N-O bond as a glycosidic-bond surrogate: Synthetic studies toward polyhydroxylated N-alkoxypiperidines

Article abstract of DOI:10.1002/chem.201202374

A series of novel polyhydroxylated N-alkoxypiperidines has been synthesized by ring-closing double reductive amination (DRA) of highly functionalized 1,5-dialdehydes with various hydroxylamines. The required saccharide-based dialdehydes were prepared efficiently from sodium cyclopentadienylide in seven steps. A two-step protocol has been developed for the DRA; it led, after deprotection, to isofagomine, 3-deoxyisofagomine, and numerous other N-alkoxy analogues. The barrier to inversion in these polyhydroxylated N-alkoxypiperidine derivatives was found by variable-temperature NMR methods to be approximately 15 kcal mol^-1. With the exception of N-hydroxyisofagomine itself, none of the compounds prepared showed significant inhibitory activity against sweet almond β-glucosidase. Copyright

Full text of DOI:10.1002/chem.201202374

DOI: 10.1002/chem.201202374
ꢀ
N O Bond as a Glycosidic-Bond Surrogate: Synthetic Studies Toward
Polyhydroxylated N-Alkoxypiperidines
Gaꢀlle Malik,[a] Angꢁlique Ferry,[a] Xavier Guinchard,[a] Thierry Cresteil,[a] and
David Crich*[a, b]
Abstract: A series of novel polyhy-
droxylated N-alkoxypiperidines has
been synthesized by ring-closing
double reductive amination (DRA) of
highly functionalized 1,5-dialdehydes
with various hydroxylamines. The re-
quired saccharide-based dialdehydes
were prepared efficiently from sodium
cyclopentadienylide in seven steps. A
two-step protocol has been developed
for the DRA; it led, after deprotection,
to isofagomine, 3-deoxyisofagomine,
and numerous other N-alkoxy ana-
logues. The barrier to inversion in
these polyhydroxylated N-alkoxypiperi-
dine derivatives was found by variable-
temperature NMR methods to be ap-
proximately 15 kcalmolꢀ1. With the ex-
ception of N-hydroxyisofagomine itself,
none of the compounds prepared
showed significant inhibitory activity
against sweet almond b-glucosidase.
Keywords: carbohydrates · glycosi-
dases · iminosugars · nitrogen het-
erocycles · reductive amination
Introduction
Knorr,[9] the development of new and improved glycosyla-
tion reactions remains an active field of research.[8b,10] Most
problems in the area are associated with a combination of
inadequate yields for the synthesis of complex oligomeric
species and imperfect anomeric selectivities, which arise
from stereochemical mismatching of donor/acceptor pairs in
some cases;[11] the products frequently require tedious time-
consuming separations with the associated decrease in yield
and difficulties in scale up. Recognizing that glycosylation
reactions are unlikely to be fully perfected in the near
future, we have initiated a program[12] to explore avenues
for the ideally iterative synthesis of novel oligosaccharide
mimetics,[13] with a focus on the development of systems that
mimic the glycosidic bond as closely as possible but that
lack the issue of stereoselectivity that plagues actual glyco-
sylation reactions.
Oligosaccharides display a wide range of biological activi-
ties, some of which may potentially be exploited for medici-
nal purposes.[1] In particular, the glycans present a wide
range of biological activities[2] and have long been consid-
ered as biological-response modifiers,[3] capable of boosting
a biological response, for example, by eliciting natural de-
fenses through the production of reactive oxygen species
and phytoallexins or by the stimulation of macrophages, al-
though they do not display cytotoxicity individually. Notably,
it has been shown that glycans are immunostimulating
agents against infectious diseases[2b,c] and cancer,[4] the activ-
ities of which are mediated by various pattern-recognition
receptors, such as complement receptor 3,[4b] lactosyl ceram-
ide,[5] scavenger receptors, and dectin-1,[6] a protein of the
lectin class. Although several syntheses of high-molecular-
weight glycans have been achieved,[6a,7] the critical reaction
in such syntheses, that is, efficient stereocontrolled glycosi-
dic-bond formation, is far from being a perfected art and
represents a major hurdle to future progress.[8] Indeed, more
than a century after the pioneering work of Koenigs and
From the various possibilities considered, we selected
oligomers based on iminosugars linked through a hydroxyla-
ꢀ
mine N O bond, illustrated for a b-1,3-glucan mimetic in
Figure 1 (X: CH2). In particular, we were attracted to the
hydroxylamine moiety because it is reported that the barrier
to inversion at the nitrogen atom in trialkyl hydroxylamines
is higher than that in simple amines but, at approximately
15 kcalmolꢀ1, is not sufficient to prevent rapid inversion at
room temperature.[14] This is in contrast with the closer
mimics of the glycosidic bond based on dialkoxyamines
(Figure 1, X: O), in which the barrier to inversion is predict-
ed to be approximately 29 kcalmolꢀ1, based on studies with
[a] Dr. G. Malik, A. Ferry, Dr. X. Guinchard, Dr. T. Cresteil,
Prof. Dr. D. Crich
Centre de Recherche de Gif
Institut de Chimie des Substances Naturelles
CNRS, Avenue de la Terrasse, 91198 Gif-sur-Yvette (France)
[b] Prof. Dr. D. Crich
Department of Chemistry
Wayne State University
5101 Cass Avenue, Detroit, MI 48202 (USA)
Fax : (+1)313-577-8822
E-mail: DCrich@chem.wayne.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202374.
Figure 1. Representative hydroxylamine-based oligosaccharide mimetics.
2168
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

FULL PAPER
model compounds,[15] and in which the nitrogen atom there-
fore constitutes an undesirable stereogenic center at room
temperature. With the low barrier to inversion, it is antici-
pated that any oligosaccharide based on the hydroxylamine
motif (Figure 1, X: CH2) would sample the full extent of
conformational space available to it at room temperature
and so would adapt for binding to lectins that are specific
for either axially or equatorially linked oligosaccharides.
Hydroxylamines are considerably less basic than amines,
which is considered to be the reason why an endocyclic hy-
droxylamine analogue of an iminosugar was found to have
only low potency for the inhibition of a glycosidase enzyme,
relative to that of the corresponding amine.[16] As our ulti-
mate goal is the synthesis of oligosaccharide mimetics rather
than glycosidase inhibitors, we viewed this lack of basicity as
an advantage, rather than as a disadvantage. Nevertheless,
there is considerable current interest in 1-N-iminosugars,
which have emerged as potent carbohydrate mimics and in
which the anomeric carbon atom has been replaced by a ni-
trogen atom and the ring oxygen atom has been replaced by
a methylene group, as glycosidase and other enzyme inhibi-
tors. For example, isofagomine, a non-natural azasugar[17]
from this class and an analogue of the natural product fago-
mine isolated from buckwheat seeds,[18] has emerged as one
of the more potent inhibitors of b-glucosidase[19] and hepatic
glycogen phosphorylase.[20]
the employment of an O-substituted hydroxylamine as the
nitrogen source rather than ammonia (Scheme 1). Interest-
ingly though, and heightening our interest in this approach,
Scheme 1. Retrosynthetic analysis based on the ring-closing double re-
ductive amination. P: permanent protection; P’: semipermanent protec-
tion.
the reductive amination of hydroxylamines with dialdehydes
was unknown, with the exception of a single example em-
ploying hydroxylamine itself that had been reported by Du
and Hindsgaul.[26] Ultimately, this approach requires a dia-
ldehyde in which any hydroxyl groups carry permanent pro-
tection (P), ideally in the form of benzyl ethers, and a hy-
droxylamine group with semipermanent protection (P’), per-
haps in the form of a 9-fluorenylmethoxycarbonyl group
(Scheme 1, X: ONHP’). However, for the purpose of the de-
velopmental work that we report here, we have worked
simply with protected polyhydric dialdehydes (Scheme 1, X:
OP).
We have embarked on a program of engineering robust
synthetic methods for the formation of hydroxylamine-
linked iminosugars (Figure 1, X: CH2) with the ultimate
goal of developing oligosaccharide mimetics. Herein, we de-
scribe in full our exploratory chemistry[12a] toward this end,
along with the glycosidase activities of the various simple N-
alkoxyiminosugars prepared in these endeavors. We also
report on the determination of the barrier to inversion of
conformation/configuration in a number of simple N-alkox-
yiminosugars, which previous to our work were only sparsely
mentioned in the literature[21] but which, subsequent to our
preliminary communication,[12a] were also reported by
Fernꢁndez-BolaÇos and co-workers.[22]
Dialdehyde synthesis starting from Cernyꢂs epoxide: Very
few approaches to saccharide-based dialdehyde scaffolds
exist in the literature beyond that first employed by Bols
and co-workers for the synthesis of isofagomine in 1994,[25a]
which started with Cernyꢂs epoxide (1,6:2,3-dianhydro-4-O-
benzyl-b-d-mannopyranose).[27] Although first reported in
the 1970s, synthetic applications of Cernyꢂs epoxide deriva-
tives have been limited, possibly due to the tendency of
these compounds to undergo isomerization under alkaline
conditions and by the lengthy methods needed for their
preparation.[27,28] Bols and co-workers prepared Cernyꢂs ep-
oxide from levoglucosan (1,6-anhydro-b-d-glucopyranose) in
a four-step sequence that suffered from low yields. The
preparation of Cernyꢂs epoxide derivatives was recently sim-
plified by Xue and Guo, who proposed a three-step se-
quence from readily available d-glucal.[29] By adopting this
protocol, we generated iodide 1 through the one-step oxida-
tive 1,6-iodocyclization of d-glucal in 86% yield (Sche-
me 2).[28e,30] Exposure of 1 to NaH and an alkylating agent
directly produced epoxide 2 with the desired regiochemistry,
as described by Arndt and Hsieh-Wilson.[31] We envisaged
the use of an O4 protecting group orthogonal to those in-
tended for the remaining hydroxyl functions in the ultimate
dialdehyde, in order to obtain piperidines that could be se-
lectively deprotected at that position. We selected the naph-
thylmethyl ether for this purpose and obtained 2 in 75%
yield. In accordance with the conditions of Noort and co-
workers for the synthesis of a perbenzylated analogue of
4,[32] a hydroxymethyl group was introduced at the C2 posi-
tion by epoxide opening with vinyl magnesium bromide[33] in
tetrahydrofuran (THF) under reflux conditions, followed by
Results and Discussion
The ideal route to the hydroxylamine-linked imino saccha-
ꢀ
ride mimetics (Figure 1, X: CH2) involves N O bond forma-
tion; however, there are very few examples of the formation
of trialkyl hydroxylamines by displacement of a leaving
group from a nitrogen atom by an alcoholACTHNUTRGNE(NUG ate) in the litera-
ture[23] and even fewer of the formation of such substances
by displacement of a leaving group from an oxygen atom by
an amine or amide. Furthermore, exploratory work along
these lines was not promising and, thus, we turned to more
classic approaches for the synthesis of N-substituted piperi-
dines, of which there are numerous examples in the litera-
ture.[24] We were inclined toward the use of a ring-closing
double reductive amination of a 1,5-dialdehyde, such as that
employed by the groups of Bols and Cardona,[25] but with
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2169

