 2.2 Evaluation of coupling constants expected for products formed by the 3
possible rearrangement mechanisms ............................................... 75

Figure 2.3 The lH NMR spectrum of 3-O-benzoyl-6-bromo-6-deoxy-2,5-di-O-
trimethylacetyl-L-manno- l ,4-1actone (29) ....................................... 78

Figure 2.4 The lH NMR spectrum of methyl 3-O-benzoy1-6-bromo-6-deoxy-2,4,5-tri-O-
trimethylacetyl-D-galactonate (40) ................................................ 84

Figure 2.5 The 'H NMR spectrum of methyl 5-O-benzoy1-4-bromo-4-deoxy-2,3,6-t1i-O—
trimethylacetyl-D-galactonate (46) ................................................ 87

ix

Chapter 3
Figure 3.1 The competitive nucleophile participated Hanessian-Hullar reaction ....... 108
Chapter 4

Figure 4.1 bromosugars generated from the acetoxonium cation facilitated bromination
serve as good synthetic precursors of iminosugars ............................. 138

Figure 4.2 A comparison of the different ring structures of D-gulono-l,4-1actone and D-
glucono- 1 ,5-lactone ................................................................ 144

Figure 4.3 The comparison of the configurations of 22 and 26: the threo-selectivity of the
weak base catalyzed pivaloyl rearrangement .................................... 149

Figure 4.4 The preference of nucleophilic attack of the t-butoxonium cation 58: to a) the
carboxonium cation; b) the secondary C-2; and c) the primary C-l ......... 160

Figure 4.5 The comparison of A) normal Ritter-type reaction and B) acyloxonium cation-
assisted Ritter-type reaction ....................................................... 163

Figure 4.6 Use of 1,2,4—butanetriol dipivaloate to clarify the active carboxonium cation
intermediate: five-membered ring or six-membered ring ..................... 164

Figure 4.7 An explanation of the regioselectivities of acyloxonium cation facilitated
bromination and the Ritter-type reaction ........................................ 166

Chapter 5

Figure 5.1 The acyloxonium cation facilitated intramolecular Ritter-type reaction in the
synthesis of sugar lactams ......................................................... 216

Figure 5.2 Chiral 1,3-dioxonium cation facilitated asymmetric bromination and Ritter-
type reaction in the synthesis of chiral building blocks ........................ 217

Figure 5.3 Chiral nitrile involved asymmetric Ritter-type reaction in the
desymmetrization of glycerol ..................................................... 217

LIST OF SCHEMES

Chapter 1
Scheme 1.1 A synthesis to castanosperrnine epimer from malic acid derivative 9
Scheme 1.2 A synthesis to swainsonine 4 from an existing bicyclic lactam ............ 10

Scheme 1.3 A synthesis to deoxymannojirimycin 2 from an achiral a-furfurylamine
derivative ........................................................................... 11

Scheme 1.4 A synthesis of L-deoxymannojirimycin 34 from dihydropyran and S-
phenylglycinol ..................................................................... 12

Scheme 1.5 A synthesis to swainsonine 4 from di-Boc protected 5-aminopentana1 13
Scheme 1.6 A synthesis to swainsonine from 4-bromobutanal ........................... 14
Scheme 1.7 A synthesis to swainsonine from a suitable enyne alcohol .................. 15

Scheme 1.8 A synthesis to deoxynojirimycin 1 from 4-methoxy-3-
(triisopropylsilyl)pyridine ....................................................... 1 6

Scheme 1.9 An asymmetric synthesis of deoxynojirimycin and castanosperrnine starting
from the same an achiral acyclicdiene ......................................... 17

Scheme 1.10 The general synthetic pathway of azasugars using azide as a nitrogen
source ............................................................................... 18

Scheme 1.11 A) Synthesis of l-deoxy-castanospermine 81 from glucose derivative; B)
Synthesis of epimer of l-deoxy-castanospermine from 2-azido-4,6-O-
benzylidene-2-deoxy-a-D-altropyranoside .................................... 1 9

Scheme 1.12 Two synthetic routes towards deoxymannojirimycin from glucose
derivative and L-gulonolactone ................................................. 21

Scheme 1.13 A synthesis of deoxynojirimycin from an epoxide with L-idofuranose
configuration ...................................................................... 23

Scheme 1.14 A synthesis of deoxynojirimycin from L-sorbose ............................ 24

Scheme 1.15 Two similar synthetic routes towards l-deoxymannojirimycin sharing a
common intermediate 6-aizdo-D-fructose ..................................... 25

Scheme 1.16 Several enatioselective synthesis of azasugars using reductive cycliation of

azide and leaving groups. A) Starting from D-gluconolactone; B) starting
from chiral lactol; C) starting from D-erythronolactone ..................... 26

xi

Scheme 1.17 Double reductive amination towards azasugars .............................. 27

Scheme 1.18 Synthesis of deoxynojirimycin by double reductive amination from 2,3,4,6-

tetra-O-benzyl-a-D-glucopyranose ............................................. 28
Scheme 1.19 The synthesis of D-hexos-S-uloses 145 and 150 as an intermediate for

azasugar synthesis ................................................................ 30
Scheme 1.20 A transformation of sugar lactones to azasugars .............................. 31
Scheme 1.21 A triple reductive amination approach to castanosperrnine ................. 32
Scheme 1.22 A triple reductive amination approach to swainsonine ...................... 33
Scheme 1.23 A norbomyl route to azasugars ................................................. 34

Scheme 1.24 A general approach to the synthesis of dideoxy and trideoxyiminoalditols

from B-D-glycosides .............................................................. 35
Scheme 1.25 A synthesis of deoxynojirimycin and castanosperrnine by a glucosylamine
pathway ............................................................................ 36
Scheme 1.26 A synthesis of castanosperrnine from a highly functional epoxide 38
Scheme 1.27 A synthesis of castanosperrnine and analogs using a "naked sugar" as the
starting material ................................................................... 39
Scheme 1.28 A synthesis to 2,5-dideoxy-2,5-imino-D-mannitol and L-iditol using double
N-alkylation of protected alditols ............................................... 40
Scheme 1.29 Synthesis of deoxynojirimycin from D—mannitol ............................. 41
Scheme 1.30 Deoxyiminoalditols from aldonolactone ....................................... 43

Scheme 1.31 A synthesis of deoxymannojirimycin analogues using N-tosyl and N-nosyl
activated aziridines derived from l-amino-l-deoxy-glucitol ............... 44

Scheme 1.32 A stereocontrolled preparation of azasugars from glycosylamine employing

anhydrosugar as the key intermediate .......................................... 45
Scheme 1.33 Syntheses deoxynojirimycin and deoxymannojirimycin based on a hi gh-
yield, ring-forming aminomercuration ......................................... 46
Chapter 2
Scheme 2.1 The regioselective cleavage of benzylidene acetals with NBS ............. 67

Scheme 2.2 The mechanism of Hanessian-Hullar reaction (a) and the water-dependent
NBS regioselective cleavage of benzylidene acetals (b) ..................... 68

xii

Scheme 2.3

Scheme 2.4

Scheme 2.5

Scheme 2.6

Scheme 2.7

Scheme 2.8

Scheme 2.9

The concept of benzoxonium cation rearrangement ......................... 69

The NBS treatment of 2,6,7-tri-O-trimethylacety1-3,5(R)-O-benzylidene-D-

glycero-D-gulo-heptono-l,4-lactone 10 ....................................... 70
Expected pathways for NBS cleavage of 3,5-(R)-O-benzylidene-2,6,7-
trimethylacetal-D-glycero-D-gulo-heptono-1,4-1actone 10 ................. 7 1

Three possible pathways that could lead to 7-bromination of 10; the third
scenario is the only one that leads to net retention ........................... 72
Selective 6—position bromination of D-gulono-1,4-lactone 26 with inversion
at C-5 ............................................................................... 77

A ring structure analysis for 3,5-O-benzylidene aldonolactones preparation
....................................................................................... 80

Selective 6-position bromination of protected D-glucono-d-lactone with
inversion at O4 to form methyl 3—O-benzoy1-6-bromo-2,4,5-tri-O-
trimethylacetyl-D-galactonate 40 ............................................... 81

Scheme 2.10 Selective 4-position bromination of protected D-galactono-g-lactone with

inversion at C-2 to form methyl 5-O-benzoyl-4-bromo-4-deoxy-2,3,6-tri-O-
trimethylacetyl-D-galactonate 46 .............................................................. 85

Scheme 2.11 Selective primary bromination of a protected L-threitol derivative with

inversion of the neighboring group ............................................. 88

Scheme 2.12 The direct bromination on the benzoxonium cation in the NBS treatment of

Chapter 3

Scheme 3.1

Scheme 3.2

Scheme 3.3

Scheme 3.4

Scheme 3.5

Scheme 3.6

2,3-O-benzy1idene- 1 ,4-di-O-benzoyl-L-threitol 54 .......................... 90

The reaction and mechanism of NBS-initiated pivaloyl rearrangement and
bromination ...................................................................... 1 10

The isomerization of bromobenzylidene lactone when BaCO3 is present in
the NBS reaction ................................................................ l 1 1

The acceleration effect of bromide ion in the NBS reaction ............... 1 13

Pyridine assisted elimination in NBS treatment of protected benzylidene
lactone ............................................................................ 115

The hydrolysis of carboxonium cation: A) The hydrolysis mechinism; B)
The regioselectivity of hydrolysis based on the stereoelectronic effect .. 1 17

The NBS treatment of benzylidenelactone with water present ............ 1 18

xiii

Scheme 3.7

Scheme 3.8

Scheme 3.9

The NBS treatment of benzylidenelactone with t-butyl alcohol in presence
...................................................................................... 119

The NBS treatment of benzylidenelactone with sodium acetate as the
competitive nucleophile ........................................................ 121

Intramolecular competitive nucleophile participated Hanessian-Hullar
reaction ........................................................................... 124

Scheme 3.10 The iminoester neighboring group participated Hanessian-Hullar reaction

Scheme 3.11

Scheme 3.12

Scheme 3.13

Chapter 4

Scheme 4.1

Scheme 4.2

Scheme 4.3

Scheme 4.4

Scheme 4.5

Scheme 4.6

Scheme 4.7

Scheme 4.8

...................................................................................... 125

Proposed intramolecular cyano group participated Hanessian-Hullar
reaction ........................................................................... 126

Attempted cyano-participated Hanessian-Hullar reaction: synthesis of 2-0-
benzoy1-3-bromo-3-deoxy-1 ,4-di-O-cyanomethy1-L-threitol ............. 1 27

Synthesis of 2-deoxy—D-ribonitrile and attempted cyano-participated
Hanessian-Hullar reaction ...................................................... 128

Acetoxonium cation generated under acidic conditions : A) dibromination
of aldonolactones; B) anhydrosugar formation .............................. 137

Synthesis of 2 ,6 dibromo- 2 ,6 dideoxy-S- O-trrmethylacetyl D-idono-1,4-
1actone4.. ..139

A) Selective monopivaloation of 2,6-dibromo-2,6-dideoxy-D-idono-1,4-
lactone 7 and B) selective tripivaloation of D-gulono-1,4-lactone 1 140

A) Synthesis of 6-bromo-6-deoxy-2,5-di-O-pivaloyl-D-galactono-1,4-
lactone 12; B) Synthesis of 6-bromo-6-deoxy—D-galactono-1,4-1actone 13
(Ref 6); C) selective tripivaloation of D-galactono-l ,4-lactone 10 ...... 142

A) Synthesis of 2,6-dibromo-2,6-dideoxy-5-O-pivaloyl-D-manno-1,4-
lactone; B) Synthesis of 2,6-dibromo-2,6-dideoxy-D-manno-1,4—lactone
(Ref 14) ........................................................................... 145

Synthesis of 2,5-anhydro-6-bromo-6-deoxy-3-O-methoxymethyl-5-O-
pivaloyl-D-gulonic amide 24 ................................................... 147

Synthesis of 6-bromo-6-deoxy—3-O-methoxymethyl-5-O—pivaloy1-D-
mannonic amide 26 ............................................................. 148

Tetraacetyl glucosyl bromide synthesis and application in glycoside
synthesis .......................................................................... 151

xiv

Scheme 4.9 Synthesis of B-D-glucose pentapivaloate from a-D-glucose and the
comparison with peracetylation under different reaction conditions 152

Scheme 4.10 The acid catalyzed anomeric isomerization of A) 0-D-
glucosepentapivaloate 29 and B) 2-deoxy-D-arabino-hexosides 34 153

Scheme 4.11 Synthesis of (It-tetrapivaloyl glucosyl bromide .............................. 154
Scheme 4.12 Synthesis of a—tetrapivaloyl mannosyl bromide ............................. 155
Scheme 4.13 Synthesis of tetrapivaloyl B- and a-glucosyl acetamides .................. 156
Scheme 4.14 Synthesis of tetraacetyl N-B- and a-glycosyl amides with gluco- and

galacto- configurations ......................................................... 157
Scheme 4.15 synthesis of (2t)-N-acy1-aminopropanediol 59 .............................. 159
Scheme 4.16 Synthesis of protected chiral bromobutanetriol 64 .......................... 161
Scheme 4.17 Synthesis of protected chiral N-acetylaminobutanetriol 65 ............... 162
Scheme 4.18 Synthesis of protected chiral N-acetylaminobutanediol 71 and 72 ....... 165

Scheme 4.19 Application of acyloxonium cation facilitated Ritter-type reaction:
synthesis of protected 6-aminogalactonolactone ............................ 169

Chapter 5

Scheme 5.1 The retrosynthetic analysis of 1,5-iminosugar synthesis by double N-
alkylation strategy ............................................................... 191

Scheme 5.2 Synthesis of azasugars from aldonolactones: a carbon chain inversion
strategy ........................................................................... 192

Scheme 5.3 Synthesis of 3,5-di-O-benzy1—2,6-dibromo-2,6—dideoxy-L-idono- -1actone 6
...................................................................................... 192

Scheme 5.4 The treatemnt of 3,5-di-O-benzyl-2,6-dibromo-2,6-dideoxy—L-idon-g—
lactone 6 with benzylamine .................................................... 193

Scheme 5.5 synthesis of 1,3,4,5-di-O-isopropy1idene-2,6-dibromo-2,6-dideoxy-L-iditol
14 and the treatment of it with benzylamine ................................. 195

Scheme 5.6 A synthetic attempt to an azasugar anolog of sialic acid as a potential
sialyltransferase inhibitor ...................................................... 195

Scheme 5.7 Synthesis of 2,6-dibromo-2,6-dideoxy-aldonolactones and alditols with (A)
D-ido and and (B) D-manno configurations ................................. 197

XV

Scheme 5.8 The isopropylidene protection of 2,6-Dibromo-2,6-dideoxy-D-idonolactone
17 .................................................................................. 198

Scheme 5.9 The isopropylidene protection of 2,6-dibromo-2,6-dideoxy-D-
mannonolactone 23 .............................................................. 200

Scheme 5.10 The isopropylidene protection of 2,6-dibromo-2,6-dideoxy-D-iditol 18

...................................................................................... 202
Scheme 5.11 The isopropylidene protection of 2,6-dibromo-2,6-dideoxy-D-mannitol 24
by acetone ........................................................................ 203
Scheme 5.12 1,2-Diacetal protection group for 1,2-diols .................................. 203

Scheme 5.131,2-Diacetal protection reaction of 2,6-dibromo-2,6-dideoxy-D-
idonolactone 17 .................................................................. 205

Scheme 5.141,2-diacetal protection reaction of 2,6-dibromo-2,6-dideoxy—D-
mannonolactone 23 .............................................................. 206

Scheme 5.15 The 1,2-diacetal protection of 2,6-dibromo-2,6-dideoxy—D-iditol 18 207

Scheme 5.16 The 1,2-diaceta1 protection of 2,6-dibromo-2,6-dideoxy—D-mannitol 24

...................................................................................... 208
Scheme 5.17 A ring-strain analysis of acetal and 1,2-diacetal protection and fused-ring
formation ......................................................................... 209
Scheme 5.18 The synthesis of 1-deoxynojirimycin(DNJ) 1 .............................. 211
Scheme 5.19 The retro-synthetic analysis of l-deoxymannojirimycin (DMJ) .......... 213
Scheme 5.20 The synthesis of l-deoxymannojirimycin (DMJ) 55 ....................... 214

xvi

Chapter 1

Literature Review: Synthesis and Structure-Activity Relationship

of Azasugars as Glycosidase Inhibitors

1.1 Introduction

Glycochemistry has been growing rapidly as a research area in recent years with
great potential for fiIture achievements. Together with glycobiology, it has great
importance in biology, biochemistry, physiology, and pathology and medical science.
Glycosyl transferases and glycosidases mediate the formation and breakdown of linkages
of glycosyl residues to each other and to other. Because of this, the synthesis of inhibitors
of these enzymes has been an important research area. Azasugars or, more correctly,
iminoalditols are a general class of glycosyl transferase and glycosidase inhibitors. They
are pyranose or furanose derivatives in which the ring oxygen atom has been replaced by
nitrogen. Great effort has been expended in developing new methods for the synthesis of

existing and new inhibitors of this class.

1.2 Glycosidases and glycosyl transferases and their inhibitors in

carbohydrate metabolism

Glycoproteins carry N- and O-glycosidically-linked carbohydrate chains with
complex structure and function on various points along the polypeptide chain [1]. The
biosynthesis of N- and O-glycan is controlled at the level of gene expression, mRNA,

enzyme protein activity and localization, and through substrate and cofactor

concentrations at the site of synthesis. Figure 1.1 illustrates the biosynthesis of N-linked
oligosaccharides in mammalian cells. This process begins from the synthesis of the
‘ ‘ J "a L id: (Glc3Man9(G1cNAc)2) precursor on the endoplasmic
reticulum (ER) membrane [2]. The completed oligosaccharide precursor is then
transferred to a protein and several monosaccharide units are removed in the ER, after
which, the newly made glycoprotein is transported to the Golgi where the final

modifications are made. There are many diseases in which cells show different

proportions of the right glycoprotein structures from normal cells, i.e. glycosylation

changes. These range from diabetes to cancers [3].

 

.52.: Synthesis of lipid-linked precursor Glycan transfer Trimming and processing

GOLGl

 

Further trimming Terminal glycosylation

Figure 1.1 Biosynthesis of the N-linked core oligosaccharide. (Figure adapted from Helenius, A.
and Aebi. M., Science, 2001, 291, 2364-69)

During the glycan processing, glycosidases and glycosyl transferases are widely

involved in the glycosidic bonds cleavage and formation. Glycosidase Mechanisms have

been extensively studied because of their extremely important roles in the glycoprotein
biosynthesis. Several reviews have been published on this topic [4-10]. As an example,

the general mechanisms for inverting and retaining B-glycosidases are Shown in Figure

1.2 [7].

3’ xk _ o/ko— I 9*

oo . oo

' OH + H‘ _
£2 6 £08,.v'oiR £3: ROH

 

 

,.(+I<)>—V\?;c>,R = "(Hm—K); ’H = n(Hm—VA
.H 5 6 OH
F1 H.
K 0 9° 0 HO 0
8T _ I ._ T
7‘ T 115
b) (Mk0 (ID/k0
OH | OH + H 5-
n(HO)—é’g({ch 7" "(Hmé’fi :\R \ k
e

 

 

0 O O O
' OH 1'4
SCH H S 06* ‘05- /
o' \
n(H0)-.—/3\,on —-__ nIHOIU,” H
\ 5—

 

 

°r°e °‘F°

Figure 1.2 General mechanisms for inverting (a) and retaining (b) glycosidases. (Figure adapted
from Zechel, D. L. and Withers, S. G., Acc. Chem. Res, 2000, 33, 11)

Carboxyl groups catalyze the glycosidic bond cleavage as general acid and base
catalysts. For the inverting glycosidases, two carboxyl groups are suitably placed with an
average distance of 10.5 A. This placement allows the substrate and a water molecule to
bind between them. A single displacement by the water molecule affords the inverted
product. In the case of retaining enzymes, the average distance between two carboxyl
groups is 5.5 A, which prevents the simultaneous placement of the substrate and a water
molecule. A double-displacement mechanism is involved to cleave the glycosidic bond
with retention of the anomeric configuration. A covalent glycosyl-enzyme intermediate is
formed in the retaining mechanism. The mechanisms of both inverting and retaining
glycosidase involve either reactive oxocarbenium—ion intermediates or transition states
with considerable oxocarbenium character, which have been targeted by the design and

synthesis of various glycosidase inhibitors.

Table 1.1 Several classes of inhibitors and their targets (Table adapted from Varki, Ajit;
Cummings, Richard; Esko, Jeffrey; Freeze, Hudson; Hart, Gerald; Marth, Jamey; Editors.
Essentials of Glycobiology. 1999, 653 pp)

 

Class of inhibitor Target

 

Metabolic inhibitor Steps involved in formation of common intermediates such as
PAPS or nucleotide sugars

 

Tunicamycin N-linked glycosylation through inhibition of Dol-PP-GlcNAc
formation; peptidoglycan biosynthesis through inhibition of
undecaprenyl-PP-GlcNAc assembly

 

Plant alkaloids N-linked glycosylation through inhibition of processing
glycosidases

 

Substrate analogs Specific glycosyltransferases or glycosidase

 

Glycoside primers Glycosylation pathways by diverting the assembly of glycans
from endogenous acceptors to exogenous primers

 

 

 

 

Several classes of glycosylation inhibitors are exploited and these target different
steps of the biosynthesis of glycoproteins (Table 1.1) [1]. From the academic point of
view, inhibitors provide an alternative approach for studying glycosylation. Mutants have
been widely used in traditional glycobiology, the isolation of similar mutants from
different organisms and cells is relatively cumbersome. Mostly small molecules, many
glycosylation inhibitors could be relatively easily introduced into many cell types to
induce glycosylation changes. By measurement and analysis of these changes, important
information related with glycosylation processes could be obtained. From a practical
point of view, glycosidase inhibitors are very promising targets of drug development.
Different inhibitors may block different glycosylation pathways by different mechanisms.
The substrate analogs could bind specific glycosyltransferases more preferentially and
tightly than the real substrate to block these enzymes. Metabolic inhibitors, on the other
hand, act on carbohydrate metabolism. For example, 3’phosphoadenyl-S’phosphosulfate
(PAPS) is a common intermediate in the metabolism. Chlorate blocks the formation
PAPS by inhibiting its sulfation step. This inhibition has no specificity of any particular
class of glycosylation or sulfation reaction. Tunicamycin inhibits N—glycosylation in
eukaryotes by another different mechanism. Because of its resemblance of the donor
nucleotide sugar, tunicamycin acts as a tight-binding competitive inhibitor and blocks the
transfer of GlcNAc-l-P from UDP-GlcNAc to dolichyl-P, thereby decreasing dolichyl-
PP-GlcNAc. Glycoside primers subvert the cellular machinery for making
oligosaccharides on endogenous proteins thereby inhibiting glycoprotein and

proteoglycan assembly.

Plant alkaloids block glycosylation by inhibiting the glycosidases involved in N-
glycan formation. (Table 1.2) [1].
Table 1.2 Several plant alkaloids as glycosylation inhibitors (From Varki, Ajit; Cummings,

Richard; Esko, Jeffrey; Freeze, Hudson; Hart, Gerald; Marth, Jamey; Editors. Essentials of
Glycobiology. 1999, 653 pp)

 

 

 

 

Alkaloid Source Target Alkaloid Source Target

HQ: '2 0” HO,
Austra- ' . Castano- ' N a-gluco-
. --"|OH a-glucosrdase I . , .
line N spermine H0 .- - srdase I, 11

H5 H OH
CHon
Deox o- N” a- lucosidase . . " N (II-manno-
yn HOW Krfunensrn 0 .
Jrnmycrn ,_ II (and 1) H0 H n srdase 1
H0 ’o HO

Hon2
Deox - . . 9“ OH
m 3; HOW NH a-mannosrdase Swain- * '9 (It-manno-
'irri'm cin I sonine N CH sidase II
J y H OH

 

 

 

 

 

 

 

 

They have different inhibition mechanism from the other classes of inhibitors.
Instead of blocking glycosylation of glycoproteins entirely like tunicamycin, they inhibit
the trimming reactions that occur after the GIC3Man9G1cNAcz oligosaccharide is attached
to a glycoprotein. One class of alkaloids inhibits the a-glucosidases involved in the initial
processing of the N-glycans and in quality control of protein folding. These include
castanosperrnine, australine and deoxynojirimycin, which cause accumulation of fully
glucosylated chains or chains containing one to two glucose residues. Another class of
alkaloids inhibits the a-mannosidases. These include swainsonine, manostatin A,
kifiinensin and deoxymannojirimycin. These inhibitors cause accumulation of Mam-

9GlcNAcz oligosaccharide chains without glucose residues on glycoproteins.

From the chemistry point of view, all of these plant alkaloid inhibitors have a
common polyhydroxylated N-hetero ring system that mimics the natural carbohydrate
substrates. These azasugar structures are generally believed to act as a transition state
analog (Figure 1.3). After protonation, the stable and charged ring nitrogen may mimic
the positive charge on the ring oxygen that arises from delocalization of charge from the

oxocarbenium cation generated during the hydrolysis reaction.

 

 

+ ROH

 

 

 

Transition state anolog
Figure 1.3 The structure and charge resemblence of azasugar and the oxocarbenium transition
state

1.3 Synthetic methodologies towards azasugars

Shown in Figure 1.4 are several azasugars for which various synthetic methodologies

   

have been developed.
OH OH
H01,“ HO .~\OH H
N .\\OH
H
n ° ..
’OH
Deoxynojirimycin 1 Deoxymannojirimycin 2 Castemosperrnine 3 Swainsonine 4

Figure 1.4 Several important azasugars as synthetic target of glycosidase inhibitors

There are three structural elements that characterize azasugars. These are the
presence of several hydroxyl groups with defined chirality, Single or fused- ring systems
and the ring nitrogen fiinctionality. The synthetic strategies that have been developed
revolve around these structural features. Several asymmetrical synthetic methodologies
such as epoxidation, dihydroxylation have been employed for the incorporation of the
hydroxylated chiral centers. Reductive amination, intramolecular nucleophilic
substitution and alkene metathesis have been widely used for the construction of the ring
system. For introducing the nitrogen atom into the ring, many synthetic methodologies
have been adopted. These include imine formation, azide substitution, nucleophilic
substitution etc. These methodologies have been widely employed and sometimes
combined in many elegant syntheses towards azasugars. We herein review the azasugar

synthesis with an emphasis on these three structural elements.

1.3.1 Synthesis towards azasugars from achiral or chirality-deficient molecules

In 1995, Leeper et. a1. [11] developed a synthesis of several polyhydroxy
indolizidines as'epimers of castanosperrnine starting from malic acid derivatives (Scheme
1.1). Starting from acetoxysuccinimide 5, the bicyclic ring system was generated by
Mitsunobu reaction, reduction and following cyclization. The key cyclization of
compound 6 was via an acyliminium ion intermediate. The following steps included
epoxidation, ring opening, elimination and dihydroxylation. These were employed to
incorporate the other hydroxylated chiral centers. By slightly changing the modification
process, three epimers of castanosperine were synthesized, represented by the 1,6-
diepicastanosperrnine 13. All the new chiralities were generated based on substrate

control, mostly by epoxidation and dihydroxylation.

 

   

O
a AcO“" N
Aco“‘" NH_’ AcOs‘TMS __
O 5 6
[c
O o
320“" N .
H“. e 820“ N d 820“" N
H\\' «'— H\\'
M30 M86 M30 i
10 M35 0
ti \2 9 8
HO“"' N
H“‘
HO 5
H6
11 12 13

Scheme 1.1 A synthesis to castanosperrnine epimer from malic acid derivative: a)
Me3SiCH=CH(CH2)zOH, DEAD, Ph3P, 87%; NaBH4, 94%; ACZO, pyr. DMAP, 85%; b)
BF3OEt2, 72%; Et3N, MeOH, H20 then BzCl, Et3N, DMAP, 84%; c) mCPBA, 76%; (1) HF
THF, H20 then MsCl, Py., 41%; e) BH3Me2S; f) aq. NH3, 37% (2 steps); g) Bu4NOAc, 25%;
h) 0504, NMO then f, 53%

In 1996, Bermejo. et. a1. [12] developed a synthetic pathway towards (i)-
swainsonine 4, taking advantage of the degradation of a-amino carboxylic acid of an
existing bicyclic lactam 14 (Scheme 1.2). With the bicylic structure Skeleton ready, most
of the synthetic steps were focused on the incorporation of the hydroxyl groups. This
included oxidative elimination, stereoselective dihydroxylation, decarboxylation,
reduction and hydroboration. However, because the starting material was achiral and

there were no asymmetric factors involved, the swainsonine 4 was synthesized as a

racemic mixture.

COZIBU 02:31.1 H02C OOHO—é
.\\\

   

16b
[c
OH
H 9%
N .\\\OH f . J \ .\\\O
N
”0H 0
Swainsonine 4 19 13

Scheme 1.2 A synthesis to swainsonine 4 from an existing bicyclic lactam: a) i: LDA, THF, -
78°C; ii: PhSeCl; iii: H202, AcOH; 70%; b) 0304, NMO, acetone, H20, tBuOH; 70%; then
CH3C(OMe)2CH3, PPTS, CH2C12, 100%; then CF3COOH, CHZCIZ, 0°C, 98%; c) (COC1)2,
98%; d) 1,2-DCE, xylene, reflux, 15h, 75%; e) BzHé, THF, then H202, NaOH; f) i: 6N HCl; ii:
Dowex-1X8
At the same time, Zhou et. a1. [13] developed an asymmetric synthesis of 1-
deoxymannojirimycin 2 from an achiral a-furfurylamine derivative 20, which could be
prepared in 40% overall yield in three steps (Scheme 1.3). A kinetic resolution was
achieved by oxidative rearrangement of the a-furfurylamine derivative 20 to yield
compound 21 with a piperidine structure. Further functional group manipulation such as
dihydoxylation was carried out on this structure to yield the final product. Many

protection and deprotection steps were also involved in this synthesis. Using the same

strategy with another a-furfurylamine derivative [14], (-)-swainsonine 4 was synthesized.

10

9

/ O
D U OCH
0 : OCHS ‘—> R0 N "/I/ 3

/ \ a NHTs Ts
NHTs + OH
’° (1° —~° O;
. OCH
HO‘W' N OCH3 E I O\\\ Pr] 3
Ts s 24
OH 22
HO so“ Id
' OAc
f
0H \ AcO _.~\\OBz e / ”(.032
H ‘—
OCH3 N OCH3
1-Deoxymannojirimycin 2 Ts Ts

Scheme 1.3 A synthesis to deoxymannojirimycin 2 from an achiral a-furfurylamine
derivative: a) Ti(OiPr)4, L-(+)—DIPT, TBHP, silica gel, CaHz, CHZCIZ, 25°C, 3 days; b)
mCPBA; c)HC(OEt)3, BF 3OEt2, THF, 0°C; 76.5% then NaBH4, CeCl3l7HzO, -30°C; 72.3%;
d)DEAD-TPP, PhCOOH, THF, r.t.; 92% then NaBH4, HCOOH, 0°C; 87%; e) (DHQ)2-
PHAL, OsO4, K3Fe(CN)6, KZCO3, t-BuOH, r.t., 2days; 85% then AczO, Pyridine, DMAP, rt;
100%; f) BBr3, CH2C12, -78°C; 72% then Na/Naphthalene, DME, -60°C; 51%
Meyers et. a1. [15] developed an asymmetric synthesis of L-deoxymannojirimycin
34 employing dihydropyran and S-phenylglycinol as their starting materials (Scheme 1.4).
Their synthesis featured the preparation of a bicyclic lactam by cyclodehydration of a
ketoacid and S-phenylglycinol developed in their lab [16]. Metallation of dihydropyran
and reaction with paraforrnaldehyde gave a primary alcohol, which was protected and
oxidized to a ketoacid 29. The cyclodehydration of the ketoacid 29 and S-phenylglycinol
gave a bicyclic lactam 30. The synthesis was continued by introduction of a double bond,

allylic oxidation, dihydroxylation and deprotection steps to yield L-deoxynojirimycin 34.

The S-phenylglycinol not only supplied the nitrogen atom, but also acted as a chiral

auxiliary. All the chiralities of the target molecule were introduced based on substrate

 

control.
OBn OBn
b Ph
0 a (l/OBn —> O —c> <3
0 O
27 28 29 3° Id
OH
HO,” ph
HO
N
H
34 31

 

Scheme 1.4 A synthesis of L-deoxymannojirimycin 34 from dihydropyran and S-
phenylglycinol: a) t-BuLi, (CH20)n, THF, -78°C, 70% then NaH, BnBr, DMF, 92%; b)
CrO3/HZSO4, THF, 70%; c) S-phenylglycinol, PhMe, heat, 80%; d) KH, PhSOzMe, THF,
then PhMe, heat, 85%, then SeOz, dioxane, heat, 64%; e) OSOJNMO, acetone/water, 83%;
r) (CH3)2C(OMe)2,CH2C12, pTSA, 75%; then BH3THF, THF, heat, 65%; g) Hz/Pd(OH)2,
EtOH then TFA, MeOH, 75%

Carretero et. a1. [17] also accomplished a stereoselective synthesis of swainsonine
from di-Boc protected 5-aminopentanal 35 (Scheme 1.5). This synthesis employed an
enzyme-facilitated kinetic resolution based on the enantioselective acylation of lipase.
Bulkier protecting group introduction improved the cyclization stereoselectivity of the
intramolecular amination of 39. After alkylation, the intramolecular acylation of 41b
through the a—sulfonylcarbanion afforded a single a-sulfonylketone which was directly

reduced to give the a-sulfonyl alcohol 42. Further steps included desulfonylation,

dihydroxylation and deprotection to give swainsonine 4.

12

 

OH OH OAC
/ /
CCHO j, Sth J; SOzPh c / sozph
”(3°92 N(Boc)2 NHBoc NHBoc
35 36 37 38
[a
TIPS /802Ph TIPS SOzPh OTIPS [sozph OTIPS
\\\\ \\\\ /
N CO Et N CO Et NH NH3
V 2 V 2 cracoo-
41b 41a 40 39
v] 9
"PS SOzF’h “'33 OTIPS
H OH
H . H +
OH —> \ J, OH
N N N
42 43 44 4

Scheme 1.5 A synthesis to swainsonine 4 from di-Boc protected 5-aminopentana1: a)
PhSOZCstoTol, piperidine, CHzClz, 0°C; b) TFA, CH2C12, rt; c) Lipase PS, vinyl acetate,
toluene, rt; d) Lipase PS, 0.1M NazHPO4, rt; then TIPSCl, imidazole, CH3CN, rt; then TF A,
CH2CI2, ft; C) MCOH, EI3N, '780C; f) BI'CH2C02EI, LII cat, K2CO3, CH3CN, 800C, g)
LHMDS, THF, 0°C; then NaBH4, MeOH, 0°C; h) Na-Hg, NazHPO4, MeOH, rt; i) 0504 cat,
Me3NO, acetone, H2041;

In 1997, Roush et. a1. [18] reported an enantioselective synthesis of (-)-
swainsonine 4 as an application of an improved chiral reagent for the anti (1-
hydroxyallylation of aldehydes (Scheme 1.6). Their synthesis began with 4-bromobutana1
45, which is available from tetrahydrofuran. The Homer-Wadsworth-Emmons reaction of
this bromobutanal 45 followed by DIBAL reduction afforded an allylic alcohol 46.
Sharpless asymmetric epoxidation and oxidation with SO3-pyridine gave the epoxy

aldehyde 48. The chain was extended by a 3-carbon unit by a-hydroxyallylation with

(S,S)-diisopropy1 tartrate-modified (E)-[y-[2-menthofury1]dimethylsilyl]-borate. The

13

silanol was then subjected to a one-pot protodesilylation-oxidation to give a diol 50,
which was protected as an acetonide. Ozonolysis of the terminal alkene provided an
aldehyde 51. The subsequent reductive amination of this aldehyde was accompanied by
spontaneous ring closure to form the indolizidine nucleus in the form of swainsonine

acetonide. This was deprotected to provide swainsonine 4.

Br CH0 3 Br \ b Br\/\/<?/\
45 46 47

[c

9H e 9H d O
Br\/\/¢Y\‘- Br\/\/¢Y\ ‘- Br CHO
OH SiMezR
50 49 48
‘ f OH

CHO 30H

Scheme 1.6 A synthesis to swainsonine from 4-bromobutana1: a) (EtO)2P(O)CH2C02Et, NaH,
THF, -30 to 23°C then DIBAL, EtZO, -78 to 23°C, 82%; b) Ti(OiPr)4, D-(-)-D1PT, TBHP,
4AM.s., -20°C, 72%; c) SO3-pyridine, DMSO, Et3N, CH2C12, 78%; d) (S,S)-Diisopropy1
Tartrate-Modified (E)-[y-[2-menthofuryl]dimethylsi1yl]-borate, toluene, 4AM.S, —78°C, 73%;
e) CF3COOH, THF, 0-23°C then KHCOJ, KF, H202, THF-MeOH, 23°C, 85%; f)
M82C(OMC)2, pPTS, CH2CI2 then 03, CH2CI2, -780C; Ph3P, 84%; g) NH4OAC, 3AMHS,
MeOH; then NaCNBH3, 71%, then 6N HCl-THF, 92%
Mukai et. a1. [19] developed a total synthesis of (i)-swainsonine based on an
endo-mode cyclization. A suitable enyne alcohol 52 was employed as the starting material
(Scheme 1.7). The nitrogen functionality was introduced by Mitsunobu reaction using

hydrazoic acid followed by reduction and tosylation. Epoxidation and ring closure

furnished the piperidine ring system 54. The subsequent steps included nucleophilic

14

addition, reduction, cyclization and dihydroxylation. After hydrolysis, this synthesis

afforded (jg-swainsonine.

| b OH OH
.1. 0 _.
\\ +. \H I. I. \\
52 53 54 55 H
[C

OTBDMS OTBDMS UOTBDMS

““0 H ‘— \\\\\\\ [1”1,
::\\/
OH

Scheme 1.7 A synthesis to swainsonine from a suitable enyne alcohol : a) HN3, DEAD, PPh3,
99%; then PPh3, H20, 99%; then TsCI, Et3N, 81%; then TBAF, 96%; then mCPBA, 58%; b)
C02(CO)3; then BF3OEt2; then CAN, 85%; c) TBDMSCI, imidazole, 99%; then n-BuLi,
(HCHO)n, 97%; d) H2, Lindlar cat., 99%; e) Na, naphthalene, then CBr4, PPh3, Et3N, 57%; t)
0504, NMO, then TBAF, then Ac20, pyridine, DMAP, 76%; then K2CO3, MeOH, 99%

 

In 2001, Comins et. a1. [20] reported an asymmetric synthesis of 1-
deoxynojirimycin 1 featuring enantiopure 2,3-dihydro-4-pyridones 60 as synthetic
intermediates (Scheme 1.8). 4-methoxy-3-(triisopropylsily1)pyridine 59 was employed as
the starting material. The addition of (benzyloxy)methylcuprate to the 1-acy1pyridinium
salt bearing a chiral auxiliary generated the desired dihydropyridone 60. The chiral
auxiliary and the 5-TIPS group were removed in a one-pot reaction to give 61. After
protection of the nitrogen, the compound went through several oxidation and reduction

processes to afford the l-deoxynojirimycin 1.

15

 

COZR‘
59 60 61 6:2
[d
O
OBn
'7‘
Cbz
1 652 64 63

Scheme 1.8 A synthesis to deoxynojirimycin 1 from 4'methoxy-3-(triisopropylsi1y1)pyridine:
a) R‘OCOCI, (R‘ = (1R,2S)-2-(l-methyl-1-pheny1ethyl)cyclohexanol), then BnOCH2(2-
Th)Cu(CN)Li2, then H3O+, 64%; b) NaOMe, then H3O+, 74%; c) n-BuLi, then BnOCOCl,
99%; d) Pb(OAc)4, toluene, reflux;78%; e) aq. HCl/EtOl-l, then Me4NBH(OAc)3,
acetone/AcOH, 83%; t) 0504, NMO; g) Pd(Ol—1)2/H2, 10% HCl

In 2003, Somfai et. a1. [21] reported another asymmetric synthesis of both (+)-1-
deoxynojirimycin 1 and (+)-castanospermine 3 starting from the same achiral acyclic
diene 66 (Scheme 1.9). Asymmetric dihydroxylation and epoxidation of 66 generated all
the chiralities of deoxynojirimycin. Suitable protecting and deprotecting strategies were
used to give 68. The epoxidation was carried out utilizing the stoichiometric protocol
since the catalytic Sharpless epoxidation only gave poor yields. The nitrogen functionality
was introduced by azide substitution of a mesylate group at 69. The subsequent reduction
and intramolecular epoxide ring opening followed by deprotection gave the (+)-1-
deoxynojirimycin 1. The synthesis of castanosperrnine took advantage of the protected
deoxynojirimycin 71 obtained during the synthesis. After selective protection and
deprotection, the primary —OH was transformed to an aldehyde by Swem oxidation to

give 73. The carbon chain was extended by the addition of an allyl group to the aldehyde

l6

using allyltrimethylsilane and a Lewis acid (Sakurai reaction) with careful control of the
reaction condition. Following dihydroxylation and NaIO4, oxidative cleavage of the diol

gave another aldehyde 75. Finally reductive amination furnished (+)-castanospermine 3.

PMBOMOE' 1,0], oer _, OJBK‘CSVOTBDPS
o

PMBO PMBO 68

#[c
OH 0
HO“ OCH 0, \ H

d9”
.1. (:1: «e— 0 ffiOTBDPST-s fworeops:
[I OH 9 H OTBDPS
1 / 71
\l~o
0,,

O
,OBn 0,, \,OBn '

   

\ A
N N
3" OTBDPS 3" o
72 73 3

Scheme 1.9 An asymmetric synthesis of deoxynojirimycin and castanosperrnine starting from
the same an achiral acyclic diene: a) AD-mix-a, CH3SOZNH2, t-BuOH, H20, 80%; then 2-
methoxypropene, p-TsOH(cat.), DMF, 97%; b) DlBA L, -78°C, CHZCIZ, 93%; then (+)-DlPT,
Ti(Oi-Pr)4, t-BuOOH, CH2C12, -20°C, 80%; then t-BuPhZSiCl, Et3N, DMAP, CH2C12, 97%;
c)DDQ, CH2C12, H20, 92%; then MsCl, i-PrzEtN, CH2C12, 100%; d) NaN3, DMF, 70°C, 91%;
then PH3P, THF, H20, 83%; e)EtOH, heat, 100%; t) HC1(37%), MeOH, 100%; g) KHMDS
(2.2equiv.), BnBr(3equiv.), THF, -78°C, 82%; then BnBr(l .3equiv.), K2C03(2.6equiv.),
CH3CN, heat, 97%; h) n-Bu4NF, THF, rt, 100%; then (COC1)2, DMSO, Et3N, CHZCIZ, -78°C,
93%; i) CHZCHCHZSiMe3, TiC14, CH2C12, ~65 to -60°C, 71%; j)OsO4, NMO, t-BuOH-THF-
H20(10:3:1); then Na104, NaHCO3, THF-H20(1:1), 84% (2 steps); k) H2, Pd/C, then TFA,
81%

To summarize, the synthesis of azasugars from achiral or chirality-deficient
molecules generally involves many functional group transformation steps.

Dihydroxylation and epoxidation are frequently used to introduce the new chiral centers.

17

Both reagent control and substrate control have been widely used in this process. The
nitrogen atom has been introduced either as a functional group in the starting material or
in the later steps of synthesis. Azide substitution and imine formation are two common
ways to introduce nitrogen functionality. The reductive intramolecular substitution and

reductive elimination are frequently used to afford the heterocycle ring systems.
1.3.2 Synthesis of azasugars from sugars or other chirality-rich molecules

Although many elegant syntheses of azasugars from achiral or chirality-deficient
molecules have been developed, the majority of the synthetic work on azasugars employs
sugars or other potentially abundant chirality-rich molecules as starting materials.
Because of the structural similarity of the azasugars to real sugars, carbohydrates should
be ideal starting points for the synthesis of the former. The major structural difference of
azasugars with real sugars is the ring nitrogen. So the introduction of a nitrogen atom into
the sugar structure is the essential synthetic task. Many nitrogen incorporation methods

have been developed and employed for this purpose. Here we review this work.

1.3.2.1 Synthesis of azasugars using azide as an intermediate

L
O

O \/
n(RO)—-: \/—\‘ —’ “TOEFL-Cl”

N3

1

L
N \0

(R0).-.—‘ \/—> ‘— (Rom—‘1?
NH2

Scheme 1.10 The general synthetic pathway of azasugars using azide as an intermediate

18

Azide has long been used as an amine precursor in synthetic chemistry. The azide
group can be relatively easily incorporated into a sugar structure by nucleophilic
substitution. The subsequent reduction generates an amino group, which can undergo an

intramolecular substitution to afford the N-heterocycle ring system (Scheme 1.10).

 

 

COzEt N3
A) Home H / /‘
= E o m ‘ H :3 H O
o —” .,
”o am’ok “9‘98 ’ok
75 Ph 77 [c
\.OH H o / H O‘NHz
e a ’ E
.\ H ‘— N mo d "'0
‘ 0sz )V "‘ 3(1),, )f
HO O HO 0
a) . TsOHzc o NCbz
i HOIII<' >lllOMe b QOMB
H6 N3 H0 MON
83
i c
0025' (EtS)2HC
Hm N NCbZ
e NCbz
HO .— BnO 1‘. 8:10
HO OH BnO OBn BnO OBn
a7 86 05

Scheme 1.11 A) Synthesis of l-deoxy-castanospermine 81 from glucose derivative: a) PCC,
powdered molecular sieves, CHZCIZ, then Ph3P=CHC02Et; b) 10% Pd/C, EtOH, H2; LiAlH4,
ether; MsCl, Et3N, CHZCIZ, 0°C; NaN3, DMF; c) NBS, BaCO3, CC14; NaOMe, MeOH; then
SnClz, MeOH, 60°C; d) NaOAc, EtOH; Cszl, NaHCO3, EtOH(50%)-H20; e) HOAc(20%)-
H20; then 10% Pd/C, MeOH, H2; B) Synthesis of epimer of l-deoxy-castanospermine from 2-
azido-4,6-O—benzylidene-2-deoxy-or-D-a1tropyranoside: a) HOAc(20%)-H20; then TsCl,
pyridine; b) 10% Pd/C, EtOH, H2; NaOAc, EtOH reflux; Cszl, NaHCO3, EtOH(50°/o)-H20; c)
EtSH, HCI, CHC13; then BnBr, NaH, DMF; d) HgClz, CdCO3, Me2C0(10%)-H20; then
Ph3P=CHC02Et, CH3CN; e) 10%Pd/C, H2, EtOH; NaOAc, EtOH, reflux, 10% Pd/C, H2,
AcOH, 48h; A020, pyridine; then BH3SMe2, THF; NaOMe, MeOH;

19

In 1987, Richardson et. al. [22,23] synthesized several polyhydroxylated
indolizidines related to castanospennine (Scheme 1.11). For l-deoxy-castanospennine 81,
3,5-0-benzylidene-l,2-O-isopropylidene-a-D-glucofuranose 76 was employed as the
starting material. The carbon chain extension was carried out by pyridinium
chlorochromate (PCC) oxidation of the free hydroxyl group followed by immediate
reaction with carboethoxymethylene triphenylphosphorane. The double bond and ester
group of 77 were reduced sequentially by catalytic hydrogenation and lithium aluminum
hydride (LiAlH4). The free primary hydroxyl group was then transformed to a leaving
group, which was subsequently displaced by an azide group using sodium azide to give
78. Afier regioselective oxidative cleavage of the benzylidene group by Hanessian-Hullar
reaction, the azide group was reduced to an amine 79. Intramolecular displacement of the
bromo group by the amino function furnished a five-membered N-heterocyclic ring 80.
After deprotection of the isopropylidene group, catalytic reductive elimination and the
final deprotection, l-deoxy-castanospermine 81 was synthesized (Scheme 1.11 A). A
different strategy was used to synthesize another epimer of l-deoxy-castanospermine.
This employed 2-azido-4,6-O-benzylidene-2-deoxy—a-D-altropyranoside 82 as the starting
material. After debenzylidination and selective tosylation of the primary hydroxyl group,
the existing azide group was catalytically reduced to form an amine, which displaced the
tosylate group by intramolecular nucleophile attack at the primary carbon. The bridged
ring system 84 was obtained and this proved to be an important intermediate. The
furanosyl ring was cleaved by thiolysis giving a monocyclic system in which the former
anomeric carbon was protected as a dithioacetal group. After deprotection, the free

aldehyde was reacted with carboethoxymethylene triphenylphosphorane to afford an

20

unsaturated chain extended product 86. Subsequent reduction and deprotection steps

yielded an epimer of l-deoxy-castanospermine 87 (Scheme 1.11B).

In 1989, Fleet et. al. [24] reported two synthetic routes towards
deoxymannojirimycin 2 using different starting materials. These are illustrated in Scheme

1.12 A and 1.12 B respectively.

$O|hlo\\\"’0’11113(__>1-8H;\\jjunob Bn ‘ "”0
HO “bk )f

 

o
as
H momm e $41,
+— N
OMe
2 91
H, o
o ' o
0.3.» TBDMSO —>TBDMSO
050 OH N30 0
,3 9, OHOOX ’5 X
to
+0 +0
e
._ 0 .0” J; o
OTBDMS
o N OTBDMS
2 97 96

Scheme 1.12 Ttwo synthetic routes towards deoxymannojirimycin from glucose derivative and
L-gulonolactone: A) a) AcOH/HZO; p-TsCl, pyridine; b) NaN3, DMF; NaH, BnBr, (Bu)4N1;
HCl, methanol; (Tf0)20, CHZCIZ, pyridine; c) Ph3P,then K2C03, H20; Cszl; (1) CF3COOH,
H20; NaBH4, methanol, H20; Pd/C, H2; B) a) acetone/dimethoxypropane, p-TsOH;
AcOH/HzO; TBDMSCI; b) (Tf0)20; NaN3, DMF; c) Pd/C, H2, methanol; d) BH3SMe2; e)
CF3COOH/H20

Starting from 1,2,5,6—di-O-isopropylidene-a-D-glucofuranose 88, the 5,6-0-
isopropylidene was selectively deprotected and the primary hydroxyl group was
derivatized to a tosylate 89. An azido group was used to introduce the nitrogen function at
the 6-position. After deprotection of the other isopropylidene group, the 2-OH was
transformed to a triflate leaving group. Reduction and intramolecular nucleophilic
substitution afforded a [3,2,1] hetero- ring system 91. The bridged ring was broken by
acid hydrolysis and NaBH4 reduction. The final catalytic hydrogenation removed all the
protecting groups to give deoxymannojirimycin 2. This compound was also synthesized
from L-gulono-l,4-lactone 93. 2,3,5,6-Di-O-isopropylidene-L-gulono-1,4-lactone was
prepared and selectively hydrolyzed. t-Butyldimethylsilyl chloride was then used to
selectively protect the primary hydroxyl group to give 94. As in the the other synthetic
route, the only free hydroxyl group was then derivatized to a triflate and subsequently
substituted by azide. Hydrogenation of the azide 95 gave an amine, which spontaneously
cyclized to the lactam 96. Borane reduction followed by acidic deprotection yielded
deoxymannojirimycin 2. The latter route had the advantage that the intermediate lactam,

which was also believed to be a suitable glycosidase inhibitor, was also formed.

Shortly after the synthesis of deoxymannojirimycin 2, they reported a synthesis of
deoxynojirimycin l [25] using a similar strategy (Scheme 1.13). An epoxide 98 with L-
idofuranose configuration was used as the starting material. Reaction of the epoxide with
sodium azide gave the ido-azide, then the free S-OH was protected by a benzyl group.
After acid hydrolysis of the 1,2-O-isopropylidene of compound 99, 2-OH was derivatized

to a triflate 100. The catalytic hydrogenation reduced the azide to an amine group, which

22

underwent intromolecular cyclization to form the bridged ring system 101. After

hydrolysis and deprotection, deoxynojirimycin 1 was obtained.

111110 I BnO
III’IOX I”’/OT

lo

 

OH OH
3
H00"- _‘\\\OH 8 BnO/I," ‘\\\O an d no N / Cbz
<— <— ————o
OH OH
fi (”212 BnO OMe
1 102 101

Scheme 1.13 A synthesis of deoxynojirimycin from An epoxide with L-idofuranose
configuration: a) NaN3, DMF; 87%; NaH, BnBr, (Bu)4Nl, THF; 91%; b) HCI, methanol, 67%;
mo; 88%; SnClz, methanol; c) NaOAc, ethanol; CBzCl; 67%; d)CF3COOH/H20, dioxane;
NaBH4, ethanol; e) Pd/C, H2, AcOH

In 1993, Tyler et. al. [26] developed a synthesis of deoxygalctonojirimycin 108
from L-Sorbose (Scheme 1.14). l,2:4,6-Di-O-isopropylidene-a-L-sorbofuranose 104 was
synthesized from L-sorbose 103. The free 3-OH was inverted by an oxidation-reduction
process to yield exclusively l,2:4,6-di-O-isopropylidene-a-L-tagatofuranose 105. Acid-
catalyzed acetal isomerization generated l,2:3,4-di-O-isopropylidene-a-L-tagatofuranose
and the free 6011 was subsequently derivatized and substituted with azide. Acetolysis
then generated the diacetate as a mixture of anomers, which was then deprotected and
selectively primary protected by a TBDMS group. Reductive amination of 107 then

afforded the deoxygalctonojirimycin 108.

23

Scheme 1.14 A synthesis of deoxynojirimycin from L-sorbose: a) Me2C(OMe)2, SnClz,
MeOCH2CH20Me; b) MeZSO, (CF3CO)20, Et3N, CHZCIZ; then NaBH4, EtOH; c) CSA,
MeZCO; then MsCl, Et3N, CHZCIZ; then NaN3, MeZSO, 80°C; d) 0.5% BF3OEt2 in Ac20, 0°C;
then MeONa, MeOH; then TBDMSCl, imidazole, DMF; e) H2, Pd/C, EtOH; then CF3C02H,
H20(3:7 v/v), RT
Two other similar synthetic routes towards l-deoxymannojirimycin were
developed by Szarek et. al. [27] and Stfitz et. al. [28]. They have very similar
intermediates, namely 6-aizdo-D-fructose 115 and the protected form 112. Szarek et. al.
started from D-mannitol (Scheme 1.15 A). Several selective protection and deprotection
steps yielded the 3,4-O-benzyl-D-mannitol 110, which could be selectively
monotosylated and then converted to an azide alcohol 111. Because of its C2 symmetry,
this process generated only one enantiomer. The selective oxidation at the 2-OH position
was accomplished by reaction with bis(tn'butyltin)oxide-bromine, which afforded the
partially protected 6-aizdo-D-fructose intermediate 112. The subsequent hydrogenation
generated l-deoxymannojirimycin 2. Alternatively, Stiitz’s synthesis (Scheme 1.15 B)

employed D-fructose 113 as the starting material. After acylation and reaction with

triphenylphosphane dibromide, the acyclic bromosugar 114 was obtained. After

24

deprotection, this bromosugar was reacted with substituted by sodium azide to yield the

similar intermediate 115. The final hydrogenation generated l-deoxymannojirimycin 2.

0 OH OH OH
“0":1W:0313"°” cm." W 2,011.1“)
”OH '1 ['11 H
BnO OBn ” OH

112 2
3)
AcO OH
0“ 0
HO \\‘OH
0 OH 3» AGO" DWJOH 3.
OH AcO "OAc—’ ': OH
OH HON
OH Br H
113 114 115 2

Scheme 1.15 Two similar synthetic routes towards l-deoxymannojirimycin sharing a common
intermediate 6-aizdo-D-fructose: A) a) 2-methoxypropene, 0°C; then NaH, BnBr; hydrolysis;
b) leqiv. p-TsCl; NaN3, DMFzHZO = 10:1, 100°C; 51% c) bis(tributyltin)oxide-
bromine(2.3eqiv.); d) Pd/C, H2, methanolzl-120 = 3:1, 0.1N HCl; 87%; B) a) A020, sulfuric
acid, 60%; then Ph3PBr2, CHZCIz, 90%; b) Deprotection then NaN3, DMF, rt, 60 ~ 65%; c)
Pd/C, H2, methanol
There have been some other synthetic works of azasugars based on using azide to
introduce the imino nitrogen. Starting from D—glucono-l,5-lactone 116, by
diisopropylidenation and introduction of an azide in the a-positon, Rapoport et. al. [29]
carried out a synthesis towards castanosperrnine 3 and 6-epicastanospennine 118
(Scheme 1.16 A). Cha et. al. [30] employed a chiral lactol 119 as the starting material.
Their synthesis towards castanosperrnine 3 (Scheme 1.16 B) featured a reductive

cyclization of azide and epoxide of 120. Using erythronolactone 121 as the starting

material (Scheme 1.16 C), Pearson et. al. synthesized (-)-swainsonine 4 [31]. This

25

synthesis also involved a similar reductive cyclization involving the azido and mesylate

groups of 122.

