Sp2-iminosugar O-, S-, and N-glycosides as conformational mimics of α-linked disaccharides; Implications for glycosidase inhibition

Article abstract of DOI:10.1002/chem.201200279

The synthesis of mimics of the α(1→6)- and α(1→4)- linked disaccharides isomaltose and maltose featuring a bicyclic sp ²-iminosugar nonreducing moiety O-, S-, or N-linked to a glucopyranoside residue is reported. The strong generalized anomeric effect operating in sp²-iminosugars determines the α-stereochemical outcome of the glycosylation reactions, independent of the presence or not of participating protecting groups and of the nature of the heteroatom. It also imparts chemical stability to the resulting aminoacetal, aminothioacetal, or gem-diamine functionalities. All the three isomaltose mimics behave as potent and very selective inhibitors of isomaltase and maltase, two α-glucosidases that bind the parent disaccharides either as substrate or inhibitor. In contrast, large differences in the inhibitory properties were observed among the maltose mimics, with the O-linked derivative being a more potent inhibitor than the N-linked analogue; the S-linked pseudodisaccharide did not inhibit either of the two target enzymes. A comparative conformational analysis based on NMR and molecular modelling revealed remarkable differences in the flexibility about the glycosidic linkage as a function of the nature of the linking atom in this series. Thus, the N-pseudodisaccharide is more rigid than the O-linked derivative, which exhibits conformational properties very similar to those of the natural maltose. The analogous pseudothiomaltoside is much more flexible than the N- or O-linked derivatives, and can access a broader area of the conformational space, which probably implies a strong entropic penalty upon binding to the enzymes. Together, the present results illustrate the importance of taking conformational aspects into consideration in the design of functional oligosaccharide mimetics. Flexibility is key! Mimicking the conformational behavior of oligosaccharides is critical to the design of selective inhibitors of glycosidases. Matching the structural features of the glycone and aglycone moieties is not sufficient if the flexibility about the intersaccharide linkage does not satisfy the natural partner, even if there is extensive overlap of the respective available conformational spaces (see scheme). Copyright

Full text of DOI:10.1002/chem.201200279

FULL PAPER
DOI: 10.1002/chem.201200279
sp2-Iminosugar O-, S-, and N-Glycosides as Conformational Mimics of
a-Linked Disaccharides; Implications for Glycosidase Inhibition
Elena M. Sꢀnchez-Fernꢀndez,[a] Rocꢁo Rꢁsquez-Cuadro,[b] Carmen Ortiz Mellet,*[b]
Josꢂ M. Garcꢁa Fernꢀndez,[a] Pedro M. Nieto,[a] and Jesffls Angulo*[a]
Abstract: The synthesis of mimics of
the a (1!4)-linked disac-
(1!6)- and a
maltase, two a-glucosidases that bind
the parent disaccharides either as sub-
strate or inhibitor. In contrast, large
differences in the inhibitory properties
were observed among the maltose
mimics, with the O-linked derivative
being a more potent inhibitor than the
N-linked analogue; the S-linked pseu-
dodisaccharide did not inhibit either of
the two target enzymes. A comparative
conformational analysis based on
NMR and molecular modelling re-
vealed remarkable differences in the
flexibility about the glycosidic linkage
as a function of the nature of the link-
ing atom in this series. Thus, the N-
pseudodisaccharide is more rigid than
the O-linked derivative, which exhibits
conformational properties very similar
to those of the natural maltose. The
G
ACHTUNGTRENNUNG
charides isomaltose and maltose featur-
ing a bicyclic sp2-iminosugar nonreduc-
ing moiety O-, S-, or N-linked to a glu-
copyranoside residue is reported. The
strong generalized anomeric effect op-
erating in sp2-iminosugars determines
the a-stereochemical outcome of the
glycosylation reactions, independent of
the presence or not of participating
protecting groups and of the nature of
the heteroatom. It also imparts chemi-
cal stability to the resulting aminoace-
tal, aminothioacetal, or gem-diamine
functionalities. All the three isomaltose
mimics behave as potent and very se-
lective inhibitors of isomaltase and
analogous
pseudothiomaltoside
is
much more flexible than the N- or O-
linked derivatives, and can access
a broader area of the conformational
space, which probably implies a strong
entropic penalty upon binding to the
enzymes. Together, the present results
illustrate the importance of taking con-
formational aspects into consideration
in the design of functional oligosac-
charide mimetics.
Keywords: carbohydrates · confor-
mation analysis · inhibitors · molec-
ular modeling · structural biology
Introduction
events. Consequently, inhibitors of these enzymes possess
strong therapeutic potential for the treatment of ailments as
varied as diabetes,[1] lysosomal storage disorders,[2] viral in-
fection,[3] and cancer.[4] Critical to these channels is the de-
velopment of compounds that can discriminate between gly-
cosidases acting on different substrates, in different tissues
or in different cell compartments. For instance, the lack of
sufficient anomeric selectivity towards a- or b-glucosidases
is largely responsible for the failure of classical iminosugar-
type inhibitors, such as 1-deoxynojirimycin (1) or castano-
spermine (2), in clinical trials (Figure 1).[5]
Glycosidases, enzymes that catalyze the cleavage of glyco-
sidic bonds in oligosaccharides and glycoconjugates, are fun-
damental to a broad range of biological processes including
the degradation of dietary polysaccharides, the lysosomal ca-
tabolism of glycoconjugates, and the biosynthesis of the oli-
gosaccharide units in glycoproteins and glycolipids, which in
their turn are involved in a plethora of cell communication
[a] Dr. E. M. Sꢀnchez-Fernꢀndez, Prof. J. M. Garcꢁa Fernꢀndez,
Dr. P. M. Nieto, Dr. J. Angulo
Instituto de Investigaciones Quꢁmicas (IIQ)
CSIC–Universidad de Sevilla
In principle, one could expect that higher selectivities
within isoenzyme series could be achieved by mimicking not
Avda. Americo Vespucio 49
41092, Sevilla (Spain)
Fax : (+34)954460565
E-mail: jesus@iiq.csic.es
[b] R. Rꢁsquez-Cuadro, Prof. C. Ortiz Mellet
Departmento de Quꢁmica Orgꢀnica
Facultad de Quꢁmica
Universidad de Sevilla
C/Prof. Garcꢁa Gonzꢀlez 1
41012, Sevilla (Spain)
Fax : (+34)954624960
E-mail: mellet@us.es
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200279.
Figure 1. Structures of glycomimetic representatives either without (1 or
2) or with (3–8) pseudoaglycone substituents.
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8527

only the glycone but also the anomeric stereochemistry and
the aglycone portion of the natural substrate.[6] If properly
arranged, once the glycone moiety is accommodated in the
catalytic (ꢀ1) site of the enzyme, an anomeric substituent
may bind to the adjacent aglycone (+1) subsite, which, de-
pending on the enzyme, can be adapted to the binding of
sugar, acyl group, or other types of moieties. The choice of
linkage type to span the glycone and aglycone binding sites
is not trivial. Replacing the glycosidic oxygen in natural gly-
cosides by other atoms, such as sulfur (3) or carbon (4), has
been shown to lead to compounds that are recognized by
glycosidases, acting as moderate to good inhibitors
(Figure 1).[7] Yet, differences in the conformation of these
molecules with respect to the natural compounds have been
reported, the origin of which has been the subject of some
controversy.[8] Moreover, such chemical modifications can
alter the stereoelectronic barriers for rotation about the gly-
cosidic linkage. 3D shapes compatible with enzyme binding
can then be obtained by conformational variations that do
not imply chair distortion, leading to differences in the pre-
ferred bound conformations.[9] Nitrogen-linked analogues
(5–7) would be particularly interesting because anomeric
amino groups can provide additional interactions with the
catalytic carboxylic groups,[10] but the instability and lack of
configurational integrity of aminoacetal (5)[11] and amino-
thioacetal functionalities (6)[12] limits this approach to carba-
sugar-type (7)[13] or nonglycosidically-linked glycomimet-
ics.[14] The use of iminosugar-type glycone moieties faces
similar instability problems for the elaboration of O- or S-
glycosides and is only compatible with the C-glycoside pro-
totype (8),[15] which implies a relatively complex chemistry
and sacrifices the contribution of the exo-anomeric effect to
the conformational equilibrium (Figure 1).[16]
Figure 2. General structure of bicyclic carbamate-type sp2-iminosugars
showing the key orbitals involved in the anomeric effect (9) and struc-
tures of the isomaltose (10–12) and maltose mimics (13–15) prepared in
this work. The numbering of the sp2-iminosugar moiety follows the num-
bering system used for the indolizidine system (e.g., as for castanosper-
mine 2).
through oxygen, sulfur, or nitrogen heteroatoms. This offers
a unique opportunity to conduct a systematic study on the
consequences of differences in the conformational behavior
within homologous series of compounds in their enzyme in-
hibitory properties, which would be of great relevance for
the proper design of second generation molecules with
better properties. To test this hypothesis, we have focused
on two closely related a-glucosidases from S. cerevisae,
oligo-a-1,6-glucosidase (EC 3.2.1.10, isomaltase) and a-1,4-
glucosidase (EC 3.2.1.20, maltase). Both enzymes belong to
family GH 13 in the CAZy classification.[21] Maltase prefer-
entially hydrolyses maltose, but not isomaltose (which acts
act as competitive inhibitor),[22] whereas isomaltase hydro-
lyzes isomaltose but not maltose (which acts as a competitive
inhibitor).[23] In a preliminary report, we noted that the gem-
diamine-type isomaltose and maltose mimics 12 and 15 were
actually stable compounds with the ability to bind isomal-
tase.[24] Herein, we compare our previous results with those
for the corresponding O-linked (10, 13) and S-linked disac-
charide mimics (11 and 14, respectively) and analyse their
inhibitory properties towards isomaltase and maltase
(Figure 2). We show that the capabilities of binding to each
enzyme are indeed strongly dependent on the chemical
nature of the compound and can be correlated to differences
in their conformational behavior in solution.
In previous work, we showed that replacing the sp3-amine
nitrogen typical of iminosugars by a pseudoamide-type ni-
trogen, with substantial sp2-character (sp2-iminosugars), re-
sults in inhibitors for which the a- or b-glycosidase anomeric
selectivity can be modulated through nonglycone interac-
tions upon the installation of appropriate substituents at
specific locations.[17,18] Interestingly, 1-deoxy-2-oxa-3-oxocas-
tanospermine derivatives with alkyl O-, S-, or N-pseudoano-
meric groups (9) acted as specific inhibitors of neutral a-glu-
cosidases, exhibiting antiproliferative activity against human
breast carcinoma cells in vitro.[19] In those compounds, the
glycone specificity is presumed to be conferred by the struc-
tural similarity of the polyhydroxylated core with d-gluco-
pyranose. The very intense anomeric effect in sp2-imino-
sugars, which is ascribable to a very efficient overlap be-
tween the p-symmetrical orbital located on the lone-pair of
the endocyclic nitrogen and the s* antibonding orbital of
Results and Discussion
ꢀ
the anomeric C X bond, imparts chemical, conformational
Synthesis: Tri-O-acetyl-protected derivatives of the 2-oxa-3-
oxo-castanospermine skeleton represented by the general
formula 9 (Figure 2) but bearing a 5-O-trichloroacetimidate
(X=O; R=C(=N)CCl3)[25] or a 5-bromo group (XR=Br)
were initially considered as suitable sp2-iminosugar pseudo-
glycosyl donors for the construction of the isomaltoside and
maltoside mimics 10 and 13. However, reactions with the
and configurational stability to the axially oriented pseudo-
aglycone substituent, and is probably responsible for the re-
markable a-selectivity (Figure 2).[20] The bicyclic piperidine-
carbamate core appears then to be an ideal scaffold for the
design of linkage-spanning a-glucosidase inhibitors in which
the glycone and aglycone portions can be brought together
8528
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
corresponding glycosyl acceptors using a variety of glycosyl-
ation promoters and conditions did not proceed to comple-
tion, resulting in extensive formation of the product of hy-
drolysis (9; XR=OH) after conventional work up. Alterna-
tively, use of the corresponding pseudoglycosyl fluoride ana-
logue 17 as precursor was attempted (Scheme 1). Compound
Scheme 2. Synthesis of the S-linked isomaltoside and maltoside mimics
11 and 14.
arrangement[31] to give d-fructose derivatives was observed
in some cases, which was found to be favored for vic-hy-
droxylamines due to their ability to undergo concomitant
oxidation-intramolecular glycosylation.[24] To limit this side-
reaction, coupling of 26 with methyl 6- and 4-aminodeoxy-
a-d-glucopyranosides (27[32] and 28[33]) in methanol was con-
ducted under a nitrogen atmosphere (Scheme 3). The a-N-
Scheme 1. Synthesis of the O-linked isomaltoside and maltoside mimics
10 and 13.