D. Crich et al.
Cernyꢂs epoxide route prompted the design of a second
strategy for dialdehyde synthesis involving oxidative cleav-
age of a suitably functionalized cyclopentene 7 (Scheme 3),
Scheme 3. Proposed 1,5-dialdehyde synthesis from a cyclopentene inter-
mediate.
as first proposed by Mehta and Mohal in 2000.[38] However,
instead of accessing the cyclopentene by fragmentation of a
complex norbornyl derivative,[39] we opted for an approach
based on allylic hydroxylation of the known optically pure
cyclopentene (+)-8, a frequent intermediate in the synthesis
of carbosugars and carbocyclic nucleosides.[40] From among
the literature approaches to (+)-8,[36] we focused on the
modification of Gellman and co-workers[40h] involving the al-
kylation of sodium cyclopentadienylide with benzyloxyme-
thylchloride (BOMCl) in N,N-dimethylformamide (DMF),
rather than the previously preferred THF, because it avoids
isomerization of the sensitive initial adduct. Immediate de-
symmetrization by reaction of the adduct with (ꢀ)-diisopi-
nocampheylborane ((ꢀ)-Ipc2BH),[41] followed by oxidative
workup, afforded the cyclopentene (+)-8 with a satisfactory
60% yield[40k] and 99% enantiomeric purity on a 40 mmol
scale (Scheme 4). After benzylation of (+)-8, the cyclopen-
Scheme 2. Preparation of highly functionalized 1,5-dialdehydes via
Cernyꢂs epoxide. NAP: naphthylmethyl; Bn: benzyl; TFA: trifluoroacetic
acid.
oxidative cleavage of the vinyl group with sodium periodate
in the presence of osmium tetroxide. The corresponding al-
dehyde was directly converted into the hydroxymethyl de-
rivative by reduction with sodium borohydride (35% over 2
steps; Scheme 2), and this was followed by dibenzylation to
give 3 in high yield. Cleavage of the 1,6-anhydro bond was
realized by a two-step procedure with trifluoroacetic acid in
acetic anhydride to afford the diacetate as a mixture of
anomers (a/b=2.6:1) and then a Zemplꢃn deacetylation, to
give 4 with an overall yield of 83% (a/b=1:0.9). Modifica-
tion of the conditions of Bols and co-workers for the subse-
quent lactol cleavage, by increasing the temperature to
1008C and adding a large excess of NaIO4 (10 equiv), gave
the pentodialdose 5 in 85% yield,[34] which consists of a mix-
ture of the dialdehyde and the diastereomeric forms of the
hydrate dihemiacetal. A sample was reduced to diol 6 with
sodium borohydride to check the efficiency of the oxidative
cleavage.[25a] Surprisingly, diol 6, which was obtained quasi-
quantitatively, was a mixture of two epimers at the C3 posi-
tion (3:1 arabino/ribo). The site of epimerization, and its
stereochemical composition, was identified subsequently
after the synthesis of diols (3R)-6 and (3S)-6 by another
strategy (see below).[35] As the two epimers of the corre-
sponding piperidine could be observed after a ring-closing
double reductive amination assay,[36] the loss of stereochemi-
cal integrity was confirmed to occur during the oxidation
step. The use of periodic acid (HIO4), a stronger oxidant, al-
lowed the reaction temperature to be lowered to 458C and
the excess of reagent to be five equivalents, while still giving
the pentodialdose in 71% yield.[37] Nevertheless, this still re-
sulted in epimerization (2:1 arabino/ribo), as confirmed by
reduction of the product to the diol. Various other oxidants,
including lead tetraacetate and iodobenzene diacetate, were
investigated but were not satisfactory.
Scheme 4. Synthesis of optically pure alcohol (+)-8 and attempted allylic
oxidation of 9a. DMP: Dess–Martin periodinane; cap: caprolactamate;
PIDA: phenyliodine diacetate; BQ: benzoquinone.
tene product 9a was submitted to various allylic oxidation
conditions,[42] none of which gave the desired product in
useful quantities.
The problematic allylic oxidation step was circumvented
by the adaptation of recent work from the Meier laborator-
y[40l] with (ꢀ)-8. Thus, (ꢀ)-8 was accessed in the same
manner as its enantiomer, but with (+)-Ipc2BH employed
Dialdehyde synthesis starting from a cyclopentene precur-
sor: The length and the complications inherent in the
2170
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

ꢀ
N O Bonds in N-Oligosaccharide Mimics
FULL PAPER
Table 1. Base-assisted elimination of the mesylate 14b.
Entry
Base
T
Conv.[a]
[%]
Ratio[a]
15b/16b/17b
1
2
proton sponge
phosphazene
PhP4
1008C
1008C
0[b]
100
–
1
0.5
0.16
[c]
3
4
5
6
7
8
tBuOK
tBuOK
1008C
RT
RT
1008C
RT
RT
100
100
100
100
100
100
1
1
1
1
1
1
0.19
0.11
1.9
8.9
0.85
0.13
0.05
0.02
1.75
9.2
Et(Me)2COK[d]
Et(Me)2COK[d]
(Et)3COK[e]
MeONa
0.86
Scheme 5. Preparation of highly functionalized cyclopentanes. Ms: meth-
ane sulfonyl; m-CPBA: meta-chloroperoxybenzoic acid.
[f]
–
[a] Conversions (conv.) and ratios were determined by 1H NMR spectro-
scopy of the crude reaction mixture. [b] No reaction. [c] Used as a 1m
solution in hexanes. [d] Used as a 2m solution in THF. [e] Used as a 2m
solution in (Et)3COH. [f] 17b was not observed under these conditions.
for the hydroboration step. In accordance with Jessel and
Meier,[40l] alcohol (ꢀ)-8 was mesylated to give 10 in excel-
lent yield (Scheme 5), which was subjected to mCPBA-
mediated epoxidation, to afford the epoxide 12 as a separa-
ble mixture of diastereomers 12a and 12b in a 1:3 ratio fa-
voring the undesired isomer 12b. Fortunately, if the mCPBA
epoxidation was conducted directly on cyclopentene (ꢀ)-8,
the reaction occurred preferentially syn to the directing hy-
droxy group and gave the two diastereomeric epoxides 11a
and 11b in a ratio of 3:1 in favor of the desired epoxide.
With Mo(CO)6 as the catalyst for this directed epoxidation
and tert-butyl hydroperoxide as the oxidant,[40g,i] the yield of
desired epoxide 11a could not be increased to 64%. After
mesylation, epoxide 12a was opened regioselectively with
benzyl or naphthylmethyl alcohols in the presence of cata-
lytic BF3·OEt2 to afford alcohols 13a and 13b as single dia-
stereoisomers in 74% and 85% yields, respectively
(Scheme 5). The resulting alcohols were then protected with
benzyl bromide or naphthylmethyl bromide in the presence
of NaH to provide the cyclitols 14a–c in 82%, 92%, and
93% yields, respectively.
(tBuOK, DMF, 1008C) led to a mixture of the three prod-
ucts in a 1:0.19:0.05 ratio (Table 1, entry 3). A decrease in
the temperature to room temperature improved this ratio in
favor of the Hofmann product 15b (Table 1, entry 4). Inter-
estingly, the use of more sterically hindered potassium alk-
oxides reversed the tendancy or led to almost identical
amounts of the three products (Table 1, entries 5–7). Sur-
prisingly enough, the best results were obtained with less
basic, less hindered sodium methoxide at room temperature,
in which case the formation of the over-elimination product
17b could be avoided and the amount of Saytzeff product
16b was limited (Table 1, entry 8). These practical condi-
tions were selected and applied to mesylates 14a–c to afford
cyclopentenes 15a, 15b, and 7, after removal of the minor
products by preparative HPLC, in 72%, 70%, and 63%
yields, respectively (Scheme 6).
For the elimination of the mesylate group in a series of
closely related compounds, Jessel and Meier[40l] identified
tBuOK in an aprotic polar solvent at 1008C as suitable con-
ditions to afford the desired Hofmann product, based on the
low nucleophilicity and steric hindrance of the base. Conse-
quently, we evaluated the action of various alkoxides on me-
sylate 14b (Table 1). Depending on the base, three products
were obtained in various ratios, namely the Hofmann and
Saytzeff products 15b and 16b, respectively, and a product,
17b, resulting from elimination of the benzyloxy group. The
Scheme 6. Synthesis of orthogonally protected cyclopentenes. Yields of
isolated products are given in parentheses.
1
ratios of products were determined by H NMR spectrosco-
The key oxidative cleavage of the double bond was first
evaluated on model cyclopentene 9a (Scheme 7). Classical
conditions with sodium periodate and osmium tetroxide in a
mixture of dioxane and water furnished, after flash chroma-
tography on silica gel, the expected dialdehyde 20a as a
mixture of dialdehyde and hemiacetal diastereoisomers. Re-
duction under the aforementioned conditions furnished diol
21a as a single diastereoisomer in 79% yield (two steps),
which validates the clean formation of the required dialde-
py of the crude reaction mixtures. The use of 1,8-bis(dime-
thylamino)naphthalene (proton sponge), a strong organic
base, led only to recovered starting material with no observ-
able elimination product (Table 1, entry 1). With the phos-
phazene superbase (PhP4), conversion was complete after
40 min, but an unsatisfactory ratio of the various elimination
products was formed (Table 1, entry 2). The use of potassi-
um alkoxides was next investigated. The Meier conditions
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2171

