Synthesis, conformational studies and mannosidase stability of a mimic of 1,2-mannobioside

Article abstract of DOI:10.1002/ejoc.200400520

Dimethyl (1S,2S,4S,5S)-4-allyloxy-5-(α-mannosyloxy)cyclohexane-1,2- dicarboxylate (3) was designed as a structural mimic of α(1,2)mannobioside (1). Its synthesis and structural analysis by NMR spectroscopy and molecular modelling are described. The result

Full text of DOI:10.1002/ejoc.200400520

FULL PAPER
Synthesis, Conformational Studies and Mannosidase Stability of a Mimic of
1,2-Mannobioside
Silvia Mari,[b] Helena Posteri,[a] Gilles Marcou,[a] Donatella Potenza,[a] Fabrizio Micheli,[c]
[b]
[b]
[a]
˜
F. Javier Canada, Jesus Jimenez-Barbero,* and Anna Bernardi*
Keywords: Carbohydrate mimics / Conformational analysis / Mannobioside
Dimethyl (1S,2S,4S,5S)-4-allyloxy-5-(α-mannosyloxy)cyclo- pseudo-disaccharide can be used as a structural mimic of
hexane-1,2-dicarboxylate (3) was designed as a structural
mimic of α(1,2)mannobioside (1). Its synthesis and structural
analysis by NMR spectroscopy and molecular modelling are
described. The results show that 3, like 1, populates two low-
energy conformations — stacked (S) and extended (E) — that
are in fast dynamic equilibrium around the glycosidic lin-
mannobioside. The mannosidase stability of 3 was found to
be significantly higher (sixfold) than that of natural manno-
bioside. This is a very useful feature of this mimic and is en-
couraging for its future development.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
kage. Thus, the data confirm the expectation that the Germany, 2004)
Introduction
by appropriate experimental work,[7Ϫ10] molecular model-
ling has also allowed us to obtain qualitative predictions of
the binding mode of the new substrates and to design
further simplifications of glycomimetic structures. We now
present our results relative to glycomimetics containing an
analog of the αMan1,2-disubstituted framework (a trans-
diaxial diol). A mimic of the 1,2-mannobioside 1 was pre-
pared by replacing the reducing-end mannose unit with the
During the past few years a number of publications and
reviews have illustrated the potential of sugar mimics in dif-
ferent fields of drug discovery.[1Ϫ3] The main advantage of
this approach is rooted in the improved drug-like character
of glycomimetics compared to natural carbohydrates. Sugar
mimics are generally more soluble and membrane penetr-
ant, less hydrophilic and less metabolically labile than the
sugars themselves.
Our groups are actively involved in the rational design
and synthesis of glycomimetics based on the concept of
scaffold replacement.[4] In brief, we have proposed the use
of conformationally stable cyclic diols to replace the non-
pharmacophoric parts of bioactive oligosaccharides, while
preserving the correct orientation of the pharmacophoric
portion of the sugar. Computational tools have been used
to predict the three-dimensional structures of virtual mimic
candidates and to compare them with the structure of the
natural counterpart. This approach has been validated for
a 3,4-disubstituted galactose scaffold (a cis diol) using the
GM1:CT recognition pair as a model system.[5,6] Supported
carbocyclic diol
2 [(1S,2S,4S,5S)-dicarboxycyclohexane
diol, DCCHD; Figure 1],[4] a conformationally stable 1,2
trans-diaxial diol.[4,11] Computer-aided molecular design
(see below) suggested that 2 can be used as a mimic of the
2-substituted α-mannose unit in 1 without altering the over-
all shape of the molecule and therefore the presentation of
the terminal (non-reducing end) residue. NMR studies con-
firmed that the resulting pseudo-mannobioside 3 has indeed
the same conformational behavior as the natural disacchar-
ide, but exhibits improved stability towards the activity of
jack-bean mannosidase.
Synthesis of the Mannobioside Mimic 3
The enantiopure trans-diaxial DCCHD scaffold
5
(Scheme 1) was synthesized in 81% overall yield starting
Dipartimento di Chimica Organica e Industriale e Centro di from the known diacid 4[12] by m-chloroperbenzoic acid
[a]
Eccellenza CISI, Universita’ di Milano
(MCPBA) double-bond oxidation, followed by Cu(OTf)2-
Via Venezian 21, 20133 Milano, Italy
promoted opening of the epoxide with allyl alcohol.[4] The
Fax: (internat.) ϩ39-02-503-14092
E-mail: anna.bernardi@unimi.it
pseudo-disaccharide skeleton was then assembled by glyco-
sylation of the acceptor 5 (Scheme 1) with tetraacetylman-
nose trichloroacetimidate (6).[13] The reaction was pro-
moted by 0.25 molar equivalents of trimethylsilyl triflate
(TMSOTf) in CH2Cl2 at Ϫ20 °C. At very short contact
times (10 min) the orthoester 7 was initially formed by nu-
[b]
[c]
Departamento de Estructura y Funcion de Proteinas, Centro
de Investigaciones Biologicas CSIC,
Ramiro de Maeztu 9, 28040 Madrid, Spain
Fax: (internat.) ϩ34-91-536-0432
E-mail: jjbarbero@cib.csic.es
GSK Medicine Research Centre
Via Fleming 4, 37135 Verona, Italy
Eur. J. Org. Chem. 2004, 5119Ϫ5125
DOI: 10.1002/ejoc.200400520
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5119

J. Jimenez-Barbero, A. Bernardi et al.
FULL PAPER
titatively afforded 3, whose structure was unequivocally
confirmed by 1H and 13C NMR spectroscopy and by
mass spectrometry.
