Investigation on the Synthesis of Shigella flexneri Specific Oligosaccharides Using Disaccharides as Potential Transglucosylase Acceptor Substrates

Article abstract of DOI:10.1021/acs.joc.5b01407

Chemo-enzymatic strategies hold great potential for the development of stereo- and regioselective syntheses of structurally defined bioactive oligosaccharides. Herein, we illustrate the potential of the appropriate combination of a planned chemo-enzymatic pathway and an engineered biocatalyst for the multistep synthesis of an important decasaccharide for vaccine development. We report the stepwise investigation, which led to an efficient chemical conversion of allyl α-d-glucopyranosyl-(1→4)-α-l-rhamnopyranosyl-(1→3)-2-deoxy-2-trichloroacetamido-β-d-glucopyranoside, the product of site-specific enzymatic α-d-glucosylation of a lightly protected non-natural disaccharide acceptor, into a pentasaccharide building block suitable for chain elongation at both ends. Successful differentiation between hydroxyl groups features the selective acylation of primary alcohols and acetalation of a cis-vicinal diol, followed by a controlled per-O-benzylation step. Moreover, we describe the successful use of the pentasaccharide intermediate in the [5 + 5] synthesis of an aminoethyl aglycon-equipped decasaccharide, corresponding to a dimer of the basic repeating unit from the O-specific polysaccharide of Shigella flexneri 2a, a major cause of bacillary dysentery. Four analogues of the disaccharide acceptor were synthesized and evaluated to reach a larger repertoire of O-glucosylation patterns encountered among S. flexneri type-specific polysaccharides. New insights on the potential and limitations of planned chemo-enzymatic pathways in oligosaccharide synthesis are provided.

Full text of DOI:10.1021/acs.joc.5b01407

Subscriber access provided by University of Pennsylvania Libraries
Featured Article
Investigation on the synthesis of Shigella flexneri–specific oligosaccharides
using disaccharides as potential transglucosylase acceptor substrates
Stéphane Salamone, Catherine Guerreiro, Emmanuelle Cambon, Jason M. Hargreaves,
Nathalie TARRAT, Magali Remaud-Siméon, Isabelle André, and Laurence A. Mulard
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.5b01407 • Publication Date (Web): 04 Sep 2015
Downloaded from http://pubs.acs.org on September 8, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Organic Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Investigation on the synthesis of Shigella flexneri–specific oligosaccharides
using disaccharides as potential transglucosylase acceptor substrates
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
,‡,¥
†,‡
§,ǁ,#
†,‡
Stéphane Salamone,
Catherine Guerreiro, Emmanuelle Cambon, Jason M. Hargreaves,
§
,ǁ,#
§,ǁ,#
§,ǁ,#
,†,‡
Nathalie Tarrat, Magali RemaudꢀSiméon, Isabelle André, and Laurence A. Mulard*
†
Institut Pasteur, Unité de Chimie des Biomolécules, 28 rue du Dr Roux, 75724 Paris cedex 15,
‡
§
France ; CNRS UMR 3523, Institut Pasteur, 75015 Paris, France ; Université de Toulouse ;
ǁ
INSA,UPS,INP ; LISBP, 135 Avenue de Rangueil, 31077 Toulouse, France ; CNRS,
#
UMR5504, Fꢀ31400 Toulouse, France ; INRA UMR792, Ingénierie des Systèmes Biologiques
et des Procédés, Fꢀ31400 Toulouse, France
Corresponding author’s eꢀmail: laurence.mulard@pasteur.fr
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract:
Chemoꢀenzymatic strategies hold great potential for the development of stereoꢀ and
regioselective syntheses of structurallyꢀdefined bioactive oligosaccharides. Herein, we illustrate
the potential of the appropriate combination of a planned chemoꢀenzymatic pathway and an
engineered biocatalyst for the multiꢀstep synthesis of an important decasaccharide for vaccine
development. We report the stepwise investigation, which led to an efficient chemical conversion
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of allyl αꢀ
D
ꢀglucopyranosylꢀ(1→4)ꢀαꢀ
L
ꢀrhamnopyranosylꢀ(1→3)ꢀ2ꢀdeoxyꢀ2ꢀtrichloroacetamidoꢀ
βꢀ ꢀglucopyranoside, the product of siteꢀspecific enzymatic αꢀ
D
D
ꢀglucosylation of a lightly
protected nonꢀnatural disaccharide acceptor, into a pentasaccharide building block suitable for
chain elongation at both ends. Successful differentiation between hydroxyl groups features the
selective acylation of primary alcohols and acetalation of a cisꢀvicinal diol, followed by a
controlled perꢀOꢀbenzylation step. Moreover, we describe the successful use of the
pentasaccharide intermediate in the [5+5] synthesis of an aminoethyl aglyconꢀequipped
decasaccharide, corresponding to a dimer of the basic repeating unit from the Oꢀspecific
polysaccharide of Shigella flexneri 2a, a major cause of bacillary dysentery. Four analogs of the
disaccharide acceptor were synthesized and evaluated to reach a larger repertoire of Oꢀ
glucosylation patterns encountered amongst S. flexneri typeꢀspecific polysaccharides. New
insights on the potential and limitations of planned chemoꢀenzymatic pathways in
oligosaccharide synthesis are provided.
Introduction
The chemical assembly of complex oligosaccharides has progressed tremendously over the
ACS Paragon Plus Environment