D. Crich et al.
were characterized after sodium borohydride mediated re-
duction. Interestingly, diols 23a, 23b, and 6 were obtained
as 6:1 mixtures of two epimers (see the Supporting Informa-
tion), but, as revealed by the subsequent epimerization-free
ring-closing cyclizations (see below), the loss of stereochem-
ical integrity occurred during the reduction step. As this
phenomenon did not occur with the 4,6-hydroxylated sub-
strates 20a–d, we assume that the stereochemical integrity
of the C3 position must have been compromised (see above;
assignment of 6, Scheme 2).
Table 2. Ring-closing double reductive amination of 20a and 20b with
benzylamine and O-benzylhydroxylamine.
Scheme 7. Preparation of the model dialdehydes 20a–d. Bz: benzoyl;
DIAD: diisopropylazodicarboxylate.
Entry Method[a] R1
R2
RCHO/R2NH2 Product Yield[b] [%]
1
2
3
4
5
6
7
8
Ref. [46] Bn
Ref. [46] NAP Bn
Bn
1.2:1
1.2:1
1.1:1
1.5:1
1.5:1
2:1
24a
24b
25a
25a
25b
25a
25b
25a
25a
25b
25b
71
71
46
54
51
hyde for the reductive double amination protocol. Neverthe-
less, we preferred the use of the more convenient conditions
of Nicolaou et al.,[43] in which osmium tetroxide and N-
methylmorpholine N-oxide (NMO) are used first, followed
by treatment with PIDA, to give diol 21a in 82% yield after
sodium borohydride mediated reduction. Application of
these conditions to the naphthylmethyl-protected cyclopen-
tene 9b, obtained analogously to 9a in 93% yield by alkyla-
tion with bromomethylnaphthalene (Scheme 7), led to diol
21b in a comparable 84% yield over two steps. With the
conditions for the oxidative ring opening established, analo-
gous dialdehyde scaffolds also derived from (+)-8 were syn-
thesized in order to evaluate the double reductive amination
on a galactose-type model (that is, 20d) or on a free hy-
droxy-containing substrate (namely, 20c). Thus, the latter
conditions were applied directly to alcohol (+)-8, and, after
reduction, diol 21c was obtained in a modest 44% yield
(two steps). The synthesis of the pentodialdose 20d began
with Mitsunobu inversion of the hydroxy group in (+)-8 in
72% yield. After protection as a naphthylmethyl ether, cy-
clopentene 19 was cleaved oxidatively and then reduced to
diol 21d in 62% yield.
A
A
A
A
A
A
B
Bn
Bn
OBn
OBn
NAP OBn
Bn OBn
NAP OBn
Bn
Bn
57
2:1
59
OBn
OBn
2.5:1
1:2.5
1:2.5
1:2.5
52[c]
81
9
10
11
B
B
NAP OBn
NAP OBn
85
0[c]
[a] Method A: R2NH2·HCl (1 equiv), NaBH3CN (5 equiv), AcOH
(20 equiv), 3 ꢄ molecular sieves, MeOH (0.04m). Method B: R2NH2·HCl
(2.5 equiv), NaBH3CN (5 equiv), AcOH (10 equiv), 3 ꢄ molecular sieves,
MeOH/THF (0.25m, 10:1). [b] Yields of isolated products. [c] Reaction
performed without molecular sieves.
Ring-closing double reductive amination (DRA): In order
to validate the strategy, we first investigated the DRA on
the disubstituted dialdehydes 20a and 20b with benzylamine
(Table 2, entries 1 and 2). These reactions proceeded
smoothly under classical Borch conditions[44] to afford the
protected piperidines 24a and 24b in 71% yields. We next
turned to the DRA of model dialdehydes 20a and 20b with
hydroxylamines, an unknown process except for a single ex-
ample with hydroxylamine.[26] With freshly prepared dialde-
hyde 20a, O-benzylhydroxylamine (Table 2, entry 3) gave
the corresponding cyclic hydroxylamine in 46% yield. When
the amount of dialdehyde was increased from 1.1 to
1.5 equivalents (Table 2, entries 4 and 5), the hydroxyla-
mines 25a and 25b could be obtained in 54% and 51%
yields, respectively. Increasing the amount of dialdehyde to
2 equivalents slightly increased the yields to approximately
60% (Table 2, entries 6 and 7), but such one-pot conditions
did not enable us to cross the 60% yield barrier. We conse-
quently changed to a two-step process that first employed
2.5 equivalents of hydroxylamine, followed by in situ cyano-
The protected trihydroxydialdehydes 22a, 22b, and 5
were obtained in high yield by NMO/OsO4/PIDA oxidation
of alkenes 15a, 15b, and 7 (Scheme 8), and the products
Scheme 8. Preparation of the dialdehydes 22a, 22b, and 5.
2172
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

ꢀ
N O Bonds in N-Oligosaccharide Mimics
FULL PAPER
borohydride-mediated reduction of the intermediate double
oxime (Table 2, method B). In these reactions, the formation
of the double oxime was monitored by LC-MS analysis, and
the sodium cyanoborohydride was added once the formation
of the double oxime was complete (approximately 2.5 h).
This resulted in the clean formation of the target hydroxyla-
mines 25a and 25b in 81% and 85% yields, respectively
(Table 2, entries 9 and 10). Unlike method A, the presence
of molecular sieves was necessary in method B for the effi-
cient generation of the double oxime (Table 2, entries 8 and
11). Having established conditions for the DRA of hydrox-
ylamines, we evaluated the scope of the reaction by varying
the nature of both the hydroxylamine and the dialdehyde
(Table 3). The use of O-benzylhydroxylamine afforded the
Table 3. Scope of the double reductive amination (DRA).[a]
Entry
Double reductive-amination product
Deprotection
conditions[c]
Deprotection product
R1
R2
R3
R4
Product Yield[b]
[%]
R5 R6
R7
R8
Product Yield[b]
[%]
[d]
1
2
3
4
5
6
7
8
H
H
H
H
OBn
ONAP
OBn
H
H
OBn
H
Bn
Bn
Bn
OBn
24a
24b
24c
25a
71
71
18
81
–
–
H
–
H
H
H
–
–
OH
–
OH
OH
OH
–
–
H
–
H
H
H
–
–
H
–
H
OBn
OBn
–
–
H
–
32
–
32
33
33
–
–
82
–
85
93
67
–
A
[e]
–
OBn
B
C
C
–
H
H
ONAP
H
ONAP
OH
H
H
H
H
OBn
OBn
OBn
OBn
OBn
25b
25c
25d
25e
85
77
84
87
–
–
–
–
–
–
9
OBn
B
C
C
E
C
C
–
C
C
C
C
D
E
E
–
H
H
H
H
H
H
–
H
H
H
H
H
H
H
–
OH OH
34
35
35
36
37
37
–
38
38
39
40a
40b
41a
41b
–
72
90
63
60
66
79
–
88
52
49
96
74
53
44
–
10
11
12
13
14
15
16
17
18
19
20
21
22
23
OH OH OBn
OH OH OBn
OH OBn OBn
OH
OH
–
OH OH OAll
OH OH OAll
OH
OH OH OH
OBn OBn OH
OH OBn OPMB
OBn OH OPMB
H
ONAP OBn
OBn
25 f
85
H
H
ONAP
H
H
H
H
ONAP
OH
H
OBn
ONAP OBn
H
H
H
OBn
OAll
OAll
OAll
OAll
OAll
OPMB
OPMB
26a
26b
26c
26d
26e
27a
27b
80
81
78
89
87
81
90
H
H
–
OAll
OAll
–
ONAP
OBn
H
OBn
H
OH
H
H
H
ONAP OBn
OBn
ONAP
OPMB
27c
27d
28
90
58
75
ONAP OPMB
H
OMe
–
–
–
24
25
26
H
H
H
ONAP
OBn
H
29a
29b
30a
69
76
75
–
–
–
–
–
–
–
OBn
H
C
H
–
OH OH
42
–
46[f]
–
[g]
ONAP
–
–
–
–
–
–
–
[g]
27
28
H
H
OBn
OBn
30b
31a
72
72
–
–
–
–
–
–
–
ONAP
H
–
F, C
–
–
–
–
29
30
H
H
OBn
OBn
31b
31c
70
63
H
–
OH OH
43
61[h]
ONAP OBn
–
–
–
–
–
[a] PMB: para-methoxybenzyl; All: allyl; TFA: trifluoroacetic acid; DDQ: 2,3-dichloro-5,6-dicyano-1,4-benzoquinone. [b] Yields of isolated products.
[c] Method A: H2, Pd(OH)2, MeOH, AcOH; method B: BBr3, CH2Cl2; method C: BCl3, CH2Cl2; method D: TFA, CH2Cl2; method E: DDQ, CH2Cl2/
H2O (20:1); method F: NaOMe, MeOH. [d] 3-Deoxyisofagomine (32a) was obtained from 24b (entry 2). [e] Isofagomine (34) was obtained from 25e
(entry 10). [f] Obtained as a 1.9:1 mixture of ethyl and methyl esters, respectively (see the Supporting Information). [g] The presence of the alkyne is not
compatible with BCl3 deprotection. [h] Obtained as a 1:1 mixture of the two anomers.
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2173

D. Crich et al.
piperidines 25a–f in high yields (77–87%) with various di-
or trisubstituted dialdehydes (Table 3, entries 4–11). In par-
ticular, we showed that the presence of an unprotected alco-
hol did not affect the DRA process (Table 3, entry 7). Curi-
ously, the DRA with the tribenzyloxydialdehyde 22a and
benzylamine resulted in the formation of the protected iso-
fagomine 24c in a significantly lower 18% yield (Table 3,
entry 3), which did not depend on the equivalents of dialde-
hyde employed, than the 87% yield obtained with O-benzyl-
hydroxylamine (Table 3, entry 9). The use of O-allyl hydrox-
ylamine afforded the N-allyloxy piperidines 26a–e in excel-
lent yield with the same set of dialdehydes (Table 3, en-
tries 13–17). Likewise, O-(para-methoxybenzyl)hydroxyla-
mine (Table 3, entries 18–22), and O-methylhydroxylamine
(Table 3, entry 23) also provided the corresponding piperi-
dines in good yields. Finally, we turned our attention to
more functionalized hydroxylamines, such as a glucose-de-
rived hydroxylamine and ester- or alkyne-bearing hydroxyla-
mines.[46] The DRA of ester- and alkyne-bearing hydroxyla-
mines furnished piperidines 29 and 30 in 69–76% yield
(Table 3, entries 24–27). Pseudo-disaccharides 31a–c were
obtained with similarly good yields (Table 3, entries 28–30).
Overall, it is clear that this protocol accommodates the pres-
ence of various functionalities on the hydroxylamine, as well
as various types of saccharide-based dialdehydes. The major-
ity of these derivatives were deprotected to give the target
N-alkoxy piperidines, all of which are analogues of isofago-
Hydrogenolysis of 24b was accomplished by using Pearl-
manꢂs catalyst in a slightly acidic medium (Table 3, entry 2)
and provided 3-deoxyisofagomine (32),[20,49] a known inhibi-
tor of b-glucosidase[49b] and hepatic glycogen phosphory-
lase,[20] albeit a poor one compared to isofagomine itself.
The BBr3 conditions also furnished compound 32[20,49] from
25a (Table 3, entry 4) in a comparable yield to that from the
hydrogenolysis (Table 3, method A, entry 2), whereas isofa-
gomine (34) could be isolated in 72% yield after purifica-
tion if liberated in this manner (Table 3, entry 9).
Synthesis of novel glycoconjugates: Carbohydrates are usu-
ally found in nature in the form of a variety of glycoconju-
gates, so we undertook a brief study of the compatibility of
the N-alkoxy piperidines with click chemistry conditions.
Accordingly, we realized a CuI-catalyzed azide alkyne cyclo-
addition with two functionalized azides, that is, ester- and
glucose-bearing azides under the conditions of Vauzeilles
and co-workers[50] (Scheme 9). Glycoconjugates 44a and 44b
ꢀ
mine characterized by the presence of an N O bond linked
to various functional groups (Table 3). Owing to the relative
instability of the N-alkoxy[45] or N-hydroxy bond[46] under
palladium-catalyzed hydrogenation conditions, we selected
BCl3, known for its compatibility with the hydroxylamine
function,[47] as the reagent of choice for benzyl- and naph-
thylmethyl-group removal (Table 3, method C). Thus, for ex-
ample, pseudo-disaccharide 43 could be obtained in 61%
yield, by removal of the benzoyl groups by Zemplꢃn deacy-
lation followed by treatment with BCl3 (Table 3, method F,
entry 29). This isomaltose mimic differs simply by the pres-
ence of one oxygen atom linked to the nitrogen atom from
a pseudo-disaccharide synthesized by Bols and co-workers
in 1996, which was reported to be a potent glycosidase in-
hibitor.[48] The simple N-hydroxyisofagomine, 40a, which is
closely related to the N-hydroxy-1-deoxymannojirimicin and
N-hydroxy polyoxygenated pyrrolidines previously described
by Py and co-workers,[47a,b] could be obtained from the
PMB-protected piperidine 27b by treatment with BCl3
under the aforementioned conditions (Table 3, entry 19). As
an alternative to BCl3, the PMB group can be orthogonally
removed by the action of TFA in CH2Cl2 to give benzylated
compound 40b in 74% yield (Table 3, method D, entry 20).
The naphthylmethyl groups can be removed in the presence
of benzyl ethers and N-OPMB groups by treatment with
DDQ, as shown for compounds 27c and 27d (Table 3, meth-
od E, entries 21 and 22). Finally, simple piperidines were ob-
tained in high yields by the action of BBr3 (Table 3, meth-
Scheme 9. Novel glycoconjugates obtained by click chemistry.
were obtained in excellent yields and were successfully de-
protected by the standard BCl3 conditions to afford 45a and
45b in 81% and 98% yields, respectively.
ꢀ
We also evaluated the effect of the N OR functionality of
41a and 41b on the glycosylation reaction and screened var-
ious glycosyl donors, in the gluco and manno series. Un-
fortunately, standard trichloroacetimidate conditions and
more conventional methods such as the Koenigs Knorr gly-
cosylation proved unsuitable for application to 25c (see the
Supporting Information). However, thioglycosides and, in
particular, methods relying on the preactivation of the
donor before addition of the acceptor were found to be
compatible with the hydroxylamine functionality. Thus, b-
mannosyl donor 46 was activated with 2-(phenylsulfinyl)pi-
peridine (BSP) and triflic anhydride at ꢀ788C in dichloro-
methane before addition of the acceptors 41a and 41b,
which led to isolation of the pseudo-disaccharides 47a and
47b. Interestingly given the known low reactivity of the 4-
OH group in pyranosyl acceptors toward glycosylation, com-
pound 47a was obtained in an acceptable 72% yield, where-
ꢀ
od B), which cleaved the hydroxylamine N O bond in addi-
tion to removing benzyl ethers, in sharp contrast to BCl3.
2174
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