Conformational Analysis of the Mannobioside Mimic 3
The conformation and dynamics of the Manα1-2-Man
linkage were studied extensively in the oligomannose oligo-
saccharide Man9GlcNAc2.[15] Computational[15,16] and ex-
perimental[17] results suggested that the linkage exists in two
distinct, flexible conformations with similar ϕ values and
different, slowly interconvertings ψ values. The two confor-
mations, which have been termed S, for ‘‘stacked’’, and E,
for ‘‘extended’’, (Figure 2) are both represented in available
X-ray structures of mannose oligosaccharides and account
for the set of four NOE contacts typically observed across
the glycosidic linkage.[17] The H1-H1Ј contact (Figure 2)
can arise from conformer S, and the H5-H1Ј and H1-H3Ј
contacts (Figure 2) from conformer E. The fourth observed
NOE contact belongs to the interglycosidic H1-H2Ј pair,
which is at NOE distance in both conformations. (The
mutually exclusive NOE contacts are marked by arrows in
Figure 2).
We recently reported[16] that modeling of 1 using the AM-
BER* force field and GB/SA solvation model gives results
that are fully compatible with previous simulations and
available experimental data. Two conformers, S (global
minimum) and E (ϩ6 kJ/mol), were located by an MC/EM
conformational search, and were found to be freely in-
terconverting and almost equally populated by MC/SD dy-
namic simulations. The same approach was then used to
predict the three-dimensional structure of the 1,2-α-manno-
bioside mimic 3; the molecule was calculated to populate
the same region of conformational space as 1. These results
are presented in Figure 3. The unconventional numbering
of the DCCHD ring shown in Figure 3 is used to facilitate
comparison with 1. An MC/EM conformational search of
3 located the S (ϕ ϭ 94°, ψ ϭ Ϫ72°; see c in Figure 3) and
E (ϕ ϭ 70°, ψ ϭ Ϫ117°; see d in Figure 3) conformers at
ϩ2.6 kJ/mol and 0.0 kJ/mol relative energy, respectively.
The dynamic simulation (10 ns, see b in Figure 3) showed
that the two conformations are in fast dynamic equilibrium
around the glycosidic linkage. Compared to 1, the confor-
mational space available to 3 is more biased towards the
region populated by extended (E) conformations. Further-
more, the E region of the conformational space appears
broad and is not centered around its local minimum at ϕ ϭ
70°, ψ ϭ Ϫ117°, but rather around larger values of ψ. This
can clearly be seen in the map of Figure 3, and has import-
ant consequences on the calculated interproton distances,
some of which display a strong dependence on the value of
the ψ torsion (cf the distances calculated by the MC/SD
dynamics with those of the isolated minima in Table 1). In
summary, computer-aided molecular design clearly sup-
ports the expectation that the pseudo-disaccharide 3 can be
used as a structural mimic of mannobioside 1.
Figure 1. Mannobioside 1 and its DCCHD-based mimic 3
cleophilic attack of 5 on the 2-acetate protecting group of
6.[14] This product could be isolated by flash chromatogra-
phy, and converted into 8 by treatment with TMSOTf in
CH2Cl2 at Ϫ20 °C.[14] When allowing the reaction between
5 and 6 to stand for a somewhat longer time (20 min), the
desired product 8 was obtained directly in 65% yield, after
chromatography. Deacetylation (NaOMe in MeOH) quan-
Scheme 1. Synthesis of the pseudo-mannobioside 3: (a) MeOH,
cat. H2SO4 (95%); (b) MCPBA, CH2Cl2 (95%); (c) allyl alcohol,
10% Cu(OTf)2 (90%); (d) 25% TMSOTf, CH2Cl2, Ϫ20 °C (65%);
(e) NaOMe, MeOH (88%)
NOESY data were collected at 400 MHz for 3 in D2O
solution at 25 °C (Table 1, column 1) and compared with
the MC/SD simulation predictions. A qualitative analysis
5120
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 5119Ϫ5125

A Mimic of 1,2-Mannobioside
FULL PAPER
Figure 2. Low-energy conformations of the 1,2-mannobioside 1: (a) the stacked conformation S (Φ ϭ 88°, Ψ ϭ Ϫ58°); (b) the extended
conformation E (Φ ϭ 72°, Ψ ϭ Ϫ138°); characteristic NOE contacts are indicated by arrows
Figure 3. Conformational studies on pseudo 1,2-α-mannobioside 3 (MMOD 5.5, AMBER*, GB/SA water solvation); an unconventional
numbering of the DCCHD moiety is used to help comparison with 1,2-mannobioside. Φ: O5ϪC1ϪO1-C2Ј; Ψ: C1ϪO1ϪC2ЈϪC1Ј; (a)
10000 steps of MC/EM conformational search; (b) 10 ns of MC/SD dynamic simulation; (c) conformer S (∆E ϭ 2.6 kJ/mol; Φ ϭ 94°,
Ψ ϭ Ϫ72°); (d) the lowest energy conformation E (Φ ϭ 70°, Ψ ϭ Ϫ117°)
Table 1. Interresidue NOE contacts observed in the D2O spectrum of 3
Proton pair[a]
NOE intensity[b]
(D2O)
Calcd. distance
(A)
Calcd. distance E
(A)
Calcd. distance S
(A)
Exp. distance range in 1
[c]
[d]
˚
˚
˚
˚
(A)
H1-H1Ј
H1-H2Ј
H1-H4Ј
H1-H3Јax
H1-H3Јeq
H2-H4Ј
H5-H2Ј
H5-H1Ј
H5-H6Јax
H5-H4Ј
w
s
w
Ϫ
s
m
m
s
m
w
3.4
2.3
3.8
4.0
2.7
3.7
3.4
2.6
2.8
4.2
3.9
2.2
4.2
4.1
3.0
4.0
3.6
2.3
2.9
4.5
2.6
2.2
4.6
4.5
4.0
4.8
3.6
4.4
2.2
3.2
2.9Ϫ3.0
2.1Ϫ2.2
no NOE
3.2Ϫ3.6
Ϫ
n.r.[e]
n.r.[e]
2.5Ϫ2.7
Ϫ
n.r.[e]
[a]
[b]
[c]
Numbering as in Figures 2 and 3.