Page 3 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1ꢀ5
past decades.
Elegant chemical synthetic methods including modular solutionꢀphase
6
7
approaches or automated stepwise solidꢀphase syntheses of oligosaccharides encompassing
more than 20 residues have been reported. Nevertheless, in the absence of general rules
governing synthetic carbohydrate chemistry, the development of stereoselective syntheses of
structurallyꢀdefined oligosaccharides remains a challenging task. Complexity is increased
owing to the impressive structural diversity of naturally occurring carbohydrates, and
therefore of synthetic targets. This is especially relevant in the case of glycans of microbial
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
origin, in which interest is growing. Efficient protecting group manipulation toward tailoredꢀ
made building blocks on the one hand and their stereoselective chemical assembly on the
other hand remain major issues, particularly with regards to the development of scalable
synthetic processes or when multiple targets are involved. In some instances, enzymes have
successfully emerged to circumvent these issues. Advantageously, synthesis using
9
glycosyltransferases proceeds selectively in the absence of protecting groups, while some
1
0
11
glycosidases could be converted into efficient glycosynthases or tranglycosidases.
Evidently, access to naturally occurring carbohydrateꢀactive enzymes is expanding, and in
recent years, enzymatic syntheses and – as an extension – chemoꢀenzymatic pathways have
gained increasing attention owing to their powerful versatility in addition to high regioꢀ and
3,4,12ꢀ14
stereoselectivity.
Purposelyꢀtailored biocatalysts could help bypass specific limitations
encountered with chemical glycosylation, we therefore chose to explore a chemoꢀenzymatic
route in the context of Shigella flexneri, a family of Gramꢀnegative enteroinvasive bacteria
and one of the causes of shigellosis – otherwise known as bacillary dysentery – in humans.
1
5
With an estimated 122,800 shigellosisꢀrelated deaths in 2010, shigellosis is one of the
diarrheal diseases causing the most burden worldwide, especially among children living in
1
6
17
developing countries. Rehydration is inadequate and resistance to antibiotics is increasing.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Above all, the renewed awareness of the burden of shigellosis in the paediatric population has
encouraged the search for novel vaccine strategies aimed at broad Shigella species and type
coverage. As an alternative to important developments involving the detoxified
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
18
lipopolysaccharide and/or analogues purified from biological sources, we and others chose
1
9ꢀ21
to investigate the impact of synthetic carbohydrates as active vaccine components.
Thus,
synthetic oligosaccharides representing functional mimics of the Oꢀspecific polysaccharide
moiety (OꢀSP) of the bacterial membraneꢀanchored lipopolysaccharides are meant to replace
2
2
the natural antigens. The efficient synthesis of tailored fragments of the Shigella OꢀSPs of
interest is a preꢀrequisite to the identification of oligosaccharides suitable for entering
vaccine design and optimization.
S. flexneri encompasses a large diversity of types and subtypes, which have been identified
based on the structure of their OꢀSPs. At least 15 of the identified OꢀSP repeating units share
the same linear tetrasaccharide backbone, which is made of three
C) and a Nꢀacetylꢀ ꢀglucosamine residue (D), 1,2ꢀtrans linked to one another. Type
specificity is associated to the phageꢀencoded site selective modification of the ABCD unit
Lꢀrhamnose residues (A, B,
D
with αꢀ
D
ꢀglucopyranosyl residues (E), Oꢀacetyl groups (Ac), and the more recently disclosed
23,24
phosphoethanolamine residues (PEtN) (Figure 1).
ACS Paragon Plus Environment

Page 5 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. S. flexneri OꢀSP backbone (ABCD) showing type/groupꢀspecific substitution and sites
of αꢀD
ꢀglucosyl (E) appendage (in blue).
αꢀ
DꢀGlucosylation can occur at any residue. It is always stoichiometric and is an essential
2
3,25
feature of the most prevalent S. flexneri serotypes sharing the {ABCD} backbone.
n
Concern for these important glucosyl sideꢀchains was addressed repeatedly in the published
chemical syntheses of S. flexneri oligosaccharides, most of which used readily available
2
6ꢀ31
tetrabenzylglucosyl donors.
However, the 1,2ꢀcis αꢀlinkage has precluded stereospecific
glucosylation and α/β mixtures were isolated despite parameter optimization. Evolving from
tetrabenzyl donors to those more diversely protected at remote positions such as analogues
32
equipped with ester protecting groups at Oꢀ6
and conformationally constrained
3
3,34,35
precursors,
have been shown to enhance α/β ratios. Alternatively, preꢀarranged
36,37
donor/acceptor systems for intramolecular aglycon delivery,
mannoseꢀtoꢀglucose
3
8
sequential oxidationꢀreduction, and precursors designed for hydrogen bond mediated
3
9
aglycon delivery, have been employed successfully. Furthermore, impressive progress in
the area of stereocontrolled chemical 1,2ꢀcis glucosylation has been achieved as exemplified
by sophisticated donors enabling neighboring group participation by means of a Cꢀ2 (S)ꢀ
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
40
41ꢀ43
auxiliary, or prepared in the form of precyclized 1,2ꢀoxathiane precursors.
Nevertheless,
despite remarkable realizations, concern emerges due to increasing donor complexity and/or
potential interferences with protecting group selection, branching patterns, subsequent chain
elongation, substituent manipulation, or final deprotection. For systems such as S. flexneri,
for which diversely substituted OꢀSPs and fragments thereof are identified as synthetic
targets, these inherent drawbacks have to be evaluated on a caseꢀtoꢀcase basis. Instead, we
chose to investigate the input of enzymatic glucosylation.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Shigella typeꢀspecific OꢀSP glucosylation takes place on the periplasmic side of the
cytoplasmic membrane on the growing membraneꢀanchored OꢀSP chain. Glucosylation
involves serotypeꢀspecific glucosyltransferases featuring several transmembrane domains and
4
4
a common membraneꢀanchored undecaprenylꢀphosphateꢀβꢀglucose donor. With this in
mind, we chose to investigate novel chemoꢀenzymatic routes to diverse αꢀ ꢀglucosylated
D
oligosaccharides by use of sucroseꢀutilizing transglucosylases. Otherwise known as
glucansucrases, the selected biocatalysts are αꢀretaining enzymes found in Glycosideꢀ
4
5
Hydrolase families 13 (GH13) and 70 (GH70) of the CAZy classification. Owing to a
remarkable plasticity, they have been evolved into biocatalysts with novel substrate
46
specificities. Moreover, relying on computerꢀaided engineering, we have identified mutants
of the amylosucrase from Neisseria polysaccharea (NpAS), which catalyze the regiospecific
αꢀ
of S. flexneri.
circulans 251 cyclodextrin glucanotransferase (CGTase, EC 2.1.4.19) was able to
accommodate nonꢀnatural rhamnoside acceptors and produce αꢀ ꢀglucosylated products,
Dꢀglucosylation of two nonꢀnatural monosaccharides into motifs of interest in the context
4
7,48
Featuring a related strategy, S. Armand et al. have shown that Bacillus
D
identified as relevant building blocks for the synthesis of various S. flexneri OꢀSP
ACS Paragon Plus Environment