ꢀ
N O Bonds in N-Oligosaccharide Mimics
FULL PAPER
Scheme 10. Diastereoselective glycosylations at positions 3 or 4 of N-
alkoxy piperidines. PMB: para-methoxybenzyl; Tol: tolyl; Tf: trifluoro-
methanesulfonyl; TTBP: tri-tert-butylpyrimidine; BSP: 1-benzenesulfinyl
piperidine.
as the glycosyl acceptor containing a hydroxy function at the
3 position furnished compound 47b in only 24% yield
(Scheme 10). Encouraged by the positive result in the glyco-
sylation of the 4-OH acceptor, the N-benzyloxypiperidine 36
was also subjected to glycosylation under the same condi-
tions and provided glycoside 48 in 55% yield.
1
Figure 2. Selected region of the H NMR spectra ([D7]DMF, 600 MHz) of
26a recorded at different temperatures ranging from ꢀ108C to 958C (d=
4.12–4.22 ppm; allylic protons Ha).
Dynamic NMR study of nitrogen inversion in N-alkoxy pi-
peridines: As discussed in the introduction, we anticipated
the N-alkoxy piperidines prepared in the course of this
study to have a barrier for inversion at the nitrogen atom of
approximately 15 kcalmolꢀ1. In the event, the 1H NMR
spectra of the N-alkoxy piperidines synthesized present
structureless and broadened resonances at room tempera-
ture, which suggests a slow conformational interconversion
on the NMR timescale. When the samples were heated to
around 1008C in DMF,[51] clean sharp spectra representing
an average conformation could be obtained, and all such
compounds were characterized under these conditions (see
the Supporting Information). When the temperature at
which the spectra were recorded was reduced to 263 K, rela-
tively well-resolved 1H and 13C NMR spectra showing dis-
tinct signals for two conformational/configurational isomers
could be obtained, in a 1:2.9 ratio for 26a, most likely in
favor of the equatorial isomer. The free energy of activation
the signals of the two exchanging species at low temperature
(Dn in Hz) and by application of Equation (1).[53]
¼
DGC ¼ 4:575 ꢁ 10ꢀ3TC½9:972 þ logðTC=DnÞꢂ
ð1Þ
¼
As the estimated barrier DGc for the inversion process
is independent of the temperature, we also calculated the
rate constant and deduced the half-life time of the process
at 293 K in the standard manner for first-order processes,
which led to the values given in Table 4. The barrier to in-
version obtained in this manner for 26a was 15.3 kcalmolꢀ1,
which is in agreement with literature values for endocyclic
six-membered-ring hydroxylamines[14a] and which corre-
sponds to a half-life for inversion of 28.4 ms at 293 K. The
Table 4. Barriers to exchange and half-life times for various N-allyloxy
piperidines determined by VT-NMR spectroscopy.
¼
(DG ) for the dynamic process was determined for com-
pound 26a by variable-temperature (VT) NMR spectrosco-
1
py. To this end, the H and 13C NMR spectra of 26a were re-
corded over a temperature range of ꢀ108C to 958C. In the
1H NMR spectrum, the most pronounced effect[52] was ob-
293K
Entry Compound 1H/13C signal Tc
Dn
[K] [Hz]
DGc
t1/2
ꢀ
served for the allylic CH2 group linked to the N O bond.
observed
[kcalmolꢀ1
]
[ms]
These two diastereotopic protons form a doublet at 958C
and become two doublets at low temperature (ꢀ108C;
Figure 2). The coalescence point for the two sets of signals
is observed at approximately 408C. The free energy of acti-
1
2
3
4
5
6
7
26a
26a
26a
26a
Ha
C1
C5
C4
303
25.5 15.3
333 404.0 15.1
343 505.7 15.4
343 592.9 15.3
15.28
293
288
¼
vation, DGc (kcalmolꢀ1), for the observed dynamic process
26a (average values)
26d
38
28.4
8.8
7.4
Ha
Ha
35.0 14.6
25.1 14.5
at the coalescence temperature (Tc in K) was then estimated
by measuring the maximum frequency difference between
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2175

D. Crich et al.
13C NMR spectrum of 26a also shows temperature depend-
ence but is relatively well resolved into two sets of signals at
room temperature. The coalescence temperature was deter-
mined to be approximately 708C (see the Supporting Infor-
mation). For reasons of resolution, the barrier to inversion
was calculated based on the frequency difference of the
methylene signals in the a position to the inverting nitrogen
site (that is, the C1 and C5 atoms; see Table 4) and was
found to be 15.1 and 15.4 kcalmolꢀ1, respectively, values
the N O bond can also be precursor to an amino group, de-
pending on the mode of deprotection, as illustrated by a de
ꢀ
novo synthesis of isofagomine. A dynamic H and 13C NMR
1
study on representative compounds revealed an inversion
barrier of approximately 14.5 kcalmolꢀ1 for trisubstituted N-
alkoxy piperidines, which indicates a half-life for inversion
of 8 ms at room temperature. Further work directed at the
synthesis of hydroxylamine oligosaccharide mimetics and
their biological applications is currently underway and will
be reported in due course.
1
that are consistent with that obtained by H NMR spectro-
scopy. The same series of VT-NMR experiments were realiz-
ed on the tri(benzyloxy) compound 26d and on the triol 38,
with both compounds showing a slightly lower inversion bar-
rier of 14.5 kcalmolꢀ1 and correspondingly shorter half-lives
for inversion. At 263 K, the tribenzyl-protected compound
26d presumably exists as a 1:2.5 mixture of isomers, whereas
the trihydroxy analogue 38 exhibits a 1:5.2 ratio of isomers
at the same temperature; we presume that the major isomer
in each case is the chair conformer with the equatorial ally-
loxy substituent. The low-millisecond half-lives for inversion
at room temperature determined for compounds 26a, 26d,
and 38 are in agreement with literature values for the con-
formation inversion of piperidines[54] and are much higher
than standard values for the inversion of pyramidal nitrogen
atoms in amines, which is consistent with the well-known in-
crease in nitrogen-inversion barriers upon introduction of
electronegative substituents.[17a] Although VT-NMR studies
have not been previously conducted on N-alkoxy piperidines
to the best of our knowledge, studies have been conducted
on N-alkoxy pyrrolidines,[55] on acyclic hydroxylami-
nes,[14b,15b] and on endocyclic six-membered-ring hydroxyla-
mines,[14a] and similar values were obtained. Overall, the
measured inversion barriers for 26a, 26d, and 38, which we
take as representative of the series as a whole, are consistent
with our initial hypothesis that oligosaccharide mimetics
based on the hydroxylamine linkage will adapt for binding
to lectins that have evolved to accommodate either axial or
equatorial glycosides.
Experimental Section
1
Detailed descriptions of experimental procedures, H and 13C NMR spec-
tra of all new compounds, and preliminary biological tests for glycosidase
inhibition are given in the Supporting Information.
General procedure for the oxidative cleavage of cyclopentene deriva-
tives: In accordance with the procedure described in the literature,[43] 4-
methylmorpholine N-oxide (1.5 equiv) and osmium tetroxide (0.02m in
water, 0.02 equiv) were added to a solution of alkene (1 equiv) in a 10:1
mixture of acetone/H2O (0.1m), and the mixture was stirred at room tem-
perature. After the starting material had been consumed (approximately
2.5 h), as monitored by TLC (4:1 CH2Cl2/methyl tert-butyl ether), PhI-
AHCTUNRTGEG(NUNN OAc)2 (1.5 equiv) was added to the reaction mixture. After being stirred
for 1–3 h, the reaction was quenched with a saturated aqueous solution
of Na2S2O3 (15 mLmmolꢀ1). The mixture was extracted with EtOAc (3ꢅ
10 mLmmolꢀ1), and the extract was washed with a saturated aqueous sol-
ution of CuSO4 (2ꢅ15 mLmmolꢀ1) and brine (2ꢅ15 mLmmolꢀ1), dried
over Na2SO4, and concentrated in vacuo. The crude residue was purified
by flash column chromatography (9:1 then 1:1 heptane/EtOAc, unless
otherwise stated) to give the expected 1,5-dialdehyde. Representative ex-
ample: 3,4,6-tri-O-benzyl-2-deoxy-d-arabinohexodialdose monohydrate
22a was prepared according to the general procedure by using 15a
(300 mg, 0.75 mmol), 4-methylmorpholine N-oxide (132 mg, 1.12 mmol),
osmium tetroxide (0.75 mL, 0.015 mmol), and PhIACTHNUGTRENUNG(OAc)2 (362 mg,
1.12 mmol) in a 10:1 mixture of acetone/H2O (7.7 mL). The reaction mix-
ture was worked up and purified as described to give dialdehyde 22a
(298 mg, 0.66 mmol, 88%) as a pale yellow oil that was directly engaged
in the double reductive amination.
General procedure for the double reductive amination with O-alkyl hy-
droxylamines: Powdered and activated 3 ꢄ molecular sieves (1 gmmolꢀ1
of dialdehyde) were added to a stirred solution of dialdehyde (1 equiv) in
anhydrous THF (2.5m) and MeOH (0.25m). O-Alkyl hydroxylamine hy-
drochloride (2.5 equiv) was added, and the mixture was stirred at room
temperature for 2.5 h. The complete formation of dioxime was confirmed
by LC-MS analysis. AcOH (10 equiv) and NaBH3CN (5 equiv) were then
added, and the mixture was stirred at room temperature. The reduction
was followed by LC-MS analysis, and if necessary (that is, if the reduction
was not complete after 3 h), a second portion of AcOH (10 equiv) and
NaBH3CN (5 equiv) were added and the mixture was stirred for the indi-
cated time. The reaction was quenched by addition of an aqueous 1m sol-
ution of NaOH (20 mLmmolꢀ1 of dialdehyde) and brine (20 mLmmolꢀ1
of dialdehyde). After extraction with EtOAc (3ꢅ30 mLmmolꢀ1 of dialde-
hyde), the combined organic layers were washed with brine (2ꢅ
30 mLmmolꢀ1 of dialdehyde), dried over Na2SO4, and concentrated in
vacuo. The crude residue was purified by flash column chromatography
with the indicated eluent to yield the expected N-alkoxypiperidine deriv-
ative. Representative example 1: (3R,4R,5R)-1,3,4-tris(benzyloxy)-5-((ben-
zyloxy)methyl)piperidine (25e) was prepared according to the general
procedure by using dialdehyde 22a (155 mg, 0.344 mmol) and O-benzyl-
hydroxylamine hydrochloride (137 mg, 0.86 mmol) in MeOH (1.5 mL)
and THF (0.15 mL). After addition of AcOH (400 mL, 6.88 mmol, in two
portions) and NaBH3CN (216 mg, 3.44 mmol, in two portions), the mix-
ture was stirred for 20 h and worked up as described. The crude product
was purified by column chromatography (95:5 heptane/EtOAc) to yield
Conclusion
A practical seven-step asymmetric synthesis of densely func-
tionalized dialdehydes was developed by starting from
sodium cyclopentadienylide. This synthetic route compares
favorably with other known procedures and presents a high
degree of flexibility through the ability to introduce orthog-
onal protecting groups or to invert the stereochemistry of a
given alcohol by a Mitsunobu sequence. Protocols for the
double reductive amination of O-alkyl hydroxylamines with
these dialdehydes were established and enabled rapid access
to numerous polyhydroxylated N-alkoxy piperidines. The
flexibility of this protocol was demonstrated by the use of
various functionalized O-alkyl hydroxylamines, including a
sugar-derived hydroxylamine. Deprotection of the com-
pounds by sequential or one-pot methods furnished various
di- and trihydroxylated N-alkoxy piperidines. Importantly,
2176
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