s: strong, m: medium, w: weak. Evaluated from ϽrϪ6Ͼ monitored during 10 ns of the MC/
SD simulation. From ref.[15] [e] Not reported.
[d]
Eur. J. Org. Chem. 2004, 5119Ϫ5125
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5121

J. Jimenez-Barbero, A. Bernardi et al.
FULL PAPER
Figure 4. NOESY spectrum of mimic 3 in D2O at 400 MHz (25 °C) with a mixing time of 800 ms; exclusive S and E conformation NOE
contacts (indicated by the arrows) are found in the spectrum (circles), thus confirming that both the predicted conformations coexist
of the NOE crosspeaks using a basic strong-medium-weak of mannobioside, and preferentially populates the extended
intensity approach allowed us to deduce the presence of (E) conformation.
both E and S conformers in fast conformational equilib-
Mannosidase Stability of 3
rium. Both sets of mutually exclusive NOE contacts were
found in the spectra — the H1-H1Ј contact originating
from the S conformers, and the H1-H3Ј and H5-H1Ј con- was investigated by submitting it to hydrolytic conditions
tacts originating from the E conformers (Figure 4). in the presence of commercial jack-bean mannosidase. The
The stability of 3 to mannosidase-catalyzed hydrolysis
Despite this qualitative approach, we could safely deduce substrate was incubated with increasing concentrations of
that the E conformer is the major one in solution, given the the enzyme at pH 4.5 (maximum activity pH for α-mannos-
strong differences between the intensities of the correspond- idases).[18] After 30 min of incubation at 37 °C the reaction
ing NOE cross-peaks. Indeed, the observed correlations are mixture was lyophilized and the products were analyzed by
in excellent agreement with the simulations (Table 1).
GC. The extent of hydrolysis was evaluated by integration
Compared to the experimental data range reported for of the peaks relative to 3 (retention time, tR, of 13.5 min)
the equivalent distances in 1[15] (Table 1, last column), our and to the aglycon 5 (tR ϭ 7.1 min). At the lowest enzyme
data confirm that the conformational space of pseudo-man- concentration (0.5 µg/mL) 4% hydrolysis of mimic 3 was
nobioside 3 is biased towards the extended conformation revealed by GC analysis, while, upon increasing the enzyme
region. This is clearly shown by the weak intensity of the concentration up to 10 µg/mL, only 10% hydrolysis was ob-
H1-H1Ј cross-peak (Table 1, entry 1, and Figure 4, inset of served. Under the same conditions 24% of the natural di-
the NOESY spectrum), a reporter signal of conformer S saccharide 1 is already hydrolysed at an enzyme concen-
(Figure 4, S), which is much stronger in 1 than in 3. Fur- tration of 0.5 µg/mL and 60% is hydrolysed at 10 µg/mL
thermore, the strong interaction observed in the NOESY enzyme concentration. Thus, 3 indeed appears to be a sub-
spectrum of 3 for the H5-H1Ј pair (Table 1, entry 8) is typi- strate of jack-bean mannosidase, but it appears to be signif-
cal of extended conformations (Figure 4, E). This picture is icantly more stable than 1 to the action of the enzyme.
also supported by the intense H1-H3Јeq cross-peak These observations prompted us to examine the ability of 3
(Table 1, entry 5, and Figure 4) and by the presence of a to perform as an inhibitor for this enzyme.
medium intensity H2-H4Ј NOE contact (Table 1, entry 6),
Investigation of 3 as a possible inhibitor of mannosidase
which is not seen in 1. This latter contact could not be ex- activity was carried out using α-mannosidase from jack-
plained by the calculated structure of the E local minimum bean and p-nitrophenol-α--mannoside (PNP-Man) as the
(Table 1, Column 3), but is in good agreement with the re- substrate (Table 2). The enzymatic reactions were run at 25
sults of the dynamic simulation (Table 1, Column 1) and is °C for 20 min and then quenched by adding a 1  carbonate
an indication that the dynamic behavior of the molecule is solution until pH 12, which inactivates the enzyme. The ex-
well reproduced by the AMBER* calculations.