Page 7 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
49
fragments.
4
7,49
As seen from our previous achievements,
the strategy that we have undertaken aims at
developing targetꢀoriented synthetic routes encompassing an early stage enzymatic step.
Therefore, acceptor substrates are fineꢀtuned in order to access a variety of glucosylation
profiles compatible with subsequent chemical elongation. Having established the concept of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
programmed chemoenzymatic pathways to microbial carbohydrates by use of purposely
47,49
tailored glycoenzymes in the case of monosaccharide acceptors,
our current effort aims at
investigating the same approach in the case of nonꢀnatural disaccharide acceptors relevant to
the design of a S. flexneri multivalent vaccine. The overall aim is to access oligosaccharides
from several independent S. flexneri OꢀSPs starting from a single disaccharide, designed for
use in acceptor reactions with properly engineered biocatalysts.
Obviously, the choice of the acceptor/transglucosylase system is a key issue governing the
implementation of efficient tailored chemoenzymatic syntheses of complex glucosylated
targets. Since the panel of known transglucosidases is highly diverse, acceptor selection
controls the choice of the whole system. The acceptor has to fulfil several criteria, such as
relevance to epidemiological data and potential for broad serotype coverage. In addition,
compatibility with the chemical manipulation of the product of enzymatic glucosylation into
S. flexneri typeꢀspecific oligosaccharides is an absolute must. Among the four possible
disaccharides found on the backbone shared by most S. flexneri OꢀSPs (Figure 1), the
Lꢀ
rhamnosylꢀ(1→3)ꢀ2ꢀNꢀacetylꢀ ꢀglucosamine moiety (CD) emerged as the one paving the way
D
to the largest diversity of glucosylated products representative of any one of multiple S.
flexneri serotypes.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Enzymatic glucosylation of acceptor substrate 1 into trisaccharide 2 catalyzed by
50
GH70 glucansucrase called GBDꢀCD2.
To fulfil requirements for chemical chain elongation, the lightly protected CD analogue 1
5
0
was preferred to unprotected disaccharide CD as an acceptor substrate (Scheme 1). Allyl
glycoside 1 was shown to act as an acceptor substrate for several glucansucrases from the
5
0
GH70 family, including GBDꢀCD2, an αꢀtransglucosylase engineered from Leuconostoc
mesenteroides NRRL Bꢀ1299 dextransucrase (DSRꢀE). In particular, as a rewarding
5
0
illustration of this powerful concept, we have recently communicated on the highꢀyielding
GBDꢀCD2ꢀmediated enzymatic glucosylation of the selected disaccharide 1 into trisaccharide
2
, which features the αꢀ
D
ꢀGlcpꢀ(1→4)ꢀαꢀLꢀRhap glucosylation pattern characteristic for S.
2
3
flexneri type II (Scheme 1), and on the efficient conversion of the latter into a
pentadecasaccharide hapten entering in the composition of a synthetic carbohydrateꢀbased
2
1
vaccine candidate against S. flexneri 2a (SF2a) infection.
ACS Paragon Plus Environment

Page 9 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Highlights of the synthesis of decasaccharide 5 from the product of enzymatic
glucosylation 2, showing the key protected intermediates 3 and 4. The development of efficient
[
2→3], [3→4], and [4→5] transformations, respectively, corresponds to issues addressed
independently in the manuscript.
Herein, we report on the stepwise investigation which led to an efficient conversion of the
glucosylation product 2 into a pentasaccharide building block 4, suitable for chain elongation
at both ends (Scheme 2), as demonstrated in the synthesis of an aminoethyl aglyconꢀequipped
51
23
dimer (5) of the basic repeating unit from SF2a OꢀSP.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Chemical structures of the CD disaccharide acceptor analogues modified at position 4C
(
6, 7) or 1 (8, 9).
D
Moreover, we also report on the synthesis of analogues of the GBDꢀCD2 acceptor substrate 1,
featuring a masked 4 ꢀOH (6 and 7) or modified at the reducing end (8 and 9) (Figure 2), and on
C
their evaluation as substrates of available glucansucrases differing in terms of substrate and
regiospecificity.
Results and Discussion
Having trisaccharide 2 produced in high yield from the easily accessible GBDꢀCD2 acceptor
5
0
substrate 1, our first objective was to demonstrate that the former could be converted into either
a donor or an acceptor for use in glycosylation reactions. Since the CD disaccharide 1 had been
designed toward this aim, the ECD trisaccharide 2 fulfils primary requirements at its D residue. It
features a temporary allyl protecting group at the reducing end, and a 2 ꢀNꢀtrichloroacetyl moiety
D
to ensure efficient anchimeric assistance in the DꢀA glycosylation step along with recovery of the
2Dꢀacetamido moiety upon final deprotection. However, trisaccharide 2 also has eight free
ACS Paragon Plus Environment