ꢀ
N O Bonds in N-Oligosaccharide Mimics
FULL PAPER
25e (156 mg, 0.298 mmol, 87%) as a colorless oil: Rf =0.46 (4:1 heptane/
EtOAc); [a]2D2 = +20.4 (c=0.96 in CHCl3); 1H NMR (500 MHz,
[D7]DMF, 373 K): d=2.10–2.14 (m, 1H; H5), 2.54 (t, J=10.6 Hz, 1H;
H2), 2.66 (t, J=11.2 Hz, 1H; H7), 3.38–3.41 (m, 1H; H7), 3.46 (t, J=
9.0 Hz, 1H; H4), 3.58–3.61 (m, 1H; H2), 3.65–3.70 (m, 2H; H6), 3.76–
Methyl 6-[(3R,4R,5R)-3,4-dihydroxy-5-(hydroxymethyl)-piperidinyl]-a,b-
d-glucopyranoside (43): A catalytic amount of NaOMe (25% solution in
MeOH, 1 drop) was added to
a stirred solution of 31b (80 mg,
0.087 mmol) in dry MeOH (2 mL). TLC analysis indicated complete con-
version of the starting material after 4 h. The solution was then concen-
trated in vacuo, and the crude material was purified by column chroma-
tography (EtOAc) to give the intermediate triol (47 mg, 0.077 mmol,
ꢀ
3.80 (m, 1H; H3), 4.50 and 4.51 (AB, JAB =12.3 Hz, 2H; CH2 Bn), 4.64
ꢀ
and 4.89 (AB, JAB =11.5 Hz, 2H; CH2 Bn), 4.68 and 4.70 (AB, JAB
=
89%) as
a
white foam: Rf =0.23 (EtOAc); [a]2D4 = +59.7 (c=1.5,
ꢀ
ꢀ
ꢀ
11.8 Hz, 2H; CH2 Bn), 4.73 (s, 2H; NO CH2 Bn), 7.27–7.38 ppm (m,
20H; H Ar); C NMR (125 MHz, [D7]DMF, 373 K): d=140.8 (Cq Bn),
MeOH); 1H NMR (500 MHz, [D7]DMF, 363 K): d=2.08–2.15 (m, 1H;
H5), 2.50 (t, J=10.7 Hz, 1H; H2), 2.63 (t, J=11.3 Hz, 1H; H7), 3.23 (t,
J=9.3 Hz, 1H; H4’), 3.37–3.40 (m, 1H; H2’), 3.39 (s, 3H; Me), 3.43–3.47
(m, 2H; H4, H7), 3.63 (t, J=9.1 Hz, 1H; H3’), 3.66–3.72 (m, 4H; H2,
H5’, H6), 3.76–3.79 (m, 1H; H3), 3.81 (dd, J=6.6, 11.3 Hz; 1H, H6’),
3.97 (brs, 1H; OH), 4.08 (dd, J=2.2, 11.3 Hz, 1H; H6’), 4.33 (brs, 1H;
13
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
140.5 (Cq Bn), 140.3 (Cq Bn), 139.9 (Cq Bn), 129.5 (CH Ar), 129.3
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(CH Ar), 129.3 (CH Ar), 129.2 (CH Ar), 128.8 (CH Ar), 128.7 (CH
ꢀ
ꢀ
ꢀ
ꢀ
Ar), 128.6 (CH Ar), 128.5 (CH Ar), 128.5 (CH Ar), 128.3 (CH Ar),
81.5 (C4), 80.1 (C3), 75.1 (NO CH2 Bn), 75.0 (CH2 Bn), 74.1 (CH2
Bn), 73.0 (CH2 Bn), 70.6 (C6), 59.0 (C2), 58.1 (C7), 40.9 ppm (C5); IR
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
OH), 4.50 (brs, 1H; OH), 4.51 and 4.52 (AB, JAB =12.1 Hz, 2H; CH2
(neat): nmax =1454, 1364, 1095, 1077, 1046, 1027, 733, 695 cmꢀ1; HRMS
(ESI-TOF): calcd for C34H38NO4 [M+H]+: 524.2801; found: 524.2789.
Representative example 2: Methyl 2,3,4-tri-O-benzoyl-6-[(3R,4R,5R)-3,4-
bis(benzyloxy)-5-((benzyloxy)methyl)-piperidinyl]-a-d-glucopyranoside
31b was prepared according to the general procedure by using dialde-
hyde 22a (113 mg, 0.25 mmol) and methyl 2,3,4-tri-O-benzoyl-6-O-
amino-a-d-glucopyranoside hydrochloride[12a] (349 mg, 0.63 mmol) in
MeOH (1 mL) and THF (0.1 mL). After addition of AcOH (290 mL,
5 mmol, in two portions) and NaBH3CN (157 mg, 2.5 mmol, in two por-
tions), the mixture was stirred for 18 h and worked up as described. The
crude product was purified by column chromatography (9:1!7:3 hep-
tane/EtOAc) to yield 31b (162 mg, 0.176 mmol, 70%) as a colorless oil:
Rf =0.63 (3:2 heptane/EtOAc); [a]2D2 = +46.4 (c=0.81, CHCl3); 1H NMR
(500 MHz, [D7]DMF, 373 K): d=2.03–2.10 (m, 1H; H5), 2.49 (t, J=
10.4 Hz, 1H; H2), 2.61 (t, J=11.3 Hz, 1H; H7), 3.37–3.40 (m, 1H; H7),
3.43 (t, J=9.1 Hz, 1H; H4), 3.57 (s, 3H; OMe), 3.60–3.64 (m, 3H; H2,
ꢀ
Bn), 4.63 and 4.89 (AB, JAB =11.3 Hz, 2H; CH2 Bn), 4.64 (d, J=3.5 Hz,
ꢀ
1H; H1’), 4.71 and 4.74 (AB, JAB =11.8 Hz, 2H; CH2 Bn), 7.26–7.42 ppm
13
ꢀ
ꢀ
(m, 15H; H Ar); C NMR (125 MHz, [D7]DMF, 363 K): d= 140.7 (Cq
Bn), 140.5 (Cq Bn), 140.3 (Cq Bn), 129.4 (CH Ar), 129.3 (CH Ar),
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
129.2 (CH Ar), 128.8 (CH Ar), 128.7 (CH Ar), 128.7 (CH Ar), 128.5
(CH Ar), 128.3 (CH Ar), 101.7 (C1’), 81.6 (C4), 80.2 (C3), 75.6 (C3’),
75.0 (CH2 Bn), 74.1 (C2’), 74.1 (CH2 Bn), 73.6 (C6’), 73.0 (CH2 Bn),
73.0 (C4’), 72.1 (C5’), 70.6 (C6), 58.9 (C2), 58.1 (C7), 55.9 (Me), 40.8 ppm
(C5); IR (neat): nmax =3392, 2925, 2856, 1454, 1364, 1102, 1073, 1049,
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1030, 749, 736, 698 cmꢀ1
; HRMS (ESI-TOF): calcd for C34H44NO9
[M+H]+: 610.3016; found: 610.3022. BCl3 (1m solution in hexanes,
450 mL, 0.45 mmol, 9 equiv) was added to a solution of the triol (30 mg,
0.049 mmol) in CH2Cl2 (5 mL), and the mixture was stirred at room tem-
perature for 5 h. After addition of MeOH (5 mL), the solution was stir-
red for 30 min and concentrated in vacuo. The crude residue was purified
by flash column chromatography (90:5:5!80:15:5 iPrOH/H2O/NH4OH)
to give 43 as a colorless oil that contained a 1:1 mixture of a and b anom-
ers (11.5 mg, 0.034 mmol, 69%), which were separated by preparative
HPLC (SunFire prep C18 OBD column, Waters, 150ꢅ19 mm, isocratic
99:1:0.1 H2O/MeOH/HCOOH, 17 mLminꢀ1, ELS detection) and charac-
terized separately: Rf =0.51 (70:20:10 iPrOH/H2O/NH4OH); a-43: tR =