3)

 

121 122 4

Scheme 1.16 Several enatioselective synthesis of azasugars using reductive cycliation of
azide and leaving groups. A) Starting from D-gluconolactone; B) starting from chiral
lactol; C) starting from D-erythronolactone

In summary, all of the synthetic strategies discussed above took advantage of the
facile and general reductive cyclization involving of azido and leaving groups. Azide
could be relatively easily installed as a nitrogen source. This amine precursor could be
selectively reduced in the appropriate stage of a synthesis to act as an in-situ generated
nucleophile. Because of such flexibility, azides have been extensively employed in
azasugar synthesis. However, the azide-facilitated synthesis has the serious drawback of

being potentially explosive.

26

1.3.2.2 Reductive amination in the synthesis of azasugars

Reductive amination has long been employed as a synthetic method to introduce

nitrogen functionality into a structure. Reitz et. al. [32-34] adopted this strategy into the

synthesis of azasugars and successfully synthesized several azasugars including 1-

deoxynojirimycin l and l-deoxymannojirimycin 2 (Scheme 1.17).

 

 

CHO
A) O CHPh2 ICHth
HO HOHZC CHZOH HOHZC CH20H HOHZC o, CHZO
0H 6 H6 H0‘
CH20H
b R=CHPh2 124
123 13%,] 127 125 126
a) CHZOH CH20H
)HO‘\‘;1::ZIIOIIOa—> 0:3';3( b
CHZOH
128 129 130 1
CHO
C) 01-]on 0CHZOH HO
1'10““ IIIIOMe a ""0Me_> b HO
OH
O
CH20H
131 132 133 2

Scheme 1.17 Double reductive amination towards azasugars: A) synthesis of 2,5-anhydro-
imino-D-glucitol 127: a) PhZCHNHz, NaCNBH3, MeOH; 68%; b) 20% Pd(OH)2/C, H2;
91%; B) synthesis of l-deoxynojirimycin l: a) nBuzsnO, MeOH; Brz; b) Dowex'so; c)
thCHNHz, NaCNBH3, MeOH; 74%; then 20% Pd(OH)2/C, H2; 90%; C) synthesis of 1-
deoxymannojirimycin 2: a) nBUZSnO, MeOH; Brz; 70%; b) Dowex-SO; quant. c)
PhZCHNHz, NaCNBH3, AcOH, MeOH; 45%; then 20% Pd(OH)2/C, H2; 90%;

27

Starting from 5-keto-D-fructose 123 (Scheme 1.17 A), a double reductive
amination using sodium cyanoborohydride gave a protected 2,5-anhydro-imino-D-glucitol
124, which was subsequently deprotected by catalytic hydrogenation to yield 2,5-
anhydro-imino-D-glucitol 127. The unexpected relatively high stereoselectivity was
assigned to the assistance of neighboring hydroxyl groups. Starting from selectively
protected l,2-O-isopropylidene-D-glucose 128, a selective oxidation of the S-OH was
achieved to yield the 5-keto-D-glucose 130 afier acid hydrolysis. A similar double
reductive amination (Scheme 1.17 B) was then carried out using the same procedure,
where l-deoxynojirimycin 1 was obtained after deprotection. Similarly, when l-O-
methyl-2,3-O-isopropylidene-D-mannofuranoside 131 was employed as the starting

material, the similar procedure yielded l-deoxymannojirimycin 2 (Scheme 1.17 C).

 

CHZOH
OBn HOB" OBn
O BnO-d /O
BnO a , b . BnO
BnO —OBn BnO \
BnO OH —0H 03" o
CH208"
134 135 136
OBn
BnO/, .eOB" H00,
c " ' d '
———> —>
OBn
N
R
R = H 137
R= Bu 138

 

Scheme 1.18 Synthesis of deoxynojirimycin by double reductive amination from 2,3,4,6-
tetra-O-benzyl-a-D-glucopyranose: a) LiAlH4, THF, l00%; b) DM SO, (CF3CO)20,
CHZClz, Et3N; c) RNHfHCOz" NaCNBH3, MeOH, molecular sieve 3 A. d) Li, NH3

28

This double reductive amination approach attracted much interest because of its
simplicity and potential industrial possibility. The 5-keto-hexose synthetic precursors for
1,5-dideoxy-l,5-iminosugars became interesting synthetic targets for azasugar synthesis.
In 1999 and 2001, two groups developed two approaches towards this kind of molecules.
Lopes et. al. [35] started with 2,3,4,6—tetra-O-benzyl-a—D-glucopyranose 134 (Scheme
1.18), which could be readily synthesized from D-glucose. A simple reduction of this
starting material yielded a 1,5-diol 135, which was subsequently oxidized to the S-ulose
136. This key intermediate was directly reacted with ammonium formate and NaCNBH3
to give the protected azasugars 137 or 138. After reductive deprotection, l-

deoxynojirimycin 1 and N-butyl-l -deoxynoj irimycin 139 were synthesized.

Murphy et. al. [36,37] employed 6-deoxyhex-5-enopyranosides as an useful
intermediate synthesis and derived D-hexos-S-uloses, which could be readily transformed
to azasugars with double reductive amination discussed above. Starting from methyl a-D—
glucoside 140 (Scheme 1.19 A), 6-deoxyhex-5-enopyranoside 141 was synthesized
through primary iodination and acetylation followed by base-catalyzed elimination. The
protection groups were converted to benzyl groups. Epoxidation followed by hydrolysis
generated the l,6-anhydro-D-glucopyranos-S-ulose 144. This unusual structure was
confirmed by NMR analysis and chemical derivatization. After several protection and
deprotection steps, 5-keto-glucose 145 was synthesized as a precursor of 1-
deoxynojirimycin 1. Similarly, 5-keto-mannose 150 was synthesized from methyl a—D-
mannoside 146 (Scheme 1.19 B). These two D—hexos-S-uloses could be transformed to 1-

deoxynojirimycin l and l-deoxymannojirimycin 2 respectively by reductive amination.

29

 

150 149 OMe

Scheme 1,19 The synthesis of D-hexos-S-uloses 145 and 150 as an intermediate for azasugar
synthesis: A) synthesis of D-xylo—hexos-S-ulose: a) PPh3, lm, 12, toluene, 80°C,3h; then TMSCI,
Py, 25°C; then DBU, DMF, 70°C, 3h; then Ac20, Py; b) NaOMe, MeOH, 25°C, 12h; then NaH,
BnBr, DMF, 0°C, 15h; c) l,l,l-trifluoroacetone, Oxone, NaHCO3, NazEDTA, CH3CN, H20;
then TBSOTf (1.5eq.), 2,6-1utidine (2eq.), CH2C12; (1) H2, Pd/C, EtOH; then TBAF, THF; B)
synthesis of D-lyxo-hexos-S-ulose: a) PPh3, lm, 12, toluene, 110°C; then Ac20, Py; b) DBU,
toluene(anhydr.), 1 10°C, 50% over 3 steps; then NaOMe, MeOH; then TMSCl, Py., 93% over 2
steps; 0) 1,1,1-trifluoroacetone, Oxone, NaHCO3, NazEDTA, CH3CN, H20, 1h, 71%; d)
MeOH, 94%

Pandit et. al. [38] used suitably protected aldonolactones as starting materials.
They developed a relatively general transformation of sugar lactones to sugar lactams
(Scheme 1.20). Sugar lactams not only can be transformed to 1,5-dideoxy-l,5-

iminosugars by reduction but, as stated earlier, also are potential glycosidase inhibitors

30

themselves. Afier aminolysis of the sugar lactones 151 to form amides 152, the free 5-
OH’s were oxidized to carbonyl groups. Reaction with ammonia in methanol generated
two isomeric hydroxyl lactams 154, which could be reduced with NaCNBH3 to yield the
protected sugar lactams 155. The reduction mechanism was believed to involve a hydride
donation by the NaCNBH3 reagent to the acyliminium ion. This strategy could be used for

sugar lactams with gluco-, galacto- and manno- configurations.

OBn
R2 0 a I
8.10 \O BflgfiXb NH2 Bn20 :0 I

R1
151: 2R -OBn 4R OBn 1528(2R)-08n(4R)-08n 153(2R1-03M4Rr03n
151a izsi-oan E4R)):OBn 15213 (281-OBn (4R)-OBn 153D (ZSl-OBn (4R)-OBn
151c (2R)-OBn (4S)-OBn 152C (2R1-08n (4S)-OBn 153C (2R1-OBn (4S)-OBn
OBn /OBn
03" e R, NH ‘1 R, NH
BnO NH .— BnO \o BnO > \o
B"° R1 HO R1
OBn
1 155a (2R)-OBn (4R)-OBn 154: (2R)-OBn (4R%OBn
5° 1551: (zs)-oan (4R).OBn 1541: (28)-OBn (4R%OBn
155:: (2R)-OBn (4syoan 1546 (ZRHDBn (4S)-OBn

Scheme 1.20 A transformation of sugar lactones to azasugars: a) NH 3, MeOH; b) DMSO,
AczO; c) NH3, MeOH; d) NaCNBH3, HCOZH; e) BH3

In 1996, Mootoo et. al. [39] developed a triple reductive amination approach to
castanosperrnine (Scheme 1.21). The synthetic strategy was to suitably place three
carbonyl groups in a molecule and perform the reductive aminations at the same time.
This approach is advantageous in preparing the fused-ring azasugars. Allylation of a
known aldehyde 157 readily derived from methyl a-D-glucoside extended the carbon

chain to the desired length. Benzyl groups were then used to protect the free hydroxyl

31

groups. The treatment of this alkene with iodonium dicollidine perchlorate gave a mixture
of iodo—tetrahydrofurans, which underwent zinc-mediated reductive elimination to an
alkenyl acetal-alcohol. This alcohol was then oxidized to a ketone 159. The ozonolysis of
the alkene and acid-hydrolysis gave the key tri-carbonyl intermediate 160, which actually
existed as a mixture of lactol isomers 161. Triple reductive amination was carried out

finally and after catalytic hydrogenation, castanosperrnine 3 was synthesized.

OHC O a BnO O 8110 —
BnO _> Bno b /O
—» BnO
BnO B O 8" Bn OMe
n OMB Bno OMe Bno OMB
157 158 159
«Brio
BnO/gCHOH BBnO OBCH0

 

3 161 160

Scheme 1.21 A triple reductive amination approach to castanosperrnine 3: a)a11yl bromide,
Sn, CH3CN-H20(10:l), ultrasound; then BnBr, NaH, n-Bu4Nl, DMF; b) lDCP, CH2C12-
MeOH; then Zn, 95%EtOl-l, heat; then Swem oxidation; c) 03, CHZCIZ, -78°C then Ph3P;
then THF-9M HCl; d) 1.3 eq. NH4HC02, 30eq NaCNBH3, MeOH; then 10% Pd/C, MeOH-
HCOOH;

This triple reductive amination approach was further employed to the synthesis of
swainsonine (Scheme 1.22) [40]. The synthesis started from 2,3:5,6-di-O-isopropylidene-
mannofuranose 162. Selective deprotection and oxidative cleavage of the diol generated
an aldehyde 163, which reacted with allyltrimethylsilane to extend the carbon chain and
the free hydroxyl group was protected by benzyl. The alkene 164 was treated according to

the standard iodocyclization-THF opening sequence to give the hydroxyl alkene 165.

32

Hydroboration and oxidation installed the protected ketoaldehyde 166. DDQ-mediated
removal of the p-methoxybenzyl provided a mixture of isomers, which underwent triple

reductive amination and further deprotection to give swainsonine 4.

><O CH0 B"°"” Bno”” OMe

n o “OPMBC OH

0‘. NOH a Hog—Z WWW”: “ "0PMB
H ,

o o 0
152 153 154 155
OH
H .1
N .I\OH
'bH

 

Scheme 1.22 A triple reductive amination approach to swainsonine 4: a) PMBCI, NaH, n-
Bu4NI, DMF; then HOAc; then Na104; b) allyltrimethylsilane, BF3OEt2; then BnBr, NaH, n-
Bu4Nl, DMF; c) IDCP, CHzClz-MeOH; then Zn, 95% EtOH, heat; d) BH3, THF, then Na202;
then Swem oxidation; e) DDQ, Et3N, CHzClz-HZO; t) NH4HC02, NaCNBH3, MeOH; g)
10%Pd-C, MeOH-HCOOH; then HCl, THF-H20

Mehta et. al. [41] reported a norbomyl route to azasugars (Scheme 1.23). A highly
functionalized cyclopentene derivative 169 was employed as the starting point, which
could be readily synthesized from norbomyl derivatives. After mesylation and elimination,
the cyclopentene ring was dihydroxylated to give a mixture of isomers 171 and 172. Then
the cyclopentane ring was broken by NaIO4 oxidation to form a ketal aldehyde as the key
intermediate. The double reductive amination finally furnished azasugar derivatives 175

and 176.

33

I OAc OAc -—OAc

HO,, Ho
HO IIIIO a 11110 _b_> HO’"! 11110 mfiquo
”OK—F H300 ’Ok HaCO ”K H3 C0 {OK

I

H3CO
1 591 70 17
JC
{OH OH fOAc
BUN ““O>< + BnN ““O>< d BHN “‘0 “no
4— 4—
""o ""0 ”"0 uno><
OCH3 OCH3 OCH3 OCH3
1 76 175 174 173

Scheme 1.23 A norbomyl route to azasugars: a) Mesylation; then NaOAc, DMF, 105°C,
6h, 77%; b) 0504, NMO(50% aq. solution), Me2C0:H20(4:1),48h, 86%; c) NaIO4(1.3eq),
DCM, 0°C; d) BnNHz, AcOH, NaCNBH3, 20h, -10°C to rt, 30% for two steps; e) KOH,
MeOH, 2h, 90%

In 2000, Hollingsworth et. al. [42] reported a general approach to the synthesis of
dideoxy and trideoxyiminoalditols from B-D-glycosides (Scheme 1.24). Afier protection
of the free hydroxyl groups, the B-D-glycosides, represented by 177, were oxidatively
ring-opened to form a keto-ester 179. Hydroxylamine was employed as the nitrogen
source to be introduced into the sugar structure. The reductive amination followed by
spontaneous cyclization formed lactam 181, while the 6-position was also reduced to
methyl group. After borane reduction, 1,6-dideoxynojirimycin 182 and 1,6-
dideoxygalactonojirimycin were synthesized in good yields. The synthetic strategy was

slightly modified to synthesize the nojirimycin lactam 185 by deacetylating the oxime 180

before the catalytic hydrogenation with hydrazine.

34

OH OAc OAc

a b O
O _> O _.> AcO /
0 A00 OM
HHC&/ 0M9 AcO 0M9 A00 e

OH OAc OAc o
177 178 179
] 0
OAc
NH e d
AGO AcO AcO OMe
OAC 0A0 O OAC O
182 131 180
if
OAc OH OH
NH h
Aoo H0 H0 NHNHz
0A0 0 0H 0 OH 0
135 184 183

Scheme 1.24 A general approach to the synthesis of dideoxy and trideoxyiminoalditols from B-
D-glycosides: a) Ac20, py.; b)CrO3, AcOH; c) NH20H, Py; (1) H2, Pd/C; e) BH3/T HF; f)
N2H4; g) H2, Pd/C; 11) A020, PY-

Reductive amination has been extensively used in many synthetic routes towards

azasugars. The easy installation of nitrogen source and cyclization to form the N-

heterocycle ring could often be achieved in a single step in the late stage of a total

synthesis, which provided flexibility of synthesis towards azasugars. However, one

inherent drawback of reductive amination is the stereochemistry control. The reductive

amination ofien involved chirality scrambling and regeneration process. Although the

stereoselectivity has been improved in some cases, mostly it is still generally not

stereospecific.

35

I. 3. 2. 3 Azasugar synthesis with N-alkylative cycli ation

As one of the best known organic reaction, N-alkylation has been also been
widely employed in azasugar synthesis. Based on its carbohydrate-like structure, it is easy
to imagine by converting the hydroxyl groups to leaving groups and double N-alkylation

may generate azasugar structure. The N-alkylation could be carried out in a single step or

multi-steps.

OH OH
0 3.. BnO 0“ COCI=3
8320 OH 9"0&/N\m‘—IL NBn
OBn OBn coc1=3
186 187
LicB
CHO
N3" ‘1 Ianon + BnO
OBn
191
at
9H OH Hg H OH Hg H OH
Ho 7 '1 f 0 r 9 HO =
COz‘Bu_> ——>
. , , N
HOW NH HO‘ N O H0\
192 193 3

Scheme 1.25 A synthesis of deoxynojirimycin and castanosperrnine by a glucosylamine
pathway: 3) benzylamine, CHCl3; then LiAlH4, THF, reflux, 5h; then trifluoroacctylation; b)
TBDMSCI, imidazole; then mesylation; then Bu4NF THF; then CH3ONa- -MeOH; c) NaBH4,
ethanol, 40°C; (I) Swem oxidation; e) Lithio t-butylacetate; t) hydrogenolysis, then (TFA- -l-l,20

60°C, 3h); g) diisobutylaluminum hydride;

In 1984, Ganem et. al. [43] developed a total synthesis of deoxynojirimycin l and
castanosperrnine 3 (Scheme 1.25). A glucosylamine pathway was employed to introduce

the amine functionality. Selectively benzylated glucose 186 was reacted with

benzylamine followed by reduction to give an amine, which was trifluoroacetylated to
fumish an amide 187. After selective silylation, mesylation and desilylation, this amide
was treated with base to furnish an epoxide 188 with inverted C-S. After deprotection of
this amide, the aminoepoxide cyclized spontaneously to a mixture of piperidine 190 and
azepane 189. The piperidine 190 was transformed to deoxynojirimycin 1 using
hydrogenolysis. The protected deoxynojirimycin 190 was oxidized to an aldehyde 191,
which was treated with lithio t-butylacetate to extend the carbon chain. Hydrogenolysis of
this amino ester 192 deprotected all the benzyl groups and gave a lactam 193, which was
further reduced to give castanosperrnine 3. By this method, the structure of

castanosperrnine 3 was first unambiguously established.

In 1991, Kibayashi et. al. [44] reported a non-carbohydrate based synthesis of
castanospermine 3 (Scheme 1.26). In this method, a highly functional epoxide 195 was
prepared by Sharpless epoxidation. The regiosepecific cleavage of the oxirane ring was
achieved by reaction with Et2A1N(CH2Ph)2 to introduced a benzyl-protected amine group.
After protection steps, the carbon chain was extended by reaction with the lithium enolate
of ethyl acetate. Although the undesired stereoisomer 199 was the major product, it could
be converted to the desired stereoisomer by Mitsunobu reaction. The primary silyl
protection groups were transformed to leaving group and a catalytic hydrogenolysis
afforded the fused-ring azasugar scaffold, which was further deprotected to give

castanosperrnine 3.

37

OH

TBDMSO —- TBDMSO TBDMSO

NBn2

O
l”
O
I
U'
I

0x6 0x0 —-> 0x0
194 195 195
*c
Ho 00 Et
MOMO ” 2 MOMQ CH0 raomso MOMQ OAc
" raomso '
TBDMSO - ~an 6 WNW d 2 ~an
076) ‘— o b ‘— OX0
199 193 197
1r

HO

MOM ”0
TBDMSO OTBDMS MOMQ, “93 H 0“
NB 9 T50 OTs h _
"2 —> NBHZ —"
o N

Ho
Ho“'

O

é“;
\“

o o
X x
200 201 3

Scheme 1.26 A synthesis of castanosperrnine from a highly functional epoxide: a) Sharpless
epoxidation; b) EtzAlN(CH2Ph)2, CHzClz, r.t.; c) AcCl(leq.), Et3N, CHZCIZ, 0°C; then
MOMC1,i-Pr2NEt,CHC13, reflux; d) LiAlH4, £120, rt; then DMSO, (cochz, Et3N, CHZCIZ, -
78°C to rt; e) AcOEt, LiN(SiMe3)2. THF, -78°C; 1) LiAlH4, EtZO, r.t.; then TBDMSCI,
imidazole, DMF, r.t.; then AcOH, Ph3P, (EtOCON=)2, C6H6, reflux; then LiA1H4, EtZO, r.t.; g)
N-Bu4NF, THF, r.t.; then TsCl, Py., rt; h) H2, Pd(OH)2, MeOH, then Et3N, MeOH, reflux; then
HCl, MeOH, reflux
Vogel et. al. [45] also developed a synthesis of castanosperrnine and analogs using
a “naked sugar” as the starting material (Scheme 1.27). Starting from this [2.2.1] bridged
ring system 202, the ketone was protected as a dibenzylacetal. Bromohydrin formation
and a rearrangement built two new chial centers with high yield and high stereoselectivity.
Baeyer-Villiger oxidation of 204 followed by methanolysis broke the bridged ring and
gave the tetrahydrofuran ring system 206. After reduction and mesylation, the nitrogen

atom was introduced by reaction with methanolic ammonia. Spontaneous N-alkylation

cyclization and following P(OEt)3-mediated condensation gave the fused-ring azasugar

scaffold 209. Then bromohydrin formation was used to functionalize the double bond.
The following epoxide formation, hydrolysis and deprotection steps afforded
castanosperrnine 3. 6-deoxycastanospermine and 6-deoxy-6-fluorocastanospermine were

also synthesized by slight modification of the final steps.

C) (3 Br c> Br c> ImeKl
[b .8- [¢ 3. & _°_, #1 d ,
OBn O —’
O OBn BnO O BnO O Bn
205

COOMe

O

    

202 203 204

    

208 207

Scheme 1.27 A synthesis of castanosperrnine and analogs using a "naked sugar" as the starting
material: a) BnOSiMe3, TMSOTf, CHZCIZ; b) Brz/CHZCIZ; then NaHCO3/ H20, -90°C, 98%;c)
Baeyer-Villiger oxidation, mCPBA, 95%; d) MeOH, SOCIZ; e) DIBAL, THF/toluene; then
MsCl, pyridine/CHZCIZ; f) 24% NH3 in EtOH/1120, 45°C, lday; then C1CH2C0C1,
pyridine/CHZCIZ; g) Ago/112804, 0°C; then P(0Et)3, 130°C; then acetylation; h) Brz, AgOAc,
AcOH/ACZO, 10°C; then MeOH/ SOClz, 20°C; then BEMP, CH3CN, 20°C; then BEMP,
CH3CN,H20 100°C; methanolysis and debenzylation;

Capitalizing on the C2 symmetry of D-mannitol, Dureault el. al. [46] developed an
enantioselective synthesis of 2,5-dideoxy-2,5-imino-D-mannitol and L-iditol. Double N-
alkylation of amine with selectively protected and derivatized D-mannitol was employed
as the key step (Scheme 1.28). Starting from 3,4-di-O-benzyl-D-mannitol 210, selective
primary benzylation followed by mesylation gave the 1,3,4,6-tetra-O-benzyl-2,5-di-O-
mesyl-D-mannitol 212. This suitably protected and derivatized compound 212 was then

treated with benzylamine. Double alkylation followed by hydrogenolysis afforded 2,5-

39

dideoxy-2,5-imino-D-iditol 214. To synthesize the more biologically important 2,5-
dideoxy-2,S-imino-D-mannitol 219, the stereochemistry at the 2- and 5- positions must be
inverted. Therefore L-iditol bis-epoxide 215 was employed as the starting material. After
epoxide ring opening and benzylation, tosylation, deacetalation and benzylation, the
similar synthetic intermediate 217 with 2-,5- inverted was obtained. The treatment of this
compound with benzylamine followed by hydrogenolysis therefore afforded 2,5-dideoxy-

2,5-imino-D-mannitol 219.

 

 

 

 

HO— BnO— BnO—
HO—l HO— MsO— BnO an
BnCH .2. BnO— _b> Bno— .9, :
—OBn ——OBn —OBn 3,10 6, OBn
N 6/
—0H L-CH POMS an
—OH OBn LOBn
210 211 212 213 214
O/ BnO— BnO—
\ e —0Ts r
O<\L 0N» BnO—
F—O O
TsO—J M50—
50 L
OBn

 

 

 

 

219

216

Scheme 1.28 A synthesis to 2.5-dideoxy-2,5-imino-D-mannitol and L-iditol using double N-
alkylation of protected alditols: a) Bu28n0 (2.1eq), toluene, reflux, 10h; then BnBr, Bu4NI,
70°C, 12h; 74%; b) MsCl, Et3N, DMAP, CH2C12, 80%; c) BnNHz (40eq), 120°C, 18; 78%; d)
H2, Pd/C, AcOH; then Dowex 50W-X8; 80%; e) NaH, BnOH, DMF, 20°C, 24h, 57%; then
TsCl, Et3N, DMAP, CH2C12, 84%; f) CF3COOH/1120 (9:1), 0°C, 2h, 86%; then
Cl3CC(NH)OBn, CH2C12/C6H,2(1 :2), CF 3SO3H, 25°C, 75%;

Again using the C2 symmetry of D-mannitol to advantage, Poitout et. al. [47] later

reported a synthesis of deoxynojirimycin 1 and analogues (Scheme 1.29).

40

 

 

 

 

0811 OH
Ho we +10ij 0“
e
HO_‘ / H...»
O\ I ’1
HO—‘ ”II/0 "I/0
BnO— j. BnO-
—-OBn "OBn—m> 223
HO OH
’_OH \ \ s
/O S ‘
—OH IIIIOH
210 220 ""OH—u>
N
222 224
OH
TBDMSO-fi \o How/0 50 3" Ho,,,, ‘50”
:83 . N a.
n —> BnO— 3:27 H
—OBn —OBn 1
-OMs /
—OTBDMS o\ 3 H0 SQH
225 226 HOIBIIIICDC;nOH _’ HO"'" OH
N
Bn H
228 229

Scheme 1.29 Synthesis of deoxynojirimycin from D-mannitol: a) Ph3P, DlAD, 130°C, 86%;
b) TBDMSCI, imidazole, DMF, 0°C, 80%; then MsCl, NEt3, CH2C12, 0°C, 98%; c) HCl,
MeOH, 20°C then NaOH, H20, 20°C, 75%; d) BnNHz, various conditions; e) H2, Pd black,
AcOH, 15h, 100%

The 1,2:5,6-dianhydro-3,4-di-O-benzyl-D-mannitol 220 and L-iditol 226 were
synthesized as the key intermediates. Starting from D-mannitol, 3,4-di-O-benzyl-D-
mannitol 210 could be synthesized based on selective protection and deprotection strategy.
Starting from this dibenzyl mannitol, Mitsunobu reaction gave the dibenzyl diepoxide
220 with D-manno configuration. This compound was treated with benzylamine under
various conditions to give protected L—deoxygulonojirimycin 221 and an azapane 222

with the D-manno configuration. The dibenzyl diepoxide 226 with L-ido configuration

was obtained from the same starting point, namely 3,4-di-O-benzyl-D-mannitol 210. This

41

was achieved by selective mesylation at 2,5-positions and di-epoxide formation. Then the
same double N-alkylation was carried out and deoxynojirimycin 1 was obtained along

with a 7-membered ring azasugar 229.

Since the N-alkylative pathway towards azasugars had been always very much
dependent on the protecting group manipulation, it would be always of interest to
synthesize these compounds with the use of fewer protecting groups. In 1993, Lundt et. a1.
[48] reported a synthesis of deoxyiminoalditols from aldonolactones (Scheme 1.30).
Their approach was to use aldonolactones as the chiral source. They could directly be
converted into bromodexoylactones by treatment with hydrogen bromide in acetic acid
(HBA). Leaving groups for nucleophilic substitution are introduced at C-2 and the
primary position. The ring closure of such bromodeoxylactones with ammonia therefore
generated deoxyiminoalditols without participation of protecting groups.
Dibromomannolactone 230 was treated with concentrated aqueous ammonia gave 3,6-
dideoxy—3,6—imino-D-alloamide 231. After hydrolysis, this amide was transformed to an
ethyl ester and subsequently reduced to form 1,4-dideoxy-1,4-imino-L-allitol 233, which
is a five-membered ring azasugar. The reaction was shown to go through a diepoxide
intermediate 236, which was attacked by ammonia to give the azasugar ring. By
interchange of the reduction and nucleophilc attack steps, the reaction was carried out
even more efficiently. Reduction of the dibromolactone 230 gave the dibromoalditol 234,
which was treated with aqueous ammonia to give the same product 233. Using no
protecting group manipulation, an azasugar was synthesized highly efficiently. With this
strategy, they further synthesized several other azasugars with D-allo, D-, L-talo, D- and

L—galacto and D-, L-ido configuration as potential glycosidase inhibitors.

42

 

. O O a N b N
HO H“ ——> HO H —> HO H
HO Br HO OH HO OH
230 231 232
1“ lo
HO H
OHOHBr N
1 1 1 e ”0 H
l l l
OHBr OH HO OH
23‘ 233
Mechanism:
OH
Brlllll IIIIO 11110 HO H
(DH—>0 N
——>HO H
IIHOH IIHOH IIIIOH HO OH
235 237 233

 

Scheme 1.30 Deoxyiminoalditols from aldonolactone: a) aq. NH3, r.t. 1h, 61%; b) aq. HBr,
reflux, 12h; 79%; then EtOH, AcCl, reflux, 3h, 100%; c) NaBH4/ EtOH, 0°C to r.t., 12h;
50%; d) NaBH4/ H20, H+ resin; 60%; e) aq. NH3, r.t., 2h, syrup 73%; cryst 51%

In 2001, Compernolle et. al. [49] developed a synthesis of deoxymannojirimycin
analogues using N-tosyl and N-nosyl activated aziridines derived from l-amino-l-deoxy-
glucitol (Scheme 1.31). After protection of l-amino-l-deoxy-glucitol, the primary
tosylation, C-2 mesylation and treatment with base gave the 1,2-tosylated aziridine 240,
which was opened by various nucleophiles from the primary position to give 2-
(tosyl)amino compound 241. After selective hydrolysis, Mitsunobu reaction furnished the

desired azasugar ring 243. This strategy could be used to synthesize many analogues of

deoxynojirimycin, which would be very useful potential glycosidase inhibitors.

43

 

Scheme 1.31 A synthesis of deoxymannojirimycin analogues using N-tosyl and N-nosyl
activated aziridines derived from l-amino—l-deoxy-glucitolz a) 4eq. Et3N, 1.5eq. p-TsCl,
CHZCIZ, 20°C, 30min; then 1.3eq. MsCl, 20°C, 1h; 92%; b) 2eq. NaH, THF, 20°C, 1h; 90%; c)
various nucleophiles, > 65% yield; d) Dowex 50X8-200, 9:1/MeOH/H20, rt. 30h-7days; e)
Ph3P, DEAD, THF, rt, 5days;

In 2004, Fuentes et. al. [50] reported a stereocontrolled preparation of azasugars
and their ethyl thioglycosides from glycosylamine. Anhydroazasugars were employed as
the key intermediates. A representative example was shown in Scheme 1.32. After several
protecting and deprotecting steps, the C-4 of glucosylamine was mesylated to give 246.
Treatment of this compound with base gave an anhydro azasugar 247, which was reduced
to give the S-membered ring azasugar 248 with galacto- configuration. This method took

advantage of the glycosylamine nitrogen functionality and could be used in some other

azasugar syntheses.

44

PMBO

b
NHP —> MSOIIIHI

 

Scheme 1.32 A stereocontrolled preparation of azasugars from glycosylamine employing
anhydrosugar as the key intermediate: a) anisaldehyde dimethyl acetal, DMF/PTSA,
50°C/20mmHg, 1h; then BzCl, py, rt, 24h; b) NaBH3CN/AcOH, rt, 2.5h; then MsCI, py, rt,
24h; c) NaOMe/DMF, 45°C/20mmHg, 15min; (1) NaBHzCN/AcOH, rt, 24h; then
NaOMe/MeOH, rt, 6h; e) MeOH/HCI, 65°C, 1h; then Dowex 50w8x.

The N-alkylation pathway is relatively straightforward in azasugar synthesis. The
general goal is to install nitrogen functionality and a leaving group in appropriate
positions. The nitrogen functionality could be installed using N-alkylation or some
suitable aminosugar could be used as starting materials. The leaving group could be
installed in various methods. The intramolecular nucleophilic N-alkylation was mostly

used to furnish the azasugar ring. To control the regio- and stereoselectivities, this N-

alkylation method ofien involved extensive manipulation of protecting strategies.

1.3.3 Other synthetic methodologies towards azasugars

Many groups have developed some other synthetic methodologies towards
azasugars. In 1985, Ganem et. al. [5]] synthesized deoxynojirimycin l and

deoxymannojirimycin 2 based on a high-yield, ring-forming aminomercuration (Scheme

45

1.33). Starting from bromosugar 250, reductive ring opening and reductive amination
gave the aminoalkene 252, which underwent aminomercuration to furnish an N-
heterocycle ring 254. After reductive oxygenation and deprotection, l-deoxynojirimycin

was obtained. l-deoxymannojirimycin could be synthesized by the similar procedure.

BnHN
BnO O Lano _. BnO
BnO BnO CHO BnO
BnO

OBn 252 OBn

25o OCH3 251
/ b
HOH C
2 NBn c BngHZC NB BnO NB"
BnO ..._ BnO " + BnO
BnO BnO
OBn OB n BngHZC OBn
255 254 253
who OBn HOHZC oa 03"
-o _, n 'NBn
BnO __> ,NBn BnO
BnO Bné) O + BnO
" CHZOH
256 OCH3 257 258

Scheme 1.33 Syntheses deoxynojirimycin and deoxymannojirimycin based on a high-yield,
ring-forming aminomercuration: a) Zn dust, 19:1 n-propyl a1cohol:water. benzylamine,
NaBH3CN, 91%; b) Hg(CF3COO)2, anhydrous THF; 61%; c) NaBl-I4-DMF-02, 70%

Hirai ct. al. [52] reported a synthesis of l-deoxymannojirimycin with a Palladium-
catalyzed cylization of urethanes. D-mannitol was transformed to a urethane, which under
catalysis of PdClz, gave the N-heterocycle ring. Although it involved many functional
transformation steps from D-mannitol to the urethane, the subsequent Pd-catalyzed
azasugar ring formation was important. Knight et. al. [53] also reported a
deoxymannojirimycin synthesis based on Pd-catalyzed N-heterocycle ring formation. The

latter was further transformed by epoxidation and dihydroxylation to form an azasugar

structure. Alkene metathesis has been also used to furnish the N-hetcrocycle structure in
the synthesis of azasugars [54-5 8]. The nitrogen fiinctionality itself does not participate in
the metathesis. However, it could be relatively easily installed in the early stage of the
synthesis. The resulted alkene could be easily functionalized by epoxidation,
dihydroxylation etc. Therefore the alkene metathesis became a powerful methodology of

azasugar synthesis, especially for fused-ring azasugars.

Enzymatic approaches have also been employed in the synthesis of azasugars [59-
67]. In these enzymatic or chemoenzymatic syntheses, aldolases were the enzymes most
often used. For example, when azide-bearing enzyme substrate 3-azido-2-hydroxy-
propanal was subjected to aldolase catalyzed reaction with dihydroxyacetone phosphate
(DHAP), the CC bond formation was catalyzed to form a nitrogen-contained
carbohydrate with high stereoselectivity. Subsequent reductive amination generated the
azasugars. Enzymatic synthesis towards azasugar could be relatively simple. However, it

still involved chemical manipulation before and after the enzymatic step.

In summation, azasugar synthesis has been extensively researched in the past two
decades and many synthetic strategies have been employed. Most of the syntheses of
azasugars are based on carbohydrates or other chirality-rich molecules. To introduce the
nitrogen function, several methodologies have been widely used. These include using
azide as an intermediate, reductive amination and N-alkylation. The synthesis of
azasugars starting from achiral or chirality-deficient molecules has also been extensively
explored. Dihydroxylation and epoxidation have been broadly used to introduce new
chiralities. The stereoselectivity have been achieved by both substrate control and reagent

control. Some other synthetic methodologies such as Pd-catalyzed heterocyclization and

47

ring-closing alkene metathesis have also been adopted in azasugar synthesis. Because of
their high selectivity, low requirement for protecting groups or activation and the mild
conditions under which they occur, enzymatic approaches to azasugars have also been
utilized. With all the achievements, there has been no universal synthetic methodology of
azasugars. Each methodology has its own advantages and disadvantages. This becomes
especially true as more and more new azasugar structures are discovered and studied
chemically and biochemically. There is still much to be done in the synthetic chemistry of

azasugars.

1.4 The structure-activity relationship (SAR) of azasugars as

glycosidase inhibitors and modified azasugars

Ever since the discovery of azasugars as glycosidase inhibitors, the structure-
activity relationship (SAR) has attracted much attention. Understanding the SAR could
help to obtain deeper insight into the glycosidation mechanism. Furthermore, given the
difficulty of synthesizing various compounds with azasugar structure, it is necessary to
deeply explore the SAR and discover the possible structure features leading to better
inhibition and lower toxicity. Apparently, it can help to rationally design and synthesize
modified azasugars as potential drug candidate. Many reviews have been published on
glycosidase mechanisms [1-10] and glycosidase inhibitions [68—76]. Here we describe the
structure-activity relationship (SAR) of azasugars as glycosidase inhibitors and some

modifications based on the SAR towards new azasugar inhibitors.

48

1.4.1 Shape and charge of azasugar glycosidase inhibitors as transition state

analogues

As mentioned earlier, the inhibition mechanism of azasugars has long been
believed to due to their resemblance with glycosidation transition state (Figure 1.3). The
protonated form of azasugars has a positive charge on the ring nitrogen, which is much
like the partial positive charge on the oxocarbenium-ion-like transition state. On the other
hand, these azasugars have sp3~hybridized anomeric carbon atom and chair conformation.
Logically, these inhibitors do not well resemble the shape of the oxycarbenium ion, which
is sp2-hybridized at the anomeric center. There are some other glycosidase inhibitors with
sp2-hybridized anomeric carbon, such as glyconolactones. Apparently, these inhibitors

are better shape mimic of glycosidation transition state (Figure 1.5).

OH OH OH
0 8 8
HO HO HO
HO ' ' HO \ 9 HO \
O O
OH OH OH
259 260 251

Figure 1.5 Gluconolactone as glucosidase inhibitor

D-Glucon0-1,5-lactone 259 is a relatively strong inhibitor of B-glucosidases
[77,78]. Leeback [79] ascribed this inhibition to the stereochemical similarities between
the lactone 259 and the oxocarbenitun-ion 261 based on their half-chair conformation.
Reese et. a1. [80] reasoned that by charge delocalization, the polar oxy group in 260 also
partially mimics the positive charge of the oxycarbenium ion intermediate. Although the

relative importance of shape and charge in the inhibition mechanism has never been

49

unambiguously established, the importance of charge was confirmed by the strong

inhibition of azasugars towards glycosidases.

On the other hand. it would be beneficial for the inhibition if the inhibitor could
resemble the transition state by both the charge and the shape. Wong et. al. [69,70]
compared the features of a set of azasugar inhibitors with their inhibition constants (Ki)

(Table 1.3).

Table 1.3 Inhibition Potencies of Various Azasugars (IIM) (From Look, G. C., Fotsch, C. H. and
Wong, C. H., Acc. Chem. Res., 1993, 26, 182-90)

 

 

 

 

 

 

 

 

 

3:331 Yeast Jack Green Bovine Aspergil- 5:13:16
B- a- bean a- coffee bean liver [5- lus niger -L-y
gluco- manno- a-galacto- galacto— [3-xylo- a
gluco- sidase sidase sidase sidase sidase fuco-
SIdase SIdase
OH
Hgflw sis 921 825
on
OH
HO ”“ 18 8 7 400 NI
HO
OH
OH 30 :1:
0H
-NH 5300 68
Ho Ho 1 0
Hogfl‘ NI NI NI 0.0053 NI
OH
OH
”°&Z: 7.8 3.3 N1 N1 250
HO OH
°” 910 .4:
no N” 19 2.8 3100 50
HO OH 20
NH
0H 0H 0H 1.4
0H
“° 5... 0.048 > 100
HO "Hz

 

 

 

 

 

 

 

 

 

 

50

Deoxynojirimycin analogs mimic the charge by the protonated ring-nitrogen. The
placement of the hydroxyl groups is also very important because of the steric
requirements to bind in the enzyme active site [81]. Thus, deoxynojirimycin is a good
inhibitor towards a- and B-glucosidase but has weak or no inhibition towards
mannosidase and galactosidase. Correspondingly, deoxymannojirimycin and
deoxygalactonojirimycin are selective inhibitors towards mannosidase and galactosidase,
respectively. Moreover, the inhibition of mannosidase by deoxynojirimycin and of
glucosidase by deoxymannojirimycin was found although it was not as strong as when the
stereochemistry was totally matched. This suggested that the matched C-2 hydroxyl
orientation is not strictly necessary for glycosidase inhibition. This could be explained by

the half-chair conformation of the oxycarbenium ion intermediate in glycosidation.

The amidine inhibitors are relatively more representative of the flattened half-
chair resemblance. Because the hydroxyl group orientation of this amidine inhibitor only
matches the gluco- configuration, only strong inhibition toward glucosidase would be
expected. In fact, this amidine inhibitor (Table 1.3) is a very strong inhibitor of several
glycosidases, including not only glucosidase, but also mannosidase and galactosidase.
This also suggests that the matching configuration is not strictly necessary for good
inhibition. To overcome some drawbacks of amidine inhibitors such as unstability and
complicated synthesis, some five-membered ring azasugar inhibitors were also developed.
Similarly, these inhibitors all have a half-chair like conformation and they could also
mimic the charge of oxycarbenium ion by the nitrogen charge. Although the ring structure
of these five-membered ring azasugar inhibitors is not directly related to that of substrates.

they turned out to be good inbibitors towards both a- and B-glucosidases. When the

51

hydroxyl group orientation of these 5-membered ring azasugars is compared with the
substrate and deoxynojirimycin, they are more like a-D—fi'ucto-furanose analogues. The
flatter five-membered ring is believed to closer mimic the half-chair conformation found
in the oxocarbenium-like transition state, resulting in good inhibition [73]. Moreover,
these five-membered ring inhibitors showed broad-spectrum inhibition towards different
enzymes. This suggested the relatively greater importance of the half-chair conformation

than the correct hydroxyl group orientation.

1.4.2 Substitution effects of azasugars as glycosidase inhibitors

To find out the N-substitution effect on the inhibition properties of azasugars, the
a-glucosidase inhibitory activity was measured in HepG2 cells for a series of N-alkylated
deoxynojirimycin [68, 82]. When the ring nitrogen of deoxynojirimycin was substituted
with a straight alkyl (or aryl) chain, the inhibition of a—glucosidase was increased. N-
methyl-, N-butyl—, N-pentyl- and N-benzyl-deoxynojirimycin showed higher activity than
the unmodified deoxynojirimycin. When the alkyl chain was extended to N-decyl-
deoxynojirimycin, the detergent-like properties were observed. Therefore clear inhibition
was only observed when the concentration was 5-10 times lower. The increase on
inhibition by N-alkylation was ascribed to the conformation change of the C-6 hydroxyl
group (Figure 1.6). While deoxynojirimycin prefers a gauche-trans conformation, the N-
alkylated derivatives prefer a gauche-gauche conformation. Interestingly, the C-1
hydroxyl group of castanosperrnine has a similar orientation correspondent to the C-6
hydroxyl group of N-alkylated deoxynojirimycin. Castanospermine inhibits a-glucosidase
I more strongly than a-glucosidase II. It was also observed that N-alkylation of

deoxynojirimycin induces a shift in specific inhibition of purified gludosidase from o-

52

glucosidase II to III-glucosidase l [83, 84]. These discoveries suggested that the
conformation of C-6 hydroxyl group plays an important role in the interaction with

glycosidases.

   

H0
H0 N
H0
H5 OH OH
Deoxynojirmycin N-alkyl-deoxynojirimycin castanosperrnine

Figure 1.6 Preferred conformation of deoxynojirimycin, N-alkyI-deoxynojirimycin and
castanospermine (Figure abstracted from: van den Broek, L. A. G. M., Vermaas, D. J.,
Heskamp, B. M., van Boeckel, C. A. A., Tan, M. C. A. A., Bolscher, .l. G. M., Ploegh, H.
L., van Kenenade, F. J., de Goede, R. E. Y. and Miedema, F ., Recueil des Travaux
Chimiques des Pays-Bas, 1993, 112, 82-94)

When the branching of the N-alkyl side chain was introduced into the
deoxynojirimycin modification, decreased inhibitory activity of glucose trimming was
observed. N-isobutyl-deoxynojirimycin was less active than N-butyl derivative, while the
N-(cycolpropylmethyl)-deoxynojirimycin was almost inactive. These results suggested

that branching of the alkyl chain might result in a sterically less favorable configuration.

By comparing the hydroxyl group orientation of all the biologically important
glycopyranoses, it is apparent that the 3-OH is the only stereochemically conserved
position. Naturally, its role in the glycosidation process is of interest. The 3-O-methy1-
deoxynojirimycin and deoxymannojirimycin was synthesized to clarify its biological role
[68]. The inactive nature of the 3-O-methyl derivatives as inhibitors of corresponding

glycosidase suggested its importance for binding of deoxynojirimycin and

53

deoxymannojirimycin to the enzymes, presumably through an interaction with a hydrogen

bond acceptor.

The ester derivatives of azasugars have also been made to act as pro-inhibitors. N-
Benzyl-6-O-butylryl-deoxynojirimycin and N-decyl-6-O-benzoyl-deoxynojirimycin
showed activity comparable to that of their respective parent N-alkyl derivatives [82].
However, N-benzyl-tetra-O-acetyl-deoxynojirimycin and N-butyl—tetra-O-acetyl-
deoxynojirimycin showed no inhibitory activity towards a-glucosidase. When no free
hydroxyl group is present, the intracellular hydrolysis of ester pro-inhibitors may be
slowed down greatly therefore no inhibition was shown. On the other hand, 6-O-butyryl-
castanosperrnine was shown to be 20 times more active in vitro than unmodified

castanosperrnine [85].

1.4.3 l-N-iminosugars as potent and selective inhibitors

In our previous discussion about glycosidase mechanism, the oxocarbenium
cation was purported to be the transition state. The positive charge is delocalized between
anomeric position and the ring oxygen. Wong et. al. [86] carried out a molecular
mechanics (MMZ) modeling study of the B-glucosidase reaction in terms of conformation
and charge distribution. They found that the positive charge was generated at the
anomeric position rather than at the ring oxygen. On the other hand, by molecular
electrostatic surface analysis of nucleoside hydrolysis, Schramm et. al. [87] concluded
that the highly localized charge between the ring heteroatom and the anomeric position is
not favored. Despite the controversy, it is apparent that the anomeric position would hear

substantial positive charge in the glycosidation transition state. It is also believed that the

54

retaining B-glucosidase mechanism includes a covalent glycosyl-enzyme intermediate
(Figure 2 b). Therefore 1-N-iminosugars were developed as a transition state analog
mimicking the positive charge at the anomeric positon. Ichikawa et. al. [88] synthesized a
series of l-N-iminosugar derivatives and evaluated their inhibitory potencies towards
different enzymes (Table 1.4). Generally, the l-N-iminosugar with gluco- and galacto-
configurations were very good inhibitors of the corresponding B-glycosidase, while they
showed much weaker inhibition of the corresponding a-egCOSidaSB. This selectivity was
expected since the anomeric charge is more important in transition state or covalent
intermediate of B-glycosidation. Furthermore, both of them showed a weaker inhibition
towards the B-glycosidase with the other configuration, which indicates that charge mimic
was more important for a transition state analog in this situation. However, when both

charge and configuration were not matched well, almost no inhibition was found.

55

 

 

 

 

 

 

 

 

:o
:0 own 82 A ooo _ A ooo .A ooo _ A on &
. 82
82A SSA 82A 82A 82A on Iz\/ oazoz em 82A 82A 82A 82A can one
.5 £
:2
82A 2. E E 2 o3 0:
ow coo—A coo—A coo—A coo—A OON 52 x0 0:
o: :0 :0
:2 :o :2 :0
o? 82A 2 8 88A 2: 2m: 82A 88A as 2. 88A o? %o;
:o :0
coo _ A U: oo no U: of 3mm
.52 ofox @ 0:
82A 82A 82A 8m 9. on. //\\}1 .5
one a: .a 9.
:z N 82A 2 a. m 88A 3 fi
010: :o :0
com coo—A coo—A Gem cog cow ox 12%?
IO
coo—A ooomA oom N56 ooomA 36 xo :0
:mZ OI :mZ IO OI
82A 2. 5 9. 88A cm «9:1 82A 2 2 E 82 own or
:z :oo: szozn.v o:
o
coo—A oom coo—A ov coo—A w.» fire ooo ooomA OOOmA ooomA 23 o6 % I
I xo
x0 .52 o:
coo—A coo—A ooo_A coo—A coo—A coo—A \i v: c: v: On 9: am $0:
\1 80:
:2 :0 12:0 0:
:z o: :
82A 82A 82A 82A 82A 82A $8M? 82A on” So? OR or 3 who
0E8 SEE. Ewe Emd 038 u.».& 03-5 5E8 Ewe Ema Bwé o_w.n
omaEmoobw flozoEE mommEmo§m .832:—

 

Aanoom 63 £23 Gem. .525 .EV.

x ..> .825 as .2 3.32:2 ..> .2333 ..> 539.22 see Bags 2%: :21 .5. V Emaoséiz me 86:28 e952,: 2 2.5

56

Bols et. al. [89-91] also synthesized and tested some l-azasugars, disaccharide
derivatives of l-azasugars and some diazasugars (azafogamines). They found that 1-
azasugar were strong glycosidase inhibitors and particularly strong inhibitors of B-
glycosidase. Their protonated form being transition state analogs also explained the
inhibition. After protonation, they generally mimic the anomeric charge of the transition
state. An argument was made about the azasugar protonation tendency when the pKa of
azafagomine (3.9) and the pKa of the protonating acid in B-glycosidase (6.7) were
compared. The enzyme might stabilize the protonated form and actually increase the
basicity of the inhibitors. F urtherrnore, the direct observation of the protonation state of a
l-N-iminosugar derivative upon binding was reported by Davies et. al. [92, 93] (Figure
1.7). The 3-D structure, combined with the pH dependence of inhibition, gave more

evidence of the proposed glycosidase mechanism.
)Ala234
P H

o'
oH/BAN“ 04C)”
H __
Wfi+H31MA,O Tyr202
I a 9"
H' X’ NH 26A ”X0

H”

U W
His101 H (Slum

Tyr66

Glu 139

 

l l l

-2 -1

Figure 1.7 Schematic representation of the interactions observed between CelSA and the
cellobio-derived isofagomine. All hydrogen atoms shown have been observed experimentally.
Only the -1 subsite interactions are shown in detail. Distances around NI are indicated. (Varrot,
A., Tarling, C. A., Macdonald, J. M., Stick, R. V., Zechel, D. L., Withers, S. G. and Davies, G.
J ., J. Am. Chem. Soc. 2003, 125, 7496-7)

57

1.4.4 Aromatic ring-fused azasugars as glycosidase inhibitors

Besides the charge distribution, the effect of the ring conformation (chair versus
half-chair) has been always an interesting topic. Vesella et. al. [75, 94-96, 101-103] and
Tatsuta et. al. [97-100] synthesized a series of aromatic ring-fused azasugars to explore

this problem (Figure 1.8).

 

 

A) OH OH 0“
Ho N §IN Ho \ ,N
HO \N N
OH OH
262 263 254
8) OH OH OH OH
Ho
N \
Ho N \ Ho N \ Ho 6
Ho \N H0 \N CH N
CH
265 266 267
C) OH OH OH
HO N \ C02M8H0 N \ HO N \ 002Me
0” OH 002m 0“ cone
266 269 27°
D) OH
HO N/N§N
HO \
OH
271 272

Figure 1.8 Some azasugars fused with A) tetrazoles; B) imidazoles; C) pyrroles; D) 1,2,3-
triazoles

The tetrazole-fused azasugars were shown to be half-chairs by X-ray analysis.

They are relatively stable in solutions over a pH range of 1-14. The measurement of

58

inhibition activities was shown to be similar with the corresponding glycono-1,5-lactones.
Their configurational selectivity was also tested for the gluco-and manno-tetrazoles
towards glucosidases and mannosidases [94-96]. Although all the enzymes were inhibited
by both tetrazoles, the tetrazoles with the correct configuration was shown to be much
more effective. Also, the inhibition of the a-glycosidases was shown to be much weaker
than that of the B-glycosidases. The results showed the configurational selectivity of these

glycosidases and inhibitors.

The imidazoles with gluco-, manno- and galactono- configuration were also
subjected to inhibition activity measurement and the results are shown in Table 1.5 [75,
97-100]. Compared with the tetrazoles, the imidazoles were found to be more potent
inhibitors of B-glycosidases. They also show little activity towards a-glucosidase. To
compare with the imidazole series, pyrroles fused azasugars were also synthesized and
subjected to the inhibition activity studies [101-102]. These compounds were shown to be

generally very weak inhibitors.

Table 1.5. leo values [uM] of the annulated imidazoles with gluco—, manno- and galacto-
configurations. (Table adapted from Heightman, T. D. and Vasella, A. T., Angew. Chem. Int. Ed.,
1999, 38, 750-70)

 

 

 

 

 

 

Enzyme(Source) Gluco- Manno- Galacto-
a-glucosidase
(b akers’ yeast) 105 >500 >500
B-glucosidase
(almond) 0.7 17.5 0.5
B'ma‘m‘iS‘dase 300 0.1 1 370
(snail)
B-galactosidase >500 >500 0 008
(E. coli) '

 

 

 

 

 

59

Vasella et. al. suggested an in-plane protonation hypothesis to explain these
results [75] (Figure 1.9). With this hypothesis, the N-l of the tetrazole would be very
important in the inhibition. Therefore, 1,2,3-triazoles fused-azasugar were synthesized
and tested of the glycosidase inhibition [103]. The 1,2,3-triazols was shown to possess
very similar conformation to their tetrazole analogues. However, the inhibition of these
triazoles was found to be much weaker than tetrazoles, supporting the in-plane or lateral

protonation at N-l of the tetrazoles.

Figure 1.9 Proposed direction of protonation A) of a glucoside perpendicular to the plane of
the ring and B) and C) of gluco-tetrazole and a glucoside in the plane of the ring

1.5 Summary

Glycochemistry and glycobiology have been developing rapidly in the last decade.
Azasugars as glycosidase inhibitors has attracted both academic and industrial interest.
Many synthetic methodologies, starting from carbohydrate or non-carbohydrate, have
been reviewed in several categories. The structure activity relationship of azasugar as

glycosidase inhibitors has been briefly discussed.

6O

REFERENCE

[1] Varki, Ajit; Cummings, Richard; Esko, Jeffrey; Freeze, Hudson; Hart, Gerald;
Marth, Jamey; Editors. Essentials of Glycobiology. 1999, 653 pp.

[2] Helenius, A.; Aebi, M. Science 2001, 291. 2364.

[3] Brockhausen ,I.; Schutzbach, J .; Kuhns, W. A C TA ANA TOMICA 1998, 161, 36.
[4] Vasella, A.; Davies, G. J .; Bohm, M. Curr. Opin. Chem. Biol. 2002, 6, 619.
[5] Withers, S. G. Carbohydr. Polym. 2001, 44, 325.

[6] Rye, C. S.; Withers, S. G. Curr. Opin. Chem. Biol. 2000, 4, 573.

[7] Zechel, D. L.; Withers, S. G. Acc. Chem. Res. 2000, 33, ll.

[8] McCarter, J .; Withers, S. G. Curr. Opin. Struct. Biol. 1994, 4, 885.

[9] Sinnott, M. L. Chem. Rev. 1990, 90, 1171.