17 was obtained in 60% yield as a stable solid by reaction
of tetracaetate 16 with pyridinium poly(hydrogen fluo-
ride).[26] In the presence of boron trifluoride-diethyl ether
complex (BF3·Et2O), 16 smoothly reacted with methyl 2,3,4-
tri-O-acetyl-a-d-glucopyranoside[27] (18) or methyl 2,3,6-tri-
O-acetyl-a-d-glucopyranoside[28] (20) in dichloromethane at
08C to give the corresponding per-O-acetylated a-(5!6’)
and a-(5!4’) pseudodisaccharides 19 and 21 in 64 and 65%
yield, respectively. Final deacetylation under standard
sodium methoxide-catalyzed conditions provided the re-
quired fully unprotected compounds 10 and 13 (Scheme 1).
sp2-Iminosugar thioglycoside derivatives have previously
been obtained from tetraacetate 16 by reaction with thiols
in the presence of BF3·Et2O.[24] However, in the case of the
6- and 4-thiousugars 22[29] and 24,[30] this procedure proved
unsuccessful. The pseudo-S-disaccharides 23 and 25 could be
obtained in satisfactory yield by using, instead, fluorocom-
pound 17 as donor (Scheme 2). Minor proportions of mono-
deacetylated products, probably arising from sulfur-assisted
regioselective deacetylation at position O-6, were detected
in the crude reaction mixtures by mass spectrometry and
1H NMR analyses. Reacetylation, column chromatographic
purification, and conventional base-catalyzed removal of the
ester protecting groups, afforded the target fully unprotected
thioodisaccharide mimics 11 and 14 (Scheme 2).
Scheme 3. Synthesis of the N-linked isomaltoside and maltoside mimics
12 and 15.
(5!6’)-linked pseudoisomaltoside 12 was thus obtained in
59%, with 25% unreacted 26 recovered. In the case of the
a N-(5!4’) maltoside mimic 15, purification required acety-
lation of the reaction mixture prior to column chromatogra-
phy and final deacetylation (Scheme 3).
The a-stereochemical outcome of the above reactions,
confirmed by the proton-proton coupling constant values
about the piperidine ring (see Experimental section) is re-
markable. Glycosidation and thioglycosidation of glycosyl
donors bearing participating groups at C-2 (C-6 in the indol-
izidine numbering system for 17) are expected to proceed
with anchimeric assistance to give preferentially the diaste-
The pseudo glucosyl fluoride 17 proved ineffective as an
sp2-iminosugar donor for the preparation of N-linked iso-
maltoside or maltoside analogues. Reaction of the unpro-
tected bicyclic carbamate 26 with ammonia or amines in
methanol has been previously shown to afford the corre-
sponding pseudoglycosylamines.[19,24] However, Amadori re-
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8529

C. Ortiz Mellet, J. Angulo et al.
reomer with a 1,2-trans relative disposition, that is, the b-
anomer for d-glucopyranosyl donors.[7a] In the case of the
sp2-iminosugar 17, however, departure of the leaving fluoro
group is probably assisted by the lone-pair of the carbamate
nitrogen to give a transient azacarbenium cation species (29;
Scheme 4). Although this intermediate might be in equilibri-
poorer inhibitor and the S-linked derivative did not inhibit
either of the two enzymes at concentrations up to 2 mm.
To ascertain whether the observed differences in the in-
hibitory properties could be correlated to differences in the
conformational behavior of the different pseudodisaccha-
rides, an NMR and computational analysis was conducted.
NMR studies: The conformation of the piperidine ring of
the sp2-iminosugar moiety in compounds 10–15 is fixed in
8
the C5 chair, whereas the pyranose ring in the substituted
4
glucopyranoside moiety adopts the equivalent C1 conforma-
tion, as inferred from the analysis of the corresponding ex-
perimental vicinal proton-proton coupling constants. There-
fore, the global geometry of these pseudodisaccharides is
fundamentally defined by the corresponding dihedral angles
Scheme 4. Generalized anomeric effect controlled (thio)glycosidation of
17.
about their pseudoglycosidic linkages, that is, a
ACHTUNGTRENNUNG
10–12 and a
AHCTUNGTRENNUNG
um with the tricyclic orthoester oxocarbenium involving the
vicinal acetoxy group (30), subsequent addition of the alco-
hol or thiol nucleophile is under strict control of the ano-
typical saccharide structures, we adopt the nomenclature
generally used for the equivalent angles in glycosides,[36]
namely f (H5-C5-X-Y’; where X=O, N, or S and Y’=C4ꢃ
or C6’) and y (C5-X-Y’-Z’; where Z’=H4’ or C5’). For the
isomaltoside mimics 10–12, the global geometry is character-
ized not only by the values of the f and y torsion angles,
meric effect, affording the a-linked pseudoACTHNUTRGNEUG(N thio)-
disaccharide as the only reaction product (Scheme 4). The
result is even more extraordinary in the case of the N-linked
pseudodisaccharides. gem-Diamines typically suffer from in-
stability under physiologically conditions[34] and the prepara-
tion of a-configured glycosylamines is particularly problem-
atic due to the overwhelming thermodynamic preference for
the b-anomer as determined by the reverse anomeric
effect.[35] Together, these results strongly support the preva-
lence of the anomeric effect as the main driving force in re-
actions involving the pseudoanomeric center in sp2-imino-
sugars.
ꢀ
but also by the rotamers around the C5’ C6’ bond, defined
by the w dihedral angle O6’-C6’-C5’-O5’.
By analogy with the parent disaccharides, three preferred
regions on the f–y potential energy surfaces of these com-
pounds would be expected, namely synf–syny, synf–antiy,
and antif–syny, the latter being less likely to be populated
in solution due to the exo-anomeric effect. Concerning the
w dihedral angle, three canonical dispositions, corresponding
to the staggered conformations gt, gg, and tg, must be con-
sidered. These regions could be unequivocally characterized
experimentally by the detection of exclusive NOE correla-
tions.
Glycosidase inhibitory activity evaluation: Compounds 10–
12 and 13–15 represent examples of pseudodisaccharides
with a hydroxylation profile and substitution pattern of ste-
reochemical complementarity with isomaltose and maltose,
respectively, differing exclusively in the nature of the inter-
glycosidic heteroatom. They can, in principle, mimic not
only the developing glucopyranosyl cation in the reaction
coordinate of enzymatic hydrolysis of a-glucosidases at the
glycone (ꢀ1) site but also the aglycone moiety at the corre-
sponding aglycone (+1) site in the cases of maltase and iso-
maltase, both enzymes being able to bind to either of the
parent disaccharides. Actually, the three isomaltose ana-
logues behaved as very selective inhibitors of isomaltase and
maltase. No inhibition of b-glucosidases or enzymes acting
on other sugar configurations was observed, and only the N-
In the case of the N-linked isomaltoside mimic 12, the
pKa of the intersaccharide amino group was found to be
1
around 12 from a H NMR pH titration experiment, indicat-
ing that at neutral pH this group is fully protonated and,
therefore, positively charged (see the Supporting Informa-
tion). This high basicity is probably a consequence of the
strong contribution of orbital interactions to the anomeric
effect, which implies an increase of electron density at the
glycosidic bond. To determine the populations of conform-
ers around the w torsion angle, the vicinal J
constants were analyzed, giving values of 2.8 Hz for H6’-
(proS) and 7.8 Hz for H6’(proR). A nonlinear least squares
ACHTUNGTREN(NGNU 5’,6’) coupling
A
ACHTUNGTRENNUNG
fitting of these experimental values with those theoretically
predicted for the three canonical conformers gt, gg, and tg
led to a distribution of populations of 80% gt and 20% gg,
with no participation of the tg rotamer. The observed NOE
contacts involving the methylene protons were in very good
agreement with the populations determined from the cou-
pling constant values. Thus, H5 at the iminosugar ring
linked derivative exhibited
a weak inhibition constant
against trehalase, an enzyme acting on the a,a’(1!1)-linked
disaccharide a,a’-trehalose. In all cases, the inhibitory poten-
cy against maltase was higher than against isomaltase, but
no strong differences between the O-, S-, and N-linked de-
rivatives were noted. The scenario totally changed in the
case of a(1!4)-linked derivatives. Whereas O-linked pseu-
domaltoside was a potent and totally selective inhibitor of
isomaltase and maltase, the N-linked analogue was a much
showed a NOE with H6’
H5/H6’(proR) NOE. On the other hand, the H4’/H6’
NOE was stronger than the H4’/H6’(proS) NOE. Interest-
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
8530
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
ingly, a significant NOE cross peak between protons H5 and
H5’ was observed, which reported on a substantial contribu-
tion of conformers with y torsion around 908, indicating
that the gt conformer must exhibit considerable flexibility
around this dihedral angle, oscillating between values of
1808 and 908 (Figure 3, left). It was also possible to identify
Figure 3. Schematic representation of the gt (left) and gg (right) confor-
mations for isomaltoside mimics. Only protons involved in diagnostic
NOE contacts are depicted. Exclusive NOEs for each conformation, as
discussed in the text, are highlighted in bold.
a weak long-range NOE contact correlating to H8a at the
sp2-iminosugar moiety and H4’ (Figure 3, right, and Fig-
ure 4a), which is only possible for the minor gg conformer.
This conclusion was further confirmed by 1D NOESY ex-
periments with selection of the resonance of H4’ (see the
Supporting Information).
1
Proton H5’ and both H6’ protons in the H NMR spec-
trum of the a
nous, precluding the measurement of the corresponding vici-
nal J(5’,6’) coupling constants. Consequently, in this case,
(5!6’) O-linked analogue 10 were isochro-
ACHTUNGTRENNUNG
the description of the conformational equilibrium relied ex-
clusively upon the observation of key NOE cross peaks in
the NOESY experiments. A very strong NOE was observed
between H5 and H6’ protons, but it was not possible to dis-
criminate in terms of intensities between both geminal pro-
tons of the hydroxymethyl group. Neither was it possible to
discriminate an observable NOE between H8a at the imino-
sugar moiety and H5’ or either H6’ protons at the glucopyr-
anosyl ring, which prevented experimental confirmation of
the y 1808 conformer. On the other hand, two clear NOE
contacts were observed between H8a and H4’, as well as be-
tween H7 and H4’, which corroborated the contribution of
Figure 4. NOE experiments for a) 12, b) 10, and c) 11. 2D NOESY ex-
periments are shown for 12 and 10 (500 MHz, 400 ms mixing time; at 25
and 358C, respectively). In the case of 11, which exhibited a much better
spectral dispersion, 1D NOESY experiments are shown (500 MHz, 358C,
2 s mixing time), with selection of H5 (bottom row) or H4’ (top row) pro-
tons. Key NOESY cross peaks are highlighted on each spectrum. See the
Supporting Information for a color version of this figure.
other protons in the aglycone glucose moiety, particularly
H4’, H3’, and H5’. All the compounds showed very strong
H5–H4’ NOE contacts, which is indicative of major contri-
butions of synf–syny conformations in solution (Figure 5a
and Figure 6). For 13 and 15, very weak NOE contacts cor-
relating H5 with H3’ and H5’ could be observed (Figure 5b
and Figure 6b), however, they were only observable at large
NOESY mixing times (800 ms), so spin diffusion through
H4’ and H3’ could not be safely ruled out. These weak NOE
contacts would be indicative of a certain degree of flexibility
around the synf–syny region of the energy map (see
below).