9.6 min (SunFire C18 5 mm column, Waters, 150ꢅ4.6 mm, isocratic
H6), 3.75 (dt, J=4.1, 8.8 Hz, 1H; H3), 3.98 and 4.01 (ABX, JAB
=
11.9 Hz, AX =3.2 Hz, BX =5.0 Hz, 2H; H6’), 4.38 (ddd, J=3.4, 4.6,
J
J
ꢀ
10.1 Hz, 1H; H5’), 4.48 and 4.49 (AB, JAB =12.3 Hz, 2H; CH2 Bn), 4.62
ꢀ
and 4.86 (AB, JAB =11.5 Hz, 2H; CH2 Bn), 4.64 and 4.67 (AB, JAB
=
ꢀ
11.8 Hz, 2H; CH2 Bn), 5.30 (d, J=3.8 Hz, 1H; H1’), 5.40 (dd, J=3.8,
10.1 Hz, 1H; H2’), 5.60 (t, J=9.8 Hz, 1H; H4’), 6.09 (t, J=9.8 Hz, 1H;
ꢀ
ꢀ
99:1:0.1 H2O/MeOH/HCOOH, 1 mLminꢀ1
,
ELS detection); [a]2D4 = +
H3’), 7.28–7.64 (m, 24H; H Ar), 7.86–7.98 ppm (m, 6H; H Ar);
1
13C NMR (125 MHz, [D7]DMF, 373 K): d=166.8 (C=O), 166.6 (C=O),
62.2 (c=0.32, H2O); H NMR (500 MHz, [D7]DMF, 363 K): d=1.77–1.84
(m, 1H; H5), 2.32 (t, J=10.6 Hz, 1H; H2), 2.37 (t, J=11.3 Hz, 1H; H7),
3.19 (t, J=9.1 Hz, 1H; H4), 3.21 (t, J=9.1 Hz, 1H; H5’), 3.38 (dd, J=
3.8, 9.8 Hz, 1H; H2’), 3.40 (s, 3H, Me), 3.42–3.46 (m, 1H; H7), 3.47–3.50
(m, 1H; H2), 3.55–3.64 (m, 3H; H3, H6, H3’), 3.67–3.71 (m, 1H; H4’),
3.76–3.80 (m, 2H; H6, H6’), 4.06 (dd, J=1.9, 11.3 Hz, 1H; H6’),
4.63 ppm (d, J=3.5 Hz, 1H; H1’); 13C NMR (125 MHz, [D7]DMF,
363 K): d=101.7 (C1’), 76.5 (C4), 75.6 (C3’), 74.1 (C2’), 73.4 (C6’), 73.0
(C5’), 72.3 (C3), 72.1 (C4’), 63.1 (C6), 61.7 (C2), 58.4 (C7), 55.8 (Me),
43.2 ppm (C5); IR (neat): nmax =3368, 2922, 1044 cmꢀ1; HRMS (ESI-
TOF): calcd for C13H25NNaO9 [M+Na]+: 362.1427; found: 362.1430; b-
43: tR =8.5 min (SunFire C18 5 mm column, Waters, 150ꢅ4.6 mm, isocrat-
ic 99:1:0.1 H2O/MeOH/HCOOH, 1 mLminꢀ1, ELS detection); [a]D24 = +
6.6 (c=0.32, H2O); 1H NMR (500 MHz, [D7]DMF, 363 K): d=1.78–1.85
(m, 1H; H5), 2.34 (t, J=10.7 Hz, 1H; H2), 2.37 (t, J=11.3 Hz, 1H; H7),
3.16–3.24 (m, 3H; H2’, H4, H5’), 3.38 (t, J=8.8 Hz, 1H; H3’), 3.41–3.50
(m, 3H; H2, H4’, H7), 3.48 (s, 3H; Me), 3.56–3.61 (m, 2H; H3, H6), 3.76
(dd, J=6.9, 11.5 Hz, 1H; H6’), 3.79 (dd, J=4.4, 10.7 Hz, 1H; H6), 4.10
(dd, J=1.9, 11.3 Hz, 1H; H6’), 4.18 ppm (d, J=7.9 Hz, 1H; H1’);
13C NMR (125 MHz, [D7]DMF, 363 K): d=105.6 (C1’), 79.0 (C3’), 76.5,
76.1, 75.4 (C4, C4’, C5’), 73.5 (C6’), 72.8 (C2’), 72.3 (C3), 63.1 (C6), 61.8
(C2), 58.5 (C7), 56.8 (Me), 43.2 ppm (C5).
ꢀ
ꢀ
ꢀ
ꢀ
166.3 (C=O), 140.6 (Cq Bn), 140.3 (Cq Bn), 140.2 (Cq Bn), 134.5 (CH
ꢀ
ꢀ
ꢀ
ꢀ
Ar), 134.5 (CH Ar), 134.3 (CH Ar), 130.9 (2Cq Bz), 130.5 (Cq Bz),
ꢀ
ꢀ
ꢀ
ꢀ
130.3 (CH Ar), 129.7 (CH Ar), 129.7 (CH Ar), 129.5 (CH Ar), 129.3
(CH Ar), 129.2 (CH Ar), 129.1 (CH Ar), 128.7 (CH Ar), 128.7 (CH
Ar), 128.6 (CH Ar), 128.4 (CH Ar), 128.4 (CH Ar), 98.5 (C1’), 81.3
(C4), 80.0 (C3), 74.8 (CH2 Bn), 74.0 (CH2 Bn), 73.3 (C2’), 72.9 (CH2
Bn), 72.7 (C3’), 72.3 (C6’), 71.8 (C4’), 70.4 (C6), 69.4 (C5’), 58.7 (C2),
57.9 (C7), 56.4 (OMe), 40.7 ppm (C5); IR (neat): nmax =2922, 2856, 1725,
1452, 1276, 1259, 1092, 1068, 1050, 1026, 706, 697 cmꢀ1; HRMS (ESI-
TOF): calcd for C55H56NO12 [M+H]+: 922.3803; found: 922.3772.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Selected examples of deprotections:
(3R,4R,5R)-1-(Hydroxy)-5-(hydroxymethyl)piperidine-3,4-diol
hydro-
chloride (40a): BCl3 (1m solution in hexanes, 870 mL, 0.867 mmol,
12 equiv) was added to a solution of 27b (40 mg, 0.072 mmol) in CH2Cl2
(7 mL), and the mixture was stirred at room temperature for 4 h. After
addition of MeOH (5 mL), the solution was stirred for 30 min and con-
centrated in vacuo. The crude residue was purified by flash column chro-
matography (95:4:1 iPrOH/H2O/NH4OH). The hydroxylamine was then
protonated by addition of a 3m HCl solution in MeOH (1 mL) and con-
centrated in vacuo to give 40a (13.8 mg, 0.069 mmol, 96%) as a white
amorphous solid (1:1 mixture of anomers): Rf =0.35 (95:4:1 iPrOH/H2O/
NH4OH); [a]2D4 = +8.0 (c=0.56, EtOH, free base); 1H NMR (300 MHz,
D2O, 300 K): d=1.89–2.01 (m, 1H; H5b), 2.32–2.44 (m, 1H; H5a), 3.20–
3.88 (m, 15H; H2, H3b, H4, H6, H7), 4.14 ppm (ddd, J=4.8, 9.1, 10.5 Hz,
1H; H3a); 13C NMR (75 MHz, D2O, 300 K): d=70.6, 70.4 (C4), 68.2
(C3b), 66.0 (C3a), 59.0, 58.7, 58.6, 57.7 (C2, C6), 56.4, 55.3 (C7), 39.8
(C5b), 37.4 ppm (C5a); IR (neat): nmax =3231, 3031, 1433, 1173, 1094,
Selected example of glycoconjugation and deprotection:
(2R,3R,4S,5S,6R)-2-(4-(3-(((3R,4R,5R)-3,4-dihydroxy-5-(hydroxymethyl)-
piperidin-1-yl)oxy)propyl)-1H-1,2,3-triazol-1-yl)-6-(hydroxymethyl)tetra-
hydro-2H-pyran-3,4,5-triol (45b): In
a schlenk flask, 30b (75 mg,
0.150 mmol, 1 equiv) and b-d-glucopyranosyl azide (31 mg, 0.150 mmol,
1 equiv) were dissolved in a 1:1 mixture of tBuOH and H2O (1.5 mL). A
375 mm aqueous solution of phenylene diamine (61 mL, 23 mmol,
0.15 equiv), a 250 mm aqueous solution of sodium ascorbate (61 mL,
15 mmol, 0.10 equiv), and a 125 mm aqueous solution of CuSO4 (61 mL,
1044, 1016, 985 cmꢀ1
;
HRMS (ESI-TOF): calcd for C6H14NO4
[MꢀHCl+H]+: 164.0923; found: 164.0916.
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2177