[10] Legler, G. Adv. Carbohydr. Chem. Biochem. 1990, 48, 319.

[11] Leeper, F. J .; Howard, S. Tetrahedron Lett. 1995, 36, 2335.

[12] Rodriguez, R.; Bermejo, F. Tetrahedron Lett. 1996, 37, 5581.

[13] Zhou, W.; Xu, Y. Tetrahedron Lett. 1996, 37, 1461.

[14] Zhou, W.; Xie, W.; Lu, 2.; Pan, X. Tetrahedron Lett. 1995, 36, 1291.

[15] Meyers, A. 1.; Price, D. A.; Andres, C. J. Synlett 1997, 533.

[16] Resek, J. E.; Meyers, A. I. Tetrahedron Lett. 1995, 36, 7051.

[17] Vicente, J. D.; Arrayas, R. G.; Canada, J .; Carretero, J. C. Synlet12000, 53.
[18] Hunt, J. A.; Roush, W. R.J. Org. Chem. 1997, 62, 1112.

[19] Mukai, C.; Sugimoto, Y.; Miyazawa, K.; Yamaguchi, S.; Hanaoka, M. J. Org. Chem.
1998, 63, 6281.

[20] Comins, D. L.; F ulp, A. B. T etrahedron Lett. 2001, 42, 6839.

[21] Somfai, P.; Marchand, P.; Torsell, S.; Lindstrom, U. M. Tetrahedron 2003, 59, 1293.

61

[22] Hendry, D.; Hough, L.; Richardson, A. C. Tetrahedron Lett. 1987, 28, 4597 and
4601.

[23] Hendry, D.; Hough, L.; Richardson, A. C. Tetrahedron 1988, 44, 6153.
[24] Fleet, G. W. J .; Ramsden, N. G.; Witty, D. R.. Tetrahedron 1989, 45, 319 and 327.

[25] Fleet, G. W. J .; Carpenter, N. M.; Petursson, S.; Ramsden, N. G. Tetrahedron Lett.
1990, 31, 409.

[26] Fumeaux, R. H.; Tyler, P. C.; Whitehouse, L. A. Tetrahedron Lett. 1993, 34, 3609.
[27] Zou, W.; Szared, W. A. Carbohydr. Res. 1994, 254, 25.

[28] Spreitz, J .; Stutz, A. E.; Wrodnigg, T. M. Carbohydr. Res. 2002, 33 7, 183.

[29] Gerspacher, M.; Rapoport, H. J. Org. Chem. 1991, 56, 3700.

[30] Kim, N.; Choi, J .; Cha, J. K. J. Org. Chem. 1993, 58, 7096.

[31] Pearson, W. H.; Hembre, E. J. J. Org. Chem. 1996, 61, 7217.

[32] Reitz, A. B.; Baxter, E. W. Tetrahedron Lett. 1990, 31, 6777.

[33] Baxter, E. W.; Reitz, A. B. Bioorg. Med. Chem. Lett. 1992, 2, 1419.

[34] Baxter, E. W. and Reitz, A. B., J. Org. Chem. 1994, 59, 3175.

[35] Matos, C. R. R.; Lopes, R. S. C.; Lopes, C. C. Synthesis 1999, 4, 571.

[36] Enright, P. M.; O’Boyle,K. M.; Murphy, P. V. Org. Lett. 2000, 2, 3929.

[37] O’Brien, J. L.; Tosin, M.; Murphy, P. V., Org. Lett., 2001, 3, 3353.

[38] Overkleeft, H. S.; van Wilterburg, J .; Pandit, U. K. Tetrahedron 1994, 50, 4215.
[39] Zhao, H.; Mootoo, D. R. J. Org. Chem. 1996, 61, 6762.

[40] Zhao, H.; Hans, S.; Chen, X.; Mootoo, D. R. J. Org. Chem. 2000, 66, 1761.

[41] Mehta, G.; Mohal, N. Tetrahedron Lett. 2000, 41, 5741.

[42] Pistia, G.; Hollingsworth, R. I. Carbohydr. Res. 2000, 328, 467.

[43] Bemotas, R. C.; Ganem, B. Tetrahedron Lett. 1984, 25, 165.

[44] Ina, H.; Kibayashi, C. Tetrahedron Lett. 1991, 32, 4147.

62

[45] Reymond, J. L.; Pinkerton, A. A.; Vogel, P. J. Org. Chem. 1991, 56, 2128.
[46] Dureault, A.; Portal, M.; Depezay, J. C. Synlett, 1991, 4, 225.

[47] Poitout, L.; Merrer, Y. L.; Depezay, J. C. Tetrahedron Lett. 1994, 35, 3293.
[48] Lundt, I.; Madsen, R. Synthsis 1993, 714 and 720.

[49] Mao, H.; Joly, G. J.; Peeters, K.; Hoornaert, G. J.; Compemolle, F. Tetrahedron
2001, 57, 6955.

[50] Fuentes, J.; Saygo, F. J.; Illangua, J. M.; Gasch, C.; Angulo, M.; Pradera, M. A.
Tetrahedron: Asymmetry 2004, 15, 603.

[51] Bemotas, R. C.; Ganem, B. Tetrahedron Lett. 1985, 26, 1123.

[52]Yokoyama, H.; Otaya, K.; Kobayashi, H.; Miyazawa, M.; Yamaguchi, S.; Hirai, Y. T.
Org. Lett. 2000, 2, 2427.

[53] Knight, J. G.; Tchabanenko, K. Tetrahedron 2003, 59, 281.
[54] Overkleefi, H. S.; Pandit, U. K. Tetrahedron Lett. 1996, 37, 547.

[55] Pandit, U. K.; Overkleefi, H. S.; Borer, B. C.; Bieraugel, H. Eur. J. Org. Chem.
1999, 64, 959.

[56] White, J. D.; Hrnciar, P. J. Org. Chem. 2000, 65, 9129.

[57] Martin, R.; Moyano, A.; Pericas, M. A.; Riera, A. Org. Lett. 2000, 2, 93.

[58] Lindsay, K. B.; Pyne, S. G. J. Org. Chem. 2002, 67, 7774.

[59] Kinast, G.; Schedel, M. Angew. Chem. Int. Ed. Engl. 1981, 20, 805.

[60] Ziegler, T.; Straub, A.; Effenberger, F. Angew. Chem. Int. Ed. Engl. 1988, 27, 716.
[61] Pederson, R. L.; Kim, M. J .; Wong, C. H. Tetrahedron Lett. 1988, 29, 4645.

[62] Pederson, R. L.; Wong, C. H. Heterocycles 1989, 28, 477.

[63] von der Osten, C. H.; Sinskey, A. J.; Barbas, C. F.; Pederson, R. L.; Wang, Y. F.;
Wong, C. H. J. Am. Chem. Soc. 1989, I I 1, 3924.

[64] Kajimoto, T.; Chen, L.; Liu, K. K.; Wong, C. H. J. Am. Chem. Soc. 1991, 113, 6678.

[65] Liu, K. K.; Kajimoto, T.; Chen, L.; Zhong, Z.; Ichikawa, Y.; Wong, C. H. J. Org.
Chem. 1991, 56, 6280.

63

[66] Henderson, 1.; Laslo, K.; Wong, C. H. Tetrahedron Lett. 1994, 35, 359.

[67] de Raadt, A.; Stutz, A. E. Tetrahedron Lett. 1992, 33, 189.

[68]van den Broek, L. A. G. M.; Vermaas, D. J .; Heskamp, B. M.; van Boeckel, C. A. A.;
Tan, M. C. A. A.; Bolscher, J. G. M.; Ploegh, H. L.; van Kenenade, F. J .; de Goede,
R. E. Y.; Miedema, F. Recl. T rav. ChimPays-B. 1993, 112, 82.

[69] Look, G. C.; Fotsch, C. H.; Wong, C. H. Acc. Chem. Res. 1993, 26, 182.

[70] Gijsen, H. J. M.; Qiao, L; Fitz, W.; Wong, C. H. Chem. Rev. 1996, 96, 443.

[71] Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetrahedron: Asymmetry
2000, 11, 1645.

[72] Compain, P.; Martin, O. R. Bioorgan. Med. Chem. 2001, 9, 3077.

[73] Wrodnigg, T. M. Monatsh. Chem. 2002, I33, 393.

[74] Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem. Rev. 2002, 102, 515.
[75] Heightman, T. D.; Vasella, A. T. Angew. Chem. Int. Ed. 1999, 38, 750.

[76] Asano, N. Glycobiology 2003, I3, 93R-104R.

[77] Ezaki, S. J. Biochem-Tokyo 1940, 32, 91.

[78] Horikoshi, K. J. Biochem-Tokyo 1942, 35, 39.

[79] Leeback, D. H. Biochem. Bioph. Res. Co. 1968, 32, 1025.

[80] Reese, E. T.; Parrish, F. W.; Ettlinger, M. Carbohydr. Res. 1971, 18, 381.

[81] Dale, M. P.; Ensley, H. E.; Kein, K.; Sastry, K. A. R.; Byers, L. Biochemsistry 1985,
24, 3503.

[82] Tan, A.; vanden Broek, L.; van Boeckel, S.; Ploegh, H.; Bolscher, J. J. Biol. Chem.
1991, 266, 14504.

[83] Zeng, Y.; Pan, Y. T.; Asano, N.; Nash, R. J.; Elbein, A. D. Glycobiology 1997, 7,
297.

[84] Szumilo, T.; Kaushal, G. P.; Elbein, A. D. Arch. Biochem. Biophys. 1986, 247, 261.

[85] Sunkara, P. S.; Taylor, D. L.; Kang, M. S.; Bowlin, T. L.; Liu, P. S.; Tyms, A. S.;
Sjoerdsma, A. Lancet 1989, 1206.

64

[86] Kajimoto, T.; Liu, K. K.-C.; Pederson, R. L.; Zhong, Z.; Ichikawa, Y.; Porco, J. A.,
Jr.; Wong, C.-H. J. Am. Chem. Soc. 1991, 113, 6187.

[87] Horenstein B. A.; Schramm V. L. Biochemistry 1993, 32, 7089 and 9917.

[88] Ichikawa, Y.; Igarashi, Y.; Ichikawa, M.; Suhara, Y. J. Am. Chem. Soc. 1998, 120,
3007.

[89] Dong, W.; Jespersen, T. M.; Bols, M.; Skrydstrup, T.; Sierks, M. R. Biochemistry
1996, 35, 2788.

[90] Bols, M.; Hazel], R.; Thomsen, 1. Chem. Eur. J. 1997, 3, 940.
[91] Bols, M. Acc. Chem. Res. 1998, 31 , 1.

[92] Varrot, A.; Tarling, C. A.; Macdonald, J. M.; Stick, R. V.; Zechel, D. L.; Withers, S.
G.; Davies, G. J. J. Am. Chem. Soc. 2003, 125, 7496.

[93]Zechel, D. L.; Boraston, A. B.; Gloster, T.; Boraston, C. M.; Macdonald, J. M.;
Tilbrook, D. M. G.; Stick, R. V.; Davies, G. J. J. Am. Chem. Soc. 2003, 125, 14313.

[94] Ermert, P.; Vasella, A. Helv. Chim. Acta. 1991, 74, 2043.

[95] Ermert, P.; Vasella, A.; Weber, M.; Rupitz, K.; Withers, S. G. Carbohydr. Res.
1993, 250, 113.

[96] Heightman, T. D.; Ermert, R; Klein, D.; Vasella, A. Helv. Chim. Acta. 1995, 78,
514.

[97] Tatsuta, K.; Miura, S.; Ohta, S.; Gunji, H. J. Antobiot. 1995, 48, 286.

[98] Tatsuta, K.; Miura, S.; Ohta, S.; Gunji, H. Tetrahedron Lett. 1995, 36, 1085.
[99] Tatsuta, K.; Miura, S. Tetrahedron Lett. 1995, 36, 6721.

[100] Tatsuta, K.; Miura, S.; Gunji, H. Bull. Chem. Soc. Jpn. 1997, 70, 427.

[101] Granier, T.; Gaiser, F.; Hintermann, L.; Vasella, A. Helv. Chim. Acta. 1997, 80,
1443.

[102] Panday, N.; Granier, T.; Vasella, A. Helv. Chim. Acta. 1998, 81, 475.

[103] Heightman, T. D.; Locatelli, M.; Vasella, A. Helv. Chim. Acta. 1996, 79, 2190.

65

Chapter 2

NBS-Initiated Dioxonium Cation Rearrangement and Applications
of the Selective Bromination and Chirality Manipulation of

Carbohydrates

ABSTRACT

In the attempt at NBS-bromination of benzylidene acetal protected D-
glucoheptonolactone derivative, a dioxonium cation rearrangement involving neighboring
pivaloyl group was discovered. The rearrangement mechanism was further explored.
Based on the experimental results, a new methodology of carbohydrate structure chirality
manipulation was developed. A series of synthetically useful selectively brominated and
protected D- and L-aldonolactone derivatives were synthesized. This methodology was
further applied in the synthesis of an orthogonally protected bromobutanetriol as a chiral

building block.
2.1 Hanessian-Hullar reaction and dioxonium cation rearrangement

In the 1960’s, Hanessian and Hullar [1-3] independently discovered that when
some benzylidene acetals were treated with N-bromosuccinimide (NBS) in carbon
tetrachloride, regioselective opening of the acetal ring occurred. In this reaction, usually
the less hindered position is brominated while the other position becomes benzoyl
protected, as shown in Scheme 2.1 a. In the cleavage of 4,6-O-benzylidene acetals of a
typical hexopyranoside 1, the product formed is not the 4—bromo, 6-benzoate (via the

more stable secondary carbocation). The 6-bromo 4-benzoate 2 is formed instead. One

66

feature of the reaction is the extent to which its outcome is dominated by sterics rather
than by electronic consideration. The formation of a cyclic 1,3 -dioxonium species 3 is the
key feature of the NBS oxidation of acetals. This reaction also works for dioxolane
benzylidene acetal and generally shows fairly good regioselectivity towards the formation
of an axial bromo group. NBS treatment of 4 therefore preferentially yielded 5a instead of

5b, although 5b should be more favored thermodynamically (Scheme 2.1 b) [4].

 

NBS, (9 B,
Ph/TO BaCO3 Ph/VO 320
a) (.340 ' (1)10 —’ ”0
con, Ho
Ho
HO OCH3 reflux OCH3 60% OCH3
1 3 2
OCH3 BOCHa OCH3
r
b) 0‘
/%/OH on + mm
03% 082 082
79% (5:1)
4 5a 51)

Scheme 2.] The regioselective cleavage of benzylidene acetals with NBS (From
Hanessian, S.; Pleassas, N. R. J. Org. Chem. 1969, 34, 1035; Bundel, D. R.; Josephson,
S. Can. J. Chem. 1978, 56, 2686.)

The mechanism is generally believed to go through a radical initiation and
benzoxonium cation formation process (Scheme 2.2 a). If water is present [5], alcohol
instead of the bromo-compound is formed (Scheme 2.2 b). The regioselectivity in this
case usually favors equatorial free hydroxyl group formation. Compound 7 was
selectively synthesized by treatment of 6 with NBS. Named as the Hanessian-Hullar
reaction, the oxidative cleavage of benzylidene acetals with NBS not only deprotects the

acetal to form another protection group, but also functionalizes the carbohydrate structure.

67

It has become an important deprotection and bromination methodology in synthetic

organic chemistry.

 

,‘fir /(T‘JSUC B
Ph 00 ph/VQ Ph k0O
6') H0 30 HO
HO HO HO
1 OCH3 OCH3 OCHa
Br6
0 ( 8'
Ph B
750 —> 2HO
HO
HO HO OCHs
OCH3 2
OCHa OCH3
b) O 0in NBS, H20 0 0in
O BaCOa. Light
Ph 6 ‘/ H
6 OCH3 Ph
OCH3
O OPiv O
—* o} H OPiv 72%
OYG 03 OH Piv=OC(CH)
z 33
Ph 7

Scheme 2.2 The mechanism of Hanessian-Hullar reaction (a) and the water-dependent NBS
regioselective cleavage of benzylidene acetals (b) (Hanessian, S.; Pleassas, N. R. J. Org.
Chem. 1969, 34, 1035, 1045 and 1053)

If there is an active neighboring group, i.e. esters, etc, the benzoxonium cation
generated by oxidative H-abstraction may rearrange to form another dioxonium cation,
which will lead to different regioselectivity of nucleophilic attack and therefore different
products (Scheme 2.3). Although this phenomenon was discovered [6-8] two decades ago,

it has not been developed and applied to the synthetic chemistry. Presumably, this is

68

because that it is difficult to design and control the rearrangement to prepare desired

molecular structure.

\ \rNu
O)\K\OCOR NBS fii Few/$0
_, _,
0 g0 R 820 O/QDR

Ph Ph

0

Scheme 2.3 The concept Of benzoxonium cation rearrangement

I I. A. IAA...
c r

2.2 Cascade dioxonium rearrangement of

benzylidene derivative under Hanessian-Hullar reaction conditions

In our research towards azasugar synthesis, D-glycero-D-gulo-heptono-1,4-
1actone 8 was used as a cheap starting material as a 7-carbon sugar lactone bearing 5
chiral centers. The 3,5-O-benzylidene-D-glycero-D-gulo-heptono-1,4-lactone 9 was
easily prepared in about 90% yield by treating the lactone with benzaldehyde and
concentrated hydrochloric acid. This 3.5-O-benzylidene glucoheptanolactone 9 was then
further protected by treatment with trimethylacetyl chloride in pyridine. The benzylidene
group of the fully protected 3,5-O-benzylidene-2,6,7-tri-O-pivaloyl D-glycero-D-gulo-
heptono-1,4-lactone 10 was then subjected to NBS-cleavage. Our primary goal was to
develop a way of introducing a bromo group at the 5-position of the carbon skeleton;
however, after separation by chromatography, the proton NMR spectrum of the main
product from the NBS oxidation of 10 indicated that the benzylidene ring had undergone
the unexpected rearrangement to form a benzoate ester ('H signals at 7.41-7.48 ppm (2H,
m), 7.56-7.62 ppm (1H, m) and 7.91—7.95 ppm (2H, m) and 13C signal at 164.70 ppm, but

it also indicated that the position of bromination was the primary position by showing two

69

dd signals at around 3.6 ppm. A single l3C signal at 28.8 ppm also showed up and
confirmed to be a methylene group by DEPT experiment. Based on these findings, the
product was tentatively assigned to be 11 (Scheme 2.4). This structure was further
confirmed by HMQC and gCOSY NMR experiments. HMQC was used to confirm the
site of bromination as the primary carbon by virtue of the connectivity of both proton
signals of the methylene group at around 3.6 ppm to the single l3C signal at 28.8 ppm.
Both proton and 13C NMR spectroscopy indicated that all three trimethylacetyl groups
were still present. Fast atom bombardment mass spectrometry also confirmed the formula

contained one benzoate, one bromo and three trimethylacetyl groups (m/z 626.17 and

 

 

628.17).
OH OH
HO . H O O H O .
HO“ PhCHO, HCI ‘ HO O PIvCI. Py
Ho" .6H 90% ' 3: s, -. 92%
8 Ph 09 OH
0in OPiv
H
PivO Q. 0 o NBS _ Br _ H O O
O s .9 C01,. reflux.75% P'VO .. .,
O 0in 820‘ ’OPiv

Piv = OCC(CH3)3

Scheme 2.4 The NBS treatment of 2,6,7—tri-O-trimethylacetyl-3,5(R)-O-benzylidene-D-
glycero-D-gulo-heptono- 1 ,4—1actone 10

The expected several pathways of NBS cleavage of 10 are illustrated in scheme
2.5. When the benzylidene acetal 10 was treated with NBS, after radical initiation, H-
abstraction and heterolytic cleavage of C-Br bond, a benzoxonium cation 12 was formed.
According to our expectation, the bromide ion would attack either the C-5 or O3 to

cleave the ring to form 5-bromo compound 13 or 3-bromo compound 14 respectively.

70

Due to the Ot-H acidity of lactones, the benzoxonium cation 10 or the 3-bromo compound
14 might undergo B-elimination to form an or, B-unsaturated lactone 15 as a by-product.
The latter elimination pathway was expected to dominate if the system was too sterically
crowded to allow bromination. Interestingly, less than 10% of the above compounds were
found in the product mixture. Instead, the major product 11 has a bromo group two
carbon atoms away from the intended site. This unexpected result was further analyzed

below.

  
 

13- bromination elimination

 

Scheme 2.5 Expected pathways for NBS cleavage of 3,5-(R)-O-benzylidem-2,6,7-
trimethylacetal-D-glycero-D-gulo—heptono-1,4-lactone 10

There are three possible rearrangement modes that could result in 7-bromination
of 10 (Scheme 2.6). The 7-pivaloyl group could have migrated to the 5-position inverting
that center (Pathway 1). Alternatively the 6-pivaloy1 group could have migrated to the 5
position and the 7-group could have migrated to the 6 position inverting both the 5 and 6
positions (Pathway 2). In the third scenario, the pivaloyl group in the C6 position could

participate in the phenyldioxolonium ring cleavage to form a new cylic dioxonium

71

species with inversion at C5. This species is then ring-opened by the C7 trimethylacetoxy
group with a second inversion (net retention) at C5 to form a 1.3-dioxonium species
between C5 and C7. This is then cleaved at the unhindered 7-position by bromide to give

the 7-bromo product (Pathway 3). In pathway 3. neither the 5 nor the 6 center is inverted.

O
O
in O \0
-, O
O
O \
Pathlvy/ [ Pathway 2

18
l i 1
1» >1, 1 490 °
0 ‘0 O \O
, 3X 38 0 9km OH E) 3i:
) 0 K O 0
Br 0 Br
Br 0
19 20 21
i i 1
A90 0 O 0 A90 0
O \O A)? o \O O ‘0
0H 63* m 6.1% 0H YE
Ph Ph " \n’
B O \n’ ‘1’ Br 0
r O=( O Br 04‘0 o O=( O
22 23 24

Scheme 2.6 Three possible pathways that could lead to 7-bromination of 10. The third
scenario is the only one that leads to net retention

72

It is relatively easy to differentiate between these three scenarios by a
consideration of the coupling constants of the H-4, H-5 and H-6 proton signals in the
NMR spectrum (Figure 2.1 and Figure 2.2). To facilitate this analysis, 2,3,5,6,7-penta—O-
trimethylacetyl-D-glycero-D—gulo-heptono-1,4-lactone 25 was synthesized from D-
glycero-D-gqu-heptono-l,4-lactone 8. Its reference spectrum was shown in Figure 2.2A.
The coupling constants observed reflect the conformation shown which is largely in
agreement with published studies on acyclic carbohydrate esters and aldonolactone side
chains [9-12] except that in this case, the terminal dihedral is a sickle conformation. In
the first scenario, if the C-5 position alone had been inverted, the H-4 / H-5 splitting and
the H-5 / H-6 splitting should be affected (Figure 2.2B). The change in the splitting
pattern for H-4 should be quite dramatic since it should go from a doublet of doublets (J =
~10 Hz) to a narrow triplet (J = ~3 Hz). The changes in splittings expected in the second
scenario where both the C-5 and C-6 positions are inverted are even more dramatic and
are shown in Figure 2.2C. In this case both H-4 and H-5 should be narrow triplets (J = ~3
Hz). The actual splittings that were observed are shown in Figure 2.2D. It is clear from
this that none of the coupling constants have changed indicating that scenario 3 was the

one that had taken place to give product 24.

73

.2: 25824“. _-o:oEo;-o~:M.D-EwobM-O-_>o_m>E-O-E-o.m£4380?55895-5-38590-m mo 83.58% mEZ I. 2: _.m Omswi

Art Ema

 

:3 Ema
86 Sam and and aim and Go.»

 

 

 

 

74

 

H H

4 14'5 9.7 Hz 5 J45 9-7 ”2
.13.4 2.7 Hz J5.6 2-7 HZ

" M

H4 J4‘5~3Hz H5 J5,6~9Hz

: J3’4~ 3Hz mJ4‘5~3Hz

 

Figure 2.2 Evaluation of coupling constants expected for products formed by the 3 possible
rearrangement mechanisms. (A) H-4 and H-5 signals from the reference spectrum of 2,3,56,7-
penta-O-trimethylacetyl-D-gl_vcero-D-gulo-heptono-l,4-lactone (25). (B) In the case of inversion
at C-5 (Scheme 2.6 pathway 1) only a narrow triplet (J = ~ 3 Hz) is expected for H4. (C) In the
case of inversion at both C-5 and C-6 (Scheme 2.6 pathway 2) the protons at these two positions
should yield narrow tripltes. (D) The H-4 and H-5 signals of the product indicating no change in
conformation compared to A (Scheme 2.6 pathway 3).

75

2.3 Selective protection, functionalization and chirality manipulation
of sugar lactones employing the NBS-initiated dioxonium cation

rearrangement

The cascade carboxonium cation rearrangement discovered in the reaction of NBS
with 3,5-O-benzylidene-2,6,7-tri-O-pivaloyl-D-glycero-D-gulo-heptono-1,4-lactone 10
showed highly steric-controlled bromination at the primary position. The presence of
active neighboring groups and the stereo constraints on the attack of incoming
nucleophiles facilitated this rearrangement. Although in this case, all the chiral centers
retained their stereo configurations, the mechanism analysis suggested great potential of
changing chiralities in other substrate molecules. Chirality manipulation is very important
in not only carbohydrate chemistry, but also general synthetic organic chemistry. To
explore this possibility, we applied this reaction to other simple, easily accessed and

important aldonolactones.

2.3.1 Selective 6-position bromination of protected D-gulono-y-lactone with

inversion at C-5

If the side chain of substrate used in the bromination reaction above were on
carbon less then bromination of the C6 position and inversion at the C5 carbon atom
should be expected based on our rationalization. This would be a very useful synthetic
stratage because the inverted center would be “righted” in any bimolecular displacement
reaction in which the original D-configuration was required. The model compound for
such a study can be visualized by simply shortening one carbon from the carbon chain of
our previous starting material D-glycero-D-gulo-heptono-1,4-lactone 8 while keeping

other chiralities unchanged. D-gulono-y-lactone 26 was the actual starting material used

76

to further explore the NBS-initiated rearrangement. 3,5-O-Benzylidene-2,6-di-O-
trimethylacetyl-D-gulono-y-lactone 28 was easily synthesized by treating D-gulono-y-
lactone with benzaldehyde under acidic condition according to the literature [13]
followed by pivaloation. The 3,5-O-benzylidene-2,6-di-O-trimethylacetyl-D-gulono-y-
lactone 28 was then treated with NBS in carbon tetrachloride under reflux to effect
oxidative cleavage of the benzylidene ring. NMR spectroscopy analysis of the major
product indicated that 6-position was brominated and the 6-O-pivaloyl group migrated to
the 5-position (Scheme 2.7). This structure was unambiguously assigned by 1H, 13C,
DEPT and gCOSY spectra. To assign the absolute configuration, a coupling constant
analysis was easily carried out. The H-H coupling constants of aldonolactone side chain
have long been used to determine the relative stereochemistry [9-12]. The coupling
constant between H4 and H5 of 2,3,5,6-tetra-O-pivaloyl-L-manno-1,4-lactone were
reported to be 9.8Hz [14]. The H-H coupling constant between H4 and H5 (J45 = 9.6 Hz)
(Figure 2.3) of the product clearly showed that 5-position was inverted to form a L-

manna-1,4-lactone derivative 29.

OH OPiv

 

OH
%=O PhCHO, HC1_ H 0 0 PivC|.Py 0%0
H026 DH Ph)\o OH 02 OP'

27
N .
Br t-Bu /h 2835;"
H O 0/4
0 85% e ) k(3|
PivO . ‘ ._ Bid 0 O

820‘ ’OPiv 2:
Ph

‘.
\‘

3O

bPiv

Scheme 2.7 Selective 6-position bromination of D-gulono-l ,4-lactone 26 with inversion at C-

77

.83 0:283».fl52:32-4-Spoon—€685-02?m.N->xoou-o-oEO£-o-_>ON=On-O-m Co 85.58% £22 F QC. m.m oSmE

 

:3 :33
0.0 0.0?
1 1 . 4 i . 1 fl
3: Sec
000.0 00—20 Obr.b OON.” ODNd 000.”
P hr b _ 1P I D F l” b P I h H D h b 1’ 1— br P F b b by b h r by P D h

 

 

1; . :

 

 

 

 

78

The mechanism of this reaction is believed to proceed through a benzoxonium
cation intermediate 30 with a spiral ring structure (Scheme 2.7). Using this method, we
made 3-O-benzoyl-6-bromo-6-deoxy-2,5-di-O-trimethylacetyl-L-manno-1,4-lactone 29 as
a differentially protected L-bromosugar lactone relatively easily. The selective inversion
of C-5 position in sugar lactone is always a challenge in the synthesis of L—sugars and
azasugars from D-sugars. Oxidation-reduction procedure and Mitsunobu reaction are
frequently used to invert the 5-position hydroxyl group. Both of them involve protecting
group manipulation. Stereoselectivity issue always accompanies the oxidation-reduction
procedure while Mitsunobu reaction generates large amount of phosphorus byproduct.
The strategy above used only easy protecting steps and Cheap reagents. It also furnished
bromo function and inverted S-position simultaneously. This strategy could be more

efficient in certain cases in synthetic chemistry.

2.3.2 Selective bromination and chirality manipulation of sugar lactone derivatives

with gluco- and galacto- configuration

The NBS bromination of 3,5-O-benzylidene-2,6-bis-O-trimethylacetyl-D-gulono-
y-lactone 28 illustrates the great potential of this methodology in Chirality manipulation of
carbohydrate structures. In the mechanism analysis, the primary neighboring ester group
attacked benzoxonium cation intermediate 30 to form a spiral ring transition state. The
bromide ion attacked the primary group to furnish the product. This stereospecific
rearrangement process may generate other important brominated sugar lactone if the
stereochemistry of the starting material is changed. Naturally, we want to extend it to
other sugar lactones such as D-glucono-O-lactone and D—galacto-y-lactone, which are

more often used and more important. The direct preparation of 3,5-O-benzylidene

79

derivatives of D-glucono-O-lactone and D-galacto-y-lactone using the previous strategy is

not possible, according to the analysis of ring structures (Scheme 2.8).

OH
OH

O U 0 O
HO Ring contraction 0
HO \O ....... > O

OH )‘05" OH

aldono-1 ,5-lactones Ph
OH OH
H O *1 0
’ O PhCHO, HCI O
HO _, O
HO OH “1% 0 OH
31 32
C3/C4 Threo- configuration cis-fused ring, low strain

   

. 33
03/04 Erythr- configuration trans-fused ring, high strain

Scheme 2.8 A ring structure analysis for 3,5-O-benzylidene aldonolactones preparation

D-glucono-O-lactone does not have a free 5-hydroxyl group. It has to go through a
ring-contraction process, which would not happen under normal benzylidenation
conditions. Even for sugar-y-lactones, only several aldonolactones are eligible for direct
3,5-O-benzylidenation. These include sugar lactones with D- and L-gulo, ido
configurations. The C3/C4 threo- configuration is necessary for the formation of a 5,6-
cis-fused ring system to avoid the high ring strain of the 5,6-trans-fused ring system. If
the aldonolactone has a C3/C4 erythro- configuration, it is not possible to prepare the 3,5-
O-benzylidene aldonolactone under normal benzylidenation conditions. To overcome this
obstacle, a simple strategy of inversion of the order of the protection steps was proposed.

Once the 2,6-positions of D-glucono-O-lactone or D-galacto-y-lactone is selectively

80

pivaloated, it should be possible to open the sugar lactone ring and form a 3,5-O-
benzylidene aldonic acid ester through a ring-switch process. These protected 3,5-0-
benzylidene aldonic acid ester could serve as good substrates for the NBS-initiated

pivaloyl rearrangement and bromination.

The selective protection and functionalization of free hydroxyl groups on
aldonolactones has long been explored in the literature [14, 15]. Although the selectivity
highly depends on the relative configuration of the aldonolactones and the protecting
reagents equivalents usage, the 2- and 6- positions generally show better reactivity. D-

glucono-O-lactone 35 was selectively protected at 2- and 6- positions to afford 2,6-di-O-

trimethylacetyl-D-glucono-O-lactone 36 by treatment with 2.2 equivalent tn'methylacetic
chloride in ~ 50% yield (Scheme 2.9). Although the yield was not extremely high due to
the competing reaction of other hydroxyl groups, the process was fairly practical because

the starting material was very cheap and only recrystallization was needed for separation.

OH
22 P Cl OP'V F>hCH(OCH,)2 IVOH
HO 9", 'V p-TsOH 011,01, OPiv
H0 \OJY—__,5°/° H00 ,. OCH,
80% 0

 

  

 

OH ”133—0 0
35
Bre H,Co
OPiv0P 01), 0
iv NBS CCI OPi , ~
PIVCI, Py OPiV Reflux 4 OE|V Plvo 082
—‘> _ t-Bu OCH ‘—’ .
90% o“ °CH3__+ 30% O s 3 P'VO'“
F0 0103/0 0 "'OPIV
38 39 40

Scheme 2.9 Selective 6-position bromination of protected D-glucono-d-lactone with inversion
at C-4 to form methyl 3-O-benzoyl-6-bromo-2,4,5-tri-O-trimethylacetyl-D-galactonate 40

81

The 2,6-di-O-trimethylacetyl-D-glucono-O—lactone 36 was then treated with
benzaldehyde dimethyl acetal under catalysis by p-toluenesulfonic acid in
dichloromethane to form the methyl 3,4-O-benzylidene-2,6-di-O-trimethylacetyl-D-
gluconate 37 as a mixture of two diastereoisomers in a 80% yield. This ring-switch
process was of interest because it opened the lactone ring and formed the acetal ring in
the same step, which greatly simplified the synthetic process. The methanol that was
generated in-situ by transacetalation was sufficient to open the lactone ring. This
suggested that afier 2,6-di-O-pivaloation, the lactone ring strain was substantially
increased because of the high steric hindrance of pivaloyl groups. Because of the trans-
relationship between the 3- and 4- free hydroxyl groups of 2,6-di-O-trimethylacetyl-D-
glucono-O-lactone, it was not possible to form a 3,5-O-benzylidene acetal between them
without opening the lactone ring. However, once the lactone ring was opened to give the
acyclic ester, a benzylidene acetal could form between any two of the three free hydroxyl
groups. Although 1,3-dioxane formation was commonly observed in the benzylidene
acetal protection of a polyol structure, in this case 1,3-dioxolane ring was formed to avoid
an axial carbon chain. The 3,4-O-benzylidene was preferred to the 4,5-O-benzylidene due
to the 3,4-threo and 4,5-erythro configuration. This mixture of two diastereoisomers 37
was then further protected by pivaloylation with the normal procedure to yield methyl
3,4-O-benzylidene-2,5,6-tri-O-trimethylacetyl-D-gluconate 38 also as a mixture of two
diastereoisomers, which was ready for NBS oxidative cleavage. In this stage, it was
unnecessary to worry about the newly formed chirality of the benzylidene carbon, which
made the product two isomers, because it would be eliminated after the NBS oxidation.

When methyl 3,4-O-benzylidene-2,5,6-tri-O-trimethylacetyl-D-gluconate 38 was treated

82

with NBS in the normal reaction condition, only one major product was obtained as
expected with ~ 80% yield. NMR spectroscopy analysis of the product indicated that the
6-position was brominated and the pivaloyl group had migrated. The location of the 3-0-
benzoyl group was assigned by 1H chemical shift and gCOSY. To determine the relative
stereochemistry, the H-H coupling constants of aldonate chain were used. The small-
large-small coupling constant set (J2,32.3Hz, J3,49.3Hz, J4,52.5Hz) matches and only
matches the acyclic sugar chain with a galacto- configuration [9-12]. The structure was
then assigned to methyl 3-O-benzoyl-6-bromo-6-deoxy-2,4,5-tri-O-trimethylacetyl-D-
galactonate 40. The reaction mechanism was believed to go through the benzoxonium
cation species 39, which rearranged to 4,6—(t-butyl-l,3-dioxonium) cation and
subsequently attacked by the bromide ion in the primary position to form the final product

40.

This strategy was then employed to D-galacto-y-lactone 41 (Scheme 2.10). After

selective pivaloation of 41, 2,6-di-O-trimethylacetyl-D-galacto-y-lactone 42 was prepared
in 65% yield. Compound 42 was then treated with benzaldehyde dimethyl acetal.
Although both 3— and 5-OH’s are free, unlike in the case of 2,6-di-O-trimethylacetyl-D-
glucono-O-lactone, they could not form a 3,5-O-benzylidene directly due to the 3,4-0-
erythro- configuration. Nevertheless, the ring-switching process occurred smoothly to
yield the methyl 4,5-O-benzylidene-2,6-di-O-trimethylacetyl-D-galaconate 43 as a
mixture of two isomers in 77% yield. This selectivity was also ascribed to its galacto-

configuration (3,4-erythro and 4,5-threo).

83

.83 Bacouom—mm-n-_boom35085-0-E-m.v.”303-9885-938590-m 733:. Co 8.58% 522 m. 2? ed 833,»

 

 

 

 

 

 

:3 :3...
o.»
— p p P p — p F L b —
4 111 1 t t 13 a j
:4: San
an.» 86

 

 

84

 

 

H 0 O , P‘VO H o PhCH(OCH3)2, 7“? °in
How P1vCl,Py. W0 stOH'CHZCb‘ OMOCH,
“0 H0 0” 65% H5 HO 6in 77% PivO OH O

43
41 42
Ph
, Ph O
. 7‘9 OPIV NBS,CCI4 319 OYt'B”
PivCI. Py Oj/‘YKKOCH3 Reflux [ O «(to ]—>
———> 5
0
90% PivO FWD 0 66‘ Piviji/vgfo
OCH
44 45a 3
(..3 1-BU
O
O \-< 4 Br OPiv
Piv __, 7
m0 PiVOMOCHa
PivO OCH, OBz OPIVO
45b 46

Scheme 2.10 Selective 4-position bromination of protected D-galactono—y-lactone with
inversion at C-2 to form methyl 5-O-benzoyl-4-bromo-4-deoxy-2,3,6-tri-O-trimethylacetyl-D-
galactonate 46

The following steps were similar with the case of methyl 3,5-O-benzylidene-2,6-
di-O-trimethylacetyl-D-gluconate 37. After further pivaloation and NBS oxidative
cleavage, we expected the bromination of 6-position with inversion at 5-position.
Unexpectedly, NMR analysis of the major product showed that 4-position was
brominated and the 5-O-position was benzoylated. A normal Hanessian-Hullar reaction
without neighboring group participation could generate this structure with a gluco-
configuration. The coupling constant analysis (J 2.31.8Hz, J 3,410.5Hz, J4.51.5Hz), however,
supported a galacto- sugar chain. The product was therefore assigned to methyl 5-O-
benzoyl-4-bromo-4-deoxy-2,3,6-tri-O-trimethylacetyl-D-galactonate 46. The mechanism
was believed to go through a benzoxonium cation 45a. The 3-pivaloyl group attacked at
the 4-position to form a t-butyl-l,3-dioxonium cation 45b, which was subsequently

attacked at the 4-position by bromide ion to give compound 46. The 4-position was

85

double-inverted so that the galacto- configuration was retained. This reaction pattern was
very different from the one with D-gluco- configuration, although in both cases, pivaloyl
group rearrangement occurred. For methyl 3,4-O-benzylidene-2,5,6-tri-O-trimethylacetyl-
D-gluconate 38, the primary bromination drove the 4,6-pivaloyl rearrangement. For
methyl 4,5-O-benzylidene-2,3,6-di-O-trimethylacetyl-D-galactonate 44, the 3-O-pivaloyl
group attacked the benzoxonium cation preferentially, after the 3,4-(t-butyl-l,3-
dioxonium) cation formation, the bromide ion attacked the 4-position to give the final
product 46. This outcome showed the remarkable stereochemical sensitivity of this NBS-

initiated dioxonium cation rearrangement.

In summary, by inverting the protection steps, we carried out this NBS-initiated
primary bromination of D-glucono-S-lactone derivative while inverting the 4-position. In
the case of D-galactono-y-lactone, 4-position was brominated and the stereochemistry
was retained. Two new selectively protected and fimctionalized sugar derivatives were
generated. Based on the simple and elegant synthesis, this methodology showed its

general applicability and practicability in synthetic carbohydrate chemistry.

86

.33 03:28390-}on350856-56.m.mixoocééEoSé-.xouconéfi 352: do 83.58% MZZ :. oi Wm 2&5

Art EGG
0.—. 0.N 0d 0.? 0.0 0.0 0. h 0.0

P—b-lp-—bLPh—#phrbhbb-r—Vthb—nbbP—P-ppb'hlh-h

 

 

 

.. 411.1141 aj

 

Ar: Enn— Avb Ema
Sn.‘ 00" 00nd and 0006

bepFF-bnh—bbhbb lPPPPbePbe—*bnp—hPrthlr

 

 

 

87

2.4 C; symmetric tartrate derivative desymmetrization using NBS-

initiated dioxonium rearrangement

Since the NBS-initiated benzoxonium cation rearrangement generally introduced
selective bromination and neighboring position chirality manipulation for aldonolactones,
we tried to generalize this methodology to other nonspecified compounds. 2,3-0-
benzylidene-1,4-di-O-trimethylacetyl-L-threitol was selected as the target molecule based
on the following reasons: First, it is readily available from commercial dimethyl or
diethyl L-tartrate; second, it is a C2 symmetric molecule, the proposed NBS oxidative
cleavage could be employed as a very good desymmetrization methodology; third, the

product could serve as a protected chiral synthon with appropriate size and fitnctionality.

 

 

 

 

VOH
tEtPhCHO,p-TSOH .. 0383+“ EtOH 5
H65 Toluene. reflux : G‘Y :0 O
80% over Y
PYh 4a Zsteps Ph 49
O
PNOVOPN Pivoqr t'B"
M 5: O NBS, CCI4, reflux§ 6‘ ‘0 O —>
93% Y 70% \6
Ph 50 Ph 51a
0
\' “f’t'B” PivOV—Br 0
"MO , , HCI, methanol_ i 7
—-> c 9, r
PM) 632 H5 0H
5‘“) 52

53

Scheme 2.] l Selective primary bromination of a protected L-threitol derivative with
inversion of the neighboring group

Diethyl L-tartrate 47 was treated with benzaldehyde and catalytic p-

toluenesulfonic acid in refluxing toluene. Co-distillation was used to drive the reaction

88

forward. The crude diethyl 2,3—O-benzylidene-L-tartrate 48 was then reduced by sodium
borohydride (NaBH4) in cold ethanol to afford 2,3-O-benzylidene-L-threitol 49 in 80%
yield over two steps. 2,3-O-Benzylidene-L-threitol 49 was treated with trimethylacetyl
chloride to give 2,3-O-benzylidene-l,4-di-O-trimethylacetyl-L-threitol 50 in 93% yield.
Compound 50 was then subjected to NBS cleavage of benzylidene acetal in refluxing
carbon tetrachloride. NMR spectroscopic analysis of the product indicated that the 2-0-
benzoyl-4-bromo-4-deoxy-1,3-di-O-trimethylacetyl-erythritol 52 was synthesized
(Scheme 2.11). The erythro- configuration of 52 was confirmed by the H-H coupling
constant analysis between H2 and H3 (J = 7.1 Hz). The mechanism was concluded to

through a dioxonium cation intermediate 51a.

Ready for functional group transformation, this optically active differentially
protected bromo-butanetriol 52 could be used as a 4—chiral-center synthon. Usually, the
desymmetrization of C2 symmetric compounds such as L-threitol often only involves
monofunctionalization. However, this monofimctionalization normally would not change
any stereochemistry, therefore it would not change the L-threo configuration. On the
other hand, the asymmetric mesa-erythritol desymmetriztion must involve an asymmetric
reagent, normally asymmetric catalyst, to generate an optically active product with
erythro- configuration. The NBS-initiated rearrangement bromination of 50 generated
optically pure product 52 with erythro- configuration from easily synthesized L-threitol
derivative. Being complimentary to the monofunctionalization of L-threitol, this method
enabled us to access the whole set of 4-carbon chiral synthons with both threo- and

erythro- configuration.

89

Acidic hydrolytic deprotection of this NBS oxidation product 52 was carried out
to deprotect the ester groups. This did not yield the desired bromobutanetriol. Instead, it
led to the 1,4-anhydroerythritol 53. The formation of this compound further confirmed the
inversion of the 3-position in the NBS-initiated rearrangement and bromination and
therefore further confirmed the rearrangement mechanism.

2.5 The Exploration of NBS-initiated Neighboring Rearrangement of

Other Protecting Group

HOVOH BZOVOBZ
é BzCI, Py. _ é

 

o o o 0
Y 88% Y
Ph 49 Ph 54
NBS, CCI4, refluxi
Br' _
" 820 y 082
o s
5- ,, AL 0 (o
320 Br W6
56 _ Ph 55 .

 

 

Scheme 2.12 The direct bromination on the benzoxonium cation in the NBS treatment of 2,3-
O-benzylidene-l ,4-di-O-benzoyl-L-threitol 54

To explore the possibility of other neighboring protecting group participating in
the NBS-induced rearrangement, 2,3-O-benzylidene-l,4-di-O-benzoyl-L-threitol 54 was
synthesized by treatment of 2,3-O-benzylidene-L-threitol with benzoyl chloride in
pyridine. Then it was reacted with NBS in refluxing carbon tetrachloride (Scheme 2.12).
NMR analysis of the major product proved it to be 3-bromo-3-deoxy-l ,2,4-tri-O-benzoyl-
L-erythritol 56. The reaction was believed to go through a normal Hanessian-Hullar

reaction pathway, in which the bromide ion attacked the benzoxonium cation 55 directly

9O

before the neighboring group rearrangement. Although benzoyl groups have been found
to undergo rearrangement in some cases [6, 7], failure to rearrange in this case suggested
that the steric environment generated by the neighboring groups was very important. If
the steric hindrance were not high enough to block the direct attack of the benzoxonium

cation by the bromo nucleophile, the rearrangement would not happen efficiently.
2.6 Conclusion

Unexpectedly, NBS oxidative cleavage of 3,5-O-benzylidene-2,6,7-tri-O-
trimethylacetyl-D-glycero-D-gulo-heptono-1,4-lactone 10 gave a 7-bromo product 24.
The structure of the product was elucidated and the mechanism of this unusual cascade
rearrangement was analyzed in detail. The concept of neighboring ester group
rearrangement was employed on several other aldonolactones to generate a series of
selectively protected and brominated D- and L-aldonolactone derivates. 3,5-0-
benzylidene-2,6-di-O-trimethylacetyl-D-gulo-1,4-lactone 28 was selectively 6-brominated
and 5-inverted to form a L-manno-1,4-lactone derivative 29. Methyl 3,4-O-benzylidene-
2,5,6-tri-O-trimethylacetyl-D-gluconate 38 and methyl 4,5-O-benzylidene-2,3,6-tri-O-
trimethylacetyl-D-galactonate 44 were selectively brominated at 6- and 4-positions,
respectively. This methodology was further applied to L-threitol derivative
desymmetrization and successfully generated orthogonally protected and functionalized,
optically pure chiral synthon 52 with erythro- configuration. Replacement of the
trimethylacetyl group with benzoyl group suppressed the neighboring group
rearrangement, further confirmed that this attracting rearrangement was largely based on

steric-control.

91

EXPERIMENTAL

General Optical activity data were obtained on a JASCO P-1010 polarimeter at
25°C. NMR spectra were obtained on a Varian VXR-SOO Spectrometer operating at
500MHz for protons. Mass spectra were obtained on a JEOL HX-l lO-HF instrument
using fast atom bombardment as ionization mode. Spectra were recorded in the positive
ion mode. IR spectra were obtained on a Nicolet 710 spectrometer in chloroform solution

except when otherwise specified.
3 ,5-O-Benzylidene-D-glycero-D-gulo-heptono-1 ,4-lactone (9):

D-glycero-D-gulo-heptono-1,4-lactone 10 g was suspended in 40 mL
benzaldehyde while 4 mL concentrated hydrochloric acid was added in slowly. The
mixture was stirred at room temperature for 2 hours and became a viscous solution. The
solution was poured into 400 mL hexanes and stirred. The precipitates were collected by
filtration and washed with 3 x 100 mL hexanes. The white powder was then dried in
vacuum to give 3,5-O-benzylidene-D-glycero-D-gulo-heptono-1,4-lactone. Yield 12.9 g,
90%. The white solid could be recrystallized from ethanol to give 3,5-O-benzylidene-D-
glycero-D-gulo-heptono-l,4-lactone 9 as white needles. M.p. 187-190°C; (lit. [16] mp.
188-191°C; [0t]D20 - 56.10 (c, 1.0 in MeOH)); SH (500 MHz; CD30D): 3.68 (1H, dd, J7,7~
11.8, 16¢ 4.7 Hz, H-7’), 3.78 (1H, dd, J63 2.5 Hz, H-7), 3.91 (1H, ddd, 15,6 9.3 Hz, H-6),
4.11 (1H, dd, J4; 1.8 Hz, H-S), 4.58 (1H, t, J33 1.9 Hz, H-4), 4.72 (1H, d, .123 4.0 Hz, H-
2), 4.76 (1H, dd, H-3), 5.69 (1H, s, CflPh), 7.47-7.50 (2H, m, ArH), 7.32-7~36 (3H, m,
ArH); 8c(125MHz; CD30D): 63.2 (07), 69.7, 70.6, 72.1. 75.8, 76.0 (C-2, C-3, C-4; C-
5, C-6), 99.8 (d, QHPh), 127.1, 128.7, 129.6 (Ar_C_H), 140.9 (s, ArQ), 178.3 (s, C=O); FT-

IR (KBr): 3450, 3300, 1770 cm".

92

3,5-O-Benzylidene-2,6,7-tri-O-pivaloyl-D-glycer0-D-gulo-heptono-1 ,4-1actone (10):

3,5—O-benzylidene-D-glucoheptonolactone (3.0 g, 10.1 mmol) was dissolved in 30
mL pyridine. The solution was cooled to 0°C and trimethylacetyl chloride (6.2 mL, 50.3
mmol) was added in drop by drop. The mixture was protected from moisture and stirred
for 24 hours. After that, the mixture was dumped into ice-saturated sodium bicarbonate
solution and then extracted with dichloromethane twice. The oil layer was collected and
all solvents were removed by vacuum. The raw product was purified by flash column
chromatography using chloroform as the elutant. The colorless solid afier removing the
solvent affords 3,5-O-Benzylidene-2,6,7-tri-O-pivaloyl-D-glycero-D-gqu-heptono-1,4-
lactone in 92% yield; M.p. 173-176 °C; FAB-MS: m/z calcd. for C29H40010 548.2622,
found 549.2699 (MH+); 8H (500MHz; CDCl3): 1.18, 1.19, 1.23 (3 x 9H, 3 x s, 3 x
C(Cfi3)3), 4.29 (1H, dd, 17.7, 12.6, 1614.1 Hz, H—7’), 4.31 (1H, dd, 145 2.0, 15,6 8.5 Hz, H-
5), 4.34 (1H dd, J14 2.2, H-4), 4.57 (1H, dd, 16.7 2.4 Hz, H-7), 4.95 (1H, dd, J23 4.2, J14
2.2 Hz, H-3), 5.30 (1H, ddd, H-6), 5.42 (1H, d, H-2), 5.55 (1H, s, CflPh), 7.32-7.40 (5H,
m, Arfl); 8c (125MHz; CDC13): 26.93, 27.13 (OCC(_(;H3)3), 38.77, 38.87 (OCQ(CH3)3),
61.22 (C-7), 68.96, 70.50, 72.64, 73.52 (C-2, C-3, C-4, C-5, C-6), 98.86 (QHPh), 125.81,
128.19, 129.28, 136.10 (ArQ), 169.97(O=_Q in lactone), 176.36, 177.55, 177.71

(ogcggmb); FT-IR: 2975, 1809, 1738 cm" .