tent of the reaction was calculated from the absorbance of
Thus, the NMR studies experimentally support the com- p-nitrophenoxide measured at a wavelength of 400 nm
putational hypothesis that 3 behaves as a structural mimic (ε400 ϭ 1.77 ϫ 104 Ϫ1cmϪ1). The concentration of p-nitro-
5122
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 5119Ϫ5125

A Mimic of 1,2-Mannobioside
FULL PAPER
phenoxide versus concentration of substrate in the reaction sented on three-dimensionally well-defined structures, such
volume was plotted and fitted to the MichaelisϪMenten as the GM1 ganglioside[20] or appropriate structural mim-
equation to give a KM value of 4.7 ϫ 10Ϫ3 , which is simi- ics,[5] they give rise to tight complexes with dissociation con-
lar to the reported value of 2.5 ϫ l0Ϫ3 .[18] Qualitative stants in the nanomolar range. The structural similarity of
inhibition studies were then performed using 3 as inhibitor 3 with the natural disaccharide 1 is therefore an important
and working with an enzyme concentration of 2.5 µg/mL element that depends critically on the nature of the dicar-
(Table 2). The inhibitory effect of 3 allows us to estimate an boxycyclohexanediol aglycon chosen to mimic the reduc-
average value for the inhibition constant, Ki, of 1.0 ϫ 10Ϫ2 ing-end mannose. The mimic 3 could thus be introduced
, assuming a competitive inhibitor model (Table 2). The into more complex oligosaccharidic structures, preserving
Ki values obtained are similar to those reported in the lit- epitope presentation, while imparting metabolic stability to
erature[18] for the known inhibitors mannono (1Ǟ4) and the synthetic construct. Furthermore, the structure of 3 al-
(1Ǟ5) lactone (1.0 ϫ 10Ϫ2  and 1.2 ϫ 10Ϫ4 , respec- lows easy conjugation to polyvalent aglycons, using either
tively).
the carboxy groups or the ether substituent. The confor-
mational behaviour of mimic 3 that we describe here could
have useful implications in recognition events where the E
conformation of the 1,2-mannobioside is selectively recog-
nized by the receptor.
An increased stability to enzymatic degradation is often
cited as one of the advantages of glycomimetics over natural
oligosaccharides in the development of drugs. The stability
of 3 to mannosidase-catalyzed hydrolysis appears, indeed,
to be significantly higher than the stability of the natural
compound 1. This is a very useful feature of this mimic and
is very encouraging for the future developments of
DCCHD-based glycomimetic drugs.
Table 2. Inhibition of jack-bean-mannosidase-catalyzed hydrolysis
of PNP-Man using 3. The inhibition experiments were carried out
for 20 min at 25 °C using 1.8 m PNP-Man in citrate buffer (pH
4.5), a constant concentration of the enzyme (2.5 µg/mL) and vari-
able concentrations of the inhibitor 3. The reaction was quenched
at pH 12 carbonate buffer and the concentration of the p-nitro-
phenoxide [PNP] was measured by UV spectroscopy. v0 (0.025 m/
min) was obtained under the same conditions but in the absence
of 3.
[3]
(m)
[PNP]
(m)
vInhibited
(m/min)
vI/v0
Ki[a]
(m)
12.8
10.1
5.1
0.28
0.31
0.32
0.41
0.014
0.015
0.016
0.021
0.565
0.626
0.651
0.841
11.9
12.3
6.8
2.5
9.7
Experimental Section
[a]
NMR spectra were recorded at 25Ϫ30 °C on Bruker AVANCE
400 and 500 MHz spectrometers. The cyclohexanediol moiety is
numbered as in Figure 3. Chemical shifts 1H and 13C NMR spectra
are expressed in ppm relative to TMS or to DSS for spectra re-
corded in D2O. The NOESY spectra of 3 were recorded with mix-
ing times of 400, 600 and 800 ms. Mass spectrometry was per-
formed with a VG 7070 EQ-HF apparatus (FAB ionization), an
Omniflex apparatus (MALDI ionization) or an Apex II ICR
FTMS (ESI ionization). Optical rotations [α]D were measured in a
1-dm path-length cell with 1 mL capacity on a PerkinϪElmer 241
polarimeter. Thin layer chromatography (TLC) was carried out
with precoated Merck F254 silica gel plates. Flash chromatography
(FC) was carried out with MachereyϪNagel silica gel 60 (230Ϫ400
mesh). Solvents were dried by standard procedures and reactions
requiring anhydrous conditions were run under nitrogen. The di-
acid 4[12] and the trichloroacetimmidate 6[13] were prepared by pub-
lished procedures.
,
where KM ϭ 4.67 m, [I] ϭ [3] and [S] ϭ [PNP-Man].
Discussion and Conclusions
The results presented here show that the enantiopure
trans-diaxial diol scaffold 5 can be used as a replacement
for a 1,2-disubstituted α-mannose residue to give, in a facile
and high-yielding synthetic sequence, a pseudo-mannobio-
side 3 which shows a high structural similarity with the
natural α-1,2-mannobioside. The conformational analysis
of 3 performed by NMR spectroscopy and supported by
computational modelling reveals that 3, like 1, populates
two low-energy conformations S (stacked) and E (extended)
Synthesis of Dimethyl (1S,2S,4S,5S)-4-Allyloxy-5-hydroxycyclohex-
ane-1,2-dicarboxylate (5): H2SO4 (0.5 mL, 0.009mol) was added to
that are in fast dynamic equilibrium. Relative to the natural a solution of the diacid 4[12] (7.02 g, 0.041mol) in dry MeOH (100
mL). The solution was stirred at reflux for about 20 h and then
concentrated under vacuum to about half of the original volume.