D. Crich et al.
7.5 mmol, 0.05 equiv) were added. The reaction mixture was then purged
with Ar, stirred for 4 h at room temperature, and concentrated in vacuo.
Purification by column chromatography (1:0!10:1 EtOAc/MeOH) gave
[5] J. W. Zimmerman, J. Lindermuth, P. A. Fish, G. P. Palace, T. T. Ste-
venson, D. E. DeMong, J. Biol. Chem. 1998, 273, 22014–22020.
[6] a) E. L. Adams, P. J. Rice, B. Graves, H. E. Ensley, H. Yu, G. D.
Brown, S. Gordon, M. A. Monteiro, E. Papp-Szabo, D. W. Lowman,
T. D. Power, M. F. Wempe, D. L. Williams, J. Pharmacol. Exp. Ther.
2008, 325, 115–123; b) G. D. Brown, S. Gordon, Nature 2001, 413,
36–37; c) A. S. Palma, T. Feizi, Y. Zhang, M. S. Stoll, A. M. Lawson,
E. Dꢇaz-Rodrꢇguez, M. A. Campanero-Rhodes, J. Costa, S. Gordon,
G. D. Brown, W. Chai, J. Biol. Chem. 2005, 281, 5771–5779.
[7] a) M. Joe, Y. Bai, R. C. Nacario, T. L. Lowary, J. Am. Chem. Soc.
2007, 129, 9885–9901; b) K.-F. Mo, H. Li, J. T. Mague, H. E. Ensley,
Carbohydr. Res. 2009, 344, 439–447; c) K.-C. Chu, C.-T. Ren, C.-P.
Lu, C.-H. Hsu, T.-H. Sun, J.-L. Han, B. Pal, T.-A. Chao, Y.-F. Lin,
S.-H. Wu, C.-H. Wong, C.-Y. Wu, Angew. Chem. 2011, 123, 9563–
9567; Angew. Chem. Int. Ed. 2011, 50, 9391–9395; d) B. Fraser-Reid,
J. Lu, K. N. Jayaprakash, J. C. Lopez, Tetrahedron: Asymmetry 2006,
17, 2449–2463; e) H. Tanaka, T. Kawai, Y. Adachi, N. Ohno, T. Ta-
kahashi, Chem. Commun. 2010, 46, 8249–8251; f) H. Tanaka, T.
Kawai, Y. Adachi, S. Hanashima, Y. Yamaguchi, N. Ohno, T. Taka-
hashi, Bioorg. Med. Chem. 2012, 20, 3898–3914; g) H. Tanaka, Y.
Nishiura, T. Takahashi, J. Org. Chem. 2009, 74, 4383–4386; h) V. Y.
Dudkin, M. Orlova, X. Geng, M. Mandal, W. C. Olson, S. J. Dani-
shefsky, J. Am. Chem. Soc. 2004, 126, 9560–9562; i) C. Unverzagt,
G. Gundel, S. Eller, R. Schuberth, J. Seifert, H. Weiss, M. Niemietz,
M. Pischl, C. Raps, Chem. Eur. J. 2009, 15, 12292–12302; j) D.
Crich, H. Li, Q. Yao, D. J. Wink, R. D. Sommer, A. L. Rheingold, J.
Am. Chem. Soc. 2001, 123, 5826–5828.
[8] a) Handbook of Chemical Glycosylation: Advances in Stereoselectiv-
ity and Therapeutic Relevance (Ed.: A. V. Demchenko), Wiley-
VCH, Weinheim, 2008; b) M. C. Galan, D. Benito-Alifonso, G. M.
Watt, Org. Biomol. Chem. 2011, 9, 3598–3610; c) D. Crich, Acc.
Chem. Res. 2010, 43, 1144–1153; d) X. Guinchard, S. Picard, D.
Crich in Modern Tools for the Synthesis of Complex Bioactive Mole-
cules (Eds.: J. Cossy, S. Arseniyadis), John Wiley & Sons, Hoboken,
NJ, 2012, Chapter 12, pp. 395–432.
[9] W. Koenigs, E. Knorr, Ber. Dtsch. Chem. Ges. 1901, 34, 957–981.
[10] a) X. Zhu, R. R. Schmidt, Angew. Chem. 2009, 121, 1932–1967;
Angew. Chem. Int. Ed. 2009, 48, 1900–1934; b) C.-H. Hsu, S.-C.
Hung, C.-Y. Wu, C.-H. Wong, Angew. Chem. 2011, 123, 12076–
12129; Angew. Chem. Int. Ed. 2011, 50, 11872–11923; c) C. Y. Wu,
C. H. Wong, Top. Curr. Chem. 2010, 301, 223–252.
44b (99 mg, 0.140 mmol, 93%) as
a yellow foam: Rf =0.12 (100:5
EtOAc/MeOH); [a]2D4 = +10.7 (c=0.81, MeOH); 1H NMR (500 MHz,
[D7]DMF, 363 K): d=1.91–1.99 (m, 2H; CH2), 2.06–2.15 (m, 1H; H5),
2.50 (t, J=10.7 Hz, 1H; H2), 2.63 (t, J=11.3 Hz, 1H; H7), 2.75–2.80 (m,
2H; CH2), 3.36–3.41 (m, 1H; H5’), 3.46 (dd, J=8.5, 9.8 Hz, 1H; H4),
3.52 (t, J=9.1 Hz, 1H; H4’), 3.58–3.66 (m, 3H; H2, H3’, H7), 3.76–3.74
(m, 3H; H6, H6’), 3.76 (t, J=6.3 Hz, 2H; CH2), 3.77–3.82 (m, 1H; H3),
3.85–3.91 (m, 1H; H6’), 3.96 (t, J=9.1 Hz, 1H; H2’), 4.11 (brs, 1H; OH),
ꢀ
4.51 and 4.53 (AB, JAB =12.1 Hz, 2H; CH2 Bn), 4.64 and 4.90 (AB, JAB
=
=
ꢀ
11.5 Hz, 2H; CH2 Bn), 4.68 (brs, 1H; OH), 4.71 and 4.75 (AB, JAB
ꢀ
11.7 Hz, 2H; CH2 Bn), 4.75 (brs, 1H; OH), 4.94 (brs, 1H; OH), 5.59 (d,
ꢀ
ꢀ
J=9.1 Hz, 1H; H1’), 7.26–7.44 (m, 15H; H Ar), 7.92 ppm (s, 1H; CH
13
ꢀ
triazole); C NMR (125 MHz, [D7]DMF, 373 K): d=148.1 (Cq triazole),
140.7 (Cq Bn), 140.4 (Cq Bn), 140.2 (Cq Bn), 129.3 (CH Ar), 129.3
(CH Ar), 129.2 (CH Ar), 128.8 (CH Ar), 128.7 (CH Ar), 128.5 (CH
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Ar), 128.5 (CH Ar), 128.3 (CH Ar), 121.7 (CH triazole), 89.5 (C1’),
ꢀ
81.5 (C4), 81.4 (C3’), 80.1 (C3), 79.2 (C5’), 75.0 (CH2 Bn), 74.3 (C2’),
ꢀ
ꢀ
74.1 (CH2 Bn), 73.0 (CH2 Bn), 71.9 (C4’, CH2), 70.5 (C6), 63.1 (C6’),
59.1 (C2), 58.2 (C7), 40.8 (C5), 29.8 (CH2), 23.4 ppm (CH2); IR (neat):
nmax =3348, 2921, 2856, 1631, 1454, 1364, 1092, 1042, 1027, 898, 734,
696 cmꢀ1; HRMS (ESI-TOF): calcd for C38H47N4O9 [MꢀH]ꢀ: 703.3343;
found: 703.3374. BCl3 (1m solution in hexanes, 380 mL, 0.38 mmol,
9 equiv) was added dropwise to
a stirred solution of 44b (30 mg,
0.043 mmol, 1 equiv) in anhydrous CH2Cl2 (4 mL) at 08C. The reaction
mixture was then stirred at room temperature for 5 h before being
quenched by addition of MeOH (3 mL) and then concentrated in vacuo.
The residue was dissolved in H2O (5 mL) and the solution was neutral-
ized with Amberlite IRA400 (OHꢀ form). The mixture was further stir-
red for 30 min and filtered. The resin was washed profusely with H2O,
and the filtrate was lyophilized to afford 45b (18.1 mg, 0.042 mmol,
98%) in pure form as a pale yellow amorphous solid: Rf =0.72 (3:2:1
EtOAc/iPrOH/H2O); [a]2D4 = +2.5 (c=1.5, MeOH); 1H NMR (500 MHz,
[D7]DMF, 363 K): d=1.74–1.87 (m, 1H; H5), 1.89–2.01 (m, 2H; CH2),
2.27–2.42 (m, 2H; H2, H7), 2.74–2.82 (m, 2H; CH2), 3.20 (t, J=9.1 Hz,
1H; H4), 3.32–3.91 (m, 16H; H2, H3, H7, H6, H3’, H4’, H5’, H6’, CH2,
4OH), 3.96 (t, J=8.8 Hz, 1H; H2’), 5.59 (d, J=9.1 Hz, 1H; H1’),
7.94 ppm (s, 1H; CH triazole); 13C NMR (125 MHz, [D7]DMF, 363 K):
ꢀ
ꢀ
ꢀ
d=148.0 (Cq triazole), 121.6 (CH triazole), 89.4 (C1’), 81.4 (C5’), 79.1
(C3’), 76.3 (C4), 74.2 (C2’), 72.2 (C3), 71.8 (C4’), 71.7 (CH2), 62.9 (C6,
C6’), 61.8 (C2), 58.3 (C7), 43.1 (C5), 29.7 (CH2), 23.4 ppm (CH2); IR
[11] a) L. Bohꢃ, D. Crich, Trends Glycosci. Glycotechnol. 2010, 22, 1–15;
b) B. Sylla, K. Descroix, C. Pain, C. Gervaise, F. Jamois, J.-C. Yvin,
L. Legentil, C. Nugier-Chauvin, R. Daniellou, V. Ferriꢆres, Carbo-
hydr. Res. 2010, 345, 1366–1370; c) N. M. Spijker, C. A. A. van
Boeckel, Angew. Chem. 1991, 103, 179–182; Angew. Chem. Int. Ed.
Engl. 1991, 30, 180–183.
(neat): nmax =3285, 1372, 1036 cmꢀ1
; HRMS (ESI-TOF): calcd for
C17H31N4O9 [M+H]+: 435.2091; found: 435.2090.
[12] a) G. Malik, X. Guinchard, D. Crich, Org. Lett. 2012, 14, 596–599;
b) A. Ferry, X. Guinchard, P. Retailleau, D. Crich, J. Am. Chem.
Soc. 2012, 134, 12289–12301.
Acknowledgements
[13] Glycomimetics: Modern Synthetic Methodologies (Ed.: R. Roy),
American Chemical Society, Washington, DC, 2005.
[14] a) F. G. Riddell, E. S. Turner, A. Boyd, Tetrahedron 1979, 35, 259–
261; b) A. Hassan, M. I. M. Wazeer, H. P. Perzanowski, S. Asrof Ali,
J. Chem. Soc. Perkin Trans. 2 1997, 411–418; c) M. Raban, D. Kost,
Tetrahedron 1984, 40, 3345–3381.
[15] a) K. Mꢈller, A. Eschenmoser, Helv. Chim. Acta 1969, 52, 1823–
1830; b) R. G. Kostyanovsky, V. F. Rudchenko, V. G. Shtamburg, I. I.
Chervin, S. S. Nasibov, Tetrahedron 1981, 37, 4245–4254.
[16] P. Bach, M. Bols, Tetrahedron Lett. 1999, 40, 3461–3464.
[17] Iminosugars: From Synthesis to Therapeutic Applications (Eds.: P.
Compain, O. R. Martin), Wiley, Chichester, 2007.
[18] M. Isobe, S. Sakamura, Agric. Biol. Chem. 1974, 38, 1111–1112.
[19] a) M. Bols, Acc. Chem. Res. 1998, 31, 1–8; b) A. Bꢈlow, I. W. Ples-
ner, M. Bols, J. Am. Chem. Soc. 2000, 122, 8567–8568.
[20] P. Jakobsen, J. M. Lundbeck, M. Kristiansen, J. Breinholt, H.
Demuth, J. Pawlas, M. P. Torres Candela, B. Andersen, N. Wester-
gaard, K. Lundgren, N. Asano, Bioorg. Med. Chem. 2001, 9, 733–
744.
J.-F. Gallard and M.-T. Martin (ICSN) are thanked for assistance with
the recording of variable-temperature NMR spectra. A.F. and G.M.
thank the ICSN for financial support.
[1] a) P. Stallforth, B. Lepenies, A. Adibekian, P. H. Seeberger, J. Med.
Chem. 2009, 52, 5561–5577; b) P. H. Seeberger, Nat. Chem. Biol.
2009, 5, 368–372.
[2] a) M. Novak, V. Vetvicka, J. Immunotoxicol. 2008, 5, 47–57; b) K.
Descroix, V. Vetvicka, I. Laurent, F. Jamois, J.-C. Yvin, V. Ferriꢆres,
Bioorg. Med. Chem. 2010, 18, 348–357; c) V. Vetvicka, J.-C. Yvin,
Int. Immunopharmacol. 2004, 4, 721–730; d) I. Chlubnovꢁ, B. Sylla,
C. Nugier-Chauvin, R. Daniellou, L. Legentil, B. Kralova, V. Fer-
rieres, Nat. Prod. Rep. 2011, 28, 937–952.
[3] J. A. Bohn, J. N. BeMiller, Carbohydr. Polym. 1995, 28, 3–14.
[4] a) G. D. Ross, V. Vetvicka, J. Yan, Y. Xia, J. Vetvickovꢁ, Immuno-
pharmacology 1999, 42, 61–74; b) J. Yan, V. Vetvicka, Y. Xia, A.
Coxon, M. C. Carroll, T. N. Mayadas, G. D. Ross, J. Immunol. 1999,
163, 3045–3052.
2178
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2168 – 2179