Page 11 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
hydroxyl groups, of which two are primary alcohols. Therefore, careful protection/deprotection
sequences were required to allow elongation at the reducing end and/or at OHꢀ3 of the C residue
of precursor 2 so as to reach more complex SF2a oligosaccharides.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Synthesis of monoꢀacetonide 17: the acetonide route. Reagents and conditions: Aꢀa)
2,2ꢀDimethoxypropane (DMP), cat pꢀtoluenesulfonic acid (pꢀTSA), DMF/acetone 1:1, 7.5 h,
86%; b) 50% aq AcOH, 0 °C, 4.5 h, 89%. Bꢀa) 2ꢀMethoxypropene, cat camphorsulphonic acid
(
CSA), acetone, 2 h, 49% for 13, 39% for 14; b) cat I , H O, MeCN, 2.75 h, 76%; c) Ac O, cat
2
2
2
DMAP, Py, 2 h, 87%.
Owing to the cis configuration of the 2 ,3 ꢀdiol moiety, the acetonideꢀprotected 17 was seen
C
C
as an ideal intermediate to the fully protected 3. The higher stability of 1,3ꢀcisꢀdioxolanes over
1,3ꢀdioxanes towards hydrolysis is a wellꢀacknowledged phenomenon, which has been
5
2ꢀ54
thoroughly exemplified.
Following optimization of the reaction conditions, the expected
chemoselectivity was obtained providing monoꢀisopropylidene 12 in yields up to 89% using
aqueous acetic acid at 0 °C for the selective hydrolysis of the 4,6ꢀacetal in the EC disaccharide
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
55
1
1, itself derived from the known methyl glycoside 10 (Scheme 3A). Hence, trisaccharide 2 was
submitted to a similar twoꢀstep process (Scheme 3B). Treatment of octaol 2 with excess 2ꢀ
methoxypropene in acetone under acidic catalysis ensured its conversion into a mixture of the
expected triꢀisopropylidene 13 and the fully protected 14, isolated in 49% and 39% yields,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
respectively. Acetalation of the 2 ,3 ꢀdiol under these kinetic conditions has been reported
E
E
5
6
previously but it was not perceived to be problematic due to the high lability of the 1,2ꢀtrans
configuration. However, the selective deprotection of the isolated mixture of compounds 13 and
14 proved troublesome. Under mild hydrolysis conditions, ranging from conventional proticꢀ
5
7,58
mediated hydrolysis,
which was efficient on the EC model 11, to less common protocols such
5
9
as the use of molecular iodine in acetonitrile or that of silica gelꢀsupported phosphomolybdic
6
0
acid, the deprotection pattern was similar despite different kinetics. As expected, the 2 ,3 ꢀOꢀ
E
E
isopropylidene was rapidly cleaved to give the sole intermediate 13. Furthermore, the 6ꢀ
membered 4 ,6 ꢀacetal was rather labile providing access to tetraol 15, as unambiguously
E
E
ascertained following isolation and perꢀacetylation into the fully protected 16. In contrast, the
,3ꢀdioxane masking the 4 ,6 ꢀdiol was comparably stable to the 1,3ꢀcisꢀdioxolane blocking
1
D
D
positions 2 and 3 . Due to this unexpected observation, the reaction had to be monitored
C
C
carefully in order to avoid complete deprotection. The I /MeCN conditions gave the best yield of
2
monoꢀacetonide 17 (76%) in only a few hours. Yet, in our hands the reaction suffered from lack
of reproducibility, with yields ranging from 37% to 76%, and a more robust pathway to the fully
protected 3 was required.
ACS Paragon Plus Environment

Page 13 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Scheme 4. Enzymatic Oꢀacetylation of disaccharide 1. Reagents and conditions: a) Vinyl acetate,
CALꢀB, THF/Py (4:1) 45 °C, 17 h, 88%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Toward this aim, the preliminary protection of the primary alcohols before installing the
2C,3 ꢀacetonide was investigated as an alternative. In order to extend the weight of the enzymatic
C
input, regioselective Oꢀacetylation by means of the wellꢀknown lipase B from Candida
antarctica (CALꢀB, also known as Novozym 435) in combination with vinyl acetate as donor
61ꢀ63
was first attempted.
Once more, while the selective 6 ꢀOꢀacetylation of the CD disaccharide
D
1 proceeded in high yield (88%) to give monoꢀacetate 18 (Scheme 4), adaptation to trisaccharide
was not as efficient. At best, 8% of the target 6 ,6 ꢀdiacetate – albeit contaminated ꢀ was
2
D
E
obtained, along with a mixture of monoꢀacetates and starting material (not shown). Since these
attempts also clearly demonstrated that the two acetate esters did not sufficiently increase
lipophilicity to allow proper isolation of the target trisaccharide, the enzymatic route was
abandoned.
Scheme 5. Synthesis of monoꢀacetonide 17 and subsequent benzylation: the ester route. Reagents
and conditions: a) BzCl, symꢀCollidine, MeCN/Acetone (1:1), ꢀ40 °C, 46 h, 65%; b) DMP, CSA,
Acetone, 2 h, 92%; c) MeONa, MeOH, 3.3 h, 94%; d) NaH, BnBr, DMF, ꢀ10 °C, 4 h, 78%.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Instead, we turned our attention to the selective chemical instalment of benzoate esters.
Optimization of the reaction conditions enabled the development of a process that accommodated
a good balance in the solubility of the starting material 2 and expected product 19, as well as a
higher reactivity of primary hydroxyls over secondary ones. Thus, octaol 2 was treated with
benzoyl chloride and a hindered nucleophilic catalyst (symꢀcollidine) in a mixture of
acetone/MeCN at ꢀ40 °C for 2 days to give dibenzoate 19 in a reproducible satisfactory 65%
yield. The subsequent 2 ,3 ꢀOꢀisopropylidenation of the latter into tetraol 20 (92%) and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
C
C
debenzoylation (94%) were uneventful and furnished monoꢀacetonide 17 in high yield (Scheme
5). When masking the remaining hydroxyl groups of the latter as benzyl ethers while avoiding Nꢀ
6
4
benzylation, the reaction was carried out at ꢀ10 °C using only a slight excess of benzyl bromide.
Indeed, available data suggest that the trichloroacetamide moiety of amino sugars is prone to Nꢀ
64
benzylation or even cleavage when treated according to various Oꢀbenzylation protocols.
Moreover, after quenching the sodium hydride in excess by addition of MeOH, the resulting
sodium methoxide was neutralized by adding acetic acid to protect the baseꢀlabile
64
trichloroacetamide during evaporation of the volatiles. Under these conditions, the fully
protected trisaccharide 3 was obtained in 78% yield.
ACS Paragon Plus Environment