ꢀ
the gg conformer around the C5’ C6’ linkage.
In the case of the S-linked isomaltoside 11, the H6’
N
proton showed a significantly stronger NOE with H4’ than
with H6’(proS), indicating a situation similar to that de-
ACHTUNGTRENNUNG
scribed above for 12, corroborating the prevalence of the gt
rotamer at the w torsion. The signal dispersion on the spec-
trum in this case was much higher than for the O-linked de-
rivative 10 and allowed the detection of a significant NOE
contact between H5 at the piperidine ring and H5’ at the
glucopyranose aglycone, revealing the presence of conform-
ers with values around 908 for the y torsion.
The distributions of conformers around the a
(5!4’) pseu-
In the case of the S-linked pseudomaltoside 14, the
NOESY spectra showed a NOE correlation between proton
H8a in the sp2-iminosugar ring and proton H3’ in the agly-
cone glucose residue (Figure 6c). The similarity of molecu-
doglycosidic linkages in the maltoside mimics 13–15 can be
fundamentally described by the observation of key NOE
contacts between proton H5 on the iminosugar ring and
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8531

C. Ortiz Mellet, J. Angulo et al.
Figure 5. Schematic representation of the major conformers identified by
NOESY spectra for the maltoside mimics. Structure c), detected only in
the case of the S-linked derivative 14, represents a synf–antiy conforma-
tion. Protons involved in diagnostic NOE contacts are depicted. Key
NOEs for each conformation, as discussed in the text, are highlighted in
bold.
lar sizes and chemical structure of all the three a
N
linked pseudodisaccharides precluded any possible spin dif-
fusion effects present only in the case of 14. On the other
hand, the S atom has the largest covalent radius among N,
O, and S, which leads to larger inter-residual distances and,
hence, lower inter-residual NOEs, reducing the probability
of spin diffusion for 14, in comparison with 13 or 15. Most
probably, the observed H8a/H3’ inter-residual NOE is re-
porting on particular conformational characteristics around
the S-pseudoglycosidic linkage (Figure 5c).
The remaining NOE cross peaks observed in the NOESY
spectra of 13–15 report trivial intraresidual spatial contacts
for the two six-membered rings. On the other hand, vicinal
JACHTUNGTRENNUNG(5’,6’) coupling constants could be estimated from the 1D
1H NMR spectra. They define the conformations of the exo-
cyclic hydroxymethyl groups at the glucopyranose rings of
the pseudodisaccharides. The experimental values were
around 2.0–2.5 and 4.0–5.0 Hz, in agreement with a typical
equilibrium distribution of rotamers around the w torsion
angle of about 70:30 gg/gt for d-glucopyranose deriva-
tives,[37] without any clear contribution of the tg conformer.
This result is particularly relevant in the case of 15 because
it indicates that protonation of the amino group at the pseu-
doglycosidic linkage, which is expected to occur at physio-
logical pH, does not induce any perceivable conformational
shift of the hydroxymethyl group of the aglycone residue.
Figure 6. NOESY spectra of a) 15, b) 13, and c) 14 (500 MHz, 258C,
800 ms mixing time). Key NOESY cross peaks are highlighted on each
spectrum. Only 14 showed a NOE cross peak between proton H8a at the
sp2-iminosugar moiety and H3’ of the glucopyranose ring. Circles on
a) and b) highlight the absence of the H8a/H3’ cross peaks in the
NOESY spectra of 15 and 13. See the Supporting Information for a color
version of this figure.
908. Although all the identified minima in the adiabatic map
of 12 were gt conformers around w, the observation of
a weak H8a–H4’ NOE (Figure 4a) was indicative of the
presence of the gg conformer in solution. This discrepancy
can be attributed to deficiencies inherent to the force field
approach, which seems to be able to identify only the most
stable conformer around a very flexible linkage, not repro-
ducing well the less populated conformational states.
The strong intensity of the H5/H6’ NOE in the NOESY
spectrum of the O-linked pseudoisomaltoside 10 (no stereo-
specific distinction was possible in this case) indicated
a major contribution from the y 1808 conformer. Interest-
ingly, in this case, the molecular modelling protocol led to
the prediction of energetically favorable gg conformers be-
sides the gt conformers (Figure 7c), which was in very good
Molecular modeling: The energy minima obtained for the a-
(5!6’)-linked pseudodisaccharides 10–12 are represented as
stick models, along with the probability maps, in Figure 7.
The pseudothioisomaltoside 11 showed an energy prefer-
ence towards the gt conformer around the w torsion, in per-
fect agreement with the NMR observations. The predicted
minimum at y 908 (Figure 7a) was highly populated in solu-
tion, as indicated by the strong H5/H5’ NOE (Figure 4c).
The observation of H5-H6’ NOEs revealed the presence of
conformers in the anti-y region (y 1808), but in lower rela-
tive proportions than the previous minimum. For the N-
linked analogue 12, the situation was rather similar to that
with 11, showing two major conformations with y 1808 or
8532
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
from the corresponding adia-
batic maps, are shown in
Figure 8. Along with the proba-
bility isocontour levels, isodis-
tance curves for the H5–H4’
and H8a–H3’ distances are rep-
resented. All f,y values within
the area of these curves should
lead to the experimental obser-
vation of a cross peak in the
NOESY spectra correlating the
two protons defining the curve.
These two distances were moni-
tored in the graphs because
they
are
characteristically
within NOE observation dis-
tance for the two main regions
of the f,y map, namely the
synf–syny and synf–antiy re-
gions, respectively. The antif–
antiy region can be discarded
from the molecular modeling
calculations on energy grounds.
These two distances were then
considered for an analysis
based on “exclusive NOEs”.
The molecular models repre-
sentative of the minima ob-
tained by the Monte Carlo pro-
tocol are represented on the
probability maps in Figure 8.
For each solution, the expected
key interresidual NOEs are
also indicated. The predicted
distances for those NOEs that
could unambiguously be ob-
served and quantified in the
NMR spectra were calculated
Figure 7. Molecular modeling of the a
a) 11, b) 12, and c) 10 obtained from the corresponding adiabatic maps (see the Supporting Information) are
G
shown. Solid lines cover those f,y regions of the conformational space showing 20, 10, 1, and 0.1% of the from the output structures of
total conformer populations. The most populated solutions from Monte Carlo calculations are represented in
stick diagrams. The observed NOEs characterizing these conformations are also indicated. See the Supporting
Information for a color version of this figure.
the MC/SD and the MC/SD/
MM calculations, respectively,
and are included in Table 1. As
typically found for carbohy-
agreement with the experimental observations of the H8a/
H4’ and H7/H4’ NOE contacts.
Together, the combined protocol (NMR and molecular
drate structures, very few NOEs of sufficient quality were
available for the determination of the structures, neverthe-
less those collected in Table 1 were in rather good agree-
ment with the resulting molecular models from the Monte
Carlo calculations.
modeling) for the a
G
cated a rather similar conformational behavior of the three
molecules in solution, with only very subtle differences in
terms of major populations (y 908 for 11 and 12, and y 1808
for 10). In any case, all three compounds showed significant
flexibility around the w and y torsions of their pseudoglyco-
sidic linkages. The scenario is actually analogous to that en-
The combined analysis of NMR and modeling data pro-
vide important information on the differential conformation
behavior of the aglycone moiety in these a
G
pseudodisaccharides depending on the type of electronega-
tive atom connecting the two six-membered rings (S, N, or
O). For the N-linked maltoside mimetic 15, the accessible
conformations are restricted to a relatively well-defined
narrow area of the energy map, centred on syn-y values
with f<0 and y<0. For the O-linked derivative 13, the ad-
countered for the a
G
maltose.[38]
The probability maps of the pseudoglycosidic torsion
angles f and y for the maltoside mimics 13–15, obtained
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8533

C. Ortiz Mellet, J. Angulo et al.
jacent region within the syn-y
area (f>0, y>0) is also acces-
sible, reflecting some increased
flexibility around the pseudo-
glycosidic linkage in compari-
son to 15, a situation that is
rather similar to that encoun-
tered for the parent a
linked disaccharide
ACHTUNGTRENNUNG
ose.[38a,39] For the S-linked deriv-
ative 14, besides a syn-y region
similar to that accessible to the
O-derivative, the computational
procedure predicted
a lower
probability for a conformation
in the anti-y region (y around
1808). Indeed, for this molecule,
a NOE cross peak of significant
intensity could be observed be-
tween protons H8a and H3’.
This NOE, which is an exclu-
sive NOE contact for the synf–
antiy region, was absent from
the NOESY spectra of 13 and
15 (Figure 6). Indeed, the situa-
tion is analogous to that en-
countered for methyl 4-thio-a-
maltoside in comparison with
methyl a-maltoside.[40]
The differential behavior of
the aglycones as a function of
the linking heteroatom can be
interpreted in terms of the re-
spective covalent radii. Thus,
the much larger covalent radius
of sulfur in comparison to nitro-
gen and oxygen, leads to larger
distances between the two con-
stituent rings, reducing signifi-
cantly the steric barrier needed
to overcome any syny$
antiy transition and facilitating
adoption of these higher energy
Figure 8. Molecular modeling of the a
G
compounds a) 14, b) 15, and c) 13, obtained from the corresponding adiabatic maps (see the Supporting Infor-
mation) are shown. Solid lines cover those f,y regions of the conformational space showing 20, 10, 1, and
0.1% of the total conformer populations. Dashed lines show relevant interproton distances represented in
levels between 2.0 and 3.0 ꢄ. The most populated solutions from Monte Carlo calculations are represented in
stick diagrams. The observed NOEs characterizing these conformations are also indicated. See the Supporting
Information for a color version of this figure.
Table 1. Relevant interproton distances [ꢄ] defining the conformations
conformations, as experimentally observed for 14. Some dif-
ferences were also observed between the N- and O-linked
derivatives 13 and 15, even when the intersaccharide atoms
(O and N, respectively) have similar covalent radii. The
more restricted accessibility of the N-linked derivative 15 to
the conformational map might be a result of the local elec-
trostatic field due to the net positive charge that is present
in the protonated form of the pseudoglycosidic linkage,
which is the major species present at physiological pH.
of the a
N
15
MC/SD[a]
MC/SD+MM[b]
experimental[c]
MC/SD
13
14
MC/SD+MM
experimental
MC/SD
MC/SD+MM
experimental
Conformational behavior–glycosidase inhibitory activity re-
lationships: Inhibition of isomaltase and maltase by the new
linkage-spanning pseudodisaccharide inhibitors was of the
competitive type. Kinetics experiments modifying the incu-
[a] Monte Carlo Simulation and Stochastic Dynamics. [b] MC/SD fol-
lowed by Multiple Minimization. [c] From initial growth rates of 1D
NOESY build-up curves and applying the isolated spin pair approxima-
tion (ISPA).