3-O-Benzoyl-7-bromo-7-deoxy-2,5,6-tri-O-pivaloyl-D-glycero-D-gulo-heptono-1 ,4-

lactone (l 1):

3,5-O-Benzylidene-2,6,7-tri-O-pivaloyl-D-glycer0-D-gqu-heptono- 1 ,4-lactone (3

g, 5.47 mmol) was dissolved in 50 mL carbon tetrachloride at 60°C and N-

93

bromosuccinimide (1.5 g, 8.43 mmol ) was added. The mixture was refluxed at 80°C oil
bath for 12 hours. TLC showed that no starting material remained. The mixture was
filtered and washed with saturated sodium bicarbonate solution. The raw product was
purified by flash column chromatography and the major fraction affords 3-O-benzoyl-7-
bromo-7-deoxy-2,5,6-tri-O-pivaloyl-D-glycero-D-gulo-heptono-1,4-lactone in 75% yield.
M.p. 138-140 °C; FAB-MS (MH+): m/z calcd. for C29H39Br010 626.1727, found
627.1799 (MH+); 8H(500MHz; CDCl3): 0.99, 1.08, 1.24 (27H, 3 x s, 3 x C(Cfl3)3), 3.58
(1H, dd, by 11.1, J6], 9.0 Hz, H-7’), 3.64 (1H, dd, 16,7 4.5 Hz, H-7), 4.94 (1H dd, J14 3.1,
14,5 9.5 Hz, H-4), 5.51 (1H, dd, 15,6 2.5 Hz, H-S), 5.54 (1H, ddd, H-6), 5.78 (1H, (1, H23 4.9
Hz, H-2), 5.92 (1H, dd, H-3), 7.41-7.48 (2H, m, ArH), 7.56-7.62 (1H, m, Arfl), 7.91-7.95
(2H, m, Ar_I_I_); 6c (125MHz; CDCl3): 26.61, 26.64, 26.83, 27.00 ( OCC(QH3)3), 28.77 (C-
7), 38.57, 38.65, 38.81 (OCQ(CH3)3), 67.48, 68.12, 69.34, 71.27, 74.44 (C-2, 3, 4, 5, 6),
127.92, 128.54, 129.67, 133.88 (ArQ), 164.70 (OQPh), 168.89 (O=Q in lactone), 176.36,

177.55, l77.71(OQC(CH3)3); FT-IR: 2975, 1818, 1743 cm".
2,3 ,5,6,7-Penta-O-t1imethylacetyl-D-glycer0-D-gulo-heptono-1 ,4-lactone (25):

D-glycero-D-gqu-heptono-1,4-lactone (4 g, 19.2 mmol) was dissolved in 100 mL
pyridine and cooled down to 0°C in ice-water bath. Trimethylacetyl chloride (30 mL,
243.6 mmol) was added in slowly and the mixture was stirred at room temperature for 24
hours. The mixture was poured into ice-saturated sodium bicarbonate water solution and
then extracted by dichloromethane. The organic layer was collected and all the solvents
were evaporated. The resulted colorless oil affords 2,3,5,6,7-penta-O-trimethylacetyl- D-
glycero-D-gulo-heptono-l,4-lactone in 91% yield. 6” (500MHz; CDC13): 1.13, 1.17, 1.21

(3x9H, s, 3 x C(cgjh), 1.23 (18H, 3, 2 x C(CH3)3), 4.03 (1H, dd, J6... 7.2, 37,7. 12.0 Hz,

94

H-7’), 4.36 (1H, dd, 16,7 4.4 Hz, H-7), 4.71 (1H, dd, J3,4 2.7, .145 9.7 Hz, H-4), 5.08 (1H,
m, J55 2.7 Hz, H-6), 5.64 (1H, dd, H-5), 5.72 (1H, (1, 123 4.5 Hz, H-2), 5.82 (1H, dd, H-3);
6c (125MHz; CDC13): 26.68, 26.81, 26.88, 26.92, 26.97 (OCC(QH3)3), 38.53, 38.63,
38.72, 38.88, 39.16 (OCQ(CH3)3), 61.73 (C-7), 68.46, 68.70, 68.97, 75.96 (C-2, 3, 4, 5,
6), 168.46 (O=_C_ in lactone), 176.15, 176.23, 176.40, 176.46, 177.91(O_(;C(CH3)3); FT-

IR: 2976, 1817, 1741 cm".
3,5-O-Benzylidene-D-gulono-y-lactone (27):

Anhydrous hydrogen chloride was bubbled through 40 mL benzaldehyde for two
minutes and D-gulono-1,4-lactone 10 g was added to the solution. The reaction mixture
was stirred for about two hours, at which time it became very thick and stirring was
stopped. Afier standing overnight, the reaction mixture was triturated with ether and
filtered. The solid was washed thoroughly with ether three times, with water three times,
and then with ether an additional two times. After drying this afforded 11.2 g (74%) of
3,5-O-benzylidene-L-gulono-l,4-lactone. Recrystallization from absolute ethanol
afforded pure material, M.p. 187-189 °C; (lit. [13] mp 188-189°C; [01023 + 61.1° (in
DMF)); 25" (500MHz; DMSO-d6) 6H 3.43 (b, t, 2H, H-6), 4.0-4.83 (m, 4H, H-2, 3, 4, 5),
4.97 (m, 1H, -CH20H), 5.68 (s, 1, CflPh), 5.97 (m, 1H, -CHOH), 7.4 (m, 5H, Arfl); 8C
(125MHz; DMSO-d6): 175.9 (s, C=O), 137.7 (s, Ar_C_), 129.0, 128.1, and 126.5 (s,
Ar_C_H), 98.2 (d, 1 QHPh), 76.2, 74.8, 70.8, 69.5 (C-2,3,4,5), 59.9 (C-6); FT-IR: 3468,

3275, 1786 cm“.
3,5-O-Benzylidene-2,6-di-O-trimethylacetyl-D-gulono-y-lactone (28):

3,5-O-benzylidene-D-gulono-y—lactone (4.0 g, 15.0 mmol) was dissolved in 40

mL dry pyridine and cooled down to 0°C. Trimethylacetyl chloride (4.62 mL, 37.5

95

mmol) was then added and the reaction mixture was warmed up to room temperature.
After being stirred for 6 hours, the solution was poured into 300 mL saturated sodium
bicarbonate solution containing ice and then extracted with dichloromethane (3 x 200
mL). The organic layer was concentrated under vacuum and trace amounts of pyridine
was removed by toluene co-distillation. The colorless oil (6.2 g) was 3,5-O-benzylidene-
2,6-O-trimethylacetyl-D-gulono-y-lactone (Yield: 95%). This material was suitable for
use without further purification. [a]D= -87.3 (c = 0.9, CHC13); FAB-MS: m/z calcd. for
C23H3003 434.1941, found 435.2017 (MH+); 5H (500MHz; CDCl3): 1.12,].14 (2 x 9H, 2
x s, 2 x C(CH3)3), 4.57-4.45 (4H, m, H-6’, H-6, H—5, H-3), 4.88 (1H, dd, J45 1.9, J3,4
4.0Hz, H-4), 5.43 (1H, (1, J23 4.2Hz, H-2), 5.60 (1H, s, CHPh), 7.35-7.46 (5H, m, ArH);
8C (125MHz; CDC13): 26.91, 27.04 (C(QH3)3), 38.72, 38.83 (OCQ(CH3)3), 62.04 (C-6)
69.62, 70.48, 72.57, 73.44 (C-2, C-3, C-4, C-S), 98.62 (_CHPh), 125.87, 128.12, 129.15,
136.40 (Ar(_3_), 170.25 (O=Q of lactone), 177.46, 177.97 (OQC(CH3)3); FT-IR: 2973,

1805, 1736, 1282, 1146 cm".
3-O-Benzoyl-6-bromo-6-deoxy-2,S-di-O-trimethylacetyl-L-manno-1 ,4-lactone (29):

3,5-O-benzylidene-2,6-di-O-trimethylacetyl-D-gulon0-y-lactone (5.0 g, 11.5
mmol) was dissolved in 100 mL carbon tetrachloride to which N-bromosuccinimide (3.70
g, 20.8 mmol) was added. The mixture was refluxed for 2 hours and then filtered. The
filtrate was diluted in 200 mL chloroform and washed with saturated sodium bicarbonate
solution. The crude product was purified by flash column chromatography to afford 3-O-
benzoyl-6-bromo-6-deoxy-2,5-di-O-trimethylacetyl-L-manno-y-lactone 9 in 85% yield.
[a]D = +13.7° (c = 0.8, CHCl3); FAB-MS: m/z calcd. for C23H29BrOg 512.1046, found

513.1116 (MH+); 8H (500MHz; CDCl3): 1.05, 1.07 (2 x 9H, 2 x s, 2 x C(CH3)3), 3.74

96

(1H, dd, .1636 11.9, .1615 2.8Hz, H-6’), 3.85 (1H, dd, 16.5 3.0Hz, H-6), 5.04 (1H, dd, J45
9.6Hz, H-4), 5.29 (1H, ddd, J5,4 9.6Hz, H-S), 5.87 (1H, dd, 123 5.0, H-2), 6.01 (1H, dd, H-
3), 7.43, 7.57, 7.92 (5H, m, Ar_H); 8c (125MHz; CDC13): 26.78 (C(QH3)3), 32.03 (C-6),
38.78, 38.83 (OC_C_(CH3)3), 66.76, 68.03, 69.04, 75.71 (C-2, C-3, C-4, C-S), 128.19,
128.65, 129.71, 133.91 (ArQ), 164.71 (O=QPh), 169.06 (O=_C_ of lactone), 176.26, 176.39

(O_C_C(CH3)3); FT-IR: 2973, 1817, 1739, 1264, 1133 cm".
2,6-Di-O-trimethylacetyl-D-glucono-5-lactone (36):

D-glucono-S-lactone 10 g was dissolved in 100 mL dry pyridine and cooled down
to 0 °C in ice bath. Trimethylacetyl chloride 15 g was added in the solution dropwise. The
solution was then stirred at 5 °C for 24 hours, after which it was poured into 500 mL
saturated sodium bicarbonate with ice and stirred vigorously. The white precipitates were
collected by filtration and washed by distilled water several times. After drying in
vacuum, the white solid was recrystallized in hexanes / ethyl acetate (2:1) twice to give
2,6-di-O-trimethylacetyl-D-gluc0no-6-lactone 9.7g (Yield: 50%). M.p. 152-154°C; [ab =
+68.6° (c = 1.0 CH3OH); 6H (500MHz; CDC13): 0.98, 0.99 (2 x 9H, 2 x s, C(Cfl3)3), 3.97
(dt, 1H, J45 6.0Hz, 15,6 3.8Hz, 15,6» 6.0Hz, H-5), 4.01 (dd, 1H, 16,5 11.4Hz, H-6’), 4.11 (dd,
1H, H-6), 4.28 (dd, 1H, 13,4 5.9Hz, H-4), 4.32 (dd, 1H, 12,3 4.6Hz, H-3), 5.12 (d, 1H, H-2);
8c (125MHz; CDCI3): 177.731, 176.444 (OQC(CH3)3), 170.567 (O=Q of lactone),
79.153, 73.672, 70.941, 67.599, 64.793 (C-2, 3, 4, 5, 6), 38.239, 38.197 (OCC_(CH3)3),

26.612, 26.429 (C(QH3)3); FT-IR (CH3OH): 3505, 1778, 1719, 1701, 1159 cm".
Methyl 3,4-O-benzylidene-2,6-di-O-trimethylacetyl-D-gluconate (37):

2,6-di-O-trimethylacetyl-D-glucono-S—lactone 5.0 g was suspended in 50 mL

dichloromethane and 5.0 g benzaldehyde dimethyl acetal was added in slowly. After that,

97

p-toluenesulfonic acid monohydrate 1.0 g was added in and the mixture was stirred for
about 10 minutes, after which the mixture became a clear solution. The solution was
stirred at room temperature overnight and then poured into sodium bicarbonate solution
with ice. The organic layer was collected and the water phase was washed with 2 x 100
mL dichloromethane. The organic phase was dried and evaporated. After column
chromatography, 5.02 g methyl 3,4-O-benzylidene-2,6-di-O-trimethylacetyl-D-gluconate
(80%) was obtained as a mixture of two isomers. For the 3,4-(R)-O-benzylidene isomer,
8H (500MHz; CDCl3): 1.19, 1.31 (2 x 9H, 2 x s, C(Cfl3)3), 2.89 (1H, b, -OH), 3.76 (3H, s,
COOCH3), 3.93 (1H, ddd, J45 8.1Hz, 15,6 3.2Hz, .155 5.4Hz, H-5), 4.07 (1H, dd, J34 5.4Hz,
H-4), 4.19 (dd, 1H, J65 12.0Hz, H-6’), 4.38 (1H, dd, H-6), 4.85 (1H, dd, 12,3 2.5Hz, H-3),
5.27 (d, 1H, H-2), 6.05 (1H, s, CflPh), 7.35 (3H, m, ArCfl), 7.43 (2H, m, ArCH); 5c
(125MHz; CDCl3): 179.168, 177.442 (OQC(CH3)3), 167.651 (O=_Q of lactone), 136.45,
129.644, 128.329, 126.753 (ArQ), 105.401 (_C_HPh), 78.458, 76.939, 73.171, 71.479,
65.591 (C-2, 3, 4, 5, 6), 52.568 (COOQH3), 38.876, 38.790 (OCQ(CH3)3), 27.075, 27.053
(C(QH3)3); For the 3,4-(S)-O-benzylidene isomer, 8H (500MHz; CDC13): 1.19, 1.23 (2 x
9H, 2 x s, C(CH3)3), 2.89 (1H, b, -OH), 3.74 (3H, s, COOCH3), 4.00 (2H, m, H-4, 5),
4.17 (dd, 1H, 16,6. 12.0Hz, H-6’), 4.42 (1H, dd, H-6), 4.81 (1H, dd, 12,3 2.5Hz, H-3), 5.33
(d, 1H, H-2), 5.92 (1H, s, CflPh), 7.35 (3H, m, ArCfl), 7.43 (2H, m, Arcw; 8c
(125MHz; CDC13): 179.168, 177.442 (OQC(CH3)3), 167.651 (O=Q of lactone), 136.45,
129.644, 128.329, 126.753 (ArQ), 105.401 (QHPh), 78.458, 76.939, 73.171, 71.479,
65.591 (C-2, 3, 4, 5, 6), 52.568 (COOQH3), 38.876, 38.790 (OCQ(CH3)3), 27.075, 27.053

(C(QH3)3); FT-IR : 3518, 2979, 1768, 1740, 1284, 1147 cm".

98

Methyl 3,4-O-benzylidene-2,5,6-tri-O-trimethylacetyl-D-gluconate (38):

Methyl 3,4-O-benzylidene-2,6-di-O-tn'methylacetyl-D-gluconate 37 was reacted
with 2 equivalent trimethylacetyl chloride in pyridine using the normal pivaloylation
procedure. 3,4-O-Benzylidene-2,5,6-tri-O-trimethylacetyl-D-gluconate was obtained also
as a mixture of two isomers in 90% yield. For the 3,4-(R)-O-benzylidene isomer, 6H
(500MHz; CDCl3): 1.14, 1.16, 1.29 (3 x 9H, 3 x s, C(Cfl3)3), 3.71 (3H, s, COOCfl3),
4.10 (dd, 1H, J 5.6 4.6Hz, J65 12.4Hz, H-6), 4.29 (dd, 1H, 13,4 5.3Hz, 14,5 7.8Hz, H-4), 4.55
(dd, 1H, 15,6 2.8Hz, H-6), 4.66 (dd, 1H, 12,3 2.3Hz, H-3), 5.14 (ddd, 1H, H-5), 5.19 (d, 1H,
H-2), 6.06 (1H, s, CflPh), 7.32 (3H, m, ArCH), 7.39 (2H, m, ArCI_I_); 6c (125MHz;
CDC]3)I 177.459, 177.155, 176.775 (OQC(CH3)3), 167.053 (O=Q of lactone), 135.955,
128.158, 126.499, 129.575 (ArQ), 105.754 (QHPh), 78.079, 75.252, 72.578, 71.655,
61.852 (C-2, 3, 4, 5, 6), 52.468 (COOQHg), 38.668, 38.632, 38.587 (OCQ(CH3)3),

26.895, 26.872, 26.827 (C(QH3)3); FT-IR: 2999, 2881, 1734, 1461, 1398, 1323 cm".
Methyl 3-O-benzoyl-6-bromo-6-deoxy-2,4,5-tri-O-trimethylacetyl-D-galactonate (40):

3,4-O-Benzylidene-2,5,6-tri-O-trimethylacetyl-D-gluconate (2 g, 3.8 mmol) and
N-bromosuccinimide (1.24 g, 6.9 mmol ) was suspended in 20 mL carbon tetrachloride.
The mixture was refluxed at 80°C oil bath for 12 hours and then filtered. TLC showed no
starting material remained. The filtrate was purified by flash column chromatography and
the major fraction affords methyl 3-O-benzoyl-6-bromo-6-deoxy-2,4,5-tri-O-
trimethylacetyl-D-galactonate 1.84 g in 80% yield. [a]D = -26.6° (c = 1.0, CHCl3); FAB-
MS: m/z calcd. for C29H41BrOlo 628.1883, found 629.1965 (MH+); 6H (500MHz; CDCl3):
1.18, 1.19, 1.25 (3 x 9H, 3 x s, C(Cfl3)3), 3.42 (dd, 1H, 15.689.3Hz, J6,6‘10.9Hz, H-6’), 3.64

(s, 3H, COOCI_I3), 3.74 (dd, 1H, 15,63.7Hz, H-6), 5.19 (d, 1H, J2.32.3Hz, H-2), 5.25 (ddd,

99

1H, 14.52.5Hz, H-5), 5.58 (dd, 1H, 13.49.3112, H-4), 5.88 (dd, 1H, H-3), 7.45 (t, 2H, J =
7.8Hz, ArCfl), 7.58 (m, 1H, ArCfl), 8.02 (dd, 2H, J = 1.2Hz, J = 8.3112, ArC_I-_I); 5c
(125MHz; CDC13): 177.120, 177.003, 175.812 (OQC(CH3)3), 167.253 (O=Q of lactone),
164.875 (O=QPh), 133.826, 129.943, 129.715, 128.685 (ArQ), 71.839, 69.782, 69.568,
68.618 (02, 3, 4, 5), 52.664 (COOCfl3), 38.913, 38.854 (OCQ(CH3)3), 28.671 (C-6)

27.008, 26.902, 26.980 (C(_C_H3)3); FT-IR (CHCl3): 2979, 1741, 1265, 1139 cm".
2,6-Di-O-trimethylacetyl-D-galacto-y—1actone (42):

D-galacto-l,4-lactone (5 g, 28 mmol) was dissolved in 50 mL dry pyridine and
cooled down to 0°C. Trimethylacetyl chloride (7.5 g, 62.2 mmol) was added to the
solution slowly. The mixture was then stirred at 5 °C for 24 hours, after which the
mixture was poured into 300 mL saturated sodium bicarbonate solution with ice and
stirred. The white precipitates were collected and washed by distilled water several times.
After drying, the crude product was recrystallized in hexanes / ethyl acetate (2: 1) to give
2,6-di-O-trimethylacetyl-D-galacto-y-lactone 6.3 g (65%). M.p. 146-148°C; [a]D = -lO.7°
(c = 1.0, CHC13); 6H (500MHz; CDCl3): 1.01, 1.06 (2 x 9H, 2 x s, C(Cfl3)3), 3.84 (dt, 1H,
J45 2.4Hz, 15,6 6.4Hz, J55 6.4Hz, H-S), 3.92 (dd, 1H, J65 11.1Hz, H-6’), 4.03 (dd, 1H, J14
8.1Hz, H-4), 4.07 (dd, 1H, H-6), 4.40 (dd, 1H, 12.3 8.8Hz, H-3), 5.38 (d, 1H, H-2); 6c
(125MHz; CDC13): 177.673, 176.842 (OQC(CH3)3), 169.415 (O=Q of lactone), 80.399,
74.204, 70.292, 65.885, 63.874 (C-2, 3, 4, 5, 6), 38.249 (OCQ(CH3)3), 26.696, 26.581

(C(_C_H3)3); FT—IR: 3791, 3096, 3077, 1797, 1733, 1624, 1146 cm".

100

Methyl 4,5-O-benzylidene-2,6-di-O-trimethylacetyl-D-galaconate (43):

2,6-di-O-trimethy1acetyl-D-galactono-y-lactone 5 g was suspended in 50 mL
dichloromethane and 5 g benzaldehyde dimethyl acetal was added in slowly. After that, p-
toluenesulfonic acid monohydrate 1 g was added in and the mixture was stirred for about
10 minutes, after which the mixture became a clear solution. The solution was stirred at
room temperature overnight and then poured into sodium bicarbonate solution with ice.
The organic layer was collected and the water phase was washed by 2 x 100 mL
dichloromethane. The organic phase was dried and evaporated. After column
chromatography, 4.86 g methyl 4,5-O-benzy1idene-2,6-di-O-trimethylacetyl-D-gluconate
(77%) was obtained as a mixture of two isomers. For the 4,5-(.S')-O-benzylidene isomer,
5” (500MHz; CDC13): 1.22, 1.26 (2 x 9H, 2 x s, C(Cfl3)3), 2.86 (1H, b, -OH), 3.72 (3H, s,
COOC_H_3), 3.99 (dd, 1H, J51,- 5.4Hz, J65 14.2Hz, H-6’), 4.13 (1H, dd, J23 2.Hz, J14 8.9Hz,
H-3), 4.31 (1H, dd, J45 5.0Hz, H-4), 4.37 (1H, dd, J52, 4.7Hz, H-6), 4.47 (1H, ddd, H-S),
5.30 (d, 1H, H-2), 5.96 (1H, s, CfiPh), 7.35 (5H, m, ArCfl); 6c (125MHz; CDCl3):
178.471, 177.023 (O_C_C(CH3)3), 168.634 (O=Q of lactone), 136.780, 129.424, 128.290,
126.369 (ArQ), 104.230 (QHPh), 78.275, 72.672, 76.272, 72.034, 63.820 (C-2, 3, 4, 5, 6),
52.471 (COO_C_H3), 38.854, 38.744 (OCQ(CH3)3), 26.897, 26.874 (C(_C_H3)3); FT-IR:

3494, 2976, 1736, 1282, 1147 cm".
Methyl 4,5-O-benzylidene-2,3,6-tri-O-trimethylacetyl-D-galaconate (44):

Methyl 4,5-O-benzylidene-2,6-di-O-trimethylacetyl-D-galactonate was reacted
with trimethylacetyl chloride in pyridine using the normal pivaloylation procedure. 4,5-O-
Benzylidene-2,3,6-tri-O-trimethy1acetyl-D-galactonate was obtained also as a mixture of

two isomers in 90% yield. For the 4,5-(R)-O-benzylidene isomer, 5H (500MHz; CDC13):

101

1.17, 1.21, 1.22 (3 x 9H, 3 x s, C(CH3)3), 3.66 (3H, s, coocgj), 4.18 (dd, 1H, 15,, 6.8Hz,
16,6. 11.8112, H-6’), 4.37 (dd, 1H,15,6 4.8Hz, H—6), 5.25 (d, 1H, 12,3 1.6Hz, H-2), 5.29 (ddd,
1H, 1.. 8.8Hz, H-S), 5.37 (dd, 1H, 13,. 7.5112, 11.4), 5.40 (dd, 1H, H-3), 6.00 (1H, s,
CflPh), 7.36 (5H, m, ArCfl); 5C (125MHz; (30013): 177.104, 176.887, 176.719
(OQC(CH3)3), 167.605 (o=_(; of lactone), 136.501, 129.606, 128.404, 126.495 (mg),
104.565 (QHPh), 77.544, 74.797, 71.809, 70.419, 64.054 (02, 3, 4, 5, 6), 52.466
(coogh), 38.973, 38.791, 38.776 (ocgcnm), 26.995, 26.975, 26.809 (C(QH3)3);

FT-IR22977,l811,1741,1278, 1131 cm".
Methyl 5-O-benzoyl-4-bromo-4-deoxy-2,3,6-tri-O-trimethy1acetyl-D-galactonate (46):

4,5-O-benzylidene-2,3,6-tri-O-trimethylacetyl-D-galactonate (1.35 g, 2.6 mmol)
and N-bromosuccinimide (0.84 g, 4.7 mmol ) was suspended in 15 mL carbon
tetrachloride. The mixture was refluxed at 80°C oil bath for 12 hours and then filtered.
TLC showed that no starting material remained. The filtrate was purified by flash column
chromatography and the major fraction affords methyl 3-O-benzoyl-6-bromo-6-deoxy-
2,4,5-tri-O-t1imethylacetyl-D-galactonate 1.02 g (66%). [ab = -15.7° (c = 1.0 CHC13);
FAB-MS: m/z calcd. for C29H41BrOlo 628.1883, found 629.1964 (MH+); 5H (500MHz;
CDC13): 1.12, 1.14, 1.25 (3 x 9H, 3 x s, C(CH3)3), 3.56 (3H, s, COOCfl3), 4.26 (dd, 1H,
J55 6.9Hz, J65 11.5Hz, H-6’), 4.30 (dd, 1H, J5_6 6.7Hz, H-6), 4.37 (dd, 1H, J45 1.5Hz, J14
10.5Hz, H-4), 5.40 (dt, 1H, H-S), 5.54 (d, 1H, J23 1.8Hz, H-2), 5.57 (dd, 1H, H-3), 7.40
(2H, m, ArCfl), 7.52 (1H, m, ArCfl), 7.97 (2H, m, ArCfl); 5c (125MHz; CDC13):
177.408, 176.815, 175.662 (OQC(CH3)3), 167.040 (O=Q of lactone), 164.694 (_COPh),
133.347, 129.721, 128.378, 128.538 (ArQ), 104.565 (QHPh), 70.941, 69.384, 66.801 (C-

2, 3, 5), 64.010 (C-6), 52.332 (COOQH3), 48.733 (C-4), 38.812, 38.757, 38.574

102

(OCQ(CH3)3), 26.916, 26.837, 26.783 (C(QH3)3); FT-IR: 2980, 1743, 1481, 1264, 1121

cm".

Diethyl 2,3-O-benzylidene-L-tartrate (48):

Diethyl L-tartrate (13.8 g, 67 mmol) was mixed with benzaldehyde (14.2 g, 134
mmol) in 100 mL toluene and 2 g p-toluenesulfonic acid monohydrate was added in. The
mixture was co-distillated under reflux to remove the water. More toluene (2 x 50 mL)
was added and evaporated until the solution became clear. The solution was washed with
saturated sodium bicarbonate solution, dried and evaporated to give a sticky liquid (19.2

g). This crude product was used in the next step without further purification.
2,3-O-Benzylidene-L-threitol (49):

Diethyl 2,3-O-benzylidene-L-tartrate (19.0 g, crude) was dissolved in 200 mL
ethanol and cooled down to 0°C, then 10 g sodium borohydride was added in several parts
in 30 minutes. After that, the reaction mixture was further stirred at 0°C for 30 minutes,
after which it was evaporated to dryness. The crude product was then purified by flash
column chromatography to give 2,3-O-benzylidene-L-threitol 11.27 g (80% over 2 steps)
as a white solid. M.p. 71-73 °C; (lit. [17] [(1]D20 + 7.4° (c 1.0, CHCl;)); 8” (500MHz;
CDC13): 2.4 (2H, b, -OH’s) 4.22 (6H, m, H-1, 2, 3, 4) 5.8 (CflPh) 7.34 (m, 5H, ArH); 6c
(125MHz; CDCl3): 135.986, 129.892, 128.223, 127.867 (ArQ) 104.439 (QHPh) 75.482,

75.268 (C-2, 3) 68.345, 68.235 (C-1, 4)
2,3-O-Benzy1idene-1,4-di-O-trimethylacetyl-L-threitol (50):

This compound was prepared from 2,3-O-benzylidene-L-threitol in 93% yield

following the same procedure of the preparation of 3,5-O-benzylidene-2,6-di-O-

103

trimethylacetyl-D-gulono-y-lactone. [ab = 4.00 (c = 0.7, CHC13); FAB-MS: m/z calcd.
for C21H3006 378.2042, found 379.2132 (MH+); 5H (500MHz; CDC13): 1.20, 1.22 (2 x
9H, 2 x S, 2 x C(CH3)3), 4.22-4.32 (6H, m, H-1, H-2, H-3, H-4), 5.97 (1H, s, CflPh),
7.36-7.46 (5H, m, ArH); 5c (125MHz; CDCI3): 27.09, 27.14 (C(QH3)3), 38.76, 38.78
(OC_Q(CH3)3), 63.45, 63.62 (C-1, C-4), 76.12, 76.76 (C-2, C-3), 104.19 (QHPh), 126.56,
128.31, 129.51, 136.89 (ArQ), 178.03, 178.07 (OQC(CH3)3); FT-IR: 2973, 1731, 1282,

1154 cm".
2-O-Benzoyl-4-bromo-4-deoxy-1,3-di-O-trimethylacetyl-erythritol (52):

This compound was prepared from 2,3-O-benzylidene-l,4-O-trimethylacetyl-L-
threitol in 70% yield following the same procedure used for the preparation of 3-0-
benzoyl-6-bromo-6-deoxy-2,5-di-O-trimethylacetyl-L-manno-1,4-lactone. [01]D = +6.7° (c
= 0.8, CHCI3); FAB-MS: m/z calcd. for C2IH29BrO6 456.1148, found 457.1212 (MH+);
8H (500MHz; CDC13): 1.17, 1.22, 1.23 (2 x 9H, 2 x s, 2 x C(CH3)3), 3.57 (1H, dd, J|~2 5.8,
J13; 11.4Hz, H-l’), 3.69 (1H, dd, J12 3.7Hz, H—l), 4.30 (1H, dd, J453 5.5, J.;-,4 12.2Hz, H-
4’), 4.45 (1H, dd, J43 3.1Hz, H-4), 5.36 (1H, ddd, J23 7.1Hz, H-2), 5.61 (l-H, ddd, H-3);
8c (125MHz; CDC13): 27.00, 27.01 (C(QH3)3), 30.60 (C-l), 38.77, 38.94 (OC_(_3(CH3)3),
61.65 (C-4), 69.43, 70.79 (C-2, C-3), 128.51, 129.68, 129.71, 133.49 (ArQH), 164.99

(O=QPh), 176.89, 177.78 (OQC(CH3)3); FT-IR: 2971, 1733, 1265, 1142 cm".
2,3-O-Benzylidene-1,4-di-O-benzoyl-L-threitol (54):

2,3-O-benzylidene-L-threitol (2 g, 9.5 mmol) was dissolved in 20 mL dry pyridine
and cooled down to 0 °C. Benzoyl chloride (4 g, 28.5 mmol) was added in slowly. After
the addition, the reaction mixture was stirred for 5 hours at 0°C, after which it was poured

into saturated sodium bicarbonate solution with ice and stirred. Dichloromethane (2 x 50

104

mL) was used for extraction and the organic layer was evaporated to give 2,3-O-
benzylidene-l,4-di-O-benzoyl-L-threitol (3.5 g, 88%) as a sticky oil. [ab = +23.7o (c =
1.0, CHC13); 6H (500MHz; CDC13): 4.54-4.66 (6H, m, H-1, H-2, H-3, H-4), 6.11 (1H, s,
CHPh), 7.36-8.04 (15H, m, ArH); 6c (125MHz; CDCl3)Z 166.088, 166.041 (OQPh),
133.199, 133.140, 130.416, 129.521, 134.392, 129.617, 129.603, 128.373, 128.317,
128.295, 126.586 (ArC), 104.328 (_C_fHPh), 76.82, 76.237 (C-2, C-3), 64.076, 64.033 (C-1,

C-4); FT-IR: 1788, 1722, 1271, 1093 cm".
3-Bromo-3-deoxy-l ,2,4-tri-O-benzoyl-L-erythritol (56):

2,3-O-benzy1idene-l,4-di-O-benzoyl-L-threitol (1.12 g, 2.68 mmol) and N-
bromosuccinimide (NBS) (0.86 g, 4.82 mmol) was suspended in 10 mL carbon
tetrachloride. The mixture was refluxed in an 80°C oil bath for 12 hours and then filtered.
TLC showed that no starting material remained. The filtrate was purified by flash column
chromatography and the major fraction affords 3-bromo-3-deoxy-1,2,4-tri-O-benzoyl-L-
erythritol (1.01 g, 76% ) as a sticky oil. [a]D = + 52.00 (c = 1.0, CHC13); 6H (500MHz;
CDCl3): 4.65-4.93 (5H, m, H-1, H-3, H-4), 5.88 (ddd, 1H, J = 3.6Hz, J = 5.9Hz, J =
9.5Hz, H-2), 7.38-8.15 (15H, m, Arfl); 6c (125MHz; CDCl3): 165.804, 165.668 (O_C_Ph),
165.064, 162.227, 134.425, 133.474, 133.274, 133.187, 130.438, 129.747, 129.674,
129.577, 128.764, 128.432, 128.380, 128.358 (ArC), 71.404 (C-2), 64.698, 63.804 (C-1,

4), 46.997 (C-3); FT-IR: 1788, 1725, 1265, 1108 cm".

105

[1]
[21
[3]
[41
[5]
[6]

[7]
[3]
[9]

REFERENCE

Hanessian, S. Carbohydr. Res. 1966, 2, 86.

Hanessian, S.; Pleassas, N. R. J. Org. Chem. 1969, 34, 1035, 1045 and 1053.
Hullar T. L.; Siskin S. B. J. Org. Chem. 1970, 35, 225.

Bundel, D. R.; Josephson, S. Can. J. Chem. 1978, 56, 2686.

Binkley, R. W.; Goewey, G. S.; Johnston, J. C. J. Org. Chem. 1984, 49, 992.

Hoffmeyer, L.; Jacobsen, S.; M015, 0.; Pedersen, C. Acta. Chem. Scand. B 1979, 33,
175.

Jacobsen, S.; Mols, O. Acta. Chem. Scand. B 1981, 35, 163 and 169.
Jacobsen, S. Acta. Chem. Scand. B 1984, 38, 157.

Horton, D.; Wander, J. D. Carbohydr. Res. 1969, 10, 279.

[10] Blanc-Muesser, M.; Defaye, J .; Horton, D. Carbohydr. Res. 1980, 87, 71.

[11] Walaszek, Z.; Horton, D. Carbohydr. Res. 1982, 105, 131.

[12] Angyal, S. J.; Le Fr, R.; Gagnaire, D. Carbohydr. Res. 1972, 23, 121.

[13] Crawford, T. C. US. patent 4,111,958, 1978.

[14] Gallo, C.; Jeroncic, L. O.; Varela 0.; de Lederkremer, R. M. J. Carbohy. Chem.

1993, 12, 841.

[15] Lundt, I.; Madsen, R. Synthesis, 1992, 1129.

[16] Richard, C. J. F.; Bruce, 1.; Hughes, D. J.; Girdhar, A.; Fleet, G. W. J.; Watkin, D. J.

Tetrahedron: Asymmetry 1993, 4, 1579.

[17] Wenger, R. M. Helv. Chim. Acta. 1983, 66, 2308.

106

Chapter 3

Exploration of Inter- and Intramolecular Competitive Nucleophiles
in the N-Bromosuccinimide-Mediated Dioxolonium Ion

Rearrangement for Introduction of New F unctionalities

ABSTRACT

To introduce new functionalities in to the sugar structure, different inter- and
intramolecular nucleophiles were employed in the N-bromosuccinimide-mediated
dioxolonium ion rearrangement. Intermolecular nucleophiles generally failed to compete
with in-situ generated bromide ion, although in certain case, the introduction of
competitive nucleophile was shown. Cyano groups do not participate in the Hanessian-
Hullar reaction as intramolecular nucleophile to give expected lactam. The results were
ascribed to the better nucleophilicity of the bromide ion generated in-situ than that of the

competitive nucleophiles introduced inter- or intramolecularly.
3.1 Introduction

In the last chapter, the rearrangement of neighboring ester groups when treating
benzylidene acetals with NBS was discussed. The reaction starts from the abstraction of
the benzylidene methine hydrogen atom by a bromine radical followed by bromination
with molecular bromine. After heterolytic cleavage of C-Br bond, a benzoxonium cation
is formed. The active neighboring group then attacks the partially positively charged

carbon initiating the rearrangement. From another point of view, the carbonyl O atom acts

107

as an intramolecular competitive nucleophile with the in-situ generated bromide ion. In
the case of the NBS treatment of 3,5-O-benzylidene-2,6,7-tri-O-trimethylacetyl-D-
glucoheptano-l,4-lactone, due to the high steric hindrance, the attack of bromide ion was
slow enough to let the complex rearrangement process occur. As a result, a primary
brominated compound was synthesized. This interesting phenomenon encouraged us to
further explore this NBS-cleavage of benzylidene group with some other nucleophiles.
Theoretically, it should be possible to add another nucleophile which could compete with
bromide ion at some stage of the NBS-cleavage of the benzylidene group. It would then
be possible to introduce some new functionality under mild reaction conditions (Figure

3.1). This is always an important process pursued by synthetic chemists.

«WM
0) 0 Br
/
nnn\nnnm<nnn MAW
H NBS, Nu: V Br \Ff
(OK, Nu \

O O 0 : <
Th if \ f—IE Nu
Ph

Figure 3.1 The competitive nucleophile participated Hanessian-Hullar reaction

3.2 The base sensitivity of NBS-oxidative cleavage of protected

benzylidene sugar lactone

3,5-O-Benzylidene-2,6,7-tri-O-trimethylacetyl-D-glucoheptano- l ,4-lactone 1
could be easily synthesized from commercially available D-glucoheptano-1,4-lactone by

two simple steps. The NBS treatment of this compound without acid scavenger induced a

108

pivaloyl rearrangement resulting in bromination of the terminal 7-position. Before we
introduced competitive nucleophiles into the reaction system, we thought it beneficial to
further explore this reaction by varying reaction conditions. This was especially true since

most of the competitive nucleophiles would be introduced as a form of base.
3.2.1 The effect of barium carbonate on the NBS reaction as an acid scavenger

From the beginning of the discovery of the Hanessian-Hullar reaction in 1960’s,
acid scavengers have been generally used to control the acidity of the reaction system [1,
2]. This was especially true from a viewpoint of industrial large scale synthesis. Barium
carbonate was very often used as a good acid scavenger because of its inert property and
easy removal. It generally gave good results. We therefore evaluated the effect of barium

carbonate in our reaction system as an acid scavenger.

As a control reaction, 3,5-O-benzylidene-2,6,7-t1i-O-trimethylacetyl-D-
glucoheptano-l,4-lactone 1 was treated with 1.8 equivalent NBS in refluxing carbon
tetrachloride without any acid scavenger. The reaction mechanism is shown in Scheme
3.1 as discussed in Chapter 2. To further understand the reaction, we carried out an easy
kinetic analysis. After two hours, we analyzed the reaction crude mixture by NMR
spectroscopy. NMR analysis showed that most of the starting materials was consumed
and transformed to a new compound. The characteristic 99 ppm signal in '3 C spectrum of
the starting benzylidene lactone was gone while a new signal at 104 ppm emerged. This
new peak could not be seen in DEPT experiment, suggesting a quaternary carbon was
formed. We believed that this new compound corresponded to the intermediate 3,5-O-
bromobenzylidene-2,6,7-t1i-O-trimethylacetyl-D-glucoheptano-1,4-1actone 3, in which

the benzylidene H atom was abstracted and substituted by a bromine atom. The NMR

109

analysis of the reaction crude mixture after 4 hours showed a complicated mixture with
weakened peak strength at 104 ppm in the '3 C spectrum, indicating that the unstable C-Br
bond was being broken and pivaloyl rearrangement started. After 6 hours, the reaction
crude mixture NMR showed that none of this intermediate was left, while lH spectrum
clearly showed the final 7-bromo product 2 emerging. After 10 hours, the reaction was

completed and a fairly pure product was prepared in ~ 80% yield.

OPiv 0in
NBS, 0014 GP" H
PM) \,~ reflux ——>PiVO/_-LHKOJ=O PM) ‘5‘ O O
oLHO s —> o
x 0‘“ ZOPiv 0“. ’7 £8 %Piv
Ph‘ 1 ms)? Br OPiv 4
Piv = 0CC(CH3)3 3 l
o
OPiv 99000 9 o
H o 80% 0 0 )Kph
Br \‘ O ‘— ' 00 Ph <— \H0 O
. \\ ’
PIvO ‘4‘ a C a O 0:3" O (0 OR
820 ©in 3' 7&0 o
2 s 5

Scheme 3.] The reaction and mechanism of NBS-initiated pivaloyl reanangement and
bromination

We then carried out this reaction with barium carbonate as the acid scavenger. All
the other reaction conditions were kept the same. A simple kinetic analysis similar to the
one described above was carried out to understand the effect of this acid scavenger
(Scheme 3.2). After 2 hours, NMR analysis of the crude reaction mixture showed a
similar result, in which the bromobenzylidene 3 was shown to be the dominant
component. However, the NMR analysis after 4 hours and 6 hours showed a very

different result. While the 104 ppm peak in l3C spectrum kept decreasing, a new peak at

110

112 ppm emerged. A DEPT experiment also showed that this was a quaternary carbon.
This carbon was not seen in the control reaction. After 8 hours, the compound with this
carbon became the dominant product. However, this compound was unstable towards
further workup by water or chromatography. After these treatments, only 5-O-Benzoyl-
2,6,7-tri-O-t1imethylacetyl-D-glucoheptano-1,4-lactone 7 was obtained. Based on the
understanding of the Hanessian-Hullar reaction, we reasoned that this unknown and
unstable compound to be another isomer of 3,5-O-bromobenzylidene-2,6,7-tri-O-

trimethylacetyl-D-glucoheptano- l ,4-lactone 7, which underwent hydrolysis to form 8.

 

OPiv .
OP” 8, 0in
H O
CCI4 reflux . . = O OPiv
\‘L OPiv .~ “ ” - - 0
Ph\ 1 Ph\\ 0 OPIV PNO/
Piv=OCC(CH3)3 5' 3 9
OPiv OPiv Ph OPiv
. H
Pivowo H20 PM) 0,.- O o Br/IVO
o" ’ Q. .9 ———-— 9 OPiv
IOPiV 1311)::0 (3in PIVO/ O
8 Br 7 0

Scheme 3.2 The isomerization of bromobenzylidene lactone when BaCO3 is present in the
NBS reaction

Presumably, in the NBS treatment of benzylidene lactone with barium carbonate
as the acid scavenger, the pivaloyl rearrangement was slowed down. The cleavage of the
C-Br bond on the bromobenzylidene group was a reversible process. While the C-Br
bond cleavage scrambled this chiral center, the reversed C—Br bond formation would re-
install this chiral center with an isomerization effect. In this substrate, the chiral center

inversion gave a more stable intermediate 7, in which the bromo sits in the equatorial

lll

position of the benzylidene ring. This isomer 7 showed a fairly different chemical shift in
'3 C spectrum at 112 ppm. This bromobenzylidene compound was labile towards
hydrolysis to form the 5-O-Benzoyl-2,6,7-tri-O-trimethylacetyl-D-glucoheptano-l,4-
lactone 8. In the control reaction, it was possible that the following pivaloyl

rearrangement was sufficiently fast so that this inversion could not be detected.

After extension of the reaction time to 18 hours, 7—bromo-7-deoxy—3-O-Benzoyl-
2,5,6-t1i-O-t1imethylacetyl-D-glucoheptano-1,4-lactone 2 was also synthesized as the
major product in a slightly lower yield compared with the control reaction. However, the
reaction rate difference generated by acid scavenger still needed to be addressed.
Compared with the control reaction, the reaction with acid scavenger was much slower.
We postulated that the in-situ generated small amount of HBr was a catalyst for the
rearrangement and bromination. A reaction mechanism accelerated by HBr is shown in
Scheme 3.3. In the NBS bromination of benzylidene protected sugar, small amount of
HBr was generated in the reaction system. This also supplied a low concentration of
bromide ion, which could undergo nucleophilic attack at the primary position of suitable
carboxonium cation intermediate to furnish the final product 2. Since the C-Br bond
heterocleavage generated a bromide ion, HBr was regenerated in this cycle. This
mechanism requests a necessary level of bromide ion to make the rearrangement and
bromination proceed smoothly. When acid scavenger such as barium carbonate was used,
the trace amount HBr was scavenged and separated from the reaction media. The
concentration of the bromide ion in the solution was dramatically decreased. Both the

pivaloyl rearrangement and the terminal bromination became much slower. As a result,

112

the isomerization by C-Br bond cleavage and re-forrnation could be seen clearly in the

system.
OPiv ()in OPiV
. H
PivO \s H O O PivO \- H O O FWD/MO
9 . . —>._ 90‘ —‘-— 9 . .
\fOf I’OPiv Br )‘§ ' - '90“ ’2)in
Ph‘ BEN Ph 9 4 PW ph Br

3 . \ 7

H O H I
Br \. . 0 O
Pivo‘ O ‘— 5 C33 \gph
320 ’OPIV Br ‘x—
2 6

Scheme 3.3 The acceleration effect of bromide ion in the NBS reaction

Although barium carbonate has been widely used in the NBS oxidative cleavage,
we have never seen any similar isomerization of the bromobenzylidene carbon center in
the literature before. This was ascribed to several reasons. First, in most research work,
the rearrangement of neighboring group was not involved. Therefore the bromination
process was relatively fast. Once the C-Br bond was cleaved, the bromide would attack a
suitable position to finish the reaction. Second, the high local steric hindrance also slowed
down the bromination process, making the isomerization possible. The identification of
this isomerization process gave us more insight into the reaction mechanism of the NBS
oxidative cleavage of the benzylidene group. Therefore control of the reaction to our

desired direction could be possible.

113

3.2.2 Pyridine assisted elimination in NBS treatment of 3,5-O-benzylidene-2,6,7-

tri-O-trimethylacetyl-D-glucoheptono-l,4-lactone

As discussed above, when barium carbonate was employed as the acid scavenger,
the NBS promoted pivaloyl rearrangement and primary bromination was slowed down.
This was ascribed to its ability of scavenging HBr and separating the bromide ion from
the reaction media. If some organic base such as pyridine were used, the produced
bromide salt might still have some solubility in the solvent and therefore keep the
concentration of the bromide ion in the solution. Therefore it could act as a good acid
scavenger without interupting the bromination reaction. Hence 3,5-O-benzylidene-2,6,7-
tri-O-trimethylacetyl-D-glucoheptano-1,4-lactone 1 was treated with NBS in refluxing
carbon tetrachloride in the presence of 2 equivalents of pyridine. After about 4 hours, the
reaction was checked by TLC and all the starting material was consumed. '3 C NMR
analysis of the crude reaction mixture showed no peaks at 104 and 112 ppm, suggesting
neither bromobenzylidene compound was present. Interestingly, only 4 major peaks were
shown in the 60~80 ppm region, which is characteristic for carbon atoms substituted by a
single oxygen function. Combined with the lH NMR spectrum, it was clear that
elimination had occurred to yield an unsaturated lactone. The mechanism is shown in
Scheme 3.4. After benzylidene H abstraction, C-Br bond formation and heterocleavage of
the C-Br bond, a 3,5-O-benzoxonium cation 4 was formed. This cation served as a good
leaving group at B-position in a lactone. Due to the a-H acidity, it was labile towards the

attack by some bases such as pyridine. The B-elimination happened smoothly to yield the

114

final (141-unsaturated sugar lactone 9. This protected a,B-unsaturated sugar lactone 9

could be a versatile building block for advanced organic synthesis.

 

Scheme 3.4 Pyridine assisted elimination in NBS treatment of protected benzylidene lactone

This elimination was expected to be a possible side product in our first trial on the
NBS treatment of glucoheptanolactone 1 (Chapter 2). However, only when pyridine was
introduced into the reaction system, was it observed. The reaction mode of NBS with
glucoheptanolactone was highly dependent on the basicity of the reaction system. When
the reaction was carried out without base, the reaction at acidic condition showed
predominantly pivaloyl rearrangement and primary bromination. When barium carbonate
was used as an acid scavenger, the reaction was kept at neutral condition. Under neutral
condition, both the pivaloyl rearrangement and the bromination were greatly slowed
down and an isomerization of the bromobenzylidene group was detected. When pyridine
was used as an acid scavenger, the reaction mixture became relatively basic because of
the good solubility of pyridine in the solvent. The basicity, although weak, was enough to
drive the B-elimination to yield an a,B-unsaturated sugar lactone 9. By simply varying the
reaction basicity with different bases, the reaction was easily tuned towards different
directions. Different useful structures were generated and they could be employed as good

building blocks for different usage.

115

3.3 The introduction of competitive nucleophiles into the NBS
bromination of 3,5-O-benzylidene-2,6,7-tri-O-pivaloyl-D-glucoheptono-

1,4-lactone

The mechanism of the NBS bromination of benzylidene compounds included
several steps: C-H abstraction, C-Br bond formation, C-Br heterocleavage to form a
benzoxonium cation and a bromide ion, (dioxonium cation rearrangement) and
nucleophilic attack of the carbon chain by the bromide ion. By introducing another
competitive nucleophile into the system, it could be possible to interfere with the last
nucleophilic step (Figure 3.1). A new functionality could be introduced into the carbon
chain directly without further transformation of bromosugar, which could be difficult in
many cases. The appropriate competitive nucleophiles could range from alcohols,

carboxylic acids, cyanides, nitrites and more.
3.3.1 The introduction of O- nucleophiles into the NBS oxidation

The simplest intermolecular competitive nucleophile would be water. The
Hanessian-Hullar reaction in the presence of water has been well studied [3]. Generally,
after the C-Br cleavage, the benzoxonium cation would be trapped by water molecular to
form unstable orthoacid. The orthoacid species would undergo proton transfer to form a
hydroxyl group and a benzoyl group (Scheme 3.5 A). Stereoelectronic effects usually
direct the regioselectivity of this hydrolysis, favoring of an equatorial hydroxyl group [4].
The protonation and breakdown of the orthoacid intermediate is the critical step. This
reaction will proceed more quickly if one of the oxygen lone pair is anti-periplanar to the

C-0 bond to be cleaved. For C-Oeq bond cleavage, one of the 0a,, lone pairs is already

116

nearly anti-periplanar to the C-Oeq bond under normal steric arrangement. However, for
the 00,, bond cleavage, this needs to proceed via a strained steric arrangement. The C-
Ocq bond cleavage is normally favored, which gives a product with an equatorial hydroxyl
group (Scheme 3.5 B). This process has been generally used as a deprotection
methodology of benzylidene protection group. It was still of importance to understand the

regioselectivity in a five-membered ring system.

(0
fig = HKEA $319553
/
R 0 R [069
«433‘— .4 (0&3
3)

HOQ

R Q00 R1 @131
95802 5133

Scheme 3.5 The hydrolysis of carboxonium cation: A) The hydrolysis mechinism; B) The
regioselectivity of hydrolysis based on the stereoelectronic effect

3,5-O-Benzylidene-2,6,7-tri-O-trimethylacetyl-D-glucoheptono-1,4-lactone 1 was
treated with NBS in refluxing carbon tetrachloride with water present. After several hours,
TLC showed a fairly clean reaction with a dominant product. NMR analysis of this
product showed its structure to be 5-O-benzoyl-2,6,7-tri-O-trimethylacetyl-D-
glucoheptano-l,4-1actone 7 with a free hydroxyl group at 3-position (Scheme 3.6). Only

trace amount of 5-OH was generated. In our system, it would be hard to predict the

117

regioselectivity based on the stereoelectronic effect discussed in Scheme 3.5 B. This was
because of the different fused ring system. In this ring system, both C-O bonds of the
orthoacid had an anti-periplanar O lone pair to facilitate the bond cleavage. The
regioselectivity here would largely be a result of thermodynamic control since the
breakage of C-05 would generate a 3-O-benzoyl group, which would be sterically
hindered by the 2-O-pivaloyl group on the same five-membered lactone ring.
Nevertheless, the remarkable regioselectivity was very interesting because it opened an

opportunity to selectively deprotect the benzylidene group ofthis system.

0"” OPiv
H O
PWO \\ NBS, H20 PivO ‘- H O
0L .~ . CCI4 reflux 01)
\‘ (f // )\ ‘c 3
Ph‘ OP” Ph 0‘ OPiv
1 4
OPiv OH 1 OPlV
Q
PiVO H O 720/ Ph
° C)
802 OPiv
Ho“ %HVP1VO/ O
O
8 10

Scheme 3.6 The NBS treatment of benzylidenelactone with water present

We continued our research to introduce other 0- nucleophiles into the reaction
system. Several alcohols were screened towards this goal. To suppress the bromination
reaction, excess alcohols were used. However, when primary and secondary alcohols
were used, the NBS cleavage of benzylidene group was shut down. Only starting material
was recovered. Although the NBS-cleavage of benzylidene acetal with free hydroxyl

groups could proceed smoothly with fair yields, the excess alcohol in this situation

118

suppressed this reaction. The major reason was ascribed to the polarity of the alcohols as
solvents, which could greatly lower the bromo radical concentration and shut down the
H-abstraction step. On the other hand, it was also possible that oxidizable alcohols could
react with NBS when the reaction time was extended. To test this, t-butyl alcohol was
employed as the competitive nucleophile. NMR study of the reaction showed that the
starting material was consumed smoothly, suggesting that NBS was still able to oxidize
the benzylidene C-H bond with t-butyl alcohol present. Interestingly, afier workup and
separation, the only recovered product was 5-O-benzoyl-2,6,7-tri-O-trimethylacetyI-D-
glucoheptano-l,4-lactone 8, indicating a hydrolysis cleavage. A mechanism was proposed
in Scheme 3.7. When the benzoxonium cation was formed, it was trapped by t-butyl
alcohol to form an orthoester. The workup and separation hydrolyzed the orthoester to

form the same product when using water as the competitive nucleophile.

- H
MW) e O 0 N83, t-BUOH> PivO

 

  

O \. . CCI4, reflux 6“
L08 ’20in )1 s.- ,
Q‘ o I -
Ph Ph OPIV
1 4
OPiv OPiv
PIvO \ H 0 0 H20 PIvO H 0 o
3ch ‘— 0“
HO [IOPIV 970‘ %Pw
P“ Gt 8
3 ' u 11

Scheme 3.7 The NBS treatment of benzylidenelactone with t-butyl alcohol in presence

119

We therefore chose carboxylate ion as our competitive nucleophile. Sodium
acetate was employed as a suitable carboxylate ion source. We treated 3,5-O-benzylidene-
2,6,7-tri-O-trimethylacetyl-D-glucoheptano-l,4-lactone 1 with NBS with 5-10 equivalent
sodium acetate present. Afier refluxed in carbon tetrachloride for 12 hours, NMR of the
crude reaction mixture suggested the major product to be 3,5-O-bromobenzylidene-2,6,7-
tri-O-trimethylacetyl-D-glucoheptano-1,4-lactone 7. In this case, sodium acetate only
played a role of acid scavenger. It lowered the acidity and the bromide ion concentration
so that slowed down the benzylidene group cleavage. Isomerization at the
bromobenzylidene C took place, shown by the '3 C NMR spectra. When the reaction time
was further extended, the NMR of the reaction mixture became very complicated,
indicating a mixture of products. However, NMR also showed the emergence of 7-
bromo-7-deoxy-3-O-benzoyl-2,5,6-tri-O—trimethylacetyl-D-glucoheptano- 1 ,4-lactone 2,
which was the rearrangement and terminal bromination product. Attempted separation of
the other products by chromatography only yielded one major product, which was
confirmed to be S-O-benzoyl-2,6,7-tri-O-trimethylacetyl-D-glucoheptano-l,4-lactone 8.
Apparently, hydrolysis of the reaction mixtures led to this single product. A mechanism
was proposed in Scheme 3.8. After the C-Br bond cleavage, a carboxonium cation and a
bromide ion were generated. In the presence of excess sodium acetate, the carboxylate ion
became competitive with the bromide ion. The carboxylate ion could trap the
carboxonium cation to form the orthoanhydride. This competitive reaction also slowed
down the rearrangement and terminal bromination reaction. On the other hand, the
formation of small amount 7-bromo-7-deoxy-3-O-benzoyl-2,5,6-tri-O-trimethylacetyl-D-
glucoheptano-l,4-lactone 2 also suggested that bromide ion was a better nucleophile than

sodium acetate since no 7-O-acetate product was detected. Under this condition, acetate

120

could not compete with the bromide ion in nucleophile substitution. Potassium benzoate
was also tested in this reaction. Although benzoate ion should have a better

nucleophilicity, the NBS reaction with potassium benzoate present showed the same

 

 

result.
OP'V OPiv
eBr
H
FWD/EC NBS NaOAc FWD 0“ O o —>
Phsko ’0in CCI4, reflux /é§ '90in 98(‘0
1 4
6
Phil on
OPiv OP'V 0in
H o PivO H 0 H o
PivO H 0 2 rwo Br ,. 0
EC ‘— ’ Pivd ‘
0“ EPIV ”1%0 OF” 326‘ I’OPiv
8 OAc 12 2

Scheme 3.8 The NBS treatment of benzylidenelactone with sodium acetate as the
competitive nucleophile

3.3.2 The introduction of N-contained nucleophiles into the NBS oxidation

Although the O-nucleophiles introduction by the NBS reaction was generally not
successful, we continued our exploration using other nucleophiles. Related to our
synthetic efforts towards iminosugars as glycosidase inhibitors, it was of more importance
to introduce N-contained nucleophiles into the sugar structure. This functionalization,
combined with further transformation, could generate versatile iminosugars. Cyanides
and nitrites were chosen to explore the NBS reaction with competitive nucleophiles

present because of their cheap prices. The successful reactions with cyanides and nitrites

l2]

could generate N-functionalized sugar structures with one carbon extension and the same

carbon number respectively.

3,S-O-Benzylidene-2,6,7-tri-O-trimethylacetyI-D-glucoheptano-1,4-lactone 1 was
treated with NBS in refluxing carbon tetrachloride in the presence of excess sodium
cyanide. NMR analysis of the crude reaction mixture was employed to monitor the
reaction. The analysis showed a similar reaction model with the NBS reaction using
barium carbonate as an acid scavenger. The pivaloyl rearrangement and bromination was
slowed down and the bromobenzylidene compound isomerization was observed. Under
extension of the reaction time, 7-bromination slowly emerged as a major reaction. In this
case, sodium cyanide only acted as an acid scavenger to neutralize the trace amount of
HBr. The low solubility of sodium cyanide in CCI4 limited its ability to act as a
competitive nucleophile; only bromination occurred after long reaction time. We
postulated that using better solvents and/or phase transfer catalysts might be able to
introduce more sodium cyanide into the reaction media so that cyanide ion could compete
with the bromide ion to induce the different reaction model. Unfortunately, neither use of
a better solvent (acetonitrile) nor addition of a phase transfer catalyst (l8-crown-6) was
successful. In either case, only starting materials were recovered. Under these reaction
conditions, the basicity and the ionic strength of the solution were substantially increased.
Therefore NBS reacted with the base before it could react with the benzylidene C-H bond
in a free radical mode, which is the initial step of the rearrangement and the bromination.
This explanation was further confirmed by employing some other base such as sodium
acetate as the competitive nucleophile instead of cyanides, in which also only starting

material was recovered.

122

Instead of cyanides, we also tried using acetone cyanohydrin, which is often used
to generate cyanide in-situ, in the NBS reaction. When the NBS reaction was carried out
with only acetone cyanohydrin in presence, the normal pivaloyl rearrangement and 7-
bromination was observed. This indicated that under neutral or slightly acid conditions,
no in-situ cyanide was generated to compete with the bromination reaction. We then used
barium carbonate to generate in-situ cyanide. Afier refluxing for several hours, NMR
analysis showed the major product to be of S-O—benzoyl-Z,6,7-tri-O-trimethylacetyl—D-
glucoheptano-l,4-lactone 8, indicating that hydrolysis had occurred. This was logical
since the in-situ cyanide generation by acetone cyanohydrin and barium carbonate could
generate some water. To avoid this problem, pyridine was used as the base. Unfortunately,
the introduction of pyridine into the NBS reaction led to B-elimination, yielding the (1J3-

unsaturated-D-glucoheptanolactone 9.

When sodium nitrite was employed as the competitive nucleophile in the NBS
reaction, the reaction mode was similar with that using sodium cyanide as the competitive
nucleophile. We reasoned that using silver nitrite could introduce more nitrite ion into the
solution by scavenging the bromide ion out of the reaction media. When the reaction was
refluxed in CCI4 for 24 hours with silver nitrite in the presence, yellow precipitates were
observed. However, NMR analysis of the reaction mixture showed only starting material.
Under this condition, NBS reacted with silver nitrite directly to form silver bromide
because of the strong bromide scavenging ability of silver salts. After that, the newly
formed compound, presumably N-nitrosuccinimide, could not activate the benzylidene C-

H and therefore the starting material was untouched.