Page 15 of 76
The Journal of Organic Chemistry
1
2
3
4
5
Scheme 6. Synthesis of pentasaccharide 24 by use of donor 22. Reagents and conditions: a) cat
+
ꢀ
[
Ir(COD){PCH (C H ) }] PF , H , THF, 2.5 h, then I , THF/H O, 5 h, 80%; b) CCl CN, cat
3 6 5 2 6 2 2 2 3
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
DBU, DCE, ꢀ10 °C, 35 min, 70%; c) TMSOTf (10 mol%), DCE, ꢀ35 °C to 0 °C, 30 min, 78%.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In order to use the desired convergent [5+5] strategy to decasaccharide 5 (Scheme 1),
conditions for a high yielding glycosylation at the DꢀA linkage were required. Until recently, the
investigation of convergent syntheses to SF2a oligosaccharides involving a disconnection at the
DꢀA linkage was ruled out in our laboratory, essentially because of problems encountered by
6
5,66
Bundle et al. in the synthesis of S. flexneri Y OꢀSP fragments.
Instead, we favoured a
However, the system described herein differs in many
instances from the halideꢀequipped oligosaccharide donors bearing a 2 ꢀNꢀphthalimido
5
1,67
disconnection at the CꢀD linkage.
D
6
5,66
participating group at their reducing end residue, which served in the S. flexneri Y studies.
Hence, the next focus was on using ECD as a donor in a [3+2] glycosylation to prepare the model
ECDAB pentasaccharide 24 (Scheme 6). To this end, trisaccharide 3 was submitted to a selective
twoꢀstep anomeric deallylation involving conventional allyl to propenꢀ1ꢀyl conversion by means
+
ꢀ
68
69
of the [Ir(COD){PCH (C H ) }] PF catalyst and subsequent iodineꢀmediated hydrolysis,
3
6
5 2
6
giving rise to hemiacetal 21 (80%). The latter was in turn activated in the form of
trichloroacetimidate (TCA) 22 (70%, 93% corrected yield) by treatment with trichloroacetonitrile
70
in the presence of a catalytic amount of DBU. Coupling of the known AB acceptor 23 and
donor 22 at low temperature in DCE containing 10 mol% TMSOTf gave pentasaccharide 24 in a
good 78% yield with complete βꢀselectivity.
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 7. Synthesis of the B(E)CD acceptor 33. Reagents and conditions: a) 50% aq TFA,
CH Cl , 75 min, 94%; b) PhC(OMe) , pꢀTSA, CH Cl , 20 min then 50% aq TFA, 45 min, 82%;
2
2
3
2
2
c) MeC(OMe) , pꢀTSA, MeCN, 35 min, then 80% aq AcOH, 0 °C, 25 min; d) 28, TMSOTf (5
3
mol%), Et O, ꢀ15 °C, 40 min, 81% from 25, via 27; e) H NNH , Py/AcOH (3:2), 0 °C, 70 min,
2
2
2
92%.
Having demonstrated that the product of enzymatic glucosylation 2 could be converted into a
potent donor, we turned our attention to the preparation of a fully protected AB(E)CD building
block 4. First, cleavage of the isopropylidene acetal in trisaccharide 3 by use of a biphasic
aqueous TFA/CH Cl system, revealed the corresponding 2 ,3 ꢀdiol 25 (94%) (Scheme 7). Next,
2
2
C
C
the 2 ꢀhydroxyl group was benzoylated through orthoester formation followed by regioselective
C
2
8
acidꢀmediated opening, yielding acceptor 26 (82%). The known rhamnosyl trichloroacetimidate
7
1
28, bearing a levulinyl ester orthogonal to the benzoyl group at position 2, was the selected
donor. However, none of the [26 + 28] glycosylation attempts, differing in terms of solvent
ACS Paragon Plus Environment