EtOAc was added, the organic phase was washed with saturated
NaHCO3 and dried with Na2SO4 and the solvent was evaporated
under reduced pressure to give dimethyl (1S,2S)-cyclohex-4-ene-
1,2-dicarboxylate in 95% yield. An analytical sample was purified
by flash chromatography: [α]D ϭ ϩ127.3 (c ϭ 1.23, CHCl3). 1H
NMR (200 MHz, CDCl3, 25 °C): δ ϭ 2.6Ϫ2.1 (m, 4 H, CH2), 2.85
(m, 2 H, H1, H2), 3.7 (s, 6 H, OCH3), 5.7 (m, 2 H, H4, H5) ppm.
13C NMR (53.3 MHz, CDCl3, 25 °C): δ ϭ 24.5, 40.96, 51.45,
counterpart, the conformational space available to the
pseudo-disaccharide appears to be biased toward the E re-
gion.
The α-1,2-mannobioside is the recognized epitope in vari-
ous important sugar-protein complexes.[19] Appropriate epi-
tope presentation is a key factor in determining ligand af-
finity to carbohydrate-binding proteins. For instance, it has
been shown recently that, although sialic acid and galac-
tose, per se, interact weakly with the cholera toxin (with
millimolar dissociation constants), when appropriately pre- 124.66, 174.79 ppm.
Eur. J. Org. Chem. 2004, 5119Ϫ5125
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5123

J. Jimenez-Barbero, A. Bernardi et al.
FULL PAPER
The crude reaction product was then sumbitted to epoxidation con-
ditions. Thus, MCPBA (1.85 g, 7.6 mmol) was added to a solution
of the diester (1.5 g, 5.5 mmol) in CH2Cl2 (13 mL). The solution
was stirred at room temperature for 3 h and the organic phase was
H, COCH3), 2.04Ϫ2.17 (m, 4 H, H3D, H6D), 3.3Ϫ3.5 (m, 2 H,
H4D, H5D), 3.40 (s, 3 H, OCH3), 3.46 (s, 3 H, OCH3), 3.57 (m, 1
H, H1D), 3.85 (dd, J ϭ 13, J ϭ 6 Hz, 1 H, H7a), 3.95 (m, 2 H, H7b,
H2D), 4.34 (m, 1 H, H5), 4.37 (dd, J ϭ 12, J ϭ 8 Hz, 1 H, H6a),
washed with saturated NaHCO3, dried and the solvents evaporated 4.49 (dd, J ϭ 12, J ϭ 6 Hz, 1 H, H6b), 5.11 (dd, J ϭ 9.2, J ϭ 1.5
under reduced pressure. The crude was purified by flash chroma-
Hz, 1 H, H9a), 5.15 (d, J ϭ 1.5 Hz, 1 H, H1), 5.27 (dd, J ϭ 16,
tography on silica gel (hexane/EtOAc, 85:15) to yield the epoxide J ϭ 1.5 Hz, 1 H, H9b), 5.65 (dd, J ϭ 3, J ϭ 2 Hz, 1 H, H2), 5.8 (t,
in 95% yield. [α]D ϭ ϩ71.6 (c ϭ 1.27, CHCl3). 1H NMR (200 MHz, J ϭ 9.7 Hz, 1 H, H4), 5.82Ϫ5.91 (m, 2 H, H3, H8) ppm. 13C-
C6D6, 25 °C): δ ϭ 1.55 (ddd, J ϭ 15, J ϭ 11, J ϭ 2.3 Hz, 1 H),
HETCOR (100.6 MHz, C6D6, 25 °C): δ ϭ 28.5, 52, 63, 70, 71, 72,
1.95 (m, 2 H), 2.30 (ddd, J ϭ 15, J ϭ 4, J ϭ 2 Hz, 1 H), 2.55 (m, 73, 75, 96, 118 ppm. MALDI-TOF: m/z ϭ 625.45 [M ϩ Naϩ].
1 H), 2.65 (m, 1 H), 2.75 (m, 1 H), 3.10 (ddd, J ϭ 11, J ϭ 11, J ϭ
Synthesis of the Pseudo-1,2-α-mannobioside 3: 1  Methanolic so-
dium methoxide (398 µL, 0.398 mmol) was added to a solution of
8 (120 mg, 0.199 mmol) in MeOH (20 mL) and stirring was con-
tinued at room temperature for 18 h. The reaction was monitored
by TLC (CHCl3/MeOH/H2O, 60:35:5). Amberlite IR-120 was then
added, the resin was removed by filtration and the solution was
concentrated. The residue was purified by flash chromatography
(CHCl3/MeOH, 9:1), to give 3 in 88% yield. [α]D ϭ ϩ58.2 (c ϭ 1,
CH3OH). 1H NMR (500 MHz, D2O, 25 °C): δ ϭ 1.75 (m, 1 H,
H3Dax), 1.82 (m, 1 H, H6Dax), 2.09 (m, 2 H, H3Deq, H6Deq), 2.88
(dt, J ϭ 11.5, J ϭ 3.4 Hz, 2 H, H4D, H5D), 3.57 (m, 1 H, H5), 3.58
(m, 1 H, H4), 3.65 (s, 3 H, OCH3), 3.66 (s, 3 H, OCH3), 3.68 (m, 1
H, H6b), 3.74 (q, J ϭ 3 Hz, 1 H, H1D), 3.77 (q, J ϭ 3.4 Hz, 1 H,
H3), 3.79 (m, 1 H, H6a), 3.94 (dd, J ϭ 3.4, J ϭ 1.8 Hz, 1 H, H2),
3.97 (q, J ϭ 3 Hz, 1 H, H2D), 4.03 (m, 1 H, CH2Oallyl), 4.09 (m,
1 H, CH2Oallyl), 4.95 (d, J ϭ 1.8 Hz, 1 H, H1), 5.20 (m, 1 H,
CH2ϭ), 5.29 (m, 1 H, CH2ϭ), 591 (m, 1 H, CHϭ) ppm.13C NMR
(125 MHz, D2O, 25 °C): δ ϭ 27, 39.2, 52.6, 61.05, 66.9, 70, 70.5,
70.6, 70.7, 73.5, 73.7, 98.7, 118.3, 134.2 ppm. MS (FABϩ): m/z ϭ
457 [M ϩ Naϩ]. HRMS (ESIϩ) for C19H30O11Naϩ: calcd.