123

Generally, introducing the competitive nucleophile into the N-bromosuccinimide-
mediated dioxolonium ion rearrangement did not give desired products with new
functionalities. This finding is largely in agreement with the literature. Jacobsen and
Pedersen and coworkers [5-11] explored the reactions of various nucleophiles with
dioxonium ions derived from carbohydrates. Various benzylidene sugar structures were
treated with triphenylmethyl fluoroborate to give benzoxonium ions and treated with
various nucleophiles. With chloride, bromide, iodide, thiocyanate and tosylate trans-
opening of the dioxonium ion took place. However, for a various number of nucleophiles
such as acetate, azide, cyanide, no trans-opening was achieved. This was ascribed to the
direct trapping of the dioxonium cation by nucleophiles to form orthoester-like structures,
which underwent hydrolysis to give cis-opening. The nucleophilic trans-opening was
shown to be competitive with the direct trapping of the dioxonium cation. The preference

was decided by the nature of the nucleophiles introduced in.

3.4 The introduction of intramolecular competitive nucleophiles into
the NBS bromination of 3,5-O-benzyIidene—2,6,7-tri-O-pivaloyl-D-

glucoheptono-1,4-lactone

R R
Nu [2 u i2 R
—> —" CH R1
0 O O 6') O
Y 2; Y 082

Ph 1"“

Scheme 3.9 Intramolecular competitive nucleophile participated Hanessian-Hullar reaction

124

Ever since the Hanessian-Hullar reaction was discovered, the neighboring group
participation was explored [12]. Normally the neighboring group acted as an
intramolecular competitive nucleophile, which could drive the reaction to different
directions and yield new functionalized structure (Scheme 3.9). Actually, the pivaloyl
rearrangement discussed in Chapter 2 was also a Hanessian-Hullar reaction participated
by intramolecular competitive nucleophiles, i.e. neighboring ester groups. Jacobsen
explored the nucleophilic trans-opening of dioxonium ions with neighboring group
participation [13-15]. Various neighboring groups such as thionobenzoate, xanthate,
carbamate, thiocarbamate, benzimidate and carbonate were employed. With an iminoester
group as the neighboring group, the nitrogen functionality could be introduced into the

sugar structure to form an aminosugar (Scheme 3.10).

H Ph
Ph
l'__-O G»\l'—-
o s o 00
0 ~ NB
0 ——> ( —> 320
HN 0 HN§( OH3N@ OH
CCI3 CCIS C130
13 14 15 16

Scheme 3.10 The iminoester neighboring group participated Hanessian-Hullar reaction

Based on our goal to the synthesis of iminosugar structures, we reasoned that an
intramolecular cyano group could act as a competitive nucleophile to give a lactam,
which could be used as iminosugar precursor (Scheme 3.11). Furthermore, the lactam

structure itself would be of much interest as a potential glycosidase inhibitor.

125

// R Br
'2 NBS I $2
CHle/Rr N CH Qi/R,
l 082

0Y0
Ph
0 OH
12 I r2
*—
HN CHVQKR1 N CHfi/R1

082 082

Scheme 3.1 1 Proposed intramolecular cyano group participated Hanessian-Hullar reaction

To test our hypothesis, we first employed 2,3-O-benzylidene-L-threitol 17 as our
starting material because of its simplicity and easy accessibility. Due to its C2 symmetry,
no regioselectivity issues needed to be addressed. When treated with sodium hydride
(NaH) and bromoacetonitrile, 2,3-O-benzylidene-l,4-di-O-cyanomethyl-L-threitoI 18 was
easily synthesized. We then treated this compound with NBS in carbon tetrachloride. A
single product was obtained in 75% yield. However, NMR analysis showed the structure
to be 2-O-benzoyl-3-bromo-3-deoxy-l,4-di-O-cyanomethyl-D-erythritol 20 (Scheme
3.12). The reaction proceeded with normal Hanessian-Hullar reaction mode without
intramolecular nucleophile participation. Using barium carbonate to slow down the
bromination did not help the intramolecular nucleophilic attack. The same product was

obtained.

126

\\“

HOVOH NCHfiOVOCHZCN

 

O O NaH, bromoacetonitrile
> O O
62% Y
Ph Ph
17 1 8
NBS, CCI4, refluxm/ *
NCHZCOWOCHZCN NCHzcowo
BzO Br BzO HNV
20 19 0

Scheme 3.12 Attempted cyano-participated Hanessian-Hullar reaction: synthesis of 2-0-
benzoyl-3-bromo-3-deoxy-l ,4-di-O-cyanomethyl-D-erythritol

We then tried to shorten the distance between the benzylidene group and the
cyano group, hoping that this could change the transition state ring structure so that
change the preference of the intramolecular nucleophile attack. We then made 2,3-O-
benzylidene-l,4-di-O-p-toluenesulfonyl-L-threitol by treatment of 2,3-O-benzylidene-L-
threitol 17 with p-toluenesulfonyl chloride in pyridine. However, even after many
attempts with different reaction conditions, the expected nucleophilic substitution of this
di-tosylate with sodium cyanide to afford a dicyano compound did not occur. Elimination
was observed as a serious side reaction. Transforming the tosyl leaving groups to iodo

groups did not improve the desired nucleophilic substitution.

We therefore tried to seek other synthetic pathway to our desired molecules to test
our idea of cyano group participated Hanessian-Hullar reaction. Sugar oximes have been
synthesized readily for many years. Transformation of the N-hydroxyl group of a sugar
oxime to bromide or ester often results elimination to give a sugar nitrile. Sugar nitriles,
appropriately protected, could serve as substrates in our exploration of the intramolecular

cyano-participated Hanessian-Hullar reaction.

127

 

OPiv
OH HO
0 HO -
O
HONHzHCI 0H 50 /° OH
OH FY 23
. 22 76%
21 PhCH(OCH3)2.

CHzClz. stOH

0 OPiv
wow A NBS, cm... 66 /o {SK/R
/ BZ ‘ \ N

N/ Ph 0
25 24

 

Scheme 3.13 Synthesis of 2-deoxy-D-ribonitrile and attempted cyano-participated
Hanessian-Hullar reaction

To shorten the synthetic precedure, we tried to combine the oxime formation and
the elimination to synthesize the sugar nitriles. 2-Deoxy-D-ribose 21 was treated with
hydroxylamine hydrogen chloride in pyridine and subsequently treated with pivaloyl
chloride to afford the 2-deoxy-5-O-trimethylacetyl-D-ribonitrile 23 in 60% yield (Scheme
3.13). This nitrile was then treated with benzaldehyde dimethyl acetal to give 3,4-0-
benzylidene-2-deoxy—S-O-trimethylacetyl-D-ribonitrile 24 as a 4:1 mixture of two
diastereoisomers in 76% yield. In the NBS-cleavage of benzylidene group, this new
chirality would be scrambled without obvious effects on the product structure. Hence 3,4-
O-benzylidene—Z-deoxy-S-O-trimethylacetyl-D-ribonitrile 24 was treated with NBS in
refluxing carbon tetrachloride. NMR analysis of the reaction showed no evidence of
formation of lactams. This meant no Ritter-type reaction occurred under our reaction
condition. The only major product was separated and analyzed. The structure was
confirmed to be 3—O-benzoyI-5-bromo-2,5-dideoxy-4-O-trimethylacetyI-L-lyxonitrile 25.
The reaction proceeded through an NBS-promoted pivaloyl rearrangement and primary
bromination. Using acid scavengers such as barium carbonate only slowed down the

bromination reaction without promoting any other side-reactions. Under our reaction

128

condition, the intramolecular cyano group was unreactive towards the in-situ formed
benzoxonium carbocation. Due to the steric hindrance, the primary pivaloyl group
rearranged to the 4-position and inverted the chiral center. The following bromination at
the primary position furnished the protected bromosugamitrile as the final product.
Although the Ritter-type reaction did not happen, the product itself was also of much
significance. Protected and brominated, this L-sugar nitrile could serve as a good chiral

building block towards azasugars by controlled reduction.

3.5 Summary

The base-sensitivity of the N-bromosuccinimide-mediated dioxolonium ion
rearrangement was discovered. While inorganic base slowed down the rearrangement and
terminal bromination by lowering down bromide ion concentration, organic base, i.e.,
pyridine induced a controlled B-elimination to generate an a,B—unsaturated sugar lactone.
Various competitive nucleophiles were used in the NBS-cleavage of benzylidene acetal of
sugar lactone to explore the introduction of new functionalities into the sugar structure.
The introduced nucleophiles generally could not compete with in-situ generated bromide
ion in the terminal nucleophile substitution. The direct trapping of the benzoxonium
cation by intermolecular competitive nucleophiles such as t-butyl alcohol and
carboxylates were also believed to be an important competing reaction. Intramolecular
attack of benzoxonium cation by cyano group did not occur. This was due to the much

lower nucleophilicity of cyano group than bromide ion.

129

EXPERIMENTAL

General Optical activity data were obtained on a JASCO P-1010 polarimeter at
25°C. NMR spectra were obtained on a Varian VXR-SOO Spectrometer operating at
500MHz for protons. Mass spectra were obtained on a JEOL HX-l 10-HF instrument
using fast atom bombardment as ionization mode. Spectra were recorded in the positive
ion mode. IR spectra were obtained on a Nicolet 710 spectrometer in chloroform solution

except when otherwise specified.

S-O-Benzoyl-2,6,7-tri~O-trimethylacetyl-D-glycero-D-gulo-heptono-1 ,4-Iactone (8):

2,6,7-Tri—O-trimethylacetyl-3,5-O-benzy1idene-D-glycero-D-gulo-heptono- l ,4-
lactone (1.0 g, 1.82 mmol) was suspended in 10 mL carbon tetrachloride and N-
bromosuccinimide (0.5 g, 2.81mmol ) was added. Water (86 mg, 4.78 mmol) was added
in. The mixture was refluxed at 80°C oil bath for 5 hours. The mixture was filtered and
washed with saturated sodium bicarbonate solution. The raw product was purified by
flash column chromatography to afford S-O-Benzoyl-2,6,7-tri-O-trimethylacetyl-D-
glycero-D-gulo-heptono-1,4-lactone 8 (0.74 g, 72% yield) as colorless oil. [a]D = -58.5°
(c = 1.0, CHCI3); 8H (500MHz; CDCI3): 1.15,1.16,1.25 (27H, 3 x s, 3 x C(Cfl3)3), 3.49
(1H, d, J = 3.5 Hz, -Ofl), 4.21 (1H, dd, In 12.1, ij‘ 8.3 Hz, H-7’), 4.58 (1H, dd, J63 3.2
Hz, H-7), 4.71 (1H dd, 13,, 2.9, .145 8.5 Hz, H-4), 4.89 (1H, m, .123 4.7 Hz, H-3), 5.34 (1H,
ddd, 15,6 8.3 Hz, H-6), 5.44 (1H, d, H-2), 5.91 (1H, dd, H-S), 7.42-7.44 (2H, m, Ar_H),
7.56-7.58 (1H, m, ArH), 8.00-8.02 (2H, m, ArH); 6c ( 125MHz; CDC13): 26.91, 27.00
(OCC(QH3)3), 38.57, 38.65, 38.81 (OCQ(CH3)3), 61.90 (C-7), 68.44, 69.98, 70.05, 70.70,

77.98 (C-2, 3, 4, 5, 6), 128.53, 129.07, 129.85 (ArQH), 133.52 (Arg), 164.85 (O_C_Ph),

130

169.80 (O=Q in lactone), 177.28, 178.21, 178.54 (OQC(CH3)3); FT-IR: 3456, 2976, 1807,

1734, 1265, 1139 cm".

(LB-Unsaturated-3-deoxy-5-O-benzoyl-2,6,7-tri-O-trimethylacetyl-D-glycero-D-gulo-

heptono- 1 ,4-lactone (9):

2,6,7-Tri-O-trimethy1acetyl-3,S-O-benzylidene-D-glycero-D-gulo-heptono- l ,4-
lactone (I g, 1.82 mmol) was suspended in carbon tetrachloride (10 mL) with N-
bromosuccinimide (0.5 g, 2.81 mmol ). Pyridine (0.29 g, 3.64 mmol) was added. The
mixture was refluxed at 80 °C for 6 hours and filtered through a thin pad of silica gel. The
filtration was evaporated to give 9 (0.75 g, 75%) as pale yellow oil. [a]D = + 4.9° (c = 0.8,
CHC13); 8H (500MHz; CDC13): 1.1 1,1.12,1.17 (3 x 9H, 3 x s, 3 x C(Cfl3)3), 4.10 (1H, dd,
17y 12.6, 1634.8 Hz, H-7’), 4.56 (1H, dd, 16,725 Hz, H-7), 5.30 (1H, d, 14.5 2.0Hz, H-4),
5.45 (1H, ddd, 15,6 7.4Hz, H-6), 5.64 (1H, dd, H-S), 738 (2H, m, ArH), 7.52 (1H, m,
ArH), 7.88 (2H, m, Arfl); 6c (125MHz; CDCI3): 26.54, 26.85, 26.93 (OCC(_(;H3)3),
38.67, 38.74, 39.15 (OCQ(CH3)3), 61.41 (C-7), 68.68, 69.63, 76.81 (C-4, 5, 6), 124.78,
127.91 (02, 3), 128.50, 129.15, 129.65 (ArQH), 133.67 (Arg), 164.89 (OQPh), 165.82
(O=_C_ in lactone), 174.11, 176.69, 177.68 (OQC(CH3)3); FT-IR: 2980, 1986, 1726, 1269,

1 116 cm".
2,3-O-Benzylidene- 1 ,4-di-O-cyanomethyl-L-threitol (18):

2,3-O-Benzy1idene-L-threitol (1.0 g, 4.8 mmol) was dissolved in 20 mL
anhydrous tetrahydrofuran and cooled down to -7 8°C. Sodium hydride (60% suspension
in mineral oil) (0.57 g, 14.3 mmol) was added slowly under stirring. Bromoacetonitrile

(1.44 g, 12.0 mmol) was added slowly. The mixture was warmed up to 0°C slowly and

131

stirred at this temperature for 2 hours. It was poured into ice and stirred vigorously. The
mixture was extracted by chloroform (3 x 50 mL). The organic layer was washed by brine
(25 mL) and water (25 mL). It was dried and evaporated to dryness. The crude product
was purified by chromatography to afford 2,3-O-benzylidene-1,4-di-O—cyanomethyl-L—
threitol 18 as colorless oil (0.85 g, 62%). [11].) = + 25.30 (c = 1.1, CHC13); 8H (500MHz;
CDC13): 3.79 (4H, d, 11,2 4.4Hz, H-1, 4), 4.21 (2H, m, H-2, 3), 4.32 (4H, m, OCfl2CN),
5.93 (1H, s, CflPh), 7.37-7.47 (5H, m, ArH); 6c ( 125MHz; CDCI3): 56.53 (OQHZCN),
71.29, 71.42 (C-1, 4), 76.32, 76.35 (C-2, 3), 104.09 (_QHPh), 115.29 (OCHng), 126.32,

128.15, 129.43 (ArQH), 136.25 (Arg); FT-IR: 3038, 2937, 2206, 1101 cm".
2-O-Benzoyl-3-bromo-3-deoxy-l ,4-di-O-cyanomethyl-D-erythritol (20):

2,3-O-benzylidene-l,4-di-O-cyanomethyl-L-threitol 18 was reacted with 2
equivalent NBS in carbon tetrachloride under reflux for 6 hours. After column
chromatography, 2-O-benzoyl-3-bromo-3-deoxy-1,4-di-O-cyanomethy1-D-erythritol 20 in
75% yield. [a]D = +22.5 (c = 0.9, CHCI3); 5H (500MHz; CDCI3): 4.00 (5H, m, H-1, 3, 4),
4.28 (4H, m, OCflzCN), 5.51 (td, 1H, J = 4.6Hz, J = 6.5Hz, H-2), 7.45 (t, 2H, J = 7.6Hz),
7.58 (t, 1H, J = 7.3112), 8.03 (d, 2H, J = 7.7Hz); 8c (125MHz; CDC13): 46.951 (03),
56.583, 56.402 (OQH2CN), 71.699, 71.374, 70.478 (C-1, 2, 4), 104.09 (QHPh), 115.320,
115.156 (OCH2QN), 129.738, 128.837, 128.513 (ArQH), 133.626 (ArQ), 165.104

(O_C_Ph); FT-IR: 1721, 1267, 1108 cm".
2-Deoxy-5-O-trimethy1acetyl-D-ribonitrile (23):

2-Deoxy-D-ribose (1.5 g, 11.2 mmol) was dissolved in 40 mL pyridine and

hydroxylamine hydrogen chloride salt (1.17 g, 16.8 mmol) was added in. The solution

132

was stirred at room temperature for 12 hours. The solution was cooled down to 0 °C and
trimethylacetyl chloride (4.72 g, 39.2 mmol) was added slowly. The solution was warmed
up to room temperature slowly and stirred for 12 hours thereafter. The solution was
poured into sodium bicarbonate with ice and stirred for 10 minutes and extracted by 3 x
50 mL chloroform. The organic phase was dried and evaporated to sticky oil. This crude
product was purified by flash column chromatography to give 2-deoxy-5-O-
trimethylacetyl-D-ribonitrile 23 as colorless oil (1.44g, 60%). [11].) = -21.70 (c = 1.0,
CHC13); 5H (500MHz; CDC13): 1.20 (9H, s, C(Cfl3)3), 2.66 (1H, dd, 12116.7 Hz,
Jzt36.8Hz, H-2’), 2.76 (1H, dd, J2,33.6Hz, H-2), 3.06 (1H, b, -Ofl), 3.59 (1H, b, -OH_),
3.79 (2H, m, H-3, H-4), 4.26 (1H, dd, 15512.2 Hz, J4,5~3.3Hz, H-5’), 4.33 (1H, dd,
1454.8Hz, H-5); 6c (125MHz; CDC13): 22.52 (C-2), 27.10 (OCC(_QH3)3), 38.96
(OCQ(CH3)3), 65.25 (C-S), 67.78, 72.06 (C-3, C-4), 118.03 (01), 179.79 (OQC(CH3)3);

FT-IR: 3448, 2976, 2439, 1709, 1286, 1165, 1071 cm".
3,4-O-Benzylidene-2-deoxy-5-O-trimethy1acetyl-D-ribonitrile (24):

2-deoxy-5-O-trimethylacety1—D-ribonitrile 23 (0.2 g, 0.93 mmol) was dissolved in
5 mL dichloromethane with benzaldehyde dimethyl acetal (0.28 g, 1.8 mmol). p-
Toluenesulfonic acid (0.03 g) was added in. The solution was stirred at room temperature
for 12 hours and poured into saturated sodium bicarbonate solution. The mixture was
extracted by chloroform (20 mL). The organic phase was dried and evaporated. The crude
mixture was purified by flash column chromatography to give 3,4-O-benzylidene-2-
deoxy-5-O-trimethylacety1-D-ribonitrile 24 (0.21 g, 76%) as a mixture of two isomers (4 :
1). For the major isomer (3,4-(S)-O-benzy1idene): 6H (500MHz; CDCI3): 1.22 (9H, s,

C(Cfl3)3), 2.66 (1H, dd, J2,2~I6.7 Hz, J2~,37.0Hz, H-2’), 2.71 (1H, dd, J2,34.4Hz, H-2), 3.62

133

(1H, dd, J5,5'10.8 Hz, 14.59.8112, H-S’), 4.08 (1H, ddd, J3,49.6Hz, H-3), 4.40 (1H, dd,
14.55.4112, H-S), 4.78 (1H, dt, H-4); 8c (125MHz; CDC13): 21.655 (C-2), 26.999
(OCC(QH3)3), 38.888 (OCQ(CH3)3), 67.648 (C-S), 74.752, 65.912 (C-3, C-4), 101.405
(QHPh), 116.153 (C-l), 136.318, 129.370, 128.330, 126.099 (ArQ), 177.181

(OQC(CH3)3); FT-IR: 2966, 2436, 1718, 1275, 1158 cm".
3-O-Benzoyl-5-bromo-2,5-dideoxy—4-O-trimethy1acetyl-L-lyxonitrile (25):

3,4-O-Benzylidene-2-deoxy-5-O-trimethylacetyl-D-ribonitri1e 24 (0.15 g, 0.5
mmol) was dissolved in 5 mL carbon tetrachloride and N-bromosuccinimide (0.18 g, 1.0
mmol) was added. The mixture was refluxed for 6 hours and filtered. The filtrate was
purified by chromatography to give 3-O-benzoyl-5-bromo-2,5-dideoxy-4-O-
trimethylacetyl-L-lyxonitrile 25 as colorless oil (0.15 g, 66%). [a]D = +1.90 (c = 0.7,
CHCI3); 5H (500MHz; CDC13): 1.25 (9H, s, C(C_1'_1_3)3), 2.81 (1H, dd, J2,2~17.3 Hz,
Jz~,34.9Hz, H-2’), 2.97 (1H, dd, J2,34.7Hz, H-2), 3.57 (1H, dd, J5,5~11.6 Hz, J4,5~4.7Hz, H-
5’), 3.71 (1H, dd, J4,s4.0Hz, H-S), 5.31 (1H, td, 13.47.2Hz, H-4), 4.54 (1H, td, H-3); 8c
(125MHz; CDC13): 19.722 (02), 26.915 (OCC(§H3)3), 29.984 (C-5), 38.958
(OCQ(CH3)3), 70.604, 67.878 (C—3, C-4), 115.427 (01), 133.896, 129.797, 128.586,
128.329 (ArQ), 164.716 (QOPh), 176.745 (OQC(CH3)3); FT-IR: 3098, 3077, 1728, 1627,

1603, 1262, 1139, 1104 cm".

134

[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]

[9]

REFERENCE

Hanessian, S. Carbohydr. Res. 1966, 2, 86.

Hanessian, S.; Pleassas, N. R. J. Org. Chem. 1969, 34, 1035, 1045 and 1053.
Binkley, R. W.; Goewey, G. S.; Johnston, J. C. J. Org. Chem, 1984, 49, 992.
King, J. F.; Allbutt, A. D. Can. J. Chem. 1970, 48, 1754.

Jacobsen, S.; Pedersen, C. Acta Chem. Scand. B. 1974, B28, 866.

Jacobsen, S.; Nielsen, B.; Pedersen, C. Acta Chem. Scand. B. 1977, B31, 59.
Jacobsen, S.; Pedersen, C. Acta Chem. Scand. B. 1977, B31, 365.

Hoffineyer, L.; Jacobsen, S.; M015, 0.; Pedersen, C. Acta Chem. Scand. B. 1979,
B33, 175.

Jacobsen, S.; M015, 0. Acta Chem. Scand. B. 1981, B35, 163.

[10] Jacobsen, S.; M015, 0. Acta Chem. Scand. B. 1981, B35, 169.

[11]Jacobsen, S.; M015, 0. Acta Chem. Scand. B. 1981, B35, 521.

[12] Ponpipom, M. M.; Hanessian, S. Carbohydr. Res. 1971, I 7, 248.

[13] Jacobsen, S. Acta Chem. Scand. B. 1984, B38, 157.

[14] Jacobsen, S. Acta Chem. Scand. B. 1986, B40, 493.

[15] Jacobsen, S. Acta Chem. Scand. B. 1986, B40, 498.

135

Chapter 4

Acid-Promoted Dioxonium Cation Facilitated Bromination and

Ritter-Type Reactions

ABSTRACT

Facilitated by dioxonium cation formation and rearrangement under strong acidic
conditions, selective bromination of aldonolactones was carried out to generate useful
bromosugars. The anomeric bromination of hexose pentapivaloates was achieved to
generate tetrapivaloyl glycosyl bromides as useful glycosyl donors. When acetonitrile
was employed as solvent and nucleophile, Ritter-type reactions occurred to afford several
glycosyl amides. Furthermore, a unique dioxonium cation assisted Ritter-type reaction
was discovered when glycerol dipivaloates were treated with strong acid in acetonitrile.
The stereoselectivity and regioselectivity of this new Ritter-type reaction were tested by
using erythritol tripivaloate and 1,2,4-butanetriol dipivaloates as the substrates. The
Ritter-type reaction was employed to easily introduce amide groups to the primary
position of aldonolactone to generate important and interesting N-contained sugar

derivative.
4.1 Introduction

In the previous chapters, we discussed the NBS-promoted acetoxonium cation
rearrangement. The dioxonium cation generated in-situ by H-abstraction of benzylidene

groups acts as a good leaving group. The intramolecular attack by an active neighboring

I36

group such as the pivaloyl group results acyl groups rearrangement and transformation of
chirality. An intermolecular nucleophile such as bromide ion furnishes a new
functionality in the molecule and drives the equilibirum of the acyl rearrangement
towards the less steric hindered specie. The acetoxonium cation generated under acidic
conditions has long been known and used as an electrophile [1-17]. The acidic
bromination of sugar lactones [7, 14] (Scheme 4.1 A) and the preparation of
anhydrosugars [5] are two prominent examples (Scheme 4.] B). In these reactions, under
acidic conditions, acetoxonium cation is formed through the participation of the acetyl
group and acts as an electrophile, which is subsequently attacked by intramolecular (-

OH) or intermolecular (Br-) nucleophiles to finish the functional group transformation.

 

A)
OH I >@—o OAc OH
HO 0 ”3' o 0 Br 0 B, o
, o HOAc . o t o o
H . H 2. _ _. H MeOH H” H‘
Br) Br ,4 —'_> ,’
HO OH 0%0 AcO Br HO ’Br
8)
01-1 >90 HO H o
H o AcOH, Hr=_ 0 \ O 0 H20 ’ o
90 \0 7 Q 4 20H
OH HO ’OH 0

Scheme 4.1 Acetoxonium cation generated under acidic conditions : A) dibromination of
aldonolactones; B) anhydrosugar formation

Our success in the control of the site of bromination and the stereochemical
course and chemoselectivity of rearrange reactions involving NBS promoted dioxonium
species suggested several important synthetic uses and methodology development
pathways for this chemistry. One such area is the development of strategies for the

preparation of azasugars. Other areas include the preparation of rare L-sugars and

137

activated but orthogonally protected alditols. To this end, we embarked on a study to
develop industrially useful methodologies for synthesis of series of differentially
protected bromosugar lactones using easily accessible starting materials. The
acetoxonium cation facilitated bromination could generate monobromo- or dibromosugar
structures. The acyl group could serve as a good protecting group in the resulted
bromosugar structure. Bearing good leaving group and selective protecting group, these

bromosugars could serve as good synthetic precursors of aminosugars (Figure 4.1).

R
0” + >63 o B e OOCR
RCOOVng L OW r Br\/LAI‘S

OOCR //
OOCR
.;:r RHN\V/J\
RN -

Figure 4.1 bromosugars generated from the acetoxonium cation facilitated bromination serve
as good synthetic precursors of iminosugars

4.2 Selective pivaloylation and bromination facilitated by dioxonium

cation formation and rearrangement

4.2.1 Synthesis of 2,6-dibromo-2,6-dideoxy-5-O-trimethylacetyl-D-idono-1,4-

lactone

D-gulono-1,4-1actone l was selectively pivaloylated at the primary position to
yield 6-O-trimethylacetyI-D-gulono-1,4-lactone 2. Compound 2 was then treated with

30% hydrogen bromide in acetic acid (HBA) to give 2,6-dibromo-2,6-dideoxy-5-O-

138

pivaloyl-D-idono-1,4-lactone (4) in 58% yield over two steps (Scheme 4.2).

t-Bu@

 

PivCl Py. H§ HBr-HOAc ‘\ \, ,,
2 H0 H6 6H
lHBfHHOAc 5
tBu l
2&0
- H
MeOH.H* 0 O o P'VO o o
<— H
58% over Br ) § 4:63" m;
Zsteps (5%?) 0° DH
4 3

Scheme 4.2 Synthesis of2,6-dibromo-2,6-dideoxy-5-O-trimethylacetyl-D-idono—1,4-1actone 4

In the strongly acidic conditions, a t-butyl-l,3-dioxonium cation 3 was formed,
and subsequently opened by the incoming Br‘nucleophile. A competitive side reaction is
the intramolecular attack of the t-butyl-l,3-dioxonium cation 5 by the 3-OH group to
form a 3,6-anhydrous sugar lactone 6. The work-up procedure (methanol
transesterification) removed the acetyl group leaving the more robust 5-O-pivaloyl group
untouched. The pivaloyl group was used as a protection group in the synthetic procedure
for several reasons. Firstly, it has a relatively better selectivity for functionalizing primary
hydroxyl groups because of its size. Secondly, the electron donating t-butyl group could
stablize the carboxonium cation intermediate and therefore increase the reaction
efficiency. Thirdly, the pivaloyl group is more tolerant to the acidic conditions used to
selectively remove the acetyl group. The net result of this procedure was the selective 5-
position pivaloylation of a 2,6-dibromo-2,6-dideoxy-D-idono-y—lactone. In other words,

the O-acylation selectivity at the primary position of sugar lactones was successfully

139

transferred to a neighboring position by an acid-catalyzed pivaloyl rearrangement.

The importance of this reaction was further demonstrated by comparing with the
direct selective pivaloylation of 2,6-dibromo-2,6—dideoxy-D-idono-1,4-lactone 7 under
normal pivaloylation condition (Scheme 4.3 A). When 7 was treated with 1 equivalent
pivaloyl chloride in the presence of pyridine in dichloromethane, we expected two
products: 3-O-pivaloy1-2,6-dibromo-2,6-dideoxy-D-idono-1,4-lactone 8 and 5-O-pivaloy1
2,6—dibromo-2,6-dideoxy-D-idono-1,4-1actone 4 with the latter one being slighly favored.
However, NMR analysis of the products showed 3-O-pivaloyl dibromolactone 8 as the

major product with only small amounts of S-O-pivaloyl dibromolactone 4 (4:8 = 1:6).

  

 

   

OPiv
1eq. PivCl, py. Br
0 dichloromethane; o +
H6 Br
7 4
Expected: major
Actual: 1
B) 0H OPiv
HO 0 FWD 0
H O 3 3e PivCl H O
\ , q—..' py. \ , crystalline. 80% Ref 18
H6 ”0H H6 ’bPiv
1 9

Scheme 4.3 A) Selective monopivaloation of2,6-dibromo-2,6-dideoxy-D-id0no- 1,4-1actone
7 and B) selective tripivaloation of D-gulono-I ,4-lactone 1

This unexpected selectivity was very different from the results of selective
tripivaloylation of sugar lactones by Gallo's group [18]. By treatment of D-gulono-1,4-
lactone 1 with 3.3-3.6 equivalent pivaloyl chloride in pyridine, 2,5,6-tri-O-pivaloyl-D-

gulono-I,4-lactone 9 was selectively obtained in 80% yield as a crystalline compound

140

(Scheme 4.3 B). The selectivity of 3-O-acylation over 5-O-acylation was fairly high.

However, our approach showed a relatively high but reversed selectivity.

This selectivity reversal could be easily understood by considering the
stereochemistry change at 2-position. D-gulono-y-lactone has 2,3-cis oriented hydroxyl
groups partially eclipsed. Once the 2-O-position is protected by a bulky pivaloyl group, it
would generate high steric hindrance towards the incoming pivaloyl group to 3-position.
For 2,6-dibromoidonolactone 7, the bromination of 2-position changes the 2,3-cis
orientation to 2,3-trans. This transformation greatly released the steric hindrance of 2-
substitution group towards the approach of the incoming pivaloyl group to 3-position.
The selectivity of pivaloylation between 3-0 and 5-0 position was therefore reversed.
Under normal pivaloylation conditions, 3-O-pivaloate 8 of 2,6-dibromoidonolactone 7
instead of 5-O-pivaloate 4 was synthesized. This further illustrates the practical utility of
the S-O-pivalation of 2,6-dibromolactone through t-butyl-1,3-dioxonium cation

rearrangement procedure.
4.2.2 Synthesis of 6-bromo-6-deoxy-2,5-di-O-pivaloyl-D-galactono-1,4-1actone

To expand this strategy to other sugar lactones, 2,6-di-O-pivaloyl-D-galactono-
1,4-lactone 11 was synthsized as discussed in chapter 2 (Scheme 4.4 A). It was then
treated with 30% hydrogen bromide in acetic acid (HBA) to yield 85% 6-bromo-6-deoxy-i
2,5-di-O-pivaloy1-D-galactono-1,4-1actone 12. The 5,6-O-t-buty1-1,3-dioxonium cation
was believed to be the active intermediate. With galacto- configuration, the trans-
orientation of the 2,3-substitution groups prevented the formation of 2,3-O-t~buty1-1,3-

dioxonium cation formation and therefore no 2-bromo compound was formed. This is in

141

consistence with the reaction of D-galactono—y-lactone 10 with HBA, which resulted in

6-bromo-6-deoxy-D-galactono-1,4-1actone 13 (Scheme 4.4 B).

 
 
  

 

 

 
 

 

 

A) OPiv
HO Br
O1)HBr-H0Ac _ 0
2 MeOH, Ht'
Piv ) 85°/ HO OPIV
° 12
1)HBr-HOAc B o
= 63% Ref6
2) MeOH, H" 2
HO OH
13
C) CH

 
 

3.3 P' Cl. .
0 eq N py > Complicated pivaloation

products

 

HO ‘OH

Scheme 4.4 A) Synthesis of 6-bromo-6-deoxy-2,5-di-O-pivaloyl-D-galactono-1,4-lactone 12;
B) Synthesis of 6-bromo-6-deoxy-D-galactono-1,4-lactone 13 (Ref 6); C) selective
tripivaloation of D-galactono-1,4-1actone 10

As a net result of the pivaloyl rearrangement to facilitate bromination at a primary
position, we selectively protected the 2 and 5-positions of 6-bromo-6-deoxy-D-
galactono—1,4-lactone 13 using pivaloyl groups. The advantage of this selectivity
transformation could also be shown by comparing it with the direct selective
pivaloylation of D-galactono-l,4-lactone. Using 3.3-3.6 equivalent pivaloyl chloride in
pyridine, the tripivaloylation of D-galactono—l,4-lactone 10 was explored [18]. Their
results showed poor selectivity of D-galactono-1,4-lactone pivaloylation by formation of
a complicated mixture, especially involving between 3- and 5- positions. The 2,6-di-O-
pivaloyl-, 2,3,6-tri-O-pivaloyl-, 2,5,6-tri-O-pivaloy1- and 2,3,5,6-tetra-O-pivaloy1-D-

galactono-1,4-lactone was obtained in 30%, 2.6%, 9.7% and 21.6% yield respectively,

142

indicating a poor selectivity between the 3- and 5- positions but with a slight preference
for the S-O-position. In the case of selective pivaloylation of 6-bromo-6-deoxy-D-
galactono-l,4-1actone, the configuration remained unchanged. We expect the similar
selectivity except that the large primary bromo- group may further reduce the selectivity
for the 5-position over the 3-position. By the acidic t-buty1-1,3-dioxonium cation
rearrangement, the 5-O-pivaloyl group was rearranged from 6-position so the 5-0-
pivaloylation selectivity came directly from the much smaller steric hindrance of the
primary position. This would be apparently more selective than the direct pivaloylation of
6-bromo-6-deoxy-D-galactono-1,4-lactone 13, in which the selectivity is based on two

less differentiated secondary hydoxyl groups.

4.2.3 Synthesis of 2,6-dibromo-2,6-dideoxy-S-O-pivaloyl—D-manno—1,4-lactone

We also tried to expand this strategy to some other sugar lactones. D-glucono- I ,5-
lactone has long been used as a chiral starting material due to its cheap price and rich
chirality. We assumed monopivaloylation of the primary hydoxyl group and further
treatment with HBA could generate 2,6-dibromo-2,6-dideoxy-S-O-pivaloyl-D-manno-
1,4-1actone as a selectively functionalized and protected chiral building block, which
could be further employed in our synthetic work towards azasugars. Experimentally, the
selectivity of primary pivaloylation was shown to be relatively lower than in that in the
case of D-gulono-l,4-lactone. Both the primary and 2-position of D-glucono-l,5-lactone
were fairly active towards pivaloylation, therefore it was necessary to use more than 1
equivalent of acylation reagent to increase the yield. Although the yield was not high, the
cheap price of the starting material made the reaction practicable. The relatively poor

selectivity could be explained by comparing the different ring structures of D-gulono-1,4-

143

lactone 1 and D-glucono-l,5-lactone 14 (Figure 4.2). The primary hydoxyl group of D-
gulon0-1,4-1actone is fairly far away from the ring structure while the 2-position is on a
five member-ring. The partially eclipse relationship between 2- and 3- positions will
highly increase the steric hindrance of 2-OH towards acylation. However, for D-glucono-
1,5-lactone, the primary position is relatively close to the ring structure, which will
decrease its acylation rate. Furthermore, the six-membered ring has a near-chair
conformation, in which the 2,3-trans hydroxyl groups have relatively low interaction. As

a result, the 2-OH is relatively more active than that in the D-gulono-1,4-1actone.

OH C-6 relatively high
9y / steric hindrance
O /'\
H O

HO 0
C-6 low steric HO) (DH
hindrance 1
- ' ' C-2 relatively low
1113112912: tenc steric hindrance

D-gulono-1.4-lactone 1 D-glucono-1.5-Iactone 14

Figure 4.2 A comparison of the different ring structures of D-gu10n0-1,4-lactone and D-glucono-
1,5-1actone

Without separation, the 6-O-pivaloyl-D-glucono-1,5-lactone was then subjected
to HBA treatment. As expected, after work-up, 2,6-dibromo-2,6-dideoxy-5-O-pivaloyl-
D-manno-1,4-lactone 17 was synthesized in a 40% yield over 2 steps (Scheme 4.5). The
relatively low yield could be understood by exploring the mechanism of this reaction. An
acyloxonium cation was believed to be the active intermediate, which would facilitate
introduction of the bromo group by acting as an electrophile. However, an important

structure aspect of D-glucono-1,5-1actone is the six-membered lactone ring. To form the

144

5,6-acyloxonium cation 16, the lactone ring has to be opened. Furthermore, the 2,3-trans
configuration of D-glucono-S-lactone also prevents the 2,3-acyloxonium cation
formation. To form this 2,3-acyloxonium cation and then introduce a bromo group into 2-
position, the lactone ring also has to be opened. After the bromo groups are introduced
into the structure, the lactone ring cyclizes again to form the final product. The necessary
ring opening process substantially retarded the bromination reaction and increased the
amount of byproduct. A serious side reaction was believed to be the intramolecular attack
of the 5,6-acyloxonium cation 18 by 3-OH, which would generate 3,6-anhydro-5-O-

pivaloyl-D-glucono-l ,4-lactone 19.

OH OPiv

 

 

A) 0 63 t-Bu
o PivCl, py. o HBr-HOAc Y
HO —» H0 : HO 0 e
”0 \o ”O \0 HO 0
0H 15 OH 18 01-1
0
14 iHBr-HOAc i
OPiv Br”! 0 GD t-Bu
0 Y PivO 1'1 0
Br , O ’ 0
Ho 0 MeOH, H HoO
40% over > (“00 ’9,
HO Br 2 steps 63 O OH
17 16 Br 19
8) OH OH
0 HBr-HOAc 3' \, O 0
”9,0 = H 50% Ref 14
\0
OH HO Br
14 20

Scheme 4.5 A) Synthesis of 2,6-dibromo-2,6-dideoxy-S-O-pivaloyl-D-manno-1 ,4-lactone;
B) Synthesis of 2,6-dibromo-2,6-dideoxy-D-manno-1 ,4-lactone (Ref 14)

The relatively low yield of this bromination reaction was also shown in the direct
dibromination of D-qucono-1,5-lactone 14 with HBA [14]. Despite the high yields of the

dibromination reaction of D-gulono-I,4-lactone (90%) and D-manno-l,4-lactone (80%),

145

the dibromination of D-glucono-I,5-lactone to 2,6-dibromo-2,6-dideoxy—D-manno-1,4-

lactone 20 proceeded in only 50% yield.

In our synthesis of 2,6-dibromo-2,6-dideoxy-5-O-pivaloyl-D-manno-1,4-1actone
17, acyloxonium cation formation was employed to introduce the bromo functionality.
Migration of the pivaloyl group to the 5-O-position transferred the pivaloylation
selectivity at the primary position to that site. Although in this case, the selectivity of
primary pivaloylation was not extremely high, the concept of transferring the selectivity
among carbohydrate hydroxyl groups was interesting and important. We also compared
our strategy with the preparation of 2,6-dibromo-2,6-dideoxy-S-O-pivaloyl-D-manno-
1,4-lactone 17 by direct selectively pivaloylation of 2,6-dibromo-2,6-dideoxy-D-manno-
1,4-1actone 20 in 60% yield [19]. Although the overall yield from D-glucono-1,5-lactone
should be around the same range, in this case the procedure of direct selective
monopivaloylation of dibromo-D-manno-y-lactone was preferred due to its easier

isolation process.

4.3 Synthesis of chiral tetrasubstituted tetrahydrofuran (T HF) ring
from 2,6-dibromo-2,6-dideoxy-S-O-pivaloyl-D-idono-1,4-1actone by a

weak-base induced pivaloyl rearrangement

In previous discussed bromination coupled with pivaloyl group rearrangement,
2,6-dibromo-2,6-dideoxy-S-O-pivaloyl-D-idono-1,4-lactone 4 could be prepared
relatively easily from D-gulono-1,4-lactone 1. This compound was then treated with
dimethoxyrnethane and phosphorus pentoxide (P205) to install a methoxymethyl (MOM)

protecting group at the 3-0- position. This orthogonally protected 2,6-dibromo-2,6-

146

dideoxy-D-idono-y-lactone 21 was then treated with ammonium hydroxide in methanol
for 6 hours. Unexpectedly, NMR analysis of the product showed that only the 2-bromo
group was substituted while the primary bromo group stayed untouched. Both of the
protecting groups were still present. Further analysis showed the structure to be a
tetrasubstituted tetrahydrofuran ring 24 (Scheme 4.6). The mechanism was also shown in

Scheme 4.6.

 

o
OPiv . L
Br 0 Dimethoxy- OPIV NH4OH, 6‘ K 1'3“
H O methane, P2054; Br 0 O MeOH Br OH NHZ
§ 92% H H 0

H04 Br M°M621 '3' 1110111522 Br
0 OH /
BrVNH—z 65% B'W/g—QANHZ
c ‘— H O
PivO ’OMOM MOMO§ B71
24 23

Scheme 4.6 Synthesis of 2,S-anhydro-6-bromo-6-deoxy-3-O—methoxymethyl-S-O-pivaloyl-D-
gulonic amide 24

When the protected lactone 21 was treated with ammonia hydroxide, the lactone
ring was aminolyzed to an open-chain structure 22. Under the catalysis of weak base, the
free 4-OH attacked the pivaloyl carbonyl and the 5-O-pivaloy1 group was subsequently
rearranged to 4-position to form 23. Then free S-OH of 23 attacked the 2-bromo group
intramolecularly to furnish a tetrahydrofuran ring 24. The pivaloyl group rearrangement
was driven by an intramolecular ring formation under mild reaction conditions. Chiral
tetrahydrofuran rings are common structural subunits in natural products. Here, by a
weak-base catalyzed ring-switch process, we generated a chiral tetrahydrofuran ring

system fully substituted with complete differentiated functional groups and protecting

147

groups. This sugar amide could be used as a good chiral building block in more

complicated natural product synthesis.

By using some other starting material such as 2,6-dibromo-2,6-dideoxy-5-O-
pivaloyl-D-manno-l,4-lactone 17, we expected the same procedure would generate
similar tetrahydrofuran ring systems with different chiralities. This could enrich the
library of this fully substituted chiral building block. However, this was found out not to
be general. 2,6-Dibromo-2,6-dideoxy-5-O-pivaloyl-D-manno-1,4-lactone 17 was treated
with dimethoxyrnethane and P205 to afford 2,6-dibromo-2,6-dideoxy—3-O-
methoxymethyl-S-O-pivaloyl-D-manno-1,4-lactone in good yield (90%). When we
treated this compound with ammonium hydroxide in methanol, however, no
tetrahydrofuran ring product was formed. The only product isolated was simply the
aminolysis product with an open-chain structure (Scheme 4.7). Increasing the reaction
temperature led to a complicated mixture instead of the desired product, presumably due

to the elimination reaction.

 

OPiv OPiv OPiv
Dimethoxy- NH
Br 0 Br 0 NH4OH, Br OH 2
\\ O methane, P20: \\ O MGOH H“
H r H —> O
90% ~ 60°/
HO Br MOMO 3, ° MOMO 26 Br
17 25

Scheme 4.7 Synthesis of 6-bromo-6-deoxy-3-O-methoxymethyl-5-O-pivaloyl-D-
mannonamide 26

148

OPiv But
-0
HC) H ‘JT“()

0 H
CHzBf
H CHzBr
G3 |

22 H c, 22'

O'
OPiv OH 0+‘Bu
/O
H

Br 0H ““2 c, H H
H‘\\ o
H
MOMO Br H H B
C 2 r GHzBr
c3

0
26 26'

 

Figure 4.3 The comparison of the configurations of 22 and 26: the threo-selectivity of the weak
base catalyzed pivaloyl rearrangement

Although the result was somewhat disappointing, it was easily understood by
comparing the configurations of the two substrates (Figure 4.3). While the 2,6-dibromo-
D-idono-y-lactone has C4, C5-threo configuration, the 2,6-dibromo-D-manno-y-Iactone
has a C4, CS-erythro configuration. The difference between these two configurations
towards acyl rearrangement could be seen from two points of view. We first analyzed the
structures of open-chain intermediate after aminolysis. For C4, C5-threo configuration,
the 4-OH is very close to the 5-O-pivaloyl carbonyl, which indicates a small energy
barrier of nucleophilic attack and therefore fast reaction. On the other hand, for C4, C5-
erythro configuration, the 4-OH is fairly far away from the 5-O-pivaloyl carbonyl. The
C4-C5 bond must rotate to a disfavored conformation to allow the 4-OH to attack the 5-
O-pivaloyl carbonyl. This indicates a big energy barrier and therefore slower reaction.
Considering the energy profile of the orthoester intermediate, we can obtain the same

result. If the open chain amide has a C4, CS-threo configuration, i.e. 22’, the interaction

149

between C4,C5—substitution groups is relatively small. However, for the C4, CS—erythro
26’, the interaction between C4, C5-substitution groups could be much larger. The energy
of the orthoester intermediate would be high enough to prevent the rearrangement from
going. As a result, the rearrangrnent of acyl group proceeds faster and smoother if the two
neighboring groups have threo- configuration compared to that if the two neighboring
groups have erythro- configuration. This configuration analysis could also be used in

some other base-catalyzed acyl rearrangement.

4.4 Anomeric bromination and Ritter-type reaction of pentaacyl

monosacchrides

Encouraged by the t-butyl-l,3-dioxonium cation facilitated rearrangement and
selective bromination under acidic conditions, we further explored its application of some
other sugar derivatives. Tetraacyl glycosyl bromides have been prepared by treating
monosacchrides pentaacetates and pentabenzoates with hydrogen bromide in acetic acid.
The mechanism has been postulated to go through an intermediate with an anomeric
positive charge. These glycosyl bromides have been generally used as good glycosyl
donors in oligasaccrides synthesis. However, in some cases, the neighboring acyl group
participation could lead to undesired side-reactions, especially the orthoester formation
(Scheme 4.8). Due to its weaker orthoester formation, tetrapivaloyl glycosyl bromides
have been used as glycosyl donors more and more. Based on the stronger ability of
stablizing the anomeric positive charge, we presumed that the treatment of glucose
pentapivaloate with hydrogen bromide in acetic acid should generate tetrapivaloyl

glucosyl bromide.

150

OAc OAc
O
O HBr-AcOH O AcO
AcO AcO
Aflk —-> Afifi ——» MO 0

i Br R8 V0

OAc
OAc O
O - AcO
AcO R0
AcO O ‘— AcO r AGO
AcO OR
OAc

00
V30 R07?

Scheme 4.8 Tetraacetyl glucosyl bromide synthesis and application in glycoside synthesis

4.4.1 Synthesis of glucose pentapivaloate and its anomeric isomerization under

acidic condition

When a-D-glucose 27 was treated with excess pivaloyl chloride in pyridine under
room temperature, the major product obtained was glucose pentapivaloate. However,
there was still serious amount of partially pivaloated product present. Even after
extension of reaction times to several days, the yield of pentapivaloate was not improved.
The uncomplete pivaloylation suggested a great steric hindrance in the perpivaloylation.
Fortunately, by simply increasing the reaction temperature to 80 °C, most of the partially
pivaloylated compound was then pivaloylated completely with ignorable side reaction.
Unexpectedly, NMR of the product showed a big coupling constant between H1 and H2.
(J 1,2 = 8.4 Hz), which undoubtly suggested a B-configuration of this glucose
pentapivaloate. To exclude the temperature effect, the product obtained at room
temperature was also analyzed by NMR, which showed the same result. Using acetic

anhydride in pyridine at room temperature, the major peracetylation product of a-D-

lSl

glucose 27 is a-glucose pentaacetate because of a relatively slower mutaroation
compared with the acylation reaction. However, in the reaction of glucose with pivaloyl
chloride in pyridine, the mutaroation became relatively faster than the acylation,

therefore, B-D-glucose pentpivaloate 29 was obtained as the dominant isomer (Scheme

4.9).
OH 0“ OPiv
o . 0/
HO PIVCI, Py. HO 0 90,0 . O
HO ——’ Ho OH '—’ P'ngozg OPiv
”20 0H 0” OPiv

100°C

Aczo, Py.
0°C to rt

Scheme 4.9 Synthesis of B-D-glucose pentapivaloate from a-D-glucose and the comparison
with peracetylation under different reaction conditions

Before we subjected glucose pentapivaloate to hydrogen bromide treatment, we
tried to selectively hydrolyze the anomeric pivaloyl protecting group to generate an
anomeric free glucose tetrapivaloate 32, which is useful in synthetic carbohydrate
chemistry. The acidic treatment of this B-glucose pentapivaloate 29 with water in
presence, however, did not yield the anomeric free glucose tetrapivaloate 32. This was
ascribed to the robustness of the pivaloyl function as a protecting group. On the other

hand, a very slight change in TLC properties was noticed and an NMR analysis of the

152

compound after acidic treatment was carried out to confirm any structural change. NMR
spectroscopy clearly showed a different compound with glucose pentapivaloate structure.
This compound showed a much smaller coupling constant between H1 and H2 (J 1,2 = 3.8
Hz), suggesting a-glucose pentapivaloate 33 was obtained. The ratio of or: B is about 10:1

(Scheme 4.10 A).

 

A) OPiv OPiv 0"“
MCCN, H20 0
. O - 0 H2304. 83% PivO
P'VSWO OH P'Vavo OPiv .— > PivO
OPiv OPiv PiVO -
32 29 33 OPIv
B) OH H+ 0“ HO
HO OMe
an —~=—- H. .. —~= .
HO OMe Ho OMe OH
34
1t “ 1t 3.
OH OH
0 HO
HO HOHo H O 0
(9 OMe H

Scheme 4. IO The acid catalyzed anomeric isomerization of A) B-D—glucosepentapivaloate 29
and B)2—deoxy-D-arabino-hexosides 34 (Ref 20)

The high efficient B to or isomerization at anomeric position was consistent with
other researcher's results. Wisniewski et. a1. [20] found that methyl 2-deoxy-D-arabino-
hexosides undergo acid-catalyzed isomerization and suggested a mechanism (Scheme
4.10 B). We believe the isomerization of B-glucose pentapivaloate to its a-anomer was
through the similar mechanism. The a- and B-glucosides very often show different
reactivity towards various reactions because of their different C-H bond orientation at the

anomeric position. The pivaloylation procedure we discussed above, combined with the

153

acid-catalyzed isomerization, could give fairly pure B- or a—anomer of glucose

pentapivaloate respectively. It could be synthetically useful in the future.

4.4.2 Synthesis of tetrapivaloyl glycosyl bromides with gluco- and manno-

 

configurations
. OPiv
OPN O
HBT’HOAC, CHZCIZ P O 0 Br
Pivo/§$/ . : IVPivo / '
PivO OPIv
OPiv O
29 38 t-Bué
OPiv 0P", B OPlV
O O t. r 87°/ Fr 0 0
' _ PivO ___,° N ,
”49m Br PivO . 5 PW
OPlV P'VO Br PIVO Bf
39 4o 41

Scheme 4.1 1 Synthesis of a-tetrapivaloyl glucosyl bromide

We then treated B-glucose pentapivaloate 29 with hydrogen bromide in acetic
acid and dichloromethane. Only one product was obtained. NMR analysis of the product
showed it to be tetrapivaloyl glucosyl bromide. The coupling constant analysis showed a
small coupling constant, suggesting the product to be tetrapivaloyl a-glucosyl bromide. A
plausible reaction mechanism is shown in Scheme 4.11. Under acidic conditions, the
anomeric pivaloyl group would leave to form a positive charge, which would then be
stabilized by the neighboring pivaloyl group through a t-butyl-l,3-dioxonium cation 38.
Bromide ion would attack the anomeric position to open the t-butyl- l ,3-dioxonium cation
ring and form the B-glucosyl bromide 39. Due to the anomeric effect, the B-glucosyl
bromide 39 was isomerized to its a-anomer 40, which is commonly used as a glucosyl

donor in oligosaccharides synthesis.

154

To expand this strategy, we also employed D-mannose as the starting material in
this precedure. When D-mannose 42 was treated with pivaloyl chloride in pyridine using
the same reaction procedure, two isomers of mannose pentapivaloate 43B and 44a were
obtained. The ratio of azfl was about 1: 6 with B-isomer as the major product. The
mixture 43 and 44 was then treated with hydrogen bromide in acetic acid and
dichloromethane without further separation to form almost exclusively tetrapivaloyl 0L-
mannosyl bromide (Scheme 4.12). Under strong acidic conditions, the anomeric pivaloyl
group would leave to form a positive charge at the anomeric position, which would then
be stabilized by t-butyl-l,3-dioxonium cation formation with 2-O-pivaloyl group. Both
a— and B-isomers of tetrapivaloyl mannosyl bromide would yield the same t-butyl-1,3-
dioxonium cation 45. Subsequently, the bromide ion would attack the anomeric position
to form a tetrapivaloyl a-mannosyl bromide 46. Due to anomeric effect, the

isomerization from or to B was not a preferred transformation.

 

OH OPiv OPiv
OH . OPiv OPiv
Ho -0 PwCI.Py._ Pivg O ‘0 0in + PivO '0
HQ 90% IV PIvO
43 44 .
42 0” \ HBr-HOAc / 0"”

Piv69 t-Bu

OPiv O
OPiv ng
PivO ‘0 86% PivO ‘ <0
PiVO ‘_— FWD

9)

46 Br 45 Br

Scheme 4. l 2 Synthesis of a-tetrapivaloyl mannosyl bromide

155

4.4.3 Synthesis of glycosyl amide by dioxonium cation facilitated anomeric Ritter-

type reaction

The synthesis of tetrapivaloyl glycosyl bromides encouraged us to further explore
the reactivity of the acyloxonium cation stabilized anomeric positive charge as an
electrophile. Theoretically, a nucleophile could attack the anomeric position to form
glycosyl derivatives with different connections. If a nitrile is present, it could attack the
anomeric positive charge to form an N-glycosyl amide, which would be formally an
anomeric Ritter-type reaction. When B-glucose pentapivaloate 29 was treated with 30%
methanesulfonic acid in acetonitrile, tetrapivaloyl a— (48) and B-glucosyl amide (47) was
isolated as the only products with a ratio of or: B = 1: 6 (Scheme 4.13). Due to the
participation of 2-O-pivaloyl group in this reaction, the attack of the anomeric position by
the nitrile group would generate a B-glucosyl amide and the a-anomer was formed by

anomeric isomerization.

 

OPiv OPiv
N?CCH3
. o . 0 H o
p PIvO 2
'vgivo 0in MeCN, MsOH: PM) a) ———>
. 66%
OPIv
38 O
29 t-Bu%
OPiv OPiv
. O . 0
PVC PIvO
' pivo NHAc <—-——’ PivO
' PivO
47 OPIv 48 NH Ac

Scheme 4.13 Synthesis of tetrapivaloyl B- and a-glucosyl acetamides

We then tested if acetyl protecting groups could also be employed in this

anomeric Ritter-type reaction. a-Glucose pentaacetate and a-galactose pentaacetate was

156

treated with 30% methanesulfonic acid in acetonitrile respectively. The experimental
results also showed clean Ritter-type reactions, yielding tetraacetate N-glycosyl amides
of gluco- and galacto- configurations. For gluco- configuration, the ratio of 51a: 51B was

1:7. For galacto-configuration, the ratio of 54a: 54B was 1:8 respectively (Scheme 4.14).