Page 17 of 76
The Journal of Organic Chemistry
1
2
3
(
Et O, CH Cl , toluene) and temperature (ꢀ15 °C, 25 °C, 70 °C), led to any tetrasaccharide
2
2
2
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
formation (not described). Alcohol 26 was recovered unaffected while donor 28 eventually
7
2
rearranged into trichloroacetamide 29. The total absence of glycosylated product was not
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
anticipated, especially since a 2 ꢀbenzoyl ester had been used in the synthesis of SF2a
C
5
1,67,73
oligosaccharides, albeit on different substrates, in our group in the past.
Interestingly, we
have previously pointed to the somewhat poor reactivity of the 2 ꢀacetamidoꢀ4 ,6 ꢀOꢀ
D
D
D
isopropylideneꢀECD acceptor 30, which closely resembles trisaccharide 26, when treated with the
7
4
known 2ꢀOꢀacetyl rhamnosyl donor 31, analogous to 28. In spite of this observation and in
contrast to the present finding, the target glycosylation product had been in this case obtained
7
5
with moderate to good yield. Whereas it was tempting to correlate the low reactivity of acceptor
7
6
30 with the known remote effect of the 2 ꢀacetamido moiety, we had no explanation for the
D
present outcome involving acceptor 26. Changing the 2 ꢀbenzoate in the latter for a 2 ꢀacetate
C
C
provided analogue 27, also obtained as a single product from diol 25 upon regioselective opening
5
0
of an intermediate orthoester. Satisfactorily, the [27 + 28] glycosylation proceeded smoothly to
give tetrasaccharide 32 in a good 81% yield (from 25), thus confirming that ECD could also
serve as an acceptor. Delevulinylation of the fully protected 32 by means of hydrazine acetate
5
0
then led to the B(E)CD acceptor 33 (92%). The striking difference in reactivity between the 2 ꢀ
C
Oꢀbenzoyl acceptor and its 2 ꢀOꢀacetyl counterpart, 26 and 27 respectively, was puzzling. We
C
postulated that it could either be a consequence of a steric clash between the donor and acceptor
2
6, or of an intramolecular hydrogen bond established between the 3 ꢀOH and the ester carbonyl
C
moiety that would only occur in the case of the benzoate from acceptor 26. It was hypothesized
that if present, the latter phenomenon would either lower the electron density on the reactive
hydroxyl group or constrain the molecule in a conformation less favorable for glycosylation. This
hypothesis was addressed first by proper NMR analysis of the two hydroxyꢀesters in an aprotic
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
solvent, which enables the detection of both intraꢀ and intermolecular hydrogen bonds involving
7
7
1
hydroxyl groups. Measuring the H NMR spectrum for each of the two samples in DMSOꢀd
6
solution at 10° intervals from 303 K to 353 K clearly showed linear upfield shifts of the 3 ꢀOH,
C
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
which were almost similar. The shift ranged between 0.002 and 0.008 ppm/°C, indicating that
7
8
both hydroxyl groups were actually engaged in intramolecular hydrogen bonding (Figure 3).
1
Figure 3. Temperature dependence of
δ
(3 ꢀOH) measured by H NMR (400 MHz) for 15 mM
C
solutions of 26 (♦) and 27 (ꢀ) in DMSOꢀd6.
ACS Paragon Plus Environment

Page 19 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Local energy minima conformations adopted by the 2 ꢀOꢀbenzoyl acceptor 26 (a and
C
b) and by the 2 ꢀOꢀacetyl acceptor 27 (c). For clarity, benzyl ethers are truncated and represented
C
as magenta circles. Distances between atoms are given in Å.
In contrast, ab initio quantum chemical calculations of the two disaccharides shed light on
their different conformational behaviors. In the case of the 2 ꢀOꢀbenzoyl trisaccharide 26, two
C
local energy minima were found upon optimization using the DFT/B3LYP/6ꢀ31G(3df,3pd) level
(Figure 4a and 4b, Supporting Information Tables S1 and S2), corresponding to geometries that
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
mainly differ in the orientation of the 3 ꢀOH. In the first situation, the hydroxyl group is involved
C
in a hydrogen bonding interaction with the carbonyl of the vicinal ester, as hypothesized (Figure
4a). In the other minimal energy conformation, which is also more stable by ꢀE = 4.8 kcal/mol,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the 3 ꢀOH interacts with the endocyclic oxygen from the branched glucosyl residue E (Figure
C
4b). In contrast, the 2 ꢀOꢀacetyl trisaccharide 27 adopts a single energy minimum conformation,
C
which closely resembles the more stable conformation in trisaccharide 26. Analogously, it is
stabilized by a hydrogen bond between the 3 ꢀOH group and the endocyclic oxygen from residue
C
E (Figure 4c, Supporting Information Table S3). All attempts to localize an energy minimum in
which the 3 ꢀOH was involved in a hydrogen bond with the acetate carbonyl moiety in
C
trisaccharide 27 were fruitless as calculations repeatedly converged toward one single energy
well. The calculated Mulliken charges on the 3 ꢀoxygen atom for the observed energy minima
C
conformations are 0.76 and 0.85 (Figures 4a and 4b, respectively) for the benzoylated 26 and
0.86 for the acetylated 27, respectively. Since the 3 ꢀOH of the most stable conformations of the
C
two hydroxylꢀesters have similar charges, it could be assumed that the hydrogen bonding had no
crucial influence on the electron density of the 3 ꢀoxygen involved in the BꢀC linkage. However,
C
the most stable conformation of the benzoylated 26 clearly showed steric hindrance at the 3 ꢀOH
C
brought by the aromatic ring of the vicinal benzoyl group, which could be held accountable for its
lower reactivity.
ACS Paragon Plus Environment