8534
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
bation time revealed that they are fast-binding inhibitors,
suggesting that no deep conformational changes take place
upon formation of the enzyme–inhibitor complex. Most
probably, the enzyme, in its ground-state conformation,
binds the inhibitor in one of its accessible conformations in
solution. The available crystallographic and molecular mod-
eling information on the complexes of isomaltase[41] and
maltase[42] with their substrates/inhibitors suggests that most
of the enthalpic contribution to binding comes from interac-
tions involving the glycone moiety with amino acid residues
at the ꢀ1 site. In the case of isomaltoside analogues, the d-
glucopyranosyl aglycone unit probably retains substantial
flexibility about the three-covalent-bond segment connect-
ing the piperidine and the pyranose ring upon complexation,
thereby limiting the entropic penalty of binding.[43] The
almost identical inhibition constants of the isomaltoside
mimetics 10–12 against isomaltase (Table 2) is in agreement
might then be ascribed to a lower accessibility of the bound
conformation when replacing the intersaccharide oxygen by
ammonium. Moving from oxygen to sulfur in this series had
even more dramatic consequences. Thus, among the six
compounds studied, compound 14 was the only one that
showed no ability for inhibition of either isomaltase or malt-
ase. The inability of a-mannosidase to accommodate thio-
substituted glycosides analogous to the natural oligosaccha-
ride substrates has previously been noted.[44] It was speculat-
ed that increased rigidity might be at the origin of this ob-
servation. However, our data indicate that the most stable
unbound conformations for the O-linked homologue 13 are
also significantly populated in the case of 14 and, indeed,
support a higher conformational flexibility for the latter, in
line with other studies on thioligosaccharides.[45] It is likely
that the increased flexibility around the longer S-pseudogly-
cosidic linkage of 14 results in a much stronger entropic
penalty for the binding process in comparison with 13 or 15,
which have structurally better defined conformations in the
unbound state.
Table 2. Inhibition constants (Ki [mm]) for the pseudodisaccharides
10–15.[a,b]
Enzyme
12
10
11
15
13
14
b-glucosidase/galactosidase
(bovine liver)
195
n.i.
n.i.
190
n.i.
n.i.
Conclusion
b-glucosidase
ACHTUNGTRENNUNG
n.i.
17
n.i.
5.5
55
n.i.
30
375
108
83
n.i.
24
209
n.i.
n.i.
n.i.
n.i.
The present results support the potential of sp2-iminosugars
as building blocks for the construction of stable pseudo O-,
S-, and N-oligosaccharide mimetics. Their unique stereoelec-
tronic characteristics allow access to unprecedented a-linked
derivatives in which the acetal functional group characteris-
tic of natural glycosidic linkages is replaced by chemically
and enzymatically stable aminoacetal, aminothioacetal or
gem-diamine functionalities. The analogy between those
groups makes the approach very attractive for the design of
homologous series of linkage-spanning inhibitors of glycosi-
dases acting on oligosaccharides for structure–activity rela-
tionship studies. We have carried out this exercise for iso-
maltose and maltose pseudodisaccharides. The inhibitory ac-
tivities towards isomaltase and maltase, two enzymes that
bind isomaltose and maltose either as substrate or inhibitor,
were correlated with their conformational behavior in solu-
tion as determined by NMR and molecular modeling. The
three isomaltose mimics exhibited conformational properties
very similar to those of the parent disaccharide, which trans-
lated into very selective inhibition of the two enzymes with
small differences depending upon the linking heteroatom. In
stark contrast, the O-, N-, and S-linked maltose mimics sig-
nificantly differed in their conformational flexibility as well
as in their isomaltase and maltase inhibitory potency. Thus,
the O-pseudodisaccharide, which matched the conforma-
tional characteristics of maltose, was a potent and very se-
lective inhibitor of both enzymes. The inhibitory activity was
greatly reduced for the more rigid N-linked analogue, and
was fully abolished for the much more flexible thiomaltoside
homologue, thereby illustrating the utmost importance of in-
corporating conformational considerations in the design of
aglycone-selective glycosidase inhibitors.
ACHTUNGTRENNUNG
53
52
36
ACHTUNGTRENNUNG
trehalase
(pig kidney)
naringinase
(Penicillium decumbes)
549
742
n.i.
702
n.i.
187
n.i.
443
n.i.
160
[a] Inhibition was competitive in all cases. [b] No inhibition (n.i.) at 1 mm
concentration was observed for any of the new compounds against a-
mannosidase (Jack beans), b-mannosidase (Helix pomatia), a-galactosi-
dase (green coffee beans), b-galactosidase (E. coli), and amyloglucosidase
(Aspergillus niger).
with this assumption and means that no significant stabiliz-
ing effects occur due to electrostatic interactions involving
the amino group in the case of the N-linked derivative 12.
The inhibition of maltase evidenced some small differences
(less than one order of magnitude) with inhibition potencies
that decreased with the trend 10>12>11, which most likely
indicates a different mode of recognition (Table 2).
In the case of maltoside mimetics 13–15 (Table 2), having
a two-covalent-bond spacer between the (pseudo)monosac-
charide constituents, the scenario was totally different. Most
probably, in this case, enzyme binding involves a defined
conformation about the f and y dihedral angles. Although
binding of different conformers depending on the intergly-
cosidic heteroatom cannot be discarded, the penalty of
freezing a given conformation will be lower if it is already
highly populated in solution and will increase with molecu-
lar flexibility. The significantly higher inhibition potency of
the O-linked maltoside mimic 13 compared with the more
rigid N-linked analogue 15, especially in the case of maltase,
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8535

C. Ortiz Mellet, J. Angulo et al.
9.5 Hz, 1H; H-8), 4.89 (dd, J
(1’,2’)=3.5 Hz, J(2’,3’)=9.5 Hz, 1H; H-2’), 4.48 (t, J
9.0 Hz, 1H; H-1a), 4.25 (dd, J(1a,1b)=9.0 Hz, J(1b,8a)=7.5 Hz, 1H; H-
1b), 4.13 (dt, J(1a,8)=J(8,8a)=9.0 Hz, J(1b,8a)=7.5 Hz, 1H; H-8a), 3.92
(ddd, J(4’,5’)=9.5 Hz, J(5’,6b’)=4.5 Hz, J(5’,6a’)=3.0 Hz, 1H; H-5’), 3.69
(dd, (5’,6a’)=3.0 Hz, (6a’,6b’)=11.0 Hz, 1H; H-6a’), 3.66 (dd, J-
(5’,6b’)=4.5 Hz, J(6a’,6b’)=11.0 Hz, 1H; H-6b’), 3.45 (s, 3H; OMe),
ACHUTGTNRNEUG(N 5,6)=JACHTGNUTRENNUNG
Experimental Section
J
G
E
E
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
General Methods: Reagents and solvents were purchased from commer-
cial sources and used without further purification, except for dichloro-
methane, which was distilled under an Ar stream over CaH2. Optical ro-
tations were measured at 208C in 1 cm or 1-dm tubes with a Perkin–
Elmer 141 MC polarimeter. IR spectra were recorded with a JASCO
E
G
ACHTUNGTRENNUNG
R
G
ACHTUNGTRENNUNG
J
R
JACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
2.13–2.03 ppm (6 s, 18H; MeCO); 13C NMR (125.7 MHz, CDCl3): d=
170.2–169.6 (MeCO), 155.8 (CO), 96.6 (C-1’), 79.1 (C-5), 72.7 (C-8),
70.8–70.4 (C-2’, C-6), 70.2 (C-3’), 69.1 (C-7), 68.6 (C-4’), 67.9 (C-5’), 67.4
(C-6’), 66.8 (C-1), 55.4 (OMe), 51.6 (C-8a), 20.7–20.5 ppm (MeCO); MS
(ESI): m/z: 656.2 [M+Na]+; elemental analysis calcd (%) for
C26H35NO17: C 49.29, H 5.57, N 2.21; found: C 49.37, H 5.40, N 2.08.
1
FTIR-4100 (ATR) spectrometer. H and 13C NMR spectra were recorded
at 500 (125.7) and 300 (75.5) MHz with Bruker AVANCE DRX 500 and
300 spectrometers. 2D COSY, 1D TOCSY, and HMQC experiments
were used to assist NMR assignments. Thin-layer chromatography (TLC)
was carried out on aluminium sheets coated with Kieselgel 30 F245 (E.
Merck), with visualisation by UV light and by charring with 10% H2SO4
or 0.1% ninhydrin in EtOH. Column chromatography was carried out on
Silica Gel 60 (E. Merck, 230–400 mesh). Electrospray mass spectra
(ESIMS) were obtained with a Bruker Esquire6000 instrument. Elemen-
tal analyses were performed at the Instituto de Investigaciones Quꢁmicas
(Sevilla, Spain). The glycosidases a-glucosidase (from yeast), b-glucosi-
dase (from almonds), b-glucosidase/b-galactosidase (from bovine liver,
cytosolic), a-galactosidase (from green coffee beans), isomaltase (from
yeast), trehalase (from pig kidney), amyloglucosidase (from Aspergillus
niger), a-mannosidase (from jack bean), b-mannosidase (from Helix po-
matia), b-galactosidase (from E. coli) used in the inhibition studies, as
well as the corresponding o- and p-nitrophenyl glycoside substrates were
purchased from Sigma Chemical Co.
Oxa-3-oxoindolizidine (10): Compound 10 was obtained by conventional
de-O-acetylation of 19 (50 mg, 0.08 mmol) with catalytic NaOMe in
MeOH. Yield: 27.5 mg (91%); Rf =0.64 (6:3:1 MeCN/H2O/NH4OH);
[a]D =+68.7 (c=0.6 in H2O); 1H NMR (500 MHz, D2O): d=5.09 (d, J-
A
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
JACHTUNGTRENNUNG
A
E
N
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
E
N
ACHTUNGTRENNUNG
Hz, 1H; H-8), 3.43–3.35 (m, 1H; H-4’), 3.36 ppm (s, 3H; OMe);
13C NMR (125.7 MHz, D2O): d=158.4 (CO), 99.4 (C-1’), 81.3 (C-5), 73.5
(C-8), 73.3 (C-3’), 72.5 (C-7), 71.2 (C-2’), 71.0 (C-6), 70.0 (C-5’), 69.4 (C-
4’), 67.6 (C-1), 66.2 (C-6’), 55.2 (OMe), 53.1 ppm (C-8a); MS (ESI): m/z:
404.1 [M+Na]+; elemental analysis calcd (%) for C14H23NO11: C 44.10, H
6.08, N 3.67; found: C 43.97, H 6.02, N 3.42.
Synthesis: The starting materials, (5R,6R,7S,8R,8aR)-5,6,7,8-tetra-O-
acetyl-2-oxa-3-oxoindolizidine (16), methyl 2,3,4-tri-O-acetyl-a-d-gluco-
pyranoside (18),[27] methyl 2,3,6-tri-O-acetyl-a-d-glucopyranoside (20),[28]
2,3,4-tri-O-acetyl-6-thio-a-d-glucopyranoside (22),[29] methyl 6-amino-6-
deoxy-a-d-glucopyranoside (27),[32] and methyl 4-amino-4-deoxy-a-d-glu-
copyranoside (28)[33] were prepared as described previously. Methyl 2,3,6-
tri-O-benzoyl-4-thio-a-d-glucopyranoside (24)[30] was prepared from
methyl 2,3,6-tri-O-benzoyl-4-O-trifluoromethanesulfonyl-a-d-galactopyr-
anoside[30] by SN2 displacement of the triflate group by potassium thio-
acetate and final S-deacetylation with hydrazine.[47]
(5R,6R,7S,8R,8aR)-6,7,8-Triacetoxy-5-(methyl
2,3,6-tri-O-acetyl-a-d-
glucopyranosid-4-yl)-2-oxa-3-oxoindolizidine (21): BF3.OEt2 (15 mL,
0.10 mmol) was added to a stirred solution of 17 (70 mg, 0.21 mmol) and
20 (67 mg, 0.21 mmol) in anhydrous CH2Cl2 (2 mL) at 08C under a nitro-
gen atmosphere. The mixture was stirred for 30 min, diluted with CH2Cl2
(25 mL), washed with saturated aqueous NaHCO3 (5 mL), dried
(Na2SO4) and concentrated. The resulting residue was purified by flash
chromatography (1:2 EtOAc/petroleum ether). Yield: 65 mg (65%);
amorphous solid; Rf =0.62 (2:1 EtOAc/petroleum ether); [a]D =+80.8
(5R,6R,7S,8R,8aR)-6,7,8-Triacetoxy-5-fluoro-2-oxa-3-oxoindolizidine
(17): Compound 16 (560 mg, 1.50 mmol) and HF-pyridine (70%, 2.8 mL)
were placed in a polyethylene vessel cooled at ꢀ408C. The reaction mix-
ture was stirred at this temperature for 80 min, diluted with Et2O
(30 mL), washed with saturated aqueous KF (15 mL) and extracted with
Et2O (3ꢅ30 mL). The organic layer was washed with saturated aqueous
NaHCO3 (10 mL), dried (Na2SO4) and concentrated. The resulting resi-
due was purified by flash chromatography (1:2 EtOAc/petroleum ether).