 

 

OAc OAc CCH 0A0
MeCN, MsOH o * 3 o
AcO 0 t AcO . “0%
Acgéfiz CW: 82% Agéa? AcO OAc ”HA0
C
‘9 5° 4; 51 a:B = 1:7
OAcOAc °A°0Ac CCH Aco 0A0
o MeCN, MsOH o " 3 0
AcO OAc 86% > A00 '2 _’ A00 NHAC
OAc OAc
52 53 J,
63 54 azB = 1:8

Scheme 4.14 Synthesis of tetraacetyl N-B- and a-glycosyl amides with gluco- and galacto-
configurations

N-Glycosyl amines and amides have long been an important subject in
carbohydrate chemistry because of their prominence in glycopeptides and glycoproteins
and their potential as selective RNA-binding agents [21, 22]. The recent discovery of N-
glycosyl amide activation further opened the possibility of N-glycosyl amide being not
only an anomeric protected saccharide, but also a glycosyl donor [23]. Many synthetic
methodologies have been discovered in the synthesis of N-glycosyl amides and amines
[24-32]. Among them, the work of Fraser-Reid using n-pentenyl glycosides was one of
the prominent ones [29-32]. In the oxidative cleavage of n-pentenyl glycosides, a positive
charge at the anomeric position was formed and attacked by a nitrile to give N-glycosyl
amides. Our easy synthesis of tetraacyl N-glycosyl amides discussed above has the
mechanistic similarity of using anomeric Ritter-type reaction. Its simplicity defines its

significance in synthetic carbohydrate chemistry.

157

4.5 Dioxonium cation in protected acyclic polyols: stereoselective

Ritter-type reactions

In our synthesis of N-glycosyl amides using anomeric Ritter-type reaction, both
the ring oxygen and the neighboring acyl group had stabilizing effects on the anomeric
positive charge, which drove the reaction forward smoothly. However, an unsolved
question was if the acyloxonium cation can provide enough stabilization to allow the
Ritter-type reaction to take place. To clarify this question and to test the generality of this
reaction, a series of acyclic polyols were selectively protected and employed in the

acyloxonium cation facilitated functional groups transformation.

4.5.1 t-Butyl-1,3-dioxonium cation as an electrophile: synthesis of N-acyl-

aminopropanediol

As the simplest acyclic ployols, glycerol 55 was selected to test our hypothesis.
Because its simple 3-carbon chain, it would be impossible for it to cyclize
intramolecularly to form a ring system under strong acidic conditions. Glycerol 55 was
treated with 2.0 equivalent pivaloyl chloride to protect two hydroxyl groups. Two regio-
isomers were obtained, which were assigned to be 1,3-di-O-pivaloyl glycerol 56 and (i)-
1,2-di-O-pivaloy1 glycerol 57 respectively with a ratio of 2:1. It was assumed that under
treatment of acid, they would form the same t-butyl-l,3-dioxonium cation species.
Without separation of the two compounds, they were treated with 30% methanesulfonic
acid in acetonitrile at room temperature for three days. TLC analysis showed clean and
complete reaction. NMR analysis confirmed the Ritter-type reaction product 59. A

mechanism is proposed in Scheme 4.15.

158

OH OH OPiv

2e PivCl. .
HOWOH q Py> PivO\)\/0Piv + PiV0\)\/OH

55%

55 55 \ / (:)—57
MeCN, MsOH
t-Bu
e
o

 

OPiv ’<
80% (
PivO\/‘\/NHAC <—— PivO o
k NICCHg
(:)—59 (i)—58

Scheme 4.15 synthesis of (i)—N-acyl-aminopropanediol 59

Under treatment with strong acid, both of the two glycerol dipivaloates would
form the same t-butyl-1,3-dioxonium cation species 58. This species was subsequently
attacked by the acetonitrile in the primary position. After attack of the nitrilinium carbon
by water generated in-situ and further proton transfer, the (:l:)-3-N-acetylamino-3-deoxy—
1,2-di-O-pivaloyl-glycerol 59 was obtained. The reaction was fairly clean and no other

serious side reaction could be detected.

Usually, for a Ritter—type reaction, the carbocation has to be substantially
stabilized and then attacked by a nitrile group. The nitrile group usually attacks the most
positively charged position. In our case, the possible reactive sites partially positively
charged were the carboxonium cation, the secondary C-2 and the primary C-l in
decreasing order (Figure 4.4). The nucleophilic attack to these positively charged
positions would generate different species. However, due to the steric hindrance, only the
primary position was possible to be attacked by the acetonitrile. This highly entropy-
controlled reaction was in agreement with the NBS-promoted pivaloyl rearrangement
discussed in Chapter 2. As a result, a protected (:t)-bromopropanediol 59 was
successfully synthesized as a racemic mixture, which could serve as a useful building

block for some synthetic work.

159

Nu:
a\ -Bu

.,<
Pix/0&0
‘7’ 8
. 5" fb

Nu:

.7 bl V;
AcHN
o l it-Bu OPiv NHAc
PivO\/‘\/ NHAc PivO\/i\/0Piv
PivO o

59 61
60

O

4.

Observed

Figure 4.4 The preference of nucleophilic attack of the t-butoxonium cation 58: to a) the
carboxonium cation; b) the secondary C-2; and c) the primary C-l .

4.5.2 The stereoselectivity of t-butyl-lJ-dioxonium cation facilitated Ritter-type
reaction: Synthesis of protected chiral bromobutanetriol and N-

acetylaminobutanetriol

As we discussed in Figure 4.4, the reaction mechanism for the dioxonium cation
facilitated amidation suggested that this reaction to be highly stereoselective and
regioselective. However, by using achiral glycerol as starting material, the
stereoselectivity of this reaction could not be confirmed due to the single chirality and
racemic property of the product. To clarify this question, we further explored these
reactions employing mesa-erythritol 62 as our starting material. Theoretically, by looking
at the relative chirality change between the two existing chiral centers, the mechanism

could be further clarified and confirmed.

Mesa-erythritol 62 was partially protected by treatment with 3.0 equivalent

pivaloyl chloride in pyridine. An easy chromatography separation was carried out to

160

separate the (i)-l,2,4-tri-O-pivaloyl-erythritol 63 as a racemic mixture. Interestingly,
only trace amount of (i)-l,2,3—tri-O—pivaloyl-erythritol was found. Compared with the
dipivaloylation of glycerol, this tripivaloylation of erythritol showed much better
selectivity. This better selectivity was also ascribed to the higher steric hindrance of the
erythritol tripivaloates. When (i)-1,2,4—tri-O-pivaloy1-erythritol 63 was treated with
hydrogen bromide in acetic acid and dichloromethane, (dz)-l-bromo-l-deoxy—2,3,4—tri-O-
trimethylacetyl-erythritol 64 was synthesized in a 75% yield (Scheme 4.16). The
plausible mechanism included a dioxonium cation formation followed by the bromide
attack at the primary position to furnish a protected bromobutanetiol. NMR analysis of
the product showed an erythro- configuration. No threo- configuration products were

found, suggesting that this reaction was stereospecific.

OPiv

“Ox —->Pivc" Py' PivO\/c'):v/\ w PIVO\/'\é/.\Br
; 0H 60% , OPiv 75% (”‘5‘ OP'V
5” 5H OPiv
62 (1)43 MeCN, PivO
MsOH ; NHAc
OPiv
(:»65

Scheme 4.16 Synthesis of protected chiral bromobutanetriol (i)-64

We then treated (2t)-l,2,4-tri-O-pivaloyl erythritol 63 with 30% methanesulfonic
acid in acetonitrile. We expected a mechanism that was similar to the HBA treatment of
(i)l,2,4-tri-O-pivaloy1 erythritol. This would generate a Ritter-type product (i)-4-N-
acetylamino-4-deoxy— l ,2,3-tri-O-pivaloyl-erythritol 65. However, even after extension of
the reaction time to several days, only a complicated mixture was obtained. NMR

analysis of the mixture showed only small amount of Ritter—type product 65,

161

characterized by the signals of the N-acetylamide group. On the other hand, NMR
suggested that some methanesulfonyl groups were introduced into the polyol structure.
Under strong acidic conditions, the acid itself served as a competitive nucleophile and
attacked the t-buty1-1,3-dioxonium cation or the carbon chain instead of acetonitrile. As a
result, only small amount of Ritter-type product was formed. Although disappointed by
the result, we reasoned that by substituting the protic acid catalyst to some Lewis acid,
this problem could be solved. When we treated (:t)-l,2,4-tri-O-pivaloy1-erythritol 63 with
3:1 acetonitrile and boron trifluoride etherate at 50 °C for 5 hours, most of the starting
material was consumed. Afier simple work-up by neutralization with sodium bicarbonate,
the Ritter-type product was obtained in 78% yield. NMR analysis of the product showed
(i)-4-N-acetylamino-4-deoxy- l ,2,3-tri-O-pivaloyl-erythritol 65a and (:t)-4-N-
acetylamino-4-deoxy-l,2,3-tri-O-pivaloyl-threitol 65b (Scheme 4.17). The ratio of

erythro- : threo- configuration was about 5:1.

 

OPiv OPiv OPiv
PiV0\/'\_/\OPiv MeCNv BF3OE‘2 = PivO\/'\/\ PivO
i o o i NHAc + NHAc
0“ 5° C' 78" OPiv OPiv
(i)-63 (i)—653 (i)—65b

Erythro- : Threo- = 5: 1

Scheme 4.17 Synthesis of protected chiral N-acetylaminobutanetriol (i)—65

In this reaction, No 3-N-acetylamino product and only small amount threo-
configuration product was detected. This suggested that this t-butyl-1,3-dioxonium cation
facilitated Ritter-reaction was not only highly regioselective, but also stereoselective. Due
to its carbocation intermediate mechanism, Ritter-type reaction usually suffers from the

unspecific stereocontrol. This has greatly limited its application in synthetic organic

162

chemistry (Figure 4.5). In our dioxonium cation facilitated Ritter-type reaction, the
carbocation was gerenated from an achiral protecting group. This carbocation in turn
activated the carbon chain without scrambling its chirality. Afier nucleophile introduction,
this carbocation was transformed back to the protecting group. This strategy successfully

solved this stereocontrol issue and could be very useful in the future synthetic work.

A) R

1 * H+ R1 RCN R1 R1
>—0H—> >69 ———> >—g_=_CR-—" >—NHCOR
R2 R R2

R2 2

OH o——<§ NHCOR OCOR
OCOR + NHCOR
R “(\(OCOR _H_, R17,'\‘...\\\0 _. R1 ...\\\ R1 .m
1 .
ij R2 R2
R2 RCN

Figure 4.5 The comparison of A) normal Ritter-type reaction and B) acyloxonium cation-
assisted Ritter-type reaction

4.5.3 Dioxonium cation facilitated functional group transformation of (S)-1,2,4-

butanetriol: more mechanism and applications

Although we have demonstrated that the dioxonium cation facilitated bromination
and Ritter-type reaction are highly regioselective and stereoselective, there remained one
issue about the intermediate carboxoniun cation to be addressed. Was the five-membered
ring or the six-membered ring carboxonium cation the real active intermediate? Although
generally the five-membered ring has been suggested as the active intermediate because
of the kinetic effect, there has been no direct evidence excluding the six-membered
carboxonium cation intermediate. We therefore reasoned that by employing (S)-I,2,4-

butanetriol 66 as the starting material, this issue could be easily and unambiguously

163

addressed (Figure 4.6). (S)-l,2,4-butanetriol 66 was a commercially available chiral
building block. Its simple chiral structure has made it a perfect platform for synthesis of
complex chiral molecules. The functional group transformations of this molecule, which

would generate new diverse chiral building blocks, are of great practical interests.

.A
Nu. /\/\/OPiv 0in
O i
E ——> Nu
69 0 OPiv
OH 53

OPiv Nu:
OPiv
67 \ Piv0/\,/\) f

G) (0 Nu

0

Figure 4.6 Use of 1,2,4-butanetriol dipivaloate to clarify the active carboxonium cation
intermediate: five-membered ring or six-membered ring

(S)-1,2,4-Butanetriol was treated with 2 equivalent pivaloyl chloride in pyridine to
afford (S)-l,4-di-O-pivaloy1-1,2,4-butanetriol 67 in 60% yield. When we treated 67 with
hydrogen bromide in acetic acid and dichloromethane, a single product was obtained.
NMR analysis showed the structure to be 4-bromo-l,3-di-O-pivaloyl-butanediol 70
(Scheme 4.18). The product suggested that the reaction went through a five-membered
ring dioxonium cation intermediate 68 (Figure 4.6), which was attacked by bromide at
the primary position to yield the product. The six-membered ring dioxonium cation 69
could only be formed between 2-OH and 4-O-pivaloyl group. If this was the active
intermediate, the bromide would attack the 4-position to yield 4-bromo-l,2-di-O-

pivaloyl-butanediol. However, this compound was not detected. The experimental results

164

confirmed that under this reaction condition, five-membered ring dioxonium cation
intermediate was the active intermediate. Furthermore, the selectively protected and
functionalized product 4-bromo-l,3-di-O-pivaloyl-butanediol 70 could serve as a good

chiral synthon.

We then treated (.S')-1,4-di-O-pivaloyl-l,2,4-butanetriol 67 with 3:1 acetonitrile
and boron trifluoride etherate at 50°C for 5 hours, after which all the starting material was
consumed. NMR analysis showed a mixture of two isomers, namely 4-N-acetylamino-
1,3-di-O-pivaloyl-l,3-butanediol 71 and 4-N-acetylamino-l,2-di-O-pivaloyl-l,2-

butanediol 72 in a ratio of 2:1 (Scheme 4.18).

 

 

OH . OH '
Perl, . OPIV
HON Py PivO\/'\/\ _ HBr-HOACA Br
0” 60% 0P” 76% ' ' OPiv
66 67 70

MeCN, BFgOEtz, 50°C 1 88%

OPiV OPiv
H .
AC N\/'\/\0Piv + PIVO NH Ac
71 2 : 1 72

Scheme 4.18 Synthesis of protected chiral N-acetylaminobutanediol 71 and 72

The unexpected Ritter-type reaction result was not consistent with the reaction
with HBA, which generated a single regioisomer 70. Undoubtedly, the Lewis acid
promoted Ritter-type reaction was through two active intermediates, the five-membered
ring l,2-t-butyl-1,3-dioxonium cation 68 and the six-membered ring 2,4-t-butyl-l,3-
dioxonium cation 69. Being attacked by acetonitrile in the 1- or 4- primary positions
respectively, they generated 71 and 72 respectively. The 2:1 ratio showed that the five-

membered ring intermediate was favored. To exclude the temperature effect, we also

165

tried this reaction at room temperature. The Ritter-type reaction proceeded slowly under
room temperature. After several days, the reaction was worked up and separated. It also

yielded two isomers 71 and 72 with essentially the same ratio.

OH

-1

’//;1 67 X

./\
NU. /\/\/OP1V . f
0 & PIvO i
E. k z
69 o __3_. 0 6) (0
‘—_
“'3 69
68
k5
HBAl ka 1 k6
OPiv OPiv NHAc
PivO PivO PivO
70 71 72
Br AcHN PivO

Figure 4.7 A explanation of the regioselectivities of acyloxonium cation facilitated
bromination and the Ritter-type reaction

Compared with the dioxonium cation facilitated selective bromination and the
Ritter-type reaction, a plausible explanation was suggested in Figure 4.7. Presumably, the
five-membered ring l,2-t-butyl-l,3-dioxonium cation 68 was kinetically much more
favored than the six-membered ring 2,4—t-butyl-1,3-dioxonium cation 69 (k. >> k2), while
the latter was thermodynamically slightly more favored than the former. In the

bromination reaction, the bromide ion was a good nucleophile, which made the

166

bromination process relatively fast (k4 > k2 & k3). The kinetic product 4-bromo-l,3-di-O-
pivaloyl-1,3-butanediol 70 was formed. However, the acetonitrile was not a nucleophile
as good as bromide ion. This made the attack by acetonitrile in the Ritter-type reaction
very slow (k5, k6 < k., k2, k3, k.;). This step became the speed-limiting step. Under this
circumstance, it was possible that the equilibrium between the five-membered ring t-
butyl-l,3-dioxonium cation 68 the six-membered ring t-butyl-l,3-dioxonium cation 69
was formed. Both of them then acted as the active intermediate to afford a mixture of two
regioisomers. This analysis was further complicated by the fact that the rates of the
dioxonium cation intermediates towards amidation were different for five-membered ring
and six-membered ring. Presumably, the five-membered ring should have a faster rate
towards amidation than the six-membered ring (k5 > k6) due to the higher ring strain.

Presumably, the 2:1 ratio of the two isomers reflected all the factors discussed above.

As a result, two regioisomers 71 and 72 of the Ritter-type reaction was obtained
when (S)-1,4-di-O-pivaloyl-l,2,4-butanetriol 67 was treated with acetonitrile and Lewis
acid. Both isomers could be employed as good chiral building blocks towards more

complex structures.

4.6 Synthetic application of dioxonium cation facilitated Ritter-type

reaction: synthesis of protected 6-aminogalactonolactone

Because of the stereospecificity and high regioselectivity towards primary
position of dioxonium cation facilitated Ritter-type reaction, we speculated that it could

be a good methodology to install an amide group at the primary position of

aldonolactones, which has been a very important task in synthetic carbohydrate chemistry.

167

To test our hypothesis, D-galactono-y-lactone was selected as our substrate because of its

readily accessibility and relatively cheap price.

D-galactono-l,4-lactone 10 was treated with 1 equivalent t-butyldimethylsilyl
chloride (TBDMSCl) in pyrindine to selectively protect the 6-position. After this reaction
was done, acetic anhydride was added into the reaction mixture to protecte the other free
hydroxyl groups. After the workup procedure, the product, majorly 2,3,5-tri-O-acetyl-6-
O-t-butyldimethylsilyl-D-galactono-1,4-lactone 73 was treated with acid to selectively
hydrolyze the 6-O-TBDMS protecting groups. These easy protection and deprotection
steps could be easily carried out without separation between any two steps. Finally, 2,3,5-

tri-O-acetyl-D-galactono- 1 ,4-lactone 74 was successfully synthesized.

Compound 74 was treated with Lewis acid in acetonitrile using the similar
reaction conditions discussed above. This reaction was relatively slow compared with the
reactions of partially protected acyclic polyols, which was ascribed to the higher steric
hindrance. After 8 hours, workup of a portion of the reaction solution gave two major
compounds well separated on TLC. NMR analysis showed one of the compounds was 6-
N-acetylamino-6-deoxy-2,3,S-tri-O-acetyl-D- galactono-1,4-1actone 75, i.e. the expected
Ritter-type reaction product. The other compound was shown to be 2,3,5,6-tetra-O-
acetyl-D-galactono-l,4-lactone 76 by NMR analysis. A mechanism was also shown in
Scheme 4.19. Under acidic conditions, nucleophilic addition of the free 6-position
hydroxyl group to acetonitrile would generate iminoester species 77. After workup with
water, the iminoester 77 was hydrolyzed to form an ester. Under our reaction condition,
this reaction could be a serious side reaction competing with Ritter-type reaction.

Fortunately, as the reaction time was extended to 20 hours, most 2,3,5,6-tetra-O-acetyl-

168

D-galactono-l,4-lactone 76 was gone and the Ritter-type product 75 became the

dominant product.

TBDMSO HO
mfif‘) 1)TBDMSCI, Py. 0 352%i2'
’20 2) ACZO Py. ' ADC ’2

 

 

o ’4,
0H 75% 0A0 88 /° Aco7 OAc
HN
AcHN OAcO Y OAc;O
BF3OEt2, MBCN 0 H2O H O
65°C, 2011 4, 2’0 ,2
58% A00 s’OAc AcO OAc AcO OAc

Scheme 4.19 Application of acyloxonium cation facilitated Ritter—type reaction: synthesis of
protected 6-aminogalactonolactone

This reaction mode encouraged us to treat the 2,3,5,6-tetra-O-acetyl-D-galactono-
1,4-1actone 76 directly with Lewis acid in acetonitrile. Presumably, the hydrolysis of 6-
position acetyl group may result in the formation of acetoxonium cation and then Ritter-
type amidation. However, even after several days, there is not any Ritter-type product
generated while the starting material stayed unchanged. This result also confirmed that
the 2,3,5,6-tetra-O-acetyl-D-galactono-1,4-lactone 76 was largely formed in the work-up

procedure.

Aminosugar has long been an important subject for synthetic organic chemistry.
The synthesis of aminosugar often involved expensive reagents, extensive protection and
deprotection and troublesome separation. It was of interest to develop new synthetic
methodologies towards different aminosugars. Our synthesis of 6-N-acetylamino-6-

deoxy-2,3,S-tri-O-acetyl-D-galactono-1,4-lactone 75 only involved easy protection and

169

deprotection steps without any expensive reagents. It could be generated in the future as a
good alternative route towards primary aminosugars. Furthermore, the easy introduction
of nitrogen functionality could greatly facilitate our synthesis towards iminosugars if

combined with other appropriate functional group transformation.

170

EXPERIMENTAL

General Optical activity data were obtained on a JASCO P-1010 polarimeter at
25°C. NMR spectra were obtained on a Varian VXR-SOO Spectrometer operating at
500MHz for protons. Mass spectra were obtained on a JEOL HX-l lO-HF instrument
using fast atom bombardment as ionization mode. Spectra were recorded in the positive
ion mode. IR spectra were obtained on a Nicolet 710 spectrometer in chloroform solution

except when otherwise specified.

2,6-Dibromo-2,6-dideoxy-5-O-trimethy1acetyl-D-idono- l ,4-lactone (4):

D-gulono-l,4-lactone (2 g, 11.2 mmol) was dissolved in 10 mL pyridine and
cooled down to 0 0C. Trimethylacetyl chloride (1.49 g, 12.4 mmol) was added slowly.
The solution was stirred at 5 °C for 12 hours. Methanol (10 mL) was added in and the
mixture was subsequently evaporated to dryness. To the residue 30% hydrogen bromide
in acetic acid (10 mL) was added. The mixture was stirred for 10 hours under room
temperature and methanol (50 mL) was added. The mixture was stirred for 10 hours more
and evaporated to dryness. The residue was partitioned between chloroform (2 x 50 mL)
and saturated sodium bicarbonate solution (100 mL). The organic layer was dried and
evaporated. The crude product was purified by column chromatography to afford 2,6-
dibromo-2,6-dideoxy-5-O-trimethyacetyl -D-idono-l,4-lactone 4 as colorless oil (2.53 g,
58% over 2 steps). [a]D = -24.2° (c = 0.8, CHC13); 5H (500MHz; CDC13): 1.21 (9H, s,
C(Cfl3)3), 3.61 (2H, d, 11643112, 15.6» 4.3Hz, H-6, 6’), 4.27 (1H, d, 12.3 2.3Hz, H-2), 4.64
(1H, dd, 13.4 4.4Hz, H-3), 5.00 (1H, dd, J45 7.0Hz, H-4), 5.36 (1H, dt, H-S); 8c (125MHz;

CDC13): 26.95 (OCC(§H3)3), 29.65 (C-6), 38.97 (OC§(CH3)3), 41.79 (C-2), 69.44, 74.00,

171

81.34 (C-3, C-4, C-5), 171.46 (C=O in lactone), 177.81 (OQC(CH3)3); FT-IR: 1784,

1734, 1147 cm".
2,6-Dibromo-2,6-dideoxy—3-O-trimethylacety1-D-idono- l ,4-lactone (8):

2,6-Dibromo-2,6-dideoxy-D-idono-l,4-lactone (1.12 g, 3.7 mmol) was dissolved
in 5 mL pyridine and cooled down to -20 °C in ice-salt bath. Trimethylacetyl chloride
(0.49 g, 4.0 mmol) was added slowly. The solution was warmed up to 0°C in one hour
and stirred for 6 hours thereafter. The reaction mixture was poured into ice and stirred for
a while. Cold hydrochloric acid (1N) was added in slowly until the pH reached 2~3.
Chloroform (100 mL) was used to extract the mixture. The organic phase was washed
with 20 mL brine, 20ml water and dried. The crude product was purified by column
chromatography to give a major product 2,6-Dibromo-2,6-dideoxy-3-O-trimethylacetyl-
D-idono-1,4-lactone 8 (0.92 g, 64% yield). 8H(500MHz; CDC13): 1.22 (9H, s, C(C_113)3),
3.42 (1H, dd, 16,6. 10.5 Hz, J55 7.0Hz, H-6’), 3.47 (1H, dd, 15.6 4.8Hz, H-6), 3.97 (1H, ddd,
14,5 3.4Hz, H-S), 4.67 (1H, d, .123 6.1Hz, H-2), 4.91 (1H, dd, J33 6.1Hz, H-4), 5.49 (1H, t,
H-3); 6c ( 125MHz; CDCl;): 26.85, 26.88, 26.93 (OCC(_CH3)3), 32.97 (C-6), 38.91
(OC_C(CH3)3), 39.66 (C-2), 68.94, 76.08, 79.43 (C-3, C-4, C-5), 169.39 (C=O in lactone),

177.34 (OQC(CH3)3); FT-IR: 2965, 1786, 1728, 1133 cm".
6-Bromo-6-deoxy-2,5-di-O-trimethy1acetyl-D-ga1actono- 1 ,4-lactone (12):

To 2,6-di-O-trimethylacetyl-D-galactono-1,4-1actone (1.0 g, 2.9 mmol) 5 mL
hydrogen bromide (30% in acetic acid) was added. The mixture was stirred for 5 hours to
give a clear solution. Methanol (20 mL) was added and stirred for another 12 hours. The

solution was evaporated and the residue was partitioned between chloroform and sodium

172

bicarbonate solution. The organic layer was dried and evaporated to give 6-bromo-6-
deoxy—2,5-di—O-trimethy1acetyl-D-galactono-l,4-1actone 12 (1.01g, 85% yield).
Compound 12 was characterized as 3-O-acetate. 5H (500MHz; CDC13): 1.19, 1.25 (2 x
9H, 2 x s, C(C_H3)3), 2.06 (3H, s, COCH3), 3.48 (1H, dd, 16.5 10.2 Hz, 15,6~7.4Hz, H-6’),
3.49 (1H, dd, 15,6 6.3Hz, H-6), 4.78 (1H, dd, J45 2.6Hz, 13,4 7.4Hz, H-4), 5.16 (1H, ddd, H-
5), 5.38 (1H, t, 123 7.5Hz, H-3), 5.55(1H, d, H-2); 5c (125MHz; CDC13): 20.67 (COQH3),
26.96, 27.12, 27.21 (OCC(Q_H3)3), 27.70 (C-6), 38.86, 39.31 (OCQ(CH3)3), 69.87, 70.07,
72.08, 72.50 (C-2, C-3, C-4, C-5), 168.03, 169.75 (QOCH3, C=O in lactone), 177.01,

177.18 (09110193); FT-IR: 2966, 1786, 1734, 1698, 1147 cm".
2,6-Dibromo-2,6-dideoxy-5-O-trimethyacety1 -D—manno-1,4-lactone (17):

This compound was prepared using essentially the same procedure as the
preparation of 2,6-Dibromo-2,6-dideoxy-3-O-trimethyacetyl-D-idono-1,4-1actone 8.
Yield: 40% over 2 steps. 5H (500MHz; CDC13): 1.20 (9H, s, C(Cfl3)3), 3.22 (1H, b, -OH),
3.68 (1H, dd, JO“). 11.7 Hz, .155 4.2Hz, H-6’), 3.83 (1H, dd, 15.62.9112, H-6), 4.40 (1H, dd,
13.4 2.8Hz, .12.; 4.4Hz, H-3), 4.61 (1H, dd, 14,5 8.9Hz, H-4), 4.76 (1H, d, H-2), 5.23 (1H,
ddd, H-S); 5c (125MHz; CDC13): 26.89 (OCC(QH3)3), 32.03 (C-6), 39.00 (OC§(CH3)3),
47.22 (C-2), 67.96, 68.82, 79.11 (C-3, C-4, C-S), 169.22 (C=O in lactone), 177.84

(O_C_C(CH3)3); FT-IR: 2943, 1796, 1738, 1147 cm".

2,6-Dibromo-2,6-dideoxy-3-O-methoxymethy1-5-O-trimethyacetyl~D-idono-1 ,4-1actone

(21):

2,6-Dibromo-2,6-dideoxy-5-O-trimethylacetyl-D-idono-1,4-lactone (1.24 g, 3.2

mmol) was dissolved in 20 mL dimethoxyrnethane and 10 mL dichloromethane and 3 g

173

dry phosphorus pentoxide. The mixture was stirred at room temperature for 2 hours and
poured into 50 mL cold saturated sodium bicarbonate solution and stirred.
Dichloromethane (2 x 20 mL more) was used to extract the water phase. The organic
phase was collected and dried. Evaporation gave 2,6—dibromo-2,6—dideoxy-3-O-
methoxymethyl-5-O-pivaloyl-D-idono-1,4-lactone 21 (1.27g, 92% yield) as colorless oil.
5“ (500MHz; CDC13): 1.22 (9H, s, C(Cfl3)3), 3.41(3H, s, CHzOCfl3), 3.49 (1H, dd, 16,6»
11.0 Hz, 155 4.2Hz, H-6’), 3.57 (1H, dd, 15,6 6.1Hz, H-6), 4.47 (1H, d, .12.; 3.9Hz, H-2),
4.59 (1H, dd, 13,4 5.4Hz, H-3), 4.72 (1H, d, .1 ~ 7.1Hz, CflzOCH3), 4.76 (1H, d,
CHJOCH3), 5.12 (1H, t, J45 5.4Hz, H-4), 5.35 (1H, ddd, H-S); 5c (125MHz; CDC13):
27.27 (OCC(_C_H3)3), 29.00 (C-6), 39.13 (OCQ(CH3)3), 40.20 (C-2), 56.92 (CHZOQH3),
69.46, 79.76, 80.39 (C-3, C-4, C-5), 97.35 (QHZOCH3), 170.23 (C=O in lactone, C-l),

177.04 (OQC(CH3)3); FT-IR: 3099, 1772, 1624, 1170, 1058 cm".
2,5-Anhydro-6-bromo-6-deoxy-3-O-methoxymethyl-5-O-pivaloyl-D-gulonamide (24):

2,6-Dibromo-2,6-dideoxy-3-O-methoxymethyl-5-O-pivaloyl-D-id0no- 1 ,4-1actone
(0.56 g, 1.3 mmol) was dissolved in 10 mL methanol and cooled down to 0°C.
Ammonium hydroxide (2 ml, 28% solution in water) was added slowly. The solution was
warmed up to room temperature and stirred for 6 hours. The solution was evaporated to
dryness and partitioned between chloroform and water. The organic phase was dried and
purified by column chromatography to give 2,5-anhydro-6-bromo-6-deoxy-3-O-
methoxymethyl-S-O—pivaloyl-D-gulonamide 24 (0.31 g, 65% yield) as colorless oil. [a]D
= -61.8° (c = 0.8, CHC13); FAB-MS: m/z calcd. for C13szBrNO6 367.0631, found
368.0711 (MH+); 5H (500MHz; CDC13): 1.18 (9H, s, C(Cfl3)3), 3.40(3H, s, CHzOCfl3),

3.45 (1H, dd, .165 10.0 Hz, 15.57.8112, H-6’), 3.50 (1H, dd, 15,6 6.4Hz, H-6), 4.42 (1H, m,

174

H-3), 4.46 (1H, d, .123 1.5Hz, H-2), 4.57 (1H, ddd, 14,5 3.2Hz, H-5), 4.78 (1H, d, .1 = 6.8Hz,
CEZOCI-h), 4.80 (1H, d, CflzOCH3), 5.24 (1H, dd, J14 1.0Hz, H-4), 5.84 (1H, b,
CONfl’z), 6.59 (1H, b, CONflz); 5c (125MHz; CDC13): 26.72, 26.85, 27.02
(OCC(QH3)3), 27.10 (C-6), 38.86 (OCQ(CH3)3), 55.92 (CHZOQH3), 75.41, 81.13, 83.00,
84.23 (C-2, C—3, C-4, C-5), 95.89 (OQHZOCH3), 172.04 (_C_IONHZ, C-l), 176.75

(O_C_C(CH3)3); FT-IR: 3098, 2973, 1737, 1690, 1627, 1144, 1033 cm".
2,6-Dibromo-2,6-dideoxy-3-O-methoxymethyl-5-O-pivaloy1-D-manno- 1 ,4-1actone (25):

This compound was prepared using essentially the same procedure for the
preparation of 2,6-dibromo-2,6-dideoxy-3~O-methoxymethyl-5~O-trimethyacety1-D—
iodono-l,4-1actone 21. Yield: 90%; 8” (500MHz; CDC13): 1.22 (9H, s, C(CH_3)3), 3.35
(3H, s, CHZOCHJ), 3.68 (1H, dd, 16,66 11.9 Hz, .155 3.3Hz, H-6’), 3.90 (1H, dd, J55 2.9Hz,
H-6), 4.47 (1H, (1, J23 4.4Hz, J33 3.3Hz, H-3), 4.70 (1H, dd, 14,5 8.2Hz, H-4), 4.72 (2H, dd,
CflzOCH3), 4.75 (1H, d, H-2), 5.16 (1H, ddd, H-5); 5c (125MHz; CDC13): 26.96
(OCC(§H3)3), 31.61 (C-6), 38.99 (OCQ(CH3)3), 45.34 (C-2), 57.30 (CHZOQH3), 67.70,
76.09, 78.88 (C-3, C-4, C-S), 98.48 (OQHZOCH3), 169.17 (C=O in lactone, C-l), 176.80

(OQC(CH3)3); FT-IR: 2966, 1778, 1632, 1175, 1068 cm".
B-D-Glucose pentapivaloate (29):

a—D-glucose (10 g, 55.6 mmol) was dissolved in 100 mL pyridine and cooled
down to 0 °C in an ice bath. Trimethylacetyl chloride (46.9 g, 389 mmol) was added in 5
minutes. The ice bath was removed and the reaction mixture was sealed and heated to 80
°C. It was stirred at this temperature for 10 hours and poured into ice with sodium

bicarbonate. After vigorous stirring, the precipitate was collected by filtration and

175

thoroughly washed with distilled water. The white solid was then dried in vacuum to give
B-D-glucose pentapivaloate 29. The total yield is > 90%. M.p. 156-158°C; (lit. [33] mp.
156-158°C; [61],,25 + 109° (c = 1, CHC13)); 8H(500MHz; CDC13): 1.08, 1.09, 1.12, 1.14,
1.17 (5 x 9H, 5 x s, C(Cfl3)3), 3.83 (1H, ddd, J51, 2.2Hz, 15,7,~ 5.3Hz, 14,5 10.2Hz, H-S), 4.06
(1H, dd, JW 11.8 Hz, H-6’), 4.12 (1H, dd, H-6), 5.12 (1H, dd, 13.4 9.3Hz, H-4), 5.18 (1H,
dd, J 1,2 8.4Hz, 12.3 9.4Hz, H-2), 5.34 (1H, t, H-3), 5.67 (1H, d, H-l); 5c (125MHz;
CDC13): 27 .02, 27.23, 27.25, 27.28, 27.34 (OCC(_C_H3)3), 38.91, 38.96, 39.05
(OCQ(CH3)3), 61.48 (C-6), 67.80, 70.15, 72.39, 72.81 (C-2, C-3, C-4, C—5), 91.87 (C-l),

176.30, 176.38, 176.97, 177.01, 177.97 (OQC(CH3)3).

a-D-Glucosepentapivaloate (33):

B-D-glucosepentapivaloate 29 (1 g, 1.7 mmol) was partially dissolved in 20 mL
acetonitrile and 5 mL 37% hydrochloric acid. This mixture was warmed to 50 °C to
become a clear solution and stirred under this temperature for 12 hours. The solution was
then evaporated and the residue partitioned between chloroform and saturated sodium
bicarbonate solution. The organic layer was collected and dried. Evaporation yielded a-
D-glucose pentapivaloate 33. Yield: 0.83g, 83%. SH (500MHz; CDC13): 1.09, 1.10, 1.15,
1.18, 1.26 (5 x 9H, 5 x s, C(Cfl3)3), 4.04-4.06 (2H, m, H-5, H-6’), 4.11 (1H, dd, 16,6.
10.6Hz, 15,6 5.2Hz, H-6), 5.09 (1H, dd, J 1.2 3.8Hz, J23 10.2Hz, H-2), 5.14 (1H, t, 13.4 9.7112,
J45 9.7Hz, H-4), 5.51 (1H, t, H-3), 6.29 (1H, d, H-l); 8c (125MHz; CDC13): 27.26, 27.28,
27.31, 27.37 (OCC(QH3)3), 38.92, 38.96, 38.99, 39.05, 39.37 (OCC_(CH3)3), 61.83 (C-6),
67.66, 69.68, 69.86, 70.50 (C-2, C-3, C-4, C-5), 88.88 (C-l), 176.08, 176.42, 177.02,

177.12, 178.22 (O§C(CH3)3).

176

2,3,4,6-Tetra-O-trimethylacetyl-a-D-g1ucosyl bromide (41):

B-D-glucose pentapivaloate 29 (1 g, 1.7 mmol) was dissolved in 2 mL
dichloromethane and 2 mL 30% hydrogen bromide in acetic acid. This solution was
stirred at room temperature for 24 hours. The solution was evaporated and the residue
was partitioned between 30 mL dichloromethane and 30 mL saturated sodium
bicarbonate solution. The organic layer was dried and evaporated to give 2,3,4,6-tetra-O-
trimethylacetyl-a-D-glucosyl bromide 41. Yield: 0.84g, 87%. (lit. [33] mp. 142-143°C;
[01]!)25 + 153.00 (c = 1, CHC13)); 5H (500MHz; CDC13): 1.04, 1.08, 1.10, 1.13 (4 x 9H, 4 x
s, C(CLI_3);;), 4.08 (2H, m, H-6, H-6’), 4.22 (1H, dt, J55 3.2Hz, 15,6 3.2Hz, 14510le, H-
5), 4.72 (1H, dd, 11.2 4.3Hz, 123, 101112, H-2), 5.13 (1H, t, 13,4 10.4Hz, H-4), 5.54 (1H, t,
H-3), 6.53 (1H, d, H-l); 8c (125MHz; CDC13): 26.82, 26.85, 26.90, 26.98 (OCC(QH3)3),
38.45, 38.55, 38.60, 38.69 (OCQ(CH3)3), 60.63 (C-6), 66.24, 69.36, 70.65, 72.34 (C-2, C-

3, C-4, C-5), 86.74 (C-l), 176.17, 176.52, 177.04, 177.65 (O_CC(CH3)3).
B-D-Mannose pentapivaloate (43):

This compound was prepared using the same procedure as described above for the
preparation of a-D-glucosepentapivaloate 33 using a-D-mannose as starting material.
Yield: 90%. 5H(500MHz; CDC13): 1.07, 1.11 1.12, 1.18, 1.26 (5 x 9H, 5 x s, C(Cfl3)3),
3.80 (1H, ddd, J55 4.1Hz, 15.6 2.0Hz, 14,5 10.0Hz, H-S), 4.10 (1H, dd, 16,66 12.4 Hz, H-6’),
4.17 (1H, dd, H-6), 5.14 (1H, dd, J23 3.2Hz, 13,4 10.2Hz, H-3), 5.42 (1H, dd, 11.2 1.0Hz, H-
2), 5.43 (1H, t, H-4), 5.80 (1H, d, H-l); 5c (125MHz; CDC13): 26.92, 26.97, 26.99, 27.01,
27.11 (OCC(C_H3)3), 38.62, 38.69 (OCQ(CH3)3), 61.38 (C-6), 64.61, 68.08, 70.90, 73.03

(C-2, C-3, C-4, C-S), 90.61 (C-l), 175.77, 176.44, 176.82, 176.98, 177.87 (OQC(CH3)3).

177

2,3,4,6-Tetra-O-trimethy1acety1-a-D-mannosy1 bromide (46):

This compound was prepared using the same procedure as described above for the
preparation of 2,3,4,6-tetra-O-trimethy1acetyl-a-D-glucosyl bromide 41 from D-mannose
pentapivaloate 43 and 44. Yield: 86%. 6” (500MHz; CDC13): 1.04, 1.10, 1.15, 1.19 (4 x
9H, 4 x s, C(Cfl3)3), 4.06-4.21 (3H, m, H-5, H-6, H-6’), 5.38 (1H, dd, J.;; 1.5Hz, 12,3
3.1Hz, H-2), 5.50 (1H, t, 13,4 10.1Hz, 14,5 10.1Hz, H-4), 5.67 (1H, dd, H-3), 6.19 (1H, d,
H-l); 6C (125MHz; CDC13): 26.91, 26.93, 26.94 (OCC(QH3)3), 38.62, 38.69, 38.74, 38.79
(OCQ(CH3)3), 60.78 (C-6), 64.17, 68.21, 71.89, 73.03 (C-2, C-3, C-4, C-5), 83.64 (C-l),

176.36, 176.46, 176.93, 177.74 (OQC(CH3)3).

N-(2,3,4,6-Tetra-O-trimethylacetyl-B-D-glucosyl) acetamide (47):

B-D-glucosepentapivaloate 29 (1.0 g, 1.7 mmol) was dissolved in 15 mL
acetonitrile and 5 mL methanesulfonic acid was added in. The reaction mixture was
stirred for 12 hours and quenched by pouring into ice with sodium bicarbonate. The
mixture was extracted by 2 x 30 mL dichloromethane. The organic layer was dried and
evaporated. The crude product was purified by column chromatography to give N-
(2,3,4,6-tetra-O-trimethy1acetyl-B-D-glucosy1) acetamide 47 as the major product. Yield:
0.61 g, 66%. SH (500MHz; CDC13): 1.05, 1.07, 1.10, 1.15 (4 x 9H, 4 x s, C(Cfl3)3), 3.79
(1H, ddd, J55 2.2Hz, J51, 4.4Hz, 14,5 10.2Hz, H-S), 4.08 (2H, 2 x dd, 16,5 12.6Hz, H-6, 6’),
4.88 (1H, t, J12 9.6Hz, 12,3 9.6Hz, H-2), 5.06 (1H, t, 134 10.2Hz, H-4), 5.23 (1H, t, Jmn
9.6Hz, H-l), 5.35 (1H, t, H-3), 6.17 (1H, d, NflAc); 5c (125MHz; CDC13): 23.14
(CH3CONH), 26.88, 26.91, 26.96, 26.99 (OCC(_CH3)3), 38.58, 38.63, 38.75, 38.84

(OCQ(CH3)3), 61.45 (C-6), 67.37, 70.29, 71.87, 73.89, 78.09 (C-1, C-2, C-3, C-4, C-5),

178

169.93 (CH3QONH), 176.31, 176.74, 177.93, 178.45 (OQC(CH3)3); FT-IR: 2988, 1786,

1755, 1406, 1168 cm".
N-(2,3,4,6-Tetra-O-acety1-D-glucosyl) acetamides (5101 and 51B):

a-D-glucosepentaacetate (1.0 g, 2.6 mmol) was dissolved in 15 mL acetonitrile
and 5 mL methanesulfonic acid was added. The reaction mixture was stirred for 12 hours
and quenched by pouring into ice with sodium bicarbonate. The mixture was extracted by
2 x 30 mL dichloromethane. The organic layer was dried and evaporated to afford a
mixture of N-(2,3,4,6-tetra—O-acety1-D-g1ucosyl) acetamide 5101 and 51B in a ratio of 1 :7.
Yield: 0.82g, 82%. [(110 = + 52.45° (c = 1.0, CHC13); For 51B: (lit. [34] mp. 156-158°C;
[01]D + 17 (c = 2, chloroform»; 8H (500MHz; CDC13): 1.93, 1.95, 1.97, 1.99, 2.01 (5 x 3H,
5 x s, COCH_3), 3.77 (1H, ddd, J55 2.2Hz, 15.6 4.7Hz, 14,5 10.2Hz, H-5), 4.02 (1H, dd, .165
12.6Hz, H-6’), 4.24 (1H, dd, H-6), 4.86 (1H, t, 11,2 9.6Hz, 12,3 9.6Hz, H-2), 4.99 (1H, t, 13.4
9.9Hz, H-4), 5.20 (1H, t, 11.1w 9.3Hz, H-l), 5.24 (1H, t, 1-1-3), 6.44 (1H, d, NflAc); 8C
(125MHz; CDC13): 20.75, 20.76, 20.84, 20.91 (OOCQH3), 23.52 (CH3CONH), 61.90 (C-
6), 68.34, 70.79, 72.95, 73.71, 78.35 (C-1,C-2, C-3, C-4, C-5), 169.78, 170.04, 170.67,

170.82, 171.15 (OQCH3); FT-IR: 3023, 1753, 1701, 1369, 1228, 1041cm".
N-(2,3,4,6-Tetra-O-acetyl-D-galactosyl) acetamides (5401 and 54B):

These compounds were prepared using essentially the same procedure as the
preparation of N-(2,3,4,6-tetra-O-acety1-D-g1ucosy1) acetamides 5101 and 51B. Yield:
86%; 5411: 54B = 1: 8. [01]D = + 62.44 (c = 1.0, CHC13); For 54B: (lit. [34] mp. 172-

174°C; [(111) + 33° (c = 1, chloroform»; 5H (500MHz; CDC13): 1.93, 1.94, 1.98, 2.00, 2.08

179

(5 x 3H, 5 x s, COCfl3), 4.02 (1H, m, H-5), 4.06-4.10 (2H, m, H-6, 6’), 5.04 (1H, dd, 11.2
9.0Hz, 12,3 10.2Hz, H-2), 5.08 (1H, dd, 13,4 3.2Hz, H-3), 5.19 (1H, t, J1,NH 9.0Hz, H-l),
5.38 (1H, dd, 14.5 1.0Hz, H-4), 6.38 (1H, d, NHAC); 5c (125MHz; CDC13): 20.77, 20.81,
20.87, 20.93 (OOCQH3), 23.52 (QH3CONH), 61.40 (C-6), 67.41, 68.54, 71.06, 72.53,
78.67 (C-l,C-2, C-3, C-4, C-5), 169.95, 170.21, 170.58, 170.59, 171.45 (OQCH3); FT-IR:

3097, 2479, 1748, 1700, 1630, 1369, 1224, 1046 cm".

1,3-Di-O-trimethy1acetyl-g1ycerol 56 and (i)-1,2-di-O-pivaloy1 glycerol (57):

Glycerol (2.32 g, 25.2 mmol) was dissolved in 50 mL pyridine and cooled down
to 0°C. Trimethylacetyl chloride (6.08 g, 50.4 mmol) was added in slowly. The reaction
mixture was stirred at 5°C for 10 hours. The mixture was poured into saturated sodium
bicarbonate solution containing ice and stirred for a while. Chloroform 2 x 50 mL was
used to extract the mixture. The organic phase was dried and evaporated to dryness. The
residue was purified by chromatography to give a mixture of 1,3-di-O-trimethylacetyl-
glycerol 56 and (i)1,2-di-O-trimethylacetyl-g1ycerol 57 in a ratio of 2: 1. Total yield: 55%.
For 56: 8” (500MHz; CDC13): 1.08 (18H, 8, C(Cfl3)3), 4.03 (1H, dd, H-2), 4.09 (4H, H-1,
1’, 3, 3’); 5c (125MHz; CDC13): 26.82 (OCC(_C_H3)3), 38.52 (OCQ(CH3)3), 64.64 (C-1, C-
3), 67.67 (C-2), 178.26 (OQC(CH3)3); For 57: 8” (500MHz; CDC13): 1.05, 1.07 (2 x 9H,
2 x s, C(Cfl3)3), 3.65 (2H, d, H-3, 3), 4.15 (1H, dd, H-l’), 4.28 (1H, dd, H-l), 5.01 (1H,
m, H-2); 8c (125MHz; CDC13): 26.77, 26.79 (OCC(QH3)3), 38.52 (OC_C_(CH3)3), 60.79
(C-3), 62.29 (C-l), 71.86 (C-2), 177.60, 177.90 (OQC(CH3)3); FT-IR: 3575, 3325, 2997,

2879, 1716, 1479, 1104 cm".

180

(i)-N-Acetyl-aminopropanediol (59):

The mixture of 1,3-di-O-trimethylacetyl-glycerol 56 and (i)l,2-di-O-pivaloy1-
glycerol 57 (0.6 g, 2.3 mmol) was dissolved in 5 mL acetonitrile and 1.5 mL
methanesulonic acid was added. The reaction mixture was stirred for 4 days at room
temperature and poured into saturated sodium bicarbonate solution with ice. Chloroform
(2 x 25ml) was used to extract the mixture. The organic layer was dried and evaporated to
afford (21:) N-acetyl-aminopropanediol 59 as colorless oil. Yield: 0.56g, 80%. FAB-MS:
m/z calcd. for C|5H27BN05 301.1889, found 302.1965 (MH+); 5H (500MHz; CDC13):
1.10, 1.11 (2 x 9H, 2 x s, C(CI_13)3), 1.89 (3H, s, HNCOCfl3), 3.39 (2H, t, 12,; 6.0Hz, H-3),
4.03 (1H, dd, J.;; 6.0Hz, My 12.1Hz, H-l’), 4.21 (1H, dd, J.; 3.7Hz, H-l), 5.00 (1H,
dddd, H-S), 6.27 (1H, b, NflAc); 8C (125MHz; CDC13): 22.90 (NHCOQH3), 26.89, 26.94
(OCC(QH3)3), 38.64, 38.68 (OCQ(CH3)3), 39.58 (C-3), 62.86 (C-l), 70.12 (C-2), 170.31
(NHQOCH3), 177.85, 178.00 (O§C(CH3)3); FT-IR: 3672, 3093, 2975, 1733, 1660, 1281,

1147 cm".

(21:)-1 ,2,4-Tri-O-pivaloyl erythritol (63):

Erythritol was partially pivaloated with 3.2 equivalent trimethylacetyl chloride
using the same procedure described above for the preparation of 1,3-di-O-
trimethylacetyl-glycerol 56 and (i)-1,2-di-O—pivaloy1-glycerol 57. Yield: 60%. 8”
(500MHz; CDC13): 1.16, 1.17 (3 x 3H, 3 x s, C(Cfl3)3), 2.92 (1H, b, -OH), 3.92 (1H, ddd,
13,4 2.9HZ, J32, 5.0Hz, 12,3 8.0Hz), 4.10 (1H, dd, J4,4~ 11.9Hz, H-4’), 4.20 (1H, dd, H-4),
4.28 (1H, dd, Jl-z 4.9Hz, 11y 12.2Hz, H-l ’), 4.41 (1H, dd, 11,2 2.9HZ, H-l), 4.99 (1H, ddd,

112); 5c (125MHz; CDC13): 26.96, 27.09 (OCC(C_H3)3), 38.76, 38.81, 38.86

181

(OCQ(CH3)3), 62.23 (04), 64.97 (01), 68.43 (03), 70.83 (C-2), 177.07, 178.49, 178.91

(O§C(CH3)3); FT-IR: 3563, 3396, 2995, 2879, 1744, 1479, 1286, 1117, 1060 cm".
(i)-4-Bromo-4-deoxy- l ,2,3-tri-O-trimethylacetyl-erythritol (64):

To 1,2,4-tri-O-pivaloy1 erythritol 63 (0.40 g, 1.1 mmol) 30% hydrogen bromide in
acetic acid (2 mL) was added. The reaction mixture was stirred at room temperature for 2
hours and evaporated. The residue was partitioned between chloroform (25 mL) and cold
saturated sodium bicarbonate solution (25 mL). The organic layer was dried, evaporated
and purified by column chromatography to afford 4-bromo-4-deoxy—l,2,3-t1i-O-
trimethylacetyl-erythritol 64 Yield: 0.36g, 75%. 8" (500MHz; CDC13): 1.04, 1.06, 1.09 (3
x 3H, 3 x s, C(CI-_13)3), 3.35 (dd, 1H, .13; 5.7Hz, J4,4v 11.5Hz, H-4’), 3.50 (dd, 1H, 13,4
3.4Hz, H-4), 3.97 (dd, 1H, J 132 5.0Hz, J W 12.3112, H-l’), 4.27 (dd, 1H, 11,2 2.8Hz, H-l),
5.10 (ddd, 1H, 12,3 7.9Hz, H-3), 5.15 (ddd, 1H, H-2); 6C (125MHz; CDC13): 26.825,
26.758, 26.714 (OCC(QH3)3), 30.381 (C-4), 38.629, 38.501, 38.483 (OCQ(CH3)3),
61.164 (C-l), 69.084, 70.017 (C-2, 3), 177.279, 176.406, 176.293 (O_C_C(CH3)3); FT-lR:

2977,2875, 1735, 1278, 1128 cm".
(i)-4-N-Acetylamino-4-deoxy-1 ,2,3-tri-O-pivaloyl-erythritol (65a):

(i)-1,2,4—t1i-O-pivaloy1-erythritol 63 (0.9 g, 2.4 mmol) was dissolved in 5 mL
acetonitrile and borontrifluoride etherate (1.5 mL) was added. The mixture was sealed
and stirred at 50 °C for 5 hours and poured into saturated sodium bicarbonate solution (50
mL) with ice. After stirring, the mixture was extracted by 2 x 25 mL chloroform. The
organic layer was dried and evaporated to dryness. Flash column chromatography

afforded (21:)-4-N-acety1amino-4-deoxy-1,2,3-tri-O-pivaloyl erythritol 65a as colorless oil.

182

Yield: 0.78g, 78%. FAB-MS: m/z calcd. for C2.H37NO7 415.2570, found 416.2651
(MH’); 8“ (500MHz; cock): 1.08, 1.12, 1.14 (3 x 3H, 3 x s, C(Cfl3)3), 1.87 (3H, s,
HNCOCH3), 3.15 (1H, dt, 13.45.2112, JNH,4~5.2H2, 14,44 14.2Hz, H-4’), 3.50 (1H, dt, 13,,
7.2112, 1101.4 7.2112, H-4), 3.97 (1H, dd, 11;; 7.3112, J.;. 11.8Hz, 11-1 ’), 4.18 (1H, 1.24.1112,
11-1), 5.07 (1H, ddd, 12,, 7.3112, H-3), 5.19 (1H, dt, 11-2), 6.04 (1H, b, NflAc); 6c
(125MHz; c0013): 22.95 (HNCOQH3), 26.92, 26.95 (OCC(_C_H3)3), 38.55, 38.79, 38.80
(OCQ(CH3)3), 39.22 (04), 62.11 (0.1), 69.61, 69.70 (02, C-3), 170.01 (HNQOCH3),

177.51, 177.62 (O_CC(CH3)3); FT-IR: 3303, 3091, 2977, 2876, 1725, 1659, 1274 cm”.
(5')- 1 ,4-Di-O-pivaloyl- l ,2,4-butanetriol (67):

Starting from (S)- l ,2,4-butanetriol, (S)-1,4-di-O-pivaloy1-1,2,4-butanetriol 67 was
prepared as described above for the dipivaloation of glycerol. Flash column
chromatography was used to purify this compound as a single isomer. Yield: 60%. [ab =
~6.2° (c = 1.0, CHC13); 8H (500MHz; CDC13): 1.13, 1.15 (2 x 9H, 2 x s, C(Cfl3)3), 1.66-
1.82 (2H, m, H-3, 3’), 2.70 (1H, b, -OH), 3.86 (1H, m, H-2), 3.98 (1H, dd, Jr; 6.0Hz, Jm»
11.0Hz, H-l’), 4.05 (1H, dd, J12 4.0Hz, H-l), 4.14 (1H, m, H-4’), 4.22 (1H, m, H-4); 5C
( 125MHz; CDC13): 27.07 (OCC(QH3)3), 32.67 (C-3), 38.63, 38.75 (OCQ(CH3)3), 60.82
(C-4), 66.89 (C-2), 68.06 (C-l), 178.52, 178.64 (OQC(CH3)3); FT-IR: 3692, 3093, 2974,

1730, 1283, 1156 cm".