Page 21 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Synthesis of tetrasaccharides B(E)CD 36 and 37. Reagents and conditions: a) 34,
TMSOTf (9 mol%), toluene, 70 °C, 60 min, 20%, contaminated; (b) TMSOTf (5 mol%), toluene,
ꢀ15 °C, 10 min, 80%.
With these new data in hand, the “steric hindrance” hypothesis was investigated more deeply
7
9
by coupling acceptor 26 with the less hindered triꢀOꢀacetyl rhamnosyl donor 34 (Scheme 8).
Despite the harsh conditions, a substantial amount of unreacted acceptor was recovered (70%),
80
while most of the donor was hydrolyzed into the known hemiacetal 35. Nevertheless, a new
product was isolated, even though impure, which was tentatively identified as the coupling
+
product 36, based on mass spectrometry analysis (HRMS (ESI ) calculated for C H Cl NO Na
8
4
92
3
23
+
[
M+Na] , 1610.5023, found m/z 1610.5002). It is noteworthy that from this point forward, some
NMR signals from the A or B rhamnoses in the B(E)CD tetrasaccharides and larger
oligosaccharides would either be broadened or absent. The coupling of donor 28 and the less
hindered diol 25, unambiguously furnished the 2 ꢀOꢀrhamnosyl tetrasaccharide 37 (Scheme 8).
C
This outcome is in total agreement with data from Kong and Du, who found that glycosylation at
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the 3ꢀOH of 2,3,4ꢀtriol rhamnoside acceptors is strongly favored, whereas in the case of 2,3ꢀdiol
8
1
rhamnosides, glycosylation occurs preferentially at the axial 2ꢀOH.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Conformation of the local energy minimum of diol 25 highlighting steric hindrance
around the 3 ꢀOH (view a) and accessibility of the 2 ꢀOH (view b). Distances between atoms are
C
C
given in Å. The 2 ꢀOH group is encircled.
C
Molecular modeling of diol 25 using the DFT/B3LYP/6ꢀ31G(3df,3pd) level indicated two
energy minima corresponding to geometries that mainly differ in the orientation of the 3 ꢀOH. In
C
the first minimal energy conformation (Supporting Information Table S4), the hydroxyl group is
involved in a hydrogen bonding interaction with the 2 ꢀoxygen. In the other major conformation
C
(Figure 5, Supporting Information Table S5), which is also more stable by 7 kcal/mol, the 3 ꢀOH
C
interacts with the endocyclic oxygen of residue E as also observed for trisaccharides 26 and 27.
ACS Paragon Plus Environment

Page 23 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
The energy difference between these two conformers is such that only the lower local minimum
is occupied. In the corresponding conformer, the charges on the oxygen atoms at positions 2 and
C
3C are similar, indicating that charge effects do not account for the observed regioselectivity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Rather, the latter is thought to arise from steric hindrance at the equatorial 3 oxygen atom
C
(
C
Figure 5a). Indeed, the 3 ꢀOH takes part in 3 hydrogen bonds. It is in close proximity of Hꢀ1E
and also interacts with the phenyl ring of the 6 ꢀbenzyl ether. In comparison to this highly
E
constrained situation, the 2 ꢀOH group is easily accessible (Figure 5b, black circle).
C
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 9. Synthesis of the target decasaccharide 5. Reagents and conditions: a) 28, TMSOTf (7
mol%), Et O, ꢀ15 °C, 86%; b) 1,5ꢀCyclooctadiene bis(methyldiphenylphosphine) iridium
2
ACS Paragon Plus Environment

Page 25 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
hexafluorophosphate ([Ir]), H , THF, 2.5 h, then I , THF/H O (3:1), 4.5 h, 93%; c) CCl CN, cat
2
2
2
3
DBU, DCE, ꢀ10 °C, 40 min, 97%; d) CF C(NPh)Cl, Cs CO , Acetone, rt, 2 h, 83%; e) 40,
3
2
3
Bromoethanol, TMSOTf (10 mol%), DCE, 0 °C, 25 min, then NaI, NaN , DMF, 80 °C, 2 h, 78%;
3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
f) H NNH , AcOH/Py (2:3), 0 °C, 90%; g) 41, TMSOTf (20 mol%), toluene, ꢀ10 °C, 2 h, 82%;
2
2
h) MeONa, MeOH, 21 h, 71%; i) H , Pd(OH) , tBuOH/CH Cl /H O, 3 d, 58%.
2
2
2
2
2
We then moved on to the preparation of the [AB(E)CD] decasaccharide (Scheme 9). To this
2
end, acceptor 33 was coupled with donor 28 allowing the formation of the pentasaccharide 38
(86%). Deallylation of the latter gave hemiacetal 39 (93%), which was in turn converted to the
key trichloroacetimidate 40 (97%). In order to equip the acceptor for siteꢀselective conjugation,
donor 40 was glycosylated with bromoethanol and the bromine atom was substituted with sodium
azide to reach the fully protected azidoethyl glycoside 43 (78%). A delevulinylation step led to
acceptor 44 (90%), which was then glycosylated with the same key donor 40. Although the
reaction was run at 40 °C, very little condensation was detected. Instead, the donor rapidly
evolved into the trichloroꢀoxazoline 45. Satisfactorily, subsequent addition of TMSOTf to reach
15 mol% allowed the full consumption of the acceptor. Yet, despite several attempts at purifying
the fully protected 46 by column chromatography using different elution systems, the latter could
not be isolated as pure material. Indeed, a MALDI analysis clearly indicated that the
+
decasaccharide was slightly contaminated with trichloroacetamide 42 (HRMS (ESI ) calculated
+
for C H Cl N O Na [M+Na] 2129.62, found m/z 2129.37), a rearranged form of donor 40.
1
11 120
6
2
26
Attempts at purifying the decasaccharide at the next step were disappointing (not shown). For
that reason, trichloroacetimidate 40 was changed for the corresponding Nꢀ
5
0
phenyltrifluoroacetimidoyl (PTFA) donor 41, obtained in 83% yield from hemiacetal 39.
Gratifyingly, the glycosylation between the new donor 41 and acceptor 44 performed at ꢀ10 °C
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
gave a good 82% yield of the desired decasaccharide 46. Methanolysis of the latter into triol 47
(72%) required unoptimized harsh conditions as previously experienced for a closely related
5
1
decasaccharide. Treatment of triol 47 with Pd(OH) /C under a hydrogen atmosphere enabled
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
the concomitant benzyl ether hydrogenolysis, trichloroacetamide hydrodechlorination and azide
reduction to give, in 58% yield following RPꢀHPLC purification, the [AB(E)CD] target 5, which
2
corresponds to a twoꢀunit segment of the S. flexneri 2a OꢀSP equipped with a linker enabling site
selective conjugation.
Figure 6. S. flexneri typeꢀspecific oligosaccharides in reach from CD following regioselective αꢀ
23
D
ꢀglucosylation at the positions indicated by the arrows.
Capitalizing on the excellent outcome of the enzymatic regioselective αꢀDꢀglucosylation of the
CD disaccharide 1 and the successful conversion of the obtained ECD 2 into relevant synthetic
fragments of the OꢀSP from S. flexneri 2a as exemplified by decasaccharide 5, a functional mimic
2
2
of the homologous OꢀSP, we questioned the potential of acceptor 1 as a precursor to αꢀ
glucosylation patterns characteristic of OꢀSPs from other S. flexneri serotypes (Figure 6). In
particular, the αꢀ ꢀglucosylation at Oꢀ4 and Oꢀ6 would open the way to oligosaccharides
Dꢀ
D
D
D
23
representative of OꢀSPs from S. flexneri type I and type IV, respectively. In the following part
of this article, we report on preliminary investigations toward this aim. To probe this hypothesis,
ACS Paragon Plus Environment