Yield: 300 mg (60%); amorphous solid; Rf =0.77 (2:1 EtOAc/petroleum
ether); [a]D =+21.6 (c=0.9 in CHCl3); 1H NMR (500 MHz, CDCl3): d=
(c=1.0 in CHCl3); 1H NMR (500 MHz, CDCl3): d=5.73 (d, J
4.5 Hz, 1H; H-5), 5.54 (dd, J(2’,3’)=10.0 Hz, J(3’,4’)=9.0 Hz, 1H; H-3’),
5.43 (t, J(6,7)=J(7,8)=10.0 Hz, 1H; H-7), 4.89 (t, J(8,8a)=10.0 Hz, 1H;
H-8), 4.87 (d, J(1’,2’)=4.0 Hz, 1H; H-1’), 4.86 (dd, J(5,6)=4.5 Hz, J-
(6,7)=10.0 Hz, 1H; H-6), 4.80 (dd, J(1’,2’)=4.0 Hz, J(2’,3’)=10.0 Hz,
1H; H-2’), 4.43 (t, J(1a,1b)=J(1a,8a)=9.0 Hz, 1H; H-1a), 4.31 (br s, 1H;
H-6a’), 4.30 (br s, 1H; H-6b’), 4.22 (dd, J(1a,1b)=9.0 Hz, J(1b,8a)=
7.5 Hz, 1H; H-1b), 4.07 (dd, J(3’,4’)=9.0 Hz, J(4’,5’)=10.0 Hz, 1H; H-
ACHTUNGTRENNUNG(5,6)=
A
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
6.17 (dd, J
(7,8)=10.0 Hz, 1H; H-7), 5.00 (ddd, J(5,6)=3.5 Hz, J(6,7)=10.0 Hz, J-
(6,F)=14.0 Hz, 1H; H-6), 4.99 (t, J(7,8)=J(8,8a)=10.0 Hz, 1H; H-8),
4.51 (dd, (1a,1b)=9.0 Hz, (1a,8a)=8.0 Hz, 1H; H-1a), 4.32 (t, J-
(1a,1b)=J(1b,8a)=9.0 Hz, 1H; H-1b), 4.17–4.09 (m, 1H; H-8a), 2.15–
2.09 ppm (3 s, 9H; MeCO); 13C NMR (125.7 MHz, CDCl3): d=170.0–
169.5 (MeCO), 154.2 (CO), 87.5 (C-5, d, J(C5,F)=211.6 Hz), 72.0 (C-8),
ACHTUNGTRENNUNG(5,F)=52.5 Hz, JACHTUNTRGEG(NNUN 5,6)=3.5 Hz, 1H; H-5), 5.60 (t, JACHTNUGTRENN(UGN 6,7)=J-
A
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
4’), 3.96–3.87 (m, 2H; H-5’, H-8a), 3.42 (s, 3H; OMe), 2.16–2.03 ppm (6 s,
18H; MeCO); 13C NMR (125.7 MHz, CDCl3): d=171.4–169.6 (MeCO),
155.6 (CO), 96.8 (C-1’), 78.9 (C-5), 72.7 (C-3’), 72.5 (C-8), 72.2 (C-4’),
71.4 (C-2’), 69.4 (C-6), 68.4 (C-7), 67.2 (C-5’), 66.9 (C-1), 62.0 (C-6’), 55.5
(OMe), 52.1 (C-8a), 21.0–20.5 ppm (MeCO); MS (ESI): m/z: 656.2
[M+Na]+; elemental analysis calcd (%) for C26H35NO17: C 49.29; H 5.57;
N 2.21; found: C 49.19; H 5.42; N 2.10.
A
A
ACHTUNGTRENNUNG
J
G
JACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
69.8 (C-6, d, JACHTUNGTRENNUNG(C6,F)=24.8 Hz), 68.6 (C-7), 67.2 (C-1), 52.0 (C-8a), 20.4
(MeCO); MS (ESI): m/z: 356.1 [M+Na]+; elemental analysis calcd (%)
for C13H16NO8F: C 46.85, H 4.84, N 4.20; found: C 46.77, H 4.71, N 4.02.
(5R,6R,7S,8R,8aR)-6,7,8-Trihydroxy-5-(methyl
a-d-glucopyranosid-4-
yl)-2-oxa-3-oxoindolizidine (13): Compound 13 was obtained by conven-
tional de-O-acetylation of 21 (50 mg, 0.08 mmol) with catalytic NaOMe
in MeOH. Yield: 26.2 mg (87%); amorphous solid; Rf =0.57 (6:3:1
(5R,6R,7S,8R,8aR)-6,7,8-Tri-O-acetyl-5-(methyl 2,3,4-tri-O-acetyl-a-d-
glucopyranosid-6-yl)-2-oxa-3-oxoindolizidine (19): BF3.OEt2 (15 mL,
0.10 mmol) was added to a stirred solution of 17 (70 mg, 0.21 mmol) and
18 (67 mg, 0.21 mmol) in anhydrous CH2Cl2 (2 mL) at 08C under a nitro-
gen atmosphere. The mixture was stirred for 30 min, diluted with CH2Cl2
(25 mL), washed with saturated aqueous NaHCO3 (5 mL), dried
(Na2SO4) and concentrated. The resulting residue was purified by flash
chromatography (2:3!1:1 EtOAc/petroleum ether). Yield: 85 mg
(64%); amorphous solid; Rf =0.31 (1:1 EtOAc/petroleum ether); [a]D =
1
MeCN/H2O/NH4OH); [a]D =+99.9 (c=0.5 in H2O); H NMR (500 MHz,
D2O): d=5.68 (d, J
H-1’), 4.58 (t, J(1a,1b)=J
6.0 Hz, 1H; H-1b), 3.85 (t, J
(8a,8)=9.0 Hz, 1H; H-8a), 3.74 (dd,
2.0 Hz, 1H; H-6a’), 3.71–3.63 (m, 2H; H-5’, H-6b’), 3.59 (t, J
(7,8)=9.5 Hz, 1H; H-7), 3.58 (t, J(3’,4’)=J(4’,5’)=9.5 Hz, 1H; H-4’),
3.55 (dd, J(5,6)=4.5 Hz, J(6,7)=10.0 Hz, 1H; H-6), 3.50 (dd, J(1’,2’)=
4.0 Hz, J(2’,3’)=10.0 Hz, 1H; H-2’), 3.47 (t, J(7,8)=J(8a,8)=9.5 Hz, 1H;
H-8), 3.34 ppm (s, 3H; OMe); 13C NMR (125.7 MHz, D2O): d=158.2
A
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
J
A
J
A
JACHTUNGTRENNUNG
ACHTUNGTRENNUNG
+97.6 (c=1.0 in CHCl3); 1H NMR (500 MHz, CDCl3): d
J(6,7)=J(7,8)=9.5 Hz, 1H; H-7), 5.49 (t, J(2’,3’)=J(3’,4’)=9.5 Hz, 1H;
H-3’), 5.45 (d, J(5,6)=4.0 Hz, 1H; H-5), 5.09 (t, J(3’,4’)=J(4’,5’)=9.5 Hz,
(7,8)=J(8a,8)=
=
5.56 (t,
A
R
ACHTUNGTRENNUNG
A
E
U
A
R
ACHTUNGTRENNUNG
A
U
A
R
ACHTUNGTRENNUNG
1H; H-4’), 4.96 (d, J
G
N
ACHTUNGTRENNUNG
8536
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
(CO), 99.2 (C-1’), 82.0 (C-5), 74.4 (C-4’), 73.9 (C-3’), 73.3 (C-8), 72.3 (C-
7), 71.2 (C-2’), 70.9 (C-6), 69.7 (C-5’), 67.5 (C-1), 60.2 (C-6’), 55.0 (OMe),
53.5 ppm (C-8a); MS (ESI): m/z: 404.0 [M+Na]+; elemental analysis
calcd (%) for C14H23NO11: C 44.10, H 6.08, N 3.67; found: C 43.86, H
5.84, N 3.39.
[M+Na]+; elemental analysis calcd (%) for C41H41NO16S: C 58.92, H
4.94, N 1.68, S 3.84; found: C 58.95, H 4.78, N 1.44, S 3.60.
(5R,6R,7S,8R,8aR)-6,7,8-Trihydroxy-5-(methyl 4-thio-a-d-glucopyrano-
sid-4-yl)-2-oxa-3-oxoindolizidine (14): Compound 14 was obtained by
conventional de-O-acetylation of 25 (108 mg, 0.13 mmol) with catalytic
NaOMe in MeOH. Yield: 50 mg (98%); amorphous solid; Rf =0.67
(6:3:1 MeCN/H2O/NH4OH); [a]D =+118.7 (c=1.0 in H2O); 1H NMR
(5R,6R,7S,8R,8aR)-6,7,8-Triacetoxy-5-(methyl 2,3,4-tri-O-acetyl-6-thio-
a-d-glucopyranosid-6-yl)-2-oxa-3-oxoindolizidine
(23):
BF3·OEt2
(500 MHz, D2O): d=5.53 (d, J
4.0 Hz, 1H; H-1’), 4.60 (t, J(1a,1b)=J
(dd, J(1a,1b)=9.0 Hz, J(1b,8a)=5.5 Hz, 1H; H-1b), 4.00 (td, J
(8a,8)=9.0 Hz, (1b,8a)=5.5 Hz, 1H; H-8a), 3.93 (dd, (6a’,6b’)=
12.0 Hz, (5’,6a’)=4.0 Hz, 1H; H-6a’), 3.86 (dd, (5’,6b’)=2.0 Hz, J-
(6a’,6b’)=12.0 Hz, 1H; H-6b’), 3.81 (dd, J(5,6)=6.0 Hz, J(6,7)=9.5 Hz,
1H; H-6), 3.80–3.72 (m, 2H; H-3’, H-5’), 3.56 (t, J(6,7)=J(7,8)=9.5 Hz,
1H; H-7), 3.50 (dd, J(1’,2’)=4.0 Hz, J(2’,3’)=10.0 Hz, 1H; H-2’), 3.49 (t,
(7,8)=J(8a,8)=9.5 Hz, 1H; H-8), 3.34 (s, 3H; OMe), 2.77 ppm (t, J-
(3’,4’)=J
(4’,5’)=11.0 Hz, 1H; H-4’); 13C NMR (125.7 MHz, D2O): d=
157.6 (CO), 99.4 (C-1’), 73.3 (C-8), 73.2 (C-3’), 73.1 (C-7), 72.2 (C-2’),
70.4 (C-6), 70.3 (C-5’), 67.1 (C-1), 61.6 (C-5), 61.1 (C-6’), 55.0 (OMe),
53.2 (C-8a), 46.0 (C-4’); MS (ESI): m/z: 420.1 [M+Na]+; elemental analy-
sis calcd (%) for C14H23NO10S: C 42.31, H 5.83, N 3.52, S 8.07; found: C
41.99, H 5.57, N 3.21, S 7.73.