(S)-4-Bromo-1 ,3-di-O-pivaloyl-1 ,3-butanediol (70):

(S)-1,4-di-O-pivaloyl-1,2,4-butanetriol 67 (0.56 g, 2.0 mmol) was dissolved in 5
mL 30% hydrogen bromide in acetic acid at room temperature. This solution was stirred

for 12 hours. Methanol (10 mL) was added and evaporated. The residue was partitioned

183

between 25 mL dichloromethane and 25 mL saturated sodium bicarbonate solution. The
organic layer was dried and evaporated to give (S)-4-bromo-1,3-di-O-pivaloyl-l,3-
butanediol 70 as a colorless oil. Yield: 0.52 g, 76%. SH (500MHz; CDC13): 1.16, 1.18 (2 x
9H, 2 x s, C(CH3)3), 199-2.05 (2H, m, H-2, 2’), 3.42 (1H, dd, 13,4'4.8Hz, J42 10.8Hz, H-
4’), 3.54 (1H, dd, 13,4 4.7Hz, H-4), 4.01-4.12 (2H, m, H-1,1’), 5.03 (1H, m, H-3), 4.22
(1H, m, H-4); 0c (125MHz; CDC13): 27.05, 27.10 (OCC(_C_H3)3), 31.64 (C-2), 33.97 (C-

4), 38.68, 38.89 (OCQ(CH3)3), 59.96 (C-l), 68.66 (C-3), 177.49, 178.24 (OQC(CH3)3).

(S)-4-N-Acetylamino-l,3-di-O-pivaloy1-1,3-butanediol (71) and (S)-4-N-acetylamino-

1 ,2-di-O-pivaloyl-1 ,2-butanediol (72):

These compounds were prepared using the same procedure as described above for
(i)-4-N-acetylamino-4-deoxy-1,2,3-tri-O-pivaloy1 erythritol 65 starting from (S)-1,4-di-
O-pivaloyl-l,2,4-butanetriol 67. Total yield: 88%. 71 : 72 = 2: 1. For 71: [a]D = -35.6° (c
= 1.0, CHC13); FAB-MS: m/z calcd. for C16H29N05 315.2046, found 316.2123 (MH+); 6H
(500MHz; CDC13): 1.03, 1.04 (2 x 9H, 2 x s, C(Cfl3)3), 1.77 (2H, m, H-2, 2’), 1.80 (3H, s,
HNCOCH3), 3.29 (2H, m, H-4, 4’), 3.93 (2H, m, H-1, 1’), 4.87 (tdd, 1H, J = 4.7Hz, J =
4.7Hz, J = 6.5Hz, J = 9.0Hz, H-3), 6.37 (t, 1H, J = 5.7Hz, LINCOCH3); 8c (125MHz;
CDC13): 22.725 (HNCOQH3), 26.821, 26.791 (OCC(_CH3)3), 30.627 (C-2), 38.585,
38.371 (OCQ(CH3)3), 42.542 (C-4), 59.819 (C-l), 69.031 (C-3), 170.161 (HNQOCH3),
178.034, 178.009 (OQC(CH3)3); FT-IR: 3704, 3298, 3094, 2974, 1730, 1656, 1283, 1153
cm". For 72: [(Ih) = + 447° (c = 1.0, c1103); FAB-MS: m/z calcd. for CszgNOs
315.2046, found 316.2123 (MH+); 8H (500MHz; CDC13): 1.08, 1.11 (2 x 9H, 2 x s,
C(Cfl3)3), 1.64, 1.73 (2 x 1H, 2 x m, H-3, 3’), 1.88 (3H, s, HNCOCH3), 2.84 (tdd, 1H,

J4,HNAc 5.4Hz, J3g4' 5.4Hz, 13.4: 9.1112, .144: 14.2Hz, H-4’), 3.43 (ddd, 1H, J3'4 6.4Hz, 13,4

184

11.8Hz, 114), 3.94 (dd, 1H, 11.26.7142, 1.,1. 11.9Hz, 11-1 ’), 4.15 (dd, 1H, 823.1112, 11-1),
5.02 (m, 1H, H-2), 6.30 (t, b, 1H, HNCOCH3); 8c (125MHz; cock): 23.053
(HNCOQH3), 26.927 (OCC(_C_H3)3), 30.585 (03), 38.689, 38.575 (OCQ(CH3)3), 35.073
(04), 64.858 (01), 68.989 (02), 170.169 (HNQOCH3), 178.394, 177.837

(O_C_C(CH3)3);FT-1R: 3703, 3091, 2975, 1731, 1654, 1283, 1149 cm".
2,3,5-Tri-O-acety1-6-O-t-buty1dimethy1silyl-D-galactone- l ,4-1actone (73):

D-galactono-1,4-lactone (2.0 g, 11.2 mmol) was dissolved in 40 mL dry pyridine
and t-butyldimethylsilyl chloride (1.86 g, 12.4 mmol) was added. The reaction mixture
was stirred for 12 hours and cooled down to 0°C. Then acetic anhydride (5.74 g, 56.2
mmol) was added slowly. The reaction mixture was stirred for 3 hours at 0°C and poured
into saturated sodium bicarbonate solution containing ice. The mixture was stirred and
the water phase was poured off. The residue was dissolved in 50 mL dichloromethane
and cold 1N hydrochloric acid was added in slowly until the pH reached ~ 1. The organic
phase was separated and washed with brine (20 mL) and distilled water (20 mL) and
dried. Evaporation followed by chromatography purification gave 2,3,5-tri-O-acetyl-6-O-
t-butyldimethylsilyl-D-galactone-1,4-lactone 73 as colorless oil. Yield: 3.52 g, 75%. [(111)
= - 45.1° (c = 1.0, CHC13); 511 (500MHz; CDC13): 0.02 (6H, s, Si(Cl_13)2), 0.82 (9H, s,
Si(C(CH_3)3)), 2.05, 2.06, 2.11 (3 x 3H, 3 x s, COC_H3), 3.70 (1H, dd, 15,9 7.7Hz, 16.6.
10.1Hz, H-6’), 3.73 (1H, dd, 15,6 5.6Hz, H-6), 4.63 (1H, dd, 13,4 6.9Hz, 14,5 2.6Hz, H-4),
4.99 (1H, ddd, H-S), 5.39 (1H, t, 123 7.0Hz, H-3), 5.56 (1H, d, H-2); 8c (125MHz;
CDC13): -5.72, -5.65 (Si(§H3)2), 18.01 (Si(Q(CH3)3)), 20.22, 20.35, 20.54 (COQH3),
25.57 (Si(C(_C_3H3)3)), 59.82 (C-6), 70.60, 72.04, 72.11, 76.64 (C-2, C-3, C-4, C-5),

168.35, 169.21, 169.55, 169.72 ( C=O’s in lactone and acetate); FT-IR: 2961, 1809, 1756,

185

1632, 1364, 1235, 1039 cm".

2,3,5-Tri-O-acetyl-D-galactone- l ,4-lactone (74):

2,3,5-tri-O-acety1-6-O-t-butyldimethylsilyl-D-galactone-1,4-1actone 73 (1.2 g, 2.5
mmol) was dissolved in 15 mL acetonitrile and 1.5 mL boron trifluoride etherate was
added in. The mixture was stirred at room temperature for 5 minutes and 5 g sodium
bicarbonate was added in to neutralize. The solution was evaporated to dryness and
partitioned between 50 mL dichloromethane and 50 mL water. The organic phase was
dried and evaporated to afford 2,3,5-tri-O-acety1-D-galactone-1,4-1actone 74. Yield:
0.77g, 88%. [01],) = + 1.1° (C = 1.0, CHC13); 6H (500MHz; CDC13): 2.00, 2.03, 2.08 (3 x
3H, 3 x s, COCH3), 3.97 (ddd, 1H, 14,5 2.3Hz, J51; 5.7Hz, 15,6 6.7Hz, H-5), 4.07 (dd, 1H,
15,5 5.7Hz, .165 11.5Hz, H-6’), 4.19 (dd, 1H, 15,6 6.7Hz, H-6), 4.37 (ddd, 1H, .1 = 0.9Hz,
13,4 5.9Hz, H-4), 5.58 (1H, d, H-2), 5.59 (dd, 1H, H-3); 5c ( 125MHz; CDC13): 20.533,
20.370, 20.135 (COQH3), 64.108 (C-6), 78.995, 72.222, 71.998, 67.248 (C-2, C-3, C-4,
C-5), 170.925, 170.132, 169.618, 168.695 ( C=O’s in lactone and acetate); FT-IR: 3495,

3097, 1804, 1749, 1629, 1371, 1225, 1040 cm".

6-N-Acety1amino-6-deoxy-2,3,5-tri-O-acety1-D-galactone-1 ,4-lactone (75):

2,3,5-t1i-O-acetyl-D-galactone-1,4-lactone 74 (0.75 g, 2.5 mmol) was dissolved in
5 mL acetonitrile and 1.5 mL boron trifluoride etherate was added. The container was
sealed and heated in an 65°C oil bath for 20 hours. The solution was poured into sodium
bicarbonate solution with ice and stirred. Chloroform (2 x 25 mL) was used to extract the
mixture and the organic layer was evaporated to dryness. The crude product was purified

by column chromatography to give 6-N-acetylamino-6-deoxy-2,3,5-tri-O-acety1-D-

186

galactone-IA-lactone 75 as colorless oil. Yield: 0.49 g, 58%. FAB-MS: m/z calcd. for
C|4H|9N09 345.1060, found 346.1136 (MW); 8” (500MHz; CDC13): 1.93 (3H, s,
CH3CONH), 2.07, 2.10, 2.12 (3 x 3H, 3 x s, OOCCH3), 3.44 (1H, ddd, J55 6.6Hz, 16,6.
14.4Hz, JNH,6‘ 5.8Hz, H-6’), 3.62 (1H, ddd, J55 4.8Hz, IN“, 6.2Hz, H-6), 4.54 (1H, dd, 13.4
6.9Hz, 14,5 2.5Hz, H-4), 5.19 (1H, m, H-5), 5.37 (1H, t, J23 7.0Hz, H-3), 5.54 (1H, d, H-
2), 6.19 (1H, t, flHAc); 5c (125MHz; CDC13): 20.40, 20.55, 20.72 (OOCQH3), 23.08
(HNOCQH3), 40.1 (06), 69.4, 72.1, 72.3, 78.6 (C-2, C-3, C-4, C-5), 168.2, 169.3, 169.8,
170.2, 170.8 ( C=O’s in lactone, amide and acetates); FT-IR: 3298, 1806, 1751, 1662,

1372, 1229, 1178 cm".

187

REFERENCE

[1] Sotlzberg, S. Adv. Carbohydr. Chem. Biochem. 1970, 25, 229.

[2] Bock, K.; Pedersen, C.; Thogersen, H. Acta. Chem. Scand., Ser. B. 1981, 35, 441.
[3] Defaye, J.; Gadelle, A.; Perderson, C. Carbohydr. Res. 1985, 136, 53.

[4] Defaye, J.; Gadelle, A.; Perderson, C. Carbohydr. Res. 1986, 152, 89.

[5] Defaye, J.; Gadelle, A.; Perderson, C. Carbohydr. Res. 1990, 205, 191.

[6] Bock, K.; Lundt, 1.; Pedersen, C. Carbohydr. Res. 1979, 68, 313.

[7] Vekemans, J. A. .1. M.; Dapperens, C. W. M.; Claessen, R.; Koten, A.M. J.;
Godefroi, E. F.; Chittemen, G. J. F. J. Org. Chem. 1990, 55, 5336.

[8] Bols, M.; Lundt, 1. Acta. Chem. Scand. 1988, B42, 67.

[9] Bock, K.; Lundt, 1.; Pedersen, C. Carbohydr. Res. 1981, 90, 17.

[10] Bock, K.; Lundt, 1.; Pedersen, C. Carbohydr. Res. 1982, 104, 79.

[1 l] Lundt, 1.; Frank, H. Tetrahedron 1994, 50, 13285.

[12] Bock, K.; Lundt, 1.; Pedersen, C.; Refit, 8.; Acta. Chem. Scand. 1986, B40, 740.
[13] Lundt, I.; Madsen, R. Synthesis 1993, 714 and 720.

[14] Lundt, 1.; Pedersen, C. Synthesis 1992, 669.

[15] Bock, K.; Lundt, 1.; Pedersen, C. Carbohydr. Res. 1981, 90, 7.

[16] Book, K.; Lundt, 1.; Pedersen, C.; Sonnichsen, R. Carbohydr. Res. 1988, I 74, 331.
[17] Lundt, 1.; Madsen, R. Synthesis 1995, 787.

[18] Gallo, C.; Jeroncic, L. O.; Varela, 0.; de Lederkremer, R. M. J. Carbohydr. Chem.
1993, 12, 841.

[19] Ferrier, Robert J.; Tyler, Peter C. Carbohydr. Res. 1985, 136, 249.

[20] Nowacki, A.; Smiataczowa, K.; Kasprzykowska, R.; Dmochowska, B.; Wisniewski,
A. Carbohydr. Res. 2002, 33 7, 265-72.

[21] Arsequell, G.; Valencia, G. Tetrahedron: Asymmetry 1999, 10, 3045.

188

[22] Grogan, M. J.; Pratt, M. R.; Marcaurelle, L. A.; Bertozzi, C. R. Annu. Rev. Biochem.
2002, 7], 593.

[23] Pleuss, N.; Kunz, H.,Angew. Chem. Int. Ed. 2003, 42, 3174.
[24] Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem, Int. Ed. Engl. 1996, 35, 1380.

[25] Likhoshertov, L. M.; Novikova, V. A.; Dervitskaja, V. A.; Kochetkov, N. K.
Carbohydr. Res. 1986, 46, Cl.

[26] Lubineau, A.; Auge, J.; Drouillat, B. Carbohydr. Res. 1995, 266, 211.
[27] Dorsey, A. D.; Barbarow, .1. E.; Trauner, D. Org. Lett. 2003, 5, 3237.
[28] Damkaci, F.; DeShong, P. J. Am. Chem. Soc. 2003, 125, 4408.

[29]Ratc1iffe, A. J.; Mootoo, D. R.; Andrews, C. W.; Fraser-Reid, B. J. Am. Chem. Soc.
1989, III, 7661.

[30] Ratcliffe, A. J.; Fraser-Reid, B. J. Chem. Soc., Perkin Trans. 1 1989, 1805.
[31]Ratc1iffe, A. J.; Konradsson, P.; Fraser-Reid, B. Carbohydr. Res. 1991, 216, 323.
[32] Handlon, A. L.; Fraser-Reid, B. J. Am. Chem. Soc. 1993, 115, 3796.

[33] Kunz, H.; Harreus, A. Ann. Chem. 1982, 41.

[34] lsac-Garcia, .1 .; Calvo-Flores, F. G.; Hernandez-Mateo, F .; Santoyo-Gonzalez, F.
Eur. J. Org. Chem. 2001, 383.

189

Chapter 5

Synthesis Efforts towards 1-Deoxynojirimycin(DNJ) and 1-

Deoxymannojirimycin (DMJ)

ABSTRACT

A strategy for the preparation of carbohydrate derivatives bearing two leaving
groups for use in the preparation of aza-sugars by di-N-alkylation of amines was
developed. The target molecules were l-deoxynojirimycin (DNJ) and 1-
deoxyrnannojirimycin (DMJ). Several fully protected dibromosugars failed to undergo di-
N-alkylation to give azasugars when treated with amines. The selective acetal protections
of dibromoaldonolactones and dibromoalditols were explored. DNJ was synthesized from
partially protected dibromoalditol by di-N-alkylation with ammonium hydroxide. DMJ

was synthesized from D-mannose using the same strategy.
5.1 Some explorations of 1,5-iminosugar syntheses

In chapter 1, we discussed many synthetic methodologies towards the synthesis of
azasugars, including 1,5-iminosugar. Despite the elegant chemistry, most of the current
methodologies are not suitable for usage in industry. A practical, environmentally benign
synthetic pathway is still desired. Based on the similar structure of azasugars with real
sugars, we employed commercially available sugars or sugar derivatives as our starting
materials. Since the structure of azasugar is defined as the replacement of ring oxygen

with nitrogen atom, our synthetic strategy starts with an introduction of a nitrogen-

190

containing nucleophile into the structure. The cheapest nitrogen source exists in different
amines. Another advantage of using amines as nucleophiles is that by using different
amines, it is relatively easy to build a library of N-substituted azasugars, which is of much
interest. SN2 nucleophilic substitution is favored in the process of introduction of
nucleophiles because it is easy to define the chiral centers involved. Therefore, leaving
group introduction in certain positions of sugar structure is desired, which often requires

different protection strategy (Scheme 5.1).

 

Scheme 5.1 The retrosynthetic analysis of 1,5-iminosugar synthesis by double N-alkylation
strategy

The 2,6-dibromination of sugar lactones has become a very important
derivatization of aldonolactones [1-7]. It is easy to imagine that disubstitution of two
bromo groups with amines to form 2,6-iminoalditols, which are 6-membered ring
azasugars. If all the substitutions (bromination and N-alkylation) happen through 8 SN;
mechanism, the stereochemistry of the aldonolactones will be retained in the resulted
azasugars because of double inversion. In a simple analysis of the sugar structure
(Scheme 5.2), it can be seen that by a carbon chain inversion strategy, l-deoxynojirimycin
(DNJ) 1 is related with L-gulono-y-lactone 2. L-gulono-y-lactone 2 is commercially

available in an affordable price; furthermore, it can also be easily synthesized from

191

abundant carbohydrate material [8]. All of the above makes it an appropriate starting

point for synthetic work.

 

 

 

— —OH
r—L
HO— C1) —1
Ho— 1:)
NHZR _
-OH _:)y L”OH
L—- Ol-l
Deoxynojirimycin 1 L-gulonolactone 2

Scheme 5.2 Synthesis of azasugars from aldonolactones: a carbon chain inversion strategy

The hemiacetal of D-glucuronic acid 3 was reduced by sodium borohydride
(NaBH4), acidic work-up of this reaction automatically cyclized the carboxylic acid 4 to
L-gulono-y-lactone 2. When 2 was treated with hydrogen bromide in acetic acid, 2,6-
dibromo-2,6-dideoxy-L-idono-y-lactone 5 was yielded in 95% crude yield. The free
hydroxyl groups in 3— and 5- positions of 5 were then benzylated in 68% yield by
treatment with benzyl 2,2,2-trichloroacetimidate under catalysis of
trifluromethanesulfonic acid to give 3,5-di-O-benzyl-2,6-dibromo-2,6-dideoxy-L-idono-y-

lactone 6 (Scheme 5.3).

HO O —> HQ ——> _ .
OH > 90% HO OH =

 

HO 3 5 OH
OH OH OH OH OH
3 4 2
HBr-HOAC/ 95%
o 0 H CCI3C(=NH)OBn o 0 H
m A TfOH, 0112012 , ;
Br OBn =08” Br 68% Br OH OH Br
6 5

Scheme 5.3 Synthesis of 3,S-di-O-benzyl-Z,6-dibromo-2,6-dideoxy-L-idono- -lactone 6

192

Derivatized and protected, 6 was treated with benzylamine in refluxing toluene

 

(Scheme 5.4).
Expected:
Br OH
O H 2 2
o BnNH2 BnHNWNHz
; ; __—__.> _ . r ———>
‘ " B .=. 5:
3' SOB" 03" r 0 7 OBn OBn
(gr 9H 0 NHBn
B HN ? =
n WNHB" _____> Eng-1&0
08 63'168" 9 OBn
Actual: Br OH
gr 9H BnNHz. Toluene, B HN §
BnHN ' ' Reflux n \ —>
WBI' _: = Br
0 68h ban 0 (3H OBn
7 10
Br 9H Br OH
BnHN ‘ A ?
Wm _:_44_°/‘1_, BnHN OBn
O 11 0

Scheme 5.4 The treatemnt of 3,5-di-O-benzyl-2,6-dibromo-2,6—dideoxy-L-idono-y-lactone 6
with benzylamine

The first expected reaction was the aminolysis of the lactone ring by benzylamine
to give amide 7. After that, under certain conditions, we expected the nucleophilic
substitution of primary bromo group by benzylamine to give 8, after this the N atom
attached to C-6 would displace the 2-bromo group in an intramolecular SN; fashion to
yield compound 9. After separation, no expected product was found. NMR analysis of the
products showed benzylamide and tetrahydrofuran ring structure. Finally, the structure
was assigned as compound 11 (N-Benzyl-Z-(4-benzyloxy-3-hydroxy~tetrahydro-furan-2-
yl)-2-bromo-acetamide). The mechanism for the formation of this product is proposed to
involve a B-elimination and Michael addition followed by an intramolecular nucleophilic

substitution. In this reaction, the activity of the primary bromo group as a leaving group

193

was too low, allowing B-elimination to occur. The intramolecular cyclization was fast

enough to drive the Michael addition.

The elimination process is specially undesired because afier elimination, two
chiralities were eliminated. It is always an issue to control the stereoselectivity when new
chiralities are built into a structure. To avoid the elimination process, it is necessary to
reduce the carbonyl group to an alcohol. This reduction does not reduce the efficiency of

the synthesis very much since it is to be carried out in later steps.

2,6-Dibromo-2,6-dideoxy-L-idono-y-lactone 5 was treated with sodium
borohydride to yield 2,6-dibromo-2,6—dideoxy-L—iditol 12. When this dibromo-L-iditol 12
was treated with excess 2,2-dimethoxypropane, the major product 1,314,5-di-O-
isopropylidene-2,6-dibromo-2,6-dideoxy-L-iditol 13. This fully protected dibromoiditol
was then treated with benzylamine in refluxing toluene. No expected product was
detected even after 3 days. Most of the starting materials remained unchanged while a

little decomposition yielded complicated mixture presumably by an elimination process

(Scheme 5.5).

The failure of nucleophilic substitution of bromo group with benzylamine in this
case could be ascribed to steric reasons. After protection by two isopropylidene groups,
steric hindrance was significantly increased in the two-ring structure, which prevented the

benzylamine molecule from coming and attacking even the primary carbon.

194

 

 

o O H
Br
_£_ a B NaBl-legyeOH > HO
B = r
' OH OH ° Br OH
5 >4 '2
O O
2.2-Dimethoxypropane Br BnNH
> a ————2—§ No Reaction
50°/o Br 0 6

.3X

Scheme 5.5 synthesis of l,3,4,S-di-O-isopropylidene-Z,6-dibromo-2.6—dideoxy-L-iditol l4 and
the treatment of it with benzylamine

We also tried to synthesize an azasugar anolog of sialic acid as a potential

sialyltransferase inhibitor using the same strategy (Scheme 5.6).

 

HO Oil: 0 0 T880 01150 0 T350 OM80” 0 NHBn
M 3’ TBDMSC" Py W BnNHZ w
‘ “ OH ¢ ‘ - ’OMS < ‘ ‘/
oro b) MSCI. Py. 680/0 0rd ”600/0 0 6 OMS
H0 14 H0 15 H0 16

Scheme 5.6 A synthetic attempt to an azasugar anolog of sialic acid as a potential
sialyltransferase inhibitor

3,5-O-Benzylidene-D-glycero-D-gulo-heptono-l,4-lactone 14 was selectively
protected at the primary position using tert-butyldimethylsilyl (TBDMS) group. The 2,5-
positions were then mesylated. Bearing two leaving groups, 3,5-O-benzylidene-6-O-tert-
butyldimethylsilyl-2,S-di-O-methanesulfonyl-D-glycero-D-gqu-heptono- 1,4-lactone 15
was treated with benzylamine in refluxing toluene. After 24 hours, no desired amine
substituted compound was recovered. Most starting material simply went through
aminolysis to form an amide while an amount of a decomposition mixture was produced.
Although mesylate is a better leaving group than bromide, the amine nucleophilic

substitution still failed to proceed. Again, this failure was ascribed to the high steric

195

hindrance existing in the structure. To avoid this difficulty, another protection strategy

was employed.

5.2 The isopropylidene protection of dibromoaldonolactones and

dibromoalditols

The research work of Lundt’s group showed that various naked 2,6-dibromosugar
alditols reacted with ammonia to give exclusively 5-membered ring system, including
iminosugars and anhydrous sugars [9]. Epoxide formation was found to be a very
important activation process in these reactions. Compared to fully protected
dibromosugar, epoxide formation greatly increased the activity of bromo group as a
leaving group. It was also clear that the naked sugar had much smaller steric hindrance.
However, while 5 and 6-member ring formations were both possible, the reaction always
went to 5-membered ring formation, which was undesired in the synthesis of 1-

deoxynojirimycin (DNJ) or other 1,5-iminosugars.

With this in mind, the involvement of protecting groups is inevitable. However,
full protection of hydroxyl groups shuts down the epoxide formation and then greatly
decreases the activity of leaving groups, especially when introducing the nitrogen source
interrnolecularly. The protecting groups should leave enough reactivity for the
introduction of a nitrogen nucleophile. Importantly and more interestingly, they should
shut down the 5-membered ring formation and switch it to 6-membered azasugar ring
formation. The sugar structure should be appropriately partially protected. To explore the

selective partial protection of activated aldonolactones and alditols, we carried out

196

systematic research on the acetal protection and 1,2-diacetal protection of

dibromoaldonolactones and dibromoalditols.

2,6-Dibromo-Z,6-dideoxy-aldonolactones and alditols with D-ido and D-manno
configurations were chosen as model compounds mainly because they are easily

accessible and affordable (Scheme 5.7).

 

 

 

 

 

 

A) OH HO—
HO 0 Br—
MO HBr-HOAc _ NaBH4 HD—
S [,2 92% 80% '—OH
HO OH —OH
16 ‘8 —Br
B) OH Br
OAc
HO 0 HBr-HOAc _ __,H00 0
HO 0 0‘3
OH e9 OH
r
19 21 %
Ho—
OAc O OH 0 Br—
Br 3 NaBH _
s o H+.CH30H_r \. —;> “0
——’ H 52% v H 81% —OH
AcO Br HO Br :2?
22 23 24

Scheme 5.7 Synthesis of 2,6-dibromo-2,6-dideoxy-aldonolactones and alditols with (A) D-ido
and and (B) D-manno configurations

2,6-Dibromo-2,6-dideoxy—D-idono-1,4-lactone 17 and iditol 18 were synthesized
from D-gulono-y-lactone 16 using the same procedure as for the L-configuration. This
procedure was also employed in the synthesis of 2,6-dibromo-2,6-dideoxy—D-manno-l,4-
lactone 23 and mannitol 24 except that the starting material was D-glucono-S-lactone 19
instead of y-lactone. The reaction between 19 and hydrogen bromide in acetic acid was

believed to go through an acetoxonium cation mechanism. In this pathway, the lactone

197

ring has to be opened to allow the acetoxonium cation formation and followed
bromination. It is necessary to extend the reaction time to let it proceed and the yield is
also lower than that of 2,6-dibromo-2,6-dideoxy—D-idonolactone 17. The reduction of
2,6-dibromo-2,6-dideoxy-D-mannolactone 23 afforded 2,6-dibromo-2,6-dideoxy—D-

mannitol 24.

5.2.1 Isopropylidene protection of dibromoaldonolactone

First, the reaction of 2,6-dibromo-2,6-dideoxy-D-idono-1,4-lactone 17 with 2,2-
dimethoxypropane under different reaction conditions was explored. The results are

shown in Scheme 5.8.

 

 

(CH3)26<OCH3>2 H o 0
stOH, CHZCIZ _ .
. 78% ' 5‘ $- Br
)(0
17 (CH3)20(OCH3)2 25
stOH, CH3OH
>4 56%
OH Br
0 :

+

1:1 0 00

man
9?
3‘
O
O
I
w

 

27

Scheme 5.8 The isopropylidene protection of 2,6-Dibrom0-2,6-dideoxy-D-idonolactone 17

2,2-Dimethoxypropane was chosen as the isopropylidene source and p-
toluenesulfonic acid was chosen as the catalyst. This reaction condition was generally
used to give the kinetic acetal products. When the reaction was carried out in
dichloromethane, 2,6-dibromo-2,6-dideoxy-3,S-O-isopropylidene-D-idono- l ,4-lactone 25

was formed in a short time. Extension of the reaction time did not change the structure of

198

the product, as shown both in thin layer chromatography (TLC) and in NMR spectra.
Although the transacetalation would release methanol in the system, the reaction
condition is not strong enough to open the lactone ring. As the kinetic product, 3,5-0-
isopropylidene product was obtained under this reaction condition. However, when we
changed the solvent to methanol without changing the other reaction conditions, a totally
different phenomenon was shown. The lactone ring was opened and two products were
formed. NMR analysis of the products showed them to be methyl 2,6-dibromo-2,6-
dideoxy-3,4-O-isopropylidene-D-idonate 26 and methyl 2,6-dibromo-2,6-dideoxy—4,5-O-
isopropylidene-D-idonate 27 respectively with a ratio of ~ 1:1. Under acidic conditions, it
was plausible for methanol to open the lactone ring and released 4-position free hydroxyl
group. Transacetalation can then occurr to protect two of the three free hydroxyl groups.
The acetal formation further drove the equilibrium between idonolactone and methyl
idonate towards the final products. Although the formation of the kinetically favored 3,5-
O-isopropylidene product 25 was possible, the lactone ring tended to go through
methanolysis to release the high ring strain of 25. The selectivity between the three free
hydroxyl groups was governed by general acetalation selectivity rules of acyclic polyols.
l,2-isopropylidene was favored over 1,3-isopropylidene because there are no strong 1,3-
diaxial interaction in the former. Between l,2-isopropylidenes, the selectivity would favor
the two hydroxyl groups with threo- configuration rather than those with erythro-
configuration. This could be understood from the conformation analysis. Normally, the
two hydroxyl groups with threo- configuration have a gauche relationship, while those
with erythro- configuration have a trans- relationship. Before the acetal ring formation,
the two hydroxyl groups with threo- configuration are closer with each other than those

with erythro- configuration. After the 5-membered acetal ring formation, the product with

199

threo- configuration has lower steric strain than those with erythro- configuration because
of the partially eclipsed relationship between two substitution groups in the latter. In this
case, both 3,4-diol and 4,5-diol had threo- relative configuration. So it is not surprising
that two isomers were formed in a ~ 1:1 ratio, 3,4-O-protected and 4,5-O-protected

respectively.

We also explored the acetal protection reaction of 2,6-dibromo-2,6-dideoxy-D-
manno-y-lactone 23. The same reaction conditions were used so that we could compare

them with each other. When we treated 23 with 2,2-dimethoxypropane under catalysis of

p-toluenesulfonic acid in dichloromethane, as we expected, the lactone ring remained
untouched while 3,5-O-benzylidene-Z,6-dibromo-2,6-dideoxy—D-manno-y—lactone 28 was

formed. This kinetically favored product 28 stayed stable under this reaction condition

(Scheme 5.9).
(CH3)20(OCH3)2
stOH. CHZCI
OH /
71%

Br 0

75%

23 (CH3)ZC(OCH3)2 i
stOH, CH30H =

 

Scheme 5.9 The isopropylidene protection of 2,6-dibromo-2,6-dideoxy-D-mannonolactone 23

When the solvent was changed to methanol, lactone ring opening occurred.
However, only one major product was obtained, which was confirmed to be methyl 2,6-

dibromo—2,6-dideoxy-3,4-O-isopropylidene-D-mannonate 29 by NMR analysis. After the

200

lactone ring was opened, 4-position hydroxyl group was released; therefore the
configuration of this acyclic sugar chain became 3,4-threo- and 4,5-erythro-. According
to our previous analysis of acetalation selectivity, 3,4-O-isopropylidene was selectively

formed.
5.2.2 Isopropylidene protection of dibromoalditol

Isopropylidene has been used extensively for the protection of 1,2- and 1,3-diols.
This is especially true in carbohydrate chemistry [10]. Generally, for a polyol system, l,2-
acetonide formation is favored over 1,3-acetonide formation, although the extent of this
preference is dependent on structures. Among l,2-acetonides, the selectivity is based on
the thermodynamic stability. Related to the azasugar synthesis, we reasoned that the
selective isopropylidene protection of dibromoalditols would generate partially protected
dibromoalditols, which would be good precursor to iminoalditols. The isopropylidene
protection of 2,6-dibromo-2,6-dideoxy—D-iditol 18 and 2,6-Dibromo-2,6—dideoxy—D-
mannitol 24 was therefore explored. We previously prepared fully isopropylidene-
protected 2,6-Dibromo-2,6-dideoxy-L-iditol 13 using 2,2-dimethoxypropane as the
carbonyl equivalent. To explore the partial acetalation of dibromoalditol, acetone was
used as both the solvent and the carbonyl reagent. Under the catalysis of p-
toluenesulfonic acid, 2,6-dibromo-2,6-dideoxy—D-iditol 18 reacted with acetone
immediately to form two isomers, NMR analysis confirmed to be 2,6-dibromo-2,6-
dideoxy-3,4-O-isopropylidene-D-iditol 31 and 2,6-dibromo-2,6-dideoxy—4,S-O-
isopropylidene-D-iditol 32 respectively. An interesting phenomenon was that the ratio
between these two isomers changed while the reaction time was extended. If the reaction

was worked-up in 10 ~ 15 minutes, the selectivity between 3,4-O-isopropylidene 31 and

201

4,5-O-isopropy1idene 32 highly favored the former with a ratio 5:1. When the reaction
time was extended to overnight, finally the selectivity between them vanished (Scheme
5.10). It was reasonable to deduce that thermodynamically the two products were very
similarly favored, however, kinetically 3,4-O-isopropy1idene was favored. One
explanation was shown in Scheme 5.10. Both the products were formed from an
isopropylidene rearrangement from 1,3-O-isopropy1idene intermediate 30. Under our
reaction conditions, 1,3-O-isopropy1idene dibromoiditol quickly rearranged to 2,6-
dibromo-2,6-dideoxy-3,4-O-isopropylidene-D-iditol 31, while the rearrangement from 31
to 4,5-O-isopropylidene 32 had a bigger energy barrier so that it took a much longer time.
As a result, by control the reaction time, we were able to selectively synthesize different

partially protected dibromoiditols.

 

 

10 minutes 5 : 1
overnight 1 : 1

Scheme 5.10 The isopropylidene protection of 2,6-dibromo—2,6-dideoxy-D-iditol 18

The reaction between 2,6-dibromo-2,6-dideoxy-D-mannitol 24 and acetone under
that catalysis of p-toluenesulfonic acid was also carried out. Since it had a 3,4-threo- and

4,5-erythro- configuration, when 3,4-O-isopropylidene product was formed, the energy

202

barrier of rearrangement to 4 5 O-isopropylidene should be fairly high As we expected

even after overnight stirring, the only major product obtained was 2 6 dibromo 2 6-

drdeoxy—3 4-O-isopropylidene—D-mannitol 33 (Scheme 5 11)

OH OH
? Br
Acetone, p-TsOH _

H0
75%

 

r OH
24

(Dun-

Scheme 5 l 1 The isopropylidene protection of 2,6-dibromo 2 6 drdeoxy- -mann1tol 24 by

acetone

5 3 The 1 2 dracetal protection of dibromoaldonolactones and

dibromoalditols
OMe
R1 OMe
H
R1 OMe HO R2
OMe R :fiR/RB
+ —H—-> 1 2
0' ; MeOH
HO ; R3
R1 0

Scheme 5 12 l,2-diacetal protection group for 1,2-diols (Adapted from Ley S. V Baeschlrn D
K Dixon D J Foster A C Ince S. J.; Priepke H. W.; Reynolds D. 1., Chemical reviews

211111, 101, 53-80)
Based on its own structure, isopropylidene generally protects l 2 diol or 1 3 drol

to form a five-membered ring or a six-membered ring respectively However in some

cases it rs necessary to protect l 2-diols while forming a six-membered ring To fulfill

thrs requrrement Ley and other groups [1 1-16] introduced 1 2 dracetals rnto protecting

203

chemistry (Scheme 5.12). Systematical research was carried out to understand the
reactivity and selectivity of different l,2-diacetal reagents. Among those l,2-diacetal
reagents, 2,3-butanedione was the most attractive one because it is commercially
available at a cheap price. Therefore we explored the l,2-diaceta1 protecting reaction of
dibromoaldonolactones and dibromoalditols using 2,3-butanedione as the protecting

reagent.

5.3.1 1,2-Diacetal protection of dibromoaldonolactones

The l,2-diacetal protection reaction of 2,6-dibromo-2,6-dideoxy—D-idono-1,4-
lactone 17 was carried out using the procedure established by Ley’s group [14,16]. In this
procedure, 2,3-butanedione was employed as the l,2-diaceta1 source while
trimethylorthoester was used as the dehydrating reagent. Methanol was usually employed
as the solvent and BF 3OEt2 was used as the Lewis acid catalyst to avoid undesired
triprotected acetal product. When 17 was subjected to this reaction condition, after 12
hours, TLC showed that two major products with similar polarities were formed. NMR
analysis confirmed them to be two regio-isomers with about 1:1 ratio that were partially
protected by l,2-diacetal at 3,4-positions (34) and 4,5-positions (35) respectively
(Scheme 5.13). It is not difficult to understand that there is no good selectivity between
3,4- and 4,5-diacetals since their relative configurations are both threo-, as discussed
before. However, when the reaction time was extended for 3 days, the selectivity between
those two changed dramatically. NMR spectroscopy showed only one major regio-isomer,
which was also confirmed to be 4,5-diacetal 35. Although there was no obvious steric
preference between 3,4-diacetal and 4,5-diacetal, the improved selectivity by extension of

reaction time could greatly improve the synthetic efficiency.

204

CH3COCOCH3.CH(OCH3)3
31:30512. CH3OH

 

H300 0 Br OCH3
+ swgfi
0 Br HO
12h 1:1 65%
72h 1'6 58%

Scheme 5.13 l,2-diacetal protection reaction of 2,6-dibromo-2,6-dideoxy-D-idonolactone 17

We also explored the 1,2-diacetal protection reaction of 2,6-dibromo-2,6-dideoxy—
D-manno-1,4-1actone 23. Same reaction condition was employed. Here we expected the
lactone ring to be opened and 3,4— or 4,5-diacetal should be formed. Based on the 3,4-
threo- and 4,5-erythro- configurations, we expected the 3,4-diacetal to be formed more
favorably. After 12 hours stirring, TLC showed only one major product; however, this
product migrated much faster on thin layer chromatography than the two 1,2-diacetal
products formed when we treat 2,6-dibromo-2,6-dideoxy—D-idono-1,4-lactone 17 with
2,3-butanedione. NMR analysis also showed no -OH signal. Further spectral analysis
suggested the structure to be 3,4,5-triprotected compound 36 (Scheme 5.14). In previous
research of reactions of acyclic polyols with 2,3-butanedione under catalysis of Lewis
acid [14-16], only partial protection of l,2-diols were formed while all the neighboring
free hydroxyl groups stay unprotected. The further cyclization only happened when protic
acid catalyst such as p-toluenesulfonic acid was used while heating was also necessary.

Thus, the full protection of this triol was interesting and could be useful in future work.

205

CH3COCOCH3.CH(OCH3)3
3130132. CH30H

 

56%

   

Scheme 5.14 l,2-diacetal protection reaction of 2,6-dibromo-2,6-dideoxy-D-manno-l,4-
lactone 23

5.3.2 1,2-Diacetal protection of dibromoalditols

The 1,2-diacetal protecting reactions with dibromoalditols were also explored.
Since the primary hydroxyl groups of 2,6-dibromoalditols are separated from 3,4,5-tiols
by 2-bromo group, we expected the participation of 1-OH in the 1,2-diaceta1 protection to
be very disfavored. We expected very similar behaviors compared to the 1,2-diacetal

protections of the dibromoaldonolactones since the stereochemistry was not changed.

When 2,6-dibromo-2,6-dideoxy-D-iditol 18 was treated with 2,3-butanedione and
trimethylorthoester under the catalysis of BF3OEt2 in dry methanol, a very complicated
mixture was obtained. NMR analysis confirmed that several methyl orthoesters of 2,6-
dibromo-Z,6-dideoxy-D-iditol were formed as major products. 3,4-Diacetal 37 and 4,5-
diacetal 38 of 2,6-dibromo-2,6-dideoxy-D-iditol, which were expected to be major
products, were not found. Extension of the reaction time did not yield diacetal protected
2,6-dibromoiditol (Scheme 5.15). This phenomenon was not consistent with reports from
literature [13-15]. Presumably, under the current reaction conditions, trimethylorthoester
reacted with 2,6-dibromoiditol very quickly to yield orthoesters. The orthoesters were

sufficiently stable under this reaction conditions so that they would not exchange with

206

2,3-butanedione to form the expected products. To avoid the orthoester formation, the
dehydrating reagent trimethylorthoester was removed from the reaction system while the
l,2-diacetal source was changed to 2,2,3,3-tetramethoxybutane. The latter could be easily
prepared from 2,3-butanedione and trimethylorthoester under acid catalysis. When we
treated 2,6-dibromo-2,6-dideoxy—D-iditol 18 with 2,2,3,3-tetramethoxybutane under
catalysis of BF3OEt2, two major products with ~1:1 ratio was yielded. NMR analysis
confirmed the structure to be 3,4-O-diacetal 37 and 4,5-O-diacetal 38 of 2,6-dibromo-2,6-
dideoxy-D-iditol respectively. Under this reaction condition, orthoesters were not formed.
Between 3,4-diaceta1 and 4,5-diacetal, there was no obvious steric difference, so they had

approximately same regioselectivity.

H H

uIIllO
uIlIIO

Br CH3COCOCH3, HC(OCH3)3
HO/\/\/\/ BFaoEtz, MeOH _ Orthoesters mixture

7

 

r OH

18
CH3(CH3O)2CC(OCH3)2CH3 620/o
OCH3

- veg? 43¢

OCH3 OCH3

UJIIII-I

Scheme 5.15 The l,2-diacetal protection of 2,6-dibromo-2,6-dideoxy-D-iditol 18

The l,2-diacetal protection reactions of 2,6-dibromo-2,6-dideoxy-D-mannitol 24
were also explored using largely the same reaction conditions employed above. When
2,3-butanedione was directly used as l,2-diacetal source combined with
trimethylorthoester as dehydrating reagent, only complicated mixtures of orthoesters were

formed, which was consistent with our experiments above. Therefore, 2,2,3,3-

207

tetramethoxybutane was used as the 1,2-diaceta1 reagent without dehydrating reagent.
Because of the 3,4-threo- and 4,5-erythro- configuration of 2,6-dibromo-2,6-dideoxy-D-
mannitol 24, we expected that 3,4-diacetal 39 should be the major product. NMR
spectroscopy of the crude product showed that the major product was 3,4,S-triprotected
compound 40 (Scheme 5.16). This triprotection was also shown in the diacetalation of
2,6-dibromo-2,6-dideoxy—D-manno-1,4-lactone 23. With this D-manno- configuration,
the triprotected compound 40 could adjust to a relatively less strained conformation so
that the bridged ring system was sufficiently stable. As a result, 3,4-diacetal was not

formed as a major product.

 

OH OH
3 CH3COCOCH3. HC(OCH3)3
3' BF3OEt2. MeOH Orthoesters
HO _5_ : mixture
Br “OH

CH3(CH3O)2CC(OCH3)2CH3

* BF30EI2, MeOH

OCH3 Br
By

0&0”
0 H3CO

OH

 

OCH3 39

Actual

Expected

Scheme 5.16 The 1,2—diacetal protection of 2,6-dibromo-2,6-dideoxy-D-mannitol 24

208

5.4 Synthesis of l-deoxynojirimycin (DNJ)

R R RI R2
OH X 2 OX0 00 X
W74 ——> *0 —> f + R1
OH "4‘" x 2 R2
OH R1 R2 R1 R2

 

 

 

 

R
slow X faster fast 2
OH R1 R1 R1
X
“MN/"1 —» afigfim’ —> aging, + asflztflgi
fast slow fast
OH R1 R1 x R1 X
”M _> RRizBO —> #31260 + RalzBo
OH 4 R4
slow faster fast

Scheme 5.17 A ring-strain analysis of acetal and l,2-diacetal protection and fused-ring
formation

Based on the literature about acetal and l,2-diacetal protections and our own
results on selective acetalation and l,2-diacetalation, we explored the synthesis of 1-
deoxynojirimycin (DNJ) 1. Our goal was to develop a facile and practicable pathway to
this azasugar using easily accessible materials and simple protection strategy. Starting
from 2,6-dbromo-2,6-dideoxy-L-iditol 12, the involvement of protecting groups was
inevitable. On the other hand, full protection of hydroxyl groups shut down the epoxide
formation and then greatly decreases the activity of leaving groups, especially when

introducing nitrogen source interrnolecularly. The protecting groups should leave enough

209

reactivity for the introduction of a nitrogen nucleophile. Importantly and more
interestingly, they should shut down the S-membered azasugar ring formation and switch
it to 6-membered azasugar ring formation. A ring systems formation analysis is shown in
Scheme 5.17. Acetal protecting groups selectively protect l,2-trans diols of an acyclic
structure, however, this makes the fused-ring formation difficult because of the high ring
strain. On the other hand, 1,2-diacetal protecting groups also selectively protect l,2-trans
diols of an acyclic structure. This makes 5,6-fiJsed ring system formation difficult,
however, this could facilitate the 6,6-fused ring formation because of the relatively small

ring strain.

Starting from commercially available L-gulono-l,4-lactone 2, 2,6-dibromo-2,6-
dideoxy-L-iditol 12 was easily prepared by dibromination followed by reduction
according to literature [3]. When treated with 2,2,3,3-tetramethoxybutane under catalysis
of BF3OEt2, a ~l:l mixture of 3,4-O- and 4,5-O-(2’,3’-dimethoxybutane-2’,3’-diyl)
protected dibromoiditol was synthesized (Scheme 5.18). This mixture was then directly
treated with tert-butyldimethylsilyl chloride to selectively protect the primary hydroxyl
groups. An inseparable mixture of 41 + 42 was then obtained. When the mixture was
treated with ammonium hydroxide in methanol, TLC showed two products that were well
discriminated. Easy separation by column chromatography and NMR study showed the
structure of the fast spreading compound as 2,3-anhydrous-6-bromo-6-dideoxy-4,5-O-
(2’,3’-dimethoxybutane-2’,3’-diyl)-1-O-t-butyldimethylsily1-L-gulitol 43, and the slow
spreading compound as 6-amino-2-bromo-2,6-dideoxy-3,4-O-(2’,3’-dimethoxybutane—
2’,3’-diyl)-l-O-t—butyldimethylsilyl-L—iditol 45. Under basic conditions, epoxides were

formed from bromo- groups and neighboring free hydroxyl groups. If an epoxide was not

210

too sterically hindered for reaction with an incoming nucleophile, e.g. an amino group, a
nitrogen-contained functional group would be introduced into the structure. Otherwise,
the epoxide would stay untouched. When reacted with ammonium hydroxide, the 3,4-O-
diacetal protected dibromoiditol formed 5,6-anhydrosugar 44, which was then reacted
with ammonia to form 45 with a yield of 21% over 3 steps. However, the 4,5-O—diacetal
protected dibromoiditol formed 2,3-anhydrosugar 43, which was too hindered for
incoming ammonia to attack and therefore stayed untouched. This interesting

differentiation greatly facilitated the separation process.

 

 

OTBS
OTBS @1030?
0H0:§;+ %)E/§h+
Br
Br OH Br OH Br Blr42
12 41 ~ 1: 1 mixture /0
c 1 Cl
\ OTBS
_ OTBS -
0 gm \0 1ones
0
fi/O . <— 00
Br OH NH2 61' O
O\ _ 0\ .i :;/E’/oh8r
45 Easily separated
21 % over 3 steps 44
(117490 23% over 3 steps
OTBS
NH- HCI
_e_> H0
95% H0

Scheme 5.18 The synthesis of l-deoxynojirimycin(DNJ) l: a) 2,2,3,3-Tetramethoxybutane,
BF 3.OEt2, MeOH, rt; b) TBDMSCl, Py, rt; c) NH4OH, MeOH, rt; d) DBU, Tol, reflux; e)
4NHCl, MeOH

211

The protected 6-amino-2—bromo-iditol 45 was then refluxed with DBU in toluene
to furnish the partially protected 2,6-iminosugar 46. No 5-membered ring formation was
detected in this process, consistent with our previous analysis. Finally, Compound 46 was
deprotected by hydrogen chloride in methanol to afford deoxynojirimycin 1

quantitatively, which matched the NMR spectral properties from the literature [17-20].

In summary, 1-deoxynojirimycin 1 was synthesized in 5 steps from 2,6-dibromo-
2,6-dideoxy-L-iditol 2 with an overall yield ~ 15%. The key step was the use of 1,2-
diacetal as a protecting group. This left the 6-bromo group reactive enough for an
intermolecular nucleophilic substitution, and facilitated the 6-membered ring formation in
an intramolular fashion as a structure tether. The whole process did not involve any azide
and heavy-metal catalysis. Only relatively cheap and easily available reagents were used.
All these features made the synthesis not only of academic interest, but also practicable.
Furthermore, it is likely that this methodology could be expanded to the synthesis of some
other important azasugars such as deoxymannojirimycin and castanosperrnine by
modifying the starting material, employing different amines as nitrogen source and/or

transforming the partially protected 2,6-iminosugar 46.

5.5 Synthesis of 1-deoxymannojirimycin(DMJ)

We further employed this strategy in the synthesis of l-deoxymannojirimycin
(DMJ). A retrosynthetic analysis was shown in Scheme 5.19. DMJ could be synthesized
from a similar 1,2-diacetal protected dibromoglucitol intermediate, which could be
obtained from dibromoglucitol. This dibromoglucitol could be synthesized from D-

manna-1,4-1actone by dibromination and reduction.

212

HO \ OTBS

OH I::> O &
NH -HCl 0
HO §0 3,
HO Br

\
OH
OH
<3 OH
OH .
Br Br

 

Scheme 5.19 The retrosynthetic analysis of l-deoxymannojirimycin(DMJ)

A synthesis of l-deoxymannojirimycin (DMJ) 55 was carried out starting from a-
D-mannose 47. The bromine oxidation of the anomeric position in buffered soluction of
47 afforded D-manno-l,4-1actone 48. Compound 48 was then subjected to HBr treatment
in acetic acid to give 2,6-dibromo-2,6-dideoxy-D-glucono-l,4-lactone 49, which was
subsequently reduced by sodium borohydride to give 2,6-dibromo-2,6-dideoxy-D-glucitol
50. The pH of the solution was controlled below 8 to prevent the anhydrosugar formation.
Compound 50 was then treated with 2,2,3,3-tetramethoxybutane under catalysis of
BF3OEt2 in methanol for 16 hours to give the 3,4—O-diacetal protected 2,6-dibromo-2,6-
dideoxy-D-glucitol 51. Because of the 4,5-erythro-configuration of 50, the selectivity of
3,4-diacetal protection was very high. The free primary -OH of 51 was then protected by
T BDMSCl to give 2,6-dibromo-2,6-dideoxy—3,4-O-(butane 2,3-diaceta1)-1-O-t-
butyldimethylsilyl-D-glucitol 52, which was appropriately protected and derivatized for
double N-alkylation. Compound 52 was treated with aqueous ammonium hydroxide in
methanol for 6 hours to give 6-amino-2-bromo-2,6-dideoxy-3,4-O-(butane 2,3-diacetal)-
l-O-t-butyldimethylsilyl-D-glucitol 53. Compound 53 was then refluxed with DBU in

toluene for 12 hours to cyclize intramolecularly to form 2,6-imino-2,6-dideoxy—3,4-O-

213

(butane 2,3-diacetal)-l-O-t-butyldimethylsilyl-D-glucitol 54, which was deprotected by

hydrochloric acid in methanol to give the l-deoxymannojirimycin (DMJ) 55.

    

OH OH
OH HO 0 Br
,0 b
OH ..3—5 ——-——>
OH 62% over
HO OH 2 steps
4., H0 48
8.188 o 99%
%2:§l\8<—e— i2 ‘—d—
o .
Br Br
0\
51
f1
OH
ON” “ NH-HCI
33% over 98% HO
4 steps H0

55

Scheme 5.20 The synthesis of l-deoxymannojirimycin (DMJ) 55: a) Brz, NaHCO3, H20; b)
HBr'HOAc; 62% over 2 steps; c) NaBH4, H+ resin, EtOH/HZO, 80%; d) 2,2,3,3-
Tetramethoxybutane, BF3OEt2, MeOH, rt; e) TBDMSCl, Py, r ; f) NH4OH, MeOH, rt; g)
DBU, Tol, reflux; 33% over 4 steps; h) 4N HCl, MeOH, 98%;

In summary, l-deoxymannojirimycin (DMJ) 55 was synthesized in 8 easy steps
from D-mannose. Similar to the synthesis of l-deoxynojirimycin (DNJ) 1, the key step
was the use of l,2-diacetal as a protecting group to selectively protect 3,4-OH’s while
leave S-OH free. The free S-OH greatly activated the primary bromo- group as a leaving
group by epoxide formation. The 6-membered diacetal ring greatly increased the
preference of intramocular N-alkylation over anhydrosugar formation. After double N-
alkylation, all the protection groups could be removed by a single acid treatment, which

also increased the efficiency of the synthesis.

214

5.6 Summary and future work

This thesis started from the NBS-initiated benzoxonium cation rearrangement.
Selective bromination and chirality manipulation of carbohydrates were accomplished.
This rearrangement was employed in tartrate derivative to generate a new chiral building
block. Inter- and intramolecular competitive nucleophiles were introduced into the
reaction system to generate new functionalities. Bromination was shown to be the most
competitive nucleophile. This methodology was extended to 1,3-dioxonium cation
facilitated selective bromination and Ritter-type amidation. A series of bromosugars and
aminosugars were generated. The stereospecificity and regioselectivity were discussed.
Finally, 1-deoxynojirimycin (DNJ) and 1-deoxymannojirimycin (DMJ) were synthesized
from partially protected dibromoalditols in a practical pathway. Some research directions

based on these results are shown below and will be carried out in the near future.

5.6.1 Intramolecular Ritter-type reaction of partially protected sugar nitriles

The t-butyl-l,3-dioxonium cation facilitated Ritter-type reaction discussed in
Chapter 4 has great potential in synthetic organic chemistry. It was employed to introduce
an amino group in a sugar lactone. If the sugar structure itself bears a cyano group, an
intramolecular Ritter—type reaction could furnish a sugar lactam structure (Figure 5.1).
Sugar nitriles could be easily synthesized from sugar oximes, which in turn could be
easily obtained from free sugars. The sugar lactams could be transformed to iminosugars

by a simple reduction.

215

O
30

N
[I a b ”d r—NH

. / mfg/rs
RO OH Y9 \R

R1
R = AC or PW R1 = Me or tBuPiv :RNH

Figure 5.1 The acyloxonium cation facilitated intramolecular Ritter-type reaction in the
synthesis of sugar lactams

 

5.6.2 Chiral 1,3-dioxonium cation facilitated asymmetric bromination and Ritter-

type reaction in the synthesis of chiral building blocks

In our t-butyl-l,3-dioxonium cation facilitated bromination and Ritter-type
amidation, glycerol was transformed to (i)3-bromo-3-deoxy-1,2-di-O-pivaloyl-glycerol
and (i)3-N-acetylamino-3-deoxy-l,2-di~O-pivaloyl-glycerol respectively. The chirality
was introduced into the achiral glycerol through the nucleophilic attack of t-butyl-l,3-
dioxonium. If chirality was introduced into the dioxonium cation, an asymmetric Ritter-
type reaction could occur to generate optical active product. Although theoretically every
chiral ester could serve as a chiral auxiliary, Mosher’s ester could be employed as a
candidate (Figure 5.2) according to its wide application in synthetic and analytic

chemistry.

OH (- Nu I
/ 0,.
RO/II’ -'/\0 RO (\gu
O G) 01
CF3
CF
Ph OMe Pfleo 3
R = H. Ac or other esters
Nu: = Br or RCN

Figure 5.2 Chiral 1,3-dioxonium cation facilitated asymmetric bromination and Ritter-type
reaction in the synthesis of chiral building blocks

Glycerol is an abundant 3—carbon building block. The asymmetric
functionalization has long been pursued. The chiral 1,3-dioxonium cation facilitated

asymmetric bromination and Ritter-type reaction could serve as a good pathway.