457.16803; found 457.16586.
4.6 Hz, 1 H), 3.28 (s, 3 H), 3.30 (s, 3 H) ppm.
The catalyst Cu(OTf)2 (0.07 mmol) was added to a solution of the
epoxide (150 mg, 0.7 mmol) in allyl alcohol (1.5 mL, 2.2 mmol).
The reaction mixture was stirred for 3 h at room temperature,
whilst monitoring by TLC (hexane/EtOAc, 6:4). After completion
of the reaction a saturated solution of NH4Cl and NH3 was added
and the mixture was extracted with EtOAc. The organic layer was
dried and the solvents evaporated under reduced pressure. The
crude mixture was purified by flash chromatography on silica gel
1
(hexane/EtOAc, 1:1) to give 5 in 90% yield. H NMR (200 MHz,
C6D6, 25 °C): δ ϭ 2.05 (m, 5 H), 2.52 (br. s, 1 H, exch), 3.40 (m,
9 H), 3.70Ϫ4.00 (m, 4 H), 5.15 (m, 2 H), 5.80 (m, 1 H) ppm. 13C
NMR (75 MHz, CDCl3, 25 °C): δ ϭ 27.0, 30.6, 39.0, 39.4, 51.9,
67.3, 69.8, 75.9, 116.9, 134.8, 175.0 ppm. MS (FABϩ): m/z ϭ 273
[M ϩ Hϩ].
˚
Mannosylation of 5. Synthesis of 8: Ground 4-A molecular sieves
were added to a solution of 5 (493 mg, 1.81 mmol) and 6[13] (1.33 g,
2.71 mmol) and the mixture was dried in vacuo overnight. Dry
CH2Cl2 (50 mL) was then added under N2 and the solution was
stirred for about 15 min. The temperature was then adjusted to
Ϫ20 °C and TMSOTf (63 µL, 0.724 mmol) was added. The solu-
tion was stirred at Ϫ20 °C for about 10 min, until TLC revealed Enzymatic Studies: Mannosidase stability assays were performed in
the disappearance of 5 (hexane/EtOAc, 6:4). The reaction was neu-
tralized by adding triethylamine and concentrated. The molecular
sieves were then removed by filtration and the product was isolated
by flash chromatography (hexane/EtOAc, 6:4), and purified a se-
cond time by chromatography (CHCl3/acetone, 95:5), to give the
20 µL samples of 5 m substrate (3 or 1) in pH 4.5 0.1  phos-
phate-citrate buffer at 37 °C. Substrates were submitted to increas-
ing concentrations of jack-bean mannosidase. Three enzymatic re-
actions were run, with 0.5, 2.5 and 10 µg protein/mL solutions
prepared from stock commercial enzyme suspension (Sigma-
M7257, 5 mg/mL). After 30 min incubation the crude mixtures
1
orthoester 7 in 86% yield. [α]D ϭ ϩ47 (c ϭ 1, CHCl3). H NMR
(400 MHz, CDCl3, 25 °C): δ ϭ 1.75 (s, 3 H, CH3), 1.8Ϫ1.99 (m, 4 were frozen and freeze-dried. In all cases lyophilized samples were
H, H3Dax,eq), 2.2Ϫ2.15 (m, H9, COCH3), 2.95Ϫ3 (m, 2 H, H4D
,
completely trimethylsilylated by treatment with trimethylsilylimi-
H5D), 3.51 (d, 1 H, H1D), 3.69 (s, 3 H, OCH3), 3.7 (s, 3 H, OCH3), dazole and pyridine at 60 °C for 30 min, and were then analyzed
3.71 (m, 1 H, H5D), 3.89 (m, 1 H, H2D), 3.92 (dd, J ϭ 12, J ϭ 5 by gas chromatography. GC analysis was performed with a
Hz, 1 H, H7a), 4.09 (dd, J ϭ 12, J ϭ 6 Hz, 1 H, H7b), 4.11 (dd,
PerkinϪElmer Autosystem. A fused silica column was employed
J ϭ 12, J ϭ 5 Hz, 1 H, H6a), 4.28 (dd, J ϭ 12, J ϭ 2.5 Hz, 1 H, (dimension 5 m ϫ 0.25 mm ϫ 0.22 µm) with SPB-1. Carrier gas:
H6b), 4.6 (m, 1 H, H2), 5.12Ϫ5.31 (m, 2 H, H9), 5.19 (s, 1 H, H3), He flow at 20 psi. Volume injected: 1 µL. Program temperature
5.25 (s, 1 H, H4), 5.45 (d, J ϭ 2 Hz, 1 H, H1), 5.8Ϫ6.0 (m, 1 H, used: 80 °C for 2.5 min, 10 °C/min to 150 °C and 15 °C/min to
H8) ppm. 13C-HETCOR (100.6 MHz, CDCl3, 25 °C): δ ϭ 34, 35, 250 °C. Injector temperature: 280 °C; FID temperature: 280 °C.