G
ACHTUNGTRENNUNG(1’,2’)=
(0.1 equiv) was added to a stirred solution of 17 (26 mg, 0.08 mmol) and
22 (40 mg, 0.12 mmol) in anhydrous CH2Cl2 (2 mL) at 08C under a nitro-
gen atmosphere. The mixture was stirred for 30 min, diluted with CH2Cl2
(25 mL), washed with saturated aqueous NaHCO3 (5 mL), dried
(Na2SO4) and concentrated. The resulting residue was purified by flash
chromatography (1:2 EtOAc/petroleum ether). Yield: 42 mg (83%); Rf =
0.45 (2:1 EtOAc/petroleum ether); [a]D =+90.6 (c=0.7 in CHCl3);
R
ACHTUNGTRENNUNG
G
E
ACHTUNGTRENNUNG
U
J
A
JACHTUNGTRENNUNG
J
E
JACHTUNGTRENNUNG
R
G
ACHTUNGTRENNUNG
R
ACHTUNGTRENNUNG
1H NMR (500 MHz, CDCl3): d=5.78 (d, J
(t, J(2’,3’)=J(3’,4’)=9.5 Hz, 1H; H-3’), 5.41 (t, J
1H; H-7), 5.09 (t, J(4’,5’)=9.5 Hz, 1H; H-4’), 4.96 (dd, J
(6,7)=10.0 Hz, 1H; H-6), 4.93 (t, J(8,8a)=10.0 Hz, 1H; H-8), 4.92 (d, J-
(1’,2’)=3.5 Hz, 1H; H-1’), 4.87 (dd, J(1’,2’)=3.5 Hz, J(2’,3’)=9.5 Hz, 1H;
H-2’), 4.55–4.49 (m, 1H; H-1a), 4.31–4.22 (m, 2H; H-1b, H-8a), 4.00
(ddd, J(4’,5’)=9.5 Hz, J(5’,6b’)=5.5 Hz, J(5’,6a’)=3.5 Hz, 1H; H-5’), 3.45
(s, 3H; OMe), 3.04 (dd, J(5’,6a’)=3.5 Hz, J(6a’,6b’)=13.5 Hz, 1H; H-
6a’), 2.66 (dd, J(5’,6b’)=5.5 Hz, J(6a’,6b’)=13.5 Hz, 1H; H-6b’), 2.13–
(5,6)=6.0 Hz, 1H; H-5), 5.48
(6,7)=J(7,8)=10.0 Hz,
(5,6)=4.5 Hz, J-
N
ACHTUNGTRENNUNG
J
G
G
ACHTUNGTRENNUNG
G
E
E
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
N
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
(5S,6S,7S,8R,8aR)-6,7,8-Trihydroxy-5-(methyl 6-amino-6-deoxy-a-d-glu-
copyranosid-6-yl)-2-oxa-3-oxoindolizidine (12): A solution of 26 (151 mg,
0.74 mmol) and 27 (143 mg, 0.74 mmol) in MeOH (1 mL) was stirred at
658C for 4 h under N2. The solvent was eliminated under reduced pres-
sure and the resulting residue was purified by flash chromatography (7:1
MeCN/H2O). Yield: 89 mg (59%); syrup; Rf =0.31 (5:1 MeCN/H2O);
[a]D =+112.5 (c=1.0 in H2O); 1H NMR (500 MHz, D2O): d=4.74 (d, J-
A
ACHTUNGTRENNUNG
2.02 ppm (6 s, 18H; MeCO); 13C NMR (125.7 MHz, CDCl3): d=170.2–
169.4 (MeCO), 155.7 (CO), 96.6 (C-1’), 72.7 (C-8), 70.8 (C-2’), 70.5 (C-
4’), 70.1 (C-6), 69.9 (C-3’), 69.6 (C-7), 68.0 (C-5’), 66.5 (C-1), 58.3 (C-5),
55.5 (OMe), 51.2 (C-8a), 31.2 (C-6’), 21.0–20.5 ppm (MeCO); MS (ESI):
m/z: 672.2 [M+Na]+; elemental analysis calcd (%) for C26H35NO16S: C
48.07, H 5.43, N 2.16, S 4.94; found: C 47.98, H 5.35, N 2.01, S 4.76.
A
ACHTUNGTRENNUNG
(5R,6R,7S,8R,8aR)-6,7,8-Trihydroxy-5-(methyl 6-thio-a-d-glucopyrano-
sid-6-yl)-2-oxa-3-oxoindolizidine (11): Compound 11 was obtained by
conventional de-O-acetylation of 23 (33 mg, 0.05 mmol) with catalytic
NaOMe in MeOH. Yield: 18 mg (91%); Rf =0.59 (6:3:1 MeCN/H2O/
NH4OH); [a]D =ꢀ16.0 (c=0.4 in H2O); 1H NMR (500 MHz, D2O): d=
G
R
ACHTUNGTRENNUNG
R
E
ACHTUNGTRENNUNG
J
U
J
A
JACHTUNGTRENNUNG
AHCTUNGTRENNUNG
J
E
G
E
ACHTUNGTRENNUNG
5.33 (d, J
1H; H-1a), 4.32 (dd, J
(td, J(1a,8)=J(8a,8)=9.0 Hz, J
(5,6)=5.5 Hz, J(6,7)=10.0 Hz, 1H; H-6), 3.67 (td, J
9.0 Hz, J(5’,6a’)=2.5 Hz, 1H; H-5’), 3.60–3.54 (m, 2H; H-3’, H-7), 3.51
(dd, 1H, J(2’,3’)=9.5 Hz, J(1’,2’)=3.5 Hz, 1H; H-2’), 3.47 (t, J(7,8)=J-
(8a,8)=9.0 Hz, 1H; H-8), 3.38 (s, 3H; OMe), 3.25 (t, J(3’,4’)=J(4’,5’)=
9.5 Hz, 1H; H-4’), 3.07 (dd, J(5’,6a’)=2.6 Hz, J(6a’,6b’)=14.0 Hz, 1H; H-
6a’), 2.64 ppm (dd, J(5’,6b’)=9.0 Hz, J(6a’,6b’)=14.0 Hz, 1H; H-6b’);
R
ACHUTGTNRENN(GU 1a,1b)=JAHCTUNGTREN(NNGU 1a,8a)=9.0 Hz,
A
ACHTUNGTRENNUNG
E
R
JACHTUNGTRENNUNG
JACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
U
A
E
N
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
(CO), 99.4 (C-1’), 73.6 (C-8), 73.0 (C-3’), 72.5 (C-7), 71.7 (C-4’), 71.3 (C-
2’), 70.2 (C-6), 68.9 (C-5’), 67.1 (C-1), 66.8 (C-5), 55.3 (OMe), 52.9 (C-
8a), 45.8 ppm (C-6’); MS (ESI): m/z: 403.1 [M+Na]+; elemental analysis
calcd (%) for C14H24N2O10: C 44.21, H 6.36, N 7.37; found: C 44.01, H
6.30, N 7.12.
A
R
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
13C NMR (125.7 MHz, D2O): d=157.9 (CO), 99.2 (C-1’), 73.5 (C-8),
73.2–72.9 (C-3’, C-7), 72.6 (C-4’), 71.3 (C-2’), 70.3 (C-6), 70.1 (C-5’), 67.3
(C-1), 61.3 (C-5), 55.0 (OMe), 52.8 (C-8a), 31.1 ppm (C-6’); MS (ESI): m/
z: 420.1 [M+Na]+; elemental analysis calcd (%) for C14H23NO10S: C
42.31, H 5.83, N 3.52, S 8.07; found: C 42.15, H 5.71, N 3.25, S 7.84.
(5S,6S,7S,8R,8aR)-6,7,8-Trihydroxy-5-(methyl 4-amino-4-deoxy-a-d-glu-
copyranosid-4-yl)-2-oxa-3-oxoindolizidine (15): A solution of 26 (70 mg,
0.34 mmol) and 28 (132 mg, 0.68 mmol, 2.0 equiv) in MeOH (1 mL) was
stirred at 658C for 24 h under N2 (TLC monitoring; 8:1 MeCN/H2O).
The solvent was removed under reduced pressure and the resulting resi-
due was acetylated by treatment with Ac2O/pyridine (1:1, 2 mL) at RT
for 3 h, then subjected to flash chromatography (3:1 EtOAc/petroleum
ether). Conventional deacetylation of the peracetylated residue with cat-
alytic NaOMe in MeOH afforded 15. Yield: 64 mg (52%); amorphous
(5R,6R,7S,8R,8aR)-6,7,8-Triacetoxy-5-(methyl
2,3,6-tri-O-benzoyl-4-
thio-a-d-glucopyranosid-4-yl)-2-oxa-3-oxoindolizidine (25): The a-S-
linked pseudodissacharide 25 was obtained following the procedure de-
scribed above using the fluoro-derivative 17 (73 mg, 0.22 mmol) and 24
(176 mg, 0.34 mmol). Column chromatography (1:2!1:1 EtOAc/petro-
leum ether). Yield: 115 mg (61%); amorphous solid; Rf =0.66 (1:1
EtOAc/petroleum ether); [a]D =+159.3 (c=1.0 in CHCl3); 1H NMR
1
solid; Rf =0.27 (6:1 MeCN/H2O); [a]D =ꢀ21.3 (c=0.5 in H2O). H NMR
(500 MHz, D2O): d=5.08 (d, JACTHNGUTREN(NUGN 5,6)=5.0 Hz, 1H; H-5), 4.51 (t, JACHTUNGTRENNUNG
JACHTUNGTNER(NNUG 1a,8a)=9.0 Hz, 1H; H-1a), 4.30 (dd, JACHTUTGNRNENUG(1a,1b)=9.0 Hz, JACHTUNGTRENNUNG
(500 MHz, CDCl3): d=8.20–7.35 (m, 15H; Ph), 6.10 (t, J
10.0 Hz, 1H; H-3’), 5.94 (d, J(5,6)=6.0 Hz, 1H; H-5), 5.25 (t, J
(7,8)=9.5 Hz, 1H; H-7), 5.24 (d, J(1’,2’)=3.5 Hz, 1H; H-1’), 5.13 (dd, J-
(1’,2’)=3.5 Hz, J(2’,3’)=10.0 Hz, 1H; H-2’), 4.87 (t, J(7,8)=J(8,8a)=
9.5 Hz, 1H; H-8), 4.83 (dd, J(5,6)=6.0 Hz, J(6,7)=9.5 Hz, 1H; H-6),
4.72 (d, J(5,6’)=3.0 Hz, 2H; H-6’), 4.35 (t, J(1a,1b)=J(1a,8a)=8.5 Hz,
1H; H-1a), 4.22–4.10 (m, 3H; H-1b, H-5’, H-8a), 3.49 (t, J(3’,4’)=J-
(4’,5’)=10.0 Hz, 1H; H-4’), 3.47 (s, 3H; OMe), 2.06–1.80 ppm (3 s, 9H;
G
ACHTUNGTRENN(UNG 3’,4’)=
R
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
E
ACHTUNGTRENNUNG
U
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
E
G
G
ACHTUNGTRENNUNG
R
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
158.5 (CO), 99.3 (C-1’), 74.7 (C-3’), 73.4 (C-8), 72.0 (C-2’), 72.2–70.5 (C-
5’, C-6, C-7), 67.3 (C-5), 66.9 (C-1), 61.0 (C-6’), 55.9 (C-4’), 54.9 (OMe),
53.2 ppm (C-8a); MS (ESI): m/z: 403.1 [M+Na]+; elemental analysis
calcd for C14H24N2O10 (380.1) C 44.21, H 6.36, N 7.37; found: C 43.94, H
6.09, N 7.18.