Page 27 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
minor modifications of the original acceptor 1 were considered, so as to identify any novel
8
2
regioselectivity by use of GBDꢀCD2 and NpAS, a glucansucrase from GH13 family, in
combination with sucrose as donor. The former naturally forms αꢀ(1→2) glucosidic bonds,
whereas the latter catalyzes the αꢀ(1→4) linkage synthesis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 10. Synthesis of the CD disaccharides 6 and 7, encompassing a masked 4 ꢀOH.
C
Reagents and conditions: a) 2ꢀMethoxypropene, DMP, cat CSA, DMF/Acetone 1:1, 36 h, 75%;
b) NaH, MeI, THF, 3 h, 89%; c) 50% aq TFA, CH Cl , 35 min, 82%; d) (COCl) , DMSO, Et N,
2
2
2
3
CH Cl , ꢀ78 °C→ rt, 75 min, 84% ; e) 80% aq AcOH, 80 °C, 15 h, 69%.
2
2
The most straightforward approach to a modified CD acceptor was to mask the 4 ꢀOH in the
C
starting allyl glycoside 1. To this end, two different targets were envisaged. The 4 ꢀOH was
C
blocked in the form of a methyl ether (6), and alternatively it was oxidized into a 4 ꢀoxo moiety,
C
1
therefore also interfering with the C ring conformation of
Lꢀrhamnose C, to give allyl glycoside
4
7
bearing conformational properties different from that of native 1 (Scheme 10). Toward this aim,
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 76
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
diꢀOꢀisopropylidenation of disaccharide 1 gave the common intermediate 48 (75%), which
8
3
underwent Oꢀmethylation with methyl iodide in THF to give the fully protected 49 (89%). It is
noteworthy that performing the reaction in THF and using sodium hydride as the base was
essential. Indeed, methylation attempts in DMF containing silver oxide or sodium hydride, or in
THF containing silver oxide, resulted in partial Nꢀmethylation of the trichloroacetamide moiety
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
4
(
not shown), as already observed by Scharf and Jütten in their synthesis of evernitrose. This
phenomenon is attributed to the strong inductive effect of the trichloroacetyl group, which
enhances the acidity of the amide proton. The fully protected 49 was then submitted to acidꢀ
mediated acetal hydrolysis, leading to the expected 4 ꢀOMe disaccharide 6 (82%). Alternatively,
C
alcohol 48 was oxidized using Swern conditions into ketone 50 (84%), which was then
deacetalated into the 4 ꢀoxo derivative 7 (69%).
C
We attempted to force glucosylation of disaccharide CD at positions different from 4 ꢀOH. To
C
this end, acceptor reaction with disaccharides 6 and 7, both lacking a free 4 ꢀOH, were performed
C
with GBDꢀCD2. The LCꢀMS followꢀup of the enzymatic conversion showed that no
glucosylation product was formed (not shown) indicating that blocking 4 ꢀOH does not facilitate
C
any alternative glucosylation at another position.
ACS Paragon Plus Environment

Page 29 of 76
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Scheme 11. Synthesis of the methyl thioglycoside 8. Reagents and conditions: a) Ac O, cat
2
DMAP, Py, 14 h, 94%; b) [Ir], H , THF, 4 h, then I , THF/H O (3:1), 16 h, 85%; c) CCl CN,
2
2
2
3
Cs CO , CH Cl , 60 min, then TMSꢀSMe, TMSOTf (5 mol%), 16 h, 50% for 53, 31% for 54; d)
2
3
2
2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MeONa, MeOH, 90 min, 91%.
A novel set of CD acceptor disaccharides, in this case differing from substrate 1 at their
reducing end, was also envisioned to eliminate the possible steric hindrance of the allyl group
with some amino acid residues constituting the active site that could prevent glucosylation by
NpAS (docking results not shown). For that reason, the allyl aglycon in disaccharide 1 was
replaced, while keeping in mind that any group introduced at the anomeric position of the D
residue should provide an access to a CD donor. On the one hand, introduction of a smaller
thiomethyl group gave disaccharide 8, and on the other hand the allyl moiety in the key precursor
1
, was simply cleaved to give the reducing analog 9, the opened form of which could lead to a
complete rearrangement of the substrate inside the enzyme active site. Hence, perꢀacetylation of
disaccharide 1 gave fully protected 51 (94%), and subsequent deallylation and reaction with
methyl trimethylsilyl sulfide gave rise to mainly oxazoline 53 (50%) but also to the expected
methyl thioglycoside 54, albeit in an unoptimized 31% yield (Scheme 11). Finally, methanolysis
gave the deprotected disaccharide 8 (91%).
ACS Paragon Plus Environment