40, 62.5, 66, 69, 70.5, 71, 73, 75, 77.5, 98, 117, 135 ppm. MS
Retention times: 3 13.5 min, 5 7.1 min,mannose 8.1 min, manno-
bioside 1 13.1 min.
(FABϩ): m/z ϭ 603 [M ϩ Hϩ].
Rearrangement of the orthoester was achieved by adding TMSOTf
(0.65 µL, 0.0035 mmol), to a cold (Ϫ20 °C) solution of the orthoes-
ter 7 (21.3 mg, 0.035 mmol) in CH2Cl2 (600 µL). The mixture was
stirred under nitrogen whilst monitoring by TLC (CHCl3/acetone,
Determination of the KM (4.67 m) and vMAX (0.13 m) param-
eters was performed in 50 µL reaction volume with 0.25 µg/mL
enzyme concentration with p-nitrophenol-α--mannoside (PNP-
Man) as substrate. Seven different substrate concentrations were
95:5). After 10 min the reaction was complete. The mixture was used (from 0.3 m to 15 m), with an incubation time of 20 min.
then treated with triethylamine and the solvents evaporated under
reduced pressure. The residue was purified by flash chromatogra-
Quenching was performed by adding 300 µL of 1  carbonate
buffer at pH 12. The absorbance of the p-nitrophenoxide released
phy on silica gel (CHCl3/acetone, 97:3) to give 8 in 50% yield. The during the enzymatic reactions was measured at 400 nm and 25 °C
same product was also obtained in 65% yield by running the man- on a UV/Vis Spectrometer (PerkinϪElmer Lambda 6).
nosylation of 5 with 6 at Ϫ20 °C for 20 min, and monitoring the Mannosidase inhibition studies were performed using a fixed
conversion of 7 by TLC (CHCl3/acetone, 95:5). [α]D ϭ ϩ38 (c ϭ
1.8 m substrate (PNP-Man) concentration and variable inhibitor
(3) concentrations in a reaction volume of 20 µL (Table 2). Enzy-
1.26, CHCl3). 1H NMR (400 MHz, C6D6, 25 °C): δ ϭ 1.74 (s, 3 H,
COCH3), 1.77 (s, 3 H, COCH3), 1.81 (s, 3 H, COCH3), 1.85 (s, 3 matic reactions were run with 2.5 µg/mL enzyme concentration in
5124  2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2004, 5119Ϫ5125

A Mimic of 1,2-Mannobioside
FULL PAPER
[8]
A. Bernardi, D. Arosio, L. Manzoni, D. Monti, H. Posteri, D.
Potenza, S. Mari, J. Jimenez-Barbero, Org. Biomol. Chem.
2003, 1, 785Ϫ792.
pH 4.5 0.1  phosphate-citrate buffer at 25 °C. After 20 min the
reactions were quenched by addition of 130 µL of 1  carbonate
solution (pH 12) and the absorbance was measured at 400 nm and
25 °C on a UV/Vis Spectrometer.
[9] [9a]
E. A. Merrit, S. Sarfaty, F. v. d. Akker, C. LЈHoir, J. A.
Martial, W. G. J. Hol, Protein Sci. 1994, 3, 166Ϫ175. [9b] E. A.
Merrit, S. Sarfaty, M. G. Jobling, T. Chang, R. K. Holmes, T.
Computational Methods: All calculations were performed using the
MacroModel/Batchmin[21] package (version 5.5) and the AMBER*
force field with the SenderowitzϪStill all-atom pyranose param-
eters.[22] Charges were taken from the force field (all-atom charge
option). Water solvation was simulated using MacroModel’s gen-
eralized Born GB/SA continuum solvent model.[23] The confor-
mational search for 3 was carried out using 10000 steps of the us-
age-directed MC/EM procedure. The glycosidic linkages and all the
extra-annular CϪC bonds were used as explicit variables during
the Monte Carlo search. Extended non-bonding cut-off distances
[9c]
R. Hirst, W. G. J. Hol, Protein Sci. 1997, 6, 1516Ϫ1528.
E.
A. Merrit, W. G. J. Hol, Curr. Opin. Struct. Biol. 1995, 5,
165Ϫ171 and references cited therein.
[10] [10a] D. Arosio, S. Baretti, S. Cattaldo, D. Potenza, A. Bernardi,
[10b]
Bioorg. Med. Chem. Lett. 2003, 13, 3831Ϫ3834.
A.
Bernardi, D. Arosio, D. Potenza, I. Sanchez-Medina, S. Mari,
F. J. Can˜ada, J. Jimenez-Barbero, Chem. Eur. J. 2004, 10,
4395Ϫ4406.
V. V. Samoshin, V. A. Chertkov, L. P. Vatlina, E. K. Dobret-
sova, N. A. Simonov, L. P. Kastorsky, D. E. Gremyachinsky,
H.-J. Schneider, Tetrahedron Lett. 1996, 37, 3981Ϫ3984.
A. Bernardi, D. Arosio, D. Dellavecchia, F. Micheli, Tetra-
hedron: Asymmetry 1999, 40, 3403Ϫ3407.
[11]
[12]
˚
(a van der Waals cut-off of 8.0 A and an electrostatic cut-off of
˚
20.0 A) were used both for the MC/EM and MC/SD calculations.
The same degrees of freedom of the MC/EM searches were used
in the MC/SD runs. All simulations were performed at 300 K, with
a dynamic time-step of 1.5 fs and a frictional coefficient of 0.1
psϪ1. One Monte Carlo step was performed after each dynamic
step. The average acceptance ratio was 6%. Two simulations of 5 ns
each were performed starting from the lowest energy E and S con-
formations. Structures were sampled every 2 ps and saved for later
evaluation. The interatomic distances reported in Table 1 under the
MC/SD header were evaluated from ϽrϪ6Ͼ, which was monitored
during the simulation (option MDDI of Batchmin).