E
N
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
MeCO); 13C NMR (125.7 MHz, CDCl3): d=169.9–165.6 (CO), 155.4
(CO), 133.6–128.4 (Ph), 97.1 (C-1’), 73.3 (C-2’), 72.5 (C-3’), 72.0 (C-8),
69.3 (C-7), 68.9 (C-6), 68.0 (C-5’), 66.0 (C-1), 63.8 (C-6’), 58.9 (C-5), 55.7
(OMe), 51.3 (C-8a), 44.5 (C-4’), 20.5–20.0 (MeCO); MS (ESI): m/z: 858.2
Glycosidase inhibition assays: Inhibitory potencies were determined
spectrophotometrically by measuring the residual hydrolytic activities of
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8537

C. Ortiz Mellet, J. Angulo et al.
the glycosidases against the respective o- (for b-glucosidase/b-galactosi-
dase from bovine liver and b-galactosidase from E. coli) or p-nitrophenyl
a- or b-d-glycopyranoside, or a,a’-trehalose (for trehalase), in the pres-
ence of the corresponding glycomimetic derivative. Each assay was per-
formed in phosphate buffer at the optimal pH for each enzyme. The Km
values for the different glycosidases used in the tests and the correspond-
ing working pH values are listed herein: a-glucosidase (yeast): Km =
0.35 mm (pH 6.8); isomaltase (yeast): Km =1.0 mm (pH 6.8); b-glucosidase
(almonds): Km =3.5 mm (pH 7.3); b-glucosidase/b-galactosidase (bovine
liver): Km =2.0 mm (pH 7.3); b-galactosidase (E. coli): Km =0.12 mm
(pH 7.3); a-galactosidase (coffee beans): Km =2.0 mm (pH 6.8); trehalase
(pig kidney): Km =4.0 mm (pH 6.2); amyloglucosidase (Aspergillus niger):
Km =3.0 mm (pH 5.5); b-mannosidase (Helix pomatia): Km =0.6 mm
(pH 5.5); a-mannosidase (jack bean): Km =2.0 mm (pH 5.5); naringinase
(Penicillium decumbens, b-glucosidase/b-rhamnosidase activity): Km =
0.6 mm (pH 6.8). The reactions were initiated by addition of enzyme to
a solution of the substrate in the absence or presence of various concen-
trations of inhibitor. After the mixture was incubated for 10–30 min at 37
or 558C, the reaction was quenched by addition of either 1 m Na2CO3 or
a solution of Glc-Trinder (Sigma, for trehalase). The absorbance of the
resulting mixture was determined at 405 or 492 nm. The Ki value and
enzyme inhibition mode were determined from the slope of the Line-
weaver–Burk plots and double reciprocal analysis.
terprotonic distances. On this set, a further multiple minimization step
was carried out, the results were clustered by comparing heavy atoms
with a cutoff of 0.5 A, 100 structures were saved within an energy
window of 20 kJmolꢀ1, and interprotonic distances were also analyzed.
Acknowledgements
The Spanish Ministerio de Ciencia
e Innovaciꢆn (contract numbers
CTQ2006–15515-CO2–01, CTQ2009–14551-C02–01, SAF2010–15670 and
CTQ2010–15848), the Fondo Europeo de Desarrollo Regional
(FEDER), the Fondo Social Europeo (FSE), and the Fundaciꢆn Ramꢆn
Areces and the Junta de Andalucꢁa, are thanked for funding. We also
thank the CITIUS (Universidad de Sevilla) for technical support. J. A.
acknowledges financial support from the Spanish Ministerio de Ciencia
e Innovaciꢆn through the Ramon y Cajal program.
[1] a) K. M. Robinson, M. E. Begovic, B. L. Rhinehart, E. W. Heineke,
J.-B. Ducep, P. R. Kastner, F. N. Marshall, C. Danzin, Diabetes 1991,
40, 825–830; b) A. J. Krentz, C. J. Bailey, Drugs 2005, 65, 385–411.
[2] a) J. Q. Fan, Biol. Chem. 2008, 389, 1–11; b) T. D. Butters, Expert
Opin. Pharmacother. 2007, 8, 427–435; c) Z. Yu, A. R. Sawkar, L. J.
Whalen, C. H. Wong, J. W. Kelly, J. Med. Chem. 2007, 50, 94–100;
d) J. Matsuda, O. Suzuki, A. Oshima, Y. Yamamoto, A. Noguchi, K.
Takimoto, M. Itoh, Y. Matsuzaki, Y. Yasuda, S. Ogawa, Y. Sakata,
E. Nanba, K. Higaki, Y. Ogawa, L. Tominaga, K. Ohno, H. Iwasaki,
H. Watanabe, R. O. Brady, Y. Suzuki, Proc. Natl. Acad. Sci. USA
2003, 100, 15912–15917; e) T. D. Butters, H. R. Mellor, K. Narita,
R. A. Dwek, F. M. Platt, Philos. Trans. R. Soc. London Ser. B 2003,
358, 927–945; f) J. Fan, S. Ishii, N. Asano, Y. Suzuki, Nat. Med.
1999, 5, 112–115; g) J. M. Benito, J. M. Garcꢁa Fernꢀndez, C. Ortiz
Mellet, Expert Opin. Ther. Pat. 2011, 21, 885–903.
[3] a) D. Durantel, C. Alotte, F. Zoulim, Curr. Opin. Investig. Drugs
2007, 8, 125–129; b) P. Greimel, J. Spreitz, A. E. Stꢇtz, T. M. Wrod-
nigg, Curr. Top. Med. Chem. 2003, 3, 513–523.
[4] a) T. M. Wrodnigg, A. J. Steiner, B. J. Ueberbacher, Anti-Cancer
Agents Med. Chem. 2008, 8, 77–85; b) Y. Nishimura Iminosugar-
based antitumoural agents. In Iminosugars: From Synthesis to Thera-
peutic Applications (Eds.: P. Compain, O. R. Martin) Wiley-VCH:
Weinheim, Germany, 2007, p. 269–294.
NMR spectroscopy: For the structural analysis, NMR spectra were re-
corded at 25–358C, for 10, 11, 12, 13, 14 and 15, in D2O with a Bruker
DRX 500 MHz NMR spectrometer. The highest temperature (358C) was
necessary for NOE experiments of 10, 11 and 12, to bring their hydrody-
namics behavior towards the extreme narrowing limit, to try to enhance
the NOEs. For sample preparation, all the compounds were lyophilized
twice with 99% D2O and once finally with 99.99% D2O (Sigma–
Aldrich). Final concentrations were approximately 1 mm in 600 mL vol-
1
umes. For H NMR titration experiments on 12, the pH in the NMR tube
was varied from 7.0 to 12.5, and 1% of acetone was used to reference
the chemical shifts. Selective 1D NOESY experiments were carried out
by using the double pulse field gradient echo sequence with different
mixing times varying from 0.2 to 2.0 s. 2D NOESY experiments were per-
formed with several mixing times between 0.2 and 1.0 s. Interproton dis-
tances were obtained from selective experiments, increasing the linearity
of NOE growth curves by normalization (PANIC methodology[46]), calcu-
lating the initial slopes by linear regression, and applying the isolated
spin pair approach (ISPA) using the rigid H5–H6 distance as a reference.
Computational methods: Calculations were performed by using Macro-
Model 9.8 and the OPLS-2005 force field. Bulk water solvation was sim-
ulated by using the generalized Born GB/SA continuum solvent model,
which treats the solvent as a fully equilibrated analytical continuum start-
ing near the van der Waals surface of the solute.[47] A dielectric constant
of 1 was used for each molecule, the sets of charges were from the force
field, and extended cutoffs of 8.0, 20.0, and 4.0 ꢄ for the van der Waals,
electrostatics, and H-bond terms were used, respectively. For each mole-
cule, 12 initial conformers were constructed (three possible conformers
around w torsion, O6-C6-C5-O5, that is, the conformers tg, gt and gg;
and four “canonical” distributions of the hydroxyl hydrogen atoms, com-
bining “clockwise” or “reverse clockwise” orientations: cc, cr, rc, or rr).
Then, a systematic torsion-driven conformational search was carried out
on torsions f (H5-C5-O4’-C4’-H4’) and y (C5-O4’-C4’-H4’), covering
from ꢀ1808 to +1808 in 208 increments. Conjugated gradients methodol-
ogy was used for energy minimization with a maximum of 1000 iterations
or a threshold for gradient convergence of 0.01 kJmolꢀ1 ꢄꢀ1. From these
12 relaxed maps, each adiabatic map was constructed by collecting the
lowest energy conformer for each grid point. Each minimum from the
adiabatic maps was subjected to a Monte Carlo conformational search
(torsional sampling MCMM), and the resulting structures within
20 kJmolꢀ1 from the absolute minimum were saved. From this set, all
molecules within 3 kcalmolꢀ1 from the minimum were subjected to a mul-
tiple minimization and the results were clustered by comparing heavy
atoms with a cutoff of 0.5 A. The global minimum was then subjected to
stochastic dynamics simulations MC/SD at 298 K, 1.5 fs time step for in-
tegration, and a total length of 2.5 ns, with an SD-to-MC ratio of 1. From
this dynamic simulation, 5000 structures were saved and analyzed for in-
[5] G. Horne, F. C. Wilson, J. Tinsley, D. H. Willians, R, Storer, Drug
Discov. Today 2011, 16, 107–118.
[6] a) J. D. McCarter, W. Yeung, J. Chow, D. Dolphin, S. G. Withers, J.
Am. Chem. Soc. 1997, 119, 5792–5797; b) W. Hakamata, M. Muroi,
K. Kadokura, T. Nishio, T. Oku, A. Kimura, S. Chibac, A. Takatsuki,
Bioorg. Med. Chem. Lett. 2005, 15, 1489–1492; c) C.-W. Ho, Y.-N.
Lin, C.-F. Chang, S.-T. Li, Y.-T. Wu, C.-Y. Wu, C.-F. Chang, S.-W.
Liu, Y.-K. Li, C.-H. Lin, Biochemistry 2006, 45, 5695–5702.
[7] a) H. Driguez, Top. Curr. Chem. 1997, 187, 85–116; b) Z. J. Witczak,
J. M. Culhane, Appl. Microbiol. Biotechnol. 2005, 69, 237–244; c) W.
Zou, Curr. Top. Med. Chem. 2005, 5, 1363–1391.
[8] a) J. F. Espinosa, F. J. CaÇada, J. L. Asensio, M. Martꢁn-Pastor, H.
Dietrich, M. Martꢁn-Lomas, R. R. Schmidt, J. Jimꢈnez-Barbero, J.
Am. Chem. Soc. 1996, 118, 10862–10871; b) R. Ravishankar, A. Sur-
olia, M. Vijayan, S. Lim, Y. Kishi, J. Am. Chem. Soc. 1998, 120,
11297–11303; c) J. F. Espinosa, M. Bruix, O. Jarreton, T. Skrydstrup,
J. M. Beau, J. Jimꢈnez-Barbero, Chem. Eur. J. 1999, 5, 442–448;
d) J. Jimꢈnez-Barbero, J. F. Espinosa, J. L. Asensio, F. J. CaÇada, A.
Poveda, Adv. Carbohydr. Chem. Biochem. 2000, 56, 235–284.
[9] A. Garcꢁa-Herrero, E. Montero, J. L. MuÇoz, J. F. Espinosa, A.
Viꢀn, J. L. Garcꢁa, J. L. Asensio, F. J. CaÇada, J. Jimꢈnez-Barbero, J.
Am. Chem. Soc. 2002, 124, 4804–4810.
[10] D. A. Kuntz, C. A. Tarling, S. G. Withers, D. R. Rose, Biochemistry
2008, 47, 10058–10068.
8538
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 8527 – 8539

a-Linked Disaccharide Mimics
FULL PAPER
[11] a) C. R. Johnson, A. Golebiowski, H. Sundram, M. W. Miller, R. L.