ꢀ
N O Bonds in N-Oligosaccharide Mimics
FULL PAPER
[21] L. Sun, P. Li, N. Amankulor, W. Tang, D. W. Landry, K. Zhao, J.
Org. Chem. 1998, 63, 6472–6475.
[22] E. Martꢇnez-Castro, A. Gonzalez-Benjumea, O. Lopez, I. Maya, E.
Alvarez, J. G. Fernꢁndez-BolaÇos, Org. Biomol. Chem. 2012, 10,
4220–4228.
[23] a) P. Truitt, L. M. Long, M. Mattison, J. Am. Chem. Soc. 1948, 70,
2829–2829; b) R. Sun, M. Lꢈ, L. Chen, Q. Li, H. Song, F. Bi, R.
Huang, Q. Wang, J. Agric. Food Chem. 2008, 56, 11376–11391; c) K.
Yamawaki, T. Nomura, T. Yasukata, K. Uotani, H. Miwa, K.
Takeda, Y. Nishitani, Bioorg. Med. Chem. 2007, 15, 6716–6732;
d) O. F. Foot, D. W. Knight, Chem. Commun. 2000, 975–976.
[24] E. W. Baxter, A. B. Reitz in Organic Reactions, Vol. 59 (Ed.: L. E.
Overman), John Wiley & Sons, New York, 2002.
127–132; e) A. Ezzitouni, P. Russ, V. E. Marquez, J. Org. Chem.
1997, 62, 4870–4873; f) A. Ezzitouni, V. E. Marquez, J. Chem. Soc.
Perkin Trans. 1 1997, 1073–1078; g) N. Gathergood, K. R. Knudsen,
K. A. Jorgensen, J. Org. Chem. 2001, 66, 1014–1017; h) M. G. Woll,
J. D. Fisk, P. R. LePlae, S. H. Gellman, J. Am. Chem. Soc. 2002, 124,
12447–12452; i) O. R. Ludek, C. Meier, Synthesis 2003, 2101–2109;
j) O. R. Ludek, T. Krꢉmer, J. Balzarini, C. Meier, Synthesis 2006,
1313–1324; k) O. R. Ludek, V. E. Marquez, Tetrahedron 2009, 65,
8461–8467; l) S. Jessel, C. Meier, Eur. J. Org. Chem. 2011, 1702–
1713.
[41] H. C. Brown, B. Singaram, J. Org. Chem. 1984, 49, 945–947.
[42] For allylic oxydation to ketone: a) with Jones reagent, see: W. G.
Salmond, M. A. Barta, J. L. Havens, J. Org. Chem. 1978, 43, 2057–
2059; b) with hypervalent iodines, see: M. Catir, H. Kilic, Synlett
2010, 1319–1322; c) with sodium chlorite, see: S. M. Silvestre,
J. A. R. Salvador, Tetrahedron 2007, 63, 2439–2445; d) with
Rh2(caprolactamate)4, see: A. J. Catino, R. E. Forslund, M. P. Doyle,
J. Am. Chem. Soc. 2004, 126, 13622–13623. For allylic oxidation to
alcohol: e) with SeO2, see: K. B. Sharpless, R. F. Lauer, J. Org.
Chem. 1972, 37, 3973–3974. f) For Pd-based oxidation, see: M. S.
Chen, N. Prabagaran, N. A. Labenz, M. C. White, J. Am. Chem. Soc.
2005, 127, 6970–6971.
[25] a) T. M. Jespersen, M. Bols, M. R. Sierks, T. Skrydstrup, Tetrahedron
1994, 50, 13449–13460; b) C. Matassini, S. Mirabella, A. Goti, F.
Cardona, Eur. J. Org. Chem. 2012, 3920–3924.
[26] M. Du, O. Hindsgaul, Synlett 1997, 395–397.
[27] a) M. Cerny, L. Kalvoda, J. Pacak, Collect. Czech. Chem. Commun.
1968, 33, 1143–1156; b) M. Cerny, J. Pecka, Collect. Czech. Chem.
Commun. 1973, 38, 132–142; c) M. Cerny, J. Stanek, J. Pacak, Col-
lect. Czech. Chem. Commun. 1969, 34, 849–858; d) M. Cerny, T.
Trnka, P. Beran, J. Pacak, Collect. Czech. Chem. Commun. 1969, 34,
3377–3382; e) T. Trnka, M. Cerny, Collect. Czech. Chem. Commun.
1971, 36, 2216–2225.
[43] K. C. Nicolaou, V. A. Adsool, C. R. H. Hale, Org. Lett. 2010, 12,
1552–1555.
[28] a) T. P. Murray, U. P. Singh, R. K. Brown, Can. J. Chem. 1971, 49,
2132–2138; b) K. Ranganayakulu, R. K. Brown, J. Org. Chem. 1974,
39, 3941–3943; c) K. Ranganayakulu, U. P. Singh, T. P. Murray,
R. K. Brown, Can. J. Chem. 1974, 52, 988–992; d) T. B. Grindley,
G. J. Reimer, J. Kralovec, R. G. Brown, M. Anderson, Can. J. Chem.
1987, 65, 1065–1071; e) C. Leteux, A. Veyrieres, F. Robert, Carbo-
hydr. Res. 1993, 242, 119–130.
[44] a) R. F. Borch, M. D. Bernstein, H. Dupont Durst, J. Am. Chem.
Soc. 1971, 93, 2897–2904; b) G. F. Painter, P. J. Eldridge, A. Falshaw,
Bioorg. Med. Chem. 2004, 12, 225–232.
[45] a) A. C. Cope, T. T. Foster, P. H. Towle, J. Am. Chem. Soc. 1949, 71,
3929–3934; b) H. Miyabe, Synlett 2012, 1709–1724.
[46] X. Guinchard, Y. Vallꢃe, J.-N. Denis, J. Org. Chem. 2007, 72, 3972–
3975.
[29] J. Xue, Z. Guo, Tetrahedron Lett. 2001, 42, 6487–6489.
[30] J. M. F. G. Aerts, H. S. Overkleeft, Novel anti-inflammatory pro-
drugs, WO 2008/153394 A2, 2008.
[31] S. Arndt, L. C. Hsieh-Wilson, Org. Lett. 2003, 5, 4179–4182.
[32] R. J. B. H. N. van den Berg, W. Donker-Koopman, J. H. van Boom,
H. M. F. G. Aerts, D. Noort, Bioorg. Med. Chem. 2004, 12, 891–902.
[33] Yields were dependent on the quality of the commercial reagent.
(The best yields were obtained with a 0.7m solution in THF supplied
from ACROS.)
[47] a) S. Desvergnes, Y. Vallꢃe, S. Py, Org. Lett. 2008, 10, 2967–2970;
b) E. Racine, C. Bello, S. Gerber-Lemaire, P. Vogel, S. Py, J. Org.
Chem. 2009, 74, 1766–1769; c) S. Py, S. Desvergnes, Y. Vallꢃe,
FR2902097, 2007.
[48] W. Dong, T. Jespersen, M. Bols, T. Skrydstrup, M. R. Sierks, Bio-
chemistry 1996, 35, 2788–2795.
[49] a) M. D. Sørensen, N. M. Khalifa, E. B. Pedersen, Synthesis 1999,
1937–1943; b) H. H. Jensen, A. Lohse, B. O. Petersen, J. O. Duus,
M. Bols, J. Chem. Soc. Perkin Trans. 1 2000, 667–670; c) H. H.
Jensen, L. Lyngbye, M. Bols, Angew. Chem. 2001, 113, 3555–3557;
Angew. Chem. Int. Ed. 2001, 40, 3447–3449.
[34] Unreacted starting material (12%) could be recovered by column
chromatography.
[35] The structures of the two epimers were elucidated after comparison
with the same epimers obtained by the cyclopentene strategy.
[36] Conditions used: dialdehyde (1 equiv), BnONH3Cl (1 equiv),
NaBH3CN (2.2 equiv), AcOH (2 equiv), 3 ꢄ molecular sieves,
MeOH (0.05m), room temperature, 20 h.
[50] a) N. Ardes-Guisot, D. S. Alonzi, G. Reinkensmeier, T. D. Butters,
C. Norez, F. Becq, Y. Shimada, S. Nakagawa, A. Kato, Y. Bleriot,
M. Sollogoub, B. Vauzeilles, Org. Biomol. Chem. 2011, 9, 5373–
5388; b) A. Baron, Y. Bleriot, M. Sollogoub, B. Vauzeilles, Org.
Biomol. Chem. 2008, 6, 1898–1901.
[37] Unreacted starting material (25%) was recovered.
[51] NMR spectra recorded in deuterated DMF.
[38] G. Mehta, N. Mohal, Tetrahedron Lett. 2000, 41, 5747–5751.
[39] G. Mehta, N. Mohal, Tetrahedron Lett. 1999, 40, 5791–5794.
[40] a) K. Biggadike, A. D. Borthwick, A. M. Exall, B. E. Kirk, S. M.
Roberts, P. Youds, A. M. Z. Slawin, D. J. Williams, J. Chem. Soc.
Chem. Commun. 1987, 255–256; b) K. Biggadike, A. D. Borthwick,
D. Evans, A. M. Exall, B. E. Kirk, S. M. Roberts, L. Stephenson, P.
Youds, J. Chem. Soc. Perkin Trans. 1 1988, 549–554; c) R. Zahler,
W. A. Slusarchyk, Hydroxymethyl(methylenecyclopentyl)Purines and
Pyrimidines, US5206244, 1992; d) G. S. Bisacchi, S. T. Chao, C.
Bachard, J. P. Daris, S. Innaimo, G. A. Jacobs, O. Kocy, P. Lapointe,
A. Martel, Z. Merchant, W. A. Slusarchyk, J. E. Sundeen, M. G.
Young, R. Colonno, R. Zahler, Bioorg. Med. Chem. Lett. 1997, 7,
[52] The assignment of the signals corresponding to the ring core is more
difficult at low temperature due to insufficent resolution of the sig-
nals concerned.
[53] J. Sandstrçm, Dynamic NMR Spectroscopy, Academic Press,
London, 1982.
[54] J. B. Lambert, R. G. Keske, J. Am. Chem. Soc. 1966, 88, 620–622.
[55] For a dynamic NMR study on N-hydroxypyrrolidine, see: J. M.
Lehn, J. Wagner, Tetrahedron 1970, 26, 4227.
Received: July 3, 2012
Revised: September 25, 2012
Published online: December 19, 2012
Chem. Eur. J. 2013, 19, 2168 – 2179
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2179