[13] [13a]
[13b]
T. Ren, D. Liu, Tetrahedron Lett. 1999, 40, 7621Ϫ7625.
R. R. Schmidt, J. Michel, Angew. Chem. Int. Ed. Engl.
[13c]
1980, 19, 731Ϫ732.
R. R. Schmidt, Angew. Chem. Int. Ed.
Engl. 1986, 25, 212Ϫ235.
[14]
[15]
W. Wang, F. Kong, J. Org. Chem. 1998, 63, 5744Ϫ5745; W.
Wang, F. Kong, Angew. Chem. Int. Ed. 1999, 38, 1247Ϫ1250.
R. J. Woods, A. Pathiaseril, M. R. Wormald, C. J. Edge, R. A.
Dwek, Eur. J. Biochem. 1998, 259, 372Ϫ386, and references
cited therein.
[16]
[17]
´
A. Bernardi, A. Colombo, I. Sanchez-Medina, Carbohydr. Res.
2004, 339, 967Ϫ973, and references cited therein.
M. R. Wormald, A. J. Petrescu, Y.-L. Pao, A. Glithero, T. El-
liot, R. A. Dwek, Chem. Rev. 2002, 102, 371Ϫ386 and refer-
ences cited therein.
Acknowledgments
[18]
Y.-T. Li, J. Biol. Chem. 1967, 242, 5474Ϫ5480.
[19] [19a]
The project was supported by the Improving Human Potential Pro-
gram under contract HPRN-CT-2002Ϫ00173 (Glycidic Scaffolds
Network). Funding from the Ministry of Science and Technology
of Spain (BQU2003-3550-C03), MIUR, and CNR is also acknowl-
edged.
Y. Bourne, V. Zamboni, A. Barre, W. J. Peumans, E. J. M.
[19b]
Van Damme, P. Rouge, Structure 1999, 7, 1473Ϫ1482.
D.
N. Moothoo, B. Canan, R. A. Field, J. Naismith, Glycobiology
[19c]
1999, 9, 539Ϫ545.
D. A. R. Sanders, D. N. Moothoo, J.
Raftery, A. J. Howard, J. R. Helliwell, J. H. Naismith, J. Mol.
Biol. 2001, 310, 875Ϫ884. [19d] C. A. Bewley, Structure 2001, 9,
931Ϫ940; L. C. Chang, C. A. Bewley, J. Mol. Biol. 2002, 318,
[1]
[19e]
P. Sears, C. H. Wong, Angew. Chem. 1999, 111, 2446Ϫ2471;
1Ϫ8.
S. R. Shenoy, L. G. Barrientos, D. M. Ratner, B. R.
Angew. Chem. Int. Ed. 1999, 38, 2300Ϫ2324, and references
O’Keefe, P. H. Seeberger, A. M. Gronenborn, M. R. Boyd,
[19f]
cited therein.
Chem. Biol. 2002, 9, 1109Ϫ1118.
C. A. Bewley, S. Ki-
[2]
[19g]
R. Thomson, M. von Izstein, in Carbohydrate Based Drug Dis-
yonaka, I. Hamachi, J. Mol. Biol. 2002, 322, 881Ϫ889.
C.
covery (Ed.: C. H. Wong), 2004, vol. 2, pp. 831Ϫ861.
R. Horstkorte, O. T. Keppler, W. Reutter, in Carbohydrate
N. Scanlan, R. Pantophlet, M. R. Wormald, E. O. Saphire, R.
Stanfield, I. A. Wilson, H. Katinger, R. A. Dwek, P. M. Rudd,
D. R. Burton, J. Virol. 2002, 76, 7306Ϫ7321.
W. B. Turnbull, B. L. Precious, S. W. Homans, J. Am. Chem.
Soc. 2004, 126, 1047Ϫ1054.
[3]
Based Drug Discovery (Ed.: C. H. Wong), 2004, vol. 2, pp.
[20]
[21]
863Ϫ873.
[4]
A. Bernardi, D. Arosio, L. Manzoni, F. Micheli, A. Pasquar-
ello, P. Seneci, J. Org. Chem. 2001, 66, 6209Ϫ6216.
A. Bernardi, P. Brocca, A. Checchia, S. Sonnino, F. Zuccotto,
F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp,
M. Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still,
J. Comp. Chem. 1990, 11, 440Ϫ467.
[5]
J. Am. Chem. Soc. 1999, 121, 2032Ϫ2036.
[6]
[22]
[23]
A. Bernardi, D. Arosio, S. Sonnino, Neurochem. Res. 2002, 27,
H. Senderowitz, W. C. Still, J. Org. Chem. 1997, 62,
539Ϫ545, and references cited therein.
A. Bernardi, D. Potenza, A. M. Capelli, A. Garcıa-Herrero, F.
1427Ϫ1438.
[7]
´
W. C. Still, A. Tempzyk, R. Hawley, T. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127Ϫ6129.
´
J. Can˜ada, J. Jimenez-Barbero, Chem. Eur. J. 2002, 8,
4597Ϫ4612.
Received July 23, 2004
Eur. J. Org. Chem. 2004, 5119Ϫ5125
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5125