Dwaihy, Tetrahedron Lett. 1994, 35, 653–654; b) H. Paulsen, K.
Todt, Adv. Carbohydr. Chem. 1968, 23, 115–232.
[12] a) J. S. Andrews, T. Weimar, T. P. Frandsen, B. Svensson, B. M.
Pinto, J. Am. Chem. Soc. 1995, 117, 10799–10804; b) L. M. Kavle-
kar, D. A. Kuntz, X. Wen, B. D. Johnston, B. Svensson, D. R. Rose,
B. M. Pinto, Tetrahedron: Asymmetry 2005, 16, 1035–1046.
[13] a) I. Cumpstey, C. Ramstadius, K. E. Borbas, D. S. Alonzi, T. D. But-
ters, Bioorg. Med. Chem. 2011, 19, 5219–5223; b) T. K. M. Shing,
H. M. Cheng, J. Org. Chem. 2010, 75, 3522–3525; c) T. K. M. Shing,
H. M. Cheng, W. F. Wong, C. S. K. Kwong, J. Li, C. B. S. Lau, P. S.
Leung, C. H. K. Cheng, Org. Lett. 2008, 10, 3145–3149.
[14] a) I. Cumpstey, J. Frigell, E. Pershagen, T. Akhtar, E. Moreno-Clavi-
jo, I. Robina, D. S. Alonzi, T. Butters, Beilstein J. Org. Chem. 2011,
7, 1115–1123; b) T. Akhtar, I. Cumpstey, Tetrahedron Lett. 2007, 48,
8673–8677.
[15] a) P. Compain, Iminosugar C-glycosides: synthesis and biological ac-
tivity. In Iminosugars: From synthesis to therapeutic applications
(Eds.: P. Compain, O. R. Martin), John Wiley & Sons, Chichester,
UK, 2007, pp. 63–86; b) P. Vogel, S. Gerber-Lemaire, L. Juillerat-
Jeanneret. Imino-C-disaccharides and analogues: synthesis and bio-
logical activity In Iminosugars: From synthesis to therapeutic applica-
tions (Eds.: P. Compain, O. R. Martin), John Wiley & Sons, Chiches-
ter, UK, 2007, pp. 87–130.
[16] J. L. Asensio, F. J. CaÇada, A. Garcꢁa-Herrero, M. T. Murillo, A.
Fernꢀndez-Mayoralas, B. A. Johns, J. Kozak, Z. Zhu, C. R. Johnson,
J. Jimꢈnez-Barbero, J. Am. Chem. Soc. 1999, 121, 11318–11329.
[17] For selected recent contributions on the synthesis of sp2-iminosugars,
see: a) M. Aguilar-Moncayo, P. Dꢁaz-Pꢈrez, J. M. Garcꢁa Fernꢀndez,
C. Ortiz Mellet, M. I. Garcꢁa-Moreno, Tetrahedron 2012, 68, 681–
689; b) M. Aguilar-Moncayo, M. I. Garcꢁa-Moreno, A. Trapero, M.
Egido-Gabꢀs, A. Llebaria, J. M. Garcꢁa Fernꢀndez, C. Ortiz Mellet,
Org. Biomol. Chem. 2011, 9, 3698–3713; c) M. Aguilar-Moncayo,
M. I. Garcꢁa-Moreno, A. E. Stꢇtz, J. M. Garcꢁa Fernꢀndez, T. M.
Wrodnigg, C. Ortiz Mellet, Bioorg. Med. Chem. 2010, 18, 7439–
7445; d) M. Aguilar-Moncayo, T. M. Gloster, J. P. Turkenburg, M. I.
Garcꢁa-Moreno, C. Ortiz Mellet, G. J. Davies J. M. Garcꢁa Fernꢀn-
dez, Org. Biomol. Chem. 2009, 7, 2738–2747; e) M. Aguilar-Mon-
cayo, M. I. Garcꢁa-Moreno, C. Ortiz Mellet, J. M. Garcꢁa Fernꢀndez,
J. Org. Chem. 2009, 74, 3595–3598.
ez, M. I. Garcꢁa-Moreno, C. Ortiz Mellet, J. Fuentes, J. C. Dꢁaz Arri-
bas, F. J. CaÇada J. M. Garcꢁa Fernꢀndez, J. Org. Chem. 2000, 65,
136–143; c) J. L. Jimꢈnez Blanco, V. M. Dꢁaz Pꢈrez, C. Ortiz Mellet,
F. Fuentes, J. M. Garcꢁa Fernꢀndez, J. C. Dꢁaz Arribas, F. J. CaÇada,
Chem. Commun. 1997, 1969–1970.
[21] B. L. Cantarel, P. M. Coutinho, C. Rancurel, T. Bernard, V. Lom-
bard, B. Henrissat, Nucleic Acid Res. 2009, 37, D233-D238.
[22] N. A. Khan, N. R. Eaton, Biochim. Biophys. Acta 1967, 146, 173–
180.
[23] R. B. Needleman, H. J. Federoff, T. R. Eccleshall, B. Buchferer, J.
Marmur, Biochemistry 1978, 17, 4657–4661.
[24] E. M. Sꢀnchez-Fernꢀndez, R. Rꢁsquez-Cuadro, M. Aguilar-Moncayo,
M. I. Garcꢁa-Moreno, C. Ortiz Mellet, J. M. Garcꢁa Fernandez, Org.
Lett. 2009, 11, 3306–3309.
[25] T. Fuchss, R. R. Schmidt, J. Carbohydr. Chem. 2000, 19, 677–691.
[26] a) S. Shoda Glycoside synthesis from glycosyl halides: glycosyl fluo-
rides. In Handbook of chemical glycosylation: Advances in stereose-
lectivity and therapeutic relevance (Ed.: A. V. Demchenko), Wiley-
VCH, Weinheim, 2008, pp. 29–58; b) J. Jꢇnnemann, J. Thiem, C.
Pedersen, Carbohydr. Res. 1993, 249, 91–94.
[27] a) D. Ellis, S. E. Norman, H. M. I. Osborn, Tetrahedron 2008, 64,
2832–2854; b) M J. Kiefel, B. Beisner, S. Bennet, I. D. Holmes, M.
Itzstein, J. Med. Chem. 1996, 39, 1314–1320.
[28] K. Sato, K. Sakai, M. Kojima, S. Akai, Tetrahedron Lett. 2007, 48,
4423–4425.
[29] D. Crich, H. Li, J. Org. Chem. 2000, 65, 801–805.
[30] L. A. Reed III, L. Goodman, Carbohydr. Res. 1981, 94, 91–100.
[31] T. M. Wrodnigg, B. Eder, Top. Curr. Chem. 2001, 215, 115–152.
[32] J. M. Garcꢁa Fernꢀndez, C. Ortiz Mellet, J. Fuentes, J. Org. Chem.
1993, 58, 5192–5199.
[33] E. J. Reist, R. R. Spencer, D. F. Calkins, B. R. Baker, L. Goodman,
J. Org. Chem. 1965, 30, 2312–2317.
[34] K. Kondo, A. Hayamitsu, E. Shitara, F. Kojima, Y. Nishimura,
Bioorg. Med. Chem. Lett. 2001, 11, 1091–1095.
[35] a) S. Chandrasekhar, ARKIVOC 2005, 37–66; b) V. G. Box, J. Mol.
Struct. 2000, 522, 145–164.
[36] A. Imberty, S. Pꢈrez, Chem. Rev. 2000, 100, 4567–4588.
[37] Y. Nishida, H. Hori, H. Ohrui, H. Meguro, J. Carbohydr. Chem.
1988, 7, 239–250.
[38] a) M. K. Dowd, P. J. Reilly, A. D. French, Biopolymers 1994, 34,
625–638; b) C. S. Pereira, D. Kony, R. Baron, M. Mꢇller, W. F. van
Gunsteren, P. H. Hꢇnenberger, Biophys. J. 2006, 90, 4337–4344.
[39] A. D. French, M. K. Dowd, J. Mol. Struct. 1983, 91-111, 183–201.
[40] K. Bock, J. Ø. Duus, S. Refn, Carbohydr. Res. 1994, 253, 51–67.
[41] K. Yamamoto, H. Miyake, M. Kusunoki, S. Osaki, FEBS J. 2010,
277, 4205–4214.
[42] a) P. M. Coutinho, M. K. Dowd, P. J. Reilly, Carbohydr. Res. 1997,
297, 309–324; b) K. Bharatham, N. Bharatham, K. H. Park, K. W.
Lee, J. Mol. Graphics Modell. 2008, 26, 1202–1212.
[43] T. P. Frandsen, M. M. Palcic, B. Svensson, Eur. J. Biochem. 2002,
269, 728–734.
[44] a) W. Zhong, D. A. Kuntz, B. Ember, H. Singh, K. W. Moremen,
D. R. Rose, G.-J. Boons, J. Am. Chem. Soc. 2008, 130, 8975–8983.
[45] G. A. Venter, R. P. Mattews, K. Naidoo, Mol. Simul. 2008, 34, 391–
401.
[18] For selected recent contributions on the therapeutic potential of sp2-
iminosugars, see: a) I. Silman, Y. Shaaltiel, D. Aviezer, J. L. Suss-
man, A. H. Futerman, C. Ortiz Mellet, Org. Biomol. Chem. 2011, 9,
4160–4167; b) Z. Luan, K. Higaki, M. Aguilar-Moncayo, L. Li, H.
Ninomiya, E. Nanba, K. Ohno, M. I. Garꢁa-Moreno, C. Ortiz Mellet,
J. M. Garcꢁa Fernꢀndez, Y. Suzuki, ChemBioChem 2010, 11, 2453–
2464; c) B. Brumshtein, M. Aguilar-Moncayo, M. I. Garcꢁa-Moreno,
C. Ortiz Mellet, J. M. Garcꢁa Fernꢀndez, I. Silman, Y. Shaaltiel, D.
Aviezer, J. L. Sussman, A. H. Futerman, ChemBioChem 2009, 10,
1480–1485; d) Z. Luan, K. Higaki, M. Aguilar-Moncayo, H. Nino-
miya, K. Ohno, M. I. Garcꢁa-Moreno, C. Ortiz Mellet, J. M. Garcꢁa
Fernꢀndez, Y. Suzuki, ChemBioChem 2009, 10, 2780–2792; e) M.
Aguilar-Moncayo, T. Takai, K. Higaki, T. Mena-Barragꢀn, Y.
Hirano, K. Yura, L. Li, Y. Yu, H. Ninomiya, M. I. Garcꢁa Moreno,
S. Ishii, Y. Sakakibara, K. Ohno, E. Nanba, C. Ortiz Mellet, J. M.
Garcꢁa Fernꢀndez, Y. Suzuki, Chem. Commun., DOI: 10.1039/
c2cc32065g.
[46] H. Hu, K. Krishnamurthy, J. Magn. Reson. 2006, 182, 173–177.
[47] W. C. Still, A. Tempczyk, R. C. Hawley, T. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127–6129.
[19] E. M. Sꢀnchez-Fernꢀndez, R Rꢁsquez-Cuadro, M. Chasseraud, A.
Ahidouch, C. Ortiz Mellet, H. Ouadid-Ahidouch, J. M. Garcꢁa Fer-
nandez, Chem. Commun. 2010, 46, 5328–5330.
[20] a) P. Dꢁaz Pꢈrez, M. I. Garcꢁa-Moreno, C. Ortiz Mellet, J. M. Garcꢁa
Fernꢀndez, Eur. J. Org. Chem. 2005, 2903–2913; b) V. M. Dꢁaz Pꢈr-
Received: January 25, 2012
Revised: February 28, 2012
Published online: June 4, 2012
Chem. Eur. J. 2012, 18, 8527 – 8539
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8539