5.6.3 Chiral nitrile facilitated asymmetric Ritter-type reaction in the asymmetric

desymmetrization of glycerol

1"
N;
r
O G) U
R1 H OR
OH
HO/Y\OH -—> /L
OH (N/
Ro/”" o ,
/ In,
o——<@ “—7 R0 (\N“
R1 OR

R = Ac or Piv; R1 = Me or t-Bu

Figure 5.3 Chiral nitrile involved asymmetric Ritter-type reaction in the desymmetrization of
glycerol

217

The 1,3-dioxonium cation facilitated Ritter-type reaction involved 1,3-dioxonium
cation as an electrophile and an incoming nitrile as nucleophile. If the nucleophile is a
chiral molecule, another version of asymmetric desymmetrization of glycerol could be
introduced (Figure 5.3). Bearing a neighboring chiral center, lactonitrile could be used as
an affordable chiral nitrile source. Since both D- and L-isomers of lactonitrile are readily
available, the Ritter-type reaction could generate an aminodiol with both R- and S-

configurations, which could be used as a good platform of organic synthesis.

218

EXPERIMENTAL

General Optical activity data were obtained on a JASCO P-lOlO polarimeter at
25°C. NMR spectra were obtained on a Varian VXR-SOO Spectrometer operating at
500MHz for protons. Mass spectra were obtained on a JEOL HX-l lO-HF instrument
using fast atom bombardment as ionization mode. Spectra were recorded in the positive
ion mode. IR spectra were obtained on a Nicolet 710 spectrometer in chloroform solution

except when otherwise specified.

2,6-Dibromo-2,6-dideoxy-L-idono-y-lactone (5):

L—gulono-l,4-lactone (5 g, 28.1 mmol) was added into 20 mL 30% hydrogen
bromide in acetic acid and stirred for 2.5 hours at room temperature. Methanol (50 mL)
was added in and the solution was stirred for 12 hours. The reaction mixture was then
evaporated and co-evaporated with 2 x 20 mL water. The residue was dissolved in 200
mL ethyl acetate and washed by 25 mL saturated sodium bicarbonate. The organic phase
was dried and evaporated to dryness to afford 2,6-Dibromo-2,6-dideoxy-L-idono-y-
lactone 5 as colorless oil. Yield: 8.11g, 95%. [a]D = + 344" (c = 0.9, CHC13); 5H
(500MHz; CDC13): 3.49-3.51 (2H, d, H-6, 6’), 3.60 (2H, b, -OH’s), 4.23 (1H, ddd, J45
3.5Hz, J53 6.1Hz, 15,5 6.1Hz, H-S), 4.45 (1H, d, .12.; 4.5Hz, H-2), 4.61 (1H, dd, J3... 5.5Hz,
H-3), 4.81 (1H, dd, H-4); 6c; (125MHz; CDC13): 32.82 (C-6), 43.56 (C-2), 69.12, 75.73,
81.42 (C-3, C-4, C-5), 172.27 (C=O in lactone); FT-IR: 3393, 3094, 1778, 1161, 1098,

1012 cm".

219

3,5-Di-O-benzyl-2,6-dibromo-2,6-dideoxy-L-idono-y-lactone (6):

2,6-dibromo-2,6-dideoxy-L-idono-y-lactone 5 (2.56 g, 8.4 mmol) was dissolved in
25 mL anhydrous dioxane and 2,2,2-trichloroacetimidic acid benzyl ester (10.6 g, 42
mmol) was added in. The solution was cooled down to 0°C and trifluoromethanesulfonic
aicd (0.05g) was added in. The mixture was then stirred under room temperature for 24
hours and poured into cold saturated sodium bicarbonate solution. The mixture was
extracted with chloroform (3 x 25 mL). The organic layer was dried and evaporated. The
residue was purified by column chromatography twice to give 3,5-di-O-benzyl-2,6-
dibromo-2,6-dideoxy-L-idono-y-lactone 6 as white solid. Yield: 2.77 g, 68%. 5“
(500MHz; CDC13): 3.40 (dd, 1H, .155 4.8Hz, J65 10.5Hz, H-6’), 3.45 (dd, 1H, J53 7.7Hz,
H-6), 4.01 (1H, m, H-5), 4.51 (dd, 1H, J = 3.8Hz, J = 5.6Hz, H-4), 4.55, 4.56 (2 x 2H, 2 x
s, OCflzPh), 4.89 (d, 1H, H-2), 4.96 (dd, 1H, H-3), 7.35 (m, 10H, Arfl); 6c ( 125MHz;
CDC13): 161.890 (_C_=O in lactone), 136.274, 128.729, 127.845, 127.461 (Ar_C_), 81.543,
79.750, 76.003 (C-3, 4, 5) 73.065, 72.905 (OQHzPh), 41.265 (C-2), 28.723 (C-6); FT-IR:

3086, 1788, 1172, 1108, 1016 cm".

N-Benzy1-2-(4-benzyloxy-3-hydroxy-tetrahydro-furan-2—yl)-2-bromo-acetamide (1 l):

3,5-Di-O-benzyl-2,6-dibromo-2,6-dideoxy-L-idono-y-lactone 6 (1.0 g, 2.1 mmol)
was dissolved in 10 mL toluene and benzylamine (1.0 g, 9.3 mmol) was added in. The
mixture was refluxed for 16 hours and evaporated to dryness. The residue was partitioned
between chloroform and brine. The organic layer was dried and evaporated to dryness.
The residue was purified by column chromatography to give of N-Benzyl-2-(4-

benzyloxy-3-hydroxy-tetrahydro-furan-2-yl)-2-bromo-acetamide 11. Yield: 0.38 g, 44%.

220

5H (500MHz; CDCl3): 3.95 (1H, dd, H-6’), 4.05 (1H, m, H-S), 4.15 (1H, dd, H-6), 4.4-4.6
(7H, m, H-2, 3, 4, OCflzPh, NHCflzPh), 7.0 (111, b, LINCO), 7.4 (1011, m, Arfl); 5c
(125MHz; CDC13): 168.11 (C=O), 137.53, 128.99, 128.78, 128.17, 127.96, 127.89 (Ar_Q),
84.41, 83.10, 75.84 (C-3, 4, 5), 72.30, 71.81 (OQHZPh, C-6), 46.25 (C-2), 44.48

(HNQHzPh); FT-IR: 3486, 2961, 1668, 1204, 1008 cm".
2,6-Dibromo-2,6-dideoxy-L-iditol (12):

2,6-Dibromo-2,6-dideoxy-L-idono-y-lactone 5 (4.82 g, 15.9 mmol) was dissolved
in 40 mL methanol and cooled to 0°C. Sodium borohydride (2.2 g, 56.5 mmol) was added
slowly. After the addition, the mixture was stirred at 0 °C for 30 minutes more. 5 mL 37%
hydrochloric acid was added and the mixture was filtered and washed by methanol. To
the filtration H+ resin was added and filtered after stirring. The filtration was evaporated
and co-evaporated with 4 x 25 mL methanol to give 2,6-dibromo-2,6-dideoxy-L-iditol 12
as white solid. Yield: 4.00 g, 82%. M.p. 108-110°C; (lit. [9] mp. 111-1 12°C; [a]D + 17°
(c = 4, MeOH)); 6H (500MHz; CD3OD): 3.49 (dd, 1H, J55 6.6Hz, .165 9.8Hz, H-6’), 3.70
(dd, 1H, 15.6 6.2Hz, H-6), 3.82 (dd, 1H, J 132 6.0Hz, 11,1: 11.6Hz, H-l’), 3.87 (dd, 1H, 12,3
2.7Hz, 13.4 6.3Hz, H-3), 3.91 (m, 1H, H-2), 3.93 (dd, 1H, 11,2 6.7Hz, H-l), 3.99 (dd, 1H,
14,5 2.9HZ, H-4), 4.29 (dt, 1H, H-S); 8c (125MHz; CD3OD): 73.469, 71.797, 70.661 (03,

4, 5), 64.14 (C-l), 60.01 (02), 35.504 (06); FT-IR (CH3OH): 3841, l472cm".
1 ,3 :4,5-Di-O-isopropylidene-2,6-dibromo-2,6-dideoxy-L-iditol (13):

2,6-dibromo-2,6-dideoxy-L—iditol (1.0 g, 3.2 mmol) was dissolved in 5 mL
anhydrous N,N-dimethylformamide (DMF) with 2,2-dimethoxypropane (1.35 g, 13.0

mmol). p-Toluenesulfonic acid (20 mg) was added and the solution was stirred at room

221

temperature for 4 hours. The reaction mixture was poured into saturated sodium
bicarbonate solution with ice and stirred. The mixture was extracted by chloroform (2 x
25 mL). The extraction was dried and evaporated to dryness. The residue was purified by
chromatography to give l,3:4,5-di-O-isopropylidene-2,6-dibromo-2,6-dideoxy—L-iditol 13
as colorless oil. Yield: 0.63 g, 50%. 5" (500MHz; CDC13): 1.36, 1.39, 1.40, 1.41 (4 x 3H,
4 x s, C(Cfl3)2), 3.27 (dd, 1H, .155 5.0Hz, 16.3 9.8Hz, H-6’), 3.40 (dd, 1H, 15,6 8.3Hz, H-
6), 3.57 (ddd, 1H, Jl-z 1.2Hz, 11,2 2.7Hz, J23 13.3Hz, H-2), 3.93 (t, 1H, 13.4 1.2Hz, 14,5
1.2Hz, H-4), 3.97 (ddd, 1H, H-S), 4.03 (m, 1H, H-l’), 4.06 (m, 1H, H-l), 4.24 (dd, 1H,
H-3); 8c (125MHz; CDC13): 101.30, 100.08 (_C_I(CH3)2), 74.54, 73.02, 61.50, 60.54 (01,

C3, C-4, C-5), 53.79 (C-2), 29.98(C-6), 29.17, 25.49, 24.75, 19.39 (C(QH3)2).

3,5-O-Benzylidene-6-O-tert-butyldimethylsilyl-2,5-di-O-methanesulfonyl-D-glycero-D-

gulo- 1 ,4-lactone (15):

3,5-O-Benzylidene-D-glycero-D-gulo-1,4—lactone (2.0 g, 6.8 mmol) was dissolved
in 40 mL pyridine. t-Butyldimethylsilyl chloride (1.12 g, 7.4 mmol) was added in. The
mixture was stirred under room temperature for 12 hours and cooled down to 0°C in ice
bath. Methanesulfonyl chloride (1.70 g, 14.9 mmol) was added slowly. The mixture was
then stirred at 0 °C for 6 hours and poured into saturated sodium bicarbonate solution
with ice. After stirring, the mixture was extracted by chloroform (3 x 25 mL). The organic
layer was dried and evaporated. The residue was purified by column chromatography to
give 3,5-O-benzylidene-6-O-tert-butyldimethylsilyl-2,5-di-O-methanesulfonyl-D-glycero-
D-gulo-l,4-1actone 15 as colorless oil (68%). 8“ (500MHz; CDCl3): 0.04, 0.06 (2 x 3H, 2
x s, Si(QH3)2), 0.88 (9H, s, SiC(C_H3)3), 3.09, 3.23 (2 x 3H, 2 x s, CHfiOy), 3.91 (1H,

dd, J37. 3.5Hz, J7,7~ 12.0Hz, H-7’), 4.09 (1H, dd, J63 2.0Hz, H-7), 4.39 (1H, dd, 14,5 2.0Hz,

222

15,6 9.0Hz, H-5), 4.58 (1H, t, 13,4 1.5Hz, H-4), 4.89 (1H, ddd, H-6), 4.94 (1H, dd, 12.3
4.0Hz, H-3), 5.54 (1H, d, H-2), 5.56 (1H, s, Cfll’h), 7.35 (3H, m, Arfl), 7.40 (2H, m,

Arfl).

2,6-Dibromo-2,6-dideoxy—D-idono-y-lactone (17):

Starting from D-gulono-1,4-lactone, 17 was prepared according the same
procedure as the preparation of 2,6-dibromo-2,6-dideoxy—L-idono-y-lactone 5. Yield:

92%. The NMR data is identical to 2,6-Dibromo-2,6-dideoxy-L-idono-y-lactone 5.

2,6-Dibromo-2,6-dideoxy—D-iditol (18):

The preparation of 2,6-Dibromo-2,6-dideoxy-D-iditol 18 was carried out
according to the preparation of its L-isomer 12. Yield: 80%. NMR data was identical to
its L-isomer. M.p. 110-112°C; [a]D = - 18.0° (c = 1.0, CH3OH); (lit. [9] mp. 111-1 12°C;

[a]D - 17° (c = 4, MeOH)).

2,6-Dibromo-2,6-dideoxy-D-manno- 1 ,4-lactone (23):

To 40 mL 30% hydrogen bromide in acetic acid D-glucono-l,5-1actone (10 g,
56.2 mmol) was added. The mixture was stirred under room temperature for 20 hours.
Methanol (100 mL) was added and the solution was stirred for 10 hours more. The
reaction mixture was evaporated. Water (2 x 25 mL) was added in and evaporated. The
residue was dissolved in 200 mL ethyl acetate and washed by saturated sodium
bicarbonate solution (50 mL), brine (25 mL) and water (25 mL). The organic layer was
dried and evaporated to dryness. The residue was crystallized from hexanes: ethyl acetate

= 5:1 to give 2,6-dibromo—2,6-dideoxy-D-marmo-1,4-1actone 23 as pale yellow solid.

223

Yield: 8.88 g, 52%. M.p. 126-128°C; [a]D = + 86.3 (c = 1.0, CHC13); (lit. [3] mp. 131-
133°C; [01].)20 + 52.2° (c = 0.7, ethyl acetate)); 5H (500MHz; CDCl;): 2.89 (2H, b, -OH’s),
3.58 (1H, dd, 15,615.1Hz, 16,3 11.0Hz, H-6’), 3.66 (1H, dd, 15,6 3.1Hz, H-6), 4.13 (1H, ddd,
14,5 8.8Hz, H-5), 4.37 (1H, dd, 13.4 3.1Hz, H-4), 4.53 (1H, dd, 12,3 4.6Hz, H-3), 4.73 (1H,
d, H-2); 8C (75.5MHz; CDC13): 36.31 (C-6), 47.28 (C-2), 66.82, 68.91, 80.98 (C-3, C-4,

C-S), 170.78 (C=O in lactone); FT -IR (CHC13): 3657, 3189, 1784 cm".
2,6-Dibromo-2,6-dideoxy-D-mannitol (24):

2,6-Dibromo-2,6-dideoxy—D-mannonolactone 23 (5 g) was dissolved in 50 mL
ethanol and cooled down to -10°C in ice-salt bath. Amberlite I-120 H+ resin (10 mL) was
added in. Then sodium borohydride (1.5 g) was added in slowly, while ice was added in
the reaction mixture to keep the temperature low. After the addition, the mixture was
stirred at 0 °C for 30 minutes more. More amberlite I-120 H+ resin was added in until the
pH ~ 3. The mixture was filtered and the filtrate was evaporated to afford 2,6-dibromo-
2,6-dideoxy-D-mannitol 24 as colorless oil. Crude yield: 4.1 g, 81%. (lit. [4] mp. 92-
9%; [(11020 - 9° (c = 1.6, ethyl acetate)); 5.. (500MHz; CD3OD): 3.52 (1H, dd, H-6’),
3.56 (1H, dd, H—6), 3.67 (1H, ddd, H-2), 3.82-3.90 (4H, m, H-1, H-l’, H-3, H-4), 3.98
(1H, ddd, H-S); 5c (125MHz; CD30D): 38.21 (C-6), 55.13 (C-2), 63.20, 69.20, 69.59,

71.25 (C-1, C-3, C-4, C-S); FT-IR (CHC13): 2940, 1470, 1420, 1067 cm".
2,6-Dibromo-2,6-dideoxy-3,5-O-isopropylidene-D-idono- 1 ,4-1actone (25):

2,6-Dibromo-2,6-dideoxy-D-idono-1,4-lactone (0.36 g, 1.2 mmol) was dissolved
in 5 mL dichloromethane and 2,2-dimethoxypropane (0.25 g, 2.4 mmol) was added in. p-

Toluenesulfonic aicd monohydrate (0.03 g) was added in. The reaction mixture was

224

stirred for 12 hours and poured into cold saturated sodium bicarbonate solution. The
mixture was extracted with chloroform (2 x 15 mL). The organic layer was dried and
evaporated to give 2,6-dibromo-2,6-dideoxy—3,S-O-isopropylidene-D-idono-1,4-lactone
25. Yield: 0.22 g, 78%. [ab = -54.9° (c = 0.9, CHC13); 5H (500MHz; CDC13): 1.33, 1.42
(2x 3H, 2 x s, C(CH3)2), 3.39 (dd, 1H, 15,616.0Hz, 16,6. 10.1Hz, H-6’), 3.51 (dd, 1H, 15,6
8.1Hz, H-6), 4.06 (s, 1H, H-2), 4.22 (ddd, 1H, J45 2.0Hz, H-S), 4.53 (d, 1H, 13.4 2.4Hz, H-
3), 4.69 (t, 1H, H-4); 50 (75.5MHz; CDC13): 19.30, 28.57 (C(QH3)2), 29.15 (C-6), 40.35
(C-2), 73.15, 70.86, 68.55 (C-3, C-4, C-5), 99.65 (C(CH3)2), 170.82 (C=O in lactone);

FT-1R(CHC13): 3092, 1788, 1175 cm".

Methyl 2,6-dibromo-2,6-dideoxy-3,4-O-isopropylidene—D-idonate (26) and Methyl 2,6-

dibromo-2,6-dideoxy-4,5-O-isopropylidene-D-idonate (27):

The reaction was carried out using the same procedure for the preparation of 2,6-
dibromo-2,6-dideoxy—3,5-O-isopropylidene-D-idonolactone 25 except methanol was used
as the solvent. A mixture of two isomers was obtained with a ratio of 1:1. Yield: 56%. SH
(500MHz; CDC13): 1.421, 1.400, 1.388, 1.378 (4x 3H, 4 x s, C(Cfl3)2), 2.63, 2.89 (2 x
1H, 2 x b, -OH’s), 4.443 (1H, dd), 4.398 (1H, d), 4.380 (1H, d), 4.296 (3H, m), 4.108
(1H, m), 4.018 (1H, dd), 3.847 (1H, m), 3.772 (3H, 5), 3.765 (3H, s), 3.467 (2H, m),
3.437 (2H, m); 6c (75.5MHz; CDC13): 168.262, 167.887 (QOOCH3), 110.984, 110.699
(C(CH3)2), 78.794, 78.462, 76.931, 76.217, 70.531, 69.957 (C-3, C-4, C-S), 53.333,
53.217 (COOQH3), 48.552, 45.776 (C-2), 34.055, 31.836 (C-6), 27.173, 27.161, 27.071,

26.791 (C(QH3)2); FT-IR (CHC13): 3494, 3096, 1740, 1373, 1216, 1061 cm".

225

2,6-Dibromo-2,6-dideoxy-3,S-O-isopropylidene—D-manno-y-lactone (28):

2,6-Dibromo-2,6-dideoxy-D-manno-1,4-1actone 23 (1.0 g, 3.3 mmol) was
dissolved in 10 mL dichloromethane and 2,2-dimethoxypropane (0.69 g, 6.6 mmol) was
added in. p-Toluenesulfonic acid monohydrate (0.05 g) was added in. The reaction
mixture was stirred under room temperature for 2 hours and subsequently poured into
cold saturated sodium bicarbonate solution and extracted by 2 x 20 mL chloroform. The
organic layer was dried, evaporated and purified by column chromatography to give
methyl 2,6-dibromo-2,6-dideoxy-3,S-O-isopropylidene-D-manno-y—lactone 28 as
colorless oil. Yield: 0.88 g, 71%. 6” (500MHz; CDC13): 1.43, 1.45 (2 x 3H, 2 x s,
C(CI_13)2), 3.52 (1H, dd, .153 6.0Hz, 16,61 11.1Hz, H-6’), 3.62 (1H, dd, 15,6 3.9Hz, H-6),
3.89 (1H, ddd, 14,5 7.2Hz, H-S), 4.46 (1H, dd, 12,3 5.8Hz, 13.4 3.6Hz, H-3), 4.61 (1H, dd, H-
4), 4.71 (1H, d, H-2); 5c (125MHz; CDCl3): 23.84, 24.02 (C(QH3)2), 32.40 (C-6), 41.69

(C-2), 67.43, 70.93, 80.32 (C-3, C-4, C-S), 102.92 (Q(CH;)Z), 170.89 (QOOCH3).

Methyl 2,6-dibromo-2,6-dideoxy-3,4-O-isopropy1idene-D-mannonate (29):

2,6-Dibromo-2,6-dideoxy—D—manno-1,4-lactone 23 (1.0 g, 3.3 mmol) was
dissolved in 10 mL methanol and 2,2-dimethoxypropane (0.69 g, 6.6 mmol) was added
in. p-Toluenesulfonic acid monohydrate (0.05 g) was added in. The reaction mixture was
stirred under room temperature for 12 hours. This mixture was poured into cold saturated
sodium bicarbonate solution and extracted by 2 x 20 mL chloroform. The organic layer
was dried, evaporated and purified by column chromatography to give methyl 2,6-
dibromo-2,6-dideoxy-3,4-O-isopropylidene—D-mannonate 29 as colorless oil. Yield: 0.93

g, 75%. 8” (500MHz; CDC13): 1.37, 1.41 (2 x 3H, 2 x s, C(Cfl3)2), 2.62 (1H, d, J ~ 6.9Hz,

226

-Ofl), 3.58 (1H, dd, 15560112, .163 10.5Hz, H-6’), 3.72 (1H, dd, J53 3.0Hz, H-6), 3.77
(1H, m, H-5), 3.79 (3H, s, COOCH3), 4.18 (1H, dd, 13,4 3.8Hz, 14,5 8.1Hz, H-4), 4.33 (1H,
d, .123 9.0Hz, H-2), 4.64 (1H, dd, H-3); 8c (125MHz; CDC13): 28.27, 28.35 (C(QH3)2),
37.29 (C-6), 45.67 (C-2), 53.24 (COOQHg), 71.51, 79.78, 80.27 (C-3, C-4, C-5), 111.96

(_C_(CH3)2), 168.19 (_C_OOCH3).

2,6-Dibromo-2,6-dideoxy-3,4-O-isopropy1idene-D-iditol (31) and 2,6-dibromo-2,6-

dideoxy-4,5-O-isopropylidene-D-iditol (32):

2,6-Dibromo-2,6-dideoxy-D-iditol 18 (0.46 g, 1.5 mmol) was suspended in 5 mL
acetone and p-toluenesulfonic acid (0.03 g) was added in. The solution was stirred at
room temperature for 15 minutes and poured into the cold saturated sodium bicarbonate
solution. The mixture was extracted with chloroform (3 x 20 mL). The organic layer was
dried and evaporated to give a mixture of 2,6-dibromo-2,6-dideoxy-3,4-O-
isopropylidene-D-iditol 31 and 2,6-dibromo-2,6-dideoxy-4,5-O-isopropylidene-D-iditol
32 in a ratio of 5:1. Yield: 0.35g, 74%. If the reaction time was extended to 12 hours, the
ratio of 31:32 became about 1:1. For 31: [01]D = -47.1 (c = 1.2, CHC13); 5H (500MHz;
CDC13): 1.41, 1.44 (2x 3H, 2 x s, C(CHfiz), 2.98 (2H, b, -OH’s), 3.47 (2H, m, H—6, 6’),
3.84 (dt, 1H, J = 1.9Hz, J = 6.6Hz, H-2), 3.90 (dd, 1H, J = 5.9Hz, J = 12.1Hz, H-l’), 3.94
(1H, m, H-l), 4.12 (1H, m, H-5), 4.21 (dd, 1H, J = 3.2Hz, J = 8.0Hz, H-4), 4.26 (dd, 1H,
J = 1.6Hz, J = 8.1Hz, H-3); 5c (75.5MHz; CDC13): 110.389 (C(CH3)2), 78.812, 76.177,
69.337 (C-3, 04, OS), 64.824 (C-l), 54.746 (C-2), 34.314 (06), 27.128, 27.000

(C(QH3)2); FT-IR (CHC13): 2943, 1373 cm"

227

2,6-Dibromo-2,6-dideoxy-3,4-O—isopropylidene-D-mannitol (33):

2,6-Dibromo-2,6-dideoxy-D-mannitol 24 (0.32 g, 1.0 mmol) was dissolved in 5
mL acetone and p-toluenesulfonic acid (0.03 g) was added in. The solution was stirred
under room temperature for 6 hours and poured into cold sodium bicarbonate solution.
The mixture was extracted by chloroform (2 x 15 mL). The organic layer was dried and
evaporated to give 2,6-dibromo-2,6-dideoxy—3,4-O-isopropylidene—D-mannitol 33 as
colorless oil. Yield: 0.27g, 75%. [0111) = -2.9 (c = 1.0, CHC13); 6H (500MHz; CDC13):
1.34, 1.42 (2 x 3H, 2 x s, C(CH3)2), 3.35 (2H, b, -OH’s), 3.53 (dd, 1H, J53~ 6.8Hz, J65
10.7Hz, H-6’), 3.71 (dd, 1H, 15,6 2.6Hz, H-6), 3.79 (ddd, 1H, 14,5 9.1Hz, H-5), 3.91 (dd,
1H, 11;; 5.6Hz, J11 12.6Hz, H-l ’), 3.96 (dd, 1H, .1 L2 4.9Hz, H-l), 4.09 (dd, 1H, J3... 4.9Hz,
H-4), 4.26 (dd, 1H, 12,3 5.5Hz, H-2), 4.42 (t, 1H, H-3); 8C (75.5MHz; CDC13): 110.962
(C(CH3)2), 80.268, 79.722, 72.642 (C-3, C-4, C-S), 63.843 (C-l), 56.133 (C-2), 37.548

(C-6), 27.775, 27.622 (C(QH3)2).

Methyl 2,6-dibromo-2,6-dideoxy-3 ,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’ ,3 ’ -diyl)-D-idonate

(34):

2,6-Dibromo-2,6-dideoxy—D-idono-1,4-lactone 23 (1.1 g, 3.6 mmol) was
dissolved in 10 mL dry methanol. Butane-2,3-dione (0.34 g, 4.0 mmol) and
trimethylorthoformate (1.2 g, 11.3 mmol) was added in. Borontrifluoride etherate (0.15 g)
was added in and the solution was stirred under room temperature for 12 hours.
Triethylamine (0.5 g) was added and the solution was evaporated. The residue was
purified by column chromatography to give two major isomers methyl 2,6-dibromo-2,6-

dideoxy-3,4-O-(1,2-diacetal)-D-idonate 34 and methyl 2,6-dibromo-2,6-dideoxy—4,5-O-

228

(1,2-diacetal)-D-idonate 35 with ~ 1:1 ratio. Total yield: 1.06 g, 65%. For 34: SH
(500MHz; CDC13): 1.31, 1.35 (2 x 3H, 2 x s, CH3OCCL13 in diacetal), 2.70 (1H, d, 15.0”
7.3Hz, —Ofl), 3.29, 3.33 (2 x 3H, 2 x s, Cfl30CCH3 in diacetal), 3.47 (1H, dd, 15.3 3.9Hz,
.165 11.7Hz, H-6’), 3.56 (1H, dd, 15.62.7Hz, H-6), 3.77 (3H, s, COOCfl3), 3.98 (1H, H-3),
4.15 (1H, dd 13,4 1.4Hz, .145, 9.7112, H-4), 4.32 (1H, ddd, H-5), 4.52 (1H, d, J23 8.9Hz, H-
2); 5c (125MHz; CDC13): 17.32, 17.03 (CH3OCQH3 in diacetal), 33.07 (C-6), 48.24 (C-
2), 449.25, 49.89 (CH3OCCH3 in diacetal), 53.29 (COO_C_H3), 70.00, 71.30, 72.61 (C-3,

C-4, 05), 100.89, 101.28 (CH3OQCH3), 168.56 (COOCH3).

Methyl 2,6-dibromo-2,6-dideoxy—4,5-O-(2 ’ ,3 ’-dimethoxybutane-2 ’,3 ’-diyl)-D-idonate

(35):

2,6-Dibromo-2,6-dideoxy-D-idono-1,4-lactone 23 (1.2 g, 3.9 mmol) was
dissolved in 10 mL dry methanol. Butane-2,3-dione (0.37 g, 4.3 mmol) and
trimethylorthoformate (1.26 g, 11.8 mmol) was added. Borontrifluoride etherate (0.15 g)
was added and the solution was stirred under room temperature for 64 hours.
Triethylamine (0.5 g) was added and the solution was evaporated. The residue was
purified by column chromatography to give methyl 2,6-dibromo-2,6-dideoxy—4,5-O-
(2’,3’-dimethoxybutane-2’,3’-diyl)-D-idonate 35. Yield: 1.03 g, 58%. 5“ (500MHz;
CDC13): 1.25, 1.26 (2 x 3H, 2 x s, CH3OCC_H3 in diacetal), 2.80 (1H, (1, 13,0" 8.8Hz, -
OH), 3.16, 3.25 (2 x 3H, 2 x s, CH30CCH3 in diacetal), 3.36 (1H, dd, 15572112, J36-
11.2Hz, H-6’), 3.48 (1H, dd, 15,6 2.6Hz, H-6), 3.75 (3H, s, COOCfl3), 3.86 (1H, dd 13.4
1.5Hz, 14,5 9.5Hz, H-4), 3.99 (1H, dt, 12,3 8.9Hz, H-3), 4.14 (1H, ddd, H-5), 4.52 (1H, d,

H-2); 5c (125MHz; CDCl3): 17.50, 17.72 (CH3OCQH3 in diacetal), 31.22 (C-6), 48.25

229

(C-2), 48.38, 48.57 (QH3OCCH3 in diacetal), 53.31 (COOQH3), 67.74, 69.47, 71.19 (C-3,

C-4, C-5), 99.42, 99.60 (CH3OQCH3), 168.66 (_QOOCH3).

Methyl 2,6-dibromo-2,6-dideoxy-3 ,4,5-O-(3 ’-methoxybutane-2 ’ ,2 ’ ,3 ’-triyl)-D-mannonate

(36):

This compound was prepared using the same procedure as for 35 except 2,6-
dibromo-2,6-dideoxy-D-manno-1,4-lactone 23 was used as the starting material. Yield:
56%. [(111) = + 17.4° (c = 1.0, CHC13); 6H (500MHz; CDC13): 1.23, 1.38 (2 x 3H, 2 x s,
CH3OCCfl3 in diacetal), 3.28 (3H, s, Cfl30CCH3 in diacetal), 3.25 (1H, dd, 15,6- 7.6Hz,
16,5 10.7Hz, H-6’), 3.31 (1H, dd, 15.6 3.2Hz, H-6), 3.80 (3H, s, COOCH3), 34.05 ((1, 1H,
J34 9.7Hz, H-4), 4.25 (m, 1H, H-5), 4.54 (s, 1H, H-2), 4.58 (d, 1H, H-3); 6c (125MHz;
CDC13): 18.466, 17.949 (CHgOCQH; in diacetal), 28.012 (06), 43.822 (C-2), 48.680
(QH30CCH3 in diacetal), 53.172 (COOQHg), 78.162, 75.590, 71.499 (03, C-4, C-5),

108.758, 99.940 (CH3OQCH3), 168.451 (QOOCH3).

2,6-Dibromo-2,6-dideoxy—3,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’ ,3 ’ -diyl)-D-iditol (37) and

2,6-dibromo-2,6-dideoxy—4,5-O-(2 ’ ,3 ’-dimethoxybutane-2 ’,3 ’-diyl)-D-iditol (38):

2,6-dibromo-2,6-dideoxy-D—iditol (1.02 g, 3.3 mmol) was dissolved in 10 mL dry
methanol with 2,2,3,3-tetramethoxymethane (0.65 g, 3.6 mmol) was added. Boron
trifluoride etherate (0.05 g) was added. The reaction mixture was stirred for 12 hours at
room temperature. Triethylamine (0.15 g) was added in. The solution was evaporated.
The residue was purified by chromatography to give a mixture of 37 + 38 in about 1:1
ratio. Yield: 62%; For 37: 6“ (500MHz; CDC13): 1.26, 1.28 (2 x 3H, 2 x s, CH3OCC_H3 in

diacetal), 2.62, 2.94 (2 x 1H, 2 x b, -O_I;l_’s), 3.23, 3.26 (2 x 3H, 2 x s, Cfl3OCCH3 in

230

diacetal), 3.47 (dd, 1H, 15.... 6.4Hz, 16,, 9.9112, H-6’), 3.52 (dd, 1H, 15,, 7.6Hz, H-6), 3.84,
3.94, 4.00 (3 x 1H, 3 x m, H-2,4,5), 4.11 (dd, 1H, J = 1.91121 = 6.2Hz, H-3), 4.14 (dd,
1H, 11:2 2.0Hz, J.,.~ 9.2112, H-l’), 4.26 (dd, 1H, 11.21.5112, H-l); 5C (125MHz; CDC13):
99.422, 99.179 (CH3OQCH3), 69.680, 69.602, 67.118, 64.605 (C-1, 3, 4, 5), 53.106 (C-
2), 48.330, 48.161 (ghoccn3 in diacetal), 32.407 (06), 17.430 (0130(1ch3 in

diacetal).

2,6-Dibromo-2,6-dideoxy-3,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’,3 ’-diyl)- l -O-t-
butyldimethylsilyl-L-iditol (41) and 2,6-dibromo-2,6-dideoxy-4,5-O—(2’,3’-

dimethoxybutane-2 ’,3 ’-diyl)— l -O-t-butyldimethylsilyl-L-iditol (42):

2,6-dibromo-2,6-dideoxy-L-iditol 12 (4.0 g, 13.0 mmol) was dissolved in 40 mL
dry methanol with 2,2,3,3-tetramethoxybutane (2.54 g, 14.3 mmol). Boron trifluoride
ethrate (0.2 g) was added. The mixture was stirred at room temperature for 12 hours.
Triethylamine (0.5 g) was added to neutralize the solution. The reaction mixture was
evaporated to dryness. The residue was dissolved in 10 mL pyridine and t-
butyldimethylsilyl chloride (2.18 g, 14.3 mmol) was added. The solution was stirred for
12 hours and poured into saturated sodium bicarbonate solution with ice. The mixture
was extracted with chloroform (2 x 50 mL). The organic phase was dried and evaporated.
The residue was purified by chromatography to give a mixture of 41 + 42. For 41: 8"
(500MHz; CDC13): 0.04, 0.06 (2 x 3H, 2 x s, 2 x SiCH3), 0.87 (9H, s, SiC(CH3)3), 1.27,
1.29 (2 x H, 2 x s, 2 x OC(Cfl3)(OCH3)), 3.24, 3.26 (2 x 3H, 2 x s, 2 x OC(CH3)(OCfl3)),
3.49 (dd, 1H, 15,5 6.7Hz, J63 9.9Hz, H-6), 3.53 (dd, 1H, 15,6 7.2Hz, H-6), 3.80 (dd, 1H,

J... 5.5Hz, 13.10.0112, H-4), 3.84 (m, 1H, H-S), 3.97 (ddd, 1H, 12,, 9.1112), 4.06 (dd, 1H,

231

H-3), 4.23 (dd, 1H, J.;; 1.5Hz, le 9.5Hz, H-l’), 4.26 (dd, 1H, 11.2 1.31-12, H-l); FT-IR

(CHC13): 3389, 2950, 1378, 1122, 1036 cm".

2,3-Anhydrous-6-bromo-6-deoxy-4,5-O-(2 ’ ,3 ’-dimethoxybutane-2 ’ ,3 ’ -diy1)-1-O-t-

butyldimethylsilyl-L-gulitol (43):

The mixture of 41 + 42 (2.1 g, 3.9 mmol) was dissolved in 20 mL methanol.
Ammonium hydroxide (28% in water, 5 mL) was added slowly while stirring. The
mixture was stirred for 12 hours at room temperature and TLC showed complete
consumption of the starting materials. The mixture was evaporated to dryness and the
residue was purified by chromatography to give 2,3-anhydrous-6-bromo-6-deoxy—4,5-O-
(2’,3’-dimethoxybutane-2’,3’-diyl)-1-O-t-butyldimethylsilyl-L-gulitol 43 (Yield: 23%
over 3 steps) and 6-amino-2-bromo-2,6-dideoxy-3,4-O-(2’,3’-dimethoxybutane-2’,3’-
diyl)-l-O-t-butyldimethylsilyl-L-iditol 45 (Yield: 21% over 3 steps). For 43: SH
(500MHz; CDC13): 0.05, 0.08 (2 x 3H, 2 x s, 2 x SiCH3), 0.88 (9H, s, SiC(CH3)3), 1.30
(6H, s, 2 x OC(Cfl3)(OCH3)), 3.22, 3.30 (2 x 3H, 2 x s, 2 x OC(CH3)(OCfl3)), 2.98 (1H,
dd, .123 4.1, 13,4 8.6Hz, H-3), 3.11 (1H, ddd, 12,. 6.0, J” 5.4Hz, H-2), 3.24 (1H, dd, 16,3
11.1, 16.5 8.6Hz, H-6), 3.46 (1H, dd, 14,5 9.7Hz, H-4), 3.48 (1H, dd, 1635 2.9HZ, H-6’), 3.65
(1H, dd, H.,y 11.7Hz, H-l), 3.83 (1H, dd, Hl’), 3.93 (1H, ddd, H-S); 8c (125MHz;
CDC13): -5.47, -5.382 (2 x SiQH3), 17.25, 17.34 (2 x OC(QH3)(OCH3)), 18.17
(OSiQ(CH3)3), 25.79 (OSiC(QH3)3), 30.84 (C-6), 47.95, 48.04 (2 x OC(CH3)(OC_H3)),
55.24, 56.69 (C-2, C-3), 61.42 ( C-l), 69.20, 69.60 (C-4, C-5), 98.66, 99.24 (2 x
OQ(CH3)(OCH3)); FT-IR (CHC13): 3557, 3021, 1462, 1373, 1252 cm"; For 45: 8“
(500MHz; CDC13): 0.05, 0.08 (2 x 3H, 2 x s, 2 x SiCH3), 0.88 (9H, s, SiC(CH3)3), 1.28,

1.30 (2 x 311, s, 2 x OC(C_I_13)(OCH3)), 3.23, 3.26 (2 x 3H, 2 x s, 2 x OC(CH3)(OCfl3)),

3.04(11-1, dd, H-6), 3.18 (1H, H-6’), 3.82 (1H, dd, J...‘ 10.1, 11,2 5.9Hz, H-l), 3.93 (1H, m,
H-5), 3.97 (1H, dd, .145 1.7, .143 9.2112, H-4), 4.05 (1H, dd, .121 8.6Hz, H-l’), 4.12 (1H,
ddd, .123 1.6112, H-2), 4.21 (1H, dd, H-3), 4.33 (3H, b, +NH3); 8c (125MHz; CDC13): -
5.39, -5.09 (2 x SiQH3), 17.46, 17.70 (2 x OC(QH3)(OCH3)), 18.25 (OSiQ(CH3)3), 25.86
(OSiC(_C_H3)3), 43.46 (C-6), 48.16, 48.20 (2 x OC(CH3)(O_C_H3)), 52.48 (C-2), 63.78,

64.91, 66.98, 72.37 (C-1, C-3, C-4, C-S), 99.25, 99.33 (2 x OQ(CH3)(OCH3)).

3,4-O-(2 ’ ,3 ’ -Dimethoxybutane-2 ’,3 ’-diyl)-6-O-t-butyldimethylsilyl- l -deoxyno j irimycin

(46):

6-amino-2-bromo-2,6-dideoxy-3,4-O-(2 ’,3 ’-dimethoxybutane-2 ’,3 ’-diyl)- l -O-t-
butyldimethylsilyl-L-iditol 45 (0.46 g, 0.97 mmol) was dissolved in 10 mL toluene and
DBU (0.30 g, 2.0 mmol) was added in. The mixture was refluxed for 16 hours and
subsequently evaporated to dryness. The residue was purified by chromatography to give
3,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’,3 ’-diyl)-6-O-t-butyldimethylsilyl- l -deoxynojirimycin
46. Yield: 0.28 g, 74%. 6” (500MHz; CDC13): 0.00, 0.02 (2 x 3H, 2 x s, 2 x SiCH3), 0.85
(9H, s, SiC(CH3)3), 1.27, 1.30 (2 x 3H, 2 x s, 2 x OC(CH_3)(OCH3)), 3.21, 3.26 (2 x 3H, 2
x s, 2 x OC(CH3)(OCfl3)), 2.55 (1H, dd, 11,1: 12.8, J.,210.1Hz, H-l), 2.65 (1H, ddd, 14.5
9.3, 15,6 2.9, 15.3 4.2Hz, H-5), 3.2 (1H, H-l ’), 3.50-3.59 (2H, m, H-3, H-4), 3.62 (1H, ddd,
12y 5.1, J23 9.2Hz, H-2), 3.71 (1H, dd, .161,~ 9.7Hz, H-6), 3.76 (1H, dd, H-6’); 8c
(125MHz; CDC13): -5.56, -5.44 (2 x SiQH3), 17.71 (2 x OC(QH3)(OCH3)), 18.21
(OSi_C_(CH3)3), 25.87 (OSiC(_C_H3)3), 47.86 (2 x OC(CH3)(O_CH3)), 49.52 (C-l), 58.19 (C-

5), 61.26 (C-6), 67.08, 69.75, 75.56 (C-3, C-4, 05), 99.38, 99.42 (2 x O§(CH3)(OCH3)).

233

l-Deoxynojirimycin (1):

3,4-0-(2 ’,3 ’ -Dimethoxybutane-2 ’,3 ’-diyl)-6-O-t—buty1dimethylsilyl- 1 -
deoxynojirimycin 46 (0.22 g, 0.6 mmol) was dissolved in 5 mL methanol and
hydrochloric acid (1 mL) was added in. The mixture was stirred at room temperature for
30 minutes and evaporated to give l-deoxynojirimycin 1 as its hydrochloric acid salt.
Yield: 0.10 g, 95%. The physical data match literatures. 5H (500MHz; D20): 3.93 (dd, J =
11.7, 3.1 Hz, 1H), 3.86 (dd, J= 11.7, 5.1 Hz, 1H), 3.77 (ddd, J: 12.2, 9.5, 5.1 Hz, 1H),
3.57 (t, J = 9.5 Hz, 1H), 3.50 (t, J = 9.5 Hz, 1H), 3.46 (dd, J = 12.2, 5.1 Hz, 1H), 3.16
(ddd, J: 3.1, 5.2, 9.5 Hz, 1H) 2.95 (t, J = 12.2 Hz, 1H); 6c (125MHz; D20): 79.1, 70.9,

70.1 (C-2, 3, 4), 62.9 (C-6), 60.9 ((3-5), 49.0 (or).

2,6-Dibromo-2,6-dideoxy-D-glucono-1 ,4-lactone (49):

D-mannose (10 g, 55.6 mmol) was dissolved in 100 mL distilled water with
sodium bicarbonate (6.8 g, 81.0 mmol). The solution was cooled down to 0°C and 3 mL
bromine was added in slowly in 30 minutes. The mixture was slowly warmed up to room
temperature and stirred for 4 days. Air was bubbled through the reaction mixture for 6
hours until the solution became pale yellow. The solution was evaporated to dryness.
Acetic acid (2 x 25 mL) was added and evaporated. To the resulted mixture 30%
hydrogen bromide in acetic acid (50 mL) was added in. The mixture was then stirred for 6
hours. Methanol (100 mL) was added and stirred for 10 hours. The solution was
evaporated and co-evaporated with water (2 x 25 mL). The residue was dissolved in 200
mL ethyl acetate and washed by sodium bicarbonate solution (50 mL), brine (25 mL) and

water (25 mL). The organic layer was dried and evaporated to give 2,6-dibromo-2,6-

234

dideoxy-D-glucono-l,4-lactone 49 as brown oil. Yield: 10.6g, 62%; (lit. [4] mp. 90-

92°C; [(11020 + 29° (c = 2, ethyl acetate)).
2,6-Dibromo-2,6—dideoxy-D-glucitol (50):

2,6-Dibromo-2,6-dideoxy-D-g1ucitol 50 and the compounds 51-55 were prepared
according to the same procedure described above for the synthesis of deoxynojirimycin.
Yield: 80%. SH (500MHz; D20): 3.53 (dd, 1H, J55 5.2Hz, 16.61 11.1Hz, H-6’), 3.59 (dd,
1H, 15,6 2.7Hz, H-6), 3.63 (m, 1H, H-3), 3.75 (m, 1H, H-S), 3.76 (dd, 1H, In 6.3Hz,
J;,.~12.7Hz, H-l’), 3.83 (dd, 1H, 1.,2 4.1Hz, H-l), 3.92 (m, 1H, H-4), 4.15 (ddd, 1H, 12,3
10.6Hz, H-2); 5c (125MHz; D20): 71.794, 70.288, 69.879 (C-3, 4, 5), 63.330 (C-l),

59.674 (C-2), 37.679 (C-6); FT-IR (CHC13): 3292, 2934, 1470, 1420, 1083 cm".
2,6-Dibromo-2,6-dideoxy-3 ,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’,3 ’-diy1)-D-glucitol (51):

6“ (500MHz; CDC13): 1.38, 1.42 (2 x 3H, 2 x s, CH3OCCI_13 in diacetal), 3.29,
3.38 (2 x 3H, 2 x s, Cfl30CCH3 in diacetal), 3.706 (2H, m), 3.963 (3H, m), 4.084 (1H,
m), 4.140 (1H, m), 4.639 (1H, m); 8c (125MHz; CDC13): 100.606, 100.230
(CH3OQCH3), 74.107, 72.959, 71.879, 64.898 (01, 3, 4, 5), 56.525 (C-2), 49.42],

48.840 (_QH3OCCH3 in diacetal), 38.354 (C-6), 16.698, 16.619 (CH3OCQH3 in diacetal).

2,6-Dibromo-2,6-dideoxy-3,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’ ,3 ’-diy1)-l -O-t-

butyldimethylsilyl-D-glucitol (52):

8” (500MHz; CDC13): 0.04 (6H, s, Si(Cfl3)2), 0.86 (9H, s, SiC(CH3)3), 1.37, 1.42
(2 x 3H, 2 x s, CH3OCCfl3 in diacetal), 3.24 (1H, m), 3.27, 3.36 (2 x 3H, 2 x s,

Cfl30CCH3 in diacetal), 3.28 (1H, m), 3.70 (2H, m), 3.77 (1H, m), 3.98 (1H, m), 4.09

235

(1H, m), 4.50 (1H, m); 8c (125MHz; CDC13): 100.606, 100.230 (CH3OQCH3), 74.107,
72.959, 71.879, 64.898 (C-1, 3, 4, 5), 56.525 (C-2), 49.421, 48.840 (CH3OCCH3 in
diacetal), 38.354 (06), 16.698, 16.619 (CH3OCQH3 in diacetal); FT-IR (CHC13): 3392,

3095, 1377, 1121cm".

2,6-Imino-2,6-dideoxy-3,4-O-(2 ’ ,3 ’-dimethoxybutane-2 ’ ,3 ’-diy1)-l-O-t-

butyldimethylsilyl-D-glucitol (54):

Yield: 33% from 2,6-Dibromo-2,6-dideoxy-D-glucitol 50. FAB-MS: m/z calcd.
for C.3H37NO(,Si 391.2390, found 392.2466 (MH+); 8H (500MHz; CDC13): 0.04, 0.05 (2 x
3H, 2 x s, 2 x SiCH3), 0.88 (9H, s, SiC(CH3)3), 1.31, 1.37 (2 x 3H, 2 x s, 2 x
OC(CH_3)(OCH3)), 2.05 (b, -OH), 2.42 (1H, ddd, .145 9.6, 15,6 2.6, 15.6. 5.3Hz, H-S), 2.73
(1H, dd, JLI- 13.7, 11321.7Hz, H-l’), 3.09 (1H, 1.3 2.7Hz, H-l), 3.28, 3.37 (2 x 3H, 2 x s,
2 x OC(CH3)(OCH3)), 3.74 (dd, 1H, 16.5 9.9Hz, H-6’), 3.82 (dd, 1H, H-6), 3.92 (dd, 1H,
12,; 2.9HZ, H-2), 3.98 (dd, 1H, 13,4 10.8Hz, H-3), 4.13 (dd, 1H, H-4); 5c (125MHz;
CDC13): 18.332 (OSiQ(CH3)3), 18.746 (2 x OC(QH3)(OCH3)), 25.866 (OSiC(QH3)3),
48.127, 47.875 (2 x OC(CH3)(OQH3)), 49.022 (C-l), 60.677 (05), 62.111 (06), 75.653,

68.1 13, 67.908 (C-3, C-4, C-5), 101.498, 101.282 (2 x OQ(CH3)(OCH3)).

l-Deoxymannojirimycin (DMJ) (55):

Yield: 98%; 8” (500MHz; D20): 3.03 (ddd, 1H,15,6 3.3Hz, 15,... 6.8112, 14,, 10.3112,
115), 3.12 (dd, 1H, 1.31.5112, J.,.~ 13.7112, H-l’), 3.29 (dd, 1H, 1,; 3.1112, H-l), 3.57
(dd, 1H, 1;. 3.1Hz, 13,. 9.5Hz, 11-3), 3.71 (dd, 1H, 16,3 12.7112, H-6’), 3.74 (dd, 1H, 11-4),
3.87 (dd, 1H, H-6), 4.12 (dt, 1H, H-2); at (125MHz; 020): 47.68 (C-l), 58.26 (05),

60.51 (C-6), 65.88, 66.03, 72.57 (C-3, C-4, C-5).

236

REFERENCE

[1] Lundt, 1. Topics in Current Chemistry 1997, 187, 1 17-56.

[2] Pedersen, C.; Bock, K.; Lundt, I. Pure and Applied Chemistry 1978, 50, 1385-400.
[3] Book, K.; Lundt, 1.; Pedersen, C. C arbohydr. Res. 1979, 68, 313-19.

[4] Bock, K.; Lundt, 1.; Pedersen, C. Carbohydr. Res. 1981, 90, 7-16 and 17-26.

[5] Book, K.; Lundt, I.; Pedersen, C. Carbohydr. Res. 1982, 104, 79-85.

[6] Bock, K.; Lundt, 1.; Pedersen, C.; Refn, S. Acta. Chem. Scand. 1984, 838, 555-61.

F14 “m3“

[7] Bock, K.; Lundt, 1.; Pedersen, C.; Refii, S. Acta. Chem. Scand. 1986, B40, 740.
[8] Hearon, W. M.; Witte, J. F. US. patent 1982, 4 pp.

[9] Lundt, 1.; Madsen, R. Synthesis 1993, 714 and 720.

[10] Clode, D. Chem. Rev. 1979, 79, 491.

[l 1] Ley S. V.; Baeschlin D. K.; Dixon D. J.; Foster A. C.; Ince S. J.; Priepke H. W.;
Reynolds D. J. Chemical reviews 2001, 101, 53-80.

[12] Ley, S. V.; Priepke, H. W. M.; Wan‘iner, S. L. Angew. Chem, Int. Ed. 1994, 33,
2290.

[13] Montchamp, J. L.; Tian, F.; Hart, M. E.; Frost, J. W. .1. Org. Chem. 1996, 61, 3897.

[14] Douglas, N. L.; Ley, S. V.; Osborn, H. M. 1.; Owen, D. R.; Priepke, H. W. M.;
Warriner, S. L. Synlett 1996, 793.

[15] Berens, U.; Leckel, D.; Oepen, S. C. J. Org. Chem. 1995, 60, 8204.

[16] Hense, A.; Ley, S. V.; Osborn, H. M. 1.; Owen, D. R.; Poisson, J. F.; Warriner, S. L.;
Wesson, K. E. J. Chem. Soc., Perkin Trans. 1 1997, 2023.

[17] Rudge, A.; Collins, 1.; Holmes, A.; Baker, R. Angewandte Chemie 1994, 106, 2416-
18.

[18] Matos, C.; Lopes, R.; Lopes, C. Synthesis 1999, 571-573.

237

[19] Somfai, P.; Marchand, P.; Torsell, S.; Lindstrom, U. Tetrahedron 2003, 59, 1293-
1299.

[20] Comins, D.; Fulp, A. Tetrahedron Lett. 2001, 42, 6839-6841.

Appendix

NMR spectra of selected compounds

239

o... QN

lei ..llrlplLlr-Lll_1 ILLl_ k W

3: can
a...” 3 3. ed as

H Ll..r|_l- _ m , _ _ _ _ _ _ _ _ _ __ _ _-l1t“l1w..-lll b-1101

 

_
_

 

 

 

 

 

 

 

512712-14“ .2

or 2:09:00 - N 839.0

 

 

 

 

240

 

 

 

 

 

 

 

 

0.. 9:69:00 - ~53ch

241

 

 

 

 

 

 

 

 

 

a

 

0 J}?

 

 

23 1

242

3 9:59:00 - N 53sz

 

 

 

 

 

 

Lrl

2:.

 

 

 

 

3 9:89:00 - N 53ch

243

E. can
as a.»

 

 

 

 

1_1- 1P rr+~1fp11C1rbb CL C _ _ rL1N HF ._1r h V _ L h F# _
11.11.. 141:1... - 1.52

. ._. . c t : __.

M r_ 1

,

 

 

mN uczanoo - N .9350

244

 

 

 

 

 

 

2:. :2. 2K

 

 

 

mN oczanoo - N .2920

245

 

 

 

o.m co_.
1 s 111111 1 l 11._111 11111-1 1-1_1- 1
d: 2.2

 

 

mN 9:69:00 - Neofimco

 

 

246

 

 

2:. one OON

 

 

 

mN uczanoo - Numfimco

247

 

 

 

 

 

 

 

 

 

 

 

 

m
ov cczoquo - Neofimco fl

 

248

 

 

 

 

 

 

 

 

 

p

ov ucaoquo - N 5520

249

E. can

 

 

 

 

 

 

o4. O.N O.” 6.? 9m o6 ON 0.»
bLL1FIF121.L11_11211_r1._ _ _ _ H _ h F p P M k_ F _ _ P _ F L _ H k _ _ F P H R w
14.. . a tfll‘ a“.

 

 

 

..3 12242.-

 

mv canon—too - Neofimco

250

 

 

 

 

 

9. 9:69:00 - Neofimco

251

 

 

 

 

 

a 1.321
_

om ucaanoo - Neoamco

252

 

 

 

 

 

 

 

 

 

 

om ucaanoo - N 55sz

253

 

 

 

3 6.... 3. 6e 3 E a.»
[F 1 1 _ _ _ _ _ H C _ _ H a ~ ~ ~ _ L, b h _ h F P g H _ F _ _ _
11.1. -(-_..11£_...$..11.2..q1 12.212211

 

 

 

Nm uczanoo - N .0320

254

 

 

 

255

 

 

 

Nm 9:3qu0 - N .0320

 

9N c.» 96 c.m c6 o.»

___Hb__L_r%LF2____~_fi___P_____

 

 

 

 

 

«N 258800 - v .03ch

256

 

 

 

 

 

 

 

 

cor

 

1

o

2.

E. 2%.
8a

.N 958800 - v .0320

257

 

 

 

o.—. ON C.” O.‘ 9m 05 OH o.»
1 L . r _ . . _1 F r . F _ F . . 1— . _ . . . . _ H . _ . _r _ . _» L _ . . _ . . . .
x1111 . 4.1 4.2 4 1 1 Jr

 

 

 

mm uczanoo - $63sz

258

 

 

 

.

E. can
on.

on 9.28.50 - Eofimzo

259

o.—. ed 0...” oi 9m ad ofi 9m

1.11..1H1L111.3..H-....._._1..r_bk..._%..._._.._H..__~_....

 

 

 

 

260

 

 

_. 8 2:828 - $2620

 

E. can

 

 

 

 

mm ucaanoo - v .6320

261

 

 

 

 

E ucaanoo - $639.0

 

262

 

 

_ ___..___ __4

 

E ucaoquo - 3639.0

263

E. :35

 

 

 

 

1.2

Nu ucaanoo - $63ch

264

1

 

 

 

Nu ucaanoo - $0320

265

 

 

 

 

 

 

 

 

 

mu 9:59:00 - $22.20

266

 

 

 

 

 

 

 

cor
.

L11.11L

E. can

1 . . . . . 1.1

 

 

mu nczanoo - v 5320

267

 

E. 5%

 

1111121411

 

 

 

 

 

vm oczanoo - 968ch

268

 

 

 

 

 

 

 

E 938800 - £65